Plato's

Matlumatical Imagination
by ROBERT s. BRUMBAUGH

The Mathematical Passages

in the Dialogues
and Their Interpretation

INDIANA UNIVERSITY PRESS

Bloomington
 Preface

PROFESSORS Richard P. McKeon, Raymond Klibansky, Newtoll

P. Stallknecht, and Guido Calogero have read this study in
manuscript, and I am most grateful to them for their sugges-
tions and comments. I hope that readers acquainted with Pro-
fessor McKeon's approach to Greek philosophy will find in this
study by one of his students and admirers a spirit and method
with which they are familiar. Over a period of years, Professor
Stallknecht has been unfailingly generous of his time and as-
sistance. His insights into the aesthetic dimension of my prob.
lems, and his distinctive analysis of the nature and history of
imagination, have added immeasurably to my own apprecia-
tion of the subject treated here. Professor Klibansky has con-
tributed a number of incisive and illuminating comments on
the Pythagorean and Platonic traditions and on some of the
textual problems crucial in mathematical passages in Plato.
Professor Calogero's general evaluation and specific comments
were most helpful. The responsibility for any shortcomings of
the text remains my own; but because this study departs at many
points from any previous work in the Platonic tradition, it is a
responsibility I should have hesitated to assume without the
encouragement and help of these scholars.
I wish to thank Professor Edward Seeber, of Indiana Univer-
sity, for his help in the preliminary editing of copy. I am par-
ticularly grateful to Mr. Walter Albee, of the Indiana Univer-
sity Press, for copy-editing the manuscript and seeing it through
the various stages of design and production and for suggesting
many improvements in the design of the figures and in the co-
herence of the text of this study. Mr. Robert Williams and the
Yale Cartographic Laboratory have been most helpful in pre-
paring drawings for the figures. Professor Ronald B. Levinson,
vii
viii Preface

of the University of Maine, has helped me to clarify and im-

prove Section 6 of Chapter III; I am sorry that his In Defense
of Plato appeared too late to be cited where it is clearly rele-
vant. The possibility of constructing a mirror which would give
a non-reversed image as the one in Figure 88 does, a possibility
that has been seriously doubted, was proven by Dr. Harrison
Pemberton and Mr. George Starr, of Yale University. Mr.
Thomas B. Rosenmeyer's criticism of my views on the Critias
has also been helpful, not only in Chapter II, where his article
is explicitly cited, but in the whole study.
My wife has been a model of patience, help, and industry;
she has copied figures, turned thickets of words into English
sentences, and listened intelligently to sporadic speeches on
Platonic mathematics. I also want to thank Meda Z. Steele for
her help in copying figtlres, and my son Robert, starting his
editorial career at the age of five, for help with various opera-
tions and problems.
Parts of this study which have already appeared as notes and
papers are included by permission of publishers and journals.
These are "Note on the Numbers in Plato's Critias," Classical
Philology, XLIII (copyright 1948, Univ. of Chicago), 40-42;
"Note on Plato Republic ix. 587D," ibid.) XLIV (copyright
1949, Univ. of Chicago), 197-99; "Colors of the Hemispheres in
Plato's Myth of Er (Republic 616E)," ibid.) XLVI (copyright
1951, Univ. of Chicago), 173-76; "Plato Republic 616E: The
Final 'Law of Nines,' " ibid.) XLIX (copyright 1954, Univ. of
Chicago), 33-34; "Teaching Plato's Republic VIII and IX," The
Classical Journal) 46 (1951), 343-48; "Early Greek Theories of
Sex Determination," Journal of Heredity) XL (1949), 49-50;
HPlato's Divided Line," The Review of Metaphysics~ V (1952),
529-34.
Preparation of typescript copies of this manuscript was as-
sisted by a Research Grant from Indiana University, and Yale
University has provided funds and facilities for preparation of
the figllre drawings.
 Acknowledgments

ACKNOWLEDGMENT of permission to include notes originally

published in journals is made on page viii. Permission to quote
from the following works has also been granted. The abbrevi-
ated title used in the text and notes is given first in each case.

Adam, Republic. James Adam, edt The Republic of Plato. 2

vols. Cambridge, 1902. Quoted by permission of Cambridge
University Press.
Bury, Timaeus} Critias} etc. R. G. Bury, trans. Timaeus Critias}
J

Cleitophon} Menexenus} and Epistolae (Loeb Classical Li-

brary). New York and London, 1929. Quoted by permission
of Harvard University Press.
Chambray. Platon J Oeuvres Completes} t. vii, pts. 1-3, La Re-
publique} texte etab. par E. Chambray. Paris, 1934. Quoted
by permission of the Societe d'Edition, "Les Belles Lettres."
Cornford, Cosmology. Francis M. Cornford. Plato's Cosmology.
New York, 1937. Quoted by permission of Kegan Paul,
Trench, Trubner & Co., Ltd.
Cornford, Republic. Francis M. Cornford. The RejJublic of
Plato. New York and London, 1945. Quoted by permission of
Oxford University Press.
Greene, Scholia. William C. Greene. Scholia Platonica. Haver-
ford, 1938. Schemata from scholia reproduced by permission
of the American Philological Association.
Heath, Euclid. Sir Thomas Heath. The Thirteen Books of Eu-
clid's Elements} 2nd ed., 3 vols. Oxford, 1928. Quoted by per-
rrlission of The Clarendon Press, Oxford.
Heath, History. Sir Thomas Heath. History of Greek Mathe-
matics} 2 vols. Oxford, 1921. Quoted by permission of The
Clarendon Press, Oxford.
ix
x Acknowledgments

Jowett, Dialogues. Benjamin Jowett, trans. The Dialogues of

Plato, 3rd ed., 5 vols. New York and London, 1892. Quoted
by permission of The Clarendon Press, Oxford.
Shorey, Republic. Paul Shorey, trans. The Republic (Loeb
Classical Library), 2 vols. New York and London, 1930.
Quoted by permission of Harvard University Press.
Taylor, Commentary. A. E. Taylor. A Commentary on Plato's
Timaeus. Oxford, 1928. Quoted by permission of The Clar-
endon Press, Oxford.
Wilson. J. Cook Wilson. "Plato, Republic, 616E." Classical Re-
view, XV (1902), 292-93. Quoted by permission of the Clas-
sical Review.

Other Works Cited by Abbreviated Title

Dupuis' Summary. J. Dupuis. Le nombre geometrique de Pla-
ton: interpretation definitive (in Theon, Exposition des Con-
naissances mathematique ... , edt with French trans. by J.
Dupuis). Paris, 1892.
Hermann, Scholia. C. Fr. Hermann, ed. Plato 15: Appendix
Platonica. Leipzig, 1875.
Taylor, Plato. A. E. Taylor. Plato: the Man and His Work, new
edt New York, 1936.
Theon. Theon Smyrnaeus. Expositio rerum mathematicarum
ad legendum Platonem utilum, E. Hiller, edt Leipzig, 1878.
Timaeus (text). Platon, Oeuvres completes, t. x, Timee, Critias,
texte etab. par A. Rivaud. Paris, 1925.
Todhunter, Euclid. The Elements of Euclid, edt by Isaac Tod-
hunter (Everyman's Library). New York, 1933.
 Contents
PAGE
Introd uction
Plato's "Mathematical" Passages 3
Types of Mathematical Metaphor 7
Orientation and Limits of the Present Study 8
Final Comment on Tactics 10

Part One
MATHEMATICAL IMAGES RELATIVELY INDEPENDENT OF THEIR
DIALECTICAL CONTEXTS

CHAPTER I Examples from Pure Mathematics of Methods

and Class-Relations
Introductory Comment: Some Simple Illustrations 15
I. Definition of Odd and Even Number 17
II. Proof of the Recollection Theory of Knowledge 19
III. The Geometer's Method of Hypothesis 32
IV. Definition of Roots and Surds 38

Part Two
MATHEMATICAL IMAGES CLOSELY DEPENDENT ON THEIR
DIALECTICAL CONTEXTS

CHAPTER II "Social Statistics": Arithmetic Detail

I. Atlantis and Its Institutions 47
II. The Social Institutions of the Laws 59
6. Mathematics and the Law 59
xi
xii Contents

PAGE
b. Administrative Logistic 61
c. Mathematics in Education for Citizenship 62
d. Statutory Mathematics: Fines and Sumptuary Laws 66
III. Arithmetic Details in Myth and Chronology 68

CHAPTER III Geometric Metaphor

VERBAL MATRICES: Introductory Remarks 72
I. Construction of a Matrix 83

IMAGES OF HARMONY AND CYCLE: The Republic 85

II. Mathematical Imagery in the Republic; the State and
the Musical Scale 85
III. The Cycle of Social Progress 87
IV. The Divided Line 91
V. Mathematics in Higher Education 104
VI. The Nuptial Number 107
a. Introductory Remarks 107
b. Problems of Translation and Pa.raphrase 109
c. The Philosophy of History 112
d. The Rulers' Problem 113
e. Why the Muses Speak Playfully 114
f. Detailed Interpretation 115
g. Interpretation Established from Text of the Passage
Alone 131
h. Historical Comments: Other Interpretations 143
VII. The Tyrant's Number 151
VIII. The Myth of Er-Astronomy 161
Q. The Unity of Republic x 161
b. The Allegorical Intention of the Myth 167
c. Detailed Interpretation 171
d. Text of the Passage: Theon's Version 196
IX. The Myth of Er-Transmigration 203

CHAPTER IV Algebraic Metaphor

Introductory Comment 209
I. Construction of the World-Soul 221
II. The Theory of Vision 230
III. The Theory of Geometrical Elements 238
xvi Figures
PAGE
26. Multiplication Table 75
27. Pythagorean Table of Contraries 76
28. Hippocratic Genetic Theory: A Combination Table 77
29. Standard Spatial Orientation of a Platonic Matrix 78
30. Political Matrix from Laws 78

31-35. Some Typical Uses of Matrices in Scholia

31. Gorgias 465C 78
32. Gorgias 477A 79
33. Repub lie 435B 79
34. Repub lie 440E 79
35. Correction of Republic Matrices 79
36. Schematization and Reduction of Three Times Three
Plane Matrix 80
37. Multiplication of Two Similar Matrices 81
38. Schematized Multiplication of Two Similar Matrices 82
39. Sophist Matrix 84
40. Changes of Topic and Locations of Mathematical Images
in Rep'ublic 87
41. Growth of State: Circular Impulsion 90
42. Growth of State: Circle under Construction 90
43. Divided Line 101
44. Alternative Construction of Divided Line, Carrying Out
Inequality of Segments 103

45-48. Figures from Scholia

45. Republic 5l0B 105
46. Republic 508A 105
47. Republic 510D 105
48. Republic 534A 105
49. Principles of Mathematical Sciences 106
50. Scholion from Proclus 137

51-61. Nuptial Number

51. Life Cycle 138
52. Acme of Human Life Cycle 138
53. The 3-4-5 Triangle 139
 Figures xvii
PAGE
54. Shape of Final Figure 140
55. Lines of Genetic Triangle Recombined into Squares 140
56. Identity of "Diagonals" 141
57·}s
58.
q uares on Diagonals 142
59. Phaedrus Matrix of Lives 142
60. List Form of Matrix Used in Phaedrus 142
61. Matrix Underlying Republic IX 143
62. Scholion Figure to Illustrate Tyrant's Number 158
63A. Parallels among Phaedrus 248, Republic VIII-IX, Re-
public III-VII 158
63B. Tyrant's Number: First Step of Calculation 159
64. Tyrant's Number: Second Step of Calculation (Squaring) 159
65. Tyrant's Number: Final Calculation 160
66. Law of Nines: Sizes 198
67. Law of Nines: Colors 199
68. Law of Nines: Colors (Alternative Version) 199
69. Size and Color (Volume and Density) 200
70. Size and Color (Ordinal): Another Law of Nines 200
71. Law of Nines: Balance o~ Velocities 201
72. Proportionality of Distance and Forward Velocity 201
73. Final Law of Nines: Exact Balance of Momentum in
System 202
74. Lack of Balance in Proclus· Alternative Text of List of
Sizes 202
75. Necessity's Spindle in Cross Section 203
76. Matrix of Character with Four Levels of Knowledge Dis-
tinguished 207
77. Matrix of Character Applied to Olympian Gods Who Are
Patrons of Types of Human Life 208
78. Scholion: Three Kinds of Ratio 218
79. Scholion: Ten Kinds of Motion (Laws 894B) 219

80-86. Construction of World Soul

80. Initial Terms of Scale 227
81. Arithmetic Means Inserted 228
82. Harmonic Means Inserted 228
83. Both Means Inserted 228
xviii Figures
PAGE
84. Matrix Form of Computation of Series 228
85. Traditional Computation 229
86. Traditional Computation in Matrix Form 229
87-93. Theory of Vision: Distorting Effects of Mirrors
87. Reversal in Plane Mirror 234
88. Rectification of Mirror to Prevent Reversal 234
89. Effect of Reversed Intellectual Vision 235
90. Inversion in Mirror 235
91. Effect of Inverted Intellectual Vision 236
92. Effect of Mirror at Oblique Angle 236
93. Effect of Oblique Intellectual Vision 237
94. The Atomic Triangles 244
95. Synthesis of Equilateral Molecular Triangles 245
96. Synthesis of Square Molecular Planes 245
97. Decomposition of Polygons into Triangles 246
98. Homoeomereity of First Atomic Triangle 247
99. Homoeomereity of Second Atomic Triangle 247
100. Transmutation Ratios of Elementary Solids 248
101. Ratios of Numbers of Bounding Planes of Elementary
Solids 248
102. "Diagonals and Their Diagonals" 258
103. Locomotion of Animals 259
104. Lack of Balance in Poseidon's Engineering 261
105. Properties of the Image of Concentric Circles in Three
Types of Context 262
106. Later Diagrams of Class-Inclusion: Euler, Leibniz 264
107. Symbolism: Figures Involving Incommensurability in
Plato's Mathematical Imagery 266
A. Meno (1) G. Republic (2)
B. Meno (2) H. Timaeus (1)
C. Theaetetus I. Timaeus (2)
D. Critias J. Timaeus (3)
E. Laws K. Republic (3)
F. Republic (1) L. Statesman
M. Republic (4)
108. Construction of Divided Line by Divisions in Mean and
Extreme Ratio 270
 Figures
PAGE
1. Greek Abacus (Schematic Representation) 20
2. Abacus Pebble-Configurations 21

3-11. Meno Discussion: Figures from Scholia

3. Initial Four-Foot Square 27
4. Calculating Square Area 27
5. Shape of "Oblong" 27
6. First Construction: Square with Area of Sixteen Square
Feet 27
7. Second Construction: Square with Area of Nine Square
Feet 28
8. Demonstration of Area of Nine-Foot Square 28
9. Final Construction: Four Four-Foot Squares Joined, with
Diagonals Drawn 28
10. Demonstration That Diagonal Divides Unit Square into
Two Half-Unit Triangles 29
11. Final Scholion Figure 29
12. Proposed Emendation of Final Scholion Figure 29
13. Later Representations of Meno Figure 30
14. Meno Hypothesis: Benecke's Interpretation 36
15. Meno Hypothesis: Heath's Interpretation 37
16. First Case of Euclid 111.35 38
17. Proof Attributed to Theaetetus: Euclid X.9 43
18. Plan of Inner City of Atlantis 53
19. Canals in Atlantean Plain 54
20. Topography of Ancient Athens 57
21. Functional Civic Plan (Laws) 58
22. Military Strength of Atlantis 59
23. Factors of 5,040: a Legislator's Manual 69
24. Statutory Fines and Sumptuary Laws (Laws) 70
25. Election Statutes (Laws) 71
xv
 Contents xiii

PAGE
CHAPTER V Mathematical Jokes: The Limits of
Mathematical Metaphor 249

APPENDIX A Effect of Context on Three Platonic

Images of Cycle 260

APPENDIX B Symbolism: The Significance of the Spe-

cific Figures Chosen as Illustrations from Pure
Mathematics 264

NOTES 273

INDEX 299
I
I

I
Introduction
Plato's UMathematical" Passages

IN THIS study, the theory has been advanced that the "mathe.-
matical" passages in Plato which have seemed nonsense or
riddles to previous students in fact describe diagrams which
Plato had designed, and were intended to accompany and clarify
his text. The result has been, if my evaluation is correct, the re-
covery of new primary source material for the study of Plato.
Recovered diagrams meant to illustrate the Timaeus and
Republic are evidently of importance and interest to anyone
who has occasion to read or teach Plato. One of the facts that
have made me think this study correct has been that these fig-
ures actually do clarify the text, even for the beginning student.
Consequently, the presentation has been made intelligible
to a reader with no special command of Platonic Greek, of
ancient mathematics, or of Plato scholarship. The text is in
English, or accompanied by English translation, throughout;
collateral information is supplied where needed; and so far as
possible the discussions have been compressed, so that one gets
to the figures at a reasonable pace. Nevertheless, it is important
to omit or evade no problem affecting the correctness of the
attribution of these illustrations to Plato, and some sections, in
spite of all attempted shortening, have had to include many
technical details of figure reconstruction.
For two thousand years certain "mathematical" passages have
puzzled Plato's readers. Humanistic scholars have generally
chosen to ignore them in favor of passages of more literary
interest. Neo-Platonic enthusiasts have tended to center atten-
8
4 Plato's Mathematical Imagination

tion on them, with the apparent expectation that some mys-

terious, occult key to Plato's philosophy and to the nature of
reality had been enciphered there. Modern treatments tend to
fall into one or the other of these traditions. On the one hand,
the humanistic scholar who reads his Plato as belles-lettres
uses the term "literary effect" to bypass explanation of obscure
mathematical passages. In literature, it may be allowable to
introduce meaningless passages to create an effect of difficulty;
in philosophy, it is not. Using this explanatory device, we should
have to decide that Immanuel Kant devoted whole volumes,
with a skill rare in literary history, to the creation of intrin-
sically nonsensical prose which admirably achieves the "literary
effect" of obscurity. A corollary of this approach is that such
passages in Plato need not be taken seriously, and might be
deleted without any effect on the rest of the work. At the oppo-
site extreme, adherents of the Neo-Platonic canons of interpre-
tation invest these passages with tremendous significance; they
believe them to contain esoteric keys to hidden doctrines which
the author did not want to express more clearly. Sometimes
they expect that such passages will be capable of interpretation
for all subjects and at all levels of generality simultaneously;
sometimes they assume that there is a hidden, specifically em-
pirical reference to a single phenomenon or number, which an
"interpretation" should discover; and sometimes, that there is
in these passages a fixed symbolism with a key, in which various
quantitative elements represent concepts, and images thus sym-
bolize encoded propositions.
These two traditions illustrate perfectly Plato's remark, in
Epistle VII, about the various ways in which a man's written
works will be misread. Some readers, he says, not understanding .1.
the writing, will feel contempt for it and ignore it; others will I
expect too much from it, and study it in the belief that mastery :1
of written formulae will make them experts in philosophy. The ~
humanistic tendency to delete or bypass mathematical passages, !.~
on the ground that they are nonfunctional or not serious, is a I
misreading of th~ former sort, whereas the Neo-Platonic con.. 1
viction that there is some sort of key to the universe hidden in
 INTRODUCTION 5

this symbolism is a clear case of the latter kind of misreading.

It is easy for the modern reader to forget that Plato was heir
to a pedagogical tradition in which the use of mathematical
illustration seems to have been a standard device for clarifying
discussion. It is equally easy to forget that behind this tradition
lay the new discovery of mathematics as a promising instrument
of science, and that successes with it in music and astronomy
must have led one to expect equally effective applications to
politics and medicine. We may even more easily forget that the
simple technological notion of including illustrations with a
book-e.g., drawing a diagram on a manuscript beside the text
it was to illustrate-had not occurred to anyone when Plato
wrote.
But if we forget these things, whole stretches of Platonic
text in which (it will be suggested in this study) the construc-
tions of diagrams to aid the reader are explained, become unin-
telligible, and what were meant as pedagogical aids become
special textual problems, serving as obstacles instead.
Contemporary references, later tradition, mathematical vo-
cabulary, and collateral material from the history of science all
indicate something about the notion of "mathematics" that the
Pythagoreans held. What they indicate appears quite clearly to
be that mathematics was not sharply separated from other
branches of natural science. The statements made about the
relations of things and numbers by the Pythagoreans make sense
only on the assumption (which their vocabulary confirms) that
the language of this mathematics was intended to be metaphori-
cal. By this I mean that the genera and differentiae into and by
which numbers were classified were not regarded as peculiar
to numbers, but were classifications derived from and shared
by nonmathematical subject matters; thus, when a surface was
called a "skin," or a line not amenable to measurement was
called "irrational," that naming reflected a conviction that there
is some common class including both skins and geometric sur-
faces, both moral stubbornness and incommensurability. Insofar
as members of the same class have common properties, theorems
about surfaces are also theorems about skins, and theorems
6 Plato's Mathematical Imagination

about incommensurables are also theorems about moral be-

havior. Conversely, problems involving skins or morality may
have analogues in pure mathematics, and it is these analogues
that are crucial in the use of mathematics as a technique of re-
search and of pedagogy.
If we can find numbers falling in the same classes as the
entities we want to study, the relations of these entities can be
illustrated and investigated by observing the correlated rela-
tions of the analogous classes of numbers. The interpretation of
such illustrations involves (1) recognition that the properties
involved are common to a broader class than quantity, in which
class this given quantitative illustration is included, and (2)
from this recognition, transference (by metaphor) of the insight
gained from the quantitative case to the analogous nonmathe-
matical cases.
Our modern view of mathematics has completely disso-
ciated "number" from such attributes as sex or justice, and
assimilated mathematics to the purely formal study of logic,
rather than to ethics, physics, or aesthetics. This dissociation
of mathematics began with Aristotle's separation of his "theo-
retk sciences," and was strengthened by later reaction against
the Neo-Pythagoreans. The effect of this location of "mathe-
matics" in our contemporary canon of sciences is to make the
contemporary historian and reader find surprisingly little ma-
terial in Platonic and pre-Platonic texts which he can recognize
as distinctively "mathematical." Much of the work which the
early Greek philosophers and scientists thought of as mathe-
matics is not "mathematics" in its twentieth-century form at all.
If, however, the contemporary reader becomes sensitive to the
relation of this early mathematical work to metaphor (and our
contemporary terms, such as "square" numbers, "roots," etc.,
still bear clear traces of such metaphorical thinking at their
time of origin), he will find it possible to approach these earlier
discussions of quantity and mathematics much more sympa-
thetically and to interpret them more accurately. Naturally
such an orientation would place great emphasis in teaching
 INTRODUCTION 7
practice on the diagram and schematism, in social as well as
natural science.
In addition to the practice of Plato in this matter, there is
clear evidence of the existence and pervasive effect of such a
pedagogical tradition in the writing of Aristotle. Although he
is consistently critical of the tendency, which he detects in
Platonists and Pythagoreans, to assimilate philosophy and mathe-
matics, Aristotle's own lecture notes are interspersed through-
out with mathematical illustration. Pedagogical diagrams
involving mathematization are referred to throughout his dis-
cussions of logic and ethics and throughout whole sections of
the physics and metaphysics, and on occasion also playa major
role in the elucidation of biology. This seems to indicate that
such diagram technique was accepted as a matter of course even
by the arch-opponent of the philosophical orientation from
which the technique originated.

Types of Mathematical Metaphor

What has been called "mathematical metaphor" may evidently

vary in respect to the branch of mathematics from which it is
drawn and, consequently, in the aspect of a subject matter to
which it directs attention. For example, a proof taken from
"pure" mathematics may illustrate a point of logic relevant to
some discussion at hand. A set of arithmetical ("statistical" in
a modern sense) descriptive details may call attention to the
integration, or lack of it, in the institutions of a society. Again,
a geometrical figure may present spatial analogues of the key
relationships of concepts presented in a lecture or discussion.
A metaphor from algebra may give a generalized formulation of
the structure proper to a certain class of formulae. This range
of source and function prevents any examination of the purely
mathematical properties of such a set of illustrations from yield-
ing any agreement as to what they illustrate; their common
function of illustrating a context may be realized through many
forms of functional quantitative structure.
8 Plato's Mathematical Imagination

Orientation and Limits of the Present Study

This study might be characterized as a "dialectical" interpre-
tation. Two features seem to justify that characterization. The
first is the consistent objective of establishing interpretations
which make philosophic sense in their contexts (which does not,
of course, relieve them of the need to make historical, mathe-
matical, and philological sense as well). This is basically, how-
ever, a philosophic rather than a philological or historical ap-
proach, since the "contexts" of these passages are conceived as
functional parts in the exposition of a philosophy. The second
feature of the study is its recognition throughout of the respon.-
siveness of the mathematical illustrations chosen to the con-
textual points they are intended to elucidate. A functional
mathematical image must reflect these contextual metllods, prin-
ciples, and distinctions adequately, even if some deviation from
the techniques of "pure" mathematics is required to seCllre such
reflection. Past interpretations seem often to have postulated
the detachability of all mathematical passages from their con-
texts, as though they enjoyed an independent nleaning, like
theorems in a system of geometry; or to have postulated that
the passages could be understood by setting them all indiscrimi-
nately into some single eclectic context.
Since the present approach is basically philosophic, the proper
criterion of its adequacy is whether or not it does, without
intruding historic or philological implausibilities, illuminate
the passages discussed, considered as parts of Plato's exposition
of a philosophy. If an alternative approach had been chosen, the
criterion might have been historical: Will the postulated inter-
pretation explain and reconcile the extant statements of all
interpreters, if we correct those statements in the light of our
knowledge of the convictions and style intruded into the origi-
nal by each? This is a separate and extensive enterprise, not
here undertaken; in the present study antecedent interpretations
are treated schematically, by criticizing or displaying the as-
sumptions operative in their construction, not individually,
except in those cases in which they have suggested some relevant
 INTRODUCTION 9

insight. If a third approach had been chosen, the passages might

have been studied as specimens of (presumably meaningful)
classical Greek in which the meanings were expressed by means
of syntactical and etymological devices analogous to those
used by other Greek authors. Such an approach would be much
more direct than either of the others, if it were only possible;
but in every case where it has been tried (and it has been tried
often) it has been necessary to resort to maxims of higher criti-
cism which have generally taken the form of intruded, un-
Platonic philosophical assumptions.
The history of Platonic scholarship goes far toward under-
scoring the need for the present mode of interpretation. The
many studies of Plato of the historical and philological sort
have uniformly, so far as the passages discussed in this study
are concerned, contributed little or even negatively to our un-
derstanding of Platonic philosophy. Any reader of the Protagoras
must agree that Plato himself believed that the proper method
of textual interpretation was not that of Prodicus but of Soc-
~s, the philosopher.
[ The main object of the present inquiry is Platonic mathe-
matics as it is revealed in mathematical imagery. The concept
of image is used here in its strict Platonic sense, in contrast
to things, names, or formulae. Passages in which mathematics
tlPpears not as illustrating but as being what is talked about, as in
the Philebus, Statesman, and Parmenides (though they have
influenced the introductory sections, treating general technique
and kinds of imagery) have been excluded. The Introduction to
my doctoral thesis (The Role of Mathematics in Plato's Dialectic,
[University of Chicago, 1942], pp. 1-7) still seems to me an ex-
act statement of the distinction intended here between names,
images, and formulae, though the present study is entirely
independent of the dissertation.
Because of this limitation, the reader will feel that supple-
mentary treatments are required to deal completely with the
Platonic mathematical vocabulary (the names used), and with
mathematical formulae) through study of which one should
recover Plato's philosophy of mathematics.
10 Plato's Mathematical Imagination

However, until some agreement about the images has been

reached, "any man who wants to may upset any argument" put
forward to explain the formulae. An application to the study
of the Parmenides of the conclusions and techniques of the
present investigation has already convinced me that it is \vorth
while even for a philosopher less interested in images than in
realities to have undertaken this scrutiny of those images as a
preliminary inquiry.

Final Comment on Tactics

The tactics of presentation used in this study have been to
present translations of the passages concerned, with translations
of relevant scholia, then discussion of alternative interpretations
and problems, ending so far as possible with final interpreta-
tions resolving these problems.
It would have been impossible to construct my own transla-
tions of these passages which did not read into them my own
notions of interpretation; and even if it had been possible to
do so, any normally cautious reader would be rightly suspicious.
It is easy to conjure a rabbit from a hat if you have already
hidden it there, and the conflicting past assertions of philologists
as to "the only meaning the Greek can bear" certainly indicate
that translations of these texts leave plenty of room for con-
flicting interpretations. I have therefore developed interpreta-
tions from other English translations, chosen for their neu-
trality. In almost every case, these were made by translators
who actually had in mind other interpretations than those
developed in this study.
Translations of scholia are included both to show what sort
of footnotes ancient scholars annexed to the mathematical pas-
sages in their manuscripts of Plato, and to prove that no set
of diagrams derived from Plato's original manuscripts was
known to Hellenistic scholars. On occasion, the text and the
scholiast's marginal diagram are sufficiently at cross-purposes to
rule out the possibility of any direct tradition.
In various introductory sections, an attempt is made to define
 INTRODUCTION 11

sharply such terms as "nlatrix" and "analogy," which are used

in the text to describe Platonic mathematical techniques and
devices. Some concepts and notation of contemporary mathe-
maticallogic have been used to give the desired generality and
precision to these statements. This use of concepts from twen-
tieth-century formal logic certainly appears anachronistic, and
in the present study is not explicitly related to the imagery that
it describes. For the present, it is offered simply as a useful
device for talking about the properties of classes of mathemati-
cal constructions and illustrations. In a later discussion of
Plato's philosophy of mathematics, however, I expect to show
that the modern concepts intruded here are essentially Platonic,
and do not really involve what seems to be an anachronism.
One final fact requires emphatic statement: To analyze or
explain an illustration is different from appreciating it directly.
. t\.n author and his intended readers have in common certain
habits of imagination and notions of pedagogy, and these read-
ers need not analyze out as separate abstract relations the various
properties of examples; they have an intuitive response to these
without analysis. The more a reader recognizes what this direct
awareness is like, the more keenly he will feel that an abstract
analysis of imagery is missing something important. Since con-
crete aptness dissolves on analysis into many abstract relations
of relevance, no one of which has the concrete vividness that we
expect a good illustration to show, a commentator who uses such
analytic technique is often accused of reading a great deal more
into the text than its author intended, or of devising analyses
with a neat or "slick" finish which most illustrations lack. Of
course abstractions are neater than concrete cases; and of course
no author sets about creating illustrative examples by separate
consideration of a tremendous number of abstract connections.
But the intuitive appreciation of the materials analyzed here
presupposes an ancient Greek intllition to which we no longer
have the key; so the choice is between this analytic approach
to interpretation and no interpretation worth mentioning.
 Part One
MATHEMATICAL 'IMAGES RELATIVELY
INDEPENDENT OF THEIR DIALECTICAL
CONTEXTS
 CHAPTER I

Examples from Pure Mathematics of Methods

and Class-Relations
Introductory Comment: Some Simple Illustrations

ONE OF the most frequent and typical intrusions of mathemati-

cal imagery into the dialogues is as illustration of class-relations,
particularly in connection with problems of method or defini-
tion. The Euthyphro,l Meno,2 Phaedo,3 and Theaetetus 4: all
contain such illustrations. Another use is the citing of mathe-
matics as an illustration of a specific art with its own definite
subject ~~tter; i~ tJ:1e Charmides,5 Statesman,6 Euthydemus,7
Gorgias,8 and Protagoras 9 there are such illustrations. The only
problems of interpretation which illustrations of this kind pre..
sent are peripheral: the terminology, intended construction, or
assumed differentiation of the branches of mathematics are
sometimes not clear. In such cases, the function of the interpre-
ter is to revive the sharpness and cogency of the intended
illustration in its original setting; but since the function of
these examples is, in context, perfectly clear, their elucidation
will not throw much light on Plato's philosophy, and therefore
very little is gained from more intensive interpretations.
The case is far otherwise with images and diagrams intended
to illustrate or summarize crucial relations of terms in:'~-t~
contextual dialectic, or the contextual method. Often in these
cases Plato's use of mathematical imagery to clarify dis-
cussions, where we no longer feel the relevance of such imagery
to do so, poses a series of complex problems; and since the
reconstruction of these diagrams provides a statement of how
their context is to be interpreted, we should spare no effort to
reconstruct them correctly.
15
16 Plato's Mathematical Imagination

The illustrations dealt with in this first chapter are all of the
first sort; therefore, apart from supplying relevant diagrams
and noting certain controversies over mathematical details of
interpretation, the treatments of these passages have been kept
brief. Even in this section, however, each of the images selected
has some relevance to the study of Plato. The sacrifice of
accuracy to dramatic consistency in the Euthyphro example,
the geometry of class-inclusion in the Meno} the construction
of the TheaetetusJ all make some such contribution.
In following the suggested method of beginning with simple
images which present no problem requiring extensive inter-
pretation, and in postulating that the later, more complex
images are similar in function to these (a postulate which can
be tested by seeing how satisfactory the interpretations are that
it suggests), we may profitably begin with the mention of a
group of "trivial" ilillstrations; cases where neither subtle
analysis of the context nor any knowledge of mathematics be-
yond its rudiments is required for interpretation. In Republic
428, for example, Socrates says that if we are looking for four
things, and have found three, the fourth must be the one that
is left. This is a simple way of illustrating the contextual point
that in the state, if courage, wisdom, and temperance have been
defined, the cardinal virtue that remains will be justice. In
Republic 458, Glaucon contrasts "geometrical" and "erotic"
necessity, the one proceeding from reason, the other from the
passions. In Republic 337, Socrates asks Thrasymachus how a
man could define 12 if he were forbidden to name any of its
factors in his definition. (Thrasymachus has just challenged
Socrates to define justice, without saying that it is good, ex-
pedient, beneficial, or "any such nonsense. H) In the Theaete-
tus,l° as a case of error, the example of the man who believes
+
that 7 5 = 11 is considered. In the SymposiumJ 11 Aristoph-
anes represents Zeus as calculating that if he divides each man
in half he will get t'ivice as many sacrifices. In Euthyphro 7,
Socrates points out that metric techniques settle disagreements
over number, weigl1t, and measure. In the Meno,12 the differ-
ence between a class and its members is illustrated by reference
 EXAMPLES FROM PURE l\fATHEMATICS OF METHODS AND CLASS-RELATIONS 17

to the difference between defining "figure" and enumerating

specific figures.
These six passages are offered as instances of uncomplicated
mathematical allusion, functional in illustrating problems of
methods in a qllantitative subject matter where their pattern
will show most clearly. There is nothing about them that is
obscure or mysterious, or nonfunctional in context. Since their
pedagogical intention, contextual relevance, and mathematical
content are very clear, these examples have been chosen because
they show most typically characteristics which, as this study is
intended to establish, all Platonic mathematical images possess.
It must be admitted, however, that an attempt to extend this list
much further-for example, by checking citations of &Ql{}fl6~ or
rrA~{}o~ in a Platonic lexicon-would quickly arouse controversy.
The present passages are atypical in the degree of their apparent
independence of relatively remote context, which gives them
perhaps less interest and artistic merit than, say, the opening of
the Timaeus with a simple enumeration of the per~ons present
for the conversation,18 or the discussion of odd and even number
in the Phaedo J• 14 but this very fact of apparent relative isolation
gives them particular pedagogical suitability as starting-points
for a program of analysis designed to lead from simpler to
more complex mathematical imagery.

I. DEFINITION OF ODD AND EVEN NUMBER

Euthyphro 12·
soc.: Then if piety is a part of justice, I suppose that we should
~ enquire what part? If 'you had pursued the enquiry in the previ-
ous cases; for instance, if you had asked me what is an even num-
ber, and what part of number the even is, I should have had no
difficulty in replying, a number which [is isoskeles~ not skalene t].
Do you not agree?

• Trans. ]o\vett, Dialogues, I, 394.

t The transliteration in brackets is substituted for Jo\vett's rendering ("repre-
sents a figure having two equal sides") because the substitution seems to give the
translation greater neutrality as to interpretation.
18 Plato's Mathematical Imagination

Then we are wrong in saying that where there is fear, there is

reverence; and we should say, where there is reverence there is
also fear. But there is not always reverence where there is fear;
for fear is a more extended notion, and reverence is a part of fear,
just as the odd is a part of number, and number is a more ex-
tended notion than the odd. I suppose that you follow me now?

Scholia *
4 (Euthyphro 12D)
The scalene triangle has three unequal sides. The isosceles,
however, has two sides equal to one another, and a third un-
equal. Since an even number is divisible into two equal numbers,
as for example 8 (into two 4's), but the odd into unequal, for
example 5, the one is called "isosceles," and the other "scalene."

Scalene: crooked and irregular. There are three kinds of triangles

-equilateral, isosceles, and scalene.

These peculiar definitions of odd and even numbers, as

"scalene" and "isosceles," respectively, have occasioned some
puzzled comment, because they seem either capricious or in-
exact. The basis of the peculiarity is dramatic, not mathe-
matical.
The Platonic dialogue is an artistic representation of a con-
versation in which the illustrations offered by the speaker are
appropriate to the character of his listener; when they are not,
the listener says so. The demands of artistic consistency require
that the speaker use examples dramatically intelligible to his
audience.
If, in this connection, we consider the character of Euthyphro,
we find a prototype of the "ageometrical" man, denied entrance
by the inscription over the door of the Academy. He consistently
substitutes piety and religious precedent for independent in-
tellectual inquiry; he is neither interested nor skilled in matters
of dialectic or mathematics. Consequently, when Socrates wants
• Greene, Scholia Platonica~ 419, 3.
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 19

to suggest, early in their discussion, that there should be some

techniques for resolving political disagreements, he cites meas-
uring and weighing as examples of demonstrative technique.15
The practical empirical aspect of mathematics, represented by
everyday use of ruler and balance, is the only sort of "mathe-
matical" reference which Euthyphro finds familiar. Conse-
quently, when odd and even numbers are cited as examples of
definable classes, we should expect any definition that he can
understand to have the same reference to everyday experience.
The obvious analogue to the ruler and scale is the abacus or
counting-board. 16 A set of pebbles on the board can be placed
in pairs on either side of the central line; the two lines of
pebbles are either equal, or one is longer. In the former case,
the pebbles outline an "equal-sided" rectangular figure; in the
latter case, an irregular quadrilateral ("scalene").
This peculiar mode of abacus-centered operational definition
does not reflect Plato's own theory of number; it is simply
Socrates' concession to the mental backwardness of his com-
panion; the images of familiar abacus pebble-configurations
seem the only counterparts Socrates can find to more accurate
definitions which would be artistically inadmissible, because
they would be beyond Euthyphro's dramatically established
powers of mathematical and abstract imagination. (See figures
I and 2.)

II. PROOF OF THE RECOLLECTION THEORY OF KNOWLEDGE

The following discussion with the slave boy ill the ]\rlcno is
one of the clearest cases of a passage giving step-by-step direc-
tions to the reader for drawing the diagram referred to. The
directions are so clear-cut that from the time the passage was
written there seems to have been no controversy over its in-
terpretation. Some unusually geometrically-mil'lded scholiast has
carefully constructed all the figures referred to, representing
each by a separate figure, rather than as subdivisions of a single,
more complex figure, as do later diagrams illustrating the pas-
sage. IT This use of separate figures makes even clearer the exact
20 Plato's Mathenlatical Imagination

Figure I

GREEK ABACUS

(Schematic Representation)

Figure 1 is a schematization of the only Greek abacus which has

been discovered (the "Salaminian table"). The table is discussed by
J. Gow, Short History of Greek Mathematics (Cambridge Univer-
sity Press, 1884), pp. 33-36. Although the reported lettering on the
table indicates that at least one of its uses was for monetary reckon-
ing,18 it has also been suggested that the table was used as a game-
board, and its design partly dictated by this function. In any case,
the existence of such computers' aids in Plato's time is certain; and
Plato's coupling of references to the game of tCE".l:ELa with com-
putation (in the Myth of Theuth and Thamus in the Phaedrus 19)
is one which is immediately suggested by the similarity of manipu-
lating counters on an abacus and pieces on a checkerboard. The
important point for the present study is that an abacus is a device
taking advantage of spatial orientation to differentiate kinds of
relation, a general technique discussed in Chapter III, Introduction,
in connection with the concept of "verbal matrices."
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 21

Figure 2
ABACUS PEBBLE-CONFIGURATIONS

ee ,.--------,
e_
ee
ee ee
e-
'-_- .1
ee L- ..I

"ISOSCELES" SET "SCALENE" SET

The analogy of sets to figures was basic in Pythagorean mathe-

matics and is still retained in the terms "square" and "cube" num-
bers in our own mathematical vocabulary. The use of a dot nota-
tion, representing each number as a spatially arranged set of units,
facilitated the formulation of such analogies, which gave the Pytha-
goreans a kind of reversed analytic geometry in that they could
represent continuous figures by analogous discrete sets. See John
Burnet, Early Greek Philosophy" 4th edition (London, 1930),
pp. 99-112, 276-309; Aristotle, Metaphysics 1092b 8 (Eurytos). See
also note 38, Chapter IV, following.

aspect of the diagram intended. These scholia figures are re-

produced on pages 27-29. They have been cross-referenced in the
text,20 so that each of Socrates' statements is referred to the
figure showing what part of the construction he is pointing to.

Meno 82 fI.*
MENO: ... If you can prove to me that what you say [that
knowledge is recollection] is true, I wish that you would.
SOC.: It will be no easy matter, but I will try to please you to
the utmost of my power. Suppose that you call one of your nu-
merous attendants, that I may demonstrate on him.
MEND: Certainly. Come hither, boy.
* Trans. Jowett, Dialogues, I, 361-65. References to figures are in brackets.
22 Plato's flt.fathematical Imagination
SOC: He is Greek, and speaks Greek, does he not?
MENO: Yes, indeed. He was born in the house.
SOC.: Attend now to the questions which I ask him, and ob-
serve whether he learns of me or only remembers.
MENO: I will.
SOC.: Tell me, boy, do you know that a figure like this is a
square? [Fig. 3]
BOY: I do.
SOC.: And do you.know that a square figure has these four
lines equal? [AB, AC, CD, BD, Fig. 3]
BOY: Certainly.
SOC.: And these lines which I have drawn through the middle
of the square are also equal? [EF" GH) Fig. 3]
BOY: Yes.
SOC.: A square may be of any size?
BOY: Certainly.
SOC.: And if one side of the figure be of two feet, and the
other side be of two feet, how much will the whole be? [ABDC,
Fig. 3] Let me explain: if in one direction the space was of two
feet, and in the other direction of one foot, the whole would be
of two feet taken once? [Fig. 4]
BOY: Yes.
SOC.: But since this side is also of two feet, there are twice two
feet?
BOY: There are.
SOC.: Then the square is of twice two feet?
BOY: Yes.
SOC.: And how many are twice two feet? Count and tell me.
BOY: Four, Socrates.
SOC.: And might there not be another square twice as large
as this, and having like this the lines equal?
BOY: Yes.
SOC.: And of how many feet will that be?
BOY: Of eight feet.
SOC.: And now try and tell me the length of the line whi~
forms the side of that double square: This is two feet-what will
that be?
BOY: Clearly, Socrates, it will be double.
SOC.: Do you observe, Meno, that I am not teaching the boy
anything, but only asking him questions, and now he fancies that
EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 23

he knows how long a line is necessary in order to produce a figure

of eight square feet; does he not?
MENO: Yes.
SOC.: And does he really know?
MENO: Certainly not.
SOC.: He only guesses that because the square is double, the
line is double.
MENO: True.
SOC.: Observe him while he recalls the steps in regular order.
(To the boy) Tell me, boy, do you assert that a double space
comes from a double line? Remember that I am not speaking of
an oblong [such as ABDC, Fig. 5] but of a figure equal every way,
and twice the size of this-t:qat is to say, of eight feet; and I want
to know whether you still say that a double square comes from a
double line?
BOY: Yes.
SOC.: But does not this line become doubled if we add an-
other such line here? [AC + CG, Fig. 6]
BOY: Certainly.
SOC.: And four such lines will make a space containing eight
feet?
BOY: Yes.
SOC.: Let us describe such a figure: Would you not say that
this is the figure of eight feet? [Fig. 6]
BOY: Yes.
SOC.: And are there not these four divisions in the figure, each
of which is equal to the figure of four feet? [ABDC" CDHG,
BDFE" FDHI, Fig. 6]
BOY: True.
SOC.: And is not that four times four?
BOY: Certainly.
SOC.: And four times is not double?
BOY: No, indeed.
SOC.: But how much?
BOY: Four times as much.
SOC.: Therefore the double line, boy, has given a space, not
twice, but four times as much.
BOY: True.
SOC.: Four times four are sixteen-are they not?
BOY: Yes.
24 Plato's iWathematical Imagination

soc.: What line would give you a space of eight feet, as this
gives one of sixteen feet;-do you see?
BOY: Yes.
SOC.: And the space of four feet is made from this half line?
[AB == ~ AE, Fig. 6]
BOY: Yes.
SOC.: Good; and is not a space of eight feet twice the size of
this, and half the size of the other? [i.e., 2· ABDC, ~ AGIE,
Fig. 6]
BOY: Certainly.
SOC.: Such a space then, will be made out of a line greater
than this one, and less than that one? [greater than AB, less than
AE, Fig. 6]
BOY: Yes, I think so.
SOC.: Very good; I like to hear you say what you think. And
now tell me, is not this a line of two feet and that of four?
BOY: Yes.
SOC.: Then the line which forms the side of eight feet ought
to be more than this line of two feet, and less than the other of
four feet?
BOY: It ought.
SOC.: Try and see if you can tell me how much it will be.
BOY: Three feet.
SOC.: Then if we add a half to this line of two, that will be
the line of three. [AB + BC, Fig. 7] Here are two and there is
one; and on the other side, here are two also and there is one
[AB, BC, and AD, DE, Fig. 7], and that makes the figure of which
you speak? [AFGE, Fig. 8 == AEFC, Fig. 7]
BOY: Yes.
SOC.: But if there are three feet this way and three feet that
way, the whole space [AEFC) Fig. 7] will be three times three
feet?
BOY: That is evident.
SOC.: And how much are three times three feet?
BOY: Nine.
SOC.: And how much is the double of four?
BOY: Eight.
SOC.: Then the figure of eight is not made out of a line of
three?
BOY: No.
EXAMPLES FRO~I PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 25
soc.: But from what line? Tell me exactly, and if you would
rather not reckon, try and show me the line.
BOY: Indeed, Socrates, I do not know.
SOC.: Do you see, lVleno, what advances he has made in his
power of recollection? He did not know at first, and he does not
know now, what is tIle side of a figure of eight feet; but then he
thought that he knew, and answered confidently as if he knew,
and had no difficulty; now he has a difficulty, and neither knows
nor fancies that he knows.
MENO: True.
SOC.: Is he not better off in knowing his ignorance?
lVIENO: I think that he is.
SOC.: If we have made hitn doubt, and given him the 'tor-
pedo's shock,' have we done him any harm?
MENO: I think not.
SOC.: We have certainly, as would seem, assisted him in some
degree to the discovery of the truth; and now he will wish to
remedy his ignorance, but then he would have been ready to tell
all the world again and again that the double space should have
a double side.
MENO: True.
SOC.: But do you suppose that he would ever have enquired
into or learned what he fancied that he knew, though he was
really ignorant of it, until he had fallen into perplexity under the
idea that he did not know, and had desired to know?
MEND: I think not, Socrates.
SOC.: Then he was the better for the torpedo's touch?
MENO: I think so.
SOC.: Mark now the farther development. I shall only ask
him, and not teach him, and he shall share the enquiry with me:
and do you watch and see if you find me telling or explaining
anything to him, instead of eliciting his opinion. Tell me, boy,
is not this a square of four feet which I have drawn? [ABED}
Fig. 9]
BOY: Yes.
SOC.: And now I add another square equal to the former one?
[BEFC} Fig. 9]
BOY: Yes.
SOC.: And a third, which is equal to either of them? [DEHG J
Fig. 9]
26 Plato's Mathematical Imagination

BOY: Yes.
SOC.: Suppose that we fill up the vacant corner? [EFIH, Fig. 9]
BOY: Very good.
SOC.: Here, then, there are four equal spaces?
BOY: Yes.
SOC.: And how many times larger is this space than this other?
BOY: Four times.
SOC.: But it ought to have been twice only, as you will re-
member.
BOY: True.
SOC.: And does not this line, reaching from corner to corner,
bisect each of these spaces? [DB, BF, FH, HD, Fig. 9]
BOY: Yes.
SOC.: And are there not here four equal lines which contain
this space? [DBFH, Fig. 9]
BOY: There are.
SOC.: Look and see how much this space is.
BOY: I do not understand.
SOC.: Has not each interior line cut off half of the four spaces?
BOY: Yes.
SOC.: And how many spaces [i.e., half-squares] are there in
this section? [I.e., in DBFH, Fig. 9]
BOY: Four.
SOC.: And how many in this? [ARDe, Fig. 10 == ABED, Fig.
9]
BOY: Two.
SOC.: And four is how many times two?
BOY: Twice.
SOC.: And this space [DBFH, Fig. 9] is of how many feet?
BOY: Of eight feet.
SOC.: And from what line do you get this figure?
BOY: From this.
SOC.: That is from the line which extends from corner to cor-
ner of the figure of four feet? [DB, Fig. 9]
BOY: Yes.
SOC.: And that is the line which the learned call the diagonal.
And if this is the proper name, then you, Meno's slave, are pre-
pared to affirm that the double space is the square of the diagonal?
[An alternative illustration of this is given in Fig. 11.]
BOY: Certainly, Socrates.
 THE "MENO" DISCUSSION: FIGURES FROM THE SCHOLIA

(Figures 3 through 11 from Greene, Scholia PIatonica, pp. 171-73.

Figure 8 and Roman lettering are my additions. Nonfunctional numbers
2', 6' are omitted in Figure 5 and nonfunctional l' marks in Figure 6.)

Figure 3 Figure 4
A..- .-.otI
B
A ,.-.. ....;Gr- .-.. B

2' E1 - - - - - - - 1 - - - - - - 1 F 21 E t--------t F

c H o
2' c o
I'
INITIAL FOUR-FOOT SQUARE CALCULATING SQUARE AREA
Area = 2' x 2' = 4 sq. ft. Area = 2' x l' = 2 sq. ft.
Figure 6

4'
A,..--_ _-..;;;;B~-_ ___. E
Figure 5

4 CJ------=D~_--..... F

SHAPE OF AN "OBLONG"

G H
FIRST CONSTRUCTION: SQUARE WITH
AREA OF SIXTEEN SQUARE FEET
Whole = 16 sq. ft.

27
 Figure 7 Figure 8

3'_"'"ToB_ _---. C
Ar-_ _....-__.
A..- .,;y.B_--..E

3'
c 0
01-----+----t------1 G II I
t
I I
I I
I
:
F G
E F
SECOND CONSTRUCTION: SQUARE WITH DElvIONSTRATION OF AREA OF
AREA OF NINE SQUARE FEET NINE-FOOT SQUARE
Area = 9 sq. ft.
The area of the 3 x 3 square is cOlnputed by sumnling the unit
squares. ABeD has already been shown to equal four of these, and
the added gnomon, CBDEFG, contains five more.

Figure 9

8
A C
First Square
Original Area 2
1
"
Added

E
0 F
Second Square Filled in
3 "-
Added
/
Square 4

G'-------~-----'"
H
FINAL CONSTRUCTION: FOUR FOUR-
FOOT SQUARES JOINED, WITH DIAGO-
NALS DRAWN

28
 Figure 10 Figure 11

A_----.B

6 7
o
DEMONSTRATION THAT
DIAGONAL DIVIDES UNIT
SQUARE INTO TWO
HALF-UNIT TRIANGLES FINAL SCHOLION FIGURE

The numbers as they stand do not make sense. Figure 12, follow-
ing, represents a suggested emendation of this figure which makes
it interpretable as an alternative intuitive demonstration of the
relative area of squares on the diagonal and the side.

Figure 12

2 4

6 8

PROPOSED EMENDATION OF
FINAL SCHOLION FIGURE

29
30 Plato's Mathematical Imagination

The final diagram represents an original excursion of the

scholiast's own; he is advancing a figure for an alternative proof
of the general theorem that "the square on the diagonal is double
the square on the side." 21 At some point, this final figure has been
corrupted in transcription, so that its numbering no longer repro-
duces this original intention. The intended proof proceeds, as does
Plato's, by computing areas through summation of equal triangles;
eight unit triangles are combined to form a square, abou! which,
by adjoining eight more unit triangles, a second square with sides
equal to the diagonal of the first is circumscribed. A second figure
is subjoined, showing that each triangle equals ~ the unit of
square area, if its side == I (Fig. 10, preceding). The area of the
inscribed square is indicated by numbers summing up the number
of the half-unit triangles through its four quadrants: These num-
bers should therefore be 2-4-6-8, rather than the 3-4-6-7 which
are in the manuscripts. 22 The areas of the two squares are thus
~ X 8 == 4, and ~ X 16 == 8, as in the preceding figure, which
corresponds to Plato's example. The diagram does not correspond
to any figure in Euclid, but resembles figures associated with a sug-
gested early method of proof of Euclid I. 47. It would be of some
possible interest to inquire f.urther where the scholiast found this
particular diagram.
Figure 13
LATER REPRESENTATIONS OF THE MEND FIGURE

G L H

B ~--t--........--t--~ M

EI""----.r----I

A o N K
JOWETT'S FIGURE HEATH'S FIGURE
(Dialogues, I, 362) (History, I, 298)
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 31

In Jowett's figure, stages are not labelled; Heath's line AN is the

side of the three-foot square which represents the slave boy's second
attempted solution. Either figure gives a clear picture of the dia-
gram as it develops during the discussion, by stages which the
scholia represent separately.

The purpose of this demonstration in context is to convince

Meno that knowledge is recollection. Meno's association with
Gorgias has lead him to accept a Sophistic notion of teaching
and learning which identifies knowledge with literal recall of
past experiences and precepts. A. E. Taylor has shown that
the Sophistic notion of virtue as a LEXV'Y) WOllld imply that it
could be learned and taught in this dogmatic way. It is this
conviction that "teaching" is dogmatic statement that causes
Meno to agree that Socrates did not "teach" anything to the
boy, though later educators not in the Sophistic tradition have
repeatedly cited this passage as an example of teaching pro-
cedllre. Socrates, convinced that learning is a matter of insight,
confronts Meno with a proof that if, as Meno maintains, all
knowledge is past experience recollected, he must postulate
both a very peclliiar "experience" and a very extended "past."
The argument need not be taken as proof of a literal belief
on Plato's part, bl.lt makes equal sense read as his development
of the implications of the Sophistic position along a line which
shows that even that position requires a theory of ideas if all
cases of knowledge are to be explained. (The converse form of
the same point is developed in the Theaetetus where no identi-
J

fication of knowledge with temporal psychological process can

be found that will apply to mathematical error and insight.)
For this demonstration, Socrates needs an example that will
be something the slave boy clearly has never learned, bllt it
IIlust also be capable of being presented in a form intuitively
evident enough for Meno and the slave to understand.
To avoid any possibility that the construction is one which
the boy may have learned, Socrates chooses an example from
the field of "higher mathematics" to convince Meno. The prop-
erties of the incommensurable were one of the fields of ad-
32 Plato's Mathematical Imagination

vanced mathematical research in Plato's time. The theorem

chosen has the added advantage that when the correct construc-
tion is completed, its truth is so evident from the diagranl, that
even a novice like Meno can see the demonstrated relation-
ship. An example from quantum mechanics would be analogous
in difficulty in a contemporary dialogue, as one from calculus
would have been in Leibniz' time. This theorem, however,
can be proved intuitively by a construction which is clear with-
out any special mathematical training. One cannot infer that
since the problem Socrates chooses is "advanced," he assumes
!\IIeno to have some technical mathematical education. This
fact is highly relevant to the mathematical passage in the Meno
discussed in the following section, since many of the interpreta-
tions proposed seem to require that Meno be a competent
mathematician.

III. THE GEOMETER'S METHOD OF HYPOTHESIS

(Meno 89 *)
". . . as regards a given area, whether it is possible for this
area to be inscribed in the {orin of a triangle in a given circle.
The answer might be "I do not yet know whether this area is such
as can be so inscribed, but I think I can suggest a hypothesis
which will be useful for the purpose; I nlean the following. If the
given area is such as, when one has applied it (as a rectangle) t
to the given straight line in the circle (the diameter), it is defi-
cient by a figure (rectangle) similar to the very figure which is
applied, then one alternative seems to me to result, while again
another results if it is impossible for what! said to be done to it.
Accordingly, by using a hypothesis, I am ready to tell you what
results with regard to the inscribing of the figure in the circle,
namely, whether the problem is possible or impossible."
• Trans. Heath, History, I, 299. .
t If the reader will accept Heath's (undoubtedly correct) identification of "the
line in the circle" as the "diameter," and tentatively disregard the parenthetical
references to rectangles, which represent Heath's scholion, derived from the later
technical sense of the tenns "applying" and "falling short," the translation has
suitable neutral accuracy. The interpretative slant represented by the references
to "rectangles" is discussed in the comment following Figure 15.
 EXA~{PLES :FROM PURE J\IATHE~fATICS OF ~IETHODS AND CLASS-RELATIONS 33

Scholia *
~lENO 86E
[Socrates is speaking of] the inscription of the triangle in the
circle.
MEND 87A
concerning the inscription of the triangle in the circle.

At present, apparently from the conviction that Plato. was

expert in pure mathematics and therefore must have introduced
many allllsions to specific mathematical problems in his writ-
ings, scllolars seem to be tending to make the objects of these
allusions more and more complex and mathematically ad..
vanced. Although this shows a comn1endably high opinion of
Plato, it is also a little reminiscent of tIle Pythagoreans' attrib..
uting all new discoveries to Pythagoras himself.
Three main interpretations of this passage have been ad-
vanced and defended as though they were mutually contradic-
tory.
Benecke and 'Gow 23 suggest that the problem is illustrated
with the same two-foot square that was used in the earlier dis-
cussion with the slave boy; in which case the inscription is
possible if the radius of the given circle equals the side of the
given area, represented as a square. They urge that a more
technical construction would be out of place, since it would in-
volve an inconsistency in the character of Meno, who is not a
mathematician, and to '\vhom the problem should therefore be
presented in intuitively evident form. One could well cite
Socrates' tempering of the rigor of llis definition of odd and
even number to the capacities of Euthyphro as an analogous
concession to a nonmathematical companion. The geometer's
condition then is: "If, when you erect on the diameter a square
equal to the area, an identical square can be constructed on
tIle remaining segment, the inscription is possible."
Heath and others prefer a more general theorem as the in-
tended ilillstration; if tl1is method is illustrated by a contempo-
rary geometric problem, they feel it should be less trivial than
... Greene, Scholia Platonica, pp. 480, 173.
34 Plato's Mathematical Imagination

the theorem involved in Benecke's interpretation. 24 In fact,

the problem of inscription in Plato's time could have been
presented in just this way, as possible under certain determinate
conditions, impossible under others. The inscription is possible
if the area can be applied to the diameter as a rectangle in such
a way that its corner point lies on the circlImference of the
circle; and, by the converse of Euclid iii. 35,25 this will happen
if the rectangle on the remaining segment of the diameter is
similar to the one applied. This gives the theorem more gener-
ality and more mathematical interest than Benecke's interpre-
tation, but at the expense of (a) presupposing more background
in geometry than Meno can consistently have and (b) dis-
connecting this figure from the ones Socrates has already drawn,
whereas his references to "this triangle," "this area," etc., seem
clearer if we allow some stIch connections. 26
Demme and, by implication, Taylor 27 believe that the iden-
tity of the theorem is irrelevant; the needs of the dialogue are
nlet if this statement is "the sort of thing that a geometer might
say." While there is no doubt that this is essentially true, it
seems likely that some specific figure was intended to illustrate
the abstract statement of method; and, in a context in ,vhich
Socrates has been illustrating his points by constructing figures,
it is reasonable to expect that he is constructing or referring
to a figure of some sort here also.
In the first place, one may certainly allow the correctness of
the third set of interpretations insofar as the mathematical
detail of the illustration falls outside of its dialectical relevance
(except for a point discussed in Appendix B), in a way typical
of such Platonic illustrations of method. It is less plausible to
assume that there was no intended technical mathematical
reference at all.
The figure of Benecke's choice represents a figure for the
simplest case of the more general theorem. In Euclid iii. 35,
the case of two intersecting lines in a circle which coincide
with perpendicular diameters is treated separately as the initial
case in the proof of the theorem. 28 There is no reason why
Socrates, if he had the more general problem in mind, should
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 35

not have presented to Meno tIle most intuitively evident special

case from a cOlnplete demonstration which Meno could not
have understood. This suggestion combines the merits of
Benecke's insistence on consistency of character and economy
of imagery, and the suggestion of the second group of inter-
preters, that some more general theorem is required to make
the example mathematically interesting or meaningful. This
case is thus analogous to the Euthyphro definition of odd and
even number by associated intuitive configurations; only so
much of a technically valid mathematical illustration is intro-
duced by Socrates as his audience can consistently be shown to
grasp.
In summary it seems tilat at least three conditions are met
by this illustration, of which each interpreter has emphasized
one. The passage must be "the sort of thing a geometer would
say," i.e., a statement illustrating a general method, quite apart
from the exact figure intended. It must, or at least should, be
something that would hold some intrinsic mathematical inter-
est, deriving from the relation of the method to the illustration
chosen (an interest that certainly would attach to the inscrip-
tion problem, where the conditiollS but not the solution were
known). It must also be dramatically appropriate, in not ex-
ceeding the posited ability of the audience to whom it is ad-
dressed and, if possible, in developing from antecedent figures
used for methodological ilillstration in preceding context. If
we take Benecke's figure, treated as a special case of the general
theorem of inscribability, the mathematical interest of the gen-
eral problem is still suggested by this special case, though con-
sidered as a theorem in its own right, apart from that context,
it is not very significant or mathematically interesting. The
general theorem, to gain its relevance to the contextual dis-
cussion, must itself be taken as a special case of a method used
by the geometer at work, and one may be able to appreciate that
technique without laboring over the detail of a postlliated,
specific example.
Still, there are many diagrallls which would satisfy all three
of these conditions. In Appendix B, following, a further con-
36 Plato's Mathematical Imagination

sideration of the intuitive connection of pure mathematical

examples to the conceptual problems in their contexts will be
presented, with an analysis to show why, out of the possible
set of instances possessing interest, appropriateness, and meth-
odological relevance, the inscription theorem ,vas the specific
example here chosen.
Figure 14

c D F

A ~-------"--------'-'B

THE "MENO" HYPOTHESIS: BENECKE'S INTERPRETATION

(A. Benecke, Ueber die geometrische Hypothesis in PIa tons

ltMenon" [Elbing, 1867])

This interpretation identifies the area to be inscribed with the

square used in the previous demonstration (so that A CDE in the
present figure is identical with ABED in Figure 9). As Socrates
refers to inscribing "this area," this interpretation visualizes him
as turning again to the other diagram and pointing to ABED. If
he drew a circle at Meno 74E, while discussing "roundness," he
might also have pointed to the circle in stating the present problem.
In this figure, if /lE is less than the diameter of the circle, AB ~
by an amount equal to itself, then a square BEDF equal to the
square A CDE can be constructed, and the area contained in the
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 37
triangle ADB} formed by drawing the diagonals AD, DB} equals
. 4CDE, inscribing the area as a triangle. The inference which shows
~4.DB equal to A CED is the same summing of half-unit triangles
used in the earlier demonstration with the slave boy.

Figure 15

l'
;

THE "MENO" HYPOTHESIS

(Heath, History, I, p. 300)

"In order, therefore, to inscribe in the circle an isosceles triangle

equal to the given area (X), we have to find a point E on the circle
such that, if ED be drawn perpendicular to AB, the rectangle
AD-DE is equal to the given area (X) ('applying' to AB a rectangle
equal to X and falling short by a figure similar to the 'applied'
figure is only another way of expressing it).... This is an equation
of the fourth degree which can be solved by means of conics, but
not by means of the straight line and circle."
If A CED is similar to DEFB} DE is a mean proportional between
AD and DB, and ACED can be inscribed as the triangle AEH.
Since AB is the diameter, DE equals DH, and DE-DH = DE2 ==
AD·DB. Conversely, if this condition is satisfied, E represents a
point on the circle of which AB is the diameter.
38 Plato's Mathematical Imagination
Figure 16

At-------t--------tC

B
F'IRST CASE OF EUCLID III. 35

The theorem is that if two straight lines intersect within a circle,

the area of the rectangle bounded by the segments of one is equal
to that of the rectangle bounded by the segments of the other. The
converse of this would establish that if, in the figure, ED is a mean
proportional between AE and EC and if ED is perpendicular to
AC~ points A)e) and D will lie on the same circle, regardless of
whether E lies at the center of the circle.
The first case: If both lines pass through the center of the circle,
the proof is evident; for, all four radii being equal, the rectangles
bounded by them are equal as well.
Compare with Benecke's diagram, Figure 14, preceding.

IV. DEFINITION OF ROOTS AND SURDS

Theaetetus 147 •
THEAET.: Yes, Socrates, there is no difficulty as you put the
question. You mean, if I ani not mistaken, something like what
occurred to me and to my friend here, your namesake Socrates,
in a recent discussion.
• Trans. Jowett, Dialogues, IV, 199-200.
 EXAMPLES FROM PURE MATHE~IATICS OF METHODS AND CLASS-RELATIONS 39

soc.: What was that, Theaetetus?

THEAET.: Theodorus was writing out for us something about
roots, such as the roots of three or five, showing that they are
incommensurable by the unit; he selected other examples up to
seventeen-there he stopped. Now as there are innumberable
roots, the notion occurred to us of attempting to include them all
under one name or class.
SOC.: And did you find such a class?
THEAET.: I think that we did; but I should like to have
your opinion.
SOC.: Let me hear.
THEAET.: We divided all numbers into two classes: those
,vhich are made up of equal factors multiplying into one another,
which we compared to square figures and called square or equi-
lateral numbers;-that was one class.
SOC.: Very good.
THEAET.: The intermediate numbers, such as three and five,
and every other number which is made up of unequal factors,
either of a greater multiplied by a less, or of a less multiplied by
a greater, and when regarded as a figure is contained in unequal
sides;-al~ these we compared to oblong figures, and called them
oblong numbers.
SOC.: Capital; and what followed?
THEAET.: The lines, or sides, which have for their squares
the equilateral plane numbers, were called by us lengths or mag-
nitudes; and the lines which are the roots of (or whose squares
are equal to) the oblong numbers, were called powers or roots;
the reason of this latter name being, that they are commensura-
ble with the former (i.e., with the so-called lengths or magni-
tudes) not in linear measurement, but in the value of the super-
ficial content of their squares; and the same about solids.
SOC.: Excellent, my boys; I think that you fully justify the
praises of Theodorus, and that he will not be found guilty of
false witness.
THEAET.: But I am unable, Socrates, to give you a similar
answer about knowledge, which is what you appear to want; and
therefore Theodorus is a deceiver after all.

The use of incommensurables to illustrate the proper method

of definition in the Theaetetus is certainly intended as a me-
40 Plato's l\tlathematical Imagination

morial to what was one of Theaetetus' important mathematical

achievements. The memorial, however, has been selected with
a view to its appropriateness in the context of the reported
conversation. The moral of the 1vhole dialogue is one that
might well be symbolized by a theorem of il1commensurability;
knowledge turns alIt, "vhatever unit of comparison we employ,
to be incommensurable with opinion. Various items of knowl-
edge, particularly the mathematical, cannot be identified with
or explained by any process of perception or physical construc-
tion. The relation of knowledge to opinion cannot be described
simply by identifying items of knowledge and opinion, but
must be stated in some other way. What is needed is to find
some common classification for all these items of knowledge,
and to find a statement of their relation to our perceptions and
habits of operation, as Theaetetlls and Young Socrates have
defined and stated the relations of magnitudes and roots.
This example has a mathematical reference which claimed
great interest from Plato, and continues to attract the attention
of historialls of mathematics. Its relevance to the dialogue itself
is only that of the other mathematical examples introduced as
paradigms of definition in the Meno and Euthyphro. 29 A good
definition ffillSt state a principle of classification, not simply
enumerate various individual cases. The development of the
treatment of irrationals, in which Plato and the Academy took
a special interest,30 provided an excellent geometrical example
of this difference. Only so much of the detail of this history as
is immediately relevant to the dialogue is included by Plato,
but this much gives almost the only clue ,ve have for filling in
a series of technical developments not discussed fully.
The proof of the irrationality of the diagonal of a unit square
cited by Aristotle,31 based on the demonstration that if the side
and diagonal are assumed to be commensurable, the same num-
ber must be both odd and even, is usually taken as a Pythag-
orean starting-point in the investigation of irrationals. With
the work of Eudoxus, carried out in collaboration with Plato,
a complete theory of irrationals, that contained in Euclid's
Elements} enters the history of geometry. Somewhere in the
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 41

intervening period, the investigations and proofs must have

been extended and generalized. But the present passage in
Plato, with its confirmation by a scholion on Euclid x.g, is the
only source of insight into this intervening period of explora-
tion and generalization.
Historians agree that the procedure of Theodorus seems to
have had the peculiarity that each case needed to be demon-
strated separately, and that what impressed Plato most in
Theaetetus' procedure was his generalization of this episodic
technique. From the theorems and methods of inquiry available
to tIle mathematician in Theodorus' time, conjectural recon-
strllctions of his proofs have been proposed, falling into three
main groups, of which one seems to fit the evidence of Plato's
account better than either of the others. 32
(1) It has been suggested that since there was available a
technique for successive approximations to the numerical values
of irrationals (the use of side- and diagonal-numbers, explained
by Theon, and attributed by him to pre-Platonic Pythago-
reans),38 Theodorus may have employed an extension of this
method. The principal objection is that such a line of inquiry
would establish only a presumption, not a proof, of the un-
ending character of the approximation, hence would not really
establish incommensurability. With the Pythagorean demon-
stration of the irrationality of the square root of two as a model,
Theodorus would hardly have been satisfied with a proof which
was so mllch less effective, nor could we explain Theaetetus'
apparent satisfaction with the validity of the proofs of the sepa-
rate cases which Plato's dialogue makes him express.
(2) The older Pythagorean proof lends itself to an extension
to the cases investigated, showing that for each irrational, the
assumption of its rationality (when it is represented as the ratio,
min, of two integers) leads to the contradiction that m is and
is not divisible by a given number. 34 This procedure probably
did occur somewhere in the transition period from Pythagoras
to Eudoxus, but its assignment to Theodorus has been chal-
lenged (and, it '\vould seem, conclusively so) on the ground that
this method would already have the generality which, according
42 Plato's Mathematical Imagination

to Plato's report, is just what Theodorus' procedure lacked.

(3) Zeuthen and Heath 35 suggest a geometrical demonstra-
tion of a type perhaps discovered and certainly most simply
illustrated by the attempt to find the greatest common measure
of a line cut in extreme and mean ratio. When the shorter seg-
ment of such a line is laid off on the longer, the difference is
itself in the same ratio to the shorter segment that the latter
had to the longer. Since the two initial segments do not have
an integral ratio, it follows that no common unit will be ap-
proached by the difference of the two segments, no matter how
often the one is subtracted from the other. This satisfies both
conditions attaching to Theodorus' proof: (a) that it is a valid
proof of incommensurability, not merely an empirical illustra-
tion of a presumption; and (b) that it requires a separate geo-
metric figure for each case, such that the same ratio can be
shown (by similar figtlres) to repeat for every subtraction of
the shorter segment (an integral length line of the figure)
from the longer (representing the given constructed irrational).
The final form of Theaetetus' generalization of Theodorus'
results was probably the source of Euclid's theorem x.g, which
is attributed specifically to Theaetetus. This connects the ratios
of the sides of similar figures with previously demonstrated
arithmetical theorems about the ratios of square numbers. 36
However, there were probably several intermediate stages
in Theaetetus' work before he arrived at this final algebraic
generalization, and Plato's passage may recall such an interme-
diate development, which supplied in a more typically geo-.
metric form the generality lacking in Theodorus' approach.
Such an intermediate step may have consisted in the application
of the extension of the Pythagorean proof, mentio11ed above as
one of the conjectured reconstructions of Theodorus' method,
to any figure in which some standard construction established
the incommensurability of the relation of diagonal and sides.
The proof would then have been developed as applying to the
square root of any number, the geometric representation of
which as an area would have been possible only under the given
conditions of construction.
 EXAMPLES FROM PURE MATHEMATICS OF METHODS AND CLASS-RELATIONS 43

Figure 17

A PROOF ATTRIBUTED TO
THEAETETUS: EUCLID x.9
(Heath, Euclid, II, pp.
29-31)

"The squares on straight lines comlnensurable in length have to

one another the ratio which a square number has to a square num-
ber; and squares which have to one another the ratio which a
square number has to a square number will also have their sides
commensurable in length. But the squares on straight lines incom-
mensurable in length have not to one another the ratio which a
square number has to a square number; and squares which do not
have to one another the ratio which a square number has to a
square number will not have their sides commensurable in length
either.
"For let A,B be commensurable in length; I say that the square
on A has to the square on B the ratio which a square number has
to a square number.
"For, since A is commensurable in length with B~ therefore A
has to B the ratio which a number has to a number (X.5).
"Let it have the ratio which C has to D.
"Since then, A:B::C:D, while the ratio of the square on A to the
square on B is duplicate of the ratio of A to B~ for similar figures
are in the duplicate ratio of their corresponding sides; (VI.20, Por.)
and the ratio of the square on C to the square on D is duplicate
of the ratio of C to D, for between two square numbers there is
one mean proportional number, and the square number has to the
square number the ratio duplicate of that which the side has to
44 PZa to's .AJathematical Imagination

the side; (VIII. I I) therefore also, as the square on A is to the square

on B, so is the square on C to the square on D.
"Next, as the square on A is to the square on B, so let the square
on C be to the square on D; I say that A is commensurable in
length with B.
"For since, as the square on A is to the square on B, so is the
square on C to the square on D, while the ratio of the square on
A to the square on B is duplicate of the ratio of A to B, and the
ratio of the square on C to the square on D is duplicate of the
ratio of C to D, therefore also, as A is to B, so is C to D.
"Therefore A has to B the ratio which the number C has to the
number D; therefore A is commensurable in length with B.
"Next, let A be incommensurable in length with B; I say that
the square on A has not to the square on B the ratio which a square
number has to a square number.
"For, if the square on A has to the square on B the ratio which
a square number has to a square number, A will be commensurable
with B;
"But it is not; therefore the square on A has not to the square on
B the ratio which a square number has to a square nurrlber.
"Again, let the square on A not have to the square on B the
ratio which a square number has to a square number; I say that A
is incommensurable in length with B.
"For, if A is commensurable with B, the square on A will have
to the square on B the ratio which a square number has to a square
number.
"But it has not; therefore, A is not commensurable in length
with B.
"Therefore, ....

"A scholium to this proposition (Schol. x No. 62) says categori-

cally that the theorem proved in it was the discovery of Theaetetus."
 Part Two

MATHEMATICAL IMAGES CLOSELY

DEPENDENT ON THEIR DIALECTICAL
CONTEXTS

...
 CHAP'TER II

~~Social Statistics": Arithmetic Detail

I. ATLANTIS AND ITS INSTITUTIONS

IN HIS description of Atlantis in the Critias) Plato gives the exact

numbers and measures of almost every phase of its geography,
public works, and political institutions. In the description of
ancient Athens, in the same dialogue, there is only one numeri-
cal detail given (the total fighting strength of the state).1 This
suggests that the use of such specific figures is a device peculiarly
appropriate to a description of the Atlantean state and that the
specific figures which Plato invents have some characteristics
intended to reflect peculiarly Atlantean principles of legislation
and technology.
The institutions and customs of ancient Athens can be and I' ,

are adequately specified by reference to a normative standard

embodied in legislative principles; the exact measurements can
be summed up by the statement that they are those which are
best adapted to proper functioning. In a disordered and only
loosely unified state such as Atlantis, on the other hand, insti-
tutional and technological details are not determined and co-
ordinated by a rational unifying plan. The closest analogue to
the structural statements made about ancient Athens (where
the structure of the society was organized around rational legis-
lation) in an account of Atlantis is, therefore, the separate
description of the institutions and public works of which this
social structure happens to be composed.
The substitution of some set of specific figures for considera-
tions of proper function in a total plan is peculiarly appropriate
to the description of the type of disunity and disorder which
47

.1:
,t8 Plato's A1athernatical Imagination

.t\.tlantis illustrates. That the specific figures give11 have in com-

mon an arithmetical characteristic ,vhicll emphasizes Atlantean
irrationality and confusion "\tvill be shown in the discussion
"vhich follows.
Poseidon seems to have been an ancestor not likely to produce
philosophic and mathematically minded offspring; for, if we
compare his ordering of circles of lalld and sea in Atlantis to
the circles of the heavens described in the Timaeus) it becomes
evident that, when this god geometrizes, he does it like a car-
penter's apprentice. And tIle institutions preserved by the de-
scendants of Poseidon 'VI10 rule Atlantis SilOW that, in fact, the
offspring have made no improvement, philosophically or mathe-
matically, on the insight of their ancestor. The key to the selec-
tion of all the numbers in the Critias is the statement that these
rulers "met alternately every fifth and every sixth year, paying
equal llonor to the odd and to the even." 2 That this sho"\vs a
total and fundanlental lack of llnderstanding of the nature of
nllmber is clear if this passage is compared with the careful
distinction of kinds of sacrifices which SllOllld be made in odd
and those which should be made in even numbers, in Laws
717A. Not only is the confusion of even and odd (which are
the basic contrary principles of the most elementary mathemati-
cal science) a sign of total lack of theoretical ability, but the
specific numbers cited here, which represent the even and the
odd, reflect this same confusion, the one being the sum and
the other the product of the first odd and the first even number.
Since in Plato's mathematical images and formulae in dialecti-
cal and cosmological contexts the basic opposition of odd and
even is observed and since in contexts dealing 1;vith legislative
detail ease of nlanipulation or religious propriety is the deter-
mining factor (and in the latter case the basic distinction is
again that of odd and even), while in mythical contexts periods
and distances are poetically dismissed as "myriads" (perhaps
composed of lesser, proportionately related periods, which are
indicated by smaller powers of ten), the absence of anything
remotely resembling "alternating fives and sixes" ill other Pla-
tonic contexts is causal, not accidental. The choice of "five and
 "SOCIAL STATISTICS": ARITH.~IETIC DETAIL 49

six alternately" by the Atlantean kings is llot only an example

of mathen1atical igll0rance on so grand a scale that they canl10t
distinguish the natures of the odd and the even but also a Sigll
of a lack of rational statesmanship so great that no real principle
of any sort is observed in the fixing of these meetings of the
rulers, the state's most important political and religious festival.
Reflecting and leading up to this final detail, where an ex-
plicit statement is given of the underlying confusion which
accounts for its selection, all the other numbers and measure-
ments cited, however casually, are (except one) either (a) multi-
ples of 6 or 5 or (b) parts of a sum, product, or ratio which
ill its entirety is a multiple of 6 or 5 or (c) 6 or 5.
Poseidon himself begat five pairs of twin sons; 3 his statue
depicts him driving six horses; 4 his engineering consists in the
construction of five circles (three of sea and t'vo of land) 5 about
a central island with a dianlcter of five stades. 6 Further, the
total widths of the circles of sea are to the total widths of those
of land in the ratio of 6:5. 7 (This consideration of total widths
is relevant, since precisely this type of relation gives the adum-
brated geometrical structure operative in Plato's assignment of
relative sizes to the rims of the wllorls described in the Myth
of Er.) 8
The divinity of Poseidon's 11ature reveals itself only in the
ease ,vith which he performs his tnechanical operations: he
established the circles of sea and land "with ease, as a god
might." 9 His descendants, like him in this respect, created
public works of a magnitude that appeared incredible,1° but
they '\Tere also like their ancestor in their partiality for 6's and
5's. "Vorking with a plain (,vhich was oblong and crooked, not
perfectly square or straight) 6,000,000 square stades in area,11
they constructed an intersecting network of canals. 12 The outer-
most canal of this net\vork, encircling the plain, had its breadth
related to its depth in the ratio of 6: I. The total length of this
ditch was 10,000 stades (since its sides were, respectively, 2,000,
~~,OOO, 2,000, and 3,000 stades long).13 In these dimensions and
details a 6: 1 ratio is used; and 6 is represented as the product of
2 and 3, 10 as the sum of 2, 3, 2, and 3. ConfllsioIl of the even
50 Plato's Mathematical Imagination

and the odd in these numbers is reflected, in the first two cases,
by the alternatives of multiplying and adding the first even and
first odd number. In this context, therefore, the representation
of 10 as a sum of 3's and 2's is not really an exception to the
rule of the prominence of 6'5 and 5'5.14
The military arrangements of Atlantis 15 afford a remarkable
array of 6'8.
As we survey these two sets of figures, a second principle of
selection is also seen to be operative: the vastness of the numbers
and distances involved is reflected by the prominence of myriads
as units of description. The ratios which give a qualitative
aspect and dialectical point to these various precise statements
of distances and numbers, however, remain, no matter what the
scale, 6's and 5's.
The way in which this principle is carried out in the principal
religious-political festival of the state has already been shown.
It is further represented, however, in the law that no king may
be sentenced to death without concurrence in the sentence of
at least six of the members of the council. 16 In the state religion,
the six steeds in the statue of Poseidon carryon this principle;
and the dimensions of Poseidon's temple are stated in such a
way that, while their basic ratio is 2: I, the immediate reduction
of stades to JtAE{}QO, which the form of statement suggests, would
yield a 6:3 ratio instead. 1T
The 11umber of Nereids (100) 18 emphasizes the continuing
influence of the cult of Poseidon, the Atlantean familiarity with
the sea, and perhaps also their tendency to make everything
bigger than they should. These are the factors emphasized by
Plato's parenthetical remark in 116£, which underscores the
deviation from tradition. But it is a deviation which merely
doubles a set of 5 times 10 and which, in conjunction with the
contextual mention of the number of steeds, shows in the state
religion the same basic confusion that is reflected in the alterna-
tion of 6's and 5'8.
In summary, the apparently random numbers so liberally
interspersed in Plato's account of the Atlantean state are not
inserted simply to give an impression of great size or simply to
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 51

create an effect of artistic verisimilitude (though, in fact, they

do perform both these fllnctions). These "random" numbers
are constructed on a dialectical al1d artistic principle in such a
way that each reflects some aspect of the rulers' basic and tradi-
tional confusion in mathematics and philosophy. Plato's selec-
tion of these objective metric statements of structural details of
i\.tlantean politics, public works, and geography illustrates, and
adds further insight and precision to, his eloquent disapproval
and con.demnation of Atlantis as a whole. Plato reveals his philo-
sophic and artistic precision and his sensitivity to the signif-
icance of detail in inventing the history of a bad state as well
as in describing the archetype of a good one.
One of the most peculiar of the Atlantean mathematical de-
tails is the 2: 1 ratio of the ground-plan of Poseidon's temple.
Although, as was pointed out above, these dimensions are stated
as 6: 3 plethra, it seems odd that Plato would even by such tenu-
ous implication criticize what is elsewhere his own favorite ratio
of 2: 1. However, the actual dimellsions of Greek temples suggest
a possible explanation.19 Though we have asserted that Plato
was not himself a skilled artisan, the fact remains that l.te was an
interested observer who found the raw material for many of his
analogies in the arts and crafts. Since the range of these illus-
trations includes fish-trapping, shipbuilding, spinning and weav-
ing, dyeing, potting, fulling, the minting of money, statue pai~t­
ing, theatrical scene-design, and other equally diverse activities,
there is no reason to believe that this interested observation did
not also extend to architecture. It must have seemed to Plato a
striking oddity of architectural design that the typical grollnd
plan of the most beautiful temples had sides that were not in
simple ratio, and that may, indeed, have been planned by a con-
struction which made them incommensurable. The aesthetic
effect of buildings planned out by simple geometrical construc-
tions which involved incommensurables may well be one of the
sources inspiring Plato's remarks in the Protagoras that the same
proportions are not always characteristic of pure functionality
and of aesthetic effect. 20 But the aesthetic value of the effect
was not lost on Plato; for the temple, like the religious myth,
52 ])lato's Jfathenultir:ol Im.agination

llas a function to which the pleasure of its appearance IS

relevant.
A temple based on the simple 2: 1 grollnd plan WOllld have
seemed to an architect of Plato's time (as it does to one today)
a dull, boxlike structure, 110t suited to its function of presenting
an object of beauty and apparent harmony to the ,vorshipper.
In his judgment of this temple, that "tllere was something
barbaric in its appearance," Plato may well be thin.king not
only of the garish external application of orichalcum and gold,
bllt also of the lack of aesthetic sul)tlety revealed in the arclli-
tectural ground plan and elevation. It is the sort of unsubtle
product one might expect of a Milesian army el1gineer, tem-
porarily turned temple-architect; and the same insensitivity to
tIle proper social integration of function that characterizes the
public works of the i\tlanteal1 engineers is reflected in their
religious architecture also.
It is interesting to speculate further on the possible relation
of architectural procedure to Plato's "diagonal of five" in the
nuptial number passage. 21 Hambidge's measurements suggest
that lines in ratio 1: V 5" may have been a sufficiently familiar
bllilder's line to have had some special designation. 22 Such
technical familiarity with the quantity, however, though it
could have helped assure Plato that the line he referred to
would be recognized, does not seem in any way to explain the
specific nomenclature that Plato has chosen.

The details of Figure 18 should be compared with Plato's pre-

ferred city planning in Laws 778C, where it is prescribed that tIle
city shall have no wall, in contrast to the 4 inner walls and to the
outer wall (50 stades in radius; not shown in Figure 18) of the
greater city of Atlantis. The tunnels permit shipping to come to
the very center of the city, aggravating all the disorders and dis-
advantages which the Athenian Stranger in the Latvs describes as
typical of a seaport town. There is no separation of residence for
soldiers, as the Acropolis in the Republic or the dispersed garrisons
in the La1l.1s would be.
If this city plan, with its central island containing the king's
residence, is compared with Ortygia and its role in the history of
 HSOCIAL STATISTICS": ARITHMETIC DETAIL 53

Figure 18

II
PLAN OF THE INNER CITY OF ATLANTIS

(R. G. Bury, Timaeus) etc.} map facing p. 286)

"..A. == central island (acropolis), 5 stades in diameter, with sacred

pillar, temple, altar, etc. (116A, C fi.)
Az == inner belt of water, 1 stade wide
B == inner belt of land, 2 stades wide, ,vith temples, gardens,
barracks, etc. (117C)
B2 == middle water-belt (2 stades)-
C == outer land belt (3 stades) with hippodrome (117C), barracks,
etc.
C2 == outer water-belt (3 stades)
bb == bridges, with gates and turrets at each end, joining j\.B, Be,
CD (116A)
tt == tunnels for ships nnder Band C (115D, E)
W\V == ring-\\Talls (4), round j\,B,C, and sacrarium in A (116A)."
 ~
<
~
~

0')
~
Z
~
<
~
~
5-4

.~ r.:r.1
r.. ~
....Z
~
~0

54
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 55
Syracuse, the similarity suggests that some of the elements of Plato's
picture of a royal residence on an island within a city derive from
contenlporary affairs in Sicily. The legendary extra-Mediterranean
location is probably derived from sailors' stories, of the sort which
had earlier inspired Homer's invention of Ogygia.
The fact that both this and the ideal cities share a circular plan
underscores Plato's point that technological accuracy is no sub-
stitute for functionality.
The image in Figure 19 is a maze rather than a regular grid;
since the purpose of some of the "transverse" canals is to com-
municate with the city, I have drawn them as they would be
planned to serve that function, rather than in the regular grid
which would be reasonable if their principal function were irriga-
tion rather than water communication.
Compare Bury's map (Timaeus, etc., facing p. 285). The use of
two large primes (29 and 31) helps to reinforce the maze effect;
here the technology has completely escaped any rationale. Note
that the city, if we interpret the canal net in the present way, is
a focus of disorder; the lot plans become more and more irregular
as one gets closer to it. A good brief discussion of the rectangular
and radial theories of city planning in Plato's time is given in
H. Diels, Antike Technik (3d ed.; Leipzig and Berlin, 1924),
pp. 15-16.
Two specific observations are suggested by this canal net: the
first is the Atlantean's characteristic waste of effort through lack
of planning; the second is the similarity of the irrigation canal net
and the rectangular grid used in city planning. Hippodamas, a
Pythagorean engineer, had used such a rectangular plan in his
redesign of the Piraeus. Plato himself, as Figure' 21 shows, favored
a radial plan. As the Laws describes the planning of the state,
radiating highways were to be built establishing direct communica-
tion between the city and the frontier. One was to pass directly
from center to periphery without (as an Atlantean dependent on
his irrigation canals would) traveling along both legs instead of
the hypotenuse of a right triangle. By describing the Atlanteans
as first digging canals establishing communication by such legs of
right triangles, then adding the net representing their hypotenuses
separately, Plato seems to bring out a criticism of the Hippodamian
plan. The radial plan seems to have been thought the more
"radical"; as usual with such new ideas, Plato adopts it, and Aris-
.~6 Plato's llIathematical Inlagination

tophanes enthusiastically opposes himself to any such unconven-

tionality.
This resemblance of the canal net to a local street plan (rein-
forced by the resemblance of the fortifications planned by the
Atlanteans and by Hippodamas) suggests that the inspiration of
the Atlantis story is not taken as a unit from any single source,
but is a combination of details selected from the contemporary
scene. In that case, most of the proposed identifications of the
sources of the story are probably correct in their identification of
sonle of its elements, but wrong in assuming that all the elements
have the same source. As one recalls the criticisms that Plato offers
of Pericles' public-works program, the exaggerated emphasis on
public works in Atlantis sounds like another criticism of the
Athenian projects. In particular, the wall and tower fortifications
and the rectangular, irrigation canal grid are replicas of Hippo-
damas' city planning for the Piraeus and Thurium. The inclusion
of the "voracious elephant" among the Atlantean fauna is probably
an invitation to the reader to compare the Atlanteans and the
Carthaginians. The elephant was peculiar to Carthage in Plato's
time, and he seems to introduce it here as a "type" (a local animal
chosen by a city for the design of its coinage). The" end of the
story, the overnight disappearance of a great sea power, is so close
to the actual fate of the Minoan civilization that there probably
were traditions preserved which suggested it. Certainly the details
of Atlantean bull-,vorship were taken from legends associating such
a cult with a hostile lVfediterranean po,ver. And so throughout
Plato's description: legend and contemporary fact have been drawn
on when either suggested detail for the composite photograph of .a
state dominated by imperialism and monomania.
In specifying the landmarks which represent the extent of the
Acropolis in ancient times, Plato picks equidistant points, bringing
out the basically circular geometrical form of the old city. In con-
trast with Atlantis, the regions are connected and functionally
allocated. In describing' the borders of the country, however, the
landmarks mentioned do not seem intended to suggest a definite
geometric figure. (The shape of Attica could not easily be derived
by normal geologic processes from a prehistoric territory of pos-
tulated geometric regularity.) Instead, these early borders of the
country underscore the continuously mountainous frontier, con-
stituting a natural defense against land attack, where Plato pre-
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 57

Figure 20

o
TOPOGRAPHY OF ANCIENT ATHENS

=
A Acropolis
B = Eridanus
C == Lycabettus
D == Ilissus
E == Pnyx

sumably intended to locate the defeat of the invading Atlanteans.

The credibility of the story is heightened throughout by explain-
ing present topography as the result of natural forces acting over a
period of time on the ancient features "recorded" in the Egyptian
inscription. Whether some ancient tradition suggested an original
connection of Pnyx, Acropolis, and Lycabettus in a single plateau,
or whether the same sort of inference is used here as in defending
the historicity of Atlantis by the mud shallows in the Atlantic,
later archaeological findings support Plato's reconstruction of the
prehistoric connection of the three hills.
In Figure 21 the city is located in the most efficient way, in
respect to its territory, for purposes of administration, defense, and
trade. Unlike Atlantis, which lies at one side of its territory, and
on a visualized nlap appears as a focus of disorder, the central
city in this plan is a center of symmetry. (The suitability of the
rectangular plan for purposes of location, which makes it func-
tional in our own culture, would not be a great advantage for a
58 Plato's Mathematical Imagination

Figure 21

<] [>

A FUNCTIONAL CIVIC PLAN (Laws)

• Centrally located capital city
Radial highways
o Regional markets and temples
District boundaries
~ Frontier garrisons

community of the size of a small Greek city state, nor would its
superiority to the radial plan in this respect seem so evident to a
culture which was not pervaded by the use of Cartesian co-ordinates
for specifying location.)

-I
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 59

Figure 22
MILITARY STRENGTH OF ATLANTIS
(Conscription is based on 60,000 military districts.)
1. PERSONNEL
Archers 120,000
Hoplites 120,000
Slingers 120,000
Javelin-throwers 180,000
Light-armed slingers 180,000
Horsemen and charioteers 240,000
Sailors 240,000

Total· 1,200,000
II. EQUIPMENT
Ships 1,200
Chariots (1/6th supplied by each
allotment) 10,000
• This is exactly 60 times the force that the ancient Athenian state maintained
(Critias 112D.; this is the only statistical detail given in connection with the
description of ancient Athens).

II. THE SOCIAL INSTITUTIONS OF THE "LAWS"

a. Mathematics and the Law

From the frequent specifications of the political need for the
right order in society, one would expect the statistical details
of the Laws to present a marked contrast to the Atlantean pat-
tern. As in the Republic} legislation is described as a way of
introducing harmony of classes and institutions into a state. 23
'Unlike the Republic} however, the practicable legal statute
must combine two types of such harmony. The order of Zeus,
which apportions wealth and excellence to merit, is an order
by geometrical proportion, the only type of structure employed
in the Republic. The arithmetical ratio, on the other hand,
gives equally to all citizens. 24 In a practicable legal code, the
two orders must be blended. Also, unlike a discussion of politi-
cal principles, a statute must make specific quantitative evalua-
GO Plato's J."Wathelnatical Imagination

tions: it is not enough to agree tllat "wedding feasts must not

be excessively expensive"; the law must state how Inllch "exces-
sive" expenses are. 25
The mathematical aspect of the Laws is pronlinent ill three
cOllnections. In the first place tIle population and administrative
organization of tl1e state are determined by logistical considera-
tions ,vhich facilitate tIle mechanics of civil administration,
representation, military conscription, tax assessment, etc. (The
careful detail with which numbers suitable for this function are
chosen COlltrasts effectively with tIle il1definite bigness of the
comparable figllres for Atlantis.) The state laws governing
,veights and measures are also determined by logistical consid-
erations designed to make all transactions easy and readily cal-
culable. These considerations are based purely on facility of
calculation, an aspect of social organization not admitted to
importance in the Republic) nor present in the state of Ancient
Athens described in the Critias. (The population figures of
tl1ese two states fall rather in the group of "powers of ten"
typical of Plato's "mythical and historical" numbers.)
In the second place, the compulsory public education, de-
signed to train students for socially effective participation in the
state, puts great emphasis 011 the mathelnatical stlldy of measure
as a technique needed to evaluate social effectiveness and the
efficacy of various social implementations of the state's political
ideal. 26
In the third place, the criminal code, which must be provided
by the legislator, since these laws are designed for an actual
rather than an ideal community, involves a very intricate appli-
cation of tnathematics in its proportioning of status and pen-
alty.27 Since there is a blending of oligarchy in the make-up of
the state, possession of property is treated as involving greater
civic importance, hence greater social responsibility. Probably
one reason for including the study of mathematics in the schools
""vas to give students the background requisite for understanding
this aspect of the state laws in the required civics course devoted
to the statutes and their explication.
The pattern which these figures represent shows a conscious,

(
t
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 61

over-all plan at work, in complete contrast to the fives and sixes

of .l\tlantis. It is a plan which adopts a different rationale for
each of the three aspects of social planning: logistic for popula..
tion and representation and for property distribution laws;
commensllrability for COllrses aimed at civic planning, which
are based on geometry; and a mixture of harmonics and calcula-
tion embodied in the fines and sumptuary la,vs of the statute
code.

b. Administrative Logistic
In the LauJs~ the scorn for calclliation as an inelegant applied
form of number theory, ,vhich Socrates expresses in the Repub-
lic) 28 is replaced by an intense interest in computation. The
mechanisms of coinage, fines, taxation, calendar, standards of
,veight and measure, as ,veIl as of election, military conscription,
irrigation and agriculture, lTIUst catc}l the dialectical propor-
tions of the political philosopher in a la\vgiver's specific metric
net, a net capable of ready logistical adjustment al1d manipula-
tion to meet the demands of new social contingencies.
For purposes of representation, defense, etc., it is necessary
to subdivide the citizens into varying numbers of equal groups.
Conseqllently, in establishing the optimum population of his
state, Plato seeks, \vith all the reSOllrces of Pythagorean number
theory, for a number \vhich will be most logistically manip-
ulable.
The number finally chosen is the 1vonderful number 5,040,
,vhich has 60 divisors, including the ntlmbers 1 throllgh 12,
except I I, which will go into 5,038 (5,040-2).29
How did Plato discover the properties of this number? To
\vork this Ollt empirically with the number notation at his dis-
posal would have been an arduous pastime. 30
Presumably Plato was aware that the total number of divisors
of a number is related to the number of prime factors it con-
tains. One way of insllring a large nllmber of divisors (includ-
ing the number 12, needed to fit the calendar divisions) ,vould
have been to multiply successive digits together, beginlling with
I and proceeding until a product suitable in size for the popu-
62 Plato's Mathematical Imagination

lation of a state resulted. As a matter of fact, since 5,040 equals

1·2·3·4·5·6·7 or 71, this is probably how Plato made his dis-
covery.a! The other divisors could have then been computed
by tabulating the various products of the constituent factors.
The same logistic emphasis is evident not only in the de-
termination of optimum population but in the laws governing
standards. 32 Since 12 has 6 divisors, it is a better number than
10 for computation and fits more nearly into the duodecimal
calendar periods. s3 Consequently, coinage, measure, and weight
are all to be established on the duodecimal system. Analogously,
multiples of 12 serve to determine voting districts and repre-
sentation in the governing assembly.
Thus, though they may dominate the mythology of the state,
the mystical numbers 10,000, 100, and 10, with their Pythago-
rean aura of perfection and their religious connotation, are
superseded in actual lawgiving by the less exotic 12 and 5,040,
which are greatly superior for purposes of political manipu-
lation.

c. Mathematics in Education for Citizenship

1. Place of mathematics in the curriculum. In Laws 818A fI.
mathematics is given its place in the curriculum of public edu-
cation. It is given this place in a context in which the Athenian
Stranger, admitting that nature has not created men in the pure
dialectical types posited in the Republic, tries to lessen the
discrepancy. This is a context which might be expected to pro-
duce several shifts in emphasis, disquieting to anyone for whom
"Plato's mathematics" is the pure deductive science described in
Republic vii as suitable training for philosophers. In fact, sev-
eral of these details in the Laws have puzzled readers and schol-
ars; for example, the strong language in which Greek ignorance
of theorems of incommensurability is reprehended, the difficulty
in seeing how the distribution of a set of bowls teaches much
worth learning about mathematics (the preparation of such
special materials would be out of all proportion to their value
if the lesson is one of simple computation), and the summary,
popularized presentation of the results of astronomical research
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 63

(which seem either to ignore or contradict Plato's earlier, more

elaborate, astronomy). The central change from the view of
mathematics developed in the Republic) the fact that measure
has usurped the place of axiomatic rigor) seems not to have been
seen as a problem, but is certairlly a more striking discrepancy
than the minor deviations noted.
This training is part of education for citizenship because
the laws of mathematics are necessities against which not even
the gods can fight. 34 In ordinary civics, computation will play
a part; but, more important, all legislation and applied ethical
decision must center around finding the right measure. Measure
is the device by which institutions and materials are made
subservient to rational control; the statutory code of the state
is an attempt to establish fines, maximum expenses, etc., which
specify this right measurement. What the curriculum is in-
tended to do is to suggest tactics for social action; it is applied
mathematics. Obviously the nature of commensurability will
be vitally important in such a technological theory; human
errors very often result from not recognizing the incompatibil-
ity of alternative plans and motives, and introducing an er-
roneous compromise. 35
The study of mathematics as measure is dismissed with a
rather brief treatment for two reasons. In the first place, the
theory has been exhaustively developed in the other late dia-
logues; 36 in the second place, the legal code (study of which
precedes this course in mathematics) is an elaborate instance
and embodiment of metric tactics at work. The course in com-
plltation, geometry, and astronomy simply brings out in a more
general form tactics that have already been applied in the legis-
lation dealing with poptllation and representation (where the
techniques are logistical), with property and penalties (where
the tactics are metric), and with state religion and natural
theology (where the theory of astronomy is written into the law
itself at some length).
The specifications of the nature of this mathematical educa-
tion are of value for a study of mathematical imagery primarily
in the way in which they suggest the sort of principle likely to
64: Plato's Mathematical Imagination

underlie this state's statistical detail, and the '\lay in which their
shifted emphasis confirms generalizations about the sensitivity
of "mathematics" in Plato to the dialectical context in which
that mathematics is presented. 37
2. Astronomy in general education. As one would expect,
the general public in the state described in Laws 818A does not
study theoretic astronomy; the schools teach chronology, but
astronomical theory is studied only by the specially talented
group who constitute the state's research academy.
The general citizen is to be informed, however, that the
results of this research give scientific weight to the conviction
that the world is not haphazard, but designed and run on a
rational plan. To correct the poplliar belief that the irregulari-
ties of apparent planetary motion indicate a basic disorder in
nature, the citizens will presumably be told, as the Athenian
Stranger tells his companions, that these stars each follow "one
road, not many." Thus science discovers that these heavenly
bodies are not literally "",vanderers," as popular belief might
suggest.
This remark can cause consternation to scholars who see in
it a direct contradiction of Plato's earlier theories of astron-
omy.3S But the notion of "one path" instead of many must be
understood here as a contrast of the theory of a single planetary
orbit (the resultant of three component motions, but perfectly
predictable and regular) 'tvith the folk-belief that the planet is
a wanderer straying down many different paths. The emphasis
of the remark falls on the regularity of celestial motion, which
science demonstrates. In this context, it would be dramatically
incongruous to have the speaker suddenly intrude technical
theoretic astronomy.
Nor does it seem necessary to explain the Athenian Stranger's
remark that he had learned this only recently as evidence that
Plato must be presenting some recent discovery of his own. The
Stranger, if we study him, is like the Elders of the Nocturnal
Council; his time has been spent in administrative and practical
affairs, in which he has had considerable experience; there is,
therefore, dramatic propriety in having his own experience

f
t
 "SOCIAL STATISTICS": ARITH1\IETIC DETAIL 65

reproduce that of the members of tIle Council, '\vho in their

old age study geometry and astronomy, and on the basis of this
research and their experience give expert advice to the state.
3. The garne of distributing bowls. School children are to
be ~aught counting and measure in an enjoyable luanner which
,viII not destroy their intellectual interest, for example, by
pairing wrestlers for games, and by distributing bowls of various
materials and sizes to other members of the grOllp.39 It has been
suggested that the objects to be distributed in this lesson were
gold, silver, and bronze coins in the bowls. This would indeed
make it seelll that Plato had anticipated one of our modern
techniques of progressive education; bllt one wonders wIlether
the Athenian Stranger ,vould not have detected some illiberal
Egyptian or Plloenician implications in this practice of playing
store. Still, it must be admitted that the handing around of
bowls as a children's game seems somewhat bizarre.
If something about these bowls is deliberately reminiscent of
the nested hemispheres of the cosmic model of Republic x, how-
ever, the game may take on added point. Perhaps the player
is Sllpposed to distribute the bowls into matched sets, of assorted
sizes, but each set consisting of the same material. (This suggests
an analogy to tIle modern set of identical cubes of different
weight used for testing in our child psychology.) The difference
in composition would then be determined by the appearance
and tIle weight of each bowl. The art of weighing is one branch
of applied mathematics with which the citizen is to be familiar.
In this game, it would actually be made the basis of an ele-
mentary training in analogy; weight A is to bowl of size B as
weight C is to bowl of size D. In addition to familiarizing the
students with one of the important branches of measure, this
game has merit as an introduction to the study of astronomical
dynamics, if that study is condllcted with such models as that
of Republic x. It is even possible that Whetl he comes to choose
anotller incarnation,40 the student trained in this way will
learn more quickly, from his vision of the different colors and
sizes of the bowls, of Necessity's spindle. Certainly it will help
him to see the cogency of the antiatheistic argument based on
66 Plato's i\1athematical Imagination

the rational order of the cosmos, as revealed throllgh the bal-

anced distribution of its circlllar motions. 41
4. The importance of knowledge of incommensurables. The
Athenian Stranger says sharply that ignorance of solid geometry
is "swinish." 42 This is strong language. Perhaps it shows merely
that Plato in his old age developed a querulous impatience with
the general lack of interest in a branch of mathematics which
his Academy had done so much to develop. Perhaps, however,
as the context of this remark suggests, it is both literally meant
and defensible. The peculiarity of pigs is the domination of
their behavior by pure appetite, uncolored by spirit or intel-
ligence. The ability to initiate and perfect social institutions
which go beyond the purely appetitive needs of citizens is one
of the attributes which differentiate a city of men fro~ a city
of pigs. This is Glaucon's criticism of an analysis of the state
set up exclusively in terms of Sllpply and demand. 43
In practical politics, the greatest problem is not so ffillCh that
of securing agreement about principles, as devising institutions
that will put them into effective operation. To do this requires
the creation of a structure associated with the function given
by the principle to be applied. All technology, including the
social, must be dominated by discovering the measure of those
constructions which embody a desired formal perfection.44
In a later book, we learn that causal efficacy proceeds from
a point to a solid. 45 Aristotle reports Plato's use of this same
analogy to describe the progression through t11e facllities of the
sou1. 46 Reason, which apprehends its object directly, must com-
bine with understanding, which analyzes the associated struc-
ture, and opinion, which supplies experience relevant to the
materials and techniques suitable to embody this strllcture, if
noetic insight is to issue in social betterment.

d. Statutory Mathematics: Fines and Sumptuary Laws

Plato's general carefulness in the choice of quantitative state-
ments and the legislation relating to population and representa-
tion in the Laws make it seem, a priori, likely that the details
of statutory fines were also carefully considered.
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 67
From the fact that the statutes assign different fines to of-
fenders of eacl1 property class, one might think that punishment
were being equalized simply by making the penalty proportion-
ate to ability to pay. In fact, ho,vever, fines based on this prin-
ciple seldom occur. The fines set reflect the same mixture of
oligarchy and democracy that characterizes the system of repre-
sentation. Members of the wealthier classes are apparently
expected to set a good example, and may be punished by an
incremental penalty added to their fines when they do not.
In certain cases of bad citizenship, the fine levied is intended to
reduce the property class of the offender; we are told explicitly
that this is to be done with persons not registering their prop-
erty (Laws 775), and we may fairly assume that the extremely
heavy fines for members of the upper economic classes who are
co\vardly 47 or who bring ullfounded charges against a state
examiner 48 (fines of 1,000 and 720 drachmae, respectively) have
the sanle function. (If this is true, it gives some basis for estimat-
ing the actual capital of each class.)
Further, a disproportionately heavy fine is levied in certain
other cases. The upper classes are filled 10 drachmae' for non-
attendance at the assembly (Laws 764); the lower two classes
pay no fine at all. Faillire to confine a madman carries an in-
cremental penalty, above proportionate liability.49
On the other hand, the Sllmptuary laws allow a disproportion-
ate expenditure by the upper classes for such occasions as
funerals. 50 The greater civic importance of these groups is rec-
ognized in a disproportionate allowance for display; the upper-
class funeral is given more importance by statute than the
lower-class one. This seems a shrewd device on the part of a
legislator bent on harnessing the profit motive to promote social
indllstry; for a certain amount of disproportionate conspicuous
display is evidently an added sign of status which will increase
the eagerness of the citizens for economic advancement. Fllrther,
there seems proportionately less liability for bachelors of the
two upper classes,51 perhaps again reflecting their greater im-
portance to the community.
The device of an incremental penalty, explicitly introduced
68 Plato's l.\lathematical In2agination

in cases taken to courts of appeal, 52 is also operative in these

statutes, adjusting privilege and penalty to ability to pay and to
the social responsibility of persons of each economic group.
In certain cases, where the offense is criminally motivated
btlt where the actual economic damage is not great, a tenfold
increase magnifies the penalty to fit the seriousness of the crime.
The increase takes effect in cases of petty pilfering 53 and late
payment of the bachelor's fine (\vhich is a form of pilfering from
the temple treasury).54 This is the scale on which the gods re-
ward and punish human behavior in the l\lyth of Er; 55 but in
general, except for those cases in which there is so great a dis-
crepancy between motive and damage, such exacting justice is
beyond the power of the human legislator.
These aritllmetic details of the legal code seem thus to be
carefllily adjusted to combine the democratic and oligarchic
principles basic in the constitlltion of the state. (See figures on
pp.69-71.)

III. ARITHMETIC DETAILS IN MYTH AND CHRONOLOGY

Whatever imperfections time and chance intrude into history

may be rectified when that history is recast in a pllrified form
and presented as a myth. The line between Plato's "history"
and "mythology" is singularly hard to draw, since his excursions
into either fielcl are dominated by a search for episodes which
serve a purpose; both the history and the nlyth have a moral
and bring it out more clearly by neglecting accidelltal distor-
tions of the pattern. Since the value of the myth is its vivid
presentation of individual events and careers, radically inac-
cessible to dialectical abstraction, the myths contain vivid cir-
cumstantial details which have often convinced Plato's readers
that they ,vere intended as histories.
The rectification of distortions is particularly apparent in
mythical Inathematical detail. Things happen in periods that
are exact "round numbers." The longest such period is repre-
sented as a "nlyriad," and its subperiods as other powers of ten.
The "almost as long" or "almost complete" period is projected
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 69

Figure 23
FACTORS OF 5,040: A LEGISL~L\TOR'S l\IANUAL
~O. OF NO. OF NO. OF PERSONS NO. OF NO. OF NO. OF PERSONS
FACTORS GROUPS IN EACH GROUP FACTORS GROUPS IN EACH GROUP

5,040 xxxi 72 70
ii 2 2,520 xxxii 80 63
iii 3 1,680 xxxiii 84 60
iv 4 1,260 xxxiv 90 56
v 5 1,008 xxv 105 48
VI 6 840 xxxvi 112 45
vii 7 720 xxxvii 120 42
v-iii 8 630 xxxviii 126 40
ix 9 560 xxxix 140 36
x 10 504 xl 144 35
xi 12 420 xli 168 30
xii 14 360 xlii 180 28
xiii 15 336 xliii 210 24
xiv 16 315 xliv 240 21
xv 18 280 xlv 252 20
xvi 20 252 xlvi 280 18
xvii 21 240 xlvii 315 16
xviii 24 210 xlviii 336 15
xix 28 180 xlix 360 14
xx 30 168 I 420 12
xxi 35 144 Ii 504 10
xxii 36 140 Iii 560 9
xxiii 40 126 liii 630 8
xxiv 42 120 liv 720 7
xxv 45 112 Iv 840 6
xxvi 48 105 lvi 1,008 5
xxvii 56 90 lvii 1,260 4
xxviii 60 84 lviii 1,680 3
xxix 63 80 lix 2,520 2
xxx 70 72 Ix 5,040 1

If the suggestion made in a later section as to the intended form

of computation of the ratios of the world-soul is correct, one can be
certain that Plato used some similar device rather than a tedious
trial-and-error series of atomic calculations to establish these sixty
factors. Such an atomic procedure would have seenlcd to him both
tedious and inelegant. If the matrix computation pattern suggested
70 Plato's Mathematical Imagination
for the Timaeus passage were applied to logistic, a simple set of
tables could produce all of the factors of any given number in a
neat, complete, and relatively elegant form, and demonstrate why
these were the only factors of that number.

Figure 24
STATUTORY FINES AND SUMPTUARY LAWS· . (Laws)
,;
BASIC FINE PROPORTIONATE FINES ACTUALLY
(LOWEST FINES, OTHER ASSESSED, OTHER
OFFENSE CLASS) CLASSES CLASSES

1 2 3 4 2 3 4

Not marrying 30 60 90 120 60 70 100

Disrespect of elders 20 40 60 80 30 50 60
Cowardice 60 120 180 240 180 300 1000
Unconfined madman 24 48 72 96 36 48 100
False charges against ex-
aminer (accuser gets
less than 1/5th of
votes) 240 480 720 960 360 480 720
Not voting (see Fig. 25) 10 20 30 40 10 30 40
Absent from assembly 10 10

SUMPTUARY LAWS
MAXIMUM EXPENSE PROPORTIONATE ACTUAL STATUTORY
ALLOWED LOWEST MAXIMUM FOR MAXIMUM FOR
ECONOMIC CLASS OTHER CLASSES OTHER CLASSSES

2 3 4 2 3 4

Wedding feasts 7.5 15 22.5 30 15 30 60

Marriage garments 50 100 150 200 60 90 120
Funeral costs 60 120 180 240 120 180 300

TAX-EXEMPT CAPITAL
PROPORTIONATE ACTUAL STATUTORY
FOR LOWEST EXEMPTION FOR EXEMPTION FOR
ECONOMIC CLASS OTHER CLASSES OTHER CLASSES
2 3 4 2 3 4

60 120 180 240 120 180 240

• Fines, expenses, and exemptions are given in drachmae.

;
 "SOCIAL STATISTICS": ARITHMETIC DETAIL 71

Figure 25
ELECTION STATUTES (Laws)
(A council of 360 members is to be elected, 90 from each class.)
DAY PROCEDURE FINE FOR NONATIENDANCE

(Class) 1 2 3 4

1 Nomination candidates from 1st class 10 10 30 40

2 Nomination candidates from 2nd class 10 10 30 40
3 Nomination candidates from 3rd class 10 30 40
4 Nomination candidates from 4th class 90 160
5 Balloting: 180 of each class elected, 90
selected for council by lot 10 10 30 4<J
A. E. Taylor Ii shows that the effect of these differential fines is
to give the lower economic classes a majority in the nomination
and challenge of representatives of the upper classes and to put
the upper economic classes in the majority for the nomination and
challenge of candidates from the lower economic classes.
• Plato, the Man and His Work (new ed.; New York, 1946), pp. 479-80.

as a power of 10 times 9. This peculiarity of mythical mathe-

matics is apparent in the temporal periods associated with the
accounts of transmigration. In Republic x the length of human
life is 100 years, the cycle of incarnations 1,000. 56 In chronology,
the war with Atlantis has taken place 9,000 years before, while
the age of Egyptian culture and its records is 8,000 years. 57 In
another transmigration myth, the cycle of recurrence is 10,000
years; 58 the foolish lover is bowled around the earth for
9,000; 59 and the philosopher is released after 3,000 years. The
state of Atlantis is described as having "myriads" of citizens,6o
troops, and acres.
The "myriad," like our "million," is a large number-word
used to express great or indefinite size. It is also a power of ten,
which makes it a suitable number for myths in which there are
many echoes of Pythagorean tradition. But it is often not
enough simply to indicate that a period, while definite, is in-
definitely large, if it comprehends subperiods to which it must
be shown as proportionately related. In these cases, the smaller
powers of ten are assigned to the subperiods.
 CHAPTER III

Geometric Metaphor
VERBAL MATRICES: Introductory Remarks

As PROFESSOR BUCHANAN has pointed out in his book Symbolic

Distance}l Plato's techniques of definition and discussion may be
visualized as instances of matrix construction. A verbal matrix
defines a term by locating it in relation to a set of other sys-
tematically ordered terms. If the meaning is not clear from
the arrangement, either the matrix is enlarged, or the term in
question is itself replaced by a matrix, differentiating the senses
in which it is used. Distinctions in scope and field of terms are
thus schematized by the spatial relations of the matrix grid
structure. The most famous such matrix, which Professor
Buchanan cites, is derived from the Republic. 2 ,_~n each case,
the_ matrix grid presents a graphic and convenient mathematical
image for the spatialization of a net of dialectical distinctions.
Since the elements of such matrices are words, which are
images, rather than numbers, which are natures,S analogies to
mathematical matrix theory are misleading, and a logic of verbal
matrices cannot possibly be built up simply by applying modern
mathematical matrix operations. Verbal matrices, however, have
certain formalizable properties and operations of their own
which provide images of the formal character of Platonic logic.
The use of matrices for combination diagrams probably '\-vas
developed by mathematicians and perhaps in the medical tradi-
tion before Plato's time. 4 In the Dialogues} frequent references
to "left" and "right," "vertical" al1d "horizontal," division sug-
gest that Plato himself was visualizing verbal matrices of the
sort described, and referring to them as an aid to the attentive
72
 GE01\IETRIC METAPHOR 73
reader. \Vhatevt:r its origillal inceptioll, the use of a lllatrix as
a notational device to present clearly relatiolls \vhich are not
evident from other cOllverltional 110tations appealed to the
Pythagorean school. A technique so effective in dealing with
factors in arithmetic, and \vitIl combinations in ratio theory,
suggested tIle desirability of exhibiting terms or concepts in a
tabular form, the location of each in tIle table illdicating its
resemblance to and difference fronl the other terms included.
Aristotle preserves one such early Pythagorean "table." 5 In this
table, ten pairs of contraries are .arranged in two colulllns: terms
in the saIne column are "alike," terms in opposite columns on
the same row are "contraries," terms in opposite columns not
on the same row are "unlike." This is essentially the same use
of spatial orientation to symbolize analogies and differences
lvhich Plato adopts in his explicit construction of matrices.
Every matrix has certain conventions in its spatial orienta-
tion. Like the divided line, distinctions of being al1d becoming
are arranged from top to bottom, the higher position being
reserved for the term most clear to intelligence and akin to
reality. The reader's left (the matrix's right) is lik'ewise the
more honorific position; and if the differentiation of columns
also embodies an application of the reality-appearance distinc-
tion, this runs toward the reader's right from his left. The basic
orientation of a matrix is therefore that of Figure 29, as in the
Sophist passage.
The relation of adjacent terms is ,one of proportion or anal-
ogy; thus, in Figure 29, A:B::C:D and A:C::B:D. In definition
by dialectic, the proportion is Plato's favorite mode of present-
ing an analogy; 6 the terms of every matrix are proportionately
related, and a matrix can be constructed from any set of pro-
portions. Often an analogy is developed between sets of terms
that are ontologically co-ordinate, as in the analogy of the state
to the body in the Laws (see Figure 30), and in such cases, the
order of columns does not indicate a being-becoming, appear-
ance-reality distinction.
The closest one can come to an analogy to quantitative matrix
maniplliation is by describing the operations of vertical and
74 Plato's Mathematical Imagination
lateral "shifting of terms in the verbal matrix. If a term is
U

shifted along a row, from its proper place to another column,

the result is a metaphor. For example, in Figure 30, if we call
the cOllncillors in the Laws heads of the state, '\ve are using the
term "heads" metaphorically. The distance of such a shift gives
an index of the viviclness of a metaphor, or of its "tension"-
the quality which leads us to describe some metaphors as
"strained." If a term is shifted between rows within a column,
the result is the "collapse" or "reduction" of a distinction. If,
for example, "head" is shifted to the row which in contextual
columns corresponds to the spirited part of the soul, the shift
suggests an erroneous identification of passion and reason; a
"being" term is explained by a context inappropriately colored
by "becoming." The mechanism of error in the Timaeus is
developed with the help of optical analogies which give such
"shifts" in space precisely this interpretation.7
In certain cases, Plato "reduces" his matrices and "multi-
plies" them. A "reduced" matrix presents a set of combinations
in a simple linear listing in which the original rows or columns
are arranged successively. Plato "multiplies" matrices in which
one set of the distinctions, either that of rows or columns, is
the same in both. The two matrices are combined by joining
them along the axis representing the distinctions which they
have in common, while the different axis of the second is used
to represent a projection of the first into the third dimension.
Matrices are also often "schematized." This is done by repre-
senting successive matrix positions by ordinal numbers, so that
the relations of tl1ese numbers summarize the original dialecti-
cal proportions which the matrices present. Thus, the operations
described above as reduction and multiplication of matrices
111ay be "schematized" as shown in figures 36 and 38.
Plato's method of schematizing matrices by simply numbering
successive positions is in principle the same as contemporary
matrix technique, which uses subscript positional numbers.
The number assigned a given position in Plato's version can
be shown to be a function of the subscripts in modern nota-
tion. 8
 c;l,

:r: = c.. =
c: -c: <Jc: <lc:
<J c.. <i <]
:r: <J <J <l i: <i <l
:c <l <1 lL <l
:r: <J <l
<J L <l
!.. <J <1 L
8L <J
:I:
::r:
:c
<l
<l c:: =
c..
E
<J <i
<i
<l
<l
<l c:: E:lL <i <J
<i
<J <l <J !L
Ii. BL. <J <l <J <l a..
<1
<l
:r: <]
<i c.. E
::c
L
<1
Bl.
E: E:
<i <J c:
<l
<]
<J
<i
<:J
=
c.. rL <i
lL
<J <:J <l
<J <J
<J
<J
<J
:I:
L
<]
m.. E: E: <i r:<J <i- <lIE:
<1 <1 <i
<1
E:
<J
=
Q:
<J <J <J
<J
<J <l
i

<J
I

<l
<l
V'..t

o i.. L r:.. c. <J c.. <J c.. <l c.. <1 c..
<J <J <I <J <J <l <J
~ <I <J <J <I <]
V'..t
<J <J <l
< <1
<{
-e- :c <1
- = <i c: <J
-<l -c.. <J E:
LJ.J
~
::I:
:c
<]
<]
c: <J <J
<J
<i <l
<l
:c <l <l <J
<J <1
::r: <J - - E: c... <i E:
:r:
:c <J E: <i <J E: <J <i <1
<l <J <1 <J
:I:I<J
::r: <J = = - E E
<J =
<J
<i
E:
<J E:
<l
:r:1<J = - c.. E:
= E: E: E:
~
~
= - -
c.. E: E E E
~ <] <1 <] <1 !L <J <J <J <J
<.0
CN
Z
o <l <l <1 lL <J <J <J
<J <J Sl <J <J
~
=' ~
tJ
<J L <J
bO Bl
~ :3
~
:I: :::c :J: i.. :I: :r: :J: :I:
6
;J
:I:
:I: ::r:
:J:
:J: i.. ::I: :c
:I:
:J:
:I:
~
:I: L
:c Ii. :I:
L

r I
 PlatoJs J.JathenLatical lrnagination

That such a table as the one in Figure 26 was in use is clear from
Aristotle's reference to the xE<paAL(j~6~ in Topics 163b.17.:1
This version is given in the "Herodianic" notation, which was
probably what i\ristotle was referring to.t
The t.able is a device for marshalling products under the "head-
ings" representing their factors; by using three rows of headings,
as in this figure, a 9 by 9 matrix can be used, as Aristotle says the
multiplication table up to ten can, for ll1ultiplication up to 1,000.
The rules for using such a table to multiply large numbers are,
as can be seen, extremely simple in the "Herodianic" notation,
'\vhich is not the case with the alphabetical number-system; to some
extent, this confirms the notion that it was the former which
~~ristotle had in mind.
No doubt actual computation had led to many short cuts and
abbreviations of the "official" Athenian number-system; presum-
ably, however, these would have been taught in a more advanced
course of "logistic" than the elementary work in which the student
mastered his ta bles up to ten times ten.

Figure 27
1\ PYTHAGORE£t\N TABLE OF CONTRARIES
(Aristotle, ivl etaphysics 986A)
LI~llT INDETER~nNATENESS

ODD EVEN
UNITY PLURALITY
RIGHT LEI"T
~rALE FEl\fALE
REST 'MOTION
STRAIGHT CURVED
LIGHT DARK
GOOD BAD
SQUARE RECTANGULAR

'"This "table" brings out two points concerning the Pythagorean

tradition that are important for the present study. In the first place,
many of the analogies suggested reappear in Plato's own mathe-
matical imagery. More itnportant, however, the arrangement in a
series of corresponding pairs, s.uch that terms on the same line are
... See Sir Tholnas Heath, Mathematics in Aristotle (Oxford, 1949), p. 93.
t See Heath, History, I, 30 ff.
 GEOMETRIC METAPHOR 77
contrary, those in the same column similar, and those in opposite
columns dissimilar, shows exactly, in an elenlentary form, the use
of spatial relations to indicate dialectical ones that characterizes
Platonic "verbal matrices." (Aristotle specifically mentions this
111atter of arrangement of the terms as an integral part of the
Pythagorean theory he is presenting here.)

Figure 28
HIPPOCRATIC GENETIC THEORY: A COMBINATION TABLE t
"Theoretical combinations of male and female sex-elements.
.tV! represents a male seed which dominates; m a male seed which is
J

dominated; F~ a female seed which dominates, and f, a female seed

l lvhich is dominated. The first symbol in each combination represents

the seed contributed by the female parent, the second symbol (in
italics) that contributed by the male. Combinations marked • are
impossible, because they violate the postulate that one and only one
seed can be dominant. Combinations marked t are impossible, be-
cause they violate the postulate that a male-contributed male seed
cannot be dominated by a female-contributed male seed, and vice-
versa."
1VIALE PARENT
M m F £

FEMALE M MlvI* Mmt lVIF* Mf

m mM mm· mF mr Ac

PARENT F FM· Fm FF· Ff

f fM fm* fFt ff·
t From R. S. Brumbaugh, "Early Greek Theories of Sex Determination," Journal
of Heredity, XL (1944), 50.

~~he theory as preserved does not use such a matrix, but examines
in turn and in systematic fashion each of the theoretically possible
combinations of the parental seeds, as though some matrix notation
had been used in calculating these combinations. It is hard not to
believe that the medical men would have recognized the suitability
of mathematical matrix technique for their analogous problem in
genetic combination.
In Figure 29 the axes from top to bottom and from left to right
represent principles of classification generating combinations, as in
the Sop·hist matrix. If three principles of order are used, the matrix
78 Plato's Mathematical Imagination

may be made three-dimensional and the axis from front to back

also used to represent a principle of classification. The dotted lines
indicate that this matrix may be extended to include an indefinite
number of columns and rows.

Figure 29
STANDARD SPATIAL ORIENTATION OF A PLATONIC MATRIX
BEING. • • • • • • • • • • • BECOMING

A B
C D

BECOMING

Figure 30
A POLITICAL MATRIX FROM THE LA WS
PARTS OF SOUL CLASSES IN STATE ORGANS OF BODY
Reason Councillors Organs of Perception (Head)
Spirit Citizens Organs of Locomotion
Appetite Craftsmen Organs of Nutrition
Compare figures 34 and 35, following. The craftsmen in this
matrix are the resident foreigners who carryon business in the state
but who do not share in military duties or other civic Uexecutive"
functions.

SOME TYPICAL USES OF MATRICES IN THE SCHOLIA

(Figures 31 through 35 from Greene, Scholia Platonica J

pp. 139, 147, 224, 226)

Figure 31
Gorgias 465C
BODY SOUL
Gymnastic Lawmaking
Medicine Judging
Cookery Sophistry
Scene Painting Rhetoric
 GEOl\<IETRIC METAPHOR 79
Figure 32
(;orgias 477A
PARTS OF ~{AN DEFECT EXCELLENCE

Soul Injustice Justice

Body Illness Health
Externals Poverty \"eaIth

Figure 33
Republic 435B
Justice Whole State Whole Man
Wisdom Rulers Spirit
Courage Soldiers Reason
Temperance ,Artisans i\ppetite

Figure 34
Republic 440£
IN THE STATE IN THE SOUL

Productive Appetitive
Protective Spirited
Deliberative Rational

Figure 35
CORRECTION OF THE REPUBLIC MATRICES
PART OF SOUL PROPER VIRTUES OF CLASSES

Rulers Soldiers Artisans

l REASON
SPIRIT
Wisdom
Courage
.......
Courage
. ......
.......
APPETITE Temperance Temperance Temperance

The diagrams of figures 33 and 34 have gained considerable

currency, but both distort Plato's definition of "temperance," which
is a virtue proper to the function of all classes in the state. The
diagram in Figure 35 is more accurate.
Since all classes share in temperance, and since rulers are selected
')n the basis of their earlier demonstrated courage as well as their
wisdoln, the scholion figure rather misses the point of the text, in
the interest of facility of schematization. If justice is to be included,
it will also be a property of all classes, perhaps best schematized as
horizontal lines of class separation in the figure.
80 Plato's J\Jathernatical Irnagination

In general, these figures do not £ollo'\V' Plato's own use of space

in their schenlatization. In Figure 31, the vertical listing has no
principle o£ order; the relation of left-right is just the reverse of
that used by Plato in the Sophist matrix. Figure 32 is thoroughly
Platonic, but in Figure 33 the erroneous schematic presentation
of temperance as exclusively the virtue of the artisan class com-
pletely thro,vs oif the part-whole principle of vertical order. In
Figure 34, the typical relations of top and bottom are gratuitously
reversed, inverting the picture. This, as well as the inexact
schematization of tenlperance, is corrected in Figure 35.

Figure 36
SCHENIA.~rIZATION i\.ND REDUC~rION OF A THREE Tlj\IES
THREE PL~t\NE l\fATRIX
1 2 3
4 5 6 147 1258 I ~ 69
7 8 9 row 1 row 2 3
l'O\V

The positions of the terms in such a matrix, if the typical spatial

orientation is respected, mean that the position of any term is itself
adequate indication of the logical relations of that term to the
others in the schema. Hence relational structu~es may be repre-
sented and treated by substituting positional numbers for the terms
thenlselves, as in the square figure. (To do this involves a con..
vention that the left·right principle of order will take priority in
assigning the numbers to the positions.) The distance indicates the
exact relation of the term in one position to that in another (if the
dinlensions of the matrix are given): for example, representing the
horizontal ordering relation by H, the vertical by V, 1 is in the
relation H I H I V or H2 I V to 6. (Here the stroke (I) is used,
as in relational calculus, to represent a "relational product." H I V,
for example, abbreviates (3x)(3y)(3z)[H(x,z). V(z,y)].
If the schema must be changed from two dimensions to one,
either because two distinctions must he represented by a single
dimension (since four or more principles are being diagrammed),
or because such a listing better represents the fact that the two
ordering relations are not co·ordinate in importance, the result is
the reduction of the matrix to a line. The linear series in the figure
represents one such reduction, in which the relative importance
of the top-bottom relation of the original matrix is elnphasized
by making this the ordering relation for each group of three.
 GEO\-LETRIC METAl-'HOk 81

Within each group, the original left-right relatioll determines the

order. Since the convention of numbering positions is independent
of this relative emphasis on the co-ordinates, the resul ting list may
hide the original diagraln from which it was derived.
In subsequent discussion, examples will be given of specific uses
in Platonic images of reduced, multiplied, and schematized verbal
matrices.
Such matrix schematization may be a device for presenting the
abstract relations bet\veen sets of comparable elements, ordered in
terms of two principles simultaneously. With the added concept of
"distance," it is a device for stating the relation of any pair of
elelnents in a set ordered in this way. Whenever sets of terms in
such relation appear in a dialectical discussion, a 111(1 trix schema

Figure 3'7
lVIUL TIPLIC~~'rION OF 'I'WO SIlVfILf\R M,\'TRICES

c
abc abc
def mno
g h p q r

or,

3 11 U12 a 13 b 1l b 12 b 13
a 21 a22 a 23 b 21 b 22 b 23
Q
a 31 a 32 a 3R b 31 b3~ b a3 1..1~1
'~ Q
l~~
82 Plato's i\Iathernatical Inlagination

provides an appropriate diagram to clarify their interrelation. What

the constitutive eleruents in the relational scheme are, however,
rnust of course be determined by substituting for the positional
nunlbers the terms which themselves figure in the dialectical
context.
T\vo schelnata may be combined into a third, when their dimen-
sions and one set of distinctions are the same, by joining the
identical r01\ S and using the second plane as a projection of the
T

first into the third dimension, as shown in Figure 37. The same
operation is shown with positional subscripts, indicating plane, row,
and column, respectively.

Figure 38
r
SCHEMATIZED 1\1UL rIPLICA.TION OF TWO SIMILAR MATRICES

:; 12
1 2 3 1 2 3 ~
4 5 f; 4 5 6 24
15
7 8 9 7 8 9 6
e 18
21
~ 9

In this figure, the letters and subscripts are replaced by positional

numbers; the joining of the planes requires a new assigning of
number, on the convention that planes are numbered from front
to back, which makes a number of the original plane used for the
third-dimensional projection equal to the ordinal number multi-
plied by 6 times the number of planes minus 1.
II GEOl\IETRIC ~IETAPHOR 83
I
\
I. CONSTRUC]'ION OF A MA1'RIX
f
Sophist 266 *
I
I STRANGER: And so there are two kinds of making and pro-
duction, the one human and the other divine.
THEAET.: True.
STR.: Then, no\V', subdivide each of the two sections which we
have already.
r-rHEAET.: How do you mean?
STR.: I luean to say that you should Inake a vertical division
of production or invention, as you have already made a lateral
one.
'T'HEAET.: I have done so.
STR.: Then, now, there are ill all four parts or segments-
two of them have reference to us and are human, and two of
them have reference to the gods and are divine.
THEAET.: True.
STR.: And again, in the division which was supposed to be
luade the other way, one part in each subdivision is the making
of the things themselves, but the two remaining parts may be
called the making of likenesses; and so the productive art is
again divided into two parts.
THEAET.: Tell me the divisions once more.
STR.: I suppose that we, and the other animals, and the ele-
Inents out of which things are made-fire, water, and the like-
are known by us to be each and all the creation of God.
THEAET.: True.
STR.: And there are images of them, which are not them, but
which correspond to them; and these are also the creation of a
wonderful skill.
THEAT.: What are they?
STR.: The appearances which spring up of themselves in sleep
or by day, such as a shadow \vhen darkness arises in a fire, or the
reflection which is produced when the light in bright and smooth
objects meets on their surface with an external light, and creates
a perception the opposite of our ordinary sight?
• Trans. Jowett, Dialogues, IV, 403-4.
84 Plato's l"'vlathen1atiral Imagination

Figure 39
'rHE SOPHIST l\{f\ TRIX

DIVINE CREATION HUMAN CREATION

Realities Natural Objects Artifacts

Shadows
Mimetic Works
Appearances and of Art
Reflections

THE1\ET.: Yes; and the images as well as the creation are

equally the work of a divine hand.
STR.: And ,vhat shall we say of human art? Do-we not make
one house by the art of building, and another by the art of draw- .
iug, which is a sort of dream created by man for those who are
awake?
THEAET.: Quite true.
STR.: And other products of human creation are also two-
fold and go in pairs; there is the thing, with which the art of
making the thing is concerned, and the image, with which imita-
tion is concerned.
THEAET.: Now I begin to understand, arid am ready to
acknowledge that there are two kinds of production, and each of
them twofold; in the lateral division there is both a divine and
a human production; in the vertical there are realities and a
creation of a kind of similitudes.
 GEOMETRIC l\rIETAPHOR 85

IMAGES OF HARMONY AND CYCLE: the Republic

II. MATHEMATICAL !l\fAGERY IN THE REPUBLIC;

THE STATE AND THE MUSICAL SCALE

Harmony characterizes those moral and political structures

which are closest to perfection. The list of images in the Repub-
lic marks the crucial stages of the theoretical construction of
such social harmonies. The image of harmony in music is
inserted as a summary of the discussion of education and art
at the COl1clusion of Book iii,9 and is again referred to in an
explanation at the end of Book iv 10 of the class structure of the
state. The way in which education inevitably introduces greater
harmony into social cycles is illustrated by the image of the up-
ward spiral in Book V,11 and the deteriorating effect of time and
chance on human institutions, leading to loss of harlnony, is
presented in the balancing down,vard spiral markillg the transi-
tion from the best state to worse states in Book viii. 12 The in-
tellectual ascent to principles is to be accomplished by the
largely mathematical curriculum of Book vii,13 founded on the
differentiation of kinds of clear kno,vledge illustrated by the
divided line at the end of Book vi. 14 The greater happiness of
the jllst man, the original point of controversy, is resolved by
the calculation of the tyrant's unhappiness in Book ix; 15 and
finally given a complete warrant by the mechanism of cosmic
justice, shown to the souls in the myth at the end of Book X,16
in which the two earlier spiral images are combined into a
cosmic and moral cycle. It is with this over-all, interconnected
pattern in mind that we must proceed to the separate treatment
of these constitutive mathematical images of the Republic.

Republic 443 *
The just man does not allow the several elements in his soul
to usurp one another's functions; he is indeed one who sets his
house in order, by self-mastery and discipline coming to be at
peace with himself, and bringing into tune those three parts,
like the terms in the proportion of a musical scale, the llighest
... Trans. Cornford, Republic, p. 142.
8G Plalo's .I.\fathe1natical lm,agination

and lowest notes and the ll1ean between them [lit. v{rtll' v1tatfl,
and flEa'll] with all the interluediate intervals.

Scholion *
The nete is the final note of a system of two tetrachords; the
hypate ... the first note of the t'VlO tetrachord system. The mese
is the note which is the termination of the first tetrachord and
the starting point of the second, as Ptolemy says, and other
musicians.
Republic 432A t
... Temperance works in a different way; it extends throughout
the whole gamut of the state, producing a consonance of all its
elements from the weakest to the strongest as measured by any
standard you like to take-wisdom, bodily strength, numbers, or
wealth. So we are entirely justified in _identifying with temper-
ance this unanimity or harmonious agreement between the na-
turally superior and inferior elements on the question which of
the two should govern, whether in the state or the individual.

The image of harmony as an evidence of organization of

parts into wholes in which they are internally related is repeat-
edly suggested in the discussion of education in the Republic.
It is given one of its most explicit statements in 443D, where
the temperate relation of classes in the state is compared to the
relation of the fixed notes of a scale. 17 It is important to note
that such "harmony" presupposes heterogeneity; the notes do
not blend to produce an average pitch midway between them,
but each, retaining its own identity reinforces that of the other.
Thus a harmony in state or soul does not suggest metaphorically
that classes or faculties become identified; on the contrary, one
condition of temperance is the presence of jllstice, which keeps
the parts distinct, and insures that each will make its own in-
dividual contriblltion.
The purpose of the science of politics' is to construct "har-
monies" in the periodic processes ,vhich constitute the temporal
careers of state and soul. The imagery of the Republic reflects
* Greene, Scholia Platonica, p. 226.
t r-Trans. Cornford, Republic) p. 126.
 GEOMETRIC ~{ETAPHOR 87
this enterprise in its initial introduction of disparate nletaphors
of llarmony and cycle, \vhich in the later itnages are gradually
sho"",'n combined.

Figure 40
CHANGES OF TOPIC AND LOCATIONS OF 1\IA.THE~IA TICAL
IlVIAGES IN THE REPUBLIC

llsterisks in the left column mark the locations of mathematical

references; it will be noted that these fall for the most part at places
'where either a summary or an anticipatory schematism would be
reasonably expected.
LOCATION
(BOOK) TOPIC OR Il\.IAGE
[ Criticism of popular definitions of justice
11-1\' Origin and functional class-structure of the state
** Analogy of temperance and harnlony
v l\Iarriage and family regulations
Image of cyclic growth of state
V Conduct of war
VI Education for rulers
Kinds of knowledge: the Divided Line
VII IVlathematical curriculum; dialectic
\'111 Variant forrrls of state and character
The Nuptial Number: cyclic decline of the state
VIII-IX Variant fornls of state and character
'The Tyrant's Number: summary of desirability of lives
of each type
x Slunming up: tendency of experience to confirnl the
argument
x 1\1ythical postscript: the afterlife
•• A.stfollomical model, l\fyth of Er: justice in the cosmos

III. "THE CYCLE OF SOCI.AL PROGRESS

Republic 424A *
l\nd, lnoreover, said 1, the state, if it once starts ,veIl, proceeds as
it were in a cycle of growth. I mean that a sound nurture and
education if kept up creates good natures in the state, and sound
I natures in turn receiving an education of this sort develop into
• Trans. Shorey, Republic, I, 331.

r
I
88 Plato's l\Jathematical I1nagination

better men than their predecessors both for other purposes and
for the production of offspring as among animals also.

Tllis refercllce to the state "growing like a circle" is 1100-

techn.ical a11d brief. In Shorey's tralls]atiol1, just qlloted, the
circle is thought of as a kind of civic life-cycle, just as allimals
have a cycle of grovvth. jowett's translation elnbodies his i11ter-
pretation; the state gains ill mOlnentum as a l1eavy lvheel would,
alternately pushed at A and B (See Fig. 41).18 J. .L~dam inter-
prets the passage as referring to a circle being drawn ,vith a
compass; the state purslles an upward arc for a time, passes its
acme, and continues through a downward one (as s110wn ill
Fig. 42). In context, it seems quite clear that the correct inter-
pretation of this "circular" effect of nature and training must
take account of some sort of reciprocity) and Jowett's notion
of the reciprocal effect of impulsions applied to a circle which
is rotating seems to do tllis best. But this judgment is con-
tingent on the interpretation of the "nuptial number," a pas-
sage ,vhich Cornford and Adam seem to have in nlind as
suggesting interpretations here. 19
The idea that tIle cycle of political history is being drawn in
this metaphor of social evolution, ,vhich underlies Adam's in-
terpretation-that the state is groV\Ting as a circle does when it
is being drawn ,vith a compass-certainly brings out a relevance
between this entire section of the discussion and the subsequent
uses of imagery of cycle.20 It does so, ho,vever, at the expense of
a rather unlikely interpretation of the metaphor given by the
H
text. In this context "cyclic increase in social excellence
stresses the reciprocity of nat'ure and nurture) a reciprocity of
which no adequate aCCOllnt is given by this interpretation. In
effect, Adam has substituted an alternative and plausible meta-
phor for the one which Plato really employed. It should be
noted also that Plato's image of a cyclic development condi-
tioned by the interaction of two contributory factors is not (as,
according to Adanl's interpretation, it ShOllld be) an arc of a
circle, but rather a spiral.
To say that this metaphor refers to the spreading Ollt of
 G:EOMETRIC }\tfETAPHOR 89
circles from a center of distllrbance, like tIle rings ill water,
completely overlooks the element of reciprocity which in this
context any appropriate Inetapll0r must incillde. It explains tIle
function of growth, but gives no functional significance to the
textual stress on circularity.
Professor Shorey's note on this passage denies that the "cir-
cularity" involved is intended to suggest any specific metaphori-
cal illustration. 21 But all the passages cited by him in tllis
connection involve the same basic metap110r of the growing
momentum of a 'tvheel spun by successive impulsions to its rim.
Lt\side from this point, since all citations involve the same
basic IIletaphor, the only question that remains is "\vhether the
passage itself is clearer or more vivid if the metaphor is taken
in a concrete rather than in the proposed abstract and generic
sense. Both the precedent of Plato's other illustrations from
techniques of the arts and his predilection for concrete vividness
of illustrative metaphor would lead one to believe that he here
intended the passage to sllggest the more concrete interpre-
tation.
The intuited fact that lnotion communicated to a point on
a wheel is distributed evenly to all points on its periphery is
prominent in the picture of the Fates turning the celestial
hen1ispheres in the Myth of Er. 22 Some clue to the relation of
cosmic mechanism and political history is latent in the similar-
ity of the mechanics of social evolutionary process to the mythi-
cal account of the coslnic turning.
Plato has a habit of introducing referellces and phrases from
familiar processes of the arts and crafts. His assumption that it
1vill help his reader to visualize things as being "like an eel-
trap" or "like a reinforced trireme hull" was no doubt a valid
one for readers who v\rere 11is contemporaries; but it is not valid
for readers today. Unfamiliar as we are with the everyday sights
of allcient Athens, and limited in our ability to reconstrllct and
\iisualize Athenian technology, the allusions become more of a
hindrance and puzzle than a help.
'I"he passage from Republic 424A, in discllssing ,vhich I
,vould accept Jo,vett's interpretation, is a case in !Joint. A pot-
90 Plato's Mathematical Irnagination

Figure 41

A B

GROWTH OF THE STATE: CIRCULAR IMPULSION

~-'his figure interprets the metaphor as "increase, like a wheel

being impelled from two directions in its rotation." In other words,
a force applied at A will increase the momentum of the whole
circle, and a second impulse applied at B will act reciprocally to
increase the velocity at A. The central point is that nurture and
education act cumulatively in accelerating the improvement of th~
state. An analogous use of momentum and impulsion appears in
the Myth of Er.
Figure 42
0_>- .... ,
"
", ,.
\
\
\
\
\
,
A c B
GROWTH OF THE STATE: CIRCLE UNDER CONSTRUCTION

In this interpretation, AB represents a time axis, CD an axis of

perfection. The circular growth of the state is thus envisaged as
a progress from A to D~ a decline from D to B. The chief objection
is that this figure completely ignores the stress in the context on
reciprocity, and the absence of allY reference to the decline in-
dicated by D-B.
 GEOMETRIC METAPHOR 91

ter's wlleel is heavy (like the flywheel of all old Ford) and there-
fore requires a good many inlpulses before it takes on high-
speed motion. Further, once it is started, its rnomentuln carries
it on its revolution, and, like the Ford fly\vheel, it is very
difficult to stop. This vVOllld be immediately suggested, by the
phrase Plato uses, to anyone \vho had ,vatched potters at ,,,,"ark
with hand-powered wheels; the increasing speed with each im-
pulsion and the inevitable persistence, once started, of the
moving wheel is far less self-evidellt to us, who buy our pottery
ready-made, than it 'llvotlld be if we watched the potter 'lvedge,
thr01V, center, alld shape his clay.
This brief passage constitlltes the first of a balanced set of
three cyclic images in the Republic. The inevitable up,vard
spiral of the state, following tIle introdllction of temperance
as the result of improved education, exactly balances the inev-
itable cyclic downward spiral, presented in the nuptial nUlllber
image of Book viii. The up'\vard and do\vnward spirals are
balanced against each other, and combined into genuinely cir-
clliar images and periods in the eschatological and cosmological
imagery which in Book x presents the relation of cosmic justice
and human freedom.
It is reasonable to suggest that for Plato there probably is
an intentional resemblance between this metaphor and the
description of the successive impulsions by \vhich the cosmic
spheres are given their irresistible momentllffi in the Myth
of Er.

IV. THE DIVIDED LINE

In dialectical illustration, as in lyric poetry, it frequently hap-

pens that many metaphors are developed simultaneollsly. It is
this very simultaneous development of many significant meta-
phorical references that gives aesthetic interest to poetry; bilt
in the attempt to carry alIt many image-class developments at
once there is bound to be what might be called "interference
of metaphor." The concrete image required for one line of
thought may not coincide exactly with the image which would
be most appropriate to another. A poet must then choose which
92 Plato's Mathematical Imagination

line of imagery is to be primary in his development, and within

that line retain as much relevance to the alternative lines of
development as possible. The actual choice of imagery can be
explained only by seeing the chosen development in its context
of aesthetically relevant alternatives.
Suppose, for example, that a poet begins a work with the
statement that "Josephine is like a violet...." What possibilities
does this establish for the next line? Violets are pretty to look
at, seclusive in behavior, perfumed, ephemeral, wild, dressed
in green, etc. The basic simile may be expanded in any of these
directions, or in any variant of them suggested by the subjec-
tive associations of violets for the poet. The different possibili-
ties of expansion are not, however, all compatible; for example,
in the next line the author cannot combine the notions that
"she will wither in the fall" and that "she is always fair to see."
Permanent and transitory properties of these alternatives re-
ciprocally interfere, and equal aesthetic emphasis cannot be
placed on both. However, whatever line of development is
chosen as central, the noncentral properties of violets remain
in the background, constituting a context which colors (if in
fact it does not establish) the "meanings" of subsequent images
in the poem.
The development of metaphorical mathematical imagery
renders itself most intelligible if we analyze it as analogous to
a poetic enterprise, and here the phenomenon of interference
may be even more marked, since the alternatives are even more
clearly disparate, than it is in the work of the lyric poet. If, for
example, four things in a set are unequal in respect to property
P, but tightly connected by proportion in respect to a related
property, Q, a linear representation of these four things as
segments cannot adequately present P' and Q as simultaneous
principles of order. The metaphor of proportion would require
the equality of two segments, whereas the metaphor of disparity
in intensity would absolutely preclude it. The diagrammatician
must then choose which of the two properties, P or Q, he in-
tends to make central and which peripheral to his illustration.
The -peripheral property is relegated to the contextual back-
 GEOMETRIC METAPHOR 93

ground and is effective in the figure actually constructed as a

felt, contextual alternative of relevant development.
A peculiarity of discussions to be illuminated by diagrams,
however, is that the property which is central in importance
may change as the discussion progresses. (Analogously, in a son-
net, the property chosen for expansion is often replaced by an
alternative in the final quatrain.) An accurate diagram can re-
flect this change in emphasis by unfolding in sequential stages,
so related that the property which is central at each stage of
illustration is the one that was central in the corresponding
stage of discussion. If, as may be the case, the schematic proper-
ties interfere, the unity of the schematism can be discovered
only by analyzing the sequential stages of construction, not the
final product of the process.
Actually, there are relatively few passages in Plato in which
shifts in emphasis require the reader, having started with one
central mathematical diagram, to shift to another. The sugges-
tion of the existence of such passages is put forward with some
reluctance, since it so easily lends itself to misuse. If we allow
the concept of separate but connected figures to dominate our
reading of mathematical passages, the way is open to bypass all
intended relations that present problems of interpretation by
treating each part of the diagram atomically, and radically dis-
connecting it from the others. It is precisely against this ten-
dency to disconnect the "mathematicar' passages from their
contexts that the present study is entirely directed. Conse-
quently, very careful controls must be set up to prevent this
technique from being overused. In defense of its being used
at all, however, the following two considerations seem con-
clusive. First, there are frequent simultaneous uses of several
mathematical metaphors in Plato's text, and it is impossible
that this should not produce some "interference"; second, there
are passages in which incompatible specifications of figures are
supplied. In addition, there are many cases in which diagrams
develop from the same elements in sequential stages, as shifts
in context suggest new interpretations of the relations of these
elements; and this creates, a priori, the expectation that a. simi-
 Plato'~ Alathernatical Imagination

lar device Inay be used in a single passage developing a complex

iluage or set of inlages.
This use of a postlllate of "interferel1ce" seems necessary and
not harlnflll so long as t\\TO conditions, to be stated, govern its
use. In the first place, a mathematical passage can shift its erll-
phasis by inclllding more tllan one figure only if the context
it illustrates contains an analogous shift of emphasis in content.
(In Section 6a of tllis chapter a more detailed discussion is
given of the ,vay to deterlnine to ,\That context the elements of
an image correspond.) The several figures ffillSt have to one
another a relation exactly analogolls to that of the portions of
context to which they correspond. In the secolld place, the
Ilotion of interference should not be employed unless the pas-
sages in question have produced antinomies of interpretation
in the Platonic tradition. By this is meant that there are al-
ternatives wIlich, though inconsistent, are firmly grollnded in
the text and deferlded by scholars. If in such a case the incom-
patibility canrlot be resolved, use of the notion of "interference"
to account for tIle antinomy seems justified. These two condi-
tions insure that there ,viII not be an evasion of real problen1s
by alleging incompatibilities where they are not explicable from
the context illustrated, and not already discovered by students
of the texts in question.
As an instance of sequential evollltion. of imagery made up
of the same elements, the sllifted mea11ing of an image of
concentric circles in the contexts of the Rep1.lblic-Tirnaeus-
Critias trilogy, discllssed in j\ppendix A, is a concrete case in
point.
The discussion of interference has been presented here be-
calIse the divided line is the first case of a geo,metrical image in
which the use of interfering metaphors appears.

Republic 509-11 *
Now take a line which has been cut into two unequal parts,
and divide each of t.hem again in the salne proportion, and sup-
pose the two tnain divisions to answer, one to the visible al1d the
'* 'Trans. Jowett. Dialogues.. III. 211-13.
 GEOI\IETRIC l\,IETAPHOR 95

other to the intelligible, and then con1pare the subdivisions in

respect of their clearness and ",rant of clearness, and you will
find that the first section [AB) Fig. 43] in the sphere of the
visible consists of images. And by inlages 1 lucan, in the first
place, shado'\vs, and in the second place, reflections in l\~ater and
in solid, snlooth and polished bodies and the like: do you under-
stand?
Yes, I understand.
Imagine, now, the other section [BG) Fig. 43J, of which this is
only the resemblance, to include the animals '\\rhich we see, and
everything that gro,\vs or is made.
Very good.
Would you not adlnit that both the sections of this division
have different degrees of truth, and that the copy is to the
original as the sphere of opinion is to the sphere of kno\vledge?
[i.e., AB: Be:: CD +DE: AB + BC]
l\tlost undoubtedly.
-Next proceed to consider the manner in which the sphere of
t he in tellectual is to be di vided.
In what manner?
Thus:-There are two subdivisions, in the lower of which
[CD, Fig. 43] the soul uses the figures gi\'en by the former divi..
sion as images; the enquiry can only be hypothetical, and instead
of going upl\rards to a principle descends t.o the other encl. In
the higher of the t\VO [DC, .Fig. ·-13], the soul passes out of 11YpO-
thesis, and goes up to a principle which is above hypothesis,
Inaking no use of images as in the forlner case, but proceeding
only in and through the ideas themselves.
I do not quite understand your nleaning, he said.
'I~hen I will try again; you will understand ll1e hetter when I
have made some prelinlinary remarks. You are alvare that stu-
dents of geometry, arithnletic, and the kindred sciences assume
the odd and the even and. the figures and three kinds of angles
and the like in their several branches of science; these are their
hypotheses, l\yhich they and everybody are supposed to kno'.v, and
therefore they do not deign to give any account of theIll either to
themselves or others; but they hegin ,vith them, and go on until
they arrive at last. and in a consistent nlanner, at their coneIu..
sian?
Yes, he said, I know.
96 l)[alo's ~\Iathcmatir,al lrnagination

~'\nd do ) ou not kno\~~ also that ~lthough they make use of the
yisible fornls and reason about them, they a~'e thinking not of
these, but of the ideals ,vhich they resemble; not of the figures
which they draw, but of the absolute square and the absolute
dialueter, and so on-the forms which they draw or make, and
,,\vhich llave shadovls and reflections in ,\\Tater of their own, are
converted by them into images, but they are really seeking to
behold the things themsel'ves, which can only be seen with the
eve, of the mind?
l.~hat is true.
And of this kind I spoke as the intelligible, although in the
search after it the soul is compelled to use hypotheses, not ascend-
ing to a first principle, because she is unable to rise above the
region of hypothesis, but flllploying the objects of vvhiLh the
shadows bela"v are resemblances in their turn as images, they
having in relation to the shadows and reflections of thenl a greater
distinctness, and therefore a higher value.
r understand, he said, that you are speaking of the provi!lce
of geometry and the sister arts.
J\ncl when .I speak of the other division of the intelligible, you
V\Till understand me to speak of that other sort of knowledge
,vhich reason 11erself attains by the power of dialectic, llsing the
hypotheses not as first principles, but only as hypotheses-that is
to say, as steps and points of departure into a ,vorlcl v/hich is
above hypotheses, in order that she l1lay soar beyond them to the
first principle of the wIloIe,: and clinging to this and then to
that ,vhjch depends on this, by successive steps she descends again
,vithout the aid of any sensible object, from ideas, through iueas,
and i.n ideas she ends.
I understand you, he replied; not perfectly, for you seC1U to
me to be describing a task which is really tremendous; but. at
any rate, I understand you to say that kno'\vledge and being,
which the science of dialectic contemplates, are clearer thon the
notions of the arts, as they are termed, '\vhich proceed froln hypo-
theses only; these are also contemplated by the understanding,
and not by the senses: yet, because they start from hypotheses
and do not ascend to a principle, thos~ who contemplate them
appear to you not to exercise the higher rc~son upon them, al-
though '\vhcn a first principle is added to th.etn they are cogniz-
able by the higher reason. i\nd the habit which is concerned
 GEO~'fETRlr: .\TETAPlIOR 97

with geometry and the cognate sciences 1 suppose that you wOllId
term understanding and not reason.~ as being intermediate be..
t,\\·een opinion and reason.
You have quite conceived Iny Ineaning, I said; and now, cor..
responding to these four divisions, let there be four faculties in
t.he soul-reason answering to the highest, understanding to the
second, faith (or cOllviction) to the third, anti perception of
shadows to the last-and let there be a scale of thern, and let us
suppose that the several faculties have clearness jn the same de-
gree that their objects have tru the
I understand, he replied, and give my assent, and accept your
arrangement.

· pl.l
R\.e ) lC ,):.1
,il· ~ ')4 r '\. *
We shall be satisfied, then, with the nan1es we gave earlier to
our four divisions: first, knovvledge; second, thinking; third, be-
lief; and fourth, imagining. The last two taken together consti-
tute the apprehension of appearances in the world of Becoming;
the first two, intelligence concerned with true Being. Finally, as
Being is to Becoming, so is intelligence to the apprehension of
appearances; and in the same relation again stand knowledge to
belief, and thinking to imagining. \Ve had better not discuss the
corresponding objects, the intelligible world and the ,vorld of
appearance, or the t\\Tofold division of each of these provinces
and the proportion in ,\rhich the divisions stand. v\Te lnight be
involved in a discussion rnany tilnes as long as the one we have
a~ready had.

Scllolion fronl Ianlblichus

The initial bisection of the line, ,vhether it is divided into equal
scgluents . . . as Iamblichus says, or into unequal, as some com-
mentaries say, represents an initial division of all things into
two classes, ,~n~~l!igible and s~nsib.1e. If the segments are equal
... then the relation of things participated in to those participat-
ing is similar, and the proportion [\vhich is the next stage of
Plato's construction] applies in the same '\Tay to both. If unequal,
as others say, then the relation of intelligibles to sCllsibles will
also be in excess [of a 1 : I ratio], and dissilnilar. . . . [The re-
ITlainder of the scholion recapitulates the proportions of the
-\!¢ rrl'ans. Cornford, Republic, p. 254.
98 Plat(/s J-1athematical lrnagination

divided line, and the things which are represented as objects of

kno,vledge by each of the four segrnents.] *
" Evidently, these commentators assume that the greater reality
al1d clarity are to be represellted by the longet' line. If the inter-
pretation of the De ,,1 nitna figure given below is correct, this
is another proof that no continuous tradition connects the
figllres of HelleIlistic scholars \vith Plato's original designs.
The point at issue is whether the analogies are to be can..
strlled as representing only relations, in which case those of
intelligibles and sensibles are similar (i.e., mathematics: imita-
tions of fornlS: : shadows: ilnitations of sensible objects), or
\vhether the difference in status of relata is also reflected in the
properties of the figure. OUf 0"""11 contemporary practice seems
to be to draw the diagram in one way, tllen il1terpret it in the
other.
rrhe figure referred to here has properties wllich callnot be
combilled by any geometrical construction. Either we can make
all four of the segments unequal, or cut tIle two divisions in
the saIne ratio; but not both. 23 Since the relation schematized
seems too complex for a single figure, two figures are in effect
needed: the one "\vould bring out differences in clarity and
adequacy of knowledge by differentiating lengths of segments;
the other would brirlg alIt analogical relations bet"\veen knowl..
edge of tllese kinds by proportioning lengtlls of segments. Evi..
del1tly this is a case of interference; the trans-spatial, trans-tem-
poral characteristics of the forms lend themselves rather
inadequately to spatialized representation. This double repre-
sentation parallels a contextual shift in emphasis. To establish
the character and proper wisdom of the philosophic rulers in
Book vi, it is l1ecessary to cont1'ast their knowledge to the erratic
opinions and llypotheses of nonphilosophic statesmen. Bllt to
achieve a cllrriclllulll for the training of these Tlliers, in Book
vii, it is llecessary to provide a transition based on the resem-
blance of kno,vledge of each kind.
The nletapll0r of ll11equal segments was used by Plato else..
• Appendix Platonica, eel. C. F. Hernlann (Leipzig, ]8~5), pp. 350-51. Other
scholia will be found 'with the figures at the end of this section.
 GEOMETRIC METAPHOR 99
where to underscore the cont'rast,24 but if this contrast is tIle
final ,vord that can be said, and if we COIlstruct the divided
line '\vith all segmellts unequal, the only interpretation one can
give the figure is that of a Stoic or Parmenidean division of the
igllorant from the ,vise; and education seems impossible. In
fact no interpreter, in the face of the stress 011 proportions in
the passage, has had the telnerity to propose this as the intended
constrllction. On the other lland, if the reference to ineqllality
is fllnctional and deliberate, its function must be to call atten-
tion to an alternative line of metaphorical developmellt, needed
to llnderstand the relation of the line chosen to the fllli situ-
ation, one aspect of ,vhich it characterizes. It is the relative posi..
tion, not the relative length, of segments in the proportion
figure on which the mathematical statements of analogy are
based. "Higher" and "lower" are the key spatial concepts in
this analogical interpretation. Actually, this is so evident in
context that relatively few Plato stuclents have been bothered
by the fact that if we interpret the proportions as relating ab-
solute lengtlls, the second and third segments of the line come
alIt equal. They are clearly uneqtlal in respect to higher-lower
position.25
The metaphor of inequality seems intended, in its other uses,
to represent a distance between the knower and the object of
knowledge; the longer the segment, the greater the distance so
represented, and the more diffuse the unity of the form to be
apprehended. On this basis, the longer segment must represent
the less adequate knowledge, and insofar as the metaphor of
ineqllality is retained (in the final figllre) in its init.ial division
into llnequal segments, the longer segment is not honorific, but
significant of a greater negative factor. (Similarly, the Pythag..
oreans tended to assign greater value to small rather than large
numbers.) 26 This consideration is decisive in determini11g the
relative lengths of the "knowledge" and "opinion" segments of
the constructible proportion figure, and it is harcl to see how any
editor could have t.llOllght tllat a modern metaphor involving
the notion "the bigger, the better" could conceivably have been
,\That a Greek of Plato's time intended.
100 Plato's Jvrathf.Jnatical IJ1u:gin.ation

i11is last p2.rag~~;\ph rClninds us that the intcrfcrCIlcc of the

two principles llJ 'FJa~!al r?p.f:st:ntation is Eot a .complete one,
for as bct"rcc;l the l.naJor dIvIsIons of tIle proportloll hgure, and
\\'ithin c;-;.cl1 d~yisioll, the metapllor or ill~qua.~}ty raay be, and
is, retalJlcd ,",v'ithollt geon1etrical ilnpossibility. ~''j
Since the ratios given ,\t·ill ""TOIl<. equally "\\'c11 in ternlS of
higher-]o\r~r jf tIle segmcllts ~re ITlade all th.e same lel1gth, the
figure is oftcll dra1'\ n in this 1vay; and it ,viII be suggcstetl that
r

Plato l1ilIlself so repre~ents it in later synthetic iltlagery, ,\There

stress is only on trle nun'zbc'r of stages relevant to classifying
kinds of \\iisdom. I-Io"\\Tcver, this representation, though it l)rings
out the silnilarities of different killds of kn()"\\'ing" central to
Hook vii. :Jeems to ignore the context of Book vi, in which the
total diffcT('nCe in kinel between these sorts of kllovvleclge is
uIlderscored.
In Sllmnlary, theIl, the present COlltentio11 is that two alterna-
tive diagrams are relevant to schen1atizillg the differences and
~~jnJilarities of kind.s of kllowlcd~'e at a point of transitioll at
\vllich the present figure appears. Otl the one hand, the relative
separation of krlo1vcr alld knt}\Vll can l)est be represented by a
scllematizatioll in VJhich ltncqnal lengths characterize knowl-
edge of each kinel, ~uld the relative lengtll of the segnlents is
the property underlying the 111ctaphor (,,,~llich is Olle of dis..
lance). all the other 11311d, analos"ical connection call best l)ring
out the comnlOll nature of all knowledge, whicl1 is fundalnental
to tIle scheme of education, al1d this analogy is based on the
nlctapllor of the relative position of segl11ellts in respect to
height on the Iirle. The two figures canIl0t be corllpletely com-
billed; at best, the property of inequality of lengtll can be trans-
ferred only to SOlne of the proportiol1ate relatiolls. The meta-
phor of proportion is chosell for full developlnent because, for
the tral1sition at hal1d, it is the most importaIlt. The metaphor
of incqlla lity is. ho,\vever, retained in the figure as a remin,ier
of a needed qllalification, lv-hiell Book vi has nlade clear. The
situation is S0111C\v!lat analogolls to the relation of tern pera11cc
and justice in the state; the sinlilarities t)et'\TC~ll knowledge of
d i l{cr\.'nt kinds must be seen and tIle ditferellces not confllsed. 2 f-;
 G)~:O \IETRTC :\ fETAPHOR 101

'The \'exed q u('stiOll of the relation of tbe t\\'O upper scgu1cnts

of tlle di" ided line. ill tcrnlS (>{ an a(lc(lll~H.C dctcrrnination of
the relations of cont.'cpts an.cl Sch~lllatisln::), "\villl1ot be separately
discussecl here.:'::; 'l~lle reason is that this elltire Fresent stlldy
of Plato's usc of SChCITlatisms to cxpla in and ill urn i 113,tc concepts
is a pl1ilosophic in(luiry into pl'cciscly tJlat problctll, alld l1oth-
iIlg briefer vvoltld be adt:qu~ltc to :'illch a discl.I,;sion.

Figure -13
- . - . . '1ilL..lI.. - ....~ .......

IE
~
Science
I c f. ScrrJtes

. .'o--..--....-h--
---~-- .. ---
f-1 ypotht~is
c f. GLulcon

Experience
c f. 'lhrasyulachus

----+---- B

i Conjt·{.ture

( f. C:ephalus,

I Pol(:'nl~r(bus

I Tlll: DIYIDED L l:"E

1"'his figure is lettered to correspond to the references in the

r preceding' text. In parentheses are indicated the naInes of speakers
l '\",ho have appeared e~rlicr in the dialogue and who seem to h~rv'c
kno\vledge of the nature of justice at each level of adequacy. There
i.s no clear agreement about what exactly constitutes conjectllre"; H
102 Plato's i\lathernatical llnagination

it rnay be kno\vlcdge in the sense of t.he Kantian nlanifold of repre-

I
sentation; or it rnay be.. in vic\v of the stress OIl iluages and reflec-
tions, that Plato had in mind the sort of thinking analyzed by
I
\rergil C~. i\ldrich in his article "Pictorial l\1eaning and Picture
'I'hinking," Kenyon Review, 5, 19-13 (pp. 175-181 in Feigl and
Sellars, Readings in J)hilosophical ~41lalys£s) New York, 1949).. In the
latter case, the criterion of Ineaning vvould be imaginability, the
inadequacy a lack of elnpirical verification of the image con-
structed; in the forIner, the inadequacy ,,,auld lie in the absellce
of any concept to unify the subjective representations into objects.
'I'he second segrnent of the line, from Plato's later account of the
ganie of guessing at sequences of shadows in the cave, seeIng. quite
clearly to involve a pure enlpiricism, similar to Hunle's interpreta-
tion of causality (as an expectation, based on past experience, that
inlpressions "vhich have been contiguous in space and time will be
so in the present case). The third segulent is quite clearly a kind of
axioluatic, deductive kllowledge best exen1plified in mathematics;
by suggesting in the figure that Glaucon may be said to have had
in mind knowledge of this type in his explanation of hUInan be-
havior by a deductive development of the social contract theory,
I intend to suggest that any subject-matter is capable of such
axiomatic fornlulation, the proper excellence of which is its coher-
ence in bringing out relations and causal connections, the weakness
of lvhich is that such coherence never guarantees the rightness of
the axioms (or at least does not seem to a Platonist to do so).
In that case, the final segment of t.he line shares the clarity of
deductive structure with the third, but is based on "better" hy-
potheses; better either in the sense that there is no possibility of
deducing a contradiction from them (Plato will later say that the
nlan who has knowledge of the good must be able to discuss it
"without tripping"), or that they take account of a wider range of
data to he explained.
rrhe representation in Figure -i'1 brings out the differences in clar-
ity of knowledge of these types but loses the sirnilarities basic to a
systeul of education. (For example, the ratio of knowledge to opinion
in this figure is 6: 1; that of conjecture to experience, 2: 1.)
'I'he figure is described by Plato in the Lau!s to ~ho\v the diffusion
of causal efficacy from the forn1 to its space-tillIe elnbodinlent.
In De anima -:10'lb, .1\ristotle points out that H[PIato] says . . .
that the un ivcrsc is derived Iron1 the idea of the one and the first
 GEOMETRIC rvrETAPHOR 103
Figure 44

VOVS

SiaVOla
2

TriO-TIS

8
ALTERNATIVE CONSTRUCTION OF THE DIVIDED LINE,
CARRYING OUT THE INEQUALITY OF SEGl\IENTS

length, breadth, and depth, and all other things in the same way.
And further, vou~ is one, and €JtlO''t~~t'Y) is two; for the latter is
generated from the one; and the number representing ()6~a is a
plane, that representing ata{}'YIaL~ a solid; and he [Plato] says these
numbers are the forms and principles, but they are composed of
elements. Sometimes things are known by reason, sometimes by
understanding, sometimes by opinion, and sometimes by percep-
tion; but the form of things is these numbers."
Compare Laws 894A: "What must happen to anything to cause
its generation? It is as if a principle being augmented comes to a
~ccond type of change, and after this goes on to the next, becoming
perceptible through this third."
In each of these analogies, the notion of a diffusion of a principle
from its existence in an act of noetic insight to its embodiment in
 104 Plato's Mathematical Imagination

a sensible medium is represented on the analogy of an arithmetical

progression. In the passage from Aristotle, this progression is
1:x:x2 :x8 • In the Laws passage, the same notion of different dimen-
sions separates the "increases" of each type.

v. MATHEMATICS IN HIGHER EDUCATION

The discussion of the mathematical curriculum in Republic

vii is more relevant to a treatment of the philosophy of mathe-
matics or of the history of mathematics than to the present
enterprise of interpreting mathematical imagery.
However, certain comments indicating the relevance of this
, curriculum are in order. In the first place, whatever the basic
philosophical detail intended, the statement is clear that .obj.c.cts
studied by the _omatltematkiAlLire the images of the ideas
il!YeStigat~CL.b:y--the"dialectician.80 Apparently the idea, in an
I understanding of which there is a grasp of some perfection or
function, has associated with it a quantitative schematism or
structure, in which an image of the form itself can be recog-
nized, and through which the form would be embodied in space
and time 81 as copies of it were constructed. (This is very remi-
niscent of the role of the Kantian schemata, through which
connection is established between the concepts of the under-
standing and the forms of intuition.) The theoretical statement
corresponds exactly to the relation of dialectic and mathe-
matics that we find Plato using in practice. The various rela-
tions of ideas are schematized in mathematical imag~tY, with a
transposition of their nature into that of quantitative analogues,
to illustrate and clarify a contextual, dialectical method or
schema.
A second point important for the present investigation is
I that the final mathematical study is harmonics, a study of the

concord and discord of numbers themselves. 82 The concords

of music will furnish images and illustratiOfs of this type of
mathematical structure. 33 As Socrates suggests in a later remark,
the goal of this education is to enable students to see things
"synoptically," and the order of the curriculum is one in which
 GEOMETRIC METAPHOR 105

FIGURES FROM SCHOLIA

(Greene, Scholia Platonica, pp. 245-46, 253)

• Figure 45
Republic SlOB
3 4
Mathematics Dialectic
from hypothesis from hypothesis
to final conclusion to non-hypothetical principles
through sensible images through the forms themselves
'You;

Figure 46
Republic 508A
Good Sun Sun's Light Darkness
Nou~ Sight Truth Likeness
Intelligibles Sensibles Knowledge Opinion
Place of Forms Place of Visibles Being Becoming
Knowledge Vision Sight Blindness

Figure 47
Republic 510D
Vision Knowledge
Opinion Intelligence
Sun The Good
Shadows Nature Mathematics I Divine
Things
Conjecture Opinion Understanding Knowledge
Becoming Being

Figure 48
Republic 534A •
R o
E KNOWLEDGE BELIEF P
A BEING . . . . . . . • . . . • • . . . • . . . . . . . . BECOMING I
S UNDERSTANDING . • . . • • . . . . . . . . . CONJECTURE N
o 2 4 I
N o
N

• Rows of dots represent horizontal brackets in the original figure.

106 Plato's Mathematical Imagination
each successive science presupposes the principles of the one
preceding, but integrates them with new attributes peculiar to
its own treatment. 34 In this order, harmonics comes last, medi-
ating the transition from mathematical to dialectical study.
Although he goes into some technical detail in specifying the
method and content of the other curricular studies, Socrates
dismisses this final subject with no technical information other
than the statements (1) that its images are the concords of
music and (2) that it must not be modelled on the empirical
study of music of various instrumentalists and Pythagoreans.
Consequently, a good deal of commentary has been devoted to
remedying this brevity of technical treatment, much of it cen-
tering on the problem of what mathematical concept "numbers
concordant and discordant in themselves" can possibly mean.
The explanation for this lack of technical description is rela-
tively easy, however; Socrates goes on to a detailed discussion
of "dialectic," and there is almost nothing to be said about
harmonics to which his subsequent remarIss do not apply. A
difference in principles (quantity in the one case, nature in the
other) is the only point of separation between the study estab-
lishing mathematical concords and the establishing of harmony
in states and souls, which is the politically relevant objective
of the course in dialectic in this plan of higher education. After

Figure 49
PRINCIPLES OF THE MATHEMATICAL SCIENCES
SUBJECT MATIER PRINCIPLES
SAME OTHER
Sets of units prime factors composite factors
Numbers odd even
Figures and solids circle & sphere straight
Straight figures and solids regular , irregular
Motions rotation translation
Translative motions regular irregular
Proportions geometric nongeometric
Nongeometric ratios harmonic arithmetic
 GEOMETRIC METAPHOR 107

the study of the structural features of dynamic systems in astron-

omy, one may safely conjecture, in view of the ideal of synop-
ticity, that the interrelations of numbers, figures, solids, and
motions already studied will be taken as one example of sys-
tematic organization.
It is entirely in keeping with the important place given to
analogies of harmony throughout the Republic that a study of
the theory of harmonic structures be the final point of a
student's instrumental education, immediately preceding the
more complex inquiry into the conditions by which harmony
can be realized in human careers and in the administration of
cities.

VI. THE NUPTIAL NUMBER

a. Introductory Remarks
A good schematic illustration of a discussion must meet two
conditions: in the first place, there should be no doubt as to
what the geometrical or arithmetic structure of the schematism
is; and, in the second, there should be some clear indication of
what element of the discussion each part of the illustration is
intended to represent. In our contemporary use of diagrams,
there are typically two distinct vocabularies at work: a mathe-
matical vocabulary of construction J and a nonmathematical
vocabulary (taken from context) of interpretation. If, however,
the mathematical vocabulary has been borrowed from non-
mathematical fields, such as biology or physics, and if a feeling
for the basic sense of the borrowed term is still operative, there
is evidently no need in every case to separate the two vocabular-
ies of interpretation and construction.. A modern mathema-
tician, applying topological concepts to neurology, might well
speak of "nerves" meaning both the technical topological ele-
ment establishing a mutual relevance of areas, and the physio-
logical element supplying connections between parts of an
organism. Insofar as he had borrowed his nomenclature from
108 Plato's Mathematical Imagination

biology because topological and biological "nerves" have the

same basic properties, his metaphorical terminology would
work equally well whether it were the abstract aspect of the
diagram or the concrete function of its intended interpretation
that the reader supplied. SCi
This point seems fairly evigent when we take a very fresh
mathematical term (ay,d we would not be surprised to find the
medical topologist giVing a diagram of "nerves" without a sepa-
rate vocabulary of interpretation appended), but as mathe-
matical terms become more shopworn and are more frequently
used in their exclusively structural sense, their nonmathemati-
cal reference is deliberately and effectively bleached out of
them, so that a term of long standing will not be felt to carry
with it any suggestion of interpretation. This point, of course,
has been already suggested in the Introduction, where the pos-
tulate of a tradition of mathematical metaphor in Plato's time
is presented.
What is peculiarly relevant, however, to the passage about to
be interpreted is more than this general sense of metaphor in
the mathematician's vocabulary: it is the ability of a given
mathematical term to indicate at the same time an intended
construction and an intended interpretation.
This use of terms with a deliberate selection which makes
them significant on any of several levels of meaning is, of course,
an aesthetic device that is familiar to contemporary authors
and readers. ss It is only in an exceptional case, however, that
a contemporary scientist or mathematician would feel at all
comfortable when confronted with it. Such an exception, if it
occurs, will do so with a newly appropriated term which still
retains a felt extra-technical reference.
However, the mathematical vocabulary of Plato's time had
very recently appropriated words from the gefteral vocabulary
for its special needs of designation. With this in mind, it does
not seem unlikely that a writer of the period ~ight make a
single term do the work both of technical specification and
interpretation. Why, for example, would it be necessary to say
that an "irrational" line is like an "irrational" man, when the
 GEOMETRIC METAPHOR 109

resemblance is already embedded in the etymology of the name

of the line? If the metaphor that seems natural to the mathe-
matical giver-of-names still remains apt for the later constructor
of diagrams, there may be no need felt to append a key to inter-
pretation, since the metaphors embodied in the technical names
themselves provide an adequate statement of the interpretation
which these structural elements are meant to represent.
This practice of making names do double work carries over
a feeling of affinity between figure and interpretation, and
probably gives a reader in the same tradition a sense of natural
fitness or translucence in the construction of the diagram.
It may also explain the intended interpretations of passages
where descriptions of construction are given, with no separate
dictionary of interpretations appended. The suggested rule of
interpretation for dealing with this device is that if no indica-
tion of interpretation is given, a term occurring in a technical
mathematical sense in a construction may be intended to stand
for the concept represented by that same term when it occurs in
a less restricted sense in the relevant context.

b. Problems of Translation and Paraphrase

Prologue to the Number •
In what way, then, will our city be moved, and in what manner
will the two classes of auxiliaries and rulers disagree among
themselves or with one another? Shall we, after the manner of
Homer, pray the Muses to tell us 'how discord first arose'? Shall
we imagine them in solemn mockery, to play and jest with us as
if we were children, and to address us in a lofty tragic vein,
making believe to be in earnest?
How would they address us?
After this manner:-A city which is thus constituted can hardly
be shaken; but, seeing that everything which has a beginning
has also an end, even a constitution such as yours will not last
forever, but will in time be dissolved. And this is the dissolution:
-In plants that grow in the earth as well as in animals that move
on the earth's surface, fertility and sterility of soul and body
occur when the circumferences of the circles of each are com-
• Trans. Jowett, Dialogues, III, 80~.
, 110 Plato's Mathematical Imagination

pieted, which in short-lived existences pass over a short space,

and in long-lived ones over a long space. But to the knowledge
of human fecundity and sterility all the wisdom and education
of your rulers will not attain; the laws which regulate them will
not be discovered by an intelligence which is alloyed with sense,
but will escape them, and they will bring children into the
world when they ought not. . ..

Translation of the Number

No "neutral" translation of the rest of this passage can be
found. In the two translations that follow the literal sense most
relevant to interpretation is presented in the first, and in the
second, the meaning most relevant to mathematical construc-
tion. Both these levels of meaning operate in the passage simul-
taneously. An appended note gives the translation which results
if the text as accepted by Proclus and later editors (except
Martin) is used.

Level I: Interpretation·
For divine becoming, there is a period comprehended by a per-
fect number; but for human, by the first in which developing
capacities, dominating and dominated, on realizing three stages
determined by four poiI!ts [in the field of these processes of]
becoming like and unlike, growing and declining, make all things
conversable with and rational in respect to one another. Of
these [the element representing referents in] a ratio of 4 to 3, in
lowest terms, married to the pempad, produces two harmonies
when thrice augmented. The one [harmony] is a resultant of
equal components, the same arithmetically in each dimension;
the other, while equal to the first in one dimension, is unequal
in the other. Each side of the former [equals] numbers produced
by squaring "the diagonal lines" that represent,a "rational" com-
ponent of the pempad, each diminished by one; [the other side
equals nurrtbers produced by squaring the "diagonals" that repre-
sent a component of the pempad which is] "irrational," [each of
these diminished by] two. The latter [harmony] equals one
hundred cubes of three. And this whole thing is a geometrical
number....
• General meanings are emphasized here.
f

I
GEOMETRIC METAPHOR. III

Level II: Construction"

... but for human, by the first [number] in which increases of
roots and squares, when they reach three stages with four bound-
ing points [in the field of processes] of becoming like and unlike,
increasing and decreasing, show all things related and in integral
ratio to one another. Of these, [that part represented by] the
integers 4 and 3, joined to five, produces two products when
made three-dimensional. The first is square, equal in each dimen-
sion; the other is equal to the square in one dimension, but
rectangular. Each side of the first [equals] numbers [represent-
ing] squares on rational diagonals of five, each minus one; or
irrational, [each minus] two. The second is one hundred cubes
of three. And this whole thing is a number summarizing the
dimensions of a figure. . ..

If we accept the version of the text given below in the notes

to the Greek text of this passage, the translation is changed:
The first harmony becomes "equal to IOOx in each dimen-
sion," instead of "the same arithmetically in each dimension."
This harmony has its side produced from "one hundred num-
bers [representing] squares on rational diagonals of five, each
H
minus one, or irrational, each minus two instead of "Each side
of the first [equals]...."

Republic 546A-D
(Chambray, Republique)
"Ea't't BE {}ELq> ~Ev yEVV'tlt'Ql JtEQ(aO~ ~v d.QL{}f!O~ JtEQtAa~~aV£L
"'EAELO~, aV{}QOO1tElq> BE EV ro
JtQdl1:q> au~~aEl~ ~"vciJ!Cva( 1'£ ')(al
a"v(lat'E,,6~EVUl, t'QEl~ &'rcoat'aoEl;, l'E't'taQa; as
oQou; A-a~ouaaL
0JlOLOVVt'WV tE xal &.VO~OLOUvrroV 'Xal au;6vtoov 'Xal <p{}LVOvrOW,
navra JtQOOt1yoQu xal Ql)'tu 1tQo~ ({)J.flAia a.3tE<FllVav· rov E:rd1'Ql'tO~
n1J{}fl~v JtEfl:rcd.at (j1Jt"YEl~ avo dQf..lovta~ Jta(>EXE'taL tQl~ aU;l}aE[~,
'tl}v ~v tOll v laaxL;, E'X(lO'tOV iocramaXt;, 't~v ae
taollrpt'll !lEV 'tn,
:TCgOf!tlX'l') aE, EXUcrtOV ~EV UgL{}f!OOV &.11:0 ataf.!E't'QOOV Q'l')'toov JtEflJtdao;,
~EOf..lEVWV EVO; Exacr'toov, uQelt'toov ~E ~'UOLV, E'Xatov ~8 'X{,~rov
l'QLa~o~.
Exao'tO'V AF: f;')(C1'tOv A2Proclus
'exC1O'''COV AF Pap. 4: 81(a'tOv AI

• Here specifically mathematical meanings are taken.

 112 Plato's Mathematical Imagination
=
(A cod. Parisinus 1807; A2 = later correction of A; F =cod. Vindobonensis
=
55: Pap. 4 Papyrus Oxyrhynchus XV 1808, column 4.)

If Plato had anticipated the controversies and interpretations

occasioned by this passage explaining the principle of political
decline, he would probably also have anticipated the example
, of some of his recent translators, and deleted it from his text. 37
For the purpose of a mathematical image in Plato's writing is
to schematize and clarify dialectical relations in the context,
and if it fails to do so, it is dialectically nonfunctional. Although
this conclusion follows immediately from an examination of
the entire set of mathematical illustrations used in the dia-
logues, accidental factors have combined to produce a few spe-
cific cases which are apparent exceptions. These exceptions have
been so disproportionately stressed that it is they which have
led to frequently repeated scholarly generalizations about
Plato's use of illustrative problems and diagrams. It must seem
paradoxical that the author capable of describing and explain-
ing the diagram of the divided line in Book vi of the Republic
should by Book viii have decided to "throw a little mathematical
dust in our eyes," with an elaborate enigma designating some
simple biological period, or signifying nothing; or that the
author who worked out the sequential development of the ar-
gument of the Republic could not resist interpolating a com-
pressed, technical, and digressive footnote on cosmological em-
bryology at a point of transition in his text where a summarizing
diagram of the antecedent discussion would have been more
naturally called for.

c. The Philosophy of History

..
There is no doubt that Plato's imagery here is intended to be
applied to human genetics as cause of the decline of the state
and is to be considered a concrete example illustrating a general
principle of the philosophy of history. But the moral of the
;philosophy present in this passage, as contrasted with the earlier
;remark about the inevitable cyclic progress of the state institut-
)ng educational reform, comes as something of a shock to the
i

I
 GEOMETRIC METAPHOR 118

tradition that Plato is an arch-conservative, and stands in ap- \

parent opposition to his repeated prohibitions of innovation in
the arts. The inability of the rulers to master completely the
recalcitrant auxiliary causes which upset the clear mathematical
genetic theory is typical of the basic moral, that no rigid attempt
at retention of form can endure in time. Adaptation, not crys-
tallization, of institutions and classes seems to be the precondi-
tion of progress. (This of course is why the rulers in each
generation must know the nature of the good, so that they do
not confuse tradition and value.) Just as cosmic process leads
each living organism to overshoot its mature acme in the tra-
jectory of a life cycle, so the flux of time and change will deform
any structure literally imposed in its plastic medium, and here
in the Republic, as well as in the later dialogues, some inter-
mediate tactics must intervene to secure the preservation of
intelligible structures in their embodime'nts as objects in the
world of change. If we read Plato's strictures on inll0vation as
attempts to avert changes in tactics rather than as attempts
literally to embalm specific institutional structures, the contra..
diction between the notion of a life-cycle of the state and the
ban on innovation seems to disappear.
d. The Rulers' Problem
The basic problem of the rulers in planning marriages is pre-
sumably not (as many interpreters have thought) astrological,
since it has already been decided that dates of marriage festivals
are to be set whenever needed to maintain the stable population
of the state. Neither is it likely that the problem referred to
here is that of determining the proper age for marriage, for
simple observation of normal human maturing has already lead
to a definite decision on that point. Nevertheless, the context
clearly indicates that this problem is closely associated with
some property of the human life-cycle. The nature of this prob-
lem is suggested by a comparison of earlier specifications in the
text of the respective ages of procreative and intellectual ma-
turity. The stages of human growth are such that cOllples must
be selected to marry and have children before they can have
 l;)lat()'~ Alathernatical Irnagination

COlllpletecl the trailling and selection which finally determine

who are best fitted by l1atllre to be ftllers. Nature thllS forces
the hand of the arrangers of rnarriage, since their calClllatiolls
can be based only on the estiulated futllre Sllccess of the couples
they select. .By the time they are 01(1 enough to nlarry, tIle
allxiliaries ,,,Till have been tested and gradecl all the basis of
their elernerltary and 111ilitary-lnathematical edllcation-in
ternlS, in other ,vords, of their dem011strated a bilities in control
of appetite and obedient exercise of spirit. I-Iow reasol1able
they "viII become as tiley lnature intellectllally can be (lecided
oilly by scoring their performance ill tIle course of theoretic
dialectic arId applied dialectic (civil adnlinistration), but by the
tilne this training is completed, they are too old to have llealthy
children.
The rulers, therefore, Blust decide suitability of parents on
tIle basis of delllonstrated nlilitary and lnathematical ability,
which Inay llot be a fair index of intellectual and administrative
capacity passed on to the chtldren. lInder these circumstances,
practical prolliems of genetics as "veIl as theoretic considerations
of possible constitlltions lead t.o the conclusion that later genera-
tions of rulers may be born who ,,~ill be too congenitally domi-
nated by spirit to carryon the tradition of responsible rule in
the aristocratic state.

e. vVhy the lVIl.ISeS Speak Playfllily

The playfulness of the Muses may be takell, together with the
lnatllematical joke in the Statesman)38 to provide instances of
what might be called Plato's 'Imathematical irony." When his-
tory or immediate experience is schelnatizec} lnathematically,
t.here is a great discrepancy between the lleatness of the theory
and t.he confusion of our perception of the imrnediate fact to
\vhich it applies. There is all ironic inappropriateness in the
forlllulation of a Ileat textbook of genetics and ellg-enics "vhich
is SlIre to l>e elnpirically inapplicable and transitory in hllman
history. Tllere is a "tragic over-pretentiollsness" in developing
all elaborate intellectllal schematism of phenomena whicll, be-
calIse of their inconstant character, ,,~il1 not behave as schema-
 GEOJ\iIETRIC METAPHOR 115

tized. '-rhis does llot luean that such sC}lenlata are false or not
useful; they are in fact the only \vay ill whicll the interacting
factors of a cOlnplex situation call be intelligibly exhibited and
analyzed. But tIle 1\-1 uses' neat schelnatic strllctllre of tIle callses
of social degradation, though 11l1derstandable and true, still
leaves its possessor llelpless to attribllte to data dependent on
perception and calculation an accllracy that these do not possess,
and hence helpless to avert the schematized historical progres-
sion.
/. Detailed Interpretation
i\. paraphrase of the passage follows. N umbers in parentheses
refer to passages in which the statenlents made in the paraphrase
are discussed in detail. Tllese passages are identified by means
of numbered side headings.
"
Indexed Paraphrase
"[he various factors which interact to determine the Ilature
f
uf a human life are not related as iIIvarial1t, proportionate re-
curring cycles, as the motions of the heavens are. (1) Although
calculation and perception are never absolutely accurate, as-
I trol10Inic calculation of celestial periods attains considerable
accuracy; but for human lnaturity arId reproduction, the sim-
plest logistical representation must take account of a process
of growth (as opposed to SilTIple IOC01TIotion) in which poten-
tialities are realized in different orders of dominance, a realiza-
tion taking place in constant interaction with a flucttlating
environment. (2) When this process of gro\vth has attained four
successive points, through three stages (like a line of three seg-
Inents, divided and bounded by fOtlr points), (3) the organism
tllat has grown and the number representing it ,viII constitllte
a rational and commensurable pattern. (4) This ratiollality is
attained at the acme of the period involved.
'1\\70 proportions which deternline the l)eriods underlying
rnarriage regulation (5) are prodllced from (6) the joining of
the line represerlting the three stages of growth of this organism
to another line, in 4: 3 ratio, by a line of length 5 (7). (The
joining thllS prodllces a figure llniting the original line of three
116 Pltlto'~ Jl[alltenLatical Jrnagination

segments \vitll the two other lines.) (8) rrhe image ,vhiell is a
geometric representation of these t'vo proportions has one
square and one rectangular plane surface. (9) TIle le11gth of
the side of the square surface is equal (numerically) (10) to
(II) (the llllInbcr representing) (12) the square of (13) tIle
"rational diagonal (14) of five" (15) minus one (16); and this
length is also equal to the number of the square of the "irra-
tional diagonal" (17) minlls two (18). [This completes the
specification of the square image (19).] The recta11gle is equal
to the square in one dimension (20); the "rectanglilar harmony"
is represented, arithnletically, as 27 x 100 (21).
This whole number, slllnmarizing tIle proportionate relations
of tIle geometrical Image ,vhich specifies the number's inter-
pretation, gives the formula or pattern of the interacting factors
which the rulers must apply, by calculatioll and perception,
to each individual couple in their arrangemellt of marriages;
and if they make mistakes, the children born are not naturally
,veIl endowed and do not have good fortune.
1. Divine vs. human generation. The basis of the f!EV-~E con-
trast between periods is not only that one cycle is divine and the
other human, but, more important, that the former can be
reliably observed and calculated (and its analogy to the "con-
cordant ratios" of harmonics is recognizable, so t11at the astrono-
mer can see these periods comprehended by a 1'EJvEIO'; number),
whereas the human periods cannot be observed or calculated so
adequately because the number comprehending them involves
a rnllltiplicity of processes and factors which require elaborate
geometric schemata, far removed from the simple proportions
of music. r-rhe common feature of divine and human becoming
-its periodicity-is reintroduced and emphasized again at the
end of the number; btlt until that reintroduction, the stress is
not on some proportional connection or si1nilal'ity of the two
periods, but on their totally diffe1"ent degree of observability
and regularity.
2. Mathern.atieaI metaphor·s of growth. " ... a{'1~11oEt.; 01,vaf.!EVUI.
1'c 'Xut ()ova(rtE'U6~(E'lat" introduces a mathematical metapl10r
which is not associatell with a specific diagram llntil later in the
 GEOMETRIC ~IE'IAPHOR 117

passage. r-[he period of human life is a pl'ocess of growtll in

,vhich potentialities are realized, and these realizations then
serve as potentialities for further development. The most rele-
vant aspect of this process for the rulers to observe is the growth
to the acme of tile period, but the two moments that comple-
ment each other in the complete cycle are noted in ai,;ovtw·v
xal c:p-frlVOV1'CDV. TIle contrasted pairs of participles emphasize
the fact that this cycle takes place in a matrix of qualitative
al1d quantitative change. The relation of "assimilation and dis-
similation" to Hgro'''~tll and decay" is one of external to internal
relations, which recalls the description of growth in the Timaeus
(in which the flux pouring through the body disrupts the
natural movements of the soul, while at the same time a more
intricate and better organization emerges as the organism grows
and progressively disengages itself from tIle immediate impact
of this flOW).39
3. The line of three segments. On encountering this phrase,
the reader of Plato's time, who did not expect his text to con-
tain illustrations, would probably have reached for his tablet
and stylus and noted the indicated "figure." ". . . t'QET;
(!.;(OaLa(jEL~ • • ." is precisely the sort of direction for construc-
tion that Plato incorporates in mathematical passages to assist
the reader in visualizing or dra,ving intended diagrams. 4o l\
more extended "direction" or "description" of the same kind
is given by the phrase discussed in (9), following.
4. Conversable and commensu·rable. The full development of
the period to its acme is marked by tIle emergence and domi-
nance of intelligence. (Tllis also parallels the account in the
Timaeus J in which the study of philosophy and cosmology are
intended eventually to re-establisll the circles in the soul in
their proper natural functioning.) 41 The "things" wllich are
"rendered conversable and commensura bIe" are both the struc-
tllre of the organisln, 1vhich attains full actllalization, and the
relations of the objects· of kno\vledge which are seen in their
true relation only by an intelligence fully perfected and "akin,"
but '\vhich exist in and constitllte such an intelligence.
.A further mathematical metaphor is also introdllced by this
118 Plato's l\lathe"latical Imagination

phrase: proportions and comn1ensurables are related to roots

ill the same ~vay that intellectual insight or maturity is related
to native capacity.
5. The tlOO harmonies. A detailed schematism, based on the
mathematical Inetaphors introduced generally in the first part
of the passage, now begins to emerge. Just as the nllmber com-
prehending the pel:od of divine becoming is a product of its
constitutive proportionate subperiods, so the nllmber of human
generation will be. In both cases, the final "nunlber" is depend-
ent for its significallce on the geometrical and chronological
relations 1vhich it summarizes; 42 the preliminary statement has
already indicated how tnllch less exactly these can be deter-
mined in humal1 affairs than in astronomy.
6. The rnode of production of the harmonies. These periods,
or "ratios," are the nllmerical representation of a three-climen-
sional diagram which is "produced from" the triangular image
discussed in (7) and illustrated in Figure 53, following. They
are produced from it in the sense that the same elements are
reconlbined in its constrllction, and the same basic relation-
ships illustrated by it. Hence Aristotle feels that he can ignore
this later elaboration completely in formulating a criticism
applying generally to the elements and principles of "Plato's
theory of revolutions." I-lis paraphrase, "meanillg by L(>l;
uV;113Et; \vhen the nllmber of the diagram is solid [(jl'EQE6~J ..."
simply indicates that the later development leaves unaltered 43
the principle being criticized.
The acttlal Inode of production, however, is by a combina-
torial technique elsewhere applied to elements by Plato (most
literally and spectaclllarly in tIle stereometrical element theory
of the Timaeus).44 Since the real elements in this problem are
the factors in hllman generation, these ffillSt be held constant,
bilt the same necessity need not apply to the mode of their
qllantitative schematization.
7. Altitude of the acrne and the 3-4-5 triangle. The ini-
tial part of his passage having indicated two aspects of human
gelleration-the process which constitlltes progress toward ma-
turity, and the insight which constitutes its natural acme-
 GEOMETRIC METAPHOR 119

Plato borrows an earlier ilnage by which the lleight of this

aCIIle was measured in Book vi, and joins it to the "three-
segment" representation of growth already illtroduced ill the
passage. The hypotenuse of 5, which connects the extremities
of these segments, illustrates a simultaneolls progress toward
lllaturity and intelligence (see Fig. 53) that anticipates the
(liscussion of the five modes in which actual maturity may be
related to man's "natural" perfection, reflected in individual
character and social organization (see Figs. 51 and 52). Both
capacity and maturity may fall ShOl~t of this natural human
perfection; the successive dominance of higher parts of the soul
rnay never be established, or the strellgth of the higher facul-
ties may not be great.
The triangular image represents "character" as a function
or, to follow Platonic metaphor, as "the offspring" of intellec-
tual ability and psychological normalcy. (An "abnormality" is
tIle continlled dornination of the psyche by a part which should
have been subordinated in natural or normal development.)
The 3-4-5 triangle was used as a "symbol" of marriage
by the Pythagoreans,45 and Plato's choice of diagram was no
doubt made with full awareness of the aptness of this tradi-
tion. 46 Not only is the suggested llletaphor of character as "off-
spring" reminiscent of the Pythagorean interpretation, but the
central aspect of the rulers' problem (hence of this passage) is
that the capacities of an individual child are inherited from his
parents. It is on the basis of his perception of the parents' ma-
ture natllres that the ruler lllllst calclllate their children's prob-
able capacities. The triangular diagram thus serves two func-
tions and has two simultaneous interpretations. The exhibition
of the natllral basis of character and maturity, which the rulers
Inust observe, is given by the immediate dialectical interpreta-
tion, while the relevance of tllese to reproduction (the mecha-
nism of heredity) is also consciollsly incorporated through the
Pythagorean tradition.
8. Dialectical interpretation. It should be noted that tIle
dialectical interpretation of the triangular image (Fig. 53, fol-
lowing) COllstitutes a perfect diagram of the dialectical transi-
 120 Plato's Mathematical Imagination

tion from Book vii to Book viii in the total argument of the
Republic. Books ii through iv have set up a definition of virtues
based on the proper, harmonious development of the three
"parts of the soul"; Books v through vii have culminated in
~.the discovery of an education which will lead to the maximum
development of the natural capacities of the rational "part."
he discussion of Books viii and ix is about to trace the degra-
ations of character and society from a model aristocracy
through the two deviations caused .by wrong dominance of the
parts of the soul, and through a subdivision of the second of
these deviations into three others, depending generically on the
intelligence shown in discriminating between more and less
real objects of the dominant appetitive part. The conclusion,
\ \in Book ix, that the tyrant is neither better, wiser, nor happier
II than the just man, returns to the beginning of Book ii. An
alternative way of conducting the same discussion, proceeding
from the tripartite soul in Book iv through kinds of character
to kinds of knowledge, is also indicated by the diagram in
Figure 53. This explains why Socrates, whose dominant interest
is in the definition of virtues and proving the identity of virtue
and knowledge, starts to follow this alternative plan of argu-
ment in Book iv, only to be recalled by the objections of "im_
practicability" raised by his audience-a recall which diverts
the Republic from a linear dialectical progression from becom-
ing to being into a cyclical progression, both beginning and
ending in becoming.
This reflection of the argument of the whole dialogue in
miniature is briefly indicated in Figure 53.
9. The shape of the figure. The following description is
designed primarily to help the reader who must draw or visual-
ize the diagrams as he goes along. It gives a purely geometrical
description of the shape of the image, the elements and dimen-
sions of which are about to be specified, and from which the
final schematic "number" will be derived as the product of its
dimensions. The description is dialectically functional, because
the emphasis on equal and unequal anticipates and suggests a
differentiation of the components of the final image into a
 GEOMETRIC METAPHOR 121

"being" component and a "becoming" component, which ac-

tually emerge as a "nature" component and a "career" com-
ponent when the interpretation has been supplied. But with
a published diagram this phrase would not be indispensable,
and its presence argues strongly against any notion that Plato
has here deliberately compressed his text with the intention of
being esoteric or enigmatic. On the contrary, he seems here to
be taking pains to clarify his procedure by instructing the
hypothetical reader with the wax tablet to set down, for his guid-
ance in visualizing the final diagram, the geometric figure shown
in Figure 54. This evidence against deliberate obscurantism
by compression substantiates the implicit thesis of the present
interpretation, that a detailed statement of dialectical referents
and antecedents of the geometric elements is omitted in this
passage simply because, in the immediate context of this same
dialogue, an extensive interpretation of those same elements
has been supplied.
10. An arithmetical ambiguity. The fact that the equality
referred to is arithmetic must be stressed here, since although
the significance of the' numbers is determined by the structure
of the diagram, the structure of the diagram need not be deter-
mined specifically by the arithmetical properties of its geo-
metrical elements. Thus, though it is arithmetically a matter of
indifference whether the two lines that combine to bound the
square represent the same or different elements in the cycle of
human life, it may be of great significance for the interpretation.
This suggests-for the first time-that the alternative construc-
tions of the square are functional in the passage, the alternatives
indicating different constituents which are, however, repre-
sented in this image by lines that are quantitatively identical.
Any other assumption leaves this alternative specification a
mystery, no less surely than the leftover gear is a mystery to the
amateur clock repairer (a mystery not really lessened by the
latter's recollection that he has already put in one gear just
like the part that is still left over).
11. The correct text of Republic 546A D. The reason why the
reading E~aLov is usually accepted at this point (and further on)
 122 Plato's Mathematical Imagination

instead of EXUOl'OV, and the probable origin of the reading E'X,(ll'OV,

t is the Neo-Platonic conviction that this whole passage must be
about astrology-a conviction that is lent some credibility if
the numbers literally specified are sufficiently frequent and
large (and the rest of the text unintelligible), but that is pat-
ently impossible on the basis of the text which has the best
manuscript authority and the least evidence of "improvement"
in the light of the Neo-Platonic tradition. 47 The Platonic
analogy of microcosm and macrocosm, which holds for Plato
only insofar as both are explicable by a principle of perfection,
is transmuted by later astrology into a simple and quantitative
one-to-one correlation (so that each planet must correspond to
some part of the body, each proportion of parts of the heavenly
cycle must be identical with a proportion of parts of the human
life-cycle, and every mathematical image illustrating a dialecti-
cal analogy must simultaneously have a valid and verifiable,
literally astronomical interpretation). That this assumption of
rigid quantitative correspondence is not always the happiest
basis for a sound textual interpretation is spectacularly illus-
trated by Proclus' "older and better" text of the astronomical
section of the Myth of Er-which is certainly not older, which is
not the text, and which is "better" only in that it avoids a sheer
impossibility for some pre-Proclean scholar who interpreted
Plato literally and astrologically.
This assumption-that microcosm and macrocosm are quanti..
tatively alike,. and that this passage is simple astrology-has
uniformly led interpreters willing to treat a mathematical image
seriously to expect a literal "cosmic number" representing the
time periods of astral influences on human embryology, or per-
haps human cultural history, or perhaps reincarnation. This ex-
pectation, in turn, has led to the assumptions that the first part
of the passage specifies the same number as the second and that
the contrast of cosmic and human periods is intended only as
one of duration, not one of kind. Neither assumption, how-
ever, is definitively established by this passage in its context or
by Plato's treatment of the same analogy elsewhere.
The acceptance of Exa'tov in its first two OCC11rrences has
 GEOMETRIC METAPHOR. 123

hinged on the assertion of higher criticism that the alternative

would make intelligible interpretation impossible. Once the
bare possibility of any alternative has been demonstrated, the
textual question must be re-examined in the light of proper
philological criteria, with less emphasis on what the passage
must have said, and more on what it did say. Examined in this
way, the evidence is very strongly in favor of €xucrtov.
12. The identity of two diagonals reconsidered. The geo-
metrical identity of the two "diagonals," one of the two or three
phrases which most interpreters have been sure they under-
stood, and with which they have begun "decipherments" of the
rest of the passage, is not clear and cannot be the one tradition-
ally assigned.
Since it is the two harmonies, arithmetically represented, that
the final "geometrical number" comprehends and the construe..
tion specifies, the final number mentioned (27 x 100) is that of
the second harmony, and the ~E refers back to ~~v as, not to
j(QOl!~Xl1 ~E. A "harmony" is arithmetically represented as the
product of its components unless otherwise specified. Plato's
EXUO'tOV, which seems unnecessary, is actually directing attention
to the fact that he is not specifying the first harmony in the
normal arithmetical way. This qualification does not apply at
all to the second harmony, which is represented by a simple
arithmetical statement. Consequently, the number 2,700 repre-
sents the product of the two dimensions of the rectangle, not
its longer side.
It will be argued in the discussion following that multiplica-
tion by 100 represents a final part of the projection of the square
figure into the third dimension, and therefore is a factor only
in the second harmony.
These two observations clear up the arithmetical dimensions
of the entire figure, and of the "diameters" that generate it. The
ratios are integral, and the relation of the equal side, X J to
*00 the unequal side, YJ is given as xy == 27. Hence one of these
factors is 3, the other 9. If x were 9, then the two "diameters"
would equal 9 plus 1 and 9 plus 2, or 10 and 11, respectively;
whereas if x is 3, they are y3 plus 1 and -J3 plus 2, or 2 and
124 Plato's Mathematical Imagination

y5, respectively. Of these two exclusive arithmetical possibili-

ties, the metaphor of "rational" and "irrationaf' clearly points
to the latter. (From an argument given in subsequent discussion,
it is possible to show that these must have been the intended
values of the two diameters, without taking account of the fact
that these values are actually established by the arithmetic of
the passage.)
The failure of previous interpretations to take account of the
relevance of 2,700 as the number of the second harmony is ex-
plained by two factors. First, the text had been altered to EXatOV
in the belief that the numbers involved were all astronomical,
and in the corollary belief that the method of statement of the
two harmonies was the same; second, by Theon's time the vo-
cabulary of geometry had settled into a literal, technical lan-
guage, with loss of awareness of the metaphors in which its
terms originated, and in this later language Plato's phrase had
come to have a technical interpretation which yields a different
arithmetical value.
13. Parents and children: Pythagorean genetics. The linear
side of the "square harmony" is really itself the product of two
factors. The analogy of squaring to human procreation has al-
ready appeared twice in the passage. The generic metaphor of
the process of human life is a growth of roots and their squares;
the first specific application is involved in the traditional Pythag-
orean interpretation of the 3-4-5 triangle (in which, presum-
ably, the invariance of total human endowment, and the joint
parental contributions to heredity, were illustrated by the equal-
ity of the squares of the sides-the "actualizations" of the auva~,u~
contributed by each parent-to the square of the hypotenuse-
the actualization of the ~omposite a{,valll~ inherited by the
child.) 48
This introduction of squaring gives a separate diagram rep-
resenting the parent-offspring relationship, which leaves the
schematism of the natures, characters, and fortunes of individual
parents and children-the schematism by which the rulers select
couples for marriage-for separate treatment in the final paral-
lelepipedon diagram (which, with only three dimensions, is
 GEOMETRIC METAPHOR 125
sufficiently complex without attempting to represent parentage
as well as nature, character, and fortune). The factor of parent-
age as determining a human life is indicated by this separate
diagram, which is made to show the transition from the sche-
matic triangle to the more accurate and specific solid image.
To some readers, Plato's assumption that children inherit the
natures of their parents seems to mean, in practice, that the
apparatus of competitive tests in the educational system is only
a fraud, designed to give the artisan class an illusion of equal
opportunity.
In fact, however, the concept of inherited aptitude as an
average, which Plato's symbolism in this passage takes over from
Pythagorean genetic theory, carries with it the necessary conse-
quence that some provision for shifting hereditary class status
be made.
A man (or woman) may fail to qualify as a ruler through
several sorts of imbalance or deficiency. Weakness, lack of
courage, stupidity, or constitutional intemperance may lead to
disqualification. There is always, therefore, the possibility that
a marriage such as that of a brilliant but irresponsible artisan
with a steady, moderately intelligent woman of his class will
produce children who have just that balance of qualities which
the state seeks in its rulers.
Again, it would follow (even if the cycle of human fertility
did not complicate the problem) that children of the ruling
class might inherit an imbalance which would exclude them
from holding rule.
The apparatus of examinations and equal opportunity is not
contradicted by the genetic theory, though it appears to be a
consequence of that theory that favorable matches between
parents of the selected upper classes are most likely to result
in an improvement of the breed. 49
14. Rational diagonals." The generic metaphor of "rational"
U

and "irrational" has already appeared twice in this passage. In its

first occurrence, "rationality" was the attribute of the mature
man who had developed his capacity for reason, and was opposed
to the fluctuation and imperfect integration of the stages of
126 Plato's Mathematical Imagination

growth through which this maturity is attained. "Rationality" in

that context was identified with the "altitude of the acme" of a
period of human life (as opposed to the chord subtending the
period, representing the successive dominance of faculties emerg-
ing with the passage of time). In the triangular diagram, this
metaphor is used more specifically in the representation of the
harmony of the "moral" and "intellectual" excellences by the
two sides of a right triangle, in concordant ratio, which jointly
determine the types of character represented by the hypotenuse
of 5, which is the "offspring" of their "marriage." The line of
4 in this triangle represents rational or intellectual attainment,
the other two lines representing the submissiveness to reason of
the other two, "irrational," parts of the soul.
As the metaphor of squaring indicates, the present diagram
exhibits generation as the inheritance of capacities, and the con-
trasting characterizations of the two diagonals parallel the
contrasting character of the capacities which each represents.
The inheritance of intelligence is illustrated separately from the
inheritance of predispositions toward a given moral character-
a separation maintained throughout the preceding discussion in
the Republic, and emphasized again in Books ix and x. In each
case, the image is that of the joining of the contributions of the
two parents, represented linearly, to produce a square. The
squares thus produced can be identified with the lines of the
triangular diagram, which are a simplified representation of the
squares, because the lines have the same number of units of
length as the squares have units of area. This is a standard
Platonic way of changing from plane to linear representation.
15. The diameter of five. The intended significance of this
designation apparently was dependent on some further, com-
monly known Pythagorean or technologically inspired diagram,
which can now only be conjectured. The "pempad" of which
these lines are diagonals must be the number 5, the arithmetical
representation of the pempad of the triangular image. As the
later tradition still clearly indicated, the problem inspiring
the differentiation of "rational" and "irrational" diagonals is
the problem of representing irrational magnitudes as num-
 GEOMETRIC METAPHOR 127

bers. The "rational" diagonal is the closest integral approxima-

tion. The "diagonals of 5" must have represented this number
in some familiar geometric presentation, analogous to Theaete-
tus' geometrical classification of numbers. In this connection, it
is worth noting that the simplest representation of marriage as a
product of male and female is a 1 x 2 rectangle, the diagonal of
which is, of course, V5.
Although this "diameter" is the tt~uvat-u~" of five, it is not
possible for Plato simply to say so. For to have said it in that
way would have made the use of the metaphor of rationality
(which specifies his intended interpretation of the final figure)
impossible. It is clear that a ~{,vaJlL~ which does not produce a
given number, but some other, is really that other number's
()uvaf.lt~. Hence the "rational" approximation to the diagonal
of a 1 x 2 rectangle is not "the rational ~vvallL~ of five" at all,
but simply "the ~vvallL; of four."
The "marriages" represented by squaring (hence in an arith-
metical statement by the numbers representing squares) are
used to show the generation of the lines representing character
and intelligence in the triangular diagram. The differentiation
of the two components of heredity will be discussed in (17).
It is possible that this phrase is an architect's familiar ab-
breviation of a designation, originally more complete, of one of
the lines which appear frequently in temple construction. The
complete description of such a line would be "the diagonal
which, squared, equals an area of five." In contemporary writ-
ing on "dynamic symmetry," we can observe a similar emergence
of an architect's geometrical vocabulary which has undergone
abbreviation. Thus, "whirling squares," "root five rectangles,"
etc., are substituted for full geometrical designations. The studies
of the dynamic symmetry of the Greek temple, paralleled by
references to Hindu temple construction, indicate that, in any
geometrical construction of temple plans, the I x y5 rectangle
would appear often enough to make the y5 a frequent line,
for which such a shortened designation might be employed.
16. Subtracting one. In other words, only three, not four,
128 Plato's Mathematical Imagination

hereditary levels of intellectual capacity are to be used in the

final diagram. Actually, only three such levels are functional in
determining classes in the ideal state, stages in education, and
types of human character, elsewhere in the Republic; appar-
ently, in Plato's view, any man in a favorable environment has
the capacity to go beyond eixaota to some form of a6;a, as he
matures and acquires experience. 5o
17. Function of the alternative statements. To keep his in..
terpretation clear, Plato cannot simply present the first harmony
in a numerical statement. It is important in understanding that
interpretation to notice that the two lines joining to determine
the "equal harmony" represent different elements in the con-
text; but this difference is arithmetically irrelevant, and would
not be made clear by the statement that "nine is the number
comprehending the first harmony." On the other hand, the
difference between integral approximations to roots (which par-
take of conversability and commensurability but do not com-
pletely exhaust the extension of the approximated lines them-
selves) and roots (which cannot be treated arithmetically until
they have been developed into their squares) provides a geo-
metrical analogy very appropriate to the distinction at hand.
18. Subtraction of two. The number of hereditary predis-
positions in respect to moral character is also represented, in
the final diagram, as less than the enumerations elsewhere
would at first lead us to expect. In fact, however, the divergences
of character among oligarch, democrat, and tyrant have a com-
mon appetitive predisposition as their native basis; insight and
environment are responsible for their differentiation.
19. The final square harmony. Figure 59, following, is the
square matrix which results from combining the linear repre-
sentations of basic types of moral predisposition and degrees of
intellectual ability.csl The same matrix appears in the tyrant's
number in Book ix, and the list of lovers in the Phaedrus is also
the same, with the rows reading from top to bottom, then from
left to right, successively. The identifications of character from
the Phaedrus list are indicated in figures 59 and 60. This dia-
gram is so lucid and so relevant to the architectonic of the
 GEOl\IETRIC METAPHOR 129

Republic that, in the last analysis, it is the best evidence that

can be offered that the interpretation here proposed is actually
a rediscovery of the intention of Plato.
20. The other dimension. The function of the third di-
mension in other Platonic images and analogies is often that of
representing the actualized spatial and temporal characteristics
of the embodiment of a structure the nature of which has been
indicated by the two-dimensional diagram.
The development of the present passage in Republic viii has
been, to this point, one of making explicit the abstract constants
which are imitated and developed in actual "periods of human
becoming." The number as a whole is to be determinative of
better and worse births, and the rulers' errors in calculation
lead to children who are "neither well endowed nor fortunate."
The projection of the square matrix into the third dimension,
represented by the unequal element in the j(Qo~~x'r) harmony,
is given as the product of two factors, the dialectical counter-
parts of which represent those additional factors which interact
with an inherited "natural disposition" to produce a particular
individual career.
In the schematization of flao~ as a function of a rational and
an irrational cp{,at~, abstraction has been made both from the
effect of 1:QoCP~, and that of TUX". But in fact the nurture of an
individual is as instrumental as his heredity in determining the
kind of life he will lead, and education has been treated as co-
ordinate with genetics in importance in all other parts of the
Republic. Fortune, on the other hand, has not played a major
role in the antecedent discussion. The sequence of physical
coincidences which presents opportunity to one man and denies
it to another has been treated principally as a variable to be
provided against in any plan-whether the plan is that of a
young man bent on becoming a tyrant, or that of a legislator
bent on seeing to it that the guardians function as disinterested
administrators. Only in the last two books of the Republic does
it begin to appear that the man who deserves opportunities is
the one who finds them; and even this is a general rule which
does not hold of every given case in any single incarnation.
130 Plato's Mathematical Imagination

The variations in nurture which lead to different develop-

ments of native capacities have in fact already been indicated
in the schematism of types of character shown in figures 59 and
60. For since the state is the individual written large, the kinds
of individuals who dominate determine also the possible num-
ber and kinds of environments. The effect of training can thus
be schematized simply by applying the 9-celled matrix of char-
acter to itself in the third dimension.
The resulting image, without the added factor of 100, is a
simple progression through powers of 3, from the 3 parts of
the soul to 9 natural endowments to the 81 types of life result-
ing from possible interactions of endowment and environment.
This final representation is a parallelepipedon" with one plane
surface a square 3 X 3, the other a rectangle with an area of 38 ,
and sides 3 X 9. (It will be shown ill subsequent discussion that
the progression 9, 92, 98 in the tyrant's number in Book ix uses
an interpretation analogous to that of the present passage, and
that the compression of the tyrant's nunlber may be explained
by Plato's assumption that the images of Book viii would al-
ready be familiar to his readers.) The elements composing this
figure are all identifiable and central in the context of the
passage, and the combination of variables represented is exactly
what a ruler's handbook of abstract genetics would discuss-
leaving to the calculation and perception of each ruler the
determination of the application of these general rules to spe-
cific cases, the point at which time and chance may play havoc
with even the best reasoned social plan.
21. The human cycle. The diagram, like the Republic as a
whole, reverts in its final step to the images of cycle with which
it began, and in which must appear the spatio-temporal particu-
larizations of the pure, atemporal dialectical proportions, the
construction and systematization of which intervene.
The periods in Plato's myths are more often than not repre-
sented as powers of ten. The largest period, determinate but
indefinitely great, appears as a "myriad," and its component
parts are hundreds, thousands, or ten. Since the function of a
myth is the illustration of arguments by fictions embodying true
 GEOMETRIC METAPHOR 131

principles but exhibiting them with the aesthetic concreteness

and individuality which argument and equation can never
reach, the astronomy and mathematics of Plato's mythology are
definite, but neither scientifically nor empirically accurate. In
the description of a divine cosmic justice built into the mecha-
nism of the universe, in the great myth with which the Republic
closes, a period of one hundred years is given as the temporal
span of human life. The sections of the Republic which inter-
vene between this final myth and the present passage fill in
testimony from experience regarding the usual correlation of
good luck and good lives; but in the dialogue as a whole, as
well as in the present passage, the display of general dialectical
relations at work in individual sensible careers requires a myth-
ical projection-whether as a separately stated "chronological"
factor in a multiplication, or as a vision of The Fates spinning
human destiny. In this passage and the whole dialogue alike,
cycles are recognized as aiming at structures which they imper-
fectly realize; and the elaborations of the relations of these
structures constitute the best means of rational description and
approximation to the unique individual trajectory marked out
by a career in space and time.
g. Interpretation Established from Text of the Passage Alone
It has been traditionally assumed that the crucial requirements
for any interpretation of this passage (Republic 546A-B) are
that it make sense philosophically and that it be faithful to the
precise terms and syntax of the text, apart from intruded as-
sumptions of higher criticism. But previous interpretations have
usually not been rigorous in applying this second test: philol-
ogists commenting on the passage have intruded assumptions
derived from sources as remote as Chaldean astrology and Hel-
lenistic embryology, while claiming to deal only with the text.
If a stricter interpretation is attempted, two basic points
must be emphasized. First, except for the "harmonies," which
usage quite clearly associates with multiplication, there is no
reason to think that all the elements of the nuptial number
must be related by addition or multiplication. On the contrary,
132 Plato's Mathematical Imagination

in a "geometrical number," we should expect the relation to

be given by geometrical construction, which would not neces-
sarily be represented arithmetically as a product or sum. When
scholars have asked, therefore, "whether the numbers !J, 4, and
5 are to be added and cubed, or multiplied and cubed, or cubed
and added," they have overlooked the possibility that these
numbers refer to geometric elements, and that the construction
need not limit itself to simple addition or multiplication. Sec-
ond, in a passage as elliptical as this one, a careful author will
use co-ordinating particles very carefully to indicate elision and
organization, so that a careful interpreter should start with
these, rather than beginning, as some have, by "deciphering"
10,000 or 4,800 or 7,500 as a number referred to in one part of
the passage. It is interesting to note in this connection that some
ancient commentators, e.g. Nestorius (cited by Proclus, In Rem
Publicam J ed. Kroll, II, 49, 409-11), recognized that "the num-
ber" might be a geometrical rather than arithmetical function
of the elements given.
The co-ordinating particles set up a carefully balanced pat-
tern, in which are contrasted with one another (1) the periods
of divine and human generation, (2) the four termini and three
segments which the latter attains, (!J) the shape of the first
harmony as opposed to the shape of the second, (4) the construc-
tion of the first harmony with rational or with irrational
"diagonals of the pempadu ; (5) the final phrase, 100·27. This
construction seems to require that (5) in the foregoing balance
(4) by giving the dimensions of the second harmony. This will
not give as large a number as some commentators, who assume
a priori that the numb~r here is astronomical, are lead to expect,
but the construction will fit better the actual use of connectives.
Also, the division of (5) into two separate factors, 100 and 27,
may well balance the subdivision of (4) Into its two alternative
or component constructions.
We can agree to assume that a "number" here means an
integer, that the passage describes a three-dimensional figure
x.x.y, and that x and yare unequal and are both derived, by
some geometrical or arithmetical elaboration, from the eIe-
 GEOMETRIC METAPHOR 138

ments that are represented by 4, 3, and 5 early in the passage.

For my own interpretation, I should also want to add the
assumption that the 100 in the final phrase is a factor incor-
porated in the third dimension of the figure as a semimythical
"temporal periodicity" factor, not derived from the antecedent
construction. But it will be seen that this assumption is needed
only to distinguish the present interpretation from a class of
other mathematically possible "solutions," none of which has
ever been seriously proposed.
Working rather informally, and using the notion "function
of" as meaning an integer resulting from arithmetical opera-
tions or summarizing a geometrical construction with given
elements, we shall see that it is possible to set up the passage as
a set of equations, and to examine the interpretations which
represent solutions of them.
We are told first that there is a number which equals the
period of human generation. We can write this
(1) n =P
At some limit, this period attains four termini bounding three
intervals.
=
(2) P 4t,3i (at some relevant limit)
Two "harmonies" (which in standard Platonic use are products
of at least two factors) are derived from (2), and another ele..
ment:
(3) hl'h 2 = f(4t 3i)
J

In (3), "fu represents "function of" as defined above. The deri-

vation of the harmonies includes two other items, but first, the
use of pythmen has the effect of making t and i in (3) arith-
metically irrelevant, by taking their 4:3 ratio in its lowest terms.
Arithmetically we can state this as
(4) t = i = 1
This ratio is then "joined to" a line or set or integer of five to
help form the harmonies. We can represent this "joining" as F,
another function; this is cautious, and avoids assuming that an
]34 Plalo's Mathematical Imagination

arithmetical sum or product must be intended by Plato's non-

technical language here.
(5) h1 ,h 2 = F(f[4,3], 5)
Yet another construction step, a "triple augmentation," is
involved in deriving the harmonies from their elements. The
exact meaning of this is also ambiguous (as we can see by look-
ing at past explications), and again it is safest to say only that
it represents a functional relation, indicated by g.
(6) hhh 2 == g(F[f[4,3], 5)
Now we are given a preliminary picture of the shapes of the
geometrical representations of the "harmonies"; the first is a
square figure, hence arithmetically a square number. Using x
as a variable to represent any "number" (in Plato's sense, i.e.,
any integer), we can write

The second harmony has x as one dimension, but is unequal in

the other.

Probably we could assert also that y is greater than x, but we

will not risk doing so just yet.
We are now told the exact values, or the exact constructions
giving the values, of the first harmony.
(9) x = (d5)2- 2
(10) x = (dr5)2 - I
From which,
(11) (d5)2 - 2 = (dr5)2 - 1

In the preceding equations x is defined in terms of the

rational and irrational diagonals of "the pempad," d5 is irra-
tional, and dr5 is the rational approximation. (It requires very
conlplex formulation to get the exact mathematical statement
that drS is the nearest rational approximation to d5, unless we
use the form in (11 ).)
 GEOMETRIC METAPHOR 135

Finally, the second harmony is identified as "one hundred

cubes of three." The identification of this with factors which
give h 2 , not y alone, is based on the use of connectives already
discussed in the text:
(12) h2 == 100 • 3 • 3 • 3
If (9) and (10) are the two dimensions of the first harmony
(arithmetically equal, but resulting from different construc-
tions), the separation of 100, 3, and "cubed" in (12) may be
exactly parallel. Let us assume, at any rate, that the factor 100
is not a function of X J so that the second harmony == 100 · x
multiplied by some third integer, z:
(13) h 2 == IOOxz
Since h2 == 2,700, we are given the value of xz:
(14) xz == 27
Since x and z are integers, x must be either 3 or 9. Working
backward and substituting, we see that the distinction of "ra_
tional" and "irrational" diagonals in (9) and (10) makes no
sense for x == 9. For x == 3, however, we get
(15) 3 == (d5)2 - 2 == (dr5)2 - 1
This checks with (11) in the following:
(16) 2 X 2 - 1 == y5 X y5- 2
By further substitutions, we arrive at a complete solution:
(17) hI == 9
(18) % == 9
==
(19) Y 9 X 100
(20) h2 = 2,700 3y=== xy
(21) 3 X 3, 3 X 9 X 100 == g (F[f[4,3], 5)
Of various functions which give this result, the choice is
limited by the two derivations of x in (9) and (10). Here,
among these possibilities, are solutions which fit the interpre-
tation given in preceding discussion:
136 Plato's Mathematical Imagination

(22) f(u)v) = (u - 1) • V
(23) =
f( 4,3) 9
(24) F(uJv)= 1 • (v - 2)2
(25) F(/[4,3], 5) == 9
(26) g(u) == 300u
(27) g(F[f[4,3], 5) == g(9) = 2,700

These are not uniquely determined by the text of the passage

alone, but from equations 12-21 it does follow necessarily that
(28) P == 9 X 2,700 == 243 X 100 = 24,300
This is "the number" summarizing dimensions of the final
geometric figure, and this solution seems to follow quite strictly
from the passage out of context, with the possible exception of
the isolation of 100 as a factor in (13). But that step does not
prevent us from showing that previous interpretations will not
fit the present analysis of the text; for, if xy =
2,700, " must
be less than 2,700 or equal to it, and standard prior interpreta-
tions uniformly make it larger.

Proclus approaches the passage with three assumptions: (I) that

the macrocosm-microcosm analogy is literal and the two careers
completely isomorphic (in effect, that astrology is a science): (2) that
Platonic mathematical images are constructed on the single prin-
ciple of (a) giving a schematization of method, (b) exhibiting that
method in pure mathematics, and (c) showing the empirical appli-
cation; he is convinced that there must be a significant geometric,
arithmetic, astronomical, and cosmological (dialectical) sense to be
had from the text; unfortunately, he assumes also (d) that a single
type of dialectic, that used by Plato in the empirical science section
of the Timaeus, is always the method operative in any context;
(3) that the "triple augmentation" producing the harmonies de-
scribes a reconstitution of the initial triangle into one with the
same ratio of sides, but bounded by "solid" numbers. Figure 50,
following, represents Procins' version of the "thrice augmented"
phrase of the text; reading the first section as a general statement
of a method of ratio construction, the solid numbers of the third
triangle in this figure are shown to yield ratios by that method. The
 GEOMETRIC METAPHOR IS7

Figure 50

3~
4

12

27

36
SCHOLION FROM PROCLUS

(Greene, Scholia Platonica, p. 257)

reading Exat'ov naturally appealed to Proclus as more amenable to

astrological and astronomical interpretation than any other. See
A. G. Laird, Plato's Geometrical Number and the Comment of
Proclus, Madison, Wis., 1918; F. Hultsch, "Excursus," in Proclus
In Platonis rem publicam commentarii, ed. G. Kroll, 2 vols.
(Leipzig, 1889-1901)~ end of Vol. II.
138 Plato's Mathematical Imagination

FIGURES 51-61: THE NUPTIAL NUMBER

The interpretation suggested is, in outline, the following:

(1) There is a period of human generation, divided into three
stages of maturity, which, in contrast to the simple periods of
astronomy, is highly complex because it involves growth as well as
simple locomotion.

Figure 51

E F

A G c'----.. . ."s H o
LIFE CYCLE

In Figure 51, arc AB = the period of a human life; arc EBeF =

the period of a human generation (from th.e maturity of a parent
to the maturity of his child); segments CB.1 BH, and HD represent
the stages of maturity (marked by the successive dominance of the
appetitive, spirited, and rational parts of the soul); and points
C, B, H, D mark the critical periods of a human life. The influence
of heredity is shown by the similarity of arcs AEB and CFD.
(2) These periods differ in height of acme as the individual
approaches more closely the human norm of intelligence.

Figure 52

A 2 o 3 B
THE ACME OF TIlE HUMAN LIFE CYCLE
 GEOMETRIC METAPHOR 189

Figure 53 shows these lines joined, with their interpretation.

Just as in the Pythagorean 3-4-5 symbol the child (square of the
hypotenuse) is the sum of the contributions of the parents (squares
of odd and even sides), so here the character of an individual is rep-
resented as the "offspring" of intelligence and maturity. Read from
the lower right to the left, the terms that appear are the key con-
cepts, in order, of Books i-iv, v-vii, and viii-ix of the Republic. Thus
the diagram shows the dialectical structure of the Republic as well
as the factors determining the characters of prospective parents.

Figure 53

Aristocrat
Knowledge
Timocrat
Hypothesis
....
::s Oligarch
~
Opinion ~.
::s Democrat
B
Conjecture Tyrant
Maturity

THE 8-4-5 TRIANGLE

The altitude of the acme CD is indicated by the divided line-

a line of four segments representing possible levels of intellectual
ability.
(3) The planning of marriages is determined by taking the "line
in 4:3 ratio" (line CD in Figure 52) and joining it, in some way
not here specified, to a line of five segments. The simplest junction
gives Figure 53, but does not produce the final diagram intended.
(4) For the benefit of the reader, since there were no figures in
the manuscript, Plato next "sketches" in the general structure of
his final diagram.
140 Plato's Mathematical Imagination
Figure 54
E F

G
A~----f

c o
SHAPE OF THE FINAL FIGURE

In Figure 54 ABDe is a square, but ABFE is a rectangle (AC =

=
AB). The "two harmonies" will be AB·AC AB2, and AB·AE =
2,700.
(5) Plato now modifies the original lines of 4 and 5 and combines
them into the square BFDE to form Figure 55.

Figure 55

c
2 3 4 5
A'"-"'--......--......--.a------oIB
E I
I
I
2 I
I
I
I
3 I
I
I
_
4 _ _ _ _ _ _ ...J1
o F
LINES OF GENETIC TRIANGLE RECOMBINED INTO SQUARES
 GEOMETRIC METAPHOR. 141

Lines AB and CD in Figure 55 equal lines CD and AB, respec-

tively, in Figure 52. EDFB is a matrix showing the different com-
binations of motivation and intelligence. AE indicates two types
of character (democrat and tyrant) which do not represent distinct
types of motivation, since the appetitive part of the soul is domi-
nant in the democrat and tyrant just as in the oligarch. CE repre-
sents a type of intelligence, ebtQaLa, so low that it is transformed
into ~6;a in any normal human experience. EtXQGLQ is like the
infant's blooming, buzzing confusion or the madman's confusion
of fact and hallucination.
(6) Figure 56 is derived from Figure 53 by the metaphor of
"squaring," which shows the hereditary nature of both motivation
and intelligence. Here one must posit a hypothetical Pythagorean
construction.

Figure 56

A 8

c o
THE IDENTITY OF THE "DIAGONALS"

In Figure 56, AC = 1, AB = the rational approximation to AD =

=
2, AD the "irrational diameter of five" = ./5. The derivation of
these values from the algebra of the passage is given in "Detailed
Interpretation" (Sec. I), preceding. Line CD in Figure 55 is gener-
ated by squaring and subtracting appropriate segments from either
AD or CD in Figure 56. (See 17-19 of Sec. f.)
Line CD of Figure 55 is equal to the 'area of Figure 58, the same
factor being represented by the line and the figure. In Figures 57-58
=
(ab)2 - 1 CD - CE (in Fig. 55)
(bd)2 - 2 = AB - AE (in Fig. 55)
The equation holds between the number of units of area of each
square and the number of units of length of each line.
142 Plaio·s iWathematical Imagination
Figures 57-58
(57) (58)

ab bd

ab bd
SQUARES ON DIAGONALS

Figure 59
PHAEDRUS ~IA'rRIX OF LIVES
NOY~ L\IANOIA ~O:IA
RATIONAL Philosopher Auxiliary Artisan
SPIRITED General [Prophet] Sophist
APPETITIVE l\tIerchant Democrat Tyrant
Soul Body Artifacts
Here we can confirm this diagranl from two other passages in
Plato. ~-"he list of lovers in Phaedrus 248 who vary in their Hvice
and forgetfulness" is shown in Figure 60.

Figure 60
LIST FORM OF MATRIX USED IN THE PHAEDRUS
Part of Soul Dominant St'rength of Recollection of Forms
R Philosopher Lovers of things of the nlind
S General (least forgetful)
A ~erchant
R Gymnast I..overs of the body
S Prophet (more forgetful)
A Poet
R Artisan Lovers of artifacts
S Sophist (most forgetful)
A 'Tyrant

rrhe "difficult COluputation" in Rep1ublic ix, where the distance

between aristocrat and tyrant is given as 3 + 2 == 9, is explained
by Figure 61.
 GEOMETRIC ~IETAPHOR 143

Figure 61

I 2 3
Aristocrat

4 5 6
Timocrat

7 8 9
Oligarch

l\IATRIX lTNDERLYING "REPUBLIC" IX

Here the proportion to be derived is Aristocrat : ~'yrant :: 1 : 9.

h. Historical Comments: Other Interpretations

J. Dllpuis, in llis "Definitive Memoir" on this passage (Republic
546A) lIas inclllded a resume of sixteenth-century through nine-
teentlI-centllry interpretations ,vhich lIas considerable interest
as a history of discussions of the passage. A translation follows.

Dtlpuis' History of Interpretations 52

I. (BAROCIUS 53) The title indicated what Barozzi thinks of
the difficulty of this passage. i\mong the authors who have at-
tempted to explain it, he cites lamblichus, Thermacides the
Pythagorean, Sebastien Fox, Raphael Volterranus (Maffei), St.
'fharnas, Jacques l... efebvre d'Etaples, and I)onatus .l'\ccacio-
laus....
For him, the nUlllher of the period is 1728, the cube of 12:
"Geometricu itaque numerum vocat Plato ipsum cubfl mille
septingenta vigenti octo."
A large number of commentators have thought to see, with
Barocius, in the base of the epitrite tIle number seven, 3 +
4.
rrhis, added to five, would be 3 + + 4 5, or 12, which raised to
its cube (LQl~ a'u;lp<)E(~) gives 1728.
"'rhe dissertation of Barocius is one of the most distinguished.
rrhe literal Latin version of the passage ... is one of the best.
II. (BODIN 54) Rene Herpin "vas a nati"\ie of the town of
i\ngers. Bodin used the name to give himself lllore freedom in
,vriting his Apologie llimself. . . . Bodin does not propose any
number.
144 Plato's Mathematical Imagination

III. (PEUCER fit) [Does not propose any interpretation.]

IV. (MERSENNE 56) Mersenne believed that "the Platonic
number is 729," which he obtained by an error in calculation.
"The hundred numbers of commensurable diagonals mean 3
and 4, which, being multiplied by 100 give 700, to which the
cube of 3, that is to say, 29 [sic] being added, makes 729, which is
the number meant by the enigma of Plato."
V. (T. TAYLOR 57) The author devotes a chapter of his study
to the geometric number. He believes the two harmonies are
10,000 and 1,000,000 and that "the whole geometric number is
a million."
VI. (LE CLERC 58) M. Le Clerc does not explain his transla-
tion nor propose any number.
VII. (SCHNEIDER 59) These two dissertations by Schneider
are very distinguished. The first is a thesis, a disputation. The
author believes, with reason, that there are two numbers, and
that the description of the second commences with rov, the actual
numerus fatalisj and he is convinced that O"Ut'Uy€(~ indicated
addition. Like Barozzi, he recognizes that by ExaTov &Qt{}~wv a1to
~LallEtQrov it is necessary to understand one hundred squares of
diagonals, and not one hundred diagonals. He thinks that the
number of destiny contains the factors 8 and 27, the final terms
of the series 1, 2, 4, 8, and 1, 3, 9, 27, but that Plato has deliber-
ately left the data needed to find the number incomplete. The
second dissertation contains the opinions of preceding commenta-
tors: Barozzi, Boulliau, Cardan, Peucer, Melanchthon, Matthias
Lauterwald, Bartholomee Bredell, Bodin, Kleuker, Lefebvre
d'Etaples" etc.
VIII. (VINCENT 60) Vincent tries next to justify his transla-
tion, and concludes thus: "In summary, the answer to this enigma
proposed by Plato is the number 216, the cube of 6, and fourth
term of the proportion 1: 8 :: 27 : 216." His first takes 216, or
3 X 72, as the small side of a right triangle, of which the two
other sides are 4 times 72 and 5 times 72, the perimeter 12 times
72, or 864, which seems to him to satisfy the conditions given,
provided one adopts the corrections that he has proposed. [In
his earlier discussion of the text and translation given by
Vincent, Dupuis says: "Vincent replaces ExaTov ~e xv~rov 't'Q(a~o~
by EXTOU l)e xu~ou l:Q(al)o~.]
IX. (MARTIN 61) Martin has adopted the solution of Vincent
with certain modifications.
 GEOMETRIC METAPHOR 145
x. (MYNAS 62) The second book being undiscoverable in the
public libraries, we believe that it ,vas never published. Minoide
Mynas, a Greek philologist, died in 1860. In his preface, he says
that the solution of the theorem of Plato was more precious to
him than the discovery of Babrias (which he had made in 1840
in a monastery on Mount Athas). Here is this solution [page 119
of the memoir]: "The creation of the world, divine primogeni-
ture, is comprised in a perfect number; for that of man, it is
otherwise: in the beginning of its development, it passes, under
the influence of dominating and dominated stars, by the three
dimensions which, combined with the four elements in affinity
and in opposition more or. less great, set in proportion and har-
mony all the parts of the nascent being. In effect, the first quater-
nary epitrite joined to the pempad anq tripled, yields two har-
monies; the one, in double ratio, perfectly equal, goes just up to
a hundred and something; the other, in triple ratio, is combined
proportionately with the first. Each term (100) of this harmony
has for diameters (i.e., factors) round digits of the pempad, the
ones less great than the others by a unit. Among these terms,
which give one hundred ternary cubes [sic] there are two incom..
mensurables. The entire number existing in geometrical propor..
tion indicates the relation of better and worse generation."
The explanation of the language of the passage, which Plato
had intended "to obscure," [see p. 131] stops at the words dom-
inating and dominated, [p. 159] which, according to the commen-
tator, relate to the planets.
The solution of Minoide Mynas, who was notwithstanding very
erudite, and who knew Greek even better than French, is are..
markable example of the strange divagations to which, even in
our day, the interpretation of the number of Plato has given rise.
XI. (ZELLER 68) In this very remarkable work, the celebrated
historian of Greek philosophy says that the cosmic period is
10,000 years, and that this is % of the political period, so that the
latter would be % of the former, and takes 7,500 years.
){II. (WEBER 64) In discussing the problem, he criticizes with
some animus the interpretation of Vincent and Martin. The
latter, in comparing to a right triangle in which the sides are 8,
4, 5, and area 6, the triangle of which the sides are 72 times
g-reater, gives to this latter triangle an area of 6 X 72, although it
is 6 X 72 2 • Weber notes this lapsus by saying: "Aream trianguli
rectanguli, cujus latera sunt 216, 280, 360, non 432, sed 72 X 482
146 Plato's Mathematical Imagination

vel 72 2 X 6 velere, nemo nisi mathematicae elementorum im-

peritus nescit. . . !"
He believes, with Hermann and Rettig, that &.QQ~l'rov ~E ~uoi:v
stands for &QQ~l'(J)v ~E aEo~Evrov ~uoi:v EXUOl'roV. Vincent and Mar-
tin had made LOllv louxL; relate to the first harmony, and
E'Xal'OV l'OOaUl'd.'XL~ to the second; he finds this interpretation
unsatisfactory: "pessimam." He thinks, with reason, that
E'Xa-rov 't'o(Jau't'ci'XL~ is short for E'Xal'OV Eita't'oVl'ci'XL~. He takes as his
harmonies the two numbers 10,000 and 7,500, with the reading
j(Qo~~'X'tl af, but he presents no other arithmetical conclusion.

From Dupuis' summary it will be seen that from 1600 to 1850,

interpreters of this passage (Republic 546A) seem generally to
have assumed that the reference intended was some specific
number important in another Platonic context, or in pure
mathematics (e.g., Cardan). This conviction seems to have
operated as an unquestioned assumption, regarding which there
was a consensus of scholarly opinion. The details of the text
were generally assumed to result from Plato's desire to write in
a deliberately enigmatic form.
In the second half of the nineteenth century the idea became
current among Plato scholars that this number was intended to
represent the relation of the Great Year, or cosmic cycle, and
the cycle of human constitutions, as society declined. The first
phrase was then construed to mean the IO,OOO-year cycle of the
Phaedrus myth, with which the political cycle was assumed to
be in 4:3 ratio (hence equal to 7,500 years). Considerably more
attention was paid to the detail of the passage than had been
in previous interpretations, apparently on the assumption that
though an enigma, the answer should somehow be written into
the riddle. Many variant details of interpretation were sug-
gested, but the consensus as to the numbers Plato had intended
was so marked that Tannery, who thought the key to the text
had been lost, believed that the loss was not too serious, since
there was general agreement that this was what the passage
must have meant. 61S
With the twentieth century, a new "common sense" is re-
flected in Plato scholarship. The impact on philosophy of new
 GEOMETRIC METAPHOR 147
empirical, scientific discoveries and the relatively greater aware-
ness of the complexities of astronomy (against which the human-
istic philologists of the previous half-century had been self-
insulated), awakened a sensitivity to the same elements in the
culture of Greece; and Plato's text comes to be appproached
with the assumption that what had before been taken as an
enigma presenting an extraordinarily naIve notion about his-
tory, may in fact have been meant as a technical, scientific pas-
sage presenting some complex and conjectural scientific hy-
pothesis, with some discoverable empirical ~eference. The "ob-
scurity" is then no longer seen as the resrilt of a nineteenth-
century romantic passion for Tiefsinnigkeit on the part of a
Plato who was a German philologist in spirit, but as the result
of our loss of the technical, scientific vocabulary used by a Plato
who was in fact more of a twentieth-century laboratory scientist.
(Compare A. E. Taylor's interpretation of the meaning of
Plato's "likely story" in the Timaeus.) Being ignorant of this
technical vocabulary, certain scholars have advanced two types
of argument: (I) they have emphasized the technical use, by
post-Platonic scholars, of terms found in Plato, and (2) they
have attempted to find in other Platonic dialogues mention of
some empirical phenomenon which this passage might be
thought to explain. A Plato recreated in the image of the early
twentieth century is evidently capable of framing hypotheses of
vastly greater complexity than his nineteenth-century counter-
part, .and indeed one mark of. a great philosopher lies in his
capacity to grasp and formulate such a complex hypothesis, so
that a priori the text is approached with the expectation that
"the number" described may be an extremely large one. Per-
haps also this "scientific" Plato is thought of as sharing the
obvious penchant of his twentieth-century counterpart for using
small standard units of measure-the gram, the second, etc.-
even when measuring large quantities.
Unfortunately, the later technical meanings of terms which
are employed in this passage represent a distinct post-Platonic
development of mathematics as a separate science, with a set of
univocal technical terms. These later usages become progres-
 Plato's Mathematical Imagination

sively less reliable guides to interpretation as we move back

toward their prespecialized origin. Further, the attempts to find
the referent of this number in other Platonic dialogues are left
fairly uncontrolled, and overlook the fact that a difference in
dialectical context will be reflected as a genuine constitutive
difference in mathematical imagery. In any case, the number
representing the scholarly consensus of this period (12,960,000) 66
cannot be established in either of these ways, without extraP9la-
tion from tradition and other texts to Babylonian astronomy
and Neo-Pythagorean astrology. But if one is allowed this
spatial and temporal latitude in extrapolation, almost any large
number could probably be defended as an "interpretation" of
the text about as plausibly as any other. (For example, we could
very well defend Thomas Taylor's thesis that the number is
1,000,000 by multiplying the great year by 100 to fit our taste for
larger numbers. This could be eruditely supported by citations
from the Neo-Platonists to show that any cosmic periodic cycle
ought to be represented as a power of the perfect number, 10.)
The development in the first half of the twentieth century
of a more critical attitude toward method and its role in history
and philosophy, the availability of the results of critical philolog-
ical studies, and work on the history of science suggest a new
"common sense" for the scholarly consensus that should under-
lie an interpretation of Plato's text. This common sense hinges
on the idea that the method by which a doctrine is constituted
should be the method by which it is interpreted. In other words,
in interpreting Plato we should self-consciously adopt the meth-
ods which Plato himself advocates for textual construction and
interpretation. This attitude, if we retain the fairly hardheaded
notion of philosophy inherited from the past generation of
scholars, takes away the mysterious fascination of the mathe-
matical imagery in the dialogues, and views it as a pedagogical
adjunct to the dialectical enterprise. It seems that the very
presuppositions which underlay the interest of the nineteenth
century in these images operated to prevent a correct interpre-
tation of them, since the texts were Platonic, but not the
maxims of interpretation.
 GEOMETRIC METAPHOR 149

The present interpretation is put forward as an example of

this reflexive application of Platonic methods to Platonic texts.
The enterprise is not a facile or easy one; ten years of intermit-
tent research lie behind the present section. In spite of this
investment of effort, I have no illusion that the interpretation
as here put forward is definitive or complete. I am convinced,
however, that it is the only right beginning. It aims to avoid the
effect of indeterminacy which is created when the interpreter,
taking his equipment from external sources, intrudes expecta-
tions and tactics that deform the text into a meaning acceptable
only to scholars of his own generation. These scholars, unaware
of their latent assumptions (since these are part of their common
sense), arrive at a consensus that the passage really means what,
had they themselves written it, they might have intended.
Adam's interpretation is currently the most widely accepted.
It will be found that my proposed interpretation agrees with
Adam's philological conclusions, except for (a) the reading
lxaOLov and (b) the interpretation of l'Ql~ aV;'l1{}e(~. Since the
final number in my interpretation is of the form x 4, there
would be no necessary disagreement with Adam. His case for
~(>l~ aV~1'l'f)E(~ as meaning "raised to the fourth power" has been
attacked as the weakest point in his treatment. The scholion on
Republic 587 may seem, however, to give some plausibility to
his unorthodox construction. Apart from these points, the ob-
jections to the interpretation are philosophic and aesthetic
rather than philological.
In the first place, Adam violates the aesthetic principle that
parts should be functional, by making the alternative construc-
tions with the two diameters of the pempad an unnecessary and
confusing duplication. Further, he does violence to the dramatic
consistency of the character of Glaucon, who in Book ix is
baffled by a simple multiplication up to 729, but is not similarly
puzzled by the elaborate computation of 4,800 X 3,600, which is
attributed to the Muses by Adam. Finally, as I think has been
shown in my calculation of the two harmonies, if the number
2,700 is the second harmony, Adam's arithmetic is impossible.
The major objection, however, seems to lie in the fact that
150 Plato's Mathematical Imagination

Adam's reconstructed diagrams confound, rather than elucidate,

the relation of this passage to Plato's similar images and to its
immediate dialectical context. At best, on this view, the nuptial
number becomes an elaborate digression, not an image giving
new pedagogical insight into the architectonic of the Republic}
and must seem to be a very unworthy and unlikely counterpart
of the brilliant contextual image of the divided line.
The earlier interpretation which I proposed (in The Role of
Mathematics in Plato's Dialectic) Appendix A) had the merit
of simplicity for the postulated construction. I took the harmo-
nies as produced from a line of four (hence in 4: 3 ratio with
the line of three stages of human growth), and from the line of
five. The square harmony I construed as a square bounded by
the line of four, minus one, and the line of five, minus two.
This entailed identifying the line of four, which suggests the
divided line, as the "rational diameter of five," and the line of
five, representing types of character or state, as the "irrational
diameter." The phrase "thrice augmented" could have been
interpreted as a projection to an equal distance in the third
dimension, giving a cube 3 x 3 x 3, and the final phrase as a
description of the entire figure, which could be "one hundred
cubes of three" viewed from the side. But the pattern of con-
nectives suggested so strongly that 2,700 was the area of the
rectangular surface, not the entire figure, that I used a more
complex projection, making the "triple augmentation" stand
for three projections to distance three in the third dimension.
The fatal defect of this, so far as I can see, is that it requires
"the irrational diameter of the pempad" to be the pempad it-
self, and that is indefensible. A further objection is that the
parallel imagery of the passage really requires a mathematical
"irrational" to represent the "irrational" contextual referent.
I anticipate, however, that some later interpretation will syn-
thesize the various factors and references of this passage into
some such simple, intended construction.
 GEOMETRIC METAPHOR 151

VII. THE TYRANT'S NUMBER

Republic 587 •
"... then the tyrant's place, I think, will be fixed at the furthest
remove from the true and proper pleasure, and the king's at the
least." "Necessarily." "Then the tyrant's life will be least pleas-
urable and the king's most." "There is every necessity of that."
"Do you know, then," said I, "how much less pleasurably the
tyrant lives than the king?" "I'll know if you tell me," he said.
"There being as it appears three pleasures, one genuine and two
spurious, the tyrant in his flight from law and reason crosses the
border beyond the spurious, cohabits with certain slavish, mer-
cenary pleasures, and the measure of his inferiority is not easy to
express except perhaps thus." "How?" he said. "The tyrant, I
believe, we found at the third remove from the oligarch, for the
democrat came in between." "Yes." "And would he not also dwell
with a phantom of pleasure in respect of reality three stages re-
moved from that other, if all that we have said is true?" "That is
SO." "And the oligarch in turn is at the third remove from the
royal man if we assume the identity of the aristocrat and the
king." "Yes, the third." "Three times three, then, by numerical
measure is the interval that separates the tyrant from true pleas-
ure." "Apparently." "The phantom [eidolon] of the tyrant's
pleasure is then by longitudinal mensuration a plane number."
"Quite so." "But by squaring and cubing it is clear what the in-
terval of this separation becomes." "It is clear," he said, u to a
reckoner." "Then taking it the other way about, if one tries to
express the extent of the interval between the king and the tyrant
in respect of true pleasure he will find on completion of the mul-
tiplication that he lives 729 times jiS happily and that the tyrant's
life is more painful by the same distance." "An overwhelming
and baffling calculation," he said, "0£ the difference between the
just and the unjust man in respect of pleasure and pain." "And
what is more, it is a true number and pertinent to the lives of
men if days and nights and months and years pertain to them."
"They certainly do," he said.

In Republic ix. 587D, Socrates begins his "marvelous and

baffiing calculation" of the unhappiness of the tyrant by stating
• Trans. Shorey, Republic, II, 397, 399.
152 Plato~s Mathematical Imagination

that the "distance in linear measure between the aristocrat and

the tyrant is nine." This is surprising since in his list of lives
there are only five, the aristocrat being the first and the tyrant
the fifth. He explains this by saying that the oligarch is l hird
from the aristocrat and the tyrant third from the oligarch( This
has been variously regarded by Plato's readers as pleasantry,
nonsense, or arithmetical sophistry, but no one has contended
that in its context it is not a non sequitur.67~dam, in his notes
(Republic, II, 360-61), suggests that we may assume some inter-
mediate characters which were not previously mentioned to
account for this "shift to a larger modulus," but he does not
suggest how or why this might be done, since he believes that
in any case the ~creal purpose of the computation is to arrive
at the number 729,'~chosen for astronomical reasons. 68 There
the matter stands.
I should like to suggest that the computation in question,
while compressed, is perfectly sequential; that Adam's conjec-
tured "intermediate constitutions and characters" are in fact
clearly implied in Plato's dialectical context; and that from a
similar context (that of the list of lovers in Phaedrus 248) one
can say whI ther~i1~i,Jlin.~_'§JJ~h .. characte~.and. what
t4,~::inte~ia.t~ . .,.ate.
The simplest mathematical image illustrating the combina-
tions of two independent factors is a square subdivided into
cells. In the classification of divine and human objects and
imitations at the end of the Sophist, Plato uses precisely such
an image for this purpose,69 and it is a usage which has prece-
dent in both mathematical and medical writers before Plato's
time.
When in the Republic we are told that the distinction be-
tween aristocrat, timocrat, and oligarch lies in the dominance
of the parts of the soul,70 and that the distinction between oli-
garch, democrat, and tyrant can be stated only in terms of the
reality of the objects of their desire,71 it is clear that the list of
five types of state about which discussion centers in Books viii
and ix involves the introduction of two distinct principles of
degradation successively. In other words, the .five constitutions
 GEOMETRIC METAPHOR 153

and characters discussed represent a gnomon-section of that

matrix which would illustrate the complete result of these
factors in their simultaneous independent operation. 72 The total
matrix would have to provide, for example, for tIle distinction
between two characters both of whom are guided by the rational
part of the soul, but in one of whom the natural strength of this
part permits contact with realities, whereas in the other its
weakness limits such contact to (artifacts (thrice removed imita-
tions»)But the path of degradation selected for schematic treat-
ment shows each step to be the same as or worse than the pre-
ceding constitution in respect to both factors, going from best
to worst in a continuous movement. The absence of the distinc-
tion of rulers-auxiliaries-artisans 73 in this scheme, a distinction
of nature and character basic to the structure of the argument
of the Republic and to the aristocratic state, is a sign of the in-
adequacy of the five-step path of degradation as a schematic
summary of the entire relevant argument.
In Phaedrus 245, nine classes of lovers are hierarchically ar-
ranged in terms of their degradation from the best lover through
the "double load of ignorance and vice." 74 The nature of each
kind of lover is indicated by a description of the typical vocation
or preoccupation of a man of the given type. When there is
not a unique or exact correspondence between vocation and
character, more than one characterizing term must be used; and
in fact there is only one case in which a single vocation is the
only one typical of a given type of lover, hence only one case
for which a single term will suffice-that of the tyrant. TIS When
this list is compared with the list of kinds of character in Books
viii and ix of the Republic, or with the list of kinds of citizen
in Books iii-vii, certain correspondences are very striking. (See
Fig. 63.) Within each triad, these characters are arranged in a
descending order of virtue defined by the dominance of the
rational, spirited, or appetitive part of the soul. Hence the cor-
respondence between the first three "kinds of lovers" and the
first three "types of human character" is exact and complete.
The successive lists of threes in the Phaedrus are arranged in a
descending order of memory, since the soul sinks through forget-
154 Plato's Mathematical Imagination

fulness as well as vice. Lovers who recognize things of the soul

as objects of love differ from those who ascend no higher than
love of the body, and from those who can see beauty only in
artifacts, in the adequacy of their recollection of the forms.
This accounts for the identity of the three classes in the ideal
state with the first, fourth, and seventh kinds of lovers in the
Phaedrus list, for these classes differ in natural intelligence, but
are alike "virtuous," and hence have the same relative domi-
nance of parts of the soul.
If the joint variation of these two factors, ignorance and vice,
is represented by a square, the resultant image is Figure 59,
preceding.
If one accepts the equivalence, which Socrates emphasizes in
Republic ix,76 of the democratic character and the mimetic
artist's occupation, the degradations of the best constitution
presented in the five-part scheme of Republic viii-ix are seen
to follow the gnomon formed by the left side and bottom of
this matrix. The reason for this selection is, as has been said,
that the other "intermediates" would not represent the pro-
gression from best to worst in direct and unqualified form,
since the ideal artisan, for example, though inferior to the
general or timocrat in wisdom, is his superior in temperance.
The differentiation of philosopher, athlete, and artisan has al-
ready been developed in Books iii and iv of the Republic; the
separation of Sophist-prophet and statesman-general is left to
a later trilogy of dialogues for its development.
From the analogy of the Republic and Phaedrus images ex-
hibited above, certain definite conclusions follow. The first is
that the distance from aristocrat to tyrant equals 9 because, as
Socrates says, the distance being represented linearly is "the
number of a plane figure." 71 The Hpl ane figure" in question
is the matrix showing the types of human character as combina-
tions of intelligence and motivation determine them. As Adam
suggests,78 therefore, the peculiarity of computation results from
the recognition of intermediate types of character, necessitated
by the dialectical schema represented. This final image, sum-
marizing the judgment of the relative happiness of the good
 GEOMETRIC METAPHOR 155

man and the tyrant, extends in scope beyond Books viii and ix
to embrace the entire antecedent discussion of the Republic.
The "intermediates" suggested by the Phaedrus diagram in-
clude the soldiers and artisans central in the discussion of Books
ii-iv, and the successive speakers in Book i were a merchant, his
son, a lover of the poets, and the Sophist Thrasymachus, speak-
ing in praise of tyrants.
The reduction of a plane image to a linear distance is a
natural step in the construction of a mathematical image which
can be at most three-dimensional, but which must represent
the interaction of more than three variable factors in the dialec-
tical context.
Given this representation of the nine types of character as
the serially ordered segments of a line, what is the significance
of the "squaring and cubing" by which the real distance in
pleasure between best and worst "is rendered clear"?
Evidently, there is more than one "dimension" to be con-
sidered in estimating relative pleasure, and in each of these
dimensions, the same nine-term order obtains, with the tyrant
at the end. From the earlier statements of the criteria needed in
forming a judgment on this matter, and from the organization
of Book ix itself, in which these three criteria are in fact suc-
cessively employed, we derive a clear indication of what these
intended dimensions are. 79
In addition to the intrinsic excellence and reality of the
pleasure sought by a man of a given type, relative prudence
and opportunity for experience are factors which determine
whether his life will be a pleasant one, or the reverse. SO <llQ6V'Yl(Jl~
in other Platonic contexts is associated with the ability to select
means which do in fact realize the ends for which they were
intended. In the present discussion, this takes the form of choos-
ing appropriate pleasures for each of the parts of the soul. If a
love of victory or a love of gain is made the criterion by which
courses of action are selected, no part of the soul realizes to the
fullest degree its own appropriate pleasure. 8 ! Since the possible
criteria for prudential choice are identical with the possible
characters of men, the same nine-term list which serves as a
 156 Plato's Mathematical Imagination

schematism of character may be used as a schematism of pru-

dence as well. The interaction of vov; and qJQ6vl1o'l~ may there-
fore be schematically represented by a 9 x 9 square matrix. In
. this matrix, each cell will represent the relative pleasure of a
life in which a man with given intelligence employs one of the
various possible criteria in his prudential choice. 82 Whether or
not all of these combinations do exist in fact (it is perhaps de-
batable how far the wise man can be misguided in respect to
his own practical interests), such a schematism is an effective
way to represent the simultaneous disparity in respect to choice
of ends and efficacious selection of means to implement them
which exists between the philosophic monarch and the tyrant.
In addition to prudence and intelligence, opportunity is
needed for a life which is in fact to be pleasant. 88 Here the
earlier argument from experience, in which it is shown that the
tyrant has of all men the least opportunity to act on his deci-
sions, is relevant. As in the nuptial number, the extension of
the image into the third dimension introduces a temporal
factor. The final cube of 729 cells is a three-dimensional matrix,
each constituent cube of which represents a combination of
nature-prudence-opportunity, which are the three factors rele-
vant to a comparison of the happiness of the life of the best man
and that of the worst. The distance between the first cell and
the last, computed by arranging them in serial order, is, as
Socrates says, 729.
It will be noted that certain elements of this image repeat
imagery found in the interpretation of the nuptial number
already proposed. This gives an explanation of the extreme
compression of the present passage, since these elements retain
the same interpretation assigned them in their earlier intro-
duction.
Glaucon is bewildered when Socrates states this interval as a
definite arithmetical number. He does not express this same
bewilderment at the sudden introduction of squaring and cub-
ing, which seems to us the most baffling part of the entire pas-
sage. This suggests that Socrates' number (729) is a logistical
device to designate an important geometrical relation in a
 GEOMETRIC METAPHOR 157
diagram, and that it is this passage from geometry back to
calculation that especially troubles Glaucon. This is made par-
ticularly plausible by the fact that the other differences between
the just man and tyrant are dismissed as dlltlXavov immediately
following this calculation of relative pleasure.
Socrates' remark that this is a true number if days and nights
and months and years pertain to human life has again been
made an argument for the astronomical interpretation, that the
entire purpose of Socrates' geometrical gyrations has been to
arrive at this number, 729, which equals the number of days
plus the number of nights in a year. If the remarks made on
the general purpose of the squaring and cubing are correct, the
correctness of such an astronomical interpretation would not be
ruled out, but it would be necessary to say that the primary
purpose of Socrates' calculations is not to indicate that this is
the measure of careers which actually take place in time. The
truth of the number would depend upon its inclusion of some
comparison of such temporal careers. The mention of four
periods of time rather than one seems gratuitous, unless these
are periods relevant to different kinds or degrees of pleasure.
(Thus the tyrant's dreams are more terrifying than those of the
just man,84 as we find later; the life of the unjust man does
not end in honor after his early years are past; S5 his daily pleas-
ures will be less; and the months in this list may suggest seasons
in some similar connection.) To interpret the passage as a cal-
culation of number of days and years still leaves most of the
passage unaccounted for, though it adds a new insight to the
function of the third dimension in the foregoing interpretation.
The notion that the metaphor here is that of computing the
area of the shadow of the tyrant's unhappiness, which increases
as the square of the distance, leaves the connection so loose
between this image and the antecedent discussion, that the
presentation of the image as though it were a summary of
antecedent dialectic is not intelligible. s6 In effect, the connec-
tion between the image thus interpreted and its contextual
passages is of the sort Coleridge describes as "fanciful" rather
than "imaginative"-the fanciful connection is based on a single,
158 Plato's Mathematical Imagination

shared" external relation, the imaginative on a likeness of the

properties of wholes. But the fanciful diagram is of very little
pedagogical use, and seems not to be typical of other instances
of mathematical imagery in Plato. This interpretation unjustifi-
ably attributes to Plato so loose a sense of dialectic that he
could substitute a bad pun on the word "shadow" for a helpful
illustration of the serious point at hand in the argument. Per-
haps the interpreter would here again invoke the "freakish"
sense of humor attributed to Plato by readers who want to ex-
plain away the Parmenides.
Figure 62
SCHOLION FIGURE TO ILLUSTRATE THE TYRANT'S NUMBER
(Greene, Scholia Platonica, p. 269)
ARISTOCRAT TIMOCRAT OLIGARCH DEMOCRAT TYRANT
2 3 6 9
This figure suggests (as does the other scholion on the passage
taken from Proclus) that as the distances become greater, some shift
of modulus is required, introducing a greater unit for their meas-
urement. (This is essentially Adam's interpretation, discussed in the
text.) There is no suggestion of the fact that the number is square
because it refers to a square plane figure.
Figure 68A
PARALLELS AMONG PHAEDRUS 248, REPUBLIC VIII-IX,
AND REPUBLIC III-VII
"PHAEDRUS" 248 "REPUBLIC" VIII-IX "REPUBLIC" III-VII
1 Philosopher 1 Aristocrat Ruler
2 General 2 Timocrat
3 Citizen-Merchant 8 Oligarch
4 Gymnast 2 Soldier
5 Prophet
6 Poet 4 Democrat
7 Artisan 3 Artisan
8 Sophist
9 Tyrant 5 Tyrant
The Phaedrus list is arranged within each section in an order
of relative virtue, the sections in an order of relative ignorance.
 GEOMETRIC METAPHOR. 159

Figure 63B
TYRANT'S NUMBER: FIRST STEP OF CALCULATION

RATIONAL 1 Philosopher 4 Gymnast 7 Artisan

VICE SPIRITED 2 General 5 Prophet 8 Sophist
APPETITIVE 3 Merchant 6 Poet 9 Tyrant

SOUL BODY ARTIFACTS

IGNORANCE

This is the Phaedrus matrix reconstructed as the geometrical

figure underlying the use of the number nine in the nuptial num-
ber passage. Its present relevance is to show why the linear measure
of the distance from the aristocrat to the tyrant is a square number;
the interaction of the two three-part principles produces 9 com-
binations, and a reduction to a single list thus comes out with 9
as the distance between the first cell and the last.

Figure 64

I ,
R 2
3

4
S 5
6

7
A 8
9 81

2 3 456 789
Prudence (Skill in Means)
TYRANT'S NUMBER: SECOND STEP OF CALCULATION (SQUARING)
160 Plato's Mathematical Imagination

The linear representation of Figure 64 is used here as the basic

line from which the final image can be constructed by "squaring
and cubing." The present figure shows the first step in expansion,
in which the ordered list of kinds of character is used to differen-
tiate pursuers of pleasure both in respect to their relative reason
and their relative prudence.

Figure 65

I
2
~""'--I--+--+-"""""'--I--+--r
3
§4
~ I--+---+-+---I--+---+---+-+--r
~5 I--+---+-+--+--+---+---+-+--r
6
t---+----+--t---I--t---+----+-+---r
7
t---+---+--I----I--+--+---+-+---r
8
t---+----+--t---I--+--+----+-+---f"""
9
2 3 4 5 6 7
Prudence
TYRANT'S NUMBER: THE FINAL CALCULATION

The final stage of calculation again uses the nine-step scheme

as an axis to represent the relative opportunity for enacting the
prudential conclusions which men of each type have. Each cell in
the figure now characterizes the pleasantness of a career in terms
of (1) reality of the ends to which it. is oriented, (2) relative skill
in determining means to those ends, and (3) probability that there
will be a chance to employ those means. The summary computation
of Socrates shows the distance between first and last cubes in this
matrix to equal 729.
 GEOMETRIC METAPHOR. 161

VIII. THE MYTH OF ER-ASTRONOMY

a. The Unity of Republic x

Book x has frequently been referred to as the most loosely con-
structed and least unified book of the Republic. It is apparently
easy to overlook the tight, underlying tripartite structure of
the argument, which finds frequent analogues in the Timaeus
cosmology and reflects the levels of the divided line. The con-
demnation of mimetic artists is expounded in a recapitulation
of the abstract dialectical proportions which summarize the as-
pects of the theory of ideas that appeared in earlier matrices.
In particular, it seems to repair the omission, in Book vi" of an
investigation into the "objective correlates" of the stages of the
divided line. 87 The transition from form to process, from pure
dialectic to personal history, is mediated by the second section,
devoted to the contemplative and moving activities of the im-
mortal soul, which would be simple in its own nature but as
we know it is multiform. 88 Finally, in the history of the journeys
that individual souls have made, in the vivid story of'the return
from death to life, we find an artist at work depicting the varied
trajectories which individuals endowed with intelligence and
immortality have actually marked out in space and time.
When in the Timaeus the perfection of the eternal model
is embodied in a beautiful, projected material process through
the mediation of the world soul, no one has ever accused the
argument of loose connection. But it is this same tripartite
argument which dominates the tenth book of the Republic.
Thus the much resented condemnation of poetry is the ultimate
judgment of pure reason; but it is a judgment by which souls
embedded in time and change cannot be guided, for their in-
sight requires such metaphysically inferior imaginative embodi-
ments as Plato has given his conception of cosmic justice in his
great poetic ~yth of immortality.
In Plato's Myth of Er 89 the properties of the model of the
universe, which all the souls are shown before they choose their
next lives, have been inadequately understood. The purpose
of this model is to show how divine justice operates throughout
 162 Plato's Mathematical Imagination

t1the world-order; the closing myth of a dialogue on justice very

, properly deals with cosmic justice in its relation to individual
choice. The justice of destiny operates in such a way that the
pattern of life which each man chooses freely is the pattern that
he himself grows to resemble; 90 if erroneous choices lead to
catastrophe, the fault is in the chooser, not in the natural
order.91
The moment of choice hangs balanced, as it were, between
the attraction of the eternal forms, always visible to reason, and
the habits and conditioning developed during the individual's
previous career. It is these antecedent habits which constitute
each man's "genius" and impel him to the choice which leads
him to meet his "destiny."
In the same way, in the universe, the motions of the present
and the past both operate and balance, producing a system in
which there is a neutral moment in the progress from past to
future: a timeless "now" in which future and past are alike
ingressive. In the cosmic scheme, the motions of same and
other are held in equilibrium, so that "justice" is served by
each sphere holding to its constant path of motion. The simi-
larities between the working of the model of the heavenly
motions and the cycle of human life lie in the polarities of
future and past, of same and other, and in their point of mo-
mentary balance, which the Law of Nines discloses.92
Thus the sizes, colors, and velocities of the moving hemi-
spheres all have a neutral equilibrium zone or point of balance
of the hemispheres following the motion of the same, and those
moved (with constantly decreasing retrograde velocity as the
outside is approached) by the motion of the other.
This image is entirely in the spirit of the Ionian philosophers
who, like Anaximander with his chariot wheels, turned their
talents as technologists to the invention of a physical mechanism
duplicating astronomical motion, and in this context of Neces-
sity that spirit is appropriate. !~L~~~~t~~~~~.Ig:!!.!!l~1_~. ~. !Jl!9JVn,
v!E~q~,~ i~ ,.,~ A~altbI.J!Jl4 . :q~tq!"~1_c~rnditiQJl. a£~ the~~-S,L)lll~J!ndJhe
~~~!~~L!!~.!~~]_Q!Q~!.j~~!, j!!~!,,}?..~~.l.~~~". ~~..~,~E~£~~!,!l,~,,~.~~9:~.~:
~xe~r~~!l'~',J)~-Jh~,.~,~,Il~~,~.,~~i!~~~,.J9p,~~ ~~.r~~,eleasa~.~<o<!~~!:~~t
 GEOMETRIC METAPHOR 163

Q,thiLQ1?l!9j.!!~~~l}.!!l. .!1l!§-.~Eec.~!,!21l is g~:g~Ogcnt gp QUI: bdici

~,!!~ t~~ ~j!!§!i~~ ,.,QfJl-MlU'.e..-A purely rational argument for the
perfection of the cosmos, based on its exemplification of natural
law in its behavior, cannot of itself create the conviction that is
needed to translate pure intellectual assent into applied pursuit
of justice. As the importance of education in poetry has shown,
true knowledge involves presentation of concrete, particular
embodiments of principles as well as their apprehension in the
ab~tract. The physiology of appetite in the Timaeus makes clear
Plato's conviction that it is only through the projection of vivid,
associated imagery that concepts can move the appetitive part
of the soul. 93
The myth represents such a concrete presentation of prin-
ciples in a form designed to carry conviction to the appetite.
A valid myth not only carries with it the sensuous vividness
which secures attention but also has an intelligible order in its
parts which is recognized and is the basis of aesthetic satis-
faction.
For the appetites, seeing is believing. The prophet is able to
devise myths which make his hearers see the concrete aesthetic
value of the principles of morality. Before the conviction needed
to secure harmony of interest and reason can be complete, how-
ever, the imagery of such myth must be checked against the con-
tent of normal sense-experience. The myth shows how concrete
things should seem if they are in fact in accord with the intel-
ligible principles of order. Perception shows whether or not
the mythical imagery coincides with the events of our concrete
experience. The final appeal of reason to appetite must there-
fore be empirical; when perception coincides with the images
accompanying thought, conviction results.
Various factors interfere with a sound judgment of this coin-
cidence of predicted application and perception. 94 This cannot
be, therefore, a completely naive empiricism; it must be safe-
guarded by conditions which insure that the perceiver is per-
ceiving rightly. These conditions are carefully met in the train-
ing in music and gymnastics which is to be the basis of public
education in an ideal state.
164 Plato's Mathematical Imagination

If in fact the cosmos is reasonably ordered and run, even in

its sheer appearance there should be discernible both a beauty
and a perceptible differentiation and integration of parts which,
when hindrances to right perception are removed, will exactly
correspond to mythical prediction and will carry complete prac-
tical conviction to the perceiver.
The myth at the end of the Republic has the double function
of showing how a just cosmos should look, and of indicating
that the appearance of the world we live in coincides clearly
with this expectation, so that mere detached inspection must
carry belief that this world is a just one.
In this connection, mere vividness is beside the point; aes-
thetic satisfaction demands further a perception in that vivid-
ness of a functional order, a perception which reason alone can
finally verify but which is immediately felt as satisfaction by an
appetite habituated to the kind of order that the projected and
associated images of thought in its experience always possess.95
The vision of the universe,96 shown to all souls before each
chooses its next life, is presented mythically as justification for
the responsibility of each soul for the consequences of its choice;
each has been given a chance to see what sort of world it will be
living in, and the prudent soul takes account of this in choosing.
The elaborate detail of the description of this cosmic model
is functional and intended to show the latent order of its appear-
ances. This latent order, through its impact on the appetitive
part of the soul, creates conviction that we can see empirically
that the world is justly run. The soul with an aesthetic sense
actively functioning will at once be aware of this latent order
as an accompaniment of its immediate perception; the less
sensitive perceiver may need to have recourse to the structural
formulae exemplified in this concrete instance before its func-
tionality is clear. Plato's account is presented by a soul of the
first type, but by the detailed report he gives of his observations,
it is expanded in such a way that the latent structure of the
system will be apparent to the less sensitive observer, willing to
reflect on the type of structure which these details of appearance
exemplify.
 GEOMETRIC METAPHOR 165

The phenomena treated in detail are the relative volumes,

colors, and velocities of the zones of th~ stars and planets. The
significance of these lists has never been quite functionally ex-
plained. Apparently, since the order of listing differs from
simple distance from the center, some latent feature of structure
is operative in each list to provide its principle of order. The
cumulative effect of these details should then be a clear indica-
tion of the type of functional structure which the apparent be-
havior of the model is exemplifying. Since the myth is a true
one, astronomical observation should suggest an empirical basis
for these principles of order. But no interpreter has been able
to t:econcile these reasonable expectations with the text to show
how in fact this bringing out of relevant structure has been
done.
Perhaps, however, interpreters have generally' tended to ap-
proach this passage with erroneous presuppositions. Any phe-
nomenon which can be empirically perceived will be observed
as an interaction of parts in a material space-time matrix.97 Any
form latent in this process is invisible and cannot be an object
of direct perception. D8 So far as pure observation can guide one,
astronomy deals with the properties of a world machine, acti-
vated by mechanical interplay of its constituent parts.D9 If, to
derive a theoretical astronomy from the data of observation,
reason infers the self-moving, divine character of planets and
stars, such an inference is observable only insofar as the me-
chanical behavior provides a model of the sort of motions that
would be expected.
Perception and appetite are, except for the feeling of aesthetic
satisfaction, insensitive to evidence of purpose or over-all plan.
We perceive separate objects influencing one another by me-
chanical contact; and, insofar as the bare imaginative picture
fits the thing itself, every object of experience must be inter-
preted as a machine. Certain physical properties of such ma-
chinery, however, are associated with the presence of a directive
function and integration in their design.
Among these perceptible evidences that some reason has
created a mechanical system one might mention (a) a sharp
166 Plato's Mathematical Imagination
differentiation of the individual parts, consequent upon each
having an appropriate contributory function; (b) an adaptation
of adjacent parts; (c) a total equilibrium of some kind in the
dynamics of the mechanism, resolving various disparate forces
into some more simple resultant motion. These are the per-
ceivable properties from which we infer a latent reason and
purpose as present in a mechanism.
The detailed appearance of the cosmic model seems to be
described in a way which underscores these properties. The
hemispheres are differentiated by mechanical restraint; each has
its own boundary, walling it off sharply from the adjoining
lones. 100 It has been shown that the order of the lists of colors
and volumes presents a kind of balance which would correspond
to the observable projection of adaptation of adjacent parts,
although this interpretation has never been proposed as the
explanation of the balancing observed. And the dynamic effect
of the machine is to translate a set of irregular and contrary im-
pulsions into a resultant of two simple circular motions with a
balanced distribution of forward and retrograde momentum per
unit of impulsive force.
Even for a disembodied soul, whose perception is entirely
accurate as long as that soul perceives without calculation or
reflection, all that one can legitimately expect is this sort of
observation of properties that are perceptible; and such proper-
ties cannot, in themselves, serve as adequate basis for a theoreti-
cal astronomy. This vision is supplemented in a myth by the
personifications of other attributes which are not perceptible.
Goddesses of harmony and justice appear who control and in-
habit the machine. But to the soul less privileged than Er to
see such supernatural personages, their presence must be in-
ferred by reason, or felt by appetite on the basis of those spatial
configurations of volume, mass, velocity, and color, to which
alone the senses can respond.
Though the general significance of this myth, and the various
details other than those of the astronomical model, have been
subjected to extensive and satisfactory interpretation, the model
itself, which plays a crucial role, has not been very well ex-
 GEOMETRIC METAPHOR 167

plained. An attempt to give a detailed and functional account

of the properties of this model is therefore still needed, and
seems to be the aspect of the present myth which falls most
naturally within the province of the interpreter of Platonic
mathematical imagery.
b. The Allegorical Intention of the Myth
The careers of the souls between incarnations, as Er describes
them, show a curious parallel to the career of the argument of
the antecedent books of the dialogue. The parallel seems to
begin with the pilgrimage across the treeless plain that follows
the experience of punishment and reward. If we note that it is
this fact of reward and punishment in the after-life that has
become the preoccupation of Cephalus, at the opening of the
dialogue, this parallel carries out the cyclical structure of the !
Republic, which begins with the old man's speculations on the
passage of the soul from life to death, and ends with the return 1
of the soul from death to life. These transitions link the prob- )'
lem of justice during a given life-span with the justice of the
cosmic order, which controls the greater cycle of reincarnation./
Life has been referred to earlier in the dialogue as a "race"
and as a "journey"; in the anticipatory enactment by the souls
of their next careers, the journey across the plain toward a place
where new insight is attained sounds like an allegorical counter-
part of the whole historical career-birth, primary education,
marriage, and war-from which men at last reach the insight,
either in the career of higher education of Book vii, or in the
reasoned judgment of the man of wide experience in Book ix,
that justice is natural and best. After a four-stage journey, end-
ing in a place ~hich clearly is above and beyond the realm of
becoming and which offers them a vantage point from which
they have the philosopher's view as "spectators of all time and
all existence," the souls are shown completing the final courses
of higher education prescribed in Book vii for rulers. In fact,
by a little forcing of the correspondences, an interpreter might
find them undergoing experiences in insight which are counter-
parts of the entire curriculum outlined earlier.
168 Plato's Mathematical Imagination

The inspection of the cosmic model gives direct insight into

the celestial motions studied in astronomy; the presence and
song of the Sirens bring them face to face with a personification
of the harmony which is the subject matter of harmonics; the
dialectical linking of these insights with the problems of human
affairs is presented through the personified goddesses of destiny,
who conduct the course of instruction and finally sum up the
lesson of the curriculum through their spokesman, the prophet.
The next episode after this education is its application to
practical affairs, just as, in the earlier discussion, the model
educational and administrative system of an ideal state having
been set up, the next consideration is the degradations of such
a state when actually transposed into the context of contingen-
cies and erosions of a career in time. Just as varieties of polity
among which a man must choose his allegiance occupy Book
viii and Book ix of the Republic, so a showcase of sample lives
of all varieties is presented to the souls following their instruc-
tion, and each must choose its next career. The factors differen-
tiating these lives offered for choice are a list familiar from the
previous contexts of the discussion, and represent combinations
of factors appealing to each part of the tripartite soul. Just as,
in the earlier context, opportunities for action beyond the con-
trol of the individual are one relevant component of the pleas-
ure of a given life, so, in this recapitulation, the fortune of the
lot is one relevant factor in the opportunity of each soul for
choice.
To underscore this parallel, Er reports a list of cases of actual
choosing in which there are examples of choosers dominated
by each of the parts of soul, ranging from the inexperienced
voracity of the holder of the first lot to the toil-hardened, dearly
bought sagacity of Odysseus, who chooses last and whose role is
clearly intended to symbolize the confirmation in concrete ex-
perience of the advantages of justice established earlier in ab-
stract argument. J

In describing the souls as drinking different amounts from

Lethe, Plato includes a mythical detail to "explain" the varia-
tion of individual intelligence, which is also recognized in the
 GEOMETRIC METAPHOR 169

Phaedrus myth, \vhere Plato says that lives differ in kind accord-
ing to their possessors' "ignorance and vice." 101
Apart from this detail, the allegorical parallel seems to end
with the pageant of souls transforming themselves into animals,
and the machinery of ratification and reincarnation completes
the frame proper to the allegorical picture.
Among its other advantages, such an allegorical interpretation
of the myth permits us to explain many of its details which on
any other interpretation seem somewhat out of place and puz-
zling. The foregoing treatments of mathematical imagery are
evidence that Plato's intuitive sensitivity leads him to choose
significant details for which we can reconstruct an intellectual
explanation by abstract analysis of the connections which cause
the chosen affinity to be felt.
The details chosen for explanation in the present treatment
are those of mathematical imagery. The arithmetical descrip-
tions of periods and the elaborate description of the dimensions
and properties of the astronomical model that is shown to the
souls are the problems peculiarly relevant to the present study;
but there seems to be no reason to exclude the possibility that
an attempt to interpret other portions of the myth along these
lines might finally disclose their intended significance as well.
As confirmation of this approach, it is worth noting that
Plato himself has done everything he could have, short of a
hopelessly anticlimactic explanation by Socrates, to indicate
to his contemporary reader that the passage at hand is intended
to be read as allegory. In the first place, the subject matter
places this story in a religious tradition in which death and
judgment after death are explained to the initiate by allegorical
pageantry. As Plato himself was aware, mythology gains in effect
and solemnity by retaining the established forms of religious
tradition. The revelation of doctrine to the mystery-religion
initiate took the form, as we have noted, of a vision of an alle-
gorical pageant or tableau. Plato's myth is told so successfully,
in a way which seems to use every device of artistry to present
the scenes described as successive visual images to the mind of
the reader, that it is almost impossible to read the passage with-
170 Plato's Mathematical Imagination

out such a visualization. Further, the relative importance of

the things the souls see is frequently indicated by giving a
visual vividness to their description, an equation of importance
with brilliance rather than with magnitude or priority in order
that is peculiarly intelligible if the author is consciously restrict-
ing himself to the tactics of the tradition in his choice of devices
for indicating the distinctions he wants to make. The reader
finds himself much in the position of the souls in the story;
a direct vision replaces inferential construction (as inspection
of the model of the heavens within the myth replaces study of
theoretical astronomy); but a correct inferential construction
results from applying metaphorical interpretation to the con-
tent of this direct aesthetic vision. (This point is later illustrated
in the discussion of the souls' intended reaction to the cosmic
model.)
As confirmation of the notion that this property of the myth
is deliberate, evidence can be cited to show that it is not typi-
cally so central a device of Platonic myth-construction. For illus-
tration, to choose a relatively trivial example, the reader is not
invited by the Platonic account to speculate on the color of the
clothing which Theuth, Thamus, the demigods of the Timaeus,
God at the helm of the cosmos in the Statesman, or Prometheus
in the Protagoras happen to be wearing, nor on the color of the
landscape in which they stand. Drlly in the myth of the "true
earth," at the end of the Phaedo, where the world is observed
from a great height above, as it is by the souls in the present
myth, is there an analogue to this importance of details of color.
We may note further that the device so regulates the tactics of
the description of Er's vision that the identification of the
planets and stars in the astronomical model is made only on the
basis of their color.
The myth of the cave, earlier in the dialogue, uses allegorical
illustration, with a point-by-point interpretation appended,
thus leading one to expect allegory from the earlier context.
One should note also that abstract concepts are presented to the
souls by personification, just as, we may suppose, such concepts
would have been displayed in the mystery pageant.
 GEOMETRIC METAPHOR 171

What seems to be conclusive evidence for this allegorical

intention, and for the reading of the astronomical part of the
allegory given below in my detailed interpretation, is that in
the Epinomis precisely this material occurs. Whether written
directly by Plato, or by a student of Plato who was thoroughly
at home in Plato's philosophy, this dialogue has exactly the
relation to the Laws that the Myth of Er bears to the Republic.
Once again, the rule of law is shown to extend throughout the
cosmic order. But in the Epinomis, flat literal statement re-
places the earlier mythical account of justice in the heavens.
In reading the detailed comments on the myth, which fol-
lows, it is worth noting specifically that (1) in the Epinomis
there are statements concerning the very large masses of the
sun and planets, (2) that from these is shown the absurdity of
any person or planet setting his or its course against the vast
momentum of nature, (3) that the motive principle postulated
for this system must be some sort of ccsoul," and that "mind"
and "volition" also have their analogues. (The third point
intrudes more of the world seen qua organism than the Myth
of Er specifically introduces.)
It may be suggested here that the analogy of Laws: Epino-
mis : : Myth of Er : Republic deserves further study. One ob-
jection to the possibility that the Epinomis is Plato's own, the
reference to "memoranda," can be neatly met if we identify
mnemonic diagrams of the sort recovered in this chapter as the
type of "memorandum" in question, and discover such diagrams
in the Laws.

c. Detailed Interpretation
Republic 617 •
Now when each company had spent seven days in the Meadow,
on the eighth they had to rise up and journey on. And on the
fourth day afterwards they came to a place whence they could see
a straight shaft of light, like a pillar, stretching from above
throughout heaven and earth, more like the rainbow than any-
thing else, but brighter and purer. To this they came after a day's
• Trans. Cornford, Republic, pp. 358-55.
172 Plato's Mathematical Imagination
journey, and there, at the middle of the light, they saw stretching
from heaven the extremities of its chains; for this light binds the
heavens, holding together all the revolving firmament, like the
under girths of a ship of war.
And from the extremities stretched the Spindle of Necessity,
by means of which all the circles revolve. The shaft of the Spindle
and the hook were of adamant, and the whorl partly of adamant
and partly of other substances. The whorl was of this fashion. In
shape it was like an ordinary whorl; but from Er's account we
must imagine it as a large whorl with the inside completely
scooped out, and within it a second smaller whorl, and a third
and a fourth and four more, fitting into one another like a nest
of bowls. For there were in all eight whorls, set one within an-
other, with their rims showing above as circles and making up
the continuous surface of a single whorl round the shaft, which
pierces right through the center of the eighth. The circle forming
the rim of the first and outermost whorl [Fixed Stars] • is the
broadest; next in breadth is the sixth [Venus]; then the fourth
[Mars]; then the eighth [Moon]; then the seventh [Sun]; then the
fifth [Mercury]; then the third [Jupiter]; and the second [Saturn]
is narrowest of all. The rim of the largest whorl [Fixed Stars] was
spangled; the seventh [Sun] brightest; the eighth [Moon] colored
by the reflected light of the seventh; the second and fifth [Saturn,
Mercury] like each other and yellower; the third Uupiter] whit-
est; the fourth [Mars] somewhat ruddy; the sixth [Venus] second
in whiteness. The Spindle revolved as a whole with one motion;
but, within the whole as it turned, the seven inner circles revolved
slowly in the opposite direction; and of these the eighth [Moon]
moved most swiftly; second in speed and all moving together, the
seventh, sixth, and fifth [Sun, Venus, Mercury]; next in speed
moved the fourth t [Mars] with what appeared to them to be a
counter-revolution; next the third [Jupiter], and slowest of all
the second [Saturn].
• The names in brackets are supplied by Cornford; Plato himself did not give
them in his text.
t Cornford has here followed, as have almost all editors, the text quoted by
Theon; the MS text, which reads "the third seemed to them to move in an opposite
direction to the fourth," makes sense neither astronomically nor mechanically in
this context. The differences between the manuscripts and Theon's quoted version
are discussed in the final part of this section.
 GEOMETRIC METAPHOR 173
The Spindle turned on the knees of Necessity. Upon each of its
circles stood a Siren, who was carried round with its movement,
uttering a single sound on one note, so that all the eight made
up the concords of a single scale. Round about, at equal dis-
tances, were seated, each on a throne, the three daughters of Ne-
cessity, the Fates, robed in white with garlands on their heads,
Lachesis, Clotho, and Atropos, chanting to the Sirens' music,
Lachesis of things past, Clotho of the present and Atropos of
things to come. And from time to time Clotho lays her right hand
on the outer rim of the Spindle and helps to turn it, while
Atropos turns the inner circles likewise with her left, and La-
chesis with either hand takes hold of inner and outer alternately.

Proclus reports an "older and better" version of the list of

sizes of hemispheres in Republic 617E. The diagram in Fig-
ure 68 is based on his "older" version and shows that the
manuscript version is preferable. The list of sizes is here given
as Proclus' "older and better" text presents it, with the cor-
responding terms, marked MSS, of the manuscript version
added in brackets where the two differ.

The circle forming the rim of the first outermost hemisphere

is the broadest; next in breadth is the seventh [sixth, MSS]; then
the eighth [fourth, MSS]; then the sixth [eighth, MSS]; then the
fourth [seventh, MSS]; then the third [fifth, MSS]; then the sec-
ond [third, MSS]; and the fifth [second, MSS] was narrowest.

Republic 616D-617D
(Chambray, Republique)

T1}V f)E aqJovf)vAou <pVO'LV elvaL 1'Otavf)s· 1:0 f.tEV axii~a OLa:7t£Q 1)
1'OU Ev'f}a~E, vOllaat ~E ~Ei E~ cI>v EAsyev 1'ol6v~e amov etvaL, OOO'TCEQ
av EL BV Evl fAeyciAq> aqJov~uAQ? 'KOtAQ? %at ESEyA'Ul-lllEvlp ~LaJ.l:n:eQE~
aAAor;, 1'OLOU1'Or;, EAa-r-roov ey'KEoL-rO tlQfA0-r-ro>v, 'Ka{}WteQ OL 'Ka~OL 0\ et~
&AAl}AO'U~ tlQflOt't'OVt'Er;" 'Kat ov-rro aT] 1:gL1'OV aAAOV 'Kat 'tEt'aQ1'OV 'Kat
(J).'Aour;, 1'E1'1'aQa~. '0'K1'00 YUQ etvaL 'tour;, ~VIl1tavrar;, a(J)OVaUAO'U~, EV
UAAl}AOlr;, EY'KELfAEVOUr;" 'KlmAO'Ur;, avoo-6'ev 'ta XELA'll qJa(vovra~, voo'tov
auvexer;, EVO~ oqJovavAo'U u1teQya~OfAEvo'U; TCEQl tftv 1)'ACl'Kcl'tllV· E'Ke[vllv
~E ~l(l IlEaou tou oy~6ou ~Laf.lTCeQEr;, EA~A<1a{}ClL. Tov fAEV o~v
3tQoo't6v 1'8 'Kat £;oo-r<11'O) a<p6V~UAOV XAa-r'U1'a-rov tOY Tot; XE(AOU;
174 Plato's Mathematical Imagination

XU'XAOV EXElV, -rOY ~€ -rOU EX-rOlJ ~EU-rEQOV, * TQl-rOV ~E -rOV tou tE-
TUQtOU, tETaQtOV ()E 'tOV tOu oy()OOU, 3t£l!1t'tOV ()E 'tOV tou £~~OIlOU,
E'X'tOV ~E 'tOV tou nE~nfOu, E~~O~OV ()E tOY tOu 'tQltOU, oy()OOY ~e
tOY tOu ~Em£Qou. Kat tOY fJEY tOU fA.EyLO'tOU nOL'XLAOY, tOY ~€ tOu
E~~61l0U AaJ.UtQ6ta'tov, TOV ()e tou oy~6ou to XQw~a dno 'tou E~~0110U
EXELV 3tQOOAUllJtOvto;, 'tOV ()E tou ()EUTEQOU 'Xat 3tElJJtfOU naQa3tAi)aLa
&.AA~AOL;, ~av{t6TEQa EXElVroV, 'tQLtOV ()E AeUXo-raTOY XQWf.la EXELY,
tEtaQl'OV ~E V1tEQU{tQOV, ~EUtEQOV ()E AEUXo'tl1'tL 'tOY E'Xl'OV. KVMEio{}at
~E ~~ OtQECPOl!£VOV tOY UTQax-rov OAOV IlEV TilY amity cpoQav, BV ~E
1'0 OAq> rcEQL<PEQOIlEvql tov~ llEV BvtO~ E1C't'a 'GMo'U~ 'tl}V EVaVttav
-r0 OAql ilQEl1a JtEQt<pEQEo{}aL, airtwv l)E TOUTOOV 't<iXLo-ra ~v tEvaL
tOY oY~oov, ()EmEQou~ ~E 'X at ufJa d.AAi}AOt~ TOV tE E~~OfJov 'Kat
E'X'tOV 'Xat nEIl1t't'Ov· TQlTOV ()E <poQ~ LEvaL, 00; a<pI.Ot <pal.vEo{}aL, E1Cava-
'XV'XAOU~EVOV 'tOY 'tE'taQtOv, 1'E1'aQ'tov l)E TOY tQI.tOY 'Xat 3tElltt'tOV 'tOV
()ElrrEQOV. ~LQE<pEo{}aL ()E au'tov BY to!; 'tii~ ·AvaY'XTl~ yovaoLv.
• Here the differences between the MSS tradition and the "older version" begin:

ex'tou, MSS e~l>6JA,ou, older version

'tE'taQ'tou, MSS oyl>6ou, older version
oyl>6ou, MSS £x'tou, older version
e~l>6J.tou, MSS 'tE'tUQ'tOU, older version
1t£J..W't'tou, MSS 'tQL'tOU, older version
't'QL'tOU, MSS ~e'lrt£Qou, older version
l)eu't£Qou, MSS 3t £J.U't'tou, older version
'tQ£'tov, Theon· 'tOv 'tQL'tOV, MSS

1. Some nonmathematical details. It is worth noting that

certain of the nonmathematical details of this myth can be
better shown to be functional if studied in the frame of refer-
ence which the notes above have suggested.
For example, the notion that the shaft of light holds the
heavens together "like the reinforcing ropes of a trireme,"
though its exact meaning has been much debated (since we are
not clear as to the details of ancient trireme reinforcement),
seems not to have been recognized as an extraordinary image
for Plato to have chosen. Unless the heavens are being viewed
under the aspect of necessity, however, they have no need what-
ever to be held together or reinforced. No such devices of
reinforcement are relevant to the theoretic astronomy of the
Republic, to the natural theology of the Laws and Epinomis,
to the proof of the existence of a calIse of mixture in the
PhilebusJ to the mathematical schematism of empirical cos-

"
 GEOMETRIC METAPHOR 175
mology in the first part of the Timaeus, or to the myths in
the Phaedrus and the Statesman. If Adam is right in his belief
that the reinforcing ropes ran around the sides of the hull
(and his picture of an ancient trireme model seems fairly con-
clusive on the point),102 tIle image carries out the idea of iron
bands of justice holding parts in place by mathematical con-
straint. This is suggested also by the statement that the shaft
and whorl are made of adamant and its alloys.
Further, the over-all analogy to a "spindle," which is in
general "like those in common use," also underscores the me-
chanical nature of the imagery. The world is seen as a machine
for spinning the threads of destiny, and this analogy may well
be carried over into details of the imagery in ways which have
thus far escaped attention.
2. The Sirens' scale. Extensive speculation has been devoted
to the relation of the described velocities of the hemispheres
and the allocation of the notes of the scale to the Sirens. These
discussions seem to lose sight of or minimize the intentional
echo here of the earlier image of political organization as a
harmony of the classes in the state, which, like v~·t'rh flEOll, and
urtcl1'l1 , accord with one another like the fixed notes of a scale. los
The Sirens must be viewed, in the light of this earlier analogy,
as personifying the projection of social harmony and justice
into the relation of parts of the cosmos itself, the final resolution
of the problem of combining images of harmony with those of
cycle. Since each Siren is presented as singing a single note, an
echo of the definition of justice as the virtue which keeps parts
specialized to their proper function is also projected into the
cosmic harmony.
One may wonder at the reason which led Plato here to use
Sirens as the mythical personification of the harmony of the
spheres. In the tradition which originated the legend, the har-
mony of the spheres seems to have been thought of as the prop-
erty of the Muses, and the Muses as patronesses of education
have already appeared in a personified form in the Republic. It
is the Muses, as goddesses of harmony, whose choir the spheres
are echoing. If the allegorical significance of the myth set forth
176 Plato's Mathematical Im.agination

above is correct, furthermore, the personification of the celestial

harmony is intended to have an educative function which the
Muses might more appropriately exercise.
In resolving this problem, the first point to note is that Plato's
Muses make their appearance in context as goddesses of intel-
lectual, not sensible, harmony. There is some ground for doubt
as to whether the study of "harmony" in the Republic has
anything to do with musical song except insofar as the beauty
of such art is due to the presence in it of an intelligible struc-
ture. Song, however, is the perceivable embodiment of the
more intellectual mathematical realm of harmonious structure
over which the Muses preside.
The substitution of Sirens for Muses is in this detail an exact
counterpart of the over-all personification of concepts as visual
(and auditory) imagery in the myth, and effects an emphasis on
the sensuous vividness of the symbol required to arrest the
attention of the inspecting souls and to register pleasurably in
their disembodied experience.
The Sirens are related to sensuous satisfaction in heard harm-
ony as the Muses are to the intellectual satisfaction of mathe-
matical insight.
3. Periods of time. In general, as analysis of the description
of Atlantis has shown, Plato exercised a meticulous sense of
fitness in his selection of arithmetical details. The myths he
recounts, some of which are explicitly identified by him as of
Pythagorean origin, show a preference throughout the choice
of detail for the number ten and its multiples and powers.
When, therefore, in such a myth, specific numerical details are
intruded which are not in this decimal tradition, it is at least
worth investigating the possibility that some feeling of the
greater appropriateness of his own figures led Plato to deviate
from the tradition.
The present myth presents many of the traditional uses of
powers of ten; in fact, the prominence of these details has been
adduced as an argument for its Pythagorean origin. It also pre-
sents many striking deviations which do n~t have the air of
having somehow been borrowed from some alternative tradi-
 GEOMETRIC METAPHOR. 177
tional use of illustrative number. Among these may be men-
tioned (1) the prominence of the number 9 in the workings of
the astronomical model; (2) the use of 7, 4, and 12 in the de-
scription of the soul's journey; (3) the appearance of 20 (2 X 10)
as the specific number of one of the lots drawn. (This last is
atypical, since it might have been more in line with tradition
to make the number 100 or 1,000, and would have been just as
probable in the story, unless we find some special relevance in
the actual figure chosen.)
The cosmic nines and the lot of Ajax will be discussed later,
in connection with their contexts. The key to the periods of
travel seems to lie (1) in the calendar, and (2) in the divided
line. After starting their progress, the souls journey on the
path we have identified as an allegorical progression from be-
coming to being through four days. At the end of the fourth
day's journey (the eleventh day), the trans-spatial and temporal
region at the center of the heavens is visible.
The training of the soul to perceive being rather than becom-
ing is presented in the earlier allegory of the cave as taking place
through four stages, given by the image of the divided line. The
fact that it has four segments is the main identifying character-
istic given in the nuptial number of a line which represents
levels of knowledge; there is probably a deliberate echo there-
fore, in the assigning of periods so that the souls have a four-
day journey.
We may recall also that the calculation of the happiness of
the just man and the tyrant is followed by Socrates' remark that
the number is true if days and nights and months and years
enter into the calculation in some way as relevant units of the
calculation. In other Platonic contexts, two different properties
of the number 12 bring it into occasional prominence. The first
of these is chronological: the number and its multiples are
easily equated with periods of time. The second is computa-
tional, and a third is theological. Of these properties, neither
the computational nor the theological seems to have any rele-
vance to the choice of 12 days as the period of time occupied by
Er's vision in the present passage. In a later discussion of the
178 Plato's Mathematical Imagination
12-celled schematism underlying the order of appearallce of all
but the first and last of the cast of transmigrating souls, a reason
will appear which will suggest why the dividing of a vision
which progresses through all stages of destiny into 12 units of
some kind would have an immediate appropriateness. This sub-
sequent use of the number, however, can hardly be developed
as the sale explanation of the detail here, since the assumption
has been made that it is antecedent context which suggests the
appropriateness of such a detail, and no 12-cell matrix appears
before this passage. In all of the mathematical imagery in which,
once the importance of harmony has been established, an at-
tempt is made to apply harmony to the cycle of temporal
process, one dimension or stage or factor of the construction has
been devoted to establishing in some degree the relevance of a
set of abstract distinctions or types to the actual periods of his-
torical time. In the nuptial number, a mythical factor represent-
ing the period of human life in the decimal tradition of Py-
thagorean myth is included. In the tyrant's number, the third
dimension is devoted to temporal factors, and a postscript points
out their relevance to the schematic calculation. A large part
of the fifth and seventh books deals with the appropriate chron-
ological ages for changes of marital and educational status.
On the basis of these precedents, one may be permitted at
least to suggest that the number 12 is intruded here as an al-
legorical time factor. The ordinary divisions of time are a
relevant measure of the stages of development of Er's vision,
but with the strange foreshortening which also appears to the
view of other disembodied souls from their point of vantage as
spectators of all existence and all rime. This foreshortening is
apparent also in the blurring of the planets in their motion in
the model into apparently concentric rings. If this myth al..
legorically recapitulates tIle earlier discussion, which in turn,
as we are told at its conclusion, recapitulates the lessons learned
by extensive human experience (except perhaps for the section
on education, which is omitted from the recapitulation at the
beginning of the Timaeus), the appropriateness of a temporal
reference seems clear. Since Plato's myth is about to show that
 GEOMETRIC METAPHOR 179

the motions of the planets-measured by years-are the mecha-

nism of human history, this suggestion of a chronological num-
ber as an appropriate detail gains in plausibility.
4. Why the planets are not identified by name. The various
planets are identified here, not by name, but by a description
of their appearance in the model. It seems incredible that Plato
could not have found for the planets contemporary names which
would have been popularly understood before the date of the
Epinomis. There is an artistry, however, in his refusal to use
such names in his myth, which justifies this peculiar circumlocu-
tion. To call a planet by its name would have put it in the
framework of an intellectual discussion of scientific astronomy.
But the time for such discussions is past; the role of intelligence
has been treated exhaustively in the first moment of this book
of the Republic. What the model of the heavens presents, with
sensuous vividness, to the souls is a direct vision of how the
threads of Destiny are spun; cosmic justice is sustained because
they are spun as they are. True, in the order of his description,
Plato indicates certain mechanical properties of this spindle
which give a necessitarian aspect to its mode of spinning, but
although these properties are purely ordinal and apparent on
inspection, the soul coming to choose another life is not re-
quired to analyze this mechanism to gain the intended insight
which, if "God is blameless," is necessary for its choice. What
is essential is the sight of the twisted continuity of these threads
between the. eternal present and the receding future, and the
role of the past in this continuity. The "Fates, robed in white"
in Plato's description, are the visual center of attention in the
panorama he is describing. It is their songs and their spinning,
with the Sirens' "music of the spheres" for contrapuntal back-
ground, that activate and explain the revolutions of the "Spin-
dle of Necessity." The colors of the hemispheres are similarly
chosen because of their vividness to the spectator of the tableau
which is the subject of Er's report. The suggestion of a sys..
tematic unity in diversity of concentric cosmic process is made
poetic by the vision of the hannonious motion of the spangled,
red, white, and yellow hemispheres.
180 Plato's Mathematical Imagination

5. Why the souls see colored rims instead of separate planets.

The reason for assigning the colors of the celestial bodies to
their zones in the machine has been explained by citing the
Greek belief that the planet is attached to and carried by the
motion of its orbit. How this relative activity manages to trans-
fer from the passive planet to the active orbit the color of the
former is not nlade clear, unless we are intended to assume that
the planet, being lazy, does not deserve such ornament.
However, the tactics of explaining this device by statements
less susceptible to interpretation than the original can be set
aside if we bear in mind what the effect of a location on the
eminence of being has on the perspective of the spectator of
the physical world moving below. The philosopher is not made
dizzy by the high altitude from which he is a spectator of all
time and all existence; his broadened concepts of space and time
are far removed from those of common men, impressed by large
estates or ancient lineage; for the size and length of these shrink
to points against the standard of the ideas of all space and all
time. The souls in this myth have ascended to this vantage
point, but, lacking the disciplined intellectual insight of the
philosopher, are represented as perceiving only those visible and
tangible features of the universe which the senses can convey.
The important philosophic concepts which inhabit this em-
inence reveal themselves in visible personified form when aware-
ness of their presence is required.
The natural effect of this new vision is a foreshortening of
time and space to a point at which even the longest celestial
revolutions can be intuitively perceived and need not be labori-
ously calculated with the help of many models, as they must be
inferentially established by the terrestrial student of descriptive
astronomy. The capacity to "see together" many distant things
in space is matched, as it is said to be elsewhere in the descrip-
tion of philosophic insight, by a "seeing in one view" of many
things far separate in time. To make the physics and astronomy
of the story sufficiently plausible to gain even poetic credibility,
we may have to assume that this new vision is actually presented
by the Fates as a model of what the souls would see from their
r GEOMETRIC METAPHOR 181

vantage point, had sight the -power of intelligence to pierce

through the outer spheres to the hidden workings of the inner
ones. This only underscores the fact that the inspection of the
world-machine is one which vastly shrinks its spatial scale and
one in which the temporal scale is correspondingly diminished.
If the ratios of speed were preserved and if the scale were set so
that the ring of Saturn revolved once in a minute, the effect of
such temporal shrinkage would require that the fixed stars travel
at about 10,000 r.p.m. But, under the impact of its propulsion,
the foreshortening of time in this model is evidently carried
much further. The blurring of individual planets into what
seem to be revolving concentric rings is the natural and ap-
parent effect of the change in temporal perspective which makes
possible any direct perception of their relative motion. All that
it signifies here is that the souls are seeing a vastly speeded-up
model of celestial process, not that there has been any inex-
plicable transfer of color to orbits from planets, nor, necessarily,
that we have a remnant of archaic Pythagorean astronomy.
The extraordinary foreshortening of time-periods in the souls'
inter-incarnation careers seems to corroborate this interpreta-
tion. The myth opens with pleasures and pains awarded in
periods of a thousand years; it closes with a four-day journey
and a one-day period of education and choice. The number
10,000 is used in myths to indicate an indefinitely large but
denumerable set or period. Consequently, no two periods Plato
could have assigned would have been more neatly calculated to
set up a contrast between the enduring and the ephemeral.
(This shift is paralleled in the change of the earlier description
of human life as "a period of one hundred yea~s" to the proph-
et's opening words, "souls that live but for a day.") The under-
scored shortening of time periods becomes functional if it is
taken as paralleling the more expansive vision of time and the
consequent foreshortening of time scale which the souls attain
as they move to a "higher" place. It may be assumed that Er,
as special observer and messenger, was in an exceptional posi-
tion ,vith respect to the correspondence between perceived and
absolute time. If a year or a hundred years seem merely part of
182 Plato's Mathematical Imagination

a single day to the soul, we may legitimately allow Er to view

events covering several centuries and return to earth in twelve
days of elapsed terrestrial time. He must see the whole process,
yet he must return to his body with his report. His return takes
place in a definite period of time, the length of which suggests
that time is somehow important; but, in the absence of any
clear anticipations of modern physics, we are not told how the
gods who have selected him as messenger accomplish this rela-
tivity of time. .
Earlier in the dialogue, when his discussion made this becom-
ing-to-being transition, Socrates himself was the subject of an
exactly similar time-sense shift, which was indicated in his state-
ment that "the time spent in such discussion seems but a mo-
ment compared to the cycle of the soul's transmigration." 104
His contemplation of the forms basic to the discussion of edu-
cation has so colored his orientation to time that the remark of
a member of his audience, who applies a normal time scale,
seems to be inappropriate and calls for explicit correction.
6. Allegorical function of the model. The concept of equilib-
rium serves in this account as the observable projection of the
principle of perfection; the balancing of parts with one another
is the sensible evidence of the presence of an organizing plan.
Equilibrium is a property of mass, not of volume or color; but
the balancing of the lists of colors and volumes seems to suggest
that these are the properties to which the concept is here ap-
plied. Mass is itself, however, related to volume and unit density,
and unit density can be inferred from color. The lists of colors
and volumes are combined to form a balanced list of masses;
they are presented separately because only one of the factors-
that of density-can have its balanced distribution verified by
empirical observation, and the volumes must be inferred from
a postulate of balanced masses to which this observable arrange-
ment of colors gives some empirical confirmation. It is the bal-
ance of this mechanism which is evidence that the parts have
been adjusted to one another by some plan.
A further function of this description is the insight it gives
into the apparent motion of the model. The impulsions of the
 GEOMETRIC METAPHOR 188

goddesses of future and past are both distributed with complete

uniformity through the model, in such a way that some irregu-
larity is created at the point where the transmitted forces are
equa1. 105 The circle at this balance point must seem, from the
mechanics of the model, to have a peculiarly erratic behavior.
As if endowed with a free will, it moves against the dominant
direction of the total system only to be brought back by the
dominating motion. These excursions, while reflecting an ir-
regularity in impulsion, are not sufficiently forceful to alter the
entire dynamics of the system. This behavior will be interpreted
as visual confirmation of the presence of freedom of will at the
moment of confluence between future and past. In this way a
present choice is not necessarily in accordance with the natural
order, but it cannot succeed in altering nature to its own plan.
The vision of destiny turning the cosmic spindle is thus, as
we should expect from its tableau form, an allegorical repre-
sentation of choices in time by which man creates his own
destiny.
7. Colors of the hemispheres. The intrusive, detailed descrip-
tion of the colors of the rims of hemispheres of the world-
machine in this myth has remained one of the least satisfactorily
explained passages in the Republic. Unlike the notorious nup-
tial number, of which each generation of scholars produces a
new interpretation, the function of this list of colors has never
been given any plausible explanation. A. E. Taylor wrote that
while he was sure the list represented some empirical counter-
part, and had some principle of order, he could not even
imagine what that counterpart and principle might be. lo6 Be-
cause of his erudition in classical and modern Platonic scholar-
ship Taylor's statement lends credence to the claim that no
plausible function has ever been suggested.
The closest approach to such an interpretation was the dis-
covery, late in the nineteenth century, that a principle of bal-
ance had been observed in the order of the hemispheres for
which colors are mentioned, so that symmetrical pairs add up to
a total ordinal value of nine. lo7 But what conceivable meaning
one is to give this meticulous "balancing" of color is left an
184 Plato's Mathematical Imagination

open question at least as difficult to resolve as the original

problem. los
Yet there is an interpretation which seems to explain both
the presence and the order of this list of colors and to render
the passage functional in context. The key to this interpretation
lies in the conviction, which grows with study of Plato's mathe-
matical imagery, that he is too precise a craftsman ever to have
intruded a clumsy metaphor treating balance as a property of
color. lo9 Balance is, as the Timaeus makes clear,llo a property of
weight, not of volume or of color. The purpose of the present
section is to present an original interpretation of the principle
of order operative in the listing of colors which (a) establishes
an empirical reference, (b) avoids attribution of this clumsiness
of metaphor construction to Plato, and (c) gives the list a
function which helps justify its sudden, emphatic intrusion in
the myth.
First, we should note that these colors are properties of parts
of a model of the world-machine, displayed by the goddess
Necessity. This goddess seems to be a personification of the
concept which, in a "necessary" account of cosmology, repre-
sents an analysis of the world under the aspect of mechanism
rather than under the aspect of organism. lll In the natural
order, organic functioning as a whole is constructed and medi-
ated by mechanical interaction between separate parts. It is
completely in keeping with this identification of the goddess
with the concept she personifies as well as with the archaic tone
proper to a religious myth, that the cosmic model shown by her
to the souls should be an archaic Ionian or Ionian-Pythagorean
device. ll2 It stresses celestial mechanics rather than the pure
mathematics of astronomical motion typical of the research in
the Academy, and more appropriate to a description of the
world qua organism. lls
In the necessary account in the Timaeus J the key concept to
the behavior of the machine as a whole is that of equilibrium,!1'
a property determined by the relative shapes and volumes of
parts, which we observe empirically on a macroscopic scale as
differences in hardness and weight. ll5 The discovery of a meta-
 GEOMETRIC METAPHOR. 185

phor of balance in the colors and volumes of Plato's lists strongly

suggests that their author intended to underscore the presence
in his model of some sort of equilibrium proper to the presence
of perfection in a world-machil1e.
If this is true, however, it is not enough for both the volumes
and colors of the hemispheres to be balanced; the point in-
tended must be that the masses are balanced. 116 Whether this
is in fact the principle of construction underlying the lists of
sizes and colors can be investigated from two directions. The
first is the consideration of the relation of color and density
("hardness") in the cnemistry of the Timaeus) as compared to
the observed colors in the present list. As we shall find, this
comparison, while it does not contradict the hypothesis that
apparent color is associated with relative density, is not exten-
sive enough to confirm it definitely. If the list of sizes represents
volume, that of colors in some way represents density, and if a
balance of masses underlies the construction of the two lists,
then a second tactic of confirmation is to construct and examine
the new ordinal list which results if we add or multiply together
the relative sizes and colors of each hemisphere. 117 The result,
which is given in figures 66-73, seems conclusive confirmation.
The final list not only shows a balancing of symmetrically paired
terms which add up to nine but represents the property in a
clearer, more evident form than either of the two sets from
which it is constructed. The probability that such a result would
come about accidentally, if the two lists so combined had not
been planned with just this result in view, can be calculated,
and is so slight that this explanation of the balancing of the
list of sizes and colors may be dismissed as a practical impossi-
bility.lls
Plato himself invites the reader to make this association of
color and physical composition by his initial remark about the
composition of the hemispheres, that "the others [i.e., those
other than the outermost circle] had varying amounts of adamant
in their composition." 119 He then presents a listing of colors
ordered, as we can demonstrate from the Timaeus color the-
ory,120 in terms of the intensitYJ not the hue, of their colors.
186 Plato's Mathematical Imagination

The color spectrum of the Timaeus theory takes as primary

colors black and white,t21 then differentiates (still as primary)
brilliant,122 white, and red,128 all acting mechanically on the
visual stream like white light, but shading in intensity from a
maximum to a minimum. The colors of the hemispheres, except
the outermost, which has special properties, are listed beginning
with brilliant and following with yellow (a composite color
intermediate in intensity between its constituents, brilliant and
white),124 a pure white, red, and a secondary whiteness (which
would be a composite color resulting from the mixture of white
and black, a grey lower in intensity than red).12G
If "the presence of more and less adamant" is revealed by a
greater and lesser intensity of color, the principle of order can
be functionally explained.
So far as it goes, the chemical theory of the Timaeus does not
contradict this association. Adamant and gold, the constituents
of the outermost hemisphere, as described in the Timaeus) are
the most dense of materials; 126 bronze, having a yellow color,
is less dense than gold; brilliance and reflecting power are the
properties of "transparent stone," which, having the most
homogeneous elements of all earth compounds, is also the most
dense, though the relative density of such stone and bronze
is not given. The other colors do not appear as identifying prop-
erties of basic types of compound except for the mention of the
color white as associated with various saps and juices.
It is not hard to see why Plato complicates his final exposition
by combining two separate lists of properties. Of the two proper-
ties relevant to a mechanical equilibrium of the world-machine,
the relative volumes of zones admit of no empirical confirma-
tion. 127 Through the posited association of density and intensity
of apparent color, however, a postulated principle of order in
the distribution of celestial masses can be tested and confirmed
empirically. The empirical fact, that an order is operative in the
relation of colors and distances of the planets, is in turn justifica-
tion for constructing a list of volumes, on the postulate that
these also are distributed according to some order, presumably
an order showing the principle of mechanical equilibrium in
operation.
 GEOMETRIC METAPHOR 187

The justice of the world, which the souls are to learn from
their inter-incarnation education, is thus at least in part evident
to every man who takes the trouble to inspect the appearance
of the heavens. Furthermore, the very colors of the planets
constitute an empirical argument for the justice of the world
order.
The principal objection that seems likely to be made to this
interpretation is that it puts an undue burden of interpretation
on the reader. Although it is extraordinarily unlikely that the
balance of the composite list is merely a result of chance, it
may be objected that this improbability is balanced by an equal
unlikelihood that Plato would expect such elaborate mind read-
ing from his intended audience. In reply, I can suggest only
that this intended audience would naturally be expected to look
for mechanical properties in an Ionian astronomical model,
and might quite reasonably be expected to take the remark
about relative proportions of adamant as a hint as to where
these properties were to be sought. Further, the myth is equally
valid and effective for the reader if he does not try to interpret
these lists of properties at all, but reads them simply as detailed
evidence that some rationale is at work in the cosmic order, a
rationale made apparent by the mechanics of this order. As a
teller of myth, Plato here needed only to suggest aesthetically
(and detailed listings are such a device of aesthetic suggestion)
the evidence of a principle at work in the mechanical details
of the cosmic order. However, as a conscientious inventor of
mythology who believed that the central points stressed in the
myth were true, Plato could not honestly have gained his
aesthetic effect by inventing a set of descriptive statements about
the observable mechanical properties of the world to which the
facts did not correspond. His aesthetic device of listings there-
fore embodies in it, for the reader not willing to take the myth
on the level of belief, but insistent on testing it by intellectual
inquiry, a basis of empirical fact to which the aesthetic proper-
_lies can actually be shown to correspond.
L-/~' 8. Motion of the hemispheres. Following his account of the
structural details of the model, Plato presents a list of the prop-
erties of its dynamic behavior. This list displays the same
188 Plato's Mathematical Imagination

principle of balance in the velocities of the hemispheres that

was the clue to the lists of their sizes and colors; 128 and if we
interpret the former lists as determining mass, as was suggested
above, this balance of velocities will also establish a balance of
momentum.
There is a contradiction in engineering between the two stages
of structural and dynamic description which must be recognized,
since it prevents the full deduction of the model's dynamic
from its structural properties. (Theon of Smyrna actually at-
tempted to make this deduction, with the aid of a reconstruction
of the model here described. 129) Though the operation of the
principle of perfection is represented in the same way, as an
equilibrium of the ordinal numbers used in the two descrip-
tions, the structural properties do not lend themselves to mo-
tions of the sort described. The effect of the relative velocities
involves a diffusion of retrograde motion from the innermost
circle, where the retrograde impulsion is applied, to the seventh
circle; and the strength of this motion decreases in proportion
to the distance from this central source. 130 This type of fluid
transmission of motion from a center of disturbance was familiar
to the Greek scientists, and most often illustrated by the be-
havior of eddies in water. It is a basic mechanism of the vortex
theories with which the formation of the cosmos was explained
by the Ionian and later Atomist philosophers. l s1 But if the
hemispheres are conceived as being rigidly defined by the iron
bands of Justice, which hold each to its proper place, this idea
of fluid transmission will not work. Consequently, a reader
anxious to construct a model of this machine would actually
have to build two models, one showing the structure of the
machine at rest, the other duplicating its apparent motion. The
latter would involve some transmission of impulsion through a
fluid medium in such a way that momentum varied with dis-
tance, and the forward and retrograde components gave result-
ant motions to the circles of the sort Plato here describes. 1s2
Although this defect in engineering might seriously disturb
an Ionian physicist, it constitutes no obstacle to the reader, even
to one who interprets the description in detail. The dynamic
 GEOMETRIC METAPHOR 189

properties of the model are taken as those given in the list;

one can easily visualize them. Having visualized, one can look
for the operation of a principle of perfection or order in the
interrelations of the motions, and even construct a table of the
momentum which hemispheres of the sort described would
have if they were moving in the system postulated by the dy-
namics section of the description. There is nothing difficult
about combining a visualization of these sizes, colors, and rela-
tive speeds in an imaginative picture, unless one has a singularly
limited and literal imagination. We shall see that the detail of
these dynamic properties and their relation to the structural
account still provide the sense and order that the myth requires,
even though some supernatural power is needed to make the
structure behave as described.
The reason for this difficulty is of course an interference be-
tween the facts of mechanics and the allegorical and poetic
demands of aesthetic suitability. In fact, though not in poetry,
the celestial spheres are not held to their courses by iron bonds;
the adamant walls bring out admirably the presence of cosmic
justice, but at the expense of requiring a divinity to make the
walled-in spheres responsive to impulsion in the fluid manner
that an empirical, atomistic account of transmission of motion
describes.133
In the motion of this model the momenta and velocities are
both balanced into three zones: one predominantly impelled by
the forward motion, one by the retrograde, the third (twice in
mass to either of the others) intermediately moved by both.
The central axis of all of these symmetrically balanced lists
falls at the fourth circle, midway between the two points of
forward and retrograde impulsion. The reader with wax tablets,
having jotted down and noted the balance of these descriptive
details, would naturally have seen that circle as occupying the
center of the stage, and wondered whether something dramatic
in its behavior might have led to its prominence in the imagery.
In fact, of course, Er singles out for special mention an anach-
ronism in the motion of this circle. The fourth circle moved
with "apparent counter-rotations" or "turnings back. The mech-
It
190 Plato's Mathematical Imagination

anism of balance, however, gives a physical explanation for

this aberrant behavior. Since this circle is at the exact point of
balance of the forward and retrograde momenta, any change in
impulsion will be most evident in its behavior; and the impul-
sions of the model are constantly changing and mechanically
erratic. A backward jump of Mars follows as an effect of a rein-
forced impulsion of the inner circle. This evidence, which Er
has previously mentioned, that the Fates turn erratically can,
however, by no means exhaust the significance of the retrograda-
tion if the phenomenon is important enough to deserve the
prominence it seems to be given. There should also be, since
this is a true myth, an identifiable empirical counterpart, and,
since the myth is allegorical and since later context requires it,
an identifiable applicability to the phenomenon of human
choice, either in this life or in our next selection of incarnation.
The empirical counterpart is, of course, the observed retro-
gradation of the planet Mars, which the Timaeus describes. 184
Cornford's commentary on that passage seems to establish def-
initely that this is what it is describing, and also that Plato
explains this apparently aberrant behavior as an exercise by
the planet of its tree-will. lslS This aspect of the empirical refer-
ent of the present passage seems, when we bear in mind that
the passage in the Timaeus centers in the problem of the rela-
tion in human choice of freedom and necessity, to be the key to
the allegorical interpretation which Plato may have had in mind
when he put the phenomenon into the foreground in his myth.
Before attempting an application of these mythical mechanics
to the central contextual problem, it will be necessary to con-
sider the model in terms of another aspect. The machine that the
souls are shown is repeatedly qualified in terms that are in-
tended to underscore the fact that a model of celestial motions
is also a model of the nature of time. More than tradition or
caprice underlies the allocation of past, present, and future
to the separate jurisdictions of the Daughters of Necessity. The
themes of their songs,136 and the impulses they give the model,
alike invite some association of cosmic mechanics, freedom, and
temporal flow from the eternal present to the indefinite future.
 GEOMETRIC METAPHOR 191

The song of the goddess of the present is about that which is;
the motion she imparts, with her right hand, is the "motion
according to the same" of the Timaeus account; 187 this motion
flows, through a zone of momentum primarily under its domina-
tion, to an intermediate lone, balanced between progression
and retrogradation, into the contrarily dominated inner zone of
the machine, the locus of the indefinite "that which is about to
be," as distinct from the locus of the perfectly determinate "that
which is." 188 In this passage from same to other, the intermedi-
ate zones are balanced much as a moment of present choice
exists suspended between present and future. The direction
taken by the intermediate zones varies with the differential activ-
ity of the goddess of the past, who sometimes reinforces the one
motion, sometimes the other. 189
Considered as a time-machine, the phenomenon of balance
takes on new significance in this context when it is equated with
the similarly balanced moment of choice which each man con-
fronts many times in his historical career, and is shown re-
enacting in the present mythical drama. In the Theaetetus,
human choice is explained by the statement that "each man has
many patterns of the sort of life he would like to lead. Of these,
he selects one; and his destiny is actually to become the type
he has selected." 140 In Republic ix, we learn that a good man
will contemplate the pattern of the citizen of the best state, a
pattern to be found only in heaven, and will live as a citizen
of that state, and no other. 141 We also learn, earlier in the Re-
public, that past choices have a predisposition to repeat; a man
who imitates something grows into a likeness of the thing imi-
tated. 142 If we now interpret the influence of the motion of the
same as representing the influence in any present moment of the
forms, always existing and knowable in an eternal present,
the motion of the other as a projection of these forms into a tem-
poral medium (where they are subject to distortion and counter-
action), and the motion of the past as a persistence of habits
influencing choice, the time-machine seems very like a model
of the human will.
If we accept this interpretation of the balanced zones of
192 Plato's Mathematical Imagination

present and future, with a moment of choice between, into

which both are alike ingressive (a double ingression which, of
course, lends some point to making this intermediate zone ex-
actly twice the momentum of either of the others), the impor-
tance of Mars is readily explained. The aberrant behavior of
this planet provides direct empirical evidence that freedom is
possible within the mechanism of the world-order. The Fates
adjust the mechanism of their machine in such a way that its
behavior allows expression to a free choice which runs counter
to the dominant momentum of the natural order. There is no
absolute determinism, because the rigid mechanism adapts itself
to implement decisions, just as the parts of an organism adjust
mechanically to new conditions which confront that organism
as a whole. Once the adjustment is made, it is irrevocable; its
mechanical consequences cannot be evaded or recalled.
A close observer of the model would also find that although
nothing prevents Mars from acting on its "desire" to move con-
trary to the natural course of things, it cannot carry out that
countermovement long. Its own momentum is so much less than
that of the whole set of turning hemispheres that its contrary
career is soon overborne and carried along by the inexorable
momentum of the total order to which its own desire was op-
posed. In the same spirit, Plato later argues that the temporary
success of injustice is no warrant for disbelieving in the over-all
perfection of 'God's management of the world; the laws of
nature, are arranged, like the rules of chess, to move each man
automatically to the status he deserves. 148 In Republic ix, the
same thing is suggested. Plato says even that if we consider men's
lives as wholes, the span of life is often long enough to show
this reward and punishment in operation. 144
Having seen this model, the soul who has learned from it
will be guided in his choice of a pattern of life by the reflection
that such a pattern should resemble the world-order in which
and with the concurrence of which the chosen role will be
enacted.
In every moment of choice, the chooser is balanced between
the eternal forms, which suggest a right standard, and the im-
 GEOMETRIC METAPHOR 19~

pulse of his past interests, which establish a certain inertia and

impel him to remain in the same course he has been following.
The future lies ahead as the locus of actualization of the career
chosen by him as a present possible pattern; whether this actual-
ization will be good or bad depends on the adequacy of the
subjective role chosen to the proper role of man, laid up as an
eternal pattern which will be furthered by nature.
On this interpretation, the lessons of human responsibility
and human freedom are presented to every soul when it is
shown the innermost workings of the world-machine; and the
choice of each after this instruction is clearly made on its own
responsibility.
This interpretation, in addition to the fact that it explains the
functionality of the elaborate details of the mythical machinery,
can be partially confirmed by a comparison with the earlier
mathematical imagery of the dialogue. If the myth sums up the
entire sweep of the dialectic, one might hope that the mathe-
matical image central to that myth would also in some way
synthesize and sum up the entire range of the antecedent mathe-
nlatical images. In this connection, we may note the recurrence
of the concept of impulsion and momentum associated in Book
iv with the wheellike growth of the state in an inevitable up-
ward spiral of progress. 14lS It has already been suggested that
this image and the downward spiral of Book viii are combined
and balanced in the present imagery of cycle. In a later note, it
will be shown that there is also a synthetic inclusion of the
tyrant's number computation in the emphasis placed in the
present image on the number nine. It has already been remarked
that the initial image of harmony is here personified by the
singing Sirens whom the souls see and that the length of the
soul's journey takes on new significance if read as a cross refer-
ence to the four stages in the image of the divided line.
This remarkable synthesis of diverse imagery, which one
could not find duplicated in any writer for whom the construc-
tion of such images did not hold aesthetic interest and value, is
also present in another Platonic passage, the construction of the
world-soul, to be interpreted in a later section. The synthesis
194 Plato's Mathematical Imagination

of antecedent images in the arithmetical and geometric cross

references of his final model helps to validate the notion pre-
sented earlier that the myth as a whole, read as allegory, per-
forms the same function for the antecedent dialectical dis-
cussion.
9. Role of the number nine. Another fact apparent to the
reader with a wax tablet filled with diagrams, but not to the
simple observer of the model, is the intrusiveness of the number
nine. Although this is the only choice by which one can establish
an arithmetical metaphor of balance based upon the ordinal
numbers from one to eight, numbers other than one to eight
might be chosen if the number nine were not considered to be
appropriate in context. The balanced character of the mecha-
nism could certainly be brought out as well by an adumbrated
scheme of ratios as by the ordered list; and such an adumbration
would not be out of place in the context of the Sirens, whose
presence reminds us of the fact that there are internal relations
in cosmic connection. Further, the use of nine as a central num-
ber breaks Plato's usual pattern of mythical arithmetical anal-
ogy; his numbers in myths and semimythical histories are else-
where presented as tens and powers of ten, presumably as a
joint effect of Pythagorean practice and the presence of "myriad"
as the largest common number-word in the language. 146 The
periods and incidental numbers, such as that of the lot of Ajax
in the present myth, preserve this preference for multiples and
powers of ten; one may therefore wonder whether there is not
some reason why the details of the mythical model do not.
If the mathematical illustrations of the Republic are intercon..
nected (and by a sort of poetic economy each is built from the
elements of its predecessors), then we may look for an explana-
tion of this use of nine in what would normally be a wholly
decimal context by considering the functions of nine in the ante-
cedent mathematical imagery. We are certainly invited to iden-
tify this number with the nine-celled matrix of human careers
and characters of the tyrant's number and of the nuptial num-
ber by substituting nine, in both of those images, as an arith-
 GEOMETRIC METAPHOR 195

metical analogue for the nine-part matrix figure whenever a

computational metaphor is used to summarize the structure of
the diagrams. The representation of each of the three dynamic
components of the time-mechanism of the universe as equivalent
to the number nine may have been felt to be appropriate be-
cause of its connection with the intended moral interpretation
of the model's behavior. The symbolic number remains the
same for past and future, and in the present is doubled through
the copresence of a given present nature and a desired future
goal. If this significant relevance to antecedent imagery is actu-
ally the explanation of the retention in a mythical context of
this particular number, representing moral balance, its signif-
icance is clearly the invariance of human character through
time. New types of character, and new ideals, are not operative
in human choice; rather, the same basic nature and super-
imposed edllcation are to be found in the present and in the
pattern which represents ingredience in the present of an ideal
future. Furthermore, the invariance extends to the relation of
character as reflected in moral choices. The use of the moral
number as a symbol of cosmic balance further underscores the
continuity, in human life, of future, present, and past. The
irreversibility of the machinery of fate insures that each choice,
helping as it does to establish the nature of the chooser, will
be rewarded with an appropriate destiny, even though there
may seem to be in some given case a temporary retrogradation
of cosmic justice, which in human life permits the apparent
securing of desired ends by means which are unnatural and in-
appropriate.
Thus from the model he is shown the soul learns the same law
that the prophet proclaims to the assembly: "The choice is
yours; God is blameless." 14'1 p'ermanent and real, as opposed to
transitory and apparent, freedom resides only in the choice in
which lessons of past experience motivate the individual to
choose on the basis of those eternal truths.
196 Plato's Mathematical Imagination

d. Text of the Passage: =* Theon's Version

The rim of the largest whorl was spangled; the seventh brightest;
the eighth colored by the reflected light of the seventh; the sec-
ond and fifth like each other and yellower; the fourth was red;
the sixth second in whiteness. The Spindle rolled forward as a
whole with one motion, that of the [daily revolution of] the cos..
mos; but, within the whole as it turned, the seven inner circles
were led around slowly in the opposite direction; and of these the
eighth itself moved most swiftly; second and all moving together-
with equal speed} the seventh, sixth, and fifth; next in speed
moved the fourth with that they said appeared to be a retrograda-
tion, more marked than that of the others; next the third, and
slowest of all the second.

Text t of Theon's Version, with Theon's Variants

KUALEoi}UL ~E Ot'QE<POIlEVOV 'tOY al'Qu%l'OV OAOV IlEV l'~V aircl}v <poQav
'tq> I)(oolliil, EV ~E OAq> 1tEQUPEQOIlEVCP 'tov~ Evro~ E1t't<l X{mAO~ -r~v
EvavtLuv t'eP OAq> f}QE~a JtEQLdYEo'3at, umwv ~E l'oln:oov 'tUXLO''tU
llEV IEvClL 'tov oy~oov, ~EUt'EQO'U~ ~E xat alA-a dAAl1AOL~ ioot'axw~ 'tov
l'E E~~OIlOV xat LOv EXtOV xal t'ov JtEll3tTOV" 'tQLtOY ~E cpoQQ. iEvat,
8v <paOL <paLVEO{}Ut, E3tUVaXUXAouIlEVOV t.ulALO't(l t'wv fi.AArov· TOV
l'EtaQ'tov, TEl'aQTov ~E tQLTOV· xal 3tEJJ1tfOV TOV ~EUl'EQOV.

Theon's Variants-
KuxAeiaitaL, MSS KVAL£uDuL, Theon
cpooo.v, MSS' cpoQo:v 'teP xOO·J.Lql, Theon
JtEQt<pEoeoit'at, MSS 3tEQLayeo'ftaL, Theon
o.AA~AOL~, MSS dAA ~AOt; too-raxro;, Theon
£x'tOV xm. 3t8J.LTC"tOV, MSS 1:Ov EX'tCW xat 'tOV Jt8~O'V, Theon
1:0'V 'tQL'tO'V, MSS "tQL'tOV, Theon
oo~ o<p'£m, MSS O'V cpam, Theon
€3t(lV(l?tU~AoUJl£VO'V, MSS e1taVaxtrXAOUJ.I.£VOV JUIAlo'tex
-.:ro'V dAArov, Theon
1:0'V 'toi1:O'V, MSS "t'QL'tO'V, Theon

Though a copyist or scholar might hesitate to tamper with

• Republic 616E, trans. Cornford, Republic, p. 354. Theon's variants have been
supplied in italics. See also the notes to the Greek text, following.
t Republic 616E-617C, Chambray, Republique.
 GEOMETRIC METAPHOR 197

the text of a passage he was sure he had not understood, tIle

empirical referents of the present passage must have presented
constant temptation to introduce adjustments bringing it into
better conformity with later commonplaces of empirical as-
tronomy. That these adjustments did take place is clear from
Proclus' "older and better" text of the sizes, already given. 148
The reading is not older, because it disagrees with Theon's text;
it is not better, except that a Hellenistic commentator can inter-
pret it more easily. What took place here was an entire re-
,vriting of the list of sizes, apparently on the principle that A. E.
Taylor has pointed out, of identifying distance with apparent
luminosity. This list is given in Figure 68, following.
One may therefore wonder whether the differences between
Theon's text and the manuscript tradition reflect any such in-
truded accommodations on either side. The variants are given
in the text and translation just preceding.
Of these, Burnet proposed to accept the qualification of the
retrogradation of Mars as "greater than that of the others"
since (1) this is needed to secure accurate correspondence with
the account of apparent planetary motions in the Timaeus, and
(2) the length of the phrase, 15 letters, is exactly what the length
of a line in the older manuscript of Plato must have been, sug-
gesting that Theon retained a line lost in later copying. 149
Burnet further suggested that since Theon says he has used the
commentary of Dercyllides, he presumably had the latter's text
at hand. It is likely that in quoting this passage, on which he
spent so much scholarly effort, Theon would have quoted more
meticulously than he does elsewhere.
The difference in the descriptions of the motion of the whole
could easily result from a transposition in copying, and since
Theon's term carries with it a notion of "translation" (linear
motion through space) as well as "rotation," it is inappropriate
here. There does not seem to be any ground for choice between
the other variant pair of verbs describing revolution. In this
context, such a verb should be capable of taking on a connota-
tion of some external impulsion or mechanical constraint, as
either of these can.
198 Plato's Mathematical Imagination

At b2,· editors have all recognized the superiority of Theon's

text to the manuscript version. At bi) Theon's addition of two
definite articles to differentiate the ordinal designations of the
hemispheres from those of their velocities may not be essential,
but seems preferable. At bJ) Theon's omission of the definite
article gives less clarity than the manuscript text, and seems
inferior.
Theon's addition of "the cosmos" to the reference to the
motion of the whole is certainly not an improvement if any
separation of cosmos and model is to be preserved in the pas-
sage, and the empirical referent is clear without it. Nor, unless
it implies some alternative type of measure, does the addition
of "equally swift" in Theon's text add anything of importance.
If "velocity" here refers to apparent celestial motion, "equal ve-
locities" of different circles would seem to mean "equal periods
of revolution," which of course implies different linear veloci-
ties. If the ordering principle is linear retrograde velocity, to
distinguish the 7th, 6th, and 5th circles in ordinal listing and to
carry out the image of "music of the spheres," the qualification
that 7, 6, and 5 are "equal in velocity as measured by period"
ought to be more explicit.
These problems must be left to future editors of the Repub-
lic, but some of Theon's variants should be given careful con-
sideration.
Figure 66
LAW OF NINES: SIZES
Order in Series 1 2 g 4 5 678
Size of Rim 18736 254
l--J I L ----1 I L-.J
(19) (17)

J. Cook Wilson's article, "Plato, Republic 616E," Classical Review}

XVI (1902), 292-3, is the basis of James Adam's interpretation in
The Republic of Plato, II, pp. 470-79, in which Adam reverses his
• This symbol and the two following (bI and bJ) are abbreviated references
to Stephanus pages; fuller designations would be 617B.2 617B.l, and 617B.J, re.
1

ferring to b2, bl, and bJ, respectively.

 GEOMETRIC METAPHOR 199

own earlier defense of Proclus' alternative text (Adam, "On Plato,

Republic 616E," Classical Review, XV [1901], 391-93). Cook Wil-
son's figures ("If the order in which breadths of rim and colors are
listed is set down for each hemisphere, the following figures result,"
loe. cit.) are reproduced as figures 66, 67, and 71.

Figure 67
LAW OF NINES: COLORS·
Order of whorls 2 3 4 5 6 7 8
Order in which their 4 6 7 5 8 2 3
colors are described
I ~--II-d= I
(18) (18)
• See text comment of Figure 66 for titation.

This is not a very clear-cut instance of balance, and, in its con-

text of balanced lists, suggests that perhaps some error has been
made in the reconstruction of the figure, hiding a more spectacular
pattern of balance. In the list of sizes the order of hemispheres gave
the balanced figure, but the order of mention was the same as the
order of whorls. Since this was not the case in the list of color, one
might try reconstructing this diagram, taking order of listing rather
than order of cosmic location as the ordering principle. The result
of this revision is shown in the following figure.
Figure 68
LAW OF NINES: COLORS (ALTERNATIVE VERSION)
No. I 284 567 8
Color I 782 5 S 4 6

I f--J --'-f-J
(18) (18)
This is simply a different form of Cook Wilson's diagram, with
the rows transposed; this form will be used in the discussion that
follows since it permits certain significant relationships of the two
listings to be more clearly indicated than in the original represen-
tation. It is preferable also because it presents both lists in the same
matrix form of:
200 Plato's Mathematical Imagination
R.OW 1 : ORDINAL PRINCIPLE 1 2 8 . . . .
R.OW 2 : COSMIC LOCATION

For any comparison of the listings, it is of course essential that

they be constructed in the same form.
Figure 69
SIZE AND COLOR (VOLUME AND DENSITY)
Width 1 8 7 3 6 2 .5 4
Color 1 7 8 2 5 3 4 6

Sums 2 15 15 5 11 5 9 10

Adding the two orders shown to balance in figures 66 and 68

gives this combination of volume and color. The result of trans-
lating these sums back into an ordinal list, from 1 to 8, is shown
in the following figure.
Figure 70
SIZE AND COLOR (ORDINAL): ANOTHER LAW OF NINES

Order in series 2 3 4 5 6 7 8
Rim and color 8 1 2 6 3 7 5 4

LJ LJ
This figure results from assigning ordinal numbers to these sums,
exhibited in Figure 69, in constructing a list of relative momen-
tum (observing, as is done in the color list, the convention that
two equal sums are numbered with the greater to the left).
This discovery might have been made by Adam and Cook Wilson if
they had realized that their ways of constructing the lists of size
and color were disparate and had reconstructed the color diagram
in its proper form.
Note that Figure 71 divides the dynamic behavior of the mech-
anism into three zones of momentum, one dominantly forward, one
retrograde, and a third (double in size) balanced between the other
two. The starred numbers represent hemispheres with the same
velocity; the balance is equally well demonstrated if we assign the
average ordinal velocity (6) to each, instead of following the right-
left convention in giving them successive ordinal numbers.
 GEOMETRIC METAPHOR. 201
Figure 71
LAW OF NINES: BALANCE OF VELOCITIES t

Hemisphere 2 3 4 5 6 7 8
Relative Speed

L.J
8 7· 6· 5- 4 3
,
2

(9) (18) (9)

t See text conlment of Figure 66 for citation.

Unless Plato's constructive ingenuity gave out at this point, one

would expect to find some balanced figure representing the com-
bination of balanced velocity and balanced mass; evidently certain
details of the myth call for a recognition of some sort of balance
of momentum. (See Fig. 73.)

Figure 72
PROPORTIONALITY OF DISTANCE AND FORWARD VELOCITY
Order of distance 8 7 6 5 4 3 2 1
Order of velocity I 2 g 4 5 6· 7 8
Sums 9 9 9 9 9 9 9 9

The passage in Laws x on the reasonableness with which radius

and velocity are kept proportionate by nature in revolving planes
or solids suggests that this sort of relation of relative distance and
speed may have been the root. or starting-point of the various
demonstrations of "arithmetical balance" of ordinal numbers in
the present passage. In an aesthetically ordered system, other prop-
erties ought also to have complements which "balance," and as these
properties approach more closely the dynamic character of the
system as a whole, their symmetrical balancing should become more
evident.
Something about this listl or perhaps an initial ordinal listing
of another property, suggested to Plato an arithmetical illustration
of some aesthetic concept with which a reader of his time would
have been familiar. Whatever this concept might have been, it was
not recognized by the Hellenistic scholars when they wrote their
explanations of this myth.
202 Plato's Mathematical Imagination
Figure 78
THE FINAL LAW OF NINES: EXACT BALANCE OF MOMENTUM
IN THE SYSTEM
Circles 8 7 6 5 4 8 2

Masses • 1 8 7 3 6 2 4 5
Speeds t I 2 3 4 5 6 7 8

Sums 2 10 10 7 11 8 11 18
Momenta (Ordinal Arrangement of Sums)
8
I
4
L-..J
5 7 2
I - ---l
6 8 , ,
• From Figure 70.
t Linear forward velocity.
This is the law showing the balanced momentum (mass and
velocity) of the system. If the series is thought of as cyclical, so that
the initial 8 follows the final 1, the principle of location of balanc-
ing terms has become the simplest possible in this diagram.

Figure 74
LACK OF BALANCE IN PROCLUS' ALTERNATIVE TEXT
OF LIST OF SIZES
Order in series 234 567 8
Rim 7 8 6 4 8 2 5
(22) (No balancing of nines) (14)
(No balancing of sides)

Note that this version shows neither a balancing of nines nor a

symmetry of sides; since the lists of colors and velocities do show
such balance, the evidence against this text seems conclusive; and
the demonstration in the foregoing figures of the balance of the
combined lists of colors and volumes based on the MS text confirms
the rejection of this "older reading."
Figure 75 shows the spindle cut down the center, to bring out
the mechanical nature of the model. If these semicircles are com-
pleted on the same scale, the result is a top view of the lips of the
hemispheres, which the souls see.
 GEOMETRIC METAPHOR. 208
Figure 75

NECESSITY'S SPINDLE IN CROSS SECTION

Scale: Width of smallest circle = 1/16 1J

IX. THE MYTH OF ER-TRANSMIGRATION •

It was indeed, said Er, a sight worth seeing, how the souls sever-
ally chose their lives-a sight to move pity and laughter and as-
tonishment; for the choice was mostly governed by the habits of
their former life. He saw one soul choosing the life of a swan;
this had once been the soul of Orpheus, which so hated all
womankind because of his death at their hands that it would not
consent to be born of woman.! And he saw the soul of Thamyras
take the life of a nightingale,2 and a swan choose to be changed
into a man, and other musical creatures do the same. The soul
which drew the twentieth lot took a lion's life; this had been
Ajax, the son of Telamon, who shrank from being born a man,
remembering the judgment concerning the arms of Achilles.8
After him came the soul of Agamemnon, who also hated mankind
because of his sufferings and took in exchange the life of an
eagle. 4 Atalanta's soul drew a lot about half-way through. She
took the life of an athlete, which she could not pass over when
s~e saw the great honours he would win. G After her he saw the
soul of Epeius, son of Panopeus, passing into the form of a crafts-
woman;6 and far off, among the last, the buffoon Thersites' soul
• Trans. Cornford, Republic, pp. 357-58. Superior figures refer to Cornford's
notes, listed on following page.
204 Plato's Mathematical Imagination

clothing itself in the body of an ape. It so happened that the last

choice of all fell to the soul of Odysseus, whose ambition was so
abated by memory of his former labours that he went about for
a long time looking for a life of quiet obscurity.· When at last
he found it lying somewhere neglected by all the rest, he chose it
gladly, saying that he would have done the same if his lot had
come first. Other souls in like manner passed from beasts into
men and into one another, the unjust changing into the wild
creatures, the just into the tame, in every sort of combination.

Commentators since Proclns have noted that this panorama

of choices of transmigration parallels the list of lovers in the
Phaedrus. This parallel has already been cited in defense of an
asserted parallel between the Phaedrus list and the tyrant's num-
ber in the Republic. It has not been explained why the parallel
is incomplete; for the relative positions of Epeius and Thersites
on the present list transpose the relative positions of mimetic
artist and craftsman as these were arranged in the Phaedrus
scheme. It is reasonable to suppose that the proximity of the
me~aphysical condemnation of the mimetic artist, at the begin-
ning of Republic x, is somehow responsible for this relegation
of the imitator to last place, behind the artisan. Had the same
metaphysical distinction of artifact and mimetic artifact been
observed in the PhaedrusJ the effect would have been the adding
·Literally, the life Odysseus chose was that of a man who did not engage in
politics and who minded his own business. Compare this with the definition of
"justice" as "minding one's own business" in Republic iv.
Cornford's notes:
1. Orpheus was torn in pieces by the Maenads, the women-worshippers of
Dionysus.
2. Another singer, who was deprived of sight and of the gift of song for challeng-
ing the Muses to a contest.
S. After Achilles' death a contest between Ajax and Odysseus for his arms ended
in the defeat and suicide of Ajax. The first mention is in Odyssey xi.543, where
the soul of Ajax, summoned from Hades, will not speak to Odysseus.
4. The conqueror of Troy, murdered by his wife Clytemnestra on his return
home.
5. Atalanta's suitors had to race with her for her hand and were killed if de-
feated. Milanion won by dropping three golden apples given him by Aphrodite,
which Atalanta paused to pick up.
6. Maker of the wooden horse in which the Greek chieftains entered Troy.
 GEOMETRIC METAPHOR 205
of a fourth level of forgetfulness, making the matrix 12-celled,
or 4 X 3. But just prior to the 9-term list, we are informed
that lovers, each following in the train of his patron deity, are
12 in number. 150 I suggest that Plato here had in mind asche-
matism of tIle 12 gods analogous to the 3 and 4 of the 34-5
triangle in the nuptial number, which, if written out analo-
gously, is as shown in Figure 77. In this diagram, rows again
represent dominant parts of the soul of the follower of each
god, while the columns represent objects of love: soul, body,
artifacts, and mimetic works, respectively. Zeus is the patron of
philosophers; Athena, in her character of "lover of wisdom and
war," of such men as the good state's auxiliaries; Hephaestus,
of craftsmen; Apollo, of poets. Ares becomes the patron of the
proud king or warrior chief listed in the 9-celled matrix of the
Phaedrus. Artemis is patroness of lovers of the chase, who are
not concerned with improvement of the body, as the gymnasts
and physicians are, but with the capture of external creatures.
Hermes, whose name in the Cratylus is found appropriate to his
attributes of thief and liar, becomes the patron god of Sophists,
who imitate intelligent discourse. Aphrodite, the goddess of
physical love, becomes the patroness of those who put their faith
and center their interest in bodily appearances, as the democratic
man typically does in the Republic. Hestia, the stay-at-home
goddess who never sees any of the forms, is well adapted to be
the prototype of the woman who, always preoccupied with her
affairs within the house, has no interest in or knowledge of the
world outside, and acquires no virtue but only petulance from
her preoccupation with domestic minutiae; this is a type of
woman whom Plato elsewhere deplores, and intends to abolish
in any good state. Poseidon, finally, the Earthshaker, ruler of
the stormy sea, driver of horses, founder and patron of the vast
and potent state of Atlantis, is, in his unbridled power, pre-
sumably the patron deity most akin to the power-seeking tyrant.
A sign of this is the lapse of the god's Atlantean descendants
into unbridled tyranny. Not only does Hestia never march in
the parade, but there would seem to be great doubt as to
whether Poseidon's steeds and chariot, designed to be ocean-
206 Plato's Mathematical 1magination
going, will ascend high enough to show their follower anything
but the foam-crested waves of the sea.
Arranging the souls which transmigrate in terms of this list,
they seem to represent a section running across the upper row,
with the exception of Atalanta, whose juxtaposition with Ajax
and Agamemnon helps to indicate their true characters. The list
of transmigrating souls now becomes:
ORPHEUS AGAMEMNON EPEIUS THERSlTES
THAMYRAS AJAX
SWAN AND OTHER ATALANTA
MUSICAL CREATURES

Here, as in the later Phaedrus list, a larger number of instances

may tend to indicate the categories of a higher type. 151
The interrelation of these characters is partially indicated by
the comment that the lot of Ajax was the twentieth.152 There is
good reason to suppose, with the ancients, that this piece of
information is a functional detail. It appears clear that 20' is
chosen because Ajax' category of lover is second in excellence,
and viewed in the mythical context of punishment and reward,
the ordinal difference is here given its eschatological tenfold
magnification.
Nor do the first and last choosers in this panorama fit badly
with the present interpretation of the Epeius-Thersites reversal.
The first man to choose represents a masculine follower of
Hestia. Incuriously accepting life in an orderly state, he appears
on the plain with no intellectual vision to guide him in making
his choice. As his exact opposite, Odysseus, drawer of the final
lot, has seen all the cities of men. He is the very antithesis of
the stay-at-home type, and he has been the hated enemy of the
god Poseidon. The life of Odysseus' choice is one in which a
man has a chance to become a follower of Zeus by cultivating
philosophy; and this choice by the hero of widest experience
gives a pragmatic confirmation to the doctrine of the relative
happiness of the just man developed theoretically in the ante-
cedent argument of the Republic.
This analysis raises the question, also posed by the operation
of subtraction of segments in the nuptial number diagram, of
the reason for the contraction of the larger schematism of 12
 GEOMETRIC METAPHOR 207
lovers to the smaller matrix of 9. It was suggested, in connection
with the nuptial number operation, that the four-part division
of knowledge, while metaphysically relevant, was not morally
so, since normal human experience would by maturity give
every man the kind of practical knowledge which is character-
ized as "opinion." In effect, this collapses the metaphysical dis-
tinction between the makers of artifacts and the makers of
mimetic artifacts. Psychologically, the poet and craftsman are
akin; neither has an art which conduces to his moral excellence
or wide experience. Both, therefore, operate on the level of
opinion; both may be regarded as lovers who have forgotten
everything about the forms once seen except those properties
observable in external bodies. The followers of Apollo, Hermes,
and Poseidon differ from those of Hephaestus, Artemis, and
Hestia neither in forgetfulness nor in vice; therefore the 12
trains of gods are adequately represented by only 9 kinds of
lovers. In this reduction, the poet finds himself identified with
the lover of beautiful bodies, the follower of Aphrodite, the
democratic mentality; and since ignorance is the major princi-
ple of organization of the list, the poet is placed a step above
the craftsman. The hunter is not included, the Sophist replacing
him; the stay-at-home drops out, and is replaced by the ignorant,
stay-at-home tyrant.
Figure 76
MATRIX OF CHARACTER WITH FOUR LEVELS OF
KNOWLEDGE DISTINGUISHED
INTELLIGENCE FORMS BODIES ARTlFACfS IMITATIONS

REASON Philosopher Auxiliary Artisan Poet

SPIRIT General Athlete Sophist Actor
APPETITE Merchant Democrat Tyrant Housewife
MOTIVATION

Though other contexts (e.g., Republic 546C) involve imagery

suggesting that no adult of normal experience will remain in the
state of being of someone whose knowledge is pure EhtQa(a, such a
life remains a metaphysical, if not a practical, possibility. It would
represent a level of ignorance below that of the artisan, whose
208 Plato s Mathematical Imagination
l

interests are centered on mimetic artifacts which lack tangible func-

tion and reality. The actor, as he projects the poet's insight into
physical performance, obviously belongs where he is placed; and
this explains why Epeius and Thersites appear in the pageant in
that order. The location of the incurious, uninformed housewife,
on a level of stupidity equal to that of the mountebank and too
out of touch with practice to make possible such social damage as
the tyrant effects, is suggested by the fact that the patroness of
keepers of the hearth, Hestia, is the only deity who never joins in
the parade of the gods, so that her followers patently have no
opportunity at all for intellectual insight.
This matrix is not necessarily one that Plato himself had in
mind, but is put forward as a schematic way of showing how the
contextual stress on the four levels of reality of the divided line
could suggest the reversal of actor and artisan.

Figure 77
MATRIX OF CHARACTER APPLIED TO OLYMPIAN GODS WHO
ARE PATRONS OF TYPES OF HUMAN LIFE
FORMS BODIES ARTIFACTS IMITATIONS

Zeus Athena Hephaestus Apollo

Ares Artemis Hermes Demeter
Hera Aphrodite Poseidon Hestia
For Plato's myth to be plausible, it is necessary only that the
range of interest and temperament of the Olympians be adequate
to the range of lives that their followers represent. However, as
the present list of patrons shows, the actually recognized personali-
ties of Olympus present a fair one-to-one correspondence with the
different categories of type of character derived from a 12-celled
reason-motivation matrix. Aphrodite seems a suitable patroness for
the democratic man who is a "lover of sights and sounds"; Demeter
is identified with the actor, perhaps on tenuous grounds, because
she communicates in the mysteries in her honor through enacted,
allegorical pageantry; Hestia, as outside the procession, seems to
require the corner location in the present scheme.
This figure is intended to suggest that no matter how closely
Plato checked the adequacy of his story of the twelve processions
and patrons, the correspondence to his purpose at hand would still
have seemed adequate.
 CHAPTER IV

A1gebraic Metaphor
Introductory Comment

PLATO chooses with care and frequently employs mathematical

illustrations based on the concept of ratio. The modes of pro-
portionate, qualitative relation may be metaphorically identi-
fied with the modes of relation of things, either as parts in
wholes or as entities juxtaposed in some sequential order. The
metaphor may he projected, in turn~ from algebraic to geo-
me~ric symbolism, in which case proportionate relations are
presented as constitutive of a metric scale or net. In this context
more than in that of geometrically mathematical imagery, which
we usually identify as "mathematical imagery" proper, a later
literalness which refuses to accept the concept of algebraic illus-
tration as metaphor has rendered passages opaque to attempted
interpretation.
The only way in which the metaphorical significance of pro-
portion can be recaptured is through the device of defining and
classifying certain more general types of relational pattern, of
which quantitative ratios provide illustration as instances of
each type. In this way, some of Plato's less understandable state-
ments about proportion can be exhibited as metaphors of the
familiar kind, in which there is "substitution of one species for
another within the same genus." But to do this involves a pre-
liminary abstraction and fonnalization ",vhich only a Platonist
enamored of the Parmenides is likely to enjoy.
In the course of this discussion, several instances will appear
of a Platonic metaphor which has later, in a more formal and
literal interpretation, provided the basic insight of a branch of
209
210 Plato's Mathematical Imagination

mathematical science. Such instances of later development will

be cited not with any intention of claiming that Plato antic-
ipated statistics, logarithms, and calculus but as evidence that
his choice of algebraic metaphors is not unmathematical, as
their contextual vocabulary could lead a modern reader to
suspect, and that in some cases the metaphor chosen has been
adopted as literal fact by later mathematicians.
In general, four entities, a, b, c, and d, are in proportion
when the relation of a to b is identical with that of b to c.
A term b is a mean when a has to b the identical relation that b
has to c, and when the set aJb has this same relation to the set
bJc. Two sets in proportional relation are "similar," though
some cases of intransitive similarity probably cannot be con-
strued as literally proportionate. It is, of course, neither neces-
sary nor desirable, as these abstract preliminary definitions indi-
cate, to interpret proportion as an essentially quantitative re-
lation.
The degree and mode of connectedness which a proportion
expresses depend on the natures of the two relations (1) be-
tween the members of the sets and (2) between the two sets
themselves. In effect, the symbol "::" is itself analogical or
equivocal, since different types of relation may correlate sets
that are analogous. Relations of "::" may be classified into two
broad types, the "translative" and "projective." A translative
relation is a correlating relation equivalent to a rigid displace-
ment of one set in space or time. If isomorphic and analogous
sets of parts are combined in identical relational nets at differ-
ent dates or positions, the analogy is so highly specific and close
that the sets and their elements are called "the same." Thus
the organs of one man are "the same" as those of another,
since the two men have "the same" structure and form. The
ultimate test of translatability is dependent on the identity of
function in context of any two structures that are the same.
We can imagine the one set replaced by the other-a given
table by another table, a saw by another saw-and this replace-
ment effects no change in the context or function.
The degree of analogy involved in the class of "projective"
 ALGEBRAIC METAPHOR 211

relations is less close, and projectively related sets are only

"metaphorically called 'the same' " or "similar." As opposed to
rigid translation, projection may involve both change of kind
and change of scale between the parts related. It is not possible
to imagine the one set as simply substitutable for the other,
because of a difference in their respective contextual functions
and their spatial and temporal scope. Perhaps it would be- more
accurate to say that the elements of sets in projective relation
are correlated on a different principle from that correlating
those in translative relation. The elements of translative sets
are connected by identical relations, and are isomorphic. With
projective sets, this identity no longer holds in the same way;
one can say only that any two correlated pairs are connected by
relations of the same relational class or type. In organized struc-
tures, the relations of elements are always "harmonious." A
"harmonious" relation may be described by its stability; it is a
relation that preserves and reinforces the identity of each of its
relata, both in itself and in its context. All relations not
"harmonious" are "hostile." Elements in hostile relation alter
both their relational patterns and their identities.
We may now summarize the nature of projection by saying
that if the relations of parts of two "similar" sets are correlated,
each correlated pair will belong to the same relation-class in
respect to its harmony or hostility. Two sets need not be isomor-
phic to be placed in such a projective correlation; the relations
between complex parts of one may be correlated with the rela-
tions of incomplex parts of the other. If the one structure is
replaced by the other, the result will be invariant in respect
to its internal unity, but altered entirely in respect to its func-
tion in context. (The nature of the modification of an analogue
to the theory of types required to avert paradox when such a
broad concept of correlation is introduced would require an
intensive study of the formal structure of Platonic logic.)
Some such preliminary abstract discussion as this is needed
as a way of explaining and correcting a peculiarity of the inter-
pretation of proportions and analogies that became characteris-
tic of Neo-Platonism. A typical Neo-Platonist would see no
212 Plato's Mathematical Imagination

sharp difference in kind between the analogical identification

of the cosmos with the human organism and that of one individ-
ual human organism with another. In the latter case, as in the
former, his interpretation of dialectical proportions presup-
poses a one-to-one correlation of all parts of both sets and an
identity of the relational connections between them. Thus, if
"man is like the heavens," it would seem legitimate in this
mode of discourse to ask where the heart, liver, etc., of the
heavens are. If he finds the constellations analogous, astrology
becomes a branch of the Neo-Platonist's philosophy. Or he may
ask-and this is a key question of Neo-Pythagorean embryology
-what seven ages in the life of man correspond exactly to the
ratios of the heavenly periods of planetary motion. The assump-
tion that the relations of all members of projectively analogous
sets are literally identical leaves room for only one key pattern
of proportions, which all subject matters share. Thus a biologist
in this tradition can make the assumption a priori that the
quantitative ratios of the periods of the planets are identical
with the set of key ratios determining organic development, and
his only task is that of filling in the blanks in this ready-made
pattern with the names of the biological parts with which the
planets are in one-to-one correlation. This leads to absurdities
ranging from the conviction that all sets with the same number
of members are essentially the same (Aristotle 1 satirizes this
belief in his criticism of the Pythagorean assertion of the essen-
tial analogy of all sets having the cardinal number 7) to the
conviction (for example, of Proclns) that each passage in Plato
is susceptible of interpretation as a theorem in every known
science. 2 Plato's own handling of such matters as the microcosm-
macrocosm analogy and the interpretation of other philosophic
texts leaves no doubt that such grotesque dialectical practice
was not his own.
For Plato, the likenesses of structures which are wholes (i.e.,
structures adapted to single functions) hinges on the fact that
all parts in any such structure must be connected in a way
which prevents either the deformation or removal of any func-
tional part, or a transposition of parts which would lead to such
 ALGEBRAIC METAPHOR 213

deformation or removal. The number and character of "parts"

will vary with the function of the "whole"; and the specific
relations through which each part, by sustaining the whole, also
preserves and sustains the other parts, will also vary. The invari-
ant basis of the Platonic projective analogy is that the unity
and value of organization demand that the constitutive organiz-
ing relations be the same in kind. Thus the collection of
analogies ceases to be an arbitrary game, because the demand
of functionality introduces a restriction to the range of com-
parable parts, and the demand of strict isomorphism among the
equations of the several sciences is replaced by the demand that
every science be inclusive and comprehensive, so that each
theorem may contribute to and be functional in the system as
a whole. A later interpretation of the mathematical images of
the Parmenides will be concerned with this relation of parts,
wholes, functions, and their associated structures. 3
A second distinction of kinds of relation constituting propor-
tion is a distinction, within each member of the analogy, of
the mode of connection of its terms. The connecting relation
between terms may be simply an identical juxtaposition, an
external relation, or actual identification of one term with an-
other within a connecting constitutive class, which is an internal
relation. The Pythagoreans and Plato differentiated multiplica-
tion from addition on the ground that a number's factors are
constitlltive of it, determining its nature in a way that a ran-
domly chosen predecessor in the integer-series, which may be
equated to the number by a suitable addition, does not.
Thus if the ratio a : b is a multiplicative relation, a is a factor
constitutive of b, and a and b have something in common, since
the character of a is reflected in the nature of b. The use of
multiplication as a symbol of marriage brings out the basic
notion of this interpretation; the children, who are the "prod-
uct" of father and mother, resemble both in their inherited
natures, just as an arithmetical product resembles both of any
pair of its factors.
On the other hand, if a : b is an additive relation, only some
sort of juxtaposition, and no real connectedness, of a and b is
214 Plato's Mathematical Imagination

created. The connecting power of a mean term between the

extremes a and c can be symbolized by noting that if the mean
is geometric, such that ax = band bx == c, then ax 2 = c; and
+
that if the mean is arithmetic, a 2x == c. The geometric rep-
resentation of ratio, in which a • c is a closed rectangular figure
+
and in which a c is simply two juxtaposed figures, reinforces
this notion of the external character of sum as opposed to
product.
Two classes may be "bound" or "connected" by the creation
of an intermediate class between them, similar to both. Like
entities fuse together, unlike separate. Thus if the two unlike
and extreme classes, a and c, are brought into contact, they will
not hold together. If, on the other hand, a is brought into con-
tact with a like class, b, and b with c, which it is like, the chain
a-b-c will hold. Terms in proportionate relation cannot be dis-
connected by reversal, translation, or any external force causing
a transposition, since any such force will have the same effect on
all similar entities on which it acts, and will therefore exert the
same transposing effect on both sides of the proportion.
The concept of analogy is also fundamental to any definition
or theory of measure. In its simplest form, measurement simply
establishes a translative relation between an arbitrary "stand-
ard" and a "magnitude" homogeneous with it, to which the
standard is applied. The relation within members is additive,
and the integers are used to count the successive translative
applications of the standard to the magnitude. Thus such meas-
urement takes the form of the proportion I : s : : x : m.
It is also possible to conceive sand m as related projectively,
and the integer-standard as a multiplicative relation rather than
an additive one. This leads to measurement of a different type,
which Plato calls "evaluation" and opposes to pure "descrip-
tion." This is the sort of measure that a craftsman uses to cor-
rect dimensions of parts which are detrimental to the function
of the whole. In such evaluation, description may still be in-
volved; indeed, without the accuracy contributed by descriptive
technique, the craftsman could not achieve the desired results.
The basic relation of such nonnative measure may be schema-
 ALGEBRAIC METAPHOR 215

tized by the proportion: I: [p : s] : : " : [q : m], where p and q

are functional parts of the wholes sand m, and the craftsman
must solve for x, the descriptive measurement of q.
Two different concepts and employments of metric nets fol-
low from these metaphors of measure. An additive or descrip-
tive metric net is a set the terms of which are connected entirely
by repetitions of a translative or additive relation, R. Thus a
spatial grid-scheme, with a spatial unit serving as the side of its
squares, and with comers as entities connected by the grid, is
an example of a descriptive net of this type. The scope of such
a net can be increased (that is, entities not falling at intersec-
tions can be described) by a further subdivision of the individ-
ual net squares. It is significant that Plato nowhere employs a
net of this simple descriptive type, although in several passages
he introduces variations of it in taking account of mechanical
interactions between parts. The appropriateness of this applica-
tion is perhaps indicated by the fact that any problem for which
the operations of solution can be represented in such a descrip-
tive net can be solved by a calculating machine. This remains
true if the operations involved include not only R, but also the
converse of R, R-l. The operations of an abacus or balance are
such descriptive calculi; and their mechanical character is dep-
recated by Plato when he separates "computation" from
"mathematics proper" in Republic vii. 4 The "normative"
metric net, on the other hand, gives the quantitative analogue
of systems of part-whole relations such as are expressed by the
artist's proportions.
The simplest example of such a net is the bare enumeration
of parts in a hierarchy, where the order reflects a relative im-
portance or scope. The combination-matrix, in which two sets
of ordered spatial positions are correlated with two sets of
ordering relations, is another such simple projective net. In
fact, the use of such symbols to schematize relational order is
the fundamental principle of the construction of mathematical
imagery. In the Laws, Plato suggests that a divine justice would
so arrange distribution of wealth that the two series, "relative
wealth" and "relative human excellence," would coincide. 1S
216 Plato's Mathematical Imagination

A normative system of measurement involves some technique

of measuring and comparing degrees of deviation from the
standard or norm. Such an "index of aberration" can be derived
by comparing the actual dimension of a part with the ideal
dimension; the ratio of these two, taken as an index of error,
would permit a comparable evaluation of technical works of
any type or size. In fact, if we grant the highly un-Platonic
axiom that most entities realize the ideal structure, modern
applications of the normal distribution curve and the standard
deviation are exact counterparts of Plato's metric metaphors of
the normative type.
The problem in legislation may seem to be that of making
tIle social privileges of wealth and the power of a citizen deviate
from some constitutionally established "normar' by a distance
comparable to that by which his actions deviate from the legally
established social norm. We should look, then, in the arithmeti..
cal details of the Laws for some such correlation of degrees of
deviation.
One other, more special usage of ratio must be considered.
This is the use of geometric progression as a.schematism or sym-
bol of organic growth and causal efficacy. The Laws, Epinomis,
and Aristotle in the De anima record such schemata.6 An in-
corporeal principle becomes causally effective in space and time
first by generating an order, then a two-dimensional plan, and
finally an embodying spatia-temporal construction. Plato sym-
bolizes this process by which an idea or insight issues in a physi-
cal creation as a I: 2: 4: 8 progression, from 0 to 3 dimensions.
A goal or ideal first manifests itself without any associated plan;
more relevant factors are ordered in respect to it, until at last a
construction issues in physical action. The organizing power of
a principle, as it projects order into structures of increasing
complexity, generates a relational net of order correlated in this
metaphorical representation with the exponential relation be-
tween its terms. The internal relations of organizing, and of
being organized by, are used as the framework of the propor-
tions which God marks off on his cosmic scale in the Timaeus,
and by which he determines the intervals of the circles of the
 ALGEBRAIC METAPHOR 217
planets. 7 The important aspect of this exponential net in its
COlltext seems to be that it carries over the symbolism of the
multiplicatively related mean as establishing an internal rela-
tion; so that the connection between any sequential entities in
the net is internal, and their relation one of "friendship" or
"linked connection." The choice of the exponential relation as
a symbol of internal connection and harmony results in a con-
struction which the modern mathematician would at once rec-
ognize as a logarithmic scale. The research into the logarithmic
spiral as typical of processes of growth in nature would no
doubt have seemed to Plato a sound and illuminating metric
metaphor. 8
In classifying ratios Plato follows the Pythagoreans in setting
aside as a special harmonious class those integral ratios which
had been found empirically to correspond to the concordant
intervals of the musical scale. The peculiarity of these ratios is
that they are simple and integral, with integers less than 10.
"Concordance" seems to result from a special kind of co-opera-
tion or reinforcement between different tones. A concord is not
the average pitch of its constituent notes, but a combination in
which each constituent reinforces the identities of the others.
The peculiar simplicity of the integral foundation of musical
theory may have been instrumental in suggesting the doctrine
attributed by Aristotle to Plato, that the number series is modu-
lar at ten, so that the first ten integers are adequate to describe
and establish the whole sys~em of relations of numbers. 9
A "concordant" relation seems to fall, as a metaphor of con-
nection, somewhere between the internal relatedness estab-
lished by a geometric mean, and an external relation such as is
symbolized by an arithmetical mean. A third type of mean, the
"harmonic," s¥mbolizes this intermediate tightness of connec-
tion.
The preceding discussion is illustrated by Figure 78, on the fol-
lowing page.
The ordering of kinds of motion in Laws x entails a classifica-
tion of relations similar to that developed in Figure 79. The kinds
of change are ordered in terms of the amount by which the identity
 .~..
:'1:

218 Plato's Mathematical Imagination

Figure 78
SCHOLION: THE THREE KINDS OF RATIO
(Greene.. Scholia PlatonicaJ pp. 167-68)
Gorgias 508A
Geometric Equality •
(1) This is justice, for the geometrical proportion is called the
order of Zeus in the Laws, and through it all things are ordered
and determined. (2) A proportion is called "arithmetic," when of
three unequal numbers, the mean exceeds and is exceeded equally,
e.g., 2:4:6. It is called "geometric" when of three unequal num·
bers, the ratio of the first to the second is also that of the second
to the third, e.g., 4:6:9; it is called "harmonic" when of three un-
equal numbers, the ratio of the greatest to the least is the same
that the larger difference has to the smaller, e.g., 2:3:6.

2 4 6 469
I - .=.2a . -_
_ .Diff. 1 =2
Diff. _ 1 1__R_at_io_=_3_/2.....11..._R_a_tio_=_3_/_2_I
ARITHMETIC GEOMETRIC

2 3 6
I Dill. = I 1 Diff. = 3 I
Diff. =2

HARMONIC, 3: 1 ratio

of the subject is affected, i.e., by the degree of difference between

the object in its final changed and original form. Since the classi-
fication of tightness of analogy was suggested as based on the kind
of relation necessary to transform one analogue into its counterpart,
the list should be the same as a list of the relations holding between
the initial state of a thing and its form as altered because of change.
The list of changes runs: rotation, translation, mixture, separation,
• This is Plato's phrase. It is used by the scholiast to identify the subject of
his comment.
 Figure 79
MOTION

I I I
BODY SOUL
I
I
PLACE
I
QUALITY
I
QUANTITY
I
BEING

J"¢
to-
cc Moving and
Rotation Mixture Increase Generation Moved by Others

Translation Separation Decrease Destruction Self-moving and

Controlling Others
SCHOLION: THE TEN KINDS OF MOTION (Laws 894B)
(Scholion figure from Greene, Scholia Platonica, p. 354)
220 Plato's Mathematical Imagination

growth, diminution, generation, destruction, self-motion externally

caused, self-motion self-caused. (0£ these, we are told that the last
two should have been put first, as they are the primary and prior
sorts of motion.)
In effect, this gives the following classification of the correlating
relations by which analogy can be established:
1. Rotary and translative relations of correlation hold where
there is a one-to-one correspondence, and such analogy is so close
that it may be stated as an identity.
2. Mixing and separative relations involve a one-to-one corre-
spondence of one analogue with a subset of the other. The two
are not interchangeable, and the analogy is less close.
3. Growth and diminution involve, on the one hand, new addi-
tions, so that correlation is still only with a part of the changed
form; but within this correlated part, the relation is projective;
whatever scale is relevant must be stretched or shrunk to show the
analogy of the one with the other.
4. Generation and destruction, as relations transforming two sets
into each other, can yield only a statement of contrast, and no
analogy at all.
5. Self-motion is a· property of organized dynamic systems, which
are analogous in their common possession of some sort of part-
whole organization, retained and self-&ustaining through changes oi
the other kinds enumerated. Thus an organism retains its identity
through changes of place, context, and growth. Analogy between
such organisms rests on the similarity of kinds of relation holding
between parts of functional wholes, and on the common resistance
of' such relational nets to change in spite of other relational changes;
the parts remain integrated throughout a motion of the whole.
Analogies based on relations of this sort, while more difficult to
exhibit as simple correlation, take account of a more basic identity
and similarity than those of other kinds.

I. CONSTRUCTION OF THE WORLD-SOUL

Timaeus 35·
The things of which he (God) composed the (world) soul, and the
manner of its composition, were as follows: (1) Between the in-
divisible Existence that is ever in the same state and the divisible
• Trans. Cornford, Cosmology, pp. 59-60, 71, 73.
 ALGEBRAIC METAPHOR 221
Existence that becomes in bodies, he compounded a third form
of Existence composed of both. (2) Again, in the case of Sameness
and in that of Difference, he also on the same principles made a
compound intermediate between that kind of them which is in-
divisible and the kind that is divisible in bodies. (3) Then, taking
the three, he blended them all into a unity, forcing the nature of
Difference, hard as it was to mingle, into union with Sameness,
and mixing them together with Existence.· And having made a
unity of the three, again he divided this whole into as many parts
as was fitting, each part being a blend of Sameness, Difference
and Existence.
And he began the division in this way. First he took one por-
tion (1) from the whole, and next a portion (2) double of this;
the third (3) half as much again as the second, and three times
the first; the fourth (4) double of the second; the fifth (9) three
times the third; the sixth (8) eight times the first; and the seventh
(27) twenty-seven times the first. Next, he went on to fill up both
the double and triple intervals, cutting off yet more parts from
the original mixture and placing them between the terms, so that
within each interval there were two means, the one (harmonic)
exceeding the one extreme and being exceeded by the other by
the same fraction of the extremes, the other (arithmetic) exceed-
ing the one extreme by the same number whereby it was exceeded
by the other.
These links gave rise to intervals of 72 and % and % within the
original intervals. And he went on to fill up all the intervals
of % (i.e. fourths) with the interval % (the tone), leaving over in
each a fraction. This remaining interval of the fraction had its
terms in the numerical proportion of 256 to 243 (semitone).
By this time the mixture from which he was cutting off t these
portions was all used up.

In his construction of the world-soul, Plato represents God

as a maker of astronomical models. In order that the constituent
motions of the system may be specified with greater clarity, the
soul is described as constructed of moving concentric bands,
like the metal bands representing equator, ecliptic, and plane-
• Cornford retains a:~ 3t£Q£ (see p. 225).
t Seenotes 11 and 12 to this chapter.
222 Plato's Mathematical Imagination

tary orbits in a Greek astronomical mode1. 10 But before he

bends and arranges the strips of the alloy of being, same, and
other, God cuts into the bands the marking of a metric scale,
to serve as a measure of their relative positions and motions. 11
He subsequently uses the framework of this same scale to adjust
the distances and velocities of the subdivisions of the circle of
the other. 12
There is a double problem that must be met in such a cosmic
construction. In the first place, same and other, by their natures
hard to mix, must be harmonized in such a way that the unity
of the world-soul will be preserved. As a corollary, the external
relations of part to part in cosmic process, which are relations
according to the other, must be given their proper places as the
mechanism by which these moving parts are forced into the
internally adjusted patterns which relate them according to
the same.
The metric scale which God employs, and etches into the
substance of the world-soul, is constructed in a manner which,
through a meticulous use of mathematical metaphor, symbolizes
the structure in which God has blended the other and the same.
There has always been a temptation for later readers to con-
strue these proportions as intended to present an empirical
description of astronomical phenomena. There is always also
the opposed temptation to treat these proportions as a whim-
sical elaboration of the poetic concept of the "harmony of the
spheres." But neither of these interpretations is quite correct.
Plato's metric scale is actually a presentation (as Proclus thought
it was) in the form of algebraic metaphor of a schematism of a
relation of parts and wholes in adjusted structures, a schematism
indicating the dialectical method appropriate to cosmology. In
this passage, differences in kind among entities appear as differ-
ences in quality between the illustrative numbers and their
ratios, and God builds into the world-soul exactly the type of
structure which reason will develop as the architectonic of ra-
tional cosmology.
The schematism is based upon a pair of different exponential
series, of the type previously referred to in the classification of
 ALGEBRAIC METAPHOR. 225
ratios, as symbolizing internal relations. It is important that the
key elements be integers, integrally related. Presumably the
tradition is quite right which identifies the sequence of even
numbers with the element of the other, the odd numbers with
that of the same. IS These two distinct constituents are mixed
or harmonized by the insertion between them of arithmetic and
harmonic means. Schematically, this is analogous to a gluing or
joining of substance between same and other; for the arithmetic
mean is a mathematical analogue of the relation holding ex-
ternally between mechanically connected objects, and the har-
monic mean also has this character of external connection.
Although such statements of interpretation as "odd numbers
symbolize the same" have a Neo-Platonic sound to them, and
therefore are automatically discounted by the modem scholar,
this is a case where the Neo-Platonists were entirely right. If
it is true, as has been already suggested, that Proclns tried to
find in every mathematical image in Plato a schematism of
method and relations which the scientist could use as a kind
of standard form in which to arrange his data, this particular
mathematical image was made to order for his technique. True,
Plato presents this as a schematism of the distinctions appro-
priate to a discussion of cosmology, and there is every reason
to think that he would dispute any attempt to apply it to other
discussions. It is peculiar to the Timaeus J and would be mis-
leading and out of place in the Republic. But, in the Timaeus,
it does just what Proclus expected it to do; and his interpreta-
tion, on which the following discussion is based, has never been
equalled or surpassed.
To counteract the overtones of this explicit Neo-Platonism,
it is only right to state that what is to be developed in the
present discussion is not a theory of mystical affinities between
numbers and things, but a display of the connections of the
terms of a cosmology in an algebraic schematization. Such an
algebraic display is thoroughly, in fact surprisingly, modern in
spirit. The concern of contemporary symbolic logic is precisely
the study and construction of such algebraic schemata. But this
modernity of spirit has been effectively masked by an antiquity
224 Plato's Mathematical Imagination

of form which depends on metaphorical interpretation of a spe..

cific instance, rather than on an ambiguous notation, to secure
its generality of statement.14
The central point in this schematism is its remarkable reHec..
tion of the fact that a right connection of separate parts, sche-
matized by the insertion of the arithmetic and harmonic means,
can lead to a fusion and emergence of these parts internally
connected in wholes. The geometric ratio reappears as a connec..
tion between sets in the developed scheme. (The geometric
ratio had been identified by Plato with the class of relations that
hold together the parts of functional wholes.) The unity, intel-
ligibility, and wholeness of the world..soul are insured by con-
tinuing the construction until a harmonious ratio holds be-
tween each adjacent set of parts.
The schematization has been developed in stages that succes-
sively direct the reader's attention to types of relation which
themselves appear successively as central in Plato's own discus..
sion of cosmology. First there is the relation of the pair of
constituents, being and becoming, into which the cosmos can
be analyzed. Next there is a discussion of the juxtaposition of
externally related elements which are the "auxiliary causes" in
making the copy resemble its model. Finally, in the Timaeus,
there emerges a set of organic patterns created from the planned
construction of these externally related connections.
The cosmologist, like the creator, founds his discussion on
the principles of being and becoming, and same and other, as
basic constituents of the cosmos. Within this frame of reference,
however, he cannot explain such phenomena as death, disease,
and war, which arise from the interference of externally related
parts within the whole; a description of creation in terms of its
likeness to a perfect model can account only for its total perfec-
tion, not for any defections in "its parts. A theory of elements
must therefore be provided, to explain the mechanism by which
being is projected into becoming. The final view, in which the
mechanism of elements in the universe is seen to create the per..
fection of the universe, and of its subordinate constitutive
wholes, shows how external relations of elements, built into
 ALGEBRAIC METAPHOR 225
organisms, can frustrate effective organization, and gives a
clearer insight into the way in which reaso,n for the most part
persuades a recalcitrant necessity.
This same schematism will therefore apply to a discussion of
astronomy, as a study of a special, particularly accurate, aspect
of cosmic order. The creator uses the same scale in subdividing
the soul of the world that he employed in its original constitu-
tion, so the schematism of creation serves as the paradigm of
the system of planetary motion as well.
This interpretation of the series of even numbers can hold
only if the frequently bracketed a{; :TtEQL is retained in the
text, as Proclus retained it. 15 The effect of this phrase is to in-
dicate that the mixture does not involve same and other in
their absolutely pure form, but a sameness and otherness al-
ready partly blended with substance, so that their natural
contrariety is lessened. The relation of otherness, as explained
in the Parmenides, is solely an external, part-ta-part relation. 16
Unless we think of this principle as already sufficiently imbued
with some sameness through its mingling with substance, rela-
tions of otherness would have to be schematized disjunctively,
as an atomic set. However, an otherness manifesting itself as an
external relation of physical or psychic entities can be schema-
tized as involving some internal connection. The phrase some-
what complicates Plato's account, but is needed to make the
intended point that God is a creator, not a dialectician, and that
his creative activity utilizes a substantial raw material as me-
dium, not the dialectician's pure, insubstantial, abstract contrary
classes. The result of mixing pure same and other would not
be a soul, but an intelligible class, assuming (quite un-Platoni-
cally) that 'God could cause these contrary classes to mix in their
pure form. The classification of the raw materials thus follows
the same method as the later schematism of cosmological struc-
ture: intermediates are inserted to connect the disparate classes
f of other and same, just as arithmetic and harnlonic means will
I
I
!
be inserted in the following schematism to connect the disparate
arithmetical classes of odd and even.
This is a remarkable synthesis of many complex mathematical
226 Plato's Mathematical Imagination

metaphors. The initial series from I to 27 gives the metaphor

of the music of the spheres. The progression to 8 and 27 intro-
duces the metaphor of progression from a principle to its em-
bodiment, through squaring and cubing. The initial contrast
of odd and even series schematically reproduces the opposition
of same and other, the ingredients blended by God in his crea-
tion of the soul. The two "bonds" created by the first insertion
of harmonic and arithmetic means reflect the two aspects of
cosmic structure central in the dialogue, the arithmetic mean
symbolizing the external relations of mechanical causality, the
harmonic approaching more closely structures such as those
which have a purposive organization and are treated in the
"likely account" early in the discussion. The reduction of the
whole series to a continuous geometric progression (except for
the AEL~Jla, where, since two metaphors conflict, the reduction
for some terms is only approximate) insures the organic unity
of the cosmos as God has connected its parts. And, if we think
of the specific numbers in this net as replaced by variables, the
schematism gives an abstract representation of the logical rela-
tional structure proper to research in cosmology.
In addition to its synthesis of this imposing set of geometrical
and algebraic metaphors, the repetition of the 2: 1 ratio of the
concordances of the initial octave scale carries with it a cyclic
or modular metaphor. (The concept of a modulus, a mathe-
matical metaphor that is natural for reiteration, seems to have
been already formulated by Plato's time. The Pythagoreans'
concept of the pythmen, their identification of 6 and 5 as "cir-
cular" numbers, and their insistence upon 10 as the natural
base of the number system show this idea emerging in arith-
metic.) 11 This matrix of doubles is also extremely attractive as a
starting-point for interpretations of the 2: 1 ratio as a principle
of projection and of the problematic "dyad." 18
The synthesizing effect of such a mathematical image as this
is so like the complex connection of image-classes in lyric poetry
that the image may be legitimately classed as an aesthetic con-
struct rather than as an abstract mathematical one. For Plato,
in his practice as well as in his theory, the two are closely akin.
 ALGEBRAIC METAPHOR 227
A thesis basic to this study, that no peculiarly technical or
esoteric reference underlies this imagery, could not well explain
a passage which demanded the solution by its reader of thirty-
four separate algebraic equations. Such an image would be less
a pedagogically useful schematism than a special problem, pre-
supposing a reader who was, for instance, an advanced student
in the Atademy. Actually, however, this calculation, as pre-
sented in the following 8 X 4 matrix, requires only that the
reader (1) be familiar with the metaphor of the "harmony of
the spheres," (2) know the ratios of the notes of the diatonic
scale, and (3) be able to multiply by two. The second condition
can be more accurately stated as follows: The reader must
know (a) that the notes of the octave are in 2: I ratio, and (b)
that the insertion of harmonic means in the octave gives musical
concords. These items of information are represented as fa-
miliar to characters in the dialogues, such as Eryximachus and
Glaucon, who know them as parts of their general culture, not
as a result of any specialized Platonic or Pythagorean training.

Figures 80-86
CONSTRUCTION OF THE WORLD-SOUL

Figure SO

INITIAL TERMS OF THE SCALE

1
2 5
4 9
8 27

The lambda-arrangement brings out the distinction of the odd

numbers, representing the same, and even numbers, representing
the other.
SAME OTHER
1
~ 2---Linear number
9 4--Plane number
27 S-801id number
228 Plato's Mathematical Imagination
Figure 81 Figure 82
ARITHMETIC MEANS HARMONIC MEANS
INSERTED INSERTED
1 3/2 2 3 4 6 8 1 4/3 2 8/3 4 16/3 8
1 2 3 6 9 18 27 1 3/2 3 9/2 9 27/2 27

The insertion of harmonic and arithmetic means creates a double

bond (according to the other and according to a mixed relation of
same and other) between the terms.
Proclus' theorem explains why there need be no specific reference
to the geometric progression:
X : harmonic mean :: arithmetic mean : Y

In other words, the insertion of these two means produces a 4-term

geometric progression.

Figure 83
BOTH MEANS INSERTED
The inserted links reduce the distinction of the two initial series,
so that the lambda-representation can be replaced by a line
1 4/8 3/2 2 8/3 3 9/2 4 16/3 6 8 9 27/2 18 27

The connections are made tighter in Figure 84 by inserting links

in 9:8 ratio between the terms which are in the ratio of 4:3, so
that the entire series approximates to a continued geometric pro-
gression, and reproduces the structure of intervals of a diatonic
scale.
Figure 84
MATRIX FORM OF COMPUTATION OF THE SERIES
9/8 9/8 256/243 9/8 9/8 9/8 256/243
I 9/8 81/64 4/3 3/2 27/16 243/128 2
2 9/4 81/32 8/3 .3 27/8 243/64 4

4 9/2 81/16 16/3 6 27/4 243/32 8

8 9 81/8 32/3 12 27/2 243/16 16

16 18 81/4 64/3 24 27
 ALGEBRAIC METAPHOR 229
In Figure 84 the relation of doubling between rows of the matrix
suggests a way to simplify the intended calculation.
The 2: 1 ratio of the octave extends the concordant structure of
the initial scale through the compass of four octaves and a sixth
necessary to complete the symbolism of extension of the odd series
r to a term which is a cube.

Figure 85
TRADITIONAL COMPUTATION
To avoid dealing with fractions, commentators traditionally treated
each term as multiplied by 384 (= 3 X 128) or 192 (= ~ X 384).
The resultant progression is:
1- 384 19- 2304
2- 432 20- 2592
3- 486 21- 2916
4- 512 22- 3072
5- 576 23- 3456
6- 648 24- 3888
7- 729 25- 4096
8- 768 26- 4608
9- 864 27- 5184
10- 972 28- 5832
11- 1024 29- 6144
12- 1152 30- 6912
13- 1296 31- 7776
14- 1458 32- 8192
15- 1536 33- 9216
16- 1728 34- 10368
17- 1944
18- 2048 Total 105,113

Figure 86
TRADITIONAL COMPUTATION IN MATRIX FORM
The difficulty of computation is considerably reduced if these
figures are also constructed in a matrix of octave width, as follows:
384 432 486 512 576 648 729 768
768 864 972 1024 1152 1296 1458 1536
1536 1728 1944 2048 2304 2592 2916 3072
3072 3456 3888 4096 4608 5184 5832 6144
6144 6912 7776 8192 9216 10368
230 Plato's Mathematical Imagination

II. THE THEORY OF VISION

Timaeus 47
Timaeus 46: • There will now be little difficulty in understand-
ing all that concerns the formation of images in mirrors and any
smooth reflecting surface. As a result of this combination of the
two fires, inside and outside, and again as a consequence of the
formation, on each occasion, at the smooth surface, of a single
fire which is in various ways changed in form, all such reflections
necessarily occur, the fire belonging to the face (seen) coalescing,
on the smooth and bright surface, with the fire belonging to the
visual ray. Left appears right because reverse parts of the visual
current come into contact with reverse parts (of the light from
the face seen), contrary to the usual rule of impact. On the con-
trary, right appears right, and left left, when the visual light
changes sides in the act of coalescing with the light with which
it does coalesce; and this happens when the smooth surface of the
mirror, being curved upward at either side, throws the right part
of the visual current to the left, and the left to the right. The
same curvature turned lengthwise to the face makes the whole
appear upside down, throwing the lower part of the ray towards
the top and the upper part towards the bottom.
Timaeus 48: • And so at the moment we speak of [when the
soul, newly joined to the body, was upset by the violent, com-
municated motions, these sensations] causing for the time being
a strong and widespread commotion and joining with that per-
petually streaming current in stirring and violently shaking the
circuits of the soul, they completely hampered the revolution of
the Same by flowing counter to it and stopped it from going on
its way and governing; and they dislocated the revolution of the
Different. Accordingly, the intervals ... were twisted in all man-
ner of ways, and all possible infractions and deformations of the
circles were caused; so that they barely held together, and, though
they moved, their motion was unregulated, now reversed, now
side-long, now inverted. It was as when a man stands on his head,
resting it on the earth, and holds his feet aloft by thrusting them
against something; in such a case right and left both of the man
• Trans. Comford, Cosmology, pp. 148, 154, 155-56. Discussion of Timaeus 47
is preceded by Cornford's translation of Timaew 46, 48, because the latter supply
material indispensable for the interpretation of the theory of vision in Timaew 47.
 ALGEBRAIC METAPHOR. 281
and of the spectators appear reversed to the other party. The
same and similar effects are produced with great intensity in the
soul's revolutions; and when they meet with something outside
that falls under the Same or the Different, they speak of it as the
same as this or different from that contrary to the true facts, and
show themselves mistaken and foolish. . . .
Timaeus' account of the genesis of knowledge 19 gives the
impression that true insight is inevitable. For the observation
of the cosmic environment stabilizes in the soul those revolu-
tions to which the movements of the cosmos correspond. Two
factors intervene which prevent this automatic acquisition of
knowledge, and provide a physiological mechanics of error. The
first of these is the shock of external relations caused by union
with the body, which interrupts the natural progress of the
soul. 20 The second is the unreliability of the senses, which
makes possible misinterpretation of our observations, so that
they seem to confirm an incorrect mode of psychic revolution. 21
In the theory of optics, here cited to explain the mechanism of
vision, Plato is providing a theory of error as well. For the
mind, whether it be warped or straight, will seem to find a
reflection of itself in visible phenomena. In a tradition which
has always used analogies of vision and illumination (and the
history of which is usually traced by optical analogies, with the
study of optics often central in that history), this epistemological
extension of the laws of sight does not seem unnatural. The
dialogues abound in examples of lawgivers, statesmen, and
philosophers who seem to be the victims of some consistent
distortion of intellectual vision. These errors in empirical intel-
ligence parallel the analogous distortions of physical sight.
The interaction of the eye of an observer and a plane mirror
typically produces an image which is reversed, with left and
right transposed. 22 The distinction of same and other finds its
natural projection, in space and motion, in the differentiation
of left and right. Thus the circle of the other always revolves to
the left. Reversed mental vision would therefore transpose the
columns of matrices, resulting in a reversed apprehension of
same and other. 28 This is distinct from inversion, which trans-
232 Plato's Mathematical Imagination

poses the rows of matrices and results in a reversed apprehen-

sion of being and becoming. 24
Moreover, if the mirror is distorted in such a way as to be
concave from top to bottom, the image is inverted. History is
full of references to "inversions" of Platonism, in which the flux
of becoming is taken as the ultimate reality, and the forms
which emerge in that flux are merely its transitory modifica-
tions, appearing for the moment from an unreal, or less real,
shadowy realm of possibility. Plato refers in Epistle vii to the
ease with which anyone who chooses can invert the logos of the
man trying to describe forms, by centering on the inadequacies
of his symbols of description, which are themselves only images.25
Thus the warped mind of an inverted Platonist must seem to
perceive reality as appearance, and appearance as reality. How
the world looks from this point of view is well brought out in
the "account according to necessity," which follows the present
passage in the Timaeus. 26
If the mirror is concave from left to right, however, the object
is not inverted, and the relative location of left and right in the
object and its image correspond. 21 This perception is therefore
adequate and suggests that the perceiving mind does more than
simply "hold a mirror up to nature"; the trained mind must
be a corrective mirror in which reflection corrects the discrep-
ancies between objects and purely sensuous images. More spe-
cifically, there must be no inversion; reality must be known as
permanence, not change. Further, there must be no reversal;
the mind must not allow the vividness of sensory accompani-
ment to distract it from the fact that things are known only by
insight into their intelligible structures. Mathematical training
is particularly suitable for correcting the undue vividness of
appearances by accustoming the mind to look for structures
themselves.
A further distortion of objects which mirrors produce in their
images occurs when the mirror is slanted or oblique.28 In this
case, the real and apparent proportions have no metric analogy;
some dimensions are elongated by the glass, others foreshort-
ened. A philosophic system seen thus obliquely loses its coher-
 ALGEBRAIC METAPHOR 288
ence and symmetry and much of its plausibility. It is to such
oblique mental vision that Aristotle's paraphrases of the Platonic
position would probably have been attributed by Plato, as
doubtless would many more recent oblique interpretations of
the Platonic philosophy.
To grasp forms, the mind must be a blank substrate. But
the mind, jostled by the intrusive body to which it is attached,
lacks the complete absence of quality that characterizes pure
space.29 And as perceptions are stamped out through the senses
on the soul, the quality of the things perceived will be identified
with characteristics which are really contributions of the per-
ceiving mind.
This passage on optics, like the introductions of proportion
in the Republic and Gorgias, provides an insight into Plato's
analogical method of discussion, and an analogy for examining
the deviations of men who, on the basis of the same experience
as the Platonist's, have followed other philosophies.

Figures 87-93
THEORY OF VISION: DISTORTING EFFECTS OF MIRRORS

A modern reader can disregard . . the mechanics and physics of

Plato's visual ray theory of perception, in dealing with mirror dis-
Y' tortions, since the effects observed are entirely independent of the
f
L postulated mechanism which produces them.
In the following figures, terms in parentheses are not parts of the
matrix reflected, but indicate the spatial orientation of the observer
to the thing seen, hence the designations and interpretations that
he attributes to its spatial relations. In figures 87-93 the mirror
images are used to schematize other philosophies as transforma-
tions of Platonism.
1;"
If we substitute the Sophist matrix for the letters in Figure 87,
the effect of this simple reversal is (since the observer's orienta-
tion, determining the co-ordinates, remains the same) to give an
idealistic emphasis by suggesting that human artifacts and imi-
tations have a primary reality, while natural objects' and images
have a derivative one.
234 Plato's Mathematical Imagination

Figure 87
REVERSAL IN A PLANE MIRROR
MATRIX SEEN DIRECTLY
(Same) (Other)
(Being) A B
(Becoming) C D

Front
Object Image
MATRIX AS SEEN IN MIRROR

Matrix Image
(Same) (Other) (Same) (Other)
(Being) A B (Being) B A
(Becoming) C D (Becoming) D C

Figure 88

Front
Object Image
RECTIFICATION OF MIRROR
TO PREVENT REVERSAL
 ALGEBRAIC METAPHOR. 235
Probably one could extend the analogy of physical sight and
intellectual vision to take account of Plato's mention of the rectify..
ing mirror in this passage by pointing out that education in the
same way "rectifies" the mind's orientation toward objects in expe-
rience when we have stabilized and corrected the revolutions in
the soul.

Figure 89
EFFECT OF REVERSED INTELLECTUAL VISION
NORMAL INTERPRETATION

(Creations [Reality]) (Constructions [Appearance])

(Objects) Natural Objects Man-made Artifacts
(Images) Images of Such Objects Human Fine Arts

REVERSED INTERPRETATION OF EXPERIENCE

(Creations [Reality]) (Imitations [Appearance])
(Objects) Man-made Artifacts Natural Objects
(Images) Human Fine Arts Images of Such Objects

Note that the transposition of columns suggests, by its reorienta-

tion of terms, that natural objects may properly be explained as a
kind of secondary artifact, a sense of "object" later to become
central in Kant's philosophy.

Figure 90

Operation
INVERSION IN MIRROR

f
236 Plato's Mathematical Imagination

Figure 91
EFFECT OF INVERTED INTELLECTUAL VISION
A. LINEAR. IMAGES
Normal Interpretation Inverted Interpretation
(Being) (Being)
eternal forms representations
abstract hypotheses opinions
opinions abstract hypotheses
represen tations eternal forms
(Becoming) (Becoming)

B. PLANE IMAGES: REVERSAL AND INVERSION

NORMAL INTERPRETATION INVERTED AND REVERSED

Nature Artifacts Mimetic Works Shadows

Shadows Mimetic Works Artifacts Nature

Note that in Figure 91 the ontological status of the inverted seg·

lllents of the divided line here corresponds exactly with the philos-

Figure 92

EFFECT OF MIRROR AT OBLIQUE ANGLE

 ALGEBRAIC METAPHOR 287
ophy of Bergson, in which process is put forward as the primary
nature of reality.
In Figure 92 an image in which the continuity of the parts de-
pends on the metaphor of height is viewed as it appears turned on
its side, in a position where this continuity is no longer operative.
The result of such obliquity is to introduce a separation of entities,
terms~ and disciplines which in the original diagram were meant
to be shown as synoptically connected. Again, the divided line is
a peculiarly suitable illustration of this distortion.

Figure 98
EFFECT OF OBLIQUE INTELLECTUAL VISION
NORMAL INTERPRETATION
(Being)
Sciences philosophy
mathematics
Arts technology
poetry
(Becoming)
OBLIQUE INTERPRETATION
(Being)
Sciences: philosophy mathematics Arts: productive fine
(Becoming)

Note that the being-becoming axis ceases to interconnect the

terms. Compare Aristotle's classification of knowledge:

Arts Sciences

Fine Productive Mathematics Metaphysics

Physics
A part of the total Aristotelian scheme is here presented; It IS
intended to be that part which shows characteristically differ-
entiated disciplines replacing the Platonic scale, hence as a typically
oblique vision of Platonism. Aquinas and Kant have classifications
of the sciences that lend themselves to this same interpretation.
288 Plato's Mathematical Imagination

III. THE THEORY OF GEOMETRICAL ELEMENTS

Timaeus 51
Timaeus 53-56: • Such being their (the elements') nature at the
time when the ordering of the universe was taken in hand, the
god then began by giving them a distinct configuration by means
of shapes and numbers. That the god framed them with the great-
est possible perfection, which they had not before, must be taken,
above all, as a principle we constantly assert; what I must now
attempt to explain to you is the distinct formation of each and
their origin. The account will be unfamiliar; but you are
schooled in those branches of learning which my explanations re-
quire, and so will follow me. In the first place, then, it is of
course obvious to anyone that fire, earth, water, and air are bod-
ies; and all body has depth. Depth, moreover, must be bounded
by surface; and every surface that is rectilinear is composed of
triangles. Now all triangles are derived from two, each having
one right angle and the other angles acute. Of these triangles,
one has on either side the half of a right angle, the division of
which is determined by equal sides (the right-angled isosceles);
the other has unequal parts of a right. angle allotted to unequal
sides (the right-angled scalene). This we assume as the first be-
ginning of fire and of the other bodies, following the account
which combines likelihood with necessity; the principles that are
yet more remote than these are known to Heaven and to such
men as Heaven favours. Now, the question to be determined is
this: What are the most perfect bodies that can be constructed,
four in number, unlike one another, but such that some can be
generated out of one another by resolution? If we can hit upon
the answer to this, we have the truth concerning the generation
of earth and fire and of the bodies which stand as proportionals
between them. For we shall concede to no one that there are vis-
ible bodies more perfect than these, each corresponding to a
single type. We must do our best, then, to construct the four
types of body that are most perfect and declare that we have
grasped the constitution of these things sufficiently for our pur-
pose.
• Trans. Cornford, Cosmology, pp.198, 212, 2HJ-14, 215-16,217-18. Consideration
of Timaeus 51 is preceded by Cornford's translation of Timaeus 53-56 because
the latter has an important bearing on the theory of geometrical elements.
 ALGEBRAIC METAPHOR 239
Now, of the two triangles, the isosceles is of one type only; the
scalene, of an endless number. Of this unlimited multitude we
must choose the best, if we are to make a beginning on our own
principles. Accordingly, if anyone can tell us a better kind that
he has chosen for the construction of these bodies, his will be the
victory, not of an enemy, but of a friend. For ourselves, however,
we postulate as the best of these many triangles one kind, passing
over all the rest; that, namely, a pair of which compose the equi-
lateral triangle. The reason is too long a story; but if anyone
should put the matter to a test and discover that it is not so, the
prize is his with all good will. So much, then, for the choice of the
two triangles, of which the bodies of fire and of the rest have been
wrought; the one isosceles (the half-square), the other having the
greater side triple in square of the lesser (the half-equilateral).
We must now be more precise upon a point that was not clearly
enough stated earlier. It appeared as though all the four kinds
could pass through one another into one another; but this ap-
pearance is delusive; for the triangles we selected give rise to four
types, and whereas three are constructed out of the triangle with
unequal sides, the fourth alone is constructed out of the isosceles.
Hence it is not possible for all of them to pass into one another
by resolution, many of the small forming a few of the greater and
vice versa. But three of them can do this:' for these are all com-
posed of one triangle, and when the larger bodies are broken up,
several small ones will be formed of the same triangles, taking on
their proper figures; and again when "several of the smaller bodies
are dispersed into their triangles, the total number made up by
them will produce a single new figure of larger size, belonging
to a single body. So much for their passing into one another. The
next thing to explain is, what sort of figure each body has, and
the numbers that combine to compose it.
First will come the construction of the simplest and smallest
figure (the pyramid). Its element is the triangle whose hypotenuse
is dOllble of the shorter side in length. If a pair of such triangles
are put together by the diagonal, and this is done three times, the
diagonals and the shorter sides resting on the same point as a
centre, in this way a single equilateral triangle is formed of tri-
angles six in number. If four equilateral triangles are put to-
gether, their plane angles meeting in groups of three make a
single solid angle, namely the one (180 degrees) that comes next
240 Plato's Mathematical Imagination
after the most obtuse of plane angles. When four such angles are
produced, the simplest solid figure is formed, whose property is
to divide the whole circumference into equal and similar parts.
A second body (the octahedron) is composed of the same (ele-
mentary) triangles when they are combined in a set of eight equi-
lateral triangles, and yield a solid angle formed by four plane
angles. With the production of six such solid angles the second
body is complete.
The third body (the icosahedron) is composed of one hundred
and twenty of the elementary triangles fitted together, and of
twelve solid angles, each contained by five equilateral triangular
planes; and it has twenty faces which are equilateral triangles.
Here one of the two elements, having generated these bodies, had
done its part. But the isosceles triangle went on to generate the
fourth body, being put together in sets of four, with their right
angles meeting at the centre, thus forming a single equilateral
quadrangle.
Six such quadrangles, joined together, produced eight solid
angles, each composed of a set of three plane right angles. The
shape of the resulting body was cubical, having six quadrangular
equilateral planes as its faces. There still remained one construc-
tion, the fifth (the dodecahedron); and the god used it for the
whole, making a pattern of animal figures thereon.

At the outset, it is essential to note that Plato is not here

drawing his illustrations from a pure stereometry of the sort
described in Republic vii. sO Such a theoretical science of solid
geometry must properly proceed from different principles pe-
culiarly appropriate to three-dimensional mathematical objects.
Any reduction of solids to the simpler elements of plane geom-
etry or arithmetic is merely an operational aid to show the
natures of the solids themselves, as geometers engage in absurd
talk about plane figures when they talk of squaring and apply-
ing, as though they were studying combinatorial technique, and
trying to build something rather than to understand it. In a
proper stereometry, "volume" must be a central concept, and
"division" must be construed as a division of volumes.
The present passage, on the contrary, treats the subject with
a technique that is throughout (except when God's motives are
 ALGEBRAIC METAPHOR. 241
specified) combinatorial, and the relevant stereometrical theo-
rems proper must, for the most part, be read in from outside
the given text. The elements of plane geometry are used as the
principles of solids, in apparent opposition to the dictum in the
Republic that each mathematical science postulates its own
contrary principles. 31 In fact, in the references to the "numbers"
of which solids are composed, the analytic reduction has been
carried a step further, and the constituent plane elements are
themselves viewed as though they were the units of arithmetic.11
This strange treatment of volumes as planes is a mathematical
analogue of the treatment of things as radically disconnected by
the removal of all bonds of proportion, as they would be "in
the absence of God." 33 The principle of a physics of such un-
integrated parts must be shock or contact, as it is in the later
physics of Descartes. 34 This is the simplest external relation that
can hold between bodies which are in no way harmonized as
parts of a common organized whole. When applied to mathe-
matics, the analytical technique by which Plato represents or-
ganisms dissolved into constitutive material elements represents
solids as dissolved into mere abstract "sets of points of contact"
-the volume is shrunk to a bounding plane, the plane to an
infinitesimal unit of area.
Even this identification of elementary entities externally re-
lated by contact requires, in the flow of process, a postulate of
definiteness and order. Pure flow, of an Anaxagorean sort, is
radically unintelligible. The intelligibility of the outcome of
mechanical process, however, requires the postulate that there
be something definite in the flux. Plato indicated this by saying
that "God has marked out" the basic units, separate from each
other, but each retentive of its own spatial identity.sli
At this point, there is an obvious objection to the claim made
that this account of Plato's involves an atypical solid geometry.
The synthesis of these planes into solids results in four of the
five regular solids and presupposes the proper advance of
stereometry through the classification of solids as regular and
irregular. This seems, in fact, the most advanced point which
the science had attained in Plato's time, and such a develop-
242 Plato's Mathematical Imagination

ment as might have served as a model for the directed research

advocated in the Republic.s6
Though from God's point of view the results of the fitting
together of these solids do exemplify an application of pure
geometrical science, and possess unusual intellectual al1d aes-
thetic merit, the process of this synthesis seems not to be de-
veloped as a proper geometrical construction should. Another
reason for the appearance of the "regular" solids here is related
to mechanics rather than to stereometry. The geometer of solids
sees why these solids are best; but a Milesian engineer would
be the man to see why the exclusive fusion of planes into regular
solids is necessary.
If the basic concept of this physics is contact or shock of
corpuscles in a plenum, the basic concepts used to describe this
process for the system as a whole are equilibrium and disequilib-
rium. It is relevant at this point to recall that the ffi.echanical
equilibrium of volumes and densities is made the ordering
principle of the model displayed by Necessity in her personifica-
tion as goddess, in the Myth of Er. 81 The formation of solids
must respect this same principle. Since the series of corpuscular
shocks is incessant from all sides, only those figures capable of
being inscribed in a sphere will have the equilibrium necessary
to attain stability and escape disintegration. Any irregular pro-
jection or asymmetry in a combination of planes under these
conditions would, when hit, cause the planes to separate again.
The idea of the atomists, that atoms are of all shapes, and
equally cohesive regardless of shape, would have seemed to
Plato to defy mechanics and geometry. One must go to the
geometer of solids to find out how many and what sort are the
solids satisfying this condition of equilibrium, but it is a
mechanically, not geometrically, imposed condition.
A further confirmation of the fact that this account is mathe-
matically atypical is found in the use of the dodecahedron. s8
God again enters explicitly to insure that the universe will not
be lacking in perfection, but, as it included all living creatures
in the earlier account of creation, it will now comprehend all
of the "fairest" solid figures. God includes this dodecahedron by
 ALGEBRAIC METAPHOR 248
using it for the ornamentation of the heavens. Only in an ac-
count "according to necessity" could Plato recognize this sort
of decoration as a legitimate mode of "inclusion." In the Repub-
lic} the work of the scene painter is criticized because it is not
"included" in nature as real objects are. 3D But immediately
after that criticism, Plato goes on to a myth in which the pic-
torial qualities of a cosmic model give the observer his clue to
its mechanism, and in this myth omits the names of the planets
presumably because their use would give an intellectual em-
phasis not in keeping with the pictorial context. All of the
regular solids are "included" in the cosmos as they might be
included in a box of mathematical models. The sculptured re-
lief in the heavens is as much a "part" of the world as are the
molded figures of the atoms. Neither is, after all, explanatory of
the important organic aspects of cosmic structure.
The analysis by which volumes are reduced to elementary
plane triangles is dependent on the prior reduction of relations
and causality to external superficial contact. As a first step, the
solid approaches its component capable of such contacts, and
is reduced to a surface. Through bisection of the angles of plane
figures the surface can be progressively analyzed into triangles
until at last repetition of the division will only go on producing
smaller triangles of the same type. Such self-reproducing tri-
angles are the end-points of division; their properties remain
homogeneous, however often the division is repeated. The unit
of contact can be described as having certain geometric proper-
ties and ratios even if it is defined as infinitesimal, and hence
can never be actually constructed.40
Since the shapes of these smallest triangles cannot be used as
the surfaces of the stable solid figures in their uncombined
form, one can explain the existence of stable solids (which the
objects in our experience attest to be a fact) only by reversing
the process of analysis through which these elements were de-
rived. The ultimate point of contact will not serve as the ulti-
mate element of physics; presumably, then, some simple aggre-
gate or complex of these smallest elements does so.
This casts some light, perhaps, on a puzzling part of Plato's
244 Plato's Mathematical Imagination

procedure. For in the course of the demonstration that the ele-

mentary triangles can aggregate to form the bounding surfaces
of regular solids, the isosceles plane is reconstituted from six
(rather than two) of these smallest scalene elements. 41 One can
see from a diagram that this synthesis exactly reverses the
antecedent analytic construction; the lines of junction in this
figure are the lines of division resulting from analysis.
From this point on, there are really two tables of elements
of which chemical and physical theory must take account. On
the one hand, there are the elementary plane surfaces bound-
ing the regular solids; on the other, there are the infinitesimal
plane triangles of contact, the existence of which represents the
least postulate needed to give an intelligible account of flux.
Though the former of these elements are most closely connected
with observed differences in phenomena, the latter are prior in
theory. The dodecahedron is relegated to a decorative design
in the heavens just because there is no simple way to show its
bounding planes as aggregates of the primary elements.
The significant point, which these detailed observations are
intended to show more specifically, is that in" this passage the
method of Plato's geometry mirrors the method of the contex-
tual dialectic, and just as the emphasis of the dialectic on anal-
ysis into elements is "uncustomary," 42 so is that of the geometry.

Figure 94

V3
I :II
THE ATOMIC TRIANGLES
 ALGEBRAIC METAPHOR. 245
Figure 95

A E c
SYNTHESIS OF THE EQUILATERAL
MOLECULAR TRIANGLES

Pairs of elementary scalene triangles are here synthesized by

joining them "diagonally," with the diagonal repeated three times.
The question raised by this construction is why triangles 5 and 6,
for example, are not simply joined along BF, BD to form a molec-
ular isosceles triangle.

Figure 96

SYNTHESIS OF THE SQUARE

MOLECULAR PLANES
246 Plato's Mathematical Imagination
Figure 97

,-------------
, 0
,,
'-
\
, c'
\
, 0'
\
\
\
\
'-
\
\
\
\

A
DECOMPOSITION OF POLYGONS INTO TRIANGLES (Euclid vi. 20; see
I. L. Heiberg, ed., Euclidis Elementa, II [Leipzig, 1884], figure
p. 133, text of Corollary, p. 138, and compare Todhunter's
figure in Euclid's Elements, p. 200.)
In th~ figure, the two polygons are similar; each triangular sub-
division of the larger has two similar sides and the same angle as
the corresponding subdivision of the smaller. The theorem is that
"similar polygons have to one another double the ratios of their
corresponding sides"; since this has already been proven for similar
triangles, the construction showing that similar polygons can be
treated as sets of such triangles proves the more general theorem.
Euclid adds, as a corollary: "The same theorem can be demon-
strated for similar quadrilaterals, that their ratio is twice that of
their similar sides. And this has been demonstrated for triangles:
and so it holds universally that similar rectilinear figures are to one
another in double the ratio of their corresponding sides." This
suggests that Euclid thought of the division of his five-sided figure
as giving a quadrilateral plus a triangle; so that, for a figure of any
number of sides, the same type of construction will show it equal
to a sum of a quadrilateral and added triangles. Further sub-
division of triangles is not needed in this proof, and is not carried
out. Since Euclid vi. is Pythagorean, it must be some construction
like this one that Plato has in mind when he says that "all plane
figures can be analyzed into triangles."
 ALGEBRAIC METAPHOR 247
Figure 98

A
-------------,""'1 0 ~
/
~
~
~
/
~
~

~
~ .
,,0
I
;' " II
~ I
~ I'
~;' I~;'

B C
HOMOEOMEREITY OF THE
FIRST ATOMIC TRIANGLE
(Comford, Plato's Cosmol-
ogy, p. 233)

"The half-square ABC is divisible into two smaller half-squares,

ABO., BOC; and Plato does in fact so divide it when he constructs
the whole square face, ABeD, of the cube." (See next figure.)

Figure 99

I
,A
I
I
I
I
I
I
I
I
I
I
I
LI ~ __- ___
B C
HOMOEOMEREITY OF THE
SECOND ATOMIC TRIANGLE
(Cornford, Plato's Cosmol-
ogy, p. 234, and note)
248 Plato's Mathematical Imagination

"In the same way, the half-equilateral, ABC, can be subdivided

into smaller half-equilaterals by bisecting the angle at C and
dropping the perpendicular ODe It is actually so subdivided into
three elementary scalenes in Plato's figure [Figure 95]; and the sub-
division can be carried on ad infinitum. ...
"Since the triangles, not the solids, are Plato's 'elements,' this
meets Aristotle's objection that not every part of a pyramid or cube
is a pyramid or cube. 'Homoeomereity" was first clearly defined by
Anaxagoras, but Empedocles had no doubt already assumed that
every part of fire was fire."

Figure 100
TRANSMUTATION RATIOS OF THE ELEMENTARY SOLIDS
EARTH • WATER AIR. FIRE
EARTH 1:1
WATER 1:1 1:2~ 1:5
AIR 2~:1 1:1 1:2
FIRE 5:1 2:1 1:1
• Earth does not transmute.

It would be most natural for a physicist or even for a practitioner

of purely theoretical solid geometry to derive these ratios from
considerations of relative volume and planes of fracture, or topo-
logical deformation. But these ratios cannot be explained by such
a derivation. As the correspondence of this figure with the next
shows, these transmutation ratios are computed on the basis of
disintegration and recombination of the elemental, triangular plane
boundaries of the solid corpuscules.

Figure 101
RATIOS OF THE NUMBERS OF BOUNDING PLANES OF THE
ELEMENTARY SOLIDS
EARTH EARTHt(6) WATER (20) AIR (8) FIRE (4)
WATER 20:20 20:8 20:4
AIR 8:20 8:8 8:4
FIRE 4:20 4:8 4:4
t Ratios of earth are not given, since its atomic bounding planes differ from
those of the other elements.
 CHAPTER V

Mathematical Jokes: the Limits of

Mathematical Metaphor
Republic vii. 434D •
"But surely," said I, "if you should ever nurture in fact your
children whom you are now nurturing and educating in word,
you would not suffer them, I presume, to hold rule in the state,
and determine the greatest matters, being themselves as irrational
as the lines so called in geometry."

Republic ix. 580D t

"Very good," said I; "this, then, would be one of our proofs, but
examine this second one and see if there is anything in it."
"What is it?" "Since," said I, "corresponding to the three types
in the city, the soul also is tripartite, ... [the calculative part can
provide t], I think, another demonstration also."

Symposium, 190B §
ARISTOPHANES: Now the sexes were three, and such as I have
described them; because the sun, moon, and earth are three; and
the man was originally the child of the sun, the woman of the
earth, and the man-woman of the moon, which is made up of
sun and earth.
Statesman 266 II
STRANGER: Every tame and herding animal has now been
split up, with the exception of two species, for I hardly think
• Trans. Shorey, Republic, II, 209.
t Trans. Shorey, ibid., p. 871.
~ The text of this passage, including the version translated in brackets, is given
in note 8 to this chapter.
§ Trans. Jowett, Dialogues, I, 816.
HTrans. Jowett, Dialogues" IV, 298.
249
250 Plato's Mathematical Imagination

that dogs should be reckoned among gregarious animals.

YOUNG SOCRATES: Certainly not; but how shall we divide
the two remaining species?
STR.: There is a measure of difference which may be appro-
priately employed by you and Theaetetus, who are students of
geometry.
Y. SOC.: What is that?
STR.: The diameter; and again, the diameter of a diameter.
Y. soc.: What do you mean?
STR.: How does a man walk, but as a diameter whose power
is two feet? [AG, Fig. 102]
Y. SOC.: Just so.
STR.: And the power of the remaining kind, being the power
of twice two feet, may be said to be the diameter of our diameter.
[eF) Fig. 102]
Y. SOC.: Certainly, and now I think that I pretty nearly un-
derstand you.
STR.: In these divisions, Socrates, I descry what would make
another famous jest.
Y. SOC.: What is it?
STR.: Human beings have come out in the same class with the
freest and airiest of creation, and have been running a race with
them.

Though a concluding discussion of mathematical jokes may

seem on first view a capricious digression from a discussion of
mathematical imagery, it is in fact the only dialectically appro-
priate conclusion for such a study. The mathematical joke typi-
cally involves correction of an overpretentiousness in the use of
the mathematical metaphor and imagery, of the expectation of
more from the image than it should be expected to give. Plato's
mathematical humor serves as a corrective to various types of
misinterpretation that result if his images are taken too seri-
ously. In the dialogues themselves they occur in locations where
such safeguard is needed against overenthusiasm in the pursuit
of mathematical metaphor; and to treat them as the conclusion
of a serious study of the use of mathematical imagery is there-
fore one way of displaying the limits of such imagery, and of
bringing out the absurdity of overstepping those limits in inter-
 MATHEMATICAL JOKES: THE LIMITS OF MATHEMATICAL METAPHOR 251

pretation. In each case, the point of the humor is to be found

in the practice of treating a metaphor overliterally; this takes
the various forms of using geometrical differentiae in biology,
where they are inappropriate, of using a geometrical vocabulary
in politics, where it does not belong, and of postulating a causal
connection on the basis of an arithmetical resemblance which
is far from justifying such a postulate. The humor of these pas-
sages is lost on the modern reader because he does not recognize
the contemporary counterparts of the literal-minded Pythag-
orean lecturer on ethics of Plato's time against whom the shafts
of wit are directed. (In a twentieth-century transposition they
would be aimed at the pedantic statistician.)
Nothing has been so diversely or implausibly interpreted by
scholars as Plato's sense of humor. A legend of a "freakish" sense
of humor that expresses itself in deliberate unintelligibility is a
godsend to readers lacking the patience or insight to interpret
difficult dialogues and passages constructively. The entire Par-
menides and the mathematical images of the Republic and the
Menexenus have all been called specimens of Platonic jest; and
the Republic as a whole has been set aside as a piece of satire.
The present study has already tried to show that such hilarity
is not the intended tone of the Republic imagery.
In the dialogues Plato actually displays a very normal and
well-developed sense of humor. The best way of beginning a
discussion of the alleged grand jokes, such as those in the
Parmenides, would have been in connection with the various
passages in the dialogues which show this humor at work. There
is nowhere in them any evidence of a humor which takes its
point from some sudden obscurity, jargon, or unintelligibility,
as jokes often do in, say, Aristophanes and Lucian. The humor
is usually ironic, and aptly illustrative of Bergson's observation
that a joke "calls attention to the physical when it is the spiritual
[or intellectual] that should be in question." The hiccough of
Aristophanes in the Symposium and his entire subsequent
speech are excellent illustrations of this principle at work. So is
Socrates' indictment of the Athenian statesmen, that "they have
filled the city so full of harbors and public buildings that they
252 Plato's Mathematical Imagination

have left no room for temperance and justice." 1 Another hu-

morous device, the failure of a professional man to adapt nor-
mally, because in a novel situation he absent-mindedly, or
overliterally, applies the categories of his profession where they
do not belong, is evident in Plato's portrait of the vain and
pompous blunders of Prodicus as he ineptly tells Socrates the
etymologies of words instead of their meanings. 2
It seems quite reasonable to believe that only a strangely
incoherent mind could display such a normal sense of humor
in the use of these devices and then deviate to a different prin-
ciple of humor that is applied with what seems to the modern
reader (who has been told that the Parmenides is a great joke)
abnormal, tedious, and monstrous exaggeration.
Plato's intended pleasantries involving the vocabulary of
mathematics are abnormal for the modern reader in still an-
other way; they seem insipid, dull, and not funny. It is there-
fore hard to see how an author capable of a brilliant parody of
Aristophanes could have amused himself by flat punning. The
interpretation of Plato's use of mathematical metaphor, how-
ever, casts some new light on the principle of this mathematical
humor. The thesis will be defended that in Plato's own time
there were professional men who, given a mathematical term
in a discussion, insisted on transferring the conversation to
mathematical subjects. Some of Plato's own passages, in praise
of geometry, for example, might be misinterpreted as advocacy
of a second-rate Pythagorean pedagogy in which mathematical
pedantries were substituted for dialectical inquiry. By an aside
showing that he himself would recognize such a literal assimila-
tion of dialectic to mathematics as humorous, as the sort of
humor we recognize in the absent-minded and constant recur-
rence to his own preoccupation of the enthusiastic specialist,
Plato detaches himself from any misinterpretation in this direc-
tion. Such asides function as safeguards to perspective; mathe-
matical analogies do not always apply, and beyond a certain
point must not be taken literally. The humor of such asides
lies in their being exactly the remarks that an absent-minded
professional man (in this case, a Pythagorean-trained, not overly
 MATIlEMATICAL JOKES: THE LIMITS OF MATHEMATICAL METAPHOR 258
intelligent, lecturer on ethics) might advance in the discussion
as his serious, though inept, contribution. Our own Pythagoreans
are statisticians, not geometers; the change of vocabulary that
this entails prevents us from seeing, in Plato's illicit identifica-
tion of the mathematical and nonmathematical senses of a term
in one of his asides, any commentary on the statements of the
misplaced professionalism of a group of experts in our own
culture. (Perhaps the final conclusion of an educational statis-
tical study that I have seen, "between men and women there is
a significant difference," is a close approximation to the inad-
vertent punning of the not-very-bright Pythagorean in Plato's
time.)
Four such passages in Plato will be discussed: the use of a
geometrical pun in Republic vii to describe young men un-
trained in mathematics, the use of another such pun in Repub-
lic ix to differentiate logic from logistic, the explanation given
by Aristophanes in the Symposium of the number of sexes, and
the geometrical differentiation in the Statesman of bipeds and
quadrupeds.
The "normal" use of mathematical metaphor, in the Platonic
and Pythagorean traditions, depends on the conviction that
physical, moral, and mathematical objects have many of the
same properties. The etymologies of the mathematician's vocab-
ulary, deriving technical terms from biology, ethics, and tech-
nology, record such recognized similarities between quantities
and objects of other types. This conviction underlies Plato's
deliberately ambiguous use of mathematical terms in such a
way that their literal, technical meaning refers to a theorem or
construction, and their broader, metaphorical sense specifies
the intended interpretation of that construction. 8 There is an-
other side, however, to this use of a common name as the sign
of an observed common nature. In some cases a discrepancy of
nature holds between quantities and other entities designated
by a single term. This happens whenever the similarity that
the common name reflects is fortuitous and nonessential, but
the discrepancy between the classes named is not. One safeguard
against the Sophistical assumption that in finding a common
254 Plato's Mathematical Imagination

name one has discovered a common nature is the right use

of humor as a corrective of conceit. 4 An overliteral faith in
verbal technical devices is corrected by displaying a case in
lvhich the result of such literal-mindedness is a patent ab-
surdity.
As with other sorts of Platonic imagery, the interpreter who
finds considerable abstract aptness in the choice and location of
mathematical wit in the text should not commit himself as to
the extent to which these reflect the author's clear-cut and
conscious plan. The author presumably has a feeling that this
aside fits well here; the interpreter can recover only objective,
abstract, intellectual relationships that are relevant, but not
quite equivalent, to this intuitive feeling of aptness. Since an
interpretation can restate felt imaginative affinities only as
abstract intellectual connections, the reader who expects an in·
tuitive and direct presentation from the interpreter is always
disappointed and incredulous; such an abstract analysis is not
what we ourselves experience when we are writing creatively,
nor is it plausible to suppose that in this respect Plato differs
much from ourselves. In the case of humor, this difference be-
tween intuitive appreciation and analytical interpretation is
peculiarly apparent.
The boys who grow up "as irrational as the lines so called in
geometry," in Republic vii, have close affinity to the antecedent
discussion of the mathematical curriculum. 5 The pun on the
irrationality of a man and a line invites an illicit identification
of education with geometrical construction. The context of the
geometrical line is purely quantitative; its construction as oper-
ation is external and mechanical. But the reasoning capacity
of the human mind cannot be elicited by the analogous external
application of courses in geometry. The pun points out that
education cannot be conceived as a kind of engineering, nor
geometry as magic; it does this by projecting the absent-minded
professionalism of the pedantic geometer to the point at which
we find him describing human nature and training with the
concepts and operations of geometry.6 By this aside, Socrates
dissociates his own intention in the proposed mathematical
t
i
j
MATHEMATICAL JOKES: THE LIMITS OF MATHEMATICAL METAPHOR 255

curriculum from that of a misinterpreter who might think its

purpose were to show all life, not under the aspect of reason,
but as composed of the elements and operations of geometry.
(In Book vii, Socrates himself uses a similar analogy of reason
I to the rational line, not to suggest a literal identification 7 but to

~ bring out a relevant resemblance between them.)

In Republic ix, there is a humorous effect in Glaucon's be-
wilderment by Socrates' lightning calculation which helps pro-
vide a clue to the intention of the earlier mild pun on logic
and logistic. 8 The relevant aspect of Socrates' "marvellous com-
putation" is the summarizing diagram he envisions, which re-
tains a clear relevance to his intended interpretation but which
seems to lose any such evident relevance when it is replaced by
a single arithmetical number representing the product of its
dimensions. 9 The humor here gains in point if we recognize
that the number, rather than the illuminating diagram from
which it was computed, would have been fastened on and pre-
sented as a literal solution of the problem by the hypothetical
Pythagorean-trained lecturer on ethics of Plato's time-just as
our modern, statistically trained moralist might insist, in all
seriousness, that the typical American family had 1% children.
This gives a clue to the humor Plato presumably saw in the
earlier mild pleasantry in the same book, cited here. Socrates
has remarked, that "since there are three parts of the soul, [the
calculative] can show, by another demonstration .. ." that the
tyrant's life is not a happy one. The humor arises from the
suggested dependence of rational demonstration on the appar-
ently fortuitous presence of the definite number, three, in the
subject matter. Here, in anticipation, Socrates is dissociating
himself from any assumption that the logical faculty is identical
with the logistica1. 10 In any use of mathematical imagery, there
is always latent the danger of this confusion. Confronted with
a complex situation, of which we have grasped several externally
related parts, reason, through human inertia, is tempted to
operate with those parts abstractly, to indulge in computation
with these externally related units, rather than to try the more
difficult task of seeing the integration of these parts in some
256 Plato's Mathematical Imagination

larger, more intelligible whole. The logistical approach loses

sight of the continuity and interdependence of the "parts,"
which it treats as the "units" of its computation.
Plato's aside underscores an ambiguity in the faculties and
techniques of "calculation." The humor lies in the implied
identity of counting and reasoning, of enumerating and under-
standing. The similarity of names leads to a silly result if we
take it as showing an identity of things.
There is also latent in this pun an element of irony: experi-
ence does not respect the preferences of the logician by keeping
itself free from confusion. The neatness and ease of manipula-
tion of the abstract schematism does not correspond to an equal
neatness and ease of control of immediate, experienced fact. If
it did so, thinking and calculating might come much closer to
being really, not just nominally, the same.
There are several reasons for restoring 1:0 AOyL(rrLXOV to the
text. Apelt's arguments (see Platons Staat, p. 525) on philological
grounds seem sound. Further, it is hard to see how the phrase
could have come in as a corruption. (One can understand the
corrupted version found in one manuscript, which lists all three
of the parts of the soul at this point.) There is also a dialectical
reason, already given, that seems equally cogent. This aside, or
one like it, is needed to anticipate and protect against a misin-
terpretation of Socrates' later calculations (a type of misinter-
pretation into which the editors who have bracketed this aside
have almost uniformly fallen). Further, this remark helps to
make intelligible the cause of Glaucon's later bewilderment.
Aristophanes' clownish theory that the s~xes are three be-
cause the sun, moon, and earth are three depends for its humor
on the familiar Aristophanic comic device of non sequitur. The
passage is important as an instance of mathematical humor both
because the error it represents is basically mathematical, a11d
because the sort of mathematical reasoning that Plato finds so
Aristophanic tends to be attributed to Plato himself by his later
expounders and admirers. 11 The notion that sets have the same
cardinal numbers because of some affinity in essential nature
becomes a principle of the microcosm-macrocosm analogy un-
 MATHEMATICAL JOKES: THE LIMITS OF MATHEMATICAL METAPHOR 257

derlying alchemy, astrology, homoeopathic medicine, numerol-

ogy, and their allied branches of speculation. In all of these
studies, passages of Plato, particularly the mathematical ones,
are cited as authority. For Plato, 110wever, though similar things
may have similar arithmetical properties, enumeration is no
guide to demonstrating essential similarity. Even the substitu-
tion of a number for a geometric diagram is accolllpanied by a
caution to the reader, since the important aspect of the diagram
does not carryover to its number.
The joke about the difference between a biped and a quad-
ruped in Statesman 266A (that the former equals the diagonal
of a diagonal of the latter) is really a joke. Its humor lies in the
unnecessary injection of l1igher mathematics into a biological
discussion to accomplish a simple differentiation of species.
This is completely compatible with Plato's view, expressed else-
where, that mathematical imagery tends to give such clear-cut
schematic pictures of a confused reality that there is something
ironic in the juxtaposition. It is also an excellent illustration
of what I have called the principle of aesthetic economy in
Plato's construction of mathematical imagery: the ingredients
of this joke are derived from the example of definition provided
by Young Socrates and Theaetetus in the first dialogue of this
trilogy.12 The phrasing hinges on just such a geometrical repre-
sentation of numbers as Theaetetus described, with the incom-
nlensurable given its full prominence. The intrusion of this
example gives some clue to Glaucon's conviction, in Republic
vii, that "only a few of the mathematicians he has known are
good reasoners," 18 and constitutes an oblique criticism by the
Stranger of the poor showing made earlier by Theaetetus in his
conversation with Socrates.
It is interesting to note here the similarity of the diagram
illustrating the present passage (see Fig. 102) to the one used
in the discussion with the slave boy in the Meno. 14 This simi-
larity seems to confirm the assertion that in his choice of the-
orems, Socrates was trying to find a problem which, to the
dilettante MenD, would have the appearance of higher mathe-
matics.
258 Plato's Mathematical Imagination

Figure 102

C
B..-----...

A~---- ... E

F
"DIAGONALS AND THEIR DIAGONALS"
(Statesman)

,;1BCD is a unit square; its diagonal, AC == the square root of two

l'he square on this diagonal, ACEF has a diagonal equal to the
J

square root of 2AC2 = 2. CF is thus the diagonal of the (square of

the) diagonal AC. The next figure will show that in the present
context, as applied to the locomotion of animals, this interpretation
of the "diagonal of a diagonal" makes sense.
Aristotle's whole analysis of locomotion (D'e incessu animalium)
applies what, from Plato's allusion here, must have been a tech-
nique devised in the Academy. The relation of the legs and the
ground is treated geometrically, in analogy to properties of right
and isosceles triangles. In Figure 103 the relative lengths of bases of
two such triangles are shown, one representing a one-step advance
by a biped of the right foot from R 1 to R 2 (where the diagonal
resulting from the advance of the right foot lengthens the base to
the diagonal of a unit square), the other a one-step advance by a
quadruped, where the diagonal advance of the right fore and left
hind feet (right fore from RF1 to RF2' left hind from LH1 to LH2)
lengthens the base to the diagonal of two squares with sides equal
to square root of two (= line AC). This figure and the Platonic
passage should be compared with Aristotle's analysis.
 I
r
MATHEMATICAL JOKES: THE LIMITS OF MATHEMATICAL METAPHOR 259

Figure 103

~ ... c
,.----------- R 2

L A
Biped Quadruped
THE LOCOMOTION OF ANIMALS

f·
 APPENDIX A

Effect of Context on Three Platonic Images of Cycle

IT HAS been suggested that the significance of the peculiar en-
gineering of Poseidon in planning Atlantis is deliberately intro-
duced as a parody of the other images of concentric systems,
one of which is prominent in each of the two preceding dia-
logues of the Republic-Timaeus-Critias trilogy. A comparison
of these three sets of circles seems to show that the construction
of Poseidon lacks the perfection proper either to an ensouled
cosmos bound by bonds of friendship and proportion, or to a
world-machine in which bonds of adamant alloy constrain the
parts in such a way that by their balance cosmic morality and
justice are preserved. The sole perfection of Poseidon's con-
struction is its mechanical regularity; being a god, he describes
his circular canals with ease.1 It is also appropriate that a system
of canals, rather than of celestial motions, be the product of the
deity who is the patron god of the sea.
If this contrast were·deliberate, we should expect some viola-
tion, in this circular system, both of the perfection of propor-
tion evident in the world-soul, and the perfection of equilib-
rium shown in the world-machine. We should expect also a
deliberate avoidance of the ratios elsewhere described as "con-
cordant," which are used to describe the structure of planned~
organic, internal relations.
In fact, the dimensions of the canals flatly violate both prin-
ciples, and avoid any aspect of proportion. Their progressive
sizes are ordinal, and although each pair of circles of land and
water is equal, they form no systematic balance in combination.
260
 APPENDIX A 261

Figure 104
LACK OF BALANCE IN POSEIDON'S ENGINEERING

WATER LAND WATER LAND WATER CENTRAL ISLAND

2 2 5

5 5 6 (5+1)

These numbers show no such symmetrical balance as those of the

astronomical model; for example, the ratio of circles of land to
those of water is 5:6; total land to water is 10:6; island to circles
is 5: 11.

The dominant ratio of the system is 6: 5, whether we compare

the outer circles to the central island, or the circles of land and
water with one another.
The accentuation of fives and sixes throughout the arith-
metical details of the context directs attention to these more
outstanding aspects of imbalance, away from the fact that a
purely ordinal, logistical technique has actually introduced an
adumbrated harmonic proportion among the paired sets of
circles. The pairing of circles is significant only to the balanced
mechanical structure, where each paired set is equal to the
others; and the proportions are significant in giving the struc-
ture of the system with each circle treated separately. However,
the system of circles here, treated in the way which brings out
harmony elsewhere, yields the series 5,1,1,2,2,3,3; while equat-
ing the pairs, which gives an effect of balance elsewhere, cannot
be carried out (because of the central island of radius 5), and,
so far as it can be constructed, gives the sets 2,4,6, as mentioned
above. 2
A program of engineering based on power without enough
insight into the combination of harmony and cycle which is
the basic metaphor of social construction elsewhere in the text,
results in precisely the lack of rationale and efficiency which the
institutions of Atlantis on a large scale exemplify.·
262 Plato's Mathematical Imagination

The remarkable property of Platonic mathematical imagery

,
I
,
in its tight integration with dialectical context completely defies
either purely mathematical or symbolist interpretation. The
relevant properties of a mathematical illustration cannot be
determined simply by an examination of the purely mathemati-
cal properties of the image, for the relevance of these properties
to the function of the illustration as a whole is differential and
variable. Nor can an interpretation by a dictionary of symbol-
ism in which each figure is correlated with a single associated
concept ever do justice to the variability of Plato's actual choice
and use of these figures.

Figure 105
PROPERTIES OF THE IMAGE OF CONCENTRIC CIRCLES IN
THREE TYPES OF CONTEXT
PROPERTY REPUBLIC TIMAEUS CRITIAS
Creator Ananke God Poseidon
Structure Balance Proportion None
(evidence of
rationale)
Material Adamant World-soul Earth &: water
Source of Religious Calculation 8c Recorded history
account vision science
Purpose Education Creation Self-defense
Context Myth, in pure Empirical Historical legend
dialectic science

A convincing proof of this point is the use made of the same

basic geometric figure-that of the set of
concentric circles-in
each dialogue of the Republic-Timaeus-Critias trilogy. In each
case, we can say that the image symbolizes the efficacy of some
cause, introducing a spatial or temporal order. The type of
order introduced and the nature of the cause are not, however,
specified by the symbolism or the geometry of concentric circu..
larity. Neither is there any single varying mathematical prop-
erty which accounts for variations in the significance of these
images. In one case, ordinal sums and products of the widths of
the individual circles are the mathematical property central· to
 APPENDIX A 263

interpretation; in another, the ratios are the crucial property;

in a third case, merely the size of the sum of widths is centrally
relevant. As in the description in the Statesman of the role of
descriptive measure in the arts, so in the art of interpreting
Plato's dialogues recognition of a total function must provide
the norm in terms of which we select those quantitative proper-
ties which are functionally relevant as the ones to be descrip-
tively presented and measured.
 APPENDIX B

Symbolism: the Significance of the Specific Figures

Chosen as Illustrations from Pure Mathematics

THOUGH IT has been suggested that in illustrating a method,

any literal theorem from pure mathematics which involves an
analogous process of reasoning will serve the desired function,
so that the peculiar interdependence of imagery and context
conspicuous in o~her cases is at a minimum, there is a tendency
for certain figures to reappear as the content of such pure illus-

Figure 106
LATER DIAGRAMS OF CLASS-INCLUSION: EULER, LEIBNIZ
Euler

HALL VIRTUE IS TEACHABLE" "NO VIRTUE IS TEACHABLE"

Leibniz

"TEACHABLE VIRTUE (EXISTS)" "NOT-(TEACHABLE VIRTUE EXISTS)"

264
 APPENDIX B 265
In the usual current schematizing of class-relations, we follow
Euler in representing all classes by the same plane figure, the circle,
then schematize class-relations by the inscription or non-inscription
of circles in each other.
Leibniz, however, whose appreciation of the value of schemata
for logic led him to propose a philosophic language in which the
propositions could be analyzed from the schematic relations of
constituent primary ideographic symbols, did not introduce Euler's
restriction to similar plane figures. For example, in his ULettre sur
la characteristique" (Couturat, Opuscules et fragments inedits de
Leibniz, Paris, 1903, pp. 29 ff.), Leibniz denotes moral qualities by
distinctive geometric figures, their inclusion or coexistence by in-
scription or superimposition. His sign for God in this language is
a triangle inscribed in a square, inscribed in a circle.
Thus the Platonic inscription figure has recurred independently
in the history of logic as a diagram signifying either class-relations
or coexistence of attributes in a single individual or class, and the
suggestion that this may explain Plato's selection of it here cannot
be dismissed as a merely fanciful positing of a resemblance which
no one else would recognize.

tration. This suggests that the choice of these figures for repeti-
tion was not purely coincidental.1
Some comment has already been made on the aptness with
which a diagram of inscription gives intuitive illustration to
a more general methodological problem of class-inclusion. In
fact, as was noted, this aptness is so great that analogous prob-
lems of class-relation are studied at the present time with the
aid of imagery of inscription.
One may therefore speculate on whether some equal intui-
tive aptness does not underlie the fact that a diagram based on
the square of a diagonal appears several times independently in
Plato's text, and is referred to several times more. In the Meno,
this figure is used to demonstrate that knowledge is innate and
can be elicited by questioning even from a young boy with no
~
mathematical education, even when the knowledge in question
lies on the advanced frontier of contemporary science. In the
I Theaetetus, this diagram underlies the construction by which

r
I
266 Plato's Mathematical Imagination
I ,
Figure 107
SYMBOLISM: FIGURES INVOLVING INCOMMENSURABILITY
IN PLATO'S MATHEMATICAL IMAGERY

~
Q7
B. Meno (2) c. Theaetetus
A. Meno (I)

I , , , , , , , ,

D. Critias E. Laws F. Republic (1)

G. Republic (2) H. Timaeus (I) I. Tirnaeus (2)

J. Timaeus (3) K. Republic (3) L. Statesman

M. Republic (4) 'ott

j
 APPENDIX B 267
Figure 107 is an attempt to bring together the various references
to the side and diagonal and to the incommensurable, to show their
frequency, and also to show in certain cases (for instance A~ B~ and
L) their similarity of geometric form.
The reader will recognize A as the final construction in Socrates'
discussion with the slave boy in the Meno (Fig. 9, Chap. I, Sec. 2).
B is one of the figures suggested to illustrate the example of the
method of hypothesis from the Meno (Fig. 14, Chap. I, Sec. 3).
C is the diagram from Euclid x.9, the theorem of Theaetetus
(Fig. 17, Chap. I, Sec. 4). D once more represents the Atlantean
canal net; in this context, one notices that the communication
canals superimposed on the irrigation grid produce regions embody-
ing all sorts of incommensurability in the relations of their bound-
ary canals (Fig. 19, Chap. II, Sec. 1). E is a pictogram to illustrate
the study of measure and incommensurability in the educational
program of the Laws (Chap. II, Sec. 2). In F, the square and its
diagonal are the examples given of the kinds of objects mathe-
maticians study in Republic vi. G is the Pythagorean genetic symbol
basic to the Nuptial Number in Republic viii; it will he recalled
that Plato's construction of this involved the squaring of an irra-
tional (Fig. 53, Chap. II~, Sec. 6). H illustrates the description of
the motion of the other in the Timaeus as "along the diagonal"
(that is, the diagonal of the rectangle formed by extensions of the
tropics and vertical tangents to the equator). I is the atomic triangle
from the Timaeus element theory, the longer side of which equals
square root of three (Fig. 94, I, Chap. IV, Sec. 3). ] represents the
synthesis of molecular triangles in progress, with two of the atomic
triangles joined "diagonally," that is, along their hypotenuses
(Fig. 95, Chap. IV, Sec. 3). K represents the allusion in Republic vii
to young men who grow up incommensurate with their responsi-
bilities, the allusion to side and diagonal balancing that of Re-
public vi (see Chap. V). L represents the joke in the Statesman
about the locomotion of bipeds and quadrupeds (Fig. 102, Chap. V).
M shows the construction of the Divided Line by section in mean
and extreme ratio, as in Euclid ii. 11, showing stages of the con-
struction.
The final diagram is Figure 43, in Chapter III, Section 4 of this
study; the construction is shown in Figure 108, of this Appendix.

lr
268 Plato's Mathematical Imagination

roots and lengths have been defined and classified, in the man-
ner in which Theaetetus wants to divide and classify opinion
and knowledge. In the Statesman, the diagram is made the basis
of a joke, the point of which is an illicit substitution of a mathe-
matical for a biological differentia. In its context this probably
has some critical reference to the way in which Theaetetus and
Young Socrates have responded in the previous conversations
recorded in the Sophist and Theaetetus.
In each of these passages, the context involves a discussion of
the relation of two kinds of knowledge. The Meno passage
brings out the difference between recollection and experience
as sources.., of insight; the Theaetetus throughout proposes
eqllivalences of opinion and knowledge which crucial cases,
usually of mathematical knowledge, show to be indefensible;
the Statesman joke involves substituting an easily understood
mathematical property for the more relevant property based
on right observation.
The consensus of these contextual passages is that knowledge
and opinion are, in some puzzling sense, mutually "incommen-
surable." A mere addition of experiences is so far from an ade-
quate account of knowledge that the hypothesis of a previous
existence in which learning took place seems more plausible.
On the other hand, the assumption that biology, an empirical
science, is to be best formulated in the theorems and concepts
of pure geometry (an assumption which some of the Pythag-
orean contemporaries of Plato no doubt seriously made) is so
ridiculous as to be a "joke."
The suggestion of the diagram is that an alternative unit or
technique of measure would permit an exact description of this
baflling relatiollship, which would resolve or circumvent the
impasse analogous to that of trying to find a linear unit of
measure which wotlld resolve the apparent incommensurability
of the diagonal and side of a square. The later stress on the
study of two- and three-dimensional problems in incommen-
surables is directed toward the better citizenship of men who
can see the interconnection and relevance of social institutions
and personal choices, as an uneducated man could not. True,
 APPENDIX B 269
the immediate applications of this education are to he to tech-
niques of social implementation (in commerce, public works,
etc.); but their importance is based on the assumption that their
influence will be far more pervasive than this limitation would
imply.
An alternative image expressing the relation of knowledge
and opinion is the "divided line" in the Republic) in which the
various segments (if established by a series of sections in ex-
treme and mean ratio) 2 are proportionate to one another,
though linearly incommensurable. Here the impasse reflected
by the alternative image of incommensurable diagonal and side
appears to have reached some contextual resolution, compara-
ble to the construction by which the incompatible lines of the
other figures are discovered to be commensurably and propor-
tionately related in square.
I suggest that it is the presence or the absence of an agree-
ment about the theory of ideas which differentiates the contexts
in which the relation of disparate levels of knowledge is pre-
sented as a simple proportion, and those in which it is presented
as a geometrical impasse. In the absence of this theory, which
provides the organizing insight of the image in the Republic,
Socrates can explain knowledge to Meno only by the postulate
of recollection; Theaetetus can in no way explain knowledge
to Socrates by following the method previously used in defining
roots and magnitudes; and the Stranger in the Statesman sug-
gests that one could not find any principle for discriminating
relevant and irrelevant subdivisions of a class.
If the objects of experience and opinion receive their char-
acter by participation in the forms, which are the proper objects
of knowledge, this discrepancy in nature may be recognized
without destroying the possibility of our making valid state-
ments about the mutual relevance and connection of the two
types of objects. This appears to be the dialectical equivalent
of the geometrical construction showing that lines linearly in-
commensurable may be compared and have their exact relation-
ship defined when they are combined as side and diagonal of
a square.
270 Plato's Mathematical Imagination
These considerations suggest that Plato's repeated choice of
this specific figure may have been guided by a felt appropriate-
ness of its intuitive structure to symbolize the dialectical im-
passe and its resolution attendant on the demonstration of the
"incommensurability" of knowledge and opinion.
What at first sight appears to run counter to this thesis is
that the imagery of diagonal and side actually appears in Book
vi and Book vii of the Republic, where the rule just suggested
would seem to exclude it. As if by open and close brackets, the
divided line is prefaced by an explanation that geometers study
"the side and diagonal as such/' and is followed by the joke
about boys and irrational lines (of which side and diagonal of
a unit square are the standard symbolic illustration). But, on
closer study, it will be noted that the geometer is not acquiring
knowledge in its highest form (in fact, Glaucon remarks later

Figure 108

B c
K
G H

F A E o
CONSTRUCTION OF THE DIVIDED LINE BY DIVISIONS IN MEAN
AND EXTREME RATIO
(Method of Euclid II. 11)

AB is the given line. The square ABeD is constructed, and AD

is bisected at E by a line drawn from B. AD is extended, and EF
is marked off equal to EB. The square AFGH is completed, with
 APPENDIX B 271
H located on AB so that AB : AH:: AH: HB. AI, equal to HB,
is laid off on AB; and HK, equal to lH, on HB. In this way AH and
HB are divided in the same ratio as AB.
Here the incommensurables, instead of presenting a problem,
have been added to produce a rational result; with the theory of
ideas in the background, the dissimilarity of knowledge and opinion
(symbolized as an incommensurability) is presented in a context
which recognizes the possibility of presenting their exact relation
in an analogical statement. The problem arises when the attempt
is made to find some element or unit which can serve as the least
common measure.

that their studies have not generally made the mathematicians

he has known better philosophers), and that the risk of having
boys grow up "incommensurate with their responsibilities" (as
Comford translates the passage) is run only if we neglect their
philosophic education.
If we then modify the statement already made about the
theory of forms to include the case in which there is a tension
between the relation of mathematics and dialectic (their rela-
tion is, of course, exactly analogous to that -of knowledge and
opinion), we can also explain this use of side and diagonal
imagery on either side of the diagram which sets up a propor-
tionate relation among the irrationall,ines that represent knowl-
edge of various kinds. Further, the idea of this symbol as mark-
ing a delimitation between an exact theory of knowledge and
contextual discussion of inquiry and instruction, processes in
which that theory may be misapplied, adds something to our
understanding of the Republic.

,.
i
 Notes
CHAPTER. I

1. Euthyphro 12.
2. Meno 82, 89.
3. Phaedo 104.
4. Theaetetus 147.
5. Charmides 166.
6. Statesman 259 fI.
7. Euthydemus 290
8. Gorgias 454.
9. Protagoras 356.
10. Theaetetus 196.
II. Symposium 190C.
12. Meno 74.
IS. Timaeus 17A.
14. Phaedo 104 fI.
15. Euthyphro 7D.
16. The existence of the multiplication table, as it must have been set
up with alphabetic number notation, would be no mean demonstration
of the possible extent, synthetic power, and pedagogical value of a matrix.
Inevitably, the difficulties of operating this notation would lead to an
interest in computational shortcuts; and in the Pythagorean pythmen and
the reported Platonic use of ten as modulus, there is evidence of an
examination of number theory along lines which would make such sim-
plification possible. The simplest aid to computation would of course be
a mechanical abacus, with some principle of positional pebble arrange-
ment. A vase painting shows such an abacus in use, and a ruled table
discovered by archaeologists has been thought to be an actual abacus
(though opinion is divided, some scholars explaining it as a game-board).
The allusions in Plato which compare computation to the game of
draughts seem to presuppose familiarity with some such computer's aid,
which would in fact resemble the arrangement of lines and counters, and
the regulated moves, of a 5 times 5 game·board. From the frequent asso-
ciation of priests and temples with banking and treasury functions, one
may plausibly infer that the abacus would be a device with which a
religious Greek was familiar. At any rate, the special tables of the money-
changers were a familiar feature of the market place. See Apology 17C. 8,
278
274 NOTES FOR. PAGES 19 TO 34

and Burnet's note, Plato's "Euthyphro," tlApology of Socrates," and

"Crito," ed. with notes by John Burnet (Oxford, 1924), p. 71. Sir Thomas
L. Heath, A History of Greek Mathematics, 2 vols. (Oxford, 1921), gives
citations (I, 46-52) which show that the use of the abacus in calculation
was common in Plato's time. See Fig. 1.
17. The Dialogues of Plato, trans. B. Jowett (3rd ed.; London, 1892),
1,298.
18. This is clear from the description of its eleven lines, each of which
is marked with the initial letter of a unit of currency.
19. Phaedrus 275.
20. Figure numbers and letters in parentheses refer to the scholia figures
given just after the text. These figures and references have been added to
jowett's translation; jowett's original figure is given in Fig. 13.
21. Meno 85D.
22. Professor Greene has written concerning this proposed emendation
of the final scholion figure "You are at liberty if you wish to quote me
as saying that the printed figure and letters [in Scholia Platonica] correctly
represent the tradition of the scholion in question, but that I concur with
your emendation... ." Mr. Daniel M. Dribin, of Arlington, Va., a student
of the history of mathematics, also concurs: "Your emendations of the
figure used by the Platonic scholiast are quite correct.•.. The figure
itself is Pythagorean, if Greek at all, and is typical of empirical proofs
of [Euclid] I. 47, which could easily be suggested by geometric designs in
tile decorations, panelling, etc. . . ." A further discussion of proofs of
Euclid I. 47 of this type is given in George J. Allman, Greek Geometry
from Thales to Euclid (Dublin and London, 1889), pp. 29-31. This ma-
terial is of particular interest in connection with Plato's "combinatorial"
approach to stereometry in the theory of elements in the Timaeus, dis-
cussed in Chap. IV, Sec. 3, following.
23. A. Benecke, Ueber die geometrische Hypothesis in Platons Menon"
It

(Elbing, 1867); J. Gow, A Short History of Greek Mathematics (Cam-

bridge, 1884), p. 175, notes 2 and 3.
24. Heath, History, I, 300.
25. Sir Thomas L. Heath, The Thirteen Books of Euclid's Elements,
3 vols. (2nd ed.; Cambridge, 1926), III, 29-31.
26. In connection with this reference to figures already drawn, see in
the text of Fig. 14 the suggestion that Socrates may have drawn a circle
earlier in the discussion.
27. C. Demme, Die Hypothesis in Platons "Menon" (Dresden, 1888);
A. E. Taylor: Plato: the Man and His Work, new ed. (New York, 1946),
p. 138, n. 3.
28. See I. Todhunter, The "Elements" of Euclid, Books i-vi and xi and
xii (London: Everyman's Library, 1933), figure p. 108.

i~
..
 NOTES FOR PAGES 40 TO 49 275
29. See Appendix B.
30. Eva Sachs, De Theaeteto Atheniensi Mathematico, diss., Berlin,
1914; Heath, History, I, 209-12 ("Theaetetus"), 322-35 (UEudoxus").
31. Aristotle, Prior Analytics, 41a. 26--27; Heath, History, I, 91.
32. Heath, History, I, 202-9 (UTheodorus"). The present discussion fol-
lows Heath's presentation and evaluation of the three proposed recon-
structions of Theodorus' demonstration.
3~. This construction is discussed in most of the interpretations of
Republic 546A, cited below; it is very well described in F. Hultsch, "Die
Pythagoreischen Reihen der Seiten und Diagonalen von Quadraten und
ihre Umbildung zu einer Doppelreihe ganzer Zahlen," Bibliotheca Mathe-
matica, Srd ser., I (1900), 8-12.
34. Heath gives such a generalized form, History, I, 204-5.
35. H. G. Zeuthen, "Sur la constitution des !ivres arithmetiques des
Elements d'Euclide et leur rapport a la question de l'irrationale," Over-
sigt over d. kgl. Dansk videnskabens Selskabs Forhandlinger, 1915,
pp. 422 H., as cited in Heath, History, I, 206-11.
36. Heath, Euclid, III, 29-31; see Fig. 17.

CHAPTER II

1. Critias 112D.
2. Ibid., I19D. 3. As Mr. Rosenmeyer has pointed out (Classical
Philology XLIV [1949], 117), my translation of this as "every five and
every six years" (Classical Philology XLIII [1948], 40) is not so literal as
"every fifth and every sixth year." But since the ordinal numbers are being
assigned to distances, not to points separated by them, I am convinced
that "every fifth year" here does mean "every five years." Compare Plato's
similar use of ordinals assigned to intervals discussed in Chap. III, sections
6 and 7, and Chap. IV, Introductory Comments. The point is particularly
clear in III, 7, where a list from first to ninth is said to have its terminal
items separated by a distance of nine units. The difference between the
number of intervals and the number of points they separate is recognized
clearly in Republic 546A, and whenever there is a choice open in con-
structing mathematical images, Plato seems to assign ordinal numbers to
the intervals.
3. Critias, lI3E.
4. Ibid., 116D.
5. Ibid., 113D.
6. Ibid., 116A.
7. Ibid., I 15D-E.
276 NOTES FOR PAGES 49 TO 61

8. Republic 616C-617D: J. Cook Wilson, "Plato, Republic, 616E,"

Class. Rev.~ XVI (1902), 292-93; J. Adam's notes in his edition of the
Republic (The Republic of Plato, 2 vols. [Cambridge, 1902]), in which
he rejects his earlier interpretation, presented in his note "On Plato,
Republic X 616E," Class. Rev.~ XV 1901), 391-98. See Chap. III, Sec. 8c,
and Appendix A.
9. Critias 113E.
10. Ibid., IISC. 5.
II. Ibid., liSA.
12. Ibid., 118D. These canals actually represen..' -an arrangement too
haphazard to fall under the rule of 6's and 5's; Plato says that there were
transverse canals every 100 stades (a total of 31 canals, counting the outer
ditch), with connecting canals which had been cut between them. The
image of a geometrical maze is used here with the arithmetical metric
frame to reinforce the notion of confusion. See Fig. 19, following.
13. Ibid., IISC. 7-D. 2.
14. The use of "myriads" has a special function, discussed below; but
the length is stated in a way which does emphasize its determination by
the sides; hence its derivation in this context from the summation of
2's and 3's, which elsewhere is presented by 5 as a symbol of the confusion
of odd and even.
15. Ibid.~ IIBE-119B. See Fig. 22.
16. Critias 120D.
17. Ibid., 116D. I: "o'Cu6Cou JA,8'V JA.TIXO;, £~Qo; 88 'tQLot tCAe'ftQOl.;••••"
18. Ibid., 116E. 1.
19. Jay Hambidge, The Parthenon and other Greek Temples (New
Haven, 1924).
20. The discrepancy between exact accuracy in construction and ap-
parent exactitude presents an aesthetic problem which the Greek archi-
tects had resolved, but which Plato does not develop or explain; in fact,
he seems to avoid analogies to architecture.
21. Republic 546 ff. See Chap. III, Sec. 6/ (12 and 14).
22. Hambidge, The Parthenon, Appendix.
23. Laws 701D, 757A.
24. Ibid., 757.
25. Ibid., 719C.
26. Ibid.~ 817 A, 746E.
27. See Fig. 24, following.
28. The sharpest distinction between logistic and arithmetic is made in
Gorgias 451B-C. Arithmetic deals with odd and even numbers, and the
,~i ('
".','.-..•.,.•.•..
:.'...•..:

j·':;.~I· , ·. :'
quantities of each (i.e., classifications and factors), whereas logistic deals ' ~,

with the same numbers, but considered as sets, not as classes (as "having
plethos"). The type of problem considered "logistical" is clarified by the '';;f
..

T
 NOTES FOR. PAGES 61 TO 64 277
scholion to Charmides 165E (Greene, Scholia J p. 115; translation and com-
ment in Heath, History, I, 14-15).
29. See Fig. 23, following (factors of 5,040).
30. A mathematician today can verify Plato's calculation easily by the
theorem that if pa_q b are the prime factors of a number, N, then the total
number of divisors of N is (a + l)(b + 1). Since 5,040 equals 24 -3 2 -5-7,
the number of factors is 5-3-2-2, or 60.
31. A. E. Taylor, Plato, p. 477, n. I.
32. Laws 746E.
33. Ibid., 771B, 746E. Similar proposals for a duodecimal number sys·
tern, based upon its superiority for calculation, continue to be put forward.
See, for example, E. Ullrich, Das Rechnen mit Duodezimahlzahlen
(Beitrage zum Programm der Realschule zu Heidelberg, 1891 [1117]),
Heidelberg, 1891; and F. Emerson Andrews, "Excursions in Numbers,"
Atlantic Monthly, CLIV (1934), 459-66.
34. Laws 817C.
35. Ibid.} 697C; Euthydemus 305C-D.
36. Cf. Philebus 24, 52B, 64E, 65D; Statesman 283-85; Laws 746E;
Parmenides l49B, 151B, 164-65.
37. Cf. Appendix A.
38. A. E. Taylor, in "A Note on Plato's Astronomy," Classical Review,
XLIX (1935), 53-56, gives a clear, brief statement of the p:r:oblem that
results if we take the doctrine of "one path" as representing a theorem
of theoretic astronomy. L. A. Post's review of J. B. Skemp's The Theory
of Motion in Plato's Later Dialogues, American Journal of Philology, LXV
(1944), 298-301, suggests an interpretation of the discussion of circular
motion in Laws 893 which would be relevant to the present passage. On
Post's interpretation, it should follow that the reference here to one path
is to some single closed curve generated as a combination of circular or
circular and linear movements. These conditions, however, are met com·
pletely by the Timaeus, where the helix generated by a point on a rotating
circle which at the same time oscillates through an angle of 46 degrees is
used to describe planetary motion. One can accept both the present inter-
pretation of "one path" and Post's rendering of the later passage.
As for the common mistake of "calling the swiftest runner the slowest,"
this is the direct result of failure to recognize that planetary motion has
two components. The retrograde component is the proper motion of each
planet, since the forward component is transmitted from the outer heaven.
To the untutored observer, it looks as though the planet with the slowest
period were falling least far behind the fixed stars in a forward race.
No elaborate system of negative nurnbers is needed to recognize that the
slower runner falls further behind the swifter than one less slow. See also
n. 130, Chap. III, following.
278 NOTES FOR PAGES 65 TO 85

39. Laws 819C.

40. See Chap. III, Sec. 8c, following.
41. See Figs. 66-74, following.
42. Laws 819D.
43. Republic ii.
44. Statesman 283-88.
45. Laws 894A; see Chap. III, Sec. 4, following.
46. Aristotle, De anima, 404a.
47. Laws 945.
48. Ibid., 948.
49. Ibid., 934C.
50. Ibid., 959.
51. Ibid., 774A.
52. Ibid., 956C.
53. Ibid., 914B.
54. Ibid., 774A.
55. See Chap. III, Sec. 8a and Sec. 8b, following.
56. Republic 615A.
57. Timaeus 23E, Critias 108E.
58. Phaedrus 249A.
59. Phaedrus 257.
60. Critias 119B if.

CHAPTER. III

1. Scott Buchanan, Symbolic Distance in Relation to Analogy and

Fiction, Psyche Miniatures, General Series, No. 39 (London, 1932), pp.
101 ff.
2. See the scholion figure (Fig. 33), following.
3. Cratylus 432C.
4. See Figs. 26-28, following.
5. See Fig. 27, following.
6. Gorgias 508A.
7. See Chap. IV, Sec. 2, following.
8. It is easy to calculate the equivalence of any given Platonic ordinal
number in a schematized matrix, and the equivalent row and column
subscripts. Since these are verbal, not numerical, matrices, the ordinal
numbers in the Platonic schematism do not respond in the same way as
the subscripts to matrix operation.
9. Republic 400B if.
10. Ibid., 443D.
11. Ibid., 424B.
 NOTES FOR PAGES 85 TO 99 279
12. Ibid.~ 546A fE.
18. Ibid., 514C-543.
14. Ibid., 509D-511.
15. Ibid., 587D.
16. Ibid., 616 ff.
17. For a further discussion of the analogy of the musical scale, see dis-
cussion of the world-soul, Chap. IV, Sec. 1; also J. F. Mountford, "The
Musical Scales of Plato's Republic," Classical Quarterly, XVII (1928),
pp. 125 fl.; and Aristoxenus, Harmonics, edited with a translation, by
Henry S. Maeran (Oxford, 1902), pp. 11 if.
18. Jowett, Dialogues, III, 686.
19. See Fig. 41, Chap. III, Sec. 8.
20. Adam, Republic, I, 214-15.
21. P. Shorey, "The Meaning of kuklos in Plato's Republic, 424A,"
Classical Philology, V (1910), 505-7.
22. Republic 616 ff.; see Chap. III, Sec. 8, and Figs. 66-74.
23. The discrepancy between the suggestion that the lengths of these seg-
ments somehow symbolize relative clarity of knowledge, and the analogies
said to hold between them, is vividly presented by Warner Fite, The
Platonic Legend (New York and London, 1934), p. 251.
To a reader of hi:s time familiar with geometry, Plato's directions for
construction involving the division and subdivision of segments "according
to the sa~e ratio" would very likely suggest a mean and extreme ratio
section. One of the most spectacular geometrical properties of this con-
struction is precisely the way in which the shorter segment, laid off along
the longer, cuts the latter again in mean and extreme proportion. The
construction once made can thus be applied indefinitely for redivision of
segments in a constant ratio. As has been suggested, this property of
recurring ratio seems to be the line of research developed by Theodorus
in his investigation of the irrationals. A diagram developed by this con-
struction will have segments of lengths which exactly satisfy the set of
proportions attributed to the figure in Socrates' summary of its properties.
This mean and extreme ratio construction of the diagram is shown in
Appendix B, Fig. 108.
24. The other uses of unequal segments referred to are the figure in
Laws x, describing the progress of causality from point to solid, and the
Platonic schematism reported by Aristotle in De anima 404a, representing
the faculties of the soul.
25. Note the use of higher and lower in the Sophist matrix, already
discussed, and the frequent use throughout the discussion in the Republic
of metaphors indicating that we have reached a higher place from which
more can be seen.
One must protest against the urge of typographical convenience, which
280 NOTES FOR PAGES 99 TO 101
has led scholars to represent this image by a line that is horizontal. The
vertical distinction of "higher" and "lower" is clearly part of the imme-
diate intuitive effect of the diagram, and to ignore it shows complete lack
of feeling for those properties of space which make possible significant
diagram and matrix representations. This point is dealt with in detail in
Chap. IV, Introductory Conlments, preceding the section dealing with
optical images in the Timaeus.
26. See the positions of "one" and "many" in the Py.thagorean table of
contraries, Fig. 27, preceding.
27. As in Fig. 44, following.
28. Compare Figs. 31-34 and the differentiation of temperance and justice
indicated in Fig. 35. ~n the present case, inequality : justice :: analogy :
temperance. It is important not to let one kind of knowledge do the work
that should belong to another; it is also important to see the resemblances
which make possible transition from one level to another.
29. The third segment of the divided line has usually puzzled inter-
preters who take it in conjunction with Aristotle's attribution of an inter-
mediate realm of mathematical objects to Plato. There is nothing intrin-
sically puzzling about it. The objects treated by hypothetical techniques
are those of the real sciences; the treatment differs only in that there has
been no dialectical investigation of the principles of these sciences. The
result is that the test of a hypothetical science is its internal consistency;
the symmetry of the system determines its acceptability ~s a hypothetical
account. The test of a true science, on the other hand, is its correspond-
ence with forms, which are grasped in an immediate act of noetic insight.
Aristotle repeats an analogous distinction in his own differentiation of
"dialectic" and "demonstration." The "sciences," therefore, are strictly a
realm of truth, tested by correspondence; whereas the systems of mathe-
matics represent a realm of symmetry tested by consistency; the imitation
of these structures in the natural order is a realm of beauty. This intro-
duction of the touchstones of the good from the Philebus to explain the
theory of truth in the Republic is mediated and justified by Socrates'
account in the Phaedo of his own technique of hypothesis.
This Aristotelian attribution of a realm of intermediates is discussed in
Aristotle1s tlMetaphysics,1I a revised text with introduction and commen-
tary by W. D. Ross, 2 vols. (Oxford, 1924), I, xxxiii-lxxvi; it is examined
in more detail by L. Robin, in La Theorie platonicienne des idees et des
nombres d1apres Aristote, Paris, 1908.
Aristotle, in treating Platonism as an anticipation of the use of formal
causes in his own philosophy, seems to introduce his own distinction be-
tween "abstraction" and "intuition" into his paraphrase of Platonic state-
ments. Aristotle does this in spite of the fact (which he elsewhere points
out) that Plato himself did not sharply separate the two faculties and their
 NOTES FOR PAGES 104 TO 108 281
objects in the way Aristotle has done. But since Aristotle's purpose is to
test the truth of Platonism as a philosophy (and Aristotle is convinced
that any true philosophy must respect this distinction), he can paraphrase
Platonism with the distinction presupposed. The result of this is a hypos-
tatized realm of intermediate mathematical entities in the paraphrase of
Plato. Robin's study brings out this point very clearly.
30. Republic 537C fI.
31. Statesman 284-86.
32. Republic 530D-531E; see Heath, History, I, 286.
33. Republic 531A.
34. Fig. 49, following.
35. Most of the following mathematical terms are taken from topology.
All seem metaphorical, and in two cases statements about their coinage
establish them as intentional metaphor. In their technical use. however,
these terms are given formal definitions which restrict them to "mathe-
matics" proper, and preclude any confusion or wrong identification of,
for example, topological and physiological "nerves." On this point, Pro-
fessor Dan E. Christie, of Bowdoin College, has written:
• . . nerve of a covering and kernel of a homomorphism are cer-
tainly used metaphorically. Whether they were coined as metaphors,
I do not know.... As for references, my dissertation, "Net Homotopy
for Compacta," Trans. Am. Math. Soc., LVI (1944), 275-308, uses
nerve as a tool but hardly goes into the subject. [My original note
had cited pp. 277, 280 of Christie's dissertation as authority for the
usage of "nerve." The letter here quoted is a correction inspired by
equal unwillingness to deprive a friend of a footnote, and to be mis-
represented in it.] Better references are either P. Alexandroff and
H. Hopf, Topologie (Berlin, 1935), p. 152, where "Nerv eines Mengen-
systems" is discussed, or P. Alexandroff, "Untersuchung iiber Gestalt
und Lage Abgeschlossener Mengen," Annals of Mathematics" 2nd series,
XXX (1929), 104: "Nun definiert aber jedes endliche Mengensystem
tJ = FI, F2 ... Fs folgendermassen einen (abstrakt zu denken) Komplex
N(~), den ich den Nerv des Mengensystems nenne..."
As for kernel, on p. 557 of the aforesaid Topologie you will find a
definition of "Kern eines Homomorphismus. . . ." It is assuredly true
that these terms have precise mathematical definitions which rule out
confusion of topological nerves and algebraic kernels with physiological
[or botanical] things.
The references cited will enable the interested reader to see how such
technical mathematical definitions operate in the cases of "nerve" and
"kernel," which seem typical of such coinage. Christie continues:
Here is another item for your file: part of a footnote from G. Joos,

t Theoretical Physics" trans. by Ira M. Freeman (Hafner, N. Y.), p. 562:

,{
'f
J
282 NOTES FOR. PAGES 108 TO 118

H • • • but that at lower temperatures an ankylosis of degrees of freedom

takes place...." (Poincare borrowed this term. [ankylosis] from pathol-

ogy to denote the "freezing up of a degree of freedom. · · .").
This example is therefore a case of documented metaphorical coinage,
as well as use.
In conversation, Professor J. W. T. Youngs, of Indiana University, has
suggested as other examples: "skeleton," ubraid," "hair," "shaving the
hairs off [a surface]," "trees," "diaphragm," "dendrite," "cactoid," "to cap
[a hole]," "osculating planes." He also told me that he was thinking of a
suitable name for a new topological concept, of an element of a given
kind, capable of a given development, and that "germ" seemed to him a
suitable metaphorical tenn. This is a second case where coinage as well
as use can be documented as definitely metaphorical.
The reader must remember that reference to "new" mathematical terms
in the text means terms that are "new" in contrast to ancient Greek
coinages, though I believe the examples given in this note have not had
their novelty and feeling of being metaphor worn off through long
circulation.
36. Compare James joyce's Finnegan's Wake" in which a principle of
polyphony or systematic ambiguity is exploited to create lines and phrases
with several simultaneous meanings, often in different languages.
37. Cornford has done this in his English, Cousin in his French, trans-
lation of the Republic.
38. Cf. Chap. V, following.
39. Timaeus 43.
40. See discussions, preceding, of Meno 82, Sophist 266.
41. Timaeus 43 if.
42. Cf. in Republic 587 Glaucon's ability to interpret the tyrant's num..
ber in its arithmetic and geometric forms. See also Chap. III, Sec. 7,
following.
43. Aristotle, Politics, 1313a2-b28. Aristotle, criticizing the theory of
revolutions and the notion of a cycle of constitutions put forward in the
Republic, quotes the phrase "of which the 4:3 base joined to the pempad
produces two harmonies when thrice augnlented," as the principle of
Plato's theory. He explains the "thrice augmented" by adding "that is,
when the number of the figure has become solid." From the passage, we
can infer (1) that Aristotle takes this passage seriously and apparently
understands it, (2) that the phrase he quotes is in his opinion that part
of the passage which serves as uprinciple" for the rest, (3) that the theory
is illustrated with a diagram, which, in the later development that
Aristotle does not feel the need to quote directly, represents a modification
of the figure described in the phrase quoted such that three factors are
needed for arithmetical statement of the relations of that final figure.
I
1
J
 NOTES FOR PAGES 118 TO 119 28!
The chief objections Aristotle urges are that the schematism is too remote
from empirical fact to have any value as a contribution to a genuine
Aristotelian science of politics; he notes that Plato himself was constrained
by the facts of the matter to leave the cycle incomplete, and not to carry
out this basic cyclic metaphor by postulating a revolution from tyranny to
aristocracy. For the most part, Aristotle is interested in the later develop-
ments of Republic viii-ix, not in this passage; but he does recognize the
passage as a schematic explanation of the principle organizing the later
discussion.
44. Timaeus 51 if.; discussed in Chap. IV, Sec. 3, following.
45. The basic idea underlying the Pythagorean use of this triangle as
a symbol of marriage seems, from its use here, and from the general
character of Greek genetics, to be correctly represented by later tradition
as stated by Heath. The point is discussed as follows by Heath (Euclid,
2nd ed., I, 417, "Popular Names for Euclidean Propositions. 1.47."):
1. The Theorem of the Bride (theorema tes nymphes).
This name is found in a MS. of Georgeus Pachymeres (1242-1310) in
the Bibliotheque N ationale at Paris; there is a note to this effect by
Tannery (La Geometric grecque, p. 105); but, as he says nothing more,
it is probable that the passage gives the mere name without any explana-
tion of it. We have, however, much earlier evidence of the supposed
connexion of the proposition with marriage. Plutarch (born about
46 A.D.) says (De Iside et Osiride, p. 373F) "We may imagine the
Egyptians (thinking of) the most beautiful of triangles (and) likening
the nature of the All to this triangle most particularly, for it is this
same triangle which Plato is thought to have employed in the Republic,
when he put together the Nuptial Figure [gamelion diagramma-dia-
gramma, though literally meaning "diagram" or "figure," was commonly
used in the sense of "proposition"] . . . and in that triangle the per-
pendicular side is 8, the base 4, and the hypotenuse, the square on
which is equal to the sum of the squares on the sides containing (the
right angle), 5. We must, then, liken the perpendicular to the male,
the base to the female and the hypotenuse to the offspring of both....
For 3 is the first odd number and is perfect, 4 is the square on an even
side, 2, while the 5 partly resembles the father and partly the mother,
being the sum of 8 and 2."
Plato used the three numbers, 3,4, 5, of the Pythagorean triangle in
the formation of his famous Geometrical Number; but Plato himself
does not call the triangle the Nuptial Triangle nor the number the
Nuptial Number. It is later writers, Plutarch, Nicomachus, and Iam-
blichus, who connect the passage about the Geometrical Number with
marriage; Nicomachus (In trod. Ar. II, 24, 11) merely alludes to "the
passage in the Republic connected with the so-called Marriage," while
284 NOTES FOR PAGES 119 TO 125

Iamblichus (In Nicom., p. 82, 20, Pistelli) only speaks of "the Nuptial
Number in the Republic."
It would appear, then, that the name "Nuptial Figure" or "Theorem
of the Bride" was originally used of one particular right-angled triangle,
namely (3,4,5). A late Arabian writer, Beha-ad-din (1547-1622), seems
to have applied the term "Figure of the Bride" to the same triangle;
the Arabs therefore seemingly followed the Greeks. The idea underlying
the use of the term, first for the triangle (3,4,5), is that of the two
parties to a marriage becoming one, just as the two squares on the sides
containing the right angle become the one square on the hypotenuse
in the said theorem.
46. See Adam, Republic, II, 264 if.
47. In general, in the texts of passages involving any sort of mathe..
matical reference or matrix construction, the manuscript tradition seems
more reliable than most editors have believed it to be.
48. The myth of the gold, silver, and iron varieties told in Republic
415B insists (1) that the differences are not in species, hence will not pre·
vent crossbreeding; (2) that the varieties will tend to breed true, but will
not necessarily do so. For example, our citizens are told, two iron parents
may produce a silver child, two silver parents an iron child, "and so for
all the other combinations." The three-part difference in these varieties
is a difference in capacity, with environment held constant, to rule wisely;
hence the axis representing intellectual capacity in the present combina-
tion diagram is anticipated. The explicit mention of all the other com·
binations after the examples given suggests that behind this story lies some
scientific explanation of the genetic phenomenon, which would be based
on a combination diagram of the possible kinds of parental ability.
49. The suggestions made to citizens, that marriages between couples of
opposite temperament are likely to produce the best children and to be
best for the state (Laws 773A), would probably have been made to the
artisan class in the Republic.
In a forthcoming article on Plato's genetic theory in the Journal of
Heredit'Y~ I have summarized evidence from the history of genetics in the
pre-Platonic period and from Plato's work which indicates the following
conclusions: (1) The concept of "normative measure," involving two
directions of deviation from a nonn, gives a genetic matrix which explains
the practical marriage regulations described in the Statesman} where in
fact these regulations are presented as instances of "nonnative measure."
(2) This same formulation makes the Myth of Metals in the Republic a
very precise piece of "popularized science"; the analogy of metals and
alloys is a fiction bypassing complex physiological phenomena, but the
expectation of variation between proper social functioning of parents and
children of every kind is exactly what the theory predicts. The provision
 NOTES FOR PAGES 128 TO 144 285
for equality of opportunity of all children, whatever their parentage, in
a just state, is therefore not a mere fabrication contradicting known
scientific fact. (3) If the view of the relation of the Republic and the main
divisions of the Statesman suggested in my "Plato Studies as Contemporary
Philosophy," Review of Metaphysics, VI, (1952), is correct, the eugenic
program that Plato would advocate for an actual human society would
be about that of the Statesman. The basic problem Plato was treating is
given a twentieth-century statement in Lee R. Dice, "Heredity and Popu-
lation Betterment," The Scientific Monthly, LXXV (1952),273-79.
50. See text of Fig. 48.
51. Compare the scholion to Phaedrus 244A (Greene, ScholiaJ p. 79).
52. J. Dupuis, "Le Nombre geometrique de Platon; Memoire definitif,"
in Theon Smyrnaeus, Exposition des connaissances mathematiques utiles
pour la lecture de Platon, ed. with a French translation by J. Dupuis,
Paris, 1892. Bibliographical citations in notes 53-64, following, are, except
for material in brackets, from Dupuis.
53. Fr. Barocius. Franciscus Barocius, Iacobi filii, patritii Veneti, eom-
mentarius in locum Platonis obscurrisimum, et hactenus a nemine reete
expositum in principio Dialogi octavi de Rep. ubi senno habetur de
numero Geometrico, de quo proverbium ,est, quod numero Platonis nihil

I
obseurius, Bologne, 1566.
54. Bodin, J. Les dix livres de Ia Republique, par J. Bodin, Angevin,
1583. Ensemble une Apologie de Ren~ Herpin, Paris, 1581.
I 55. Peucer, G. Les divins ou Commentaire des principales sortes de
~, devinations, distingue en quinze livres, lesquels les ruses et impostures de
( Satan sont descouvertes, solidement refutees et separees d'avec Ies sainctes
Propheties et d'avec les predictions naturelles, tr. par Simon Goullart,
1 Anvers, 1854. Tome IX, Chap. viii.
56. Mersenne, P. Traite de l'hannonie universelle par Ie sieur de Sennes,
Paris, 1627, t. II, theoreme xiii, p. 430.
I 57. [Thomas Taylor, The Theoretic Arithmetic of the Pythagoreans,
with an introductory essay by Manly Hall, Los Angeles, 1934, pp. 148-57.

I The interpretation here corrects some details of Taylor's earlier presenta-

tion, in his translation of the RepublicJ cited by Dupuis; but Dupuis'
summary applies equally well to the conclusions of this later version.]
58. Le Clerc, J.-V. Pensees de Platon sur Ia religion, la morale, la
politique, Paris, 1819, p. 310.
59. Schneider, C. E. Chr. De numero Platonis commentationes duae.
Quarum prior novam jus explicationem continet, posterior aliorum de eo
opiniones recenset. Breslau, 1821.
60. Vincent, A.-J.-H. Notice sur trois manuserits grees relatifs a la
musique. Supplement a la note L, sur Ie nombre nuptial. (Notices et
extraits des mss.... Vol. XVI, pt. 2, pp. 184-194.)
I

I
I
286 NOTES FOR PAGES 145 TO 152
61. Martin, Th.-H. Histoire de l'Arithmetique, Ie nombre nuptial et Ie
nombre parfait de Platon (Extrait de la revue archeologique, 13 8 annee.)
Paris, 1857.
62. Mynas, C. M. Diagramme de la creation du monde de Platon decou-
vert et explique en grec ancien et en £ran~ais apres 2250 ans. Paris, 1848.
63. [Eduard Zeller, Plato and the Older Academy, trans. by S. F. Alleyne
and A. Goodwin, new ed. (London, 1888), note 110, pp. 423-28.]
64. Weber, Otto. Gymnasium zu Cassel, Programm vom Schuljahre,
1861-2. Inhalt: De numero Platonis seripsit Dr. Otto Weber, Cassel, 1862.
65. P. Tannery, uY a-t-il un nombre geometrique de Platon?" Rev. des
1!:.t. Grecques, LXX (1903), 173-79.
66. Ivor Thomas, ed., Greek Mathematical Works, 1: From Thales to
Euclid (London: Loeb Classical Library, 1939), pp. 399-401, D. c, sum-
marizing Laird's and Adam's interpretations. Of interest in connection
with the latter is G. Kafka, "ZU J. Adams Erklarung der Platonischen
Zahl," Philologus, LXXXIII (1914), 109-21.
The reader must remember, as a counterweight to Thomas and Kafka,
that no number of the size proposed makes any sense if we try to interpret
it as a period of a political cycle. Social changes do not take 36,000 years,
particularly in societies as small and turbulent as the Greek city-states.
From the postulates in context in the Republic, the first step of the down-
ward cycle must occur in less than 160 years, and the full progression in
less than 800. Any realistic study of history would show that changes in
social forms do not take anything like Zeller's more modest 7,500-year
period. Neither does the recession from an ancient "Golden Age"
approach these large figures; the time between the existence of Plato's
Athens and a mythical "Ancient Athens" which was a living embodiment
of Republic i-v was 9,000 years, and the great cycle is 10,000; numbers
that, even translated into days, are much too small for Adam or Thomas
Taylor (not to mention Sosigenes, cited by Proclus as making the number
something larger than 33 trillion). Nor is there any warrant for treating
this as the life-cycle of the human race. Men have survived many catas-
trophes, and presumably always will, because the perfection of the cosmos
requires that there be human beings ··in it. So even if this were what the
Platonic text said (and certain specific comments already made have given
some of my reasons for thinking it is not), we should be very far from
seeing what it could possibly mean.
67. Since the list is usually visualized as written out in a line:
(1) aristocrat
(2) timocrat
(3) oligarch
(4) democrat
(5) tyrant
 NOTES FOR PAGES 152 TO 161 287
In this arrangement the interval is readily seen to be 1:5, not 1:9.
68. Since 729 equals 364~ X 2, it equals the number of days and
nights in a year. Cf. Adam, Republic, II, 358-59.
69. Sophist 266. See Chap. III, Sec. I, preceding.
70. Republic 581C. 3.
71. Ibid., 586E-587C.
72. If a total matrix is

a b c
A Aa Ab Ac
B Ba Bb Bc
C Ca Cb Cc

a gnomon-section is Aa-Ba-Ca-Cb-Cc.
73. The auxiliary differs from a timocrat, and the artisan from an
oligarch, because both artisan and auxiliary are temperate, i.e., motivated
by reason, not by ambition or avarice.
74. Phaedrus 248.
75. Thus there are three characterizations given of the best kind of
lover, two each of the second through the eighth kinds, and only one of
the ninth. The same principle repeats in the listing of choices of life in
Republic X (see Chap. III, Sec. 9). This suggests that the emendation
!~,
l.
.·."i'.t.,.,
.. <11> at Phaedrus 248D. 6 (Platonis Opera, ed. J. Burnet, II [Oxford,
1910]), which spoils the principle, is not desirable.

J 76. Republic 557C. To lend credence to the parallel between Republic

ix and Phaedrus 248 here suggested, it may be noted that in Republic x,
+' ~
the souls changing into animals (Orpheus, Thamyras, Ajax, Agamemnon,
Atalanta, and Epeius and Thersites) represent, as Proclus and Adam have
I
noted, the musician-ruler-athlete-artisan-mimetic artist of the Phaedrus list,

1 and they are presented in almost the same order. See Chap. III, Sec. 9.
77. Republic 587C.
78. Adam, Republic, II, 360-61.
79. Republic 582A. (See Fig. 65.)
80. Ibid., 582D.
81. Ibid., 580D if.
82. See Fig. 64, following.
83. Republic 577C-580B.
84. Ibid., 571C-572B, 575A.
85. Ibid., 612C-614A.
86. A. G. Laird, UNote on Plato, Republic, 587C-E," Classical Philology,
XI (1916), 465-68.
87. Republic 534A.
88. Ibid., 611C.
288 NOTES FOR PAGES 161 TO 169

89. Ibid., 6l4B to end.

90. Theaetetus l76E.
91. Republic 617D.
92. See below, Section (7), (8).
93. The mechanism by which reason modifies appetite through the
presentation of images of its concepts, which in themselves have no ap-
petitive appeal, is presented in the account of the function of the liver,
Timaeus 71B-E. The strange notions of physiology in this passage should
not blind the reader to the importance of the basic psychological insight
into the relation of thought, imagination, and desire.
94. Some of the factors preventing uniform coincidence of a priori
expectation and actual perception are discussed in Timaeus 43D-44D; see
also the discussion of optics in the Timaeus, Chap. IV, Sec. 2, following.
95. Republic, 400-1.
96. The suggestion that what the souls see is really a model of the
universe, which Necessity holds in her lap, seems to have been made first
by J. A. Stewart, in The Myths of Plato. This suggestion has two argu-
ments in its favor. In the first place, it reconciles the position of the souls
with what they see; from the place of choice, which is on a trans-celestial
plain, the girders of the world can be seen stretching out into space, but
a bird's-eye cross-sectional vision of the cosmic hemispheres would not
be possible. In the second place, it accords with Plato's technique in the
Timaeus, where God's creation of the world is presented by a description
of the techniques that a maker of astronomical models would use in con-
structing a machine to show the motions of stars and planets. Further,
only in a model could the celestial spheres be seen in cross section, so
that their relative structure and the mechanism would not be hidden by
the all-enclosing outermost sphere. Of course, there must be complete
correspondence between the model and the world it represents, or the
model ceases to have value as a guide to the choice of life which the
souls who are shown it are about to make.
The present interpretation is based on the conviction that what the
souls see is in fact such a model as Stewart suggests.
97. Cf. Timaeus, mechanics of perception, 44D-47C; role of space in
thought, 52B. Since both perception and imagination are limited by space,
neither structures nor principles can be directly perceived or imagined,
but must be recognized by reason.
98. Cf. the basis of the differentiation of reason and opinion, Timaeus
5ID.
99. Timaeus 57E-58C.
100. Cf. Heraclitus, Frag. 24 in J. Burnet, Early Greek Philosophy (4th
ed.; London, 1938), p. 135. See also the definition of justice in Republic iv.
101. Phaedrus 248E.
 NOTES FOR PAGES 175 TO 185 289
102. Adam, Republic, II, 443, Fig. ii; 445-7. Since this study was written,
however, I have decided that the evidence is by no means "conclusive,"
as I had thought. In discussion with Mr. Lawrence Hall, of Reed's Cove
Boatyard, Orr's Island, Maine, and several of his friends, I found that
experts in small sailing craft agree that any functional reinforcing cable
or strap must pass from side to side beneath the hull, not from stem to
stem. To quote one boatwright, the cable in Adam's model "may be a
fender, or something of that sort, but it couldn't be a reinforcing cable;
boats don't break that way." Mr. Hall is convinced that this would neces-
sarily be true for a Greek trireme. As a Platonist, one should probably
defer to the judgment of experts in a matter of their own art, and reject
Adam's interpretation of the hypozomata.
103. Cf. Chap. III, Sec. 1, preceding.
104. Republic 498D.
105. Ibid., 617B.
106. A. E. Taylor, A Commentary on Plato's uTimaeus" (Oxford, 1928),
pp. 161-62, n. 2: "The MSS text [of this list of colors] should therefore
be accepted as the only authenticated one. I confess I am entirely at a
loss to understand what the order it gives is meant to represent. It is hard
to acquiesce in Adam's view that it is a mere jeu d' esprit; it ought to
stand for something which can be detected in the 'appearances,' but for
what?"
107. See Figs. 66-74.
108. As Taylor remarks in the note just quoted, Adam has no definite
function to suggest for this phenomenon of balancing.
109. The interpretations of Chapters II-IV of this study show that
Plato's images are meticulously constructed. See Figs. 66-75, which, as
calculation will show, would be very unlikely to balance as they do if
they had not been constructed with this result in view.
110. E.g., Timaeus 52D-E.
Ill. Timaeus 68E is a clear, brief statement of this distinction.
112. See Adam, loco cit.; John Burnet, Opt cit., pp. 188, 190-91, 304, n. 1.
113. Sir Thomas Heath, History, I, 335, summarizes his conclusion, de-
veloped in his studies of Greek astronomy, that Aristotle reintroduced the
Ionian notion of astronomy as celestial mechanics in his notion of coun-H

teracting spheres." The mathematical description of the world qua

organism is given by Plato in the proportions used in the construction of
the world-soul, Timaeus 35A-36E.
114. Timaeus 52D-53C, 57E-58C.
115. Ibid., 58D If.
116. Mass equals volume and density. The term is not used in its tech-
nical modern sense, but as representing the concept that heavier bodies
are harder to lift or move, or, once put in motion, to stop, a concept
290 NOTES FOR PAGE 185
which Plato uses, for example, in Republic iVa to describe the inevitability
of the growth of a state, once it has started gaining such "momentum. It

117. See Figs. 69, 70.

118. I would like to correct here a statement made in my article,
"Colors of the Hemispheres in Plato's Myth of Er (Republic 616E),"
Classical Philology XLVI (1951), 178-76. On p. 174 I say that the prob-
ability of the properties of symmetry and balance appearing by chance in
lists of digits from 1 to 8 "can easily be calculated." In footnote 10 (ibid.,
276), an approximation of this probability is calculated. But the approxi-
mation assumes the independence of certain factors which are not in fact
independent, and the correct calculation is "easy" only for a professional
calculator.
The conditions stated in the footnote cited are too far from those of
this particular problem to give the right result. Actually, the correct figure
is arrived at as follows. Given any list of digits from 1 to 8, take the
first one and its complement (9 minus this first digit). How many other
pairs of positions adding to nine would we describe as "symmetrical"?
Tak.ing the digit which lies in one position of a second "symmetrical"
pair, what is the chance that the other position of that pair will contain
its complement? Similarly for third pairs, given the first two. (If the first
three pairs meet these specifications for symmetry, the fourth pair must
also meet them.) Of these symmetrical arrangements, which will always
have the sum of the first four digits equal to those of the last four? In
what proportion of the patterns will only half the symmetrical arrange-
ments approximate this balance?
The answers to these questions must be worked out separately for each
location of the complement of the first digit in the list. The result is that
the chance of having both symmetry and balance in a single such list,
constructed at random, is less than 1/10 (actually, 9/98). To have four
independent lists show these properties if they are random in construction
has a probability of 1/10.4: (The renumbering in order of size of the sums
of two such lists interferes with both properties sufficiently to justify
treating the fourth balanced figure [of masses] as independent.) The sym-
metry of the fifth list, also independent, has a likelihood of 1/7. The
probability that these properties are only coincidence is therefore 1/70,000.
Ordinarily, we assume that in a given single case an event with a prob-
ability of 1/10,000 or less does not in fact happen. (Ordinary statistical
procedures make a much less rigorous assumption.) The fact that the
principle of symmetry appears in a clearer form in the fifth list (if the
reader will grant that balance of adjacent pairs is the most clear and
natural way of representing mutual adaptation) is, since we admit the
first and last digits as "adjacent," 3/105. But, having already included the
condition that this fifth list is symmetrical, we must use here only the
 NOTES FOR PAGES 185 TO 186 291

probability that the favored type of symmetry will appear in the fifth of
five symmetrical lists, which is 1/7. The final figure, therefore, expressing
the probability that the properties present in these lists would appear in
'such a set of lists constructed by chance, is 1/490,000. Though this is
considerably less than the figure that would result if the properties were
treated as independent for each pair in each list (the calculation in my
footnote would apply exactly to certain techniques of list construction,
but presumably not to those that Plato used), the revision makes no
practical difference. It is still mathematically shown that the appearance
of these properties is so unlikely to be coincidental that we can disregard
that explanation.
The reader is asked to verify the fact that in the present calculation
the problem has been stated in a way which evaluates the chance of an
interpreter's being able to find these properties in lists when he is looking
for them. In other words, conditions are used which operate against find-
ing a high improbability; a very liberal notion of "symmetry" is em-
ployed, and the interpreter is assumed to look in each list for any pattern
that can be called "symmetrical." By assuming slightly different conditions
(for instance, that fewer arrangements are admitted as having the property
of "symmetry," or that we are dealing with the construction of lists by
drawing digits and position numbers from a hat), one can set up condi-
tions that seem to apply to this problem, but, not being exactly enough
applicable, give too high an improbability.
119. Republic 616C. 6-8.
120. Timaeus 67C. 5-68E. 9.
121. Ibid., 68A.
122. Ibid., 68B. 1-5.
123. Republic, 617A. 2-3.
124. Ibid., 617A. 4.
125. Timaeus 68C. 2-4. In connection with this entire passage on color
perception, and with the device of arranging colors by intensity in the
Myth of Er, the descriptions of color mixture and the emphasis on in-
tensity correspond closely to the first-hand color experience of a man at
least partially red-green color blind. (This is especially true of the descrip-
tion, in the Timaeus passage just cited, of bright green as the result of
mixing grey and yellow.) It is therefore possible that many of the pecu-
liarities of the color description do not reflect general Greek practice, but
idiosyncrasies of the color vision of the author; and that, in turn, may
help to explain why Plato feels this whole field is controversial, and
difficult to secure agreement about.
126. Timaeus 59B-60D.
127. Unless, as Taylor suggests in the quotation cited in n. 106, the
"older version" of the text given by Proclus (discussed by Adam) was an
292 NOTES FOR PAGES 188 TO 190

attempt to judge this property from apparent luminosity, a very unscien-

tific undertaking.
128. Cf. Figs. 66-75.
129. Theon says that he himself built a model of the astronomical
machine described in the Myth of Er. Presumably this was one of the
water-driven orreries which Hellenistic mechanics had devised; see Heath,
History, II, 428-29.
Though the variant xUALEaihn in Theon's text would describe a planet
moving within its zone in a cycloid, nothing in Plato's description of the
model corresponds to the planet as distinct from its zone or orbit, and of
course no motion compounded of rotation and translation fits the me-
chanics of concentric hemispheres in the machine.
The difference here between Theon and the manuscript tradition sug-
gests that both may be alternative restorations of a corrupt original. In
that case, it is possible that Plato originally gave some indication here of
how the transmission of momentum in the model came about.
130. This diffusion of velocities is somewhat obscured by the fact that
whereas the swiftest circle has a forward velocity, the velocities of the
others are measured by length of period and are compounded of forward
and retrograde motions. The diffusion of the retrograde impulsion applied
to the inner circle is shown by the proportionate lengthening of period as
distance from the center increases. Taking the retrograde component as
the proper motion of each planet (since the celestial revolution is given
it from outside), we can see why, in the Laws treatment of astronomy
discussed in Chap. II, preceding, ordinary common sense is blamed for
"praising the slowest runner as the fastest"; this follows if we think of
planetary motion as having only a forward component, and the planet
with the longest period as coming closest to "keeping up with" the motion
of the stars in a celestial race. See n. 38, Chap. II.
131. The place of the vortex concept in Ionian cosmology is well de-
veloped in Burnet, Early Greek Philosophy.
132. Evidently the Hellenistic notion of fluid transmission is the most
natural one to apply in designing a model that will duplicate the motions
Plato describes and that will respect the diffusion of impulsions applied
to the outside and center of the system. At any rate, Professor Wallis
Hamilton, of Northwestern University, in reply to an inquiry as to how
a model could most easily be made that would duplicate the dynamics
of Plato's world-machine as I described them, has suggested a model with
cardboard rings floating on water and an inverted can in the center
setting up the retrograde motion by its rotation.
133. Timaeus 57E-58C.
134. Ibid., 38D.
135. See also F. M. Cornford, Plato's Cosmology: The UTimaeus" of Plato
 NOTES FOR PAGES 190 TO 197 295
translated with a running commentary (New York. 1937). pp. 74-115.
136. Republic 617C. 5-6.
137. Timaeus 36C.
138. Ibid., 37E-38B.
139. Republic 617C. 6.
140. Theaetetus 176E-177A.
141. Republic 592A-B.
142. Ibid., 395.
143. Laws 903E.
144. Republic 612B-613E.
145. If we further assume an intentional or unintentional connection
between the various images of cycle in the Republic) the image of the
momentum of a repeatedly impelled wheel, introduced in Book iv, may
be cited as an indication that the inexorability of this machine is the
result of its tremendous momentum; the massiveness of the cosmic hemi-
spheres makes it impossible for any human agency to stop their motion.
This compares very accurately with the "necessary" physics, in which
process is treated in terms of the velocities and densities.
However, the fact that we are inclined to look for correspondences
among the cyclic images introduced previously and in the present image
reinforces the initial assumption that the nature of the model does not
preclude the possibility of free choice. For both the cycles of evolution and
degradation are presented as spirals. If a state at any time directs its atten-
tion to educational reform, it will improve just as inevitably as it will de-
cline when, neglecting dialectic and impelled only by the motives inher-
ited from its immediate past history, it does not. The political situation,
as presented in these parallel images, seems to be one in which the state
at each moment may either blindly follow its trajectory to an ultimate
j tyranny, or through new insights initiate a reform leading it back to an
aristocracy. Certainly we should not expect a model of cosmic process to
{ render what has been presented as political fact a physical impossibility.
146. The eschatological details, however, retain the emphasis on ten
characteristic of the Orphic-Pythagorean tradition. For the importance of
ten in Plato's myths generally, see Chap. II, Sec. 3, preceding.
147. Republic 617E.
,
148. J. Adam, "On Plato, Republic 616E," Classical Review, XV (1901),
i 891-98 (which must be read with the fact in mind that Adam later rejects
~ ProcIns' readings), and A. E. Taylor, Commentary, p. 161, n. 2, suggests

I
\
the interpretation which seems to have inspired this text. See Fig. 69,
following, for a demonstration that this text does not follow the balanced
pattern of the other lists of details in the passage.
149. In his Early Greek Philosophy, p. 304, n. I, Burnet defends his

I decision to recognize Theon's qualification of Mars's retrogradation. See

294 NOTES FOR PAGES 205 TO 222
A. C. Clark, The Descent of Manuscripts (Oxford, 1918), pp. 383-418.
I have accepted Burnet's hypothesis in this section, though most con-
temporary scholars would question it. Certainly, his argument needs quali-
fication; but one is tempted to think that something of the sort is the
case, particularly in view of the occurrence of five consecutive variants at
intervals of 15 letters, between Theon's text and the manuscript version.
150. Phaedrus 252C.
151. See Sec. 7, note 75, in this chapter.
152. Republic 620B.

CHAPTER IV

1. Aristotle, Metaphysics 1092a.

2. For this principle of multiple interpretation at work, see The
Timaeus and Critias or Atlanticus, trans. Thomas Taylor, reprint ed.
(The Bollingen Series; New York, 1944), Introduction; and The Commen-
taries of Proclus on the tlTimaeus" of Plato, trans. Thomas Taylor, 2 vols.
(London, 1820).
3. An inquiry into the use of mathematics in connection with meta-
physics, as in the ParmenidesJ was originally planned as part of the present
study, but has proven so complex and extensive that it will require a
separate work. I discuss the dialectical context of this problem in Chap. III
of The Rdle of Mathematics in Plato's Dialectic (Chicago, 1942).
4. See the discussion of logistic and mathematics preceding, particu-
larly Chap. II, Sec. 2b.
5. Laws 757B.
6. Laws 894A; Epinomis 990D-99IC; Aristotle, De anima 404a.
7. See subsequent discussion in Chap. IV.
8. E.g., D'Arcy W. Thompson, On Growth and Form (Cambridge,
1917).
9. Aristotle, Physics 206b32.
10. Timaeus 86C. The frontispiece in Cornford, Plato's Cosmology, is
a photograph of such a metal-band astronomical model.
II. Still following the analogy between God's creation of the cosmos
and an artisan building a metal astronomical model, Plato seems here to
envision the "cutting off of strips" as marking off and cutting into the
metal band the intervals of a metric scale. Such a scale is evidently an
important part of a model of this kind. To believe, as most translators
and editors have, that the big band is exhausted by being "cut up" into
smaller pieces, leaves no material from which the outermost circles of the
cosmos are made; and such interpretations render the text at 37C, where
the strip is split lengthwise after this "cutting," quite unintelligible. By
 ,
.I.!
NOTES FOR PAGES 222 TO 241 295
what may be a happy arithmetical coincidence, if we take a wide metal
band incised with a scale from 1 to 27 and split it into three, the third
strip can be cut up to give exactly the lengths needed for the rings within
the model, representing planetary orbits, if the circumferences of these
=
correspond to the basic scale (since I + 2 + 3 + 4 + 8 + 9 27).
12. Timaeus 36C-D.
18. The best English source for an idea of Proclus' full interpretation
is the translation cited in n. 2 of his commentary on the Timaeus.
14. A good deal of Pythagorean "number-symbolism" seems to have
resulted from the Pythagoreans' having had no simple notation for a
mathematical or logical variable. Consequently, the first (i.e., smallest)
integers having a given property are used with the convention that these
represent any term or number that has the same relevant property. In
dealing with opposed pairs of properties, such as the "same" and other"
U

in the Timaeus, specific odd and even integers may be used, the former
standing for terms in scientific analogies representing the "same," the
latter for those representing the "other." (This suggestion answers Aris-
totle's question as to why the Pythagoreans identify a given property with
the first integer that has it, and makes possible a sensible interpretation of
such Pythagorean maxims as "justice is four.")
15. Comford, Ope cit., pp. 59-66; A. E. Taylor, Commentary, pp. 106-8.
16. Parmenides 164C-166.
17. Heath, History, I, 115-17: "The pythmen and the rule of nine or
seven."
18. Cf. L. Robin, Theorie platonicienne (Paris, 1908).
19. Timaeus 47.
20. Ibid., 44A, 64D.
21. Ibid., 51E ft.
22. Ibid., 46B, see Fig. 87, following.
23. Ibid., see Fig. 89, following.
24. Ibid., see Fig. 91, following.
25. Epistle VII, 848C ff.
26. Timaeus 47A ff.
27. Ibid., 46B. 7-13.
28. Ibid., 43B, gOD, 92A.
29. Cf. Timaeus 50D if.
30. Republic vii, 528A.
31. See the table of principles, Figure 49, preceding.
32. Timaeus 54D.
88. Ibid., 58B.
84. Compare the "shock" physics of Descartes, "The World," in Ex tracts
from Descartes' Writings" trans. H. A. P. Torrey (New York, 1892).
85. Timaew 58B-C.
296 NOTES FOR PAGE 242
36. Eva Sachs, Die funf platonischen Korper (Berlin, 1917); Heath,
Euclid, III, notes on Book xiii; Euclid, Elements, xiii.
37. See Chap. III, Sec. Sa and Fig. 66, preceding; the role of equilib-
rium is presented in Timaeus 52D-53C, 57E-58C.
38. Timaeus 55C. If we read the earlier postulate of God's "marking
out triangles" as equivalent to the assumption that an integration of
places produces a space (treating a place as the locus of some homo-
geneous property, the space as a field within which there is constant,
continuous interaction and qualitative flow), this embroidery takes on
more point. The use of points of light to mark outlines of animal figures
in the sky is technologically very like the mathematician's postulate of
infinitesimal "places" which provide the groundwork necessary for a
description of objects in space.
The concept of a limit and the problem of defining it make their ap-
pearance with the Pythagorean attempt to build magnitudes from non..
extended spatial elements; Zeno's paradoxes demonstrated some of the
defects in logic of their formulation. By Anaxagoras' time, it was evidently
felt that some justification must be given for reasoning from the proper-
ties of finite things to the properties of their infinitesimal parts, and
Anaxagoras' concept of homoeomereity is his justification for such infer-
ence. Cornford's demonstration of an analogous property in Plato's ele-
mentary triangles seems a valid and needed one. If space is continuous,
the process of decomposing plane figures can go on indefinitely; but, once
the process of analysis yields figures similar to the ones analyzed, the
inference from finite to infinitesimal is a reasonable one.
Particularly in the light of A. N. Whitehead's recent analyses of space,
place, and location, Plato's development of this geometrical element theory
is a philosophically suggestive one. Plato's statement that "we assume God
marks out the elementary triangles" translates into more contemporary
terms as the postulate that "space equals an integration of places," taking
"place" as a limit. That postulate connects mathematical and empirical
space, since a region (a nexus of places) can be observed and described,
but a pure empty space cannot. Pure space can be "known" only by the
psychological act of emptying consciousness of all particular perception
and thought; space remains. This is exactly the process recommended by
Bergson for apprehending pure "duration," and the similarity suggests
that Platonic "space" has dynamic properties.
The construction of volumes from bounding "places" is a neat mathe-
matical formalization of the psychological construction (explained in non-
mathematical terms in Aristotle, De anima, Book iii) by which we associate
notions of touch and motion with perception of plane visual images, and
perceive volumes and solids.
A comparison of this entire passage in the Timaeus to the relation of
 NOTES FOR PAGES 243 TO 255 297
the monad and the infinitesimal in Leibniz' philosophy is extremely
interesting.
39. Republic x, 60lD if.
40. Cornford, Cosmology, pp. 231-39, where this concept of the homoe-
omereity of the elemental triangles is developed. See Figs. 98-99, following.
41. Timaeus 54E; see Fig. 95.
42. Ibid., S3e: <illitEL AOyCP.

CHAPTER V

1. Gorgias 518E-519A.
2. Prodicus in Protagoras 340C-341C.
3. See HAmbiguity of metaphor," Chap. III, Sec. 6a, preceding.
4. Euthydemus, throughout; also the Protagoras passage already cited,
D.2.
5. Republic vii. 534D.
6. A. E. Taylor (in Plato) reads many of the earlier dialogues as refu-
tations of the Sophistic idea that virtue can be, as it were, externally and
mechanically applied.
7. See above, Chap. III, Sec. 6e and 6f (14).
8. The text is:
'EJtEL6"" £031:£0 MAt;, Ii\' ~' ~ycb, 6L1lQ1ltUL xa-ru -rQia £t8T1, 06100 xal 'lJUX1)
~'Vo; ~XaC1tOU tQl.xn, 6e;E-ras,·, 00; ~JArol 8oxet, xal ~-rsoav Wt63EL;I,V.

• ae;EtW W (Chambray), Chambray, Adam, Shorey

'to loytC1tL)Gav ~e~EtW A (Chambray)
AOYLO''tLXTtV l)e;etal q1 (Adam)
loyuJ'nxTt ~e~E1:aL q2 (Adam)
Aoyunl.XOv 6e;E1:at A2, F, D, M (Shorey), Apelt
AoyurnxOv en:LituJ.tTt.'ttXoV ituJ,tLxOV -Par. 1642 (Shorey)
= =
Chambray Chambray, Republique; Adam Adam, Republic, II; Shorey =
Shorey, Republic~ II; A (Chambray) = cod. Parisinus 1807; ql (Adam) = cod.
Monacensis 237; q2 (Adam) = later corrections of and additions to ql; Apelt =
Platons Staat, trans. and ed. O. Apelt (Der Philosophischen Bibliothek, Bd. 80,
Leipzig, 1920); A2 (Shorey) = later corrections of and additions to cod. Parisinus
1807; F (Shorey) = cod. Vindobonensis 55; D (Shorey) = cod. Venetus 185; M
(Shorey) = cod. Malatestianus XXVIII, 4; Par. 1642 (Shorey) =
cod. Parisinus
1642; W (Chambray) = cod. Vindobonensis 54. The translation given in my text
follows Apelt and Shorey's AI, F, D, H.
9. See discussion of the bewilderment of Glaucon, Chap. III, Sec. 7,
preceding.
10. Compare the differentiation of "logic" (dialectic) and "logistic"
throughout Republic vii.
298 NOTES FOil PAGES 256 TO 269
11. See the analysis of characteristic Neo-Platonic interpretation of
analogy, Chap. IV, Introductory Comments.
12. Theaetetus 146, discussed in Chap. I, Sec. 4, preceding, and in
Appendix B.
13. Republic vii. 5S1E.
14. Meno 83 ft.; Chap. I, Sec. 2, preceding; and Appendix B.

APPENDIX It.

1. Here, as was suggested in connection with the Phaedrus, the power

of Poseidon is his most striking attribute; and the mixture of a divine
nature inherited from him with a mortal nature produces descendants
characterized by their desire for domination and power.
2. See Chap. III, Sec. 8, preceding.
3. See Chap. II, Sec. 1, preceding, for discussion of the confusion of
same and other represented by this use of 5 and 6; and compare with
the use of same and other as orienting co-ordinates of matrices, e.g.,
Sophist 266A, discussed in Chap. III, Sec. 1, preceding.

APPENDIX B

1. The figures involving imagery of quadrilaterals and their diagonals

are presented together for comparison in Fig. l07A-M, following.
2. See Chap. III, Sec. 4, preceding; and Fig. 108.
 INDEX
I

Academy, 18, 40, 66, 185, 227, 258 Atomists, 188 I:

Achilles, 20S Atropos, 17g II
I
Acropolis, 52, 56, 57 Attica, 56

I Adam, J., 88, 149, 150, 152, 154, 158, 175,

198, 199, 200, 276, 279, 284, 286, 287, Barocius, F., 148, 285
1 289, 291, 29~, 297
Agarnenrnnon,203,204, 206,287
Benecke, A., 33, 34, 35, 36, 38, 274
Bergson, H., 237,251,296
Ajax, 177, 194,20~, 204,206,287
Aldrich, V. Ct, 102
Bodin, J., 143, 285
Brumbaugh, R. S., 9, 77, 150, 284, 285,
I
Alexandroff, P., 281 290, 294
Alleyne, S. F., 286 Buchanan, S., 72, 278
Allman, G. Jt, 274 Burnet, J., 21, 197, 274, 286, 287, 288,
Anaxagoras, 241, 248, 296 289,292,293,294
Anaximander,162 Bury, R. G., SS, 55
Andrews, F. E., 277
Apelt, 0., 256, 297 Carthage, 56
Aphrodite, 204, 205, 207, 208 Cephalus, 101, 167
Apollo, 205, 207, 208 Chambray, t., 111-12, 173-74, 196, 297
Aquinas, 287 Christie, D. E., 281
Ares, 205, 208 Clark, A. C., 294
Aristophanes, 16, 55, 251, 252, 256 Clotho, 178
Aristotle, 6, 7, 21, 40, 66, 73, 76, 77, 102, Clytemnestra, 204
104, 118, 212, 216, 217, 283, 237, 248, Coleridge, S. Tt, 157
258, 275, 278, 279, 280, 281, 282, 288, Comford, F. M., 85-6, 88, 97, 171-78,
289, 294, 296 190, 196, 208-4, 220-21, 2~o-31, 238-
Aristoxenus, 279 40,247-48,271,282,292,294,295,296,
Artemis, 205, 207, 208 297
Atalanta, 20~, 204, 206, 287 Cousin, V 0' 282
Athena, 205 Couturat, L., 265
Athenian Stranger, 52, 62, 68, 64, 65, 66
Athens, ancient, 47, 57,59,60,286 Demeter, 208
Athens, 55, 56, 76 Demme, C., 84,274
Atlantis, 47-59, 60, 61, 71, 176, 205, 260, Dercyllides, 197
261, 267 Descartes, 241, 295
299
300 INDEX
i
Dice, L. R., 285
Diels, H., 55
Hopf, H., 281
Hultsch, F., 137, 275 ,•
i

Dribin, D. M., 274

Dupuis, J., 143-46, 285-86
Hume, 102

Iamblichus, 97
I
Elders of the Nocturnal Council, 64-65 Ilissus, river, 57
Empedocles, 248
Epeiw, 203, 204, 206, 287
Er. See Myth of Er
Ionians, 162, 184, 187, 188,289

Joos, G., 281

r
Eridanus, river, 57
Eryximachus, 227
Euclid, 30, 34, 38, 40, 41, 42, 43-44, 246,
Jowett, B., 17-18, 19-26, 30, 31, 38-39,
83-84, 88, 89, 94-97, 109-10, 249-50,
274, 279
1
267, 270, 274, 275, 283 Joyce, J., 282
Eudoxus, 40,41,275
Euler, 264, 265 Kafka, G., 286
Eurytos,21 Kant, E., 4, 102, 104, 235, 287
Euthyphro, 18, 19,33,273 Kroll, G., 132, 137

Fates, 89, 131, 179, 190, 192 Lachesis, 173

Fite, W., 279 Laird, A. G., 137, 286, 287
Freeman, I. M., 281 Le Clerc, J.-V., 144, 285
Leibniz, 32, 264, 265, 297
Glaucon, 16, 66, 101, 102, 149, 156, 157, Lethe, river, 168
227,255,256,257,270,282,297 Lucian, 251
Goodwin, A., 286 Lycabettus, Mt., 57
Gorgias, 81
Gow, J., 20,33,274 Macran, H. S., 279
Greene, W. C., 18, 27-29, 33, 78-79, 86, Maenads, 204
105,137,158,218,274,277,285 Martin, Tn.-H., 110, 145, 146,286
Meno,31, 32,33, 34,35,257,269
Hall, L. S., 289 Mersenne, P., 144, 285
Hall, M., 285 Milanion, 204
Hambidge, J., 52, 276 Milesian engineers, 52, 242
Hamilton, W., 292 Mountford, J. F., 279
Heath, Sir T. M., 30, 31, 32-33, 86, 37, Muses, 114-15, 149, 175, 176
42,48-44,76,274,275, 277, 281,28S- Mynas, C. M., 145, 286
84, 289, 292, 295 Myth of Er, 68, 89,90,91, 122, 161-205,
Heiberg, I. L., 246 242, 289-94
Hephaestus, 205, 207, 208
Hera, 208 Necessity, 65, 162, 173, 179, 242, 287
Heraclitus, 288 Neo-Platonists, 3, 4, 122, 148, 211-12,
Hermann, C. F., 98, 146 223,298
Hermes, 205, 207, 208 Neo-Pythagoreans,6, 148,212,225
Herpin, R., 143, 285 Nereids, 50
Hestia, 205, 206, 207, 208 Nestorius, 132
Hippocratic school, 77
Hippodamas, 55, 56 Odysseus, 168, 204, 206
Homer, 55 Ogygia,55
 INDEX 801
Orpheus, 20~, 206, 287 Poincare, J. H., 282
Ortygia,52 Polemarchus, 101
Poseidon, 48, 49, 50, 51, 205, 206, 207,
Peucer, 144, 285 208, 260, 261
Piraeus, 55, 56 Post, L. A., 277
Plato, works: Apology~ 278, 274; Char- Proclus, 110, 122, 132, 136-37, 158, 173,
mides~ 15, 278, 277; Cratylus~ 205, 197, 199, 202, 204, 212, 222, 223, 225,
278; Critiasl 47-59, 60, 266, 275, 276, 228,286,291,294,295
278, 294; Crito~ 274; Epinomis~ 171, Prodicus, 9, 252, 297
174, 179, 216, 294; Epistle YII, 4, 232, Prometheus, 170
295; Euthydemus, 15, 278, 277, 297; Pythagoras, 33, 41
Euthyphro 15, 16, 17-19, 85, 40, 278,
l
Pythagoreans, 5, 7, 20, 21, 33, 40, 41, 42,
274; Gorgiasl 15, 78, 79, 218,238,278, 55, 61, 62, 71, 73, 76, 77, 99, 106, 119,
276, 278,297; Laws, 48,52,55,58,59- 124, 126, 139, 141, 176, 178, 181, 185,
71, 78, 74, 78, 102-3, 104, 171, 174, 194, 213, 217, 226, 227, 246, 251, 252,
201, 215, 216, 217, 219, 266, 267, 276, 253, 255, 267, 268, 273, 280, 283, 293,
277, 278, 279, 284, 292, 293, 294; 295, 296
Menexenus, 251; Meno, 15, 16, 19-38,
40, 257, 265, 266, 267, 268, 278, 274, Robin, L., 280, 281, 295
282, 298; Parmenides,' 9-10, 158, 209, Rosenmeyer, T., 275
225, 251, 252, 277, 294, 295; Phaedo, Ross, W. D., 280
15, 17, 170, 278, 280; Phaedrus1 20,
128,142,146,152-55,158-59,175,204, Sachs, E., 275, 296
206, 274, 27~ 285,286,287,294,298; Schneider, C. E., 144,285
Philebus,9, 174, 277, 280; Protagoras~ Shorey, P., 87-88, 89, 151, 249, 279, 297
. . 9, 15, 51, 60, 170, 273, 297; Republic~ Sicily, 55
L3, 16, 52, 59, 60, 61, 62, 63, 65, 71, 72, Sirens, 173, 175, 176, 179, 193, 194
79, 85-208, 215, 223, 233, 240, 241, Skemp, J. B., 277
242, 243, 249, 251, 253, 254, 255, 257, Socrates. See Plato, works
266, 267, 269, 270, 271, 275, 276, 278, Sophists. See Gorgias, Prodicus, Thrasy-
279, 280, 281, 282, 283, 284, 285, 286, machus, and Plato, works (Euthyde-
287, 288, 289, 290, 291, 293, 294, 297, mus, Gorgiasl Phaedrus, Protagorasl
298;) Republic-Timaeus-Critias tril- Republic)
ogy, 94, 260, 262; Sophist, 73, 76, 77, Sosigenes, 286
80, 88-4, 152, 233, 268, 279, 282, 287, Stewart, J. A., 288
298; Statesman, 9, 15, 114, 170, 175, Syracuse, 55
249-50,253,257,258,263,266,267,268,
269,273, 277, 278, 281, 284, 285; Sym- Tannery, P., 146, 286
posium, 16, 249, 252, 253, 273; Theae- Taylor, A. E., 81, 84, 71, 147, 188, 191,
tetus, 15, 16, 31, 38-44, 191, 265, 266, 274,277,289,291,298,295,297
26t 268, 27t 27~ 278, 282, 288, 289, Taylor, T., 144, 148, 285, 286, 294
291, 292, 293, 295, 297, 298; Timaeus, Thamus, 20, 170
3, 17,48, 53, 55, 70, 74, 117, 118, 147, Thamyras,203,204, 206,287
161, 163, 170, 175, 178, 184, 185, 186, Theaetetus, 40, 41, 42, 48, 44, 127, 257,
190, 191, 197, 209-48, 266, 267, 273, 267,268,269,275
274, 27t 278, 280, 282, 283,288,289, Theodorus, 41, 42, 275, 279
291,292,293,294,295,296,297 Theon, 41, 124, 188, 196-98, 285, 292,
Pnyx, 57 293,294
302 INDEX
J
Thersites, 203, 204, 206, 287
Theuth, 20, 170
Weber, 0., 145, 286
Whitehead, A. N. J 296
j
f
Thomas, I., 286 Wilson, J. C., 198, 199, 200, 201, 276
Thompson, D'A. W., 294 I
Thrasymachus, 16, 101, 155 l'
Jrhurium,56 Young Socrates, 40, 257,268
Timaeus, 231 Youngs,J. VV. Jr., 282
Todhunter, I., 246, 274
Torrey, H. A. P., 295
Zeller, E., 145, 286
Ullrich, E., 277 Zeno, 296
Zeus, 16, 59, 205, 206
Vincent, A.-J.-H., 144, 145, 146, 285 Zeuthen, H. G., 42, 275