THE MUSICAL SCALES OF PLATO'S REPUBLIC. L Tue object of this article is to discuss, defend, and supplement the only definite piece of evidence we possess which deals with the musical scales (or Harmoniai) referred to by Plato in the Republic (3980-3990). In this first section I shall consider the list of scales given by Aristides Quintilianus and ‘suggest the source from which itis derived ; in the second part the employment of certain abriormal intervals will be established and elucidated; and finally the evidence of the preceding sections will be reviewed from the standpoint of the Pythagorean musical theory. Aristides Quintilianus! is the only one of our authorities on Greek music to state precisely that such and such a sequence of intervals comprised this or that Harmonia, From p. x8 to p. 2x of his Hep! nouzeas he gives an account, of those divisions of the tetrachord which the theory of Aristoxenus had classified under the names of Enharmonic, Chromatic, and Diatonic genera. ‘Then suddenly he seems to have forgotten the genera, and proceeds (pp. 21-2) “There are also other divisions of the tetrachord which the very oldest of the Greeks (oi wdvy srakaravo) used for the Harmoniai. Sometimes the Harmoniai filled up a whole octave ; in some cases they exceeded the six-tone scale, but often they were of lesser extent, for they did not always include all the notes. We will give the reason later. They constructed the Lydian scale by diesis, ditone, tone, diesis, diesis, ditone, diesis (and this was a complete scale). The Dorian they constructed by tone, diesis, diesis, ditone, tone, iesis, diesis, ditone (and this scale was a tone in excess of the octave). The Phrygian was tone, diesis, diesis, ditone, tone, diesis, diesis, tone. This scale also was a complete octave. The Ionian was constructed by diesis, diesis, Gitone, a three-semitone interval, and a tone. This scale was a tone less than an octave. The Mixolydian was two successive dieses, a tone, a tone, diesis, diesis, and 2 three-tone interval. It was a complete scale. ‘The so-called Syntonos Lydian was diesis, diesis, ditone, and an interval of three semitones. ‘Thus the enharmonic diesis is to be heard in all these scales (Sicow 8& viv én) révray éxovertor hp évappévier). For the sake of clearness a diagram of the scales is given below. It is of these scales that Plato (é deios UXdrav) makes mention in the Refublic, where he says that the Mixolydian and the Syntono- lydian are threnodic, and the Ionian and Lydian are convivial and too relaxed (Mav drevnteas)” te arora sny beer ha Plug wel ein nse guerionr Couple searely aects the value of is evidencs since ams friberen Scheien sehr” "verschedeoea, bia boot i a compilation from older writers. Wertes6 J. F. MOUNTFORD The diagram referred to by Aristides is preserved in the MSS., and consists of the notation for the scales, and bears out fully the verbal account of the text. If we take the word diesis to mean a quarter-tone, we obtain the subjoined table: Lydian Dorian Phrygian lonian Mixolydian Syntonolydian 2 ik Aristides makes a very definite and important claim for these remarkable scales. For they are indeed remarkable: in no other theoretical writer do we find such a series, and the extant fragments of Aristoxenus would not lead us to suppose that such musical phenomena ever existed in Greece. Hitherto investigators generally have cither regarded them with distrust or have been bewildered by them. Laloy (Arisfoxéne) is unusually courageous in suggesting that Aristides had access to pre-Aristoxenian works on music, but he has discreetly left the matter there, and has not traced out the full implications of his view. Gevaert (Prodlémes Musicaux d'Aristote) would perhaps accept the passage of Aristides if he were permitted to alter the text, Macram (Avis: toxenus) takes the seales to be elementary attempts at theorizing about the Harmoniai, though such an opinion is inseparable from Macran's own very mistaken theory of the Harmonia, onto (Modes) looked upon them as no better than impudent forgeries. All of these attitudes, however, would seem not only to be unfair to Aristides, but to hinder progress in the knowledge of early Greek music. The value of his evidence needs 2 reconsideration (1) There are no solid a prior’ grounds for rejecting these scales, We may not like the look of them, and they may not satisfy our preconceived notions of what the Platonic Harmoniai ought to belike, Yet neither the size Of the intervals nor their relations one with another need cause any comment. ‘One has only to turn to Carl Engel’s National Music (pp. 24-80) to realize that a remarkable variety of ‘gapped’ scales and uofamiliar intervals is to be found in different parts of the world. (2) There is nothing vastly surprising in the manner in which Aristides thas introduced this digression on the old Harmoniai. It arises from a state~ ‘ment that in the Harmoniai there were divisions of the tetrachord which were not found in the Aristoxenian theory of the gevera, That particular point is not developed. Instead we have an account of the interval sequence of the ‘Harmoniai, and at the end we are biden to notice that in each of them the ‘enharmonic diesis is to be heard; not a word is added concerning the structure of the tetrachords in these scales, Yet the inference from this * wobbling" is, not necessarily that the evidence of Aristides is untrustworthy. Tt is well within the range of possibility that Aristides found this digression in his authority, but we are not forced to press that point. Even if the blemish be 2THE MUSICAL SCALES OF PLATO'S REPUBLIC — 27 due to a lack of skill on the part of the unfortunate compiler it is perfectly explicable. Aristides, we muy suppose, reached a point where his authority in a passing reference to the Harmoniai pointed ont that their tetrachordal structure placed them outside the theory of the gexeva. ‘Then remembering that elsewhere there was a fuller account of these Harmoniai, Aristides promptly inserted it without hiding the suture completely. G) In the excerpt itself a promise is made to explain the reason of certain anomalies in the Harmoniai at a later stage; that promise is not kept. We may of course believe that the forger would draw attention to his roguery in {this unskilful manner; yet we may equally well believe that Aristides’ authority promised an explanation, and actually did give one in a later past of his work, bbut that Aristides did not carry his quotation far enough (@) Fortunately we are not dependent on general probabilities in discussing these scales ; we can put them to several tests: @) Aristoxemus (quoted by Plutarch, Tlepl poveveis, cap. 15, p. 11360) tells us that Lamprocles of Athens, differing from his contemporaries, believed the ‘disjunctive tone’ of the Mixolydian to be at the top of the scale. If we ‘enquire why it was left to Lamprocies (at the beginning of the fifth century ¥.G,) to decide this matter (which should have been obvious to everyone if the Mixolydian Harmonia had been exactly the same as the Aristoxenian ‘octave species ' of the same mame) the answer is to be found in the scales as given by Aristides. There we find that at the top of the Mixolydian there was a large interval of three whole tones. That large undivided interval obscured the trae position of the ‘disjunctive tone,’ and Lamprocles was the first to decide (at Jeast theoretically) how the large interval should be broken up. It was to be divided not into tone and ditone, but into a ditone followed by atone. With- ‘out the Mixolydian scale of Aristides, however, this reference to Lamprocles would have been unintelligible. (0) The small fragment of the music to the Orestes of Euripides which has been preserved is mutilated, but not so badly that we cannot discover the scale in which it is written. It is worthy of notice that the scale used in that fragment corresponds to the first six intervals of the Phrygian scale of Aristides and—this is the remarkable feature—to no other scale of which we have knowledge. (©) Elsewhere? I have demonstrated in detail that there is a distinct correspondence between these seales and the Aristoxenian ‘ octave-species * (fn 700 61d rac’x). ‘The octave-species of Aristoxenus were a well-balanced, homogeneous series of scales, but that they existed in this form ab initio inherently improbable, The scales of Aristides, however, supply us with their eatlier and less homogencous shape before they had been subjected to the analysis of musical theorists such as Lamprocles and Aristoxeaus, (5) We have thus far seen that the balance of probabilities is in favour of the authenticity of these scales, and that they are congruent with other data ' Journ Hel, Stdics, Vol, XL, (3920, p27128 J. F. MOUNTFORD which without them would lack explanation. We may go further now and enquire into the source whence Aristides obtained his list. (a) The mention of six tones as being the equivalent of an octave at once stamps the passage as coming from some work of the Aristoxenian school, fer the Aristoxenian theory alone supported such an erroneous equivalence. (8) In a passage to be quoted later we shall see that Aristoxenus himself was interested in the structure of early Greek music as represented by the mythical Olympus. For the present we shall adduce a passage of Plutarch (Mepl novowrie, cap. 17, 11368). In reference to the Harmoniai he writes as follows: rovrev 8) vie dpyoviin, Tix piv Opmybucis ros obons, vis 8 aenorudens, cledras 6 Thdrev waparrredueros airas tiv Aupiori dx woheuscois dobpdare xai avibpoaw dipuitovoas cidero. ob wh Bt dyvorjoas, hs 'Apeoriterés drat & 1$ Seurépp rv Movarstv, 8n wat ey dncivaus 1 yphowwor Sy pis oderear puracinn, We have here definite evidence that Aristoxenus himself in one of his works did actually discuss the Harmoniai and Plato's treatment of them. Is it therefore possible that the passage of Aristides is based on Aristoxenus—even on that part of the Movewed which Plutarch had in mind? (0 In the digression of Aristides stress is Iaid on the fact that in all the scales one may hear the enharmonic diesis. The exalting of the enharmonic above the other genera is rather characteristic of Aristoxeniis. Subsequent accounts of the Aristoxenian theory, like the Pythagorean textbooks, do not lay any special stress on the enharmonic. But take a glance at p. 23 of the Elements of Avistoxenus himself. There we find him inveighing against the chromatic sweetness in vogue amongst his contemporaties, and praising the ancient enharmonic, ‘which, so far from being contemptible, is perhaps the most noble of all styles’ (eal ax} $avnordrn ye dAAa oyebiv eadNlorn). Further on we shall have occasion to see how interested Aristoxenus was in the origin of this favourite genus. Now the account of the Harmoniai in “Aristides clearly comes from a writer who was fond of the enharmonic and had an axe to grind. The possibility that Aristides is basing his account on a work of Aristoxenus is consequently greater. (@) The reader of the ‘Apuowed Srocycia is frequently annoyed by the habit Aristoxenus has of referring to his previous discussions of a topic and of postponing his reasons to a later stage How he mentions a subject merely to say that he will return to deal with it more fully at a later stage may be illustrated from pp. 8, 15, 18, 19, 23, and 28 of the first book alone, The unfulfilled promise in the Aristides’ passage is quite congruent (to say the least of it) with the structure of a treatise by Aristoxenus. ‘To one who considers the cumulative evidence of these four points it will pethaps not seem rash to suggest that behind the account of Aristides we can posit the authority of Aristoxenus. It is not necessary to imagine that we have here the ipsissima verba of Aristotle's pupil. Indeed, all through bis work ‘Aristides has a uniform style which bears little resemblance to the cramped phraseology of Aristoxenus; and the mention of Plato as 4 ios IIdrawTHE MUSICAL SCALES OF PLATO'S REPUBLIC 129 is typical of Aristides’ references (o the philosopher. A compiler may surely be allowed some liberty of presentation; even Pluterch, who in his Tept ovzweis is a less adept compiler than Aristides, does not seem to have been careful to preserve the precise words of his authority. The really important point for us is that, though the passage of Aristides is short, there are sufficient indications to justify the suggestion that we have here to deal with information which comes from a reputable source. Remembering in addition that the scales themselves satisfy certain crucial tests, we must be prepared to treat Aristides with more respect than has hitherto been accorded him. His scales should be looked upon as representative of the scales Plato had in mind when writing the well-knowa passage of the Republic. 1 The scales of Aristides belong professedly to the enharmonic class, and no indication is given by him that there were parallel chromatic and diatonic forms. One might suppose that such existed, since Aristoxenus (p. 19) says the diatonic scales are the first and the oldest, and the enharmonic the most recondite (dvirarey). But how far Aristoxenus is theorizing and how far he had a historical basis for his statement is not clear. For the pre-Aristoxenian period of Greek music it is best to avoid speculation and confine ourselves to certain definite pieces of evidence. (2) Aristides writes (p. 28): ‘We must also mention the eklysis, the spondeiasmos, and the ekbole, for the ancients made use of these intervals in the differences between the Harmoniai (mods ras &sagopas rv dpuomiv). The fall of a three-quartertone interval (rpisv Siicewr dovrOérav Aveots) was called fan shlysis and a rise of the same interval was called a spondeiasmas.! The name of a tise of five-quartertones was eddole, On account of their rare use they were also called ‘affections of intervals (ravra 8¢ eal rd@y rv Baer draw Bid 79 arrdnoy ris xprjeeus roacmyopetera). Platarch also (cap. 29, P. 11418) mentions the esis and the eébole, saying that they were created (rremoincévas) by Polymnestus, who represents nothing more definite than extreme antiquity. The importance of the interval of three-quartertones in carly times is furthermore evidenced by Aristoxenus (p. 37) when he points ‘out that originally some of the keys (révo.) differed from each other by that interval? To one who has only conceived of Greek music as it is presented in the fully developed Aristoxenian theory this kind of evidence must be discon- certing, for none of these intervals are taken into account. Some sort of consolation may be derived from the hint of Aristides that these intervals were not of common use; but the fact remains that there were certain intervals 1 Why the same foterval should have ono found In the enharmonic grt, But the wate- ‘name fi is ascending and another ifdascending ment is abaurd i ise, He meant perbape inset lear, ‘hat they are connected with the eaharmoaic " Bacchive the Elder (37) alto mentions the gyms since ooe is a dese more, andthe other a ‘pis aad anol, Ho says that they are only cient eta than 8 toce,130 J. F. MOUNTFORD occurring in carly Greek music which do not come within the scope of the Aristoxenian theory of the genera. So far we have merely the names of these three intervals, and concerning the ekiysis and cétofe there is nothing more definite to be added. But some- thing of importance concerning the spondciasmos can be derived from a passage of Aristoxenus preserved by Plutarch (cap. 13, p. 1134F) “As Aristoxenus says, Olympus is considered by musicians as the inventor (eipenis) of the enharmonic genus, for before his time all melodies were diatonic or chromatic. And his discovery is believed to have come about the following way: Moving in the diatonic genus, Olympus took his melodies frequently to the diatonic Parhypate (F), sometimes starting from the Paramese (B), sometimes from the Mese (A), skipping the diatonic Lichanus (G). He noticed the beauty of the elect, admired the scale con- structed on this analogy (18 é¢ 19s dvadoyias ewveargeis otierqua), and com: posed melodies in it in the Dorian scale (ni 709 Awpiov révo). For in this ‘way he did not touch upon the characteristic notes of the-diatonic or chromatic, for enharmonic genera, Such was the character of his first enharmonic com positions, Musicians consider the most ancient of them is the Spondeion Song, in which is manifested the characteristic of none of the ordinary divisions of the totrachord, unless, indeed, one considers the higher-spondeiasmos (ovrrove- ‘repos o.=in the upper tetrachord ?) and judges it to be diatonic. But it clear that such an equation would be false and unmelodic—false because the interval of the sponideiasmos is a diesis less than the tone interval which adjoins the Mese, and unmelodic because if one equated the interval of the spondeiasmas to the value of a tone the result would be two ditones in succession, one undivided and the other divided. Such, then, was the nature of the first enharmonic melodies. For the pair of small intervals (roxvéy) which is now used in the tetrachord Meson does not appear to be employed by Olympus. It is easy to realize this if one listens to someone playing the ‘tute in the ancient manner, for he endeavours to keep the semitone of the tetrachord ‘Meson undivided. Later on, in Lydian and Phrygian melodies, this semitone was divided.” ‘As it stands this passage is incoherent either through the fault of Aristoxenus or the poor arrangement of Plutarch. If there is not a confusion of thought, there is at least illogicality of presentation; we may be glad that the information we wish to derive from the passage is independent of these defects. ‘Aristoxenus commences to explain the ‘ origin ’ of the enharmonic genus. ‘He supposes that Olympus (another synonym for antiquity) used two downe ward progressions: B-F and A-F, and in each case omitted to insert Lichanus—i.e. G. But neither progression has any relation to the enharmonic unless the Hypate (E) is also added; and thie note is implied by Aristoxenus when he speaks of the tetrachord Meson (E, F,G,A). Thus far we obtain the scale in descending order: B, A, F, E. As Aristoxenus says, there is no noteTHE MUSICAL SCALES OF PLATO'S REFUBLIC 131 which is characteristic of any of the three genera: the A and E are immovable notes ($06y701 éardires), and the F could be regarded either as the diatonic or chromatic Parhypate or as the enharmonic Lichanus. Towards the end of the paragraph, after a digression, Aristoxenus explains that it was only after ‘Olympus had foreshadowed the undivided interval of a ditone in making his progression B, A, F, E, that the semitone (F-E) was divided, and so formed a real enharmonic scale. The susnetotal of which is that Olyshpus did not himself discover the enharmonic! In the middle of this fantastic explanation of the origin of the enharmonic there is inserted « note on one of the ‘enharmonic’ compositions of Olympus which were believed to be extant in the timeof Aristotle." In that composition the spondeiasmos was used, and we can deduce from the statements of Ais toxenus the position which it occupied in the sesle* It mast have been immediately above the sequence we have deduced previously: the scale ascending was E, F, A,B, C+, ie. 42, 1,4. As Aristoxenus points out, if the spondeiasmos had been equal to a tone we should have had the sequence att Here, then, we have another scale to place beside those of Aristides as genuinely representative of early Greek music such as Plato admired? This scale presumably is some form of the Dorian, since Aristoxenus was thinking of the Dorian compositions of Olympus. To what genus it belongs and how it is to be correlated with the scales of Aristides are matters which will become clearer in the sequel. ML In the cvidence we have just considered in the preceding sections we have been concerned with certain technical terms, such as diatonic, enharmonic, tone, and disis. These terms need some discussion if we are to arrive at 4 reasonably true view of early Greek music. In the Aristoxenian theory the words diesis, semitone, and tone are of frequent occurrence in the discussions of the genera. The enharmonic diesis is regarded as an interval, which, if taken four times, will make up a whole tone. On that supposition the scales of Aristides were transcribed above, For the Aristoxenians the range of possible sounds resembled a straight line, any point of which could be a note and any section an interval On this linear view of sound any interval was capable of division into an infinite number of parts, since a line could be considered as an infinite number of points. The modern view of sound and intervals is quite different. We know that an interval is determined by the different rates of vibration of the notes 2 Ch Pate VET, cap. 5 340. completed by the upper I, 26 the words de rit ‘of matic epoken of as arodadfur draheries aera to imply. Tho reader who cout pares this scale withthe Donan of Aristides wal otic that the coal of Actdee would descend that fact ha already 0. BE. 2930). sand Clooaidee, cep. 2.1 which form it, aud that the exact meastremont of an interval demands the use ‘of ratios, since the size of the interval is expressed by a comparison of two rates of vibration. From this it follows that, if two intervals are added together, the resultant interval is expressed by the product of the ratios of the constituent intervals, and the halving of an interval is expressed by the square oot of the original ratio. The Aristoxenian tone is the interval between a Perfect Fourth and a Perfect Fifth, and its ratio is therefore 9:8. This ratio, however, has no rational square root and no rational fourth root. There cannot, therefore, be a mathematically exact semitone or quartertone. You may obtain for musical purposes an interval which is very nearly a quarter- tone, but you will never get one which, when it is taken four times, makes up the interval of a 9:8 tone exactly. Thus the sAristoxcnian theory, forall its surface rigidity, is vague on the question of the precise nature of the intervals, it discusses. For practical purposes in Greece such vagueness may not have mattered; just as we talk of a semitone without adding parentheses concert temperament.’ But we are far removed from Greece, and this laxity in the Aristoxenian system is serious. Fortunately there are writers who deal with the intervals of Greek music from a point of view other than that of Aristoxenus. In the work of Claudius Ptolemaeus we have preserved the evaluations of the intervals of the different genera from Archytas downwards: All these evaluations are the work of Pythagoreans, who knew that an interval is to be expressed by a ratio and not by a linear measurement. For the enharmonic genus we find : Archytas (400.0). ns 2827 36235, Eratosthenes (240 8.c.) 40:39 39:38) Didymus (10 4.0.2). 5233838230 Ptolemy (1504.0) 40245-24223 For the chromatic genus: Archytas aBi27 24g aag Eratosthenes zor1g 19718 Didymus 16:15 25224 Ptolemy (uadaxdy) oo. one 28:27 »— (odrvove8) . aes ar For the diatonic genus and its varieties: ‘Archytas 28:27 Eratosthenes, 243 Didymus 15 Plolemy (uadaxé») we 2020 (rovaion) «.. 27 Crromaiony 243 (@irrore) wn 15 eC HrTHE MUSICAL SCALES OF PLATO'S REPUBLIC 133, ‘What do these evaluations represent? Certainly not any gradual change which took place in the genera during the ages between Plato and Ptolemy, for in the enharmonie genos the evalvations of Didymus are much closer to those of Archytas than they are to those of Eratosthenes. Neither does it follow, because Ptolemy attributes an evaluation to Didymus, that such a series of intervals only came into vogue during the time of Didymus; it may ‘equally well be that Didymus was the earliest authority known to Ptolemy for this particular evaluation. Much of the Pythagorean learning again was traditional, and the fact that an evaluation is found for the first time in Plolemy does not necessarily mean that it represents a new development in the music of Ptolemy's own day; indeed, the Tonic diatonic of Ptolemy is identical with the diatonic of Archytas. From these evaluations, however, we can Tearn what kind of tetrachord would be called enharmonic or chromatic by Archytas, Eratosthenes, Didymus, or Ptolemy. Each one of these authorities has taken what scemed to him a typical tetrachord in cach of the genera, and has given us his ratios for it ‘The amazing thing about these evaluations is the great divergence between them. Let us take the two smaller intervals of the enharmonic genus. They are of various sizes, which progress as follows: 40545 40539 39738 36535 Gaign BHi30 B8ia7 2423 The smallest of them is appreciably less than a quarter of a g:8 tone; the largest is larger than a third of a g:8 tone. The chromatic intervals range from 28: 27, which is as small as some of the enkarmonic intervals, to 12:11, which is distinctly larger than half a 9:8 tone, and larger than some of the intervals used in the diatonic genus. The diatonic semitone intervals range from 28:27 to 12:11, and the diatonic tone intervals ftom 11:10 to 8: Now can we draw any information from these tables? I think so. It will be noticed that, although in the case of the enharmonic there is a great varicty in the size of the small intervals, there is 2 consensus of opinion that the large interval of that genus is $:4, or a major third, Eratosthenes' evaluation is the only exception to this, and I think we may say that the really stable feature of the enharmonic genus as given in these evaluations is the size of the large interval ; the remaining intervals vary among themselves. Similarly in the chromatic genus the largest interval is in the majority of cases a lesser third of the value 6:5; the other intervals vary considerably. ‘The feature of the various forms of the diatonic is the presence of two intervals of a tone in each tetrachord, the size of the tone ranging between 10:9 and 8 : In examining the scales given here by Ptolemy, we find that the distinction of genus, except for the principles just deduced, is by no means as clear as the Aristoxenian theory postulates. In the evaluations of Archytas, to which no fone can raise any objection on the ground of date, we find that the first interval in all the three genera is 28:27 yet the Aristoxenian theory would demand that the first interval of the enharmonic genus should be smaller than134 J. F. MOUNTFORD the first interval of the chromatic. As far as the evidence of Archytas goes, the difference of genus consists, nbt in the size of the smallest interval, but in the size of the largest. Again we notice that in no case are the two smaller intervals of the tetrachords of the same size in any one evaluation. In some cases, like the chromatic of Didymus, there is a distinct difference; yet the Aristoxenian theory has little to say about divisions of the tetrachord in which the two smaller intervals are not of the same size, In short, I would maintain, firstly, that the distinction of genus depends essentially on the size of the largest interset ; and, secondly, that, as far as the smalley infervals ave concerned, one genus merged imperceptibly into another, The rigid Aristoxenian theory mis- represents or slurs over these points, and modern students of Greek masic seem to have been misled by it. There are, however, in the work of Atis- toxenus himself signs of the rigid theory in the making. On p. 19 he says : “Our next step will be to enumerate the genera into which melody in general may be divided. There are apparently three in number (paiveras 8 cis rela). Any melody that is harmonized on one principle (xéhor rap eis ‘ratr® jipuorpéras) is diatonic or chromatic or enharmonic.’ On p. 35 he gives some interesting sidelights: “Of these parts of Harmonics, one and the first is to define the genera, and to show what are the permanent and what are the changeable elements presupposed by this distinction. None of our predecessors have drawn this distinction at all; nor can this be wondered at, for they confined their attention to the enharmonic. genus, to the neglect of the other two. Students of instruments, it is true, could not fal to distinguish each gens by ear, but none of them reflected even on the question—at what point does the enharmonic begin to pass into the chromatic? For their ability to discriminate extended not to all “shades” (yeéas), inasmuch as they were not acquainted with all styles of musical composition or trained to exercise 2 nice discrimination. ‘These reasons sufficiently explain why the genera have not as yet been definitely distinguished ; but itis evident that we must supply this deficiency.”* How this deficiency was remedied may be seen from the pedantic passage (pp. 22-27) of the fragments of Aristoxenus. He set himself to shot off the genera into watertight compartments. Furthermore, he only professed to deal ‘with melodies which were constituted on ‘one principle’; that is to say, his theory did not attempt to fit any melody which contained both enbiarmonic and chromatic intervals. Nor does his theory take into account melodies in which the two enharmonic intervals were of different sizes. To explain the fact that all chromatic and diatonic semitones are not of the same size, an elaborate theory of ‘colours’ or ‘shades’ was superimposed on to the theory of the genera. Tf we are to understand the nature of the Harmoniai, we must be prepared to banish this artificial and elaborate theory of Aristoxenus and get behind it to the phenomena which it tries to explain. Doubtless in the actual practice 1 acres tration,THE MUSICAL SCALES OF PLATO'S REPUBLIC 135, of musicians the Dorian Harmonia was sometimes of an enharmonic nature, sometimes diatonie or chromatic, but the boundary lines were not meticulously defined. Indeed, one tetrachord of the scale might be of a chromatic nature ‘while the other was diatonic ; so long as the general sequence of large and small intervals remained the same and their relations to each other were stable, a Greek composer would have very great liberty in his choice of nuance. Regarded from this point of view, the scales of Aristides will seem to have more life. We are no longer compelled to think of them all ax employing that vague monstrosity the ‘quartertone"; we need not wonder how a Greek singer managed to sing two of the quartertones in succession. Tf one looks at the enharmonic tetrachord of Archytas, itis clear that the two sinall intervals ‘were not precisely the same, and that the really important thing was that they should together make up a semitone, leaving a major third to complete the tetrachord. These ratios of Archytas represent for us the kind of intervals ‘which were in use in the time of Plato, and we can apply them to the scales of Aristides with very little difficulty. As is well known, the interval of the Perfect Fourth was fundamental in Greck music: when we find the sequence “diesis, diesis,ditone,’ in the list of Aristides we can justifiably substitute the tetrachord of Archytas, As a check we can make use of the fact that the ratio representing the octave is 2:1. In some cases the ditone is divided up into two separate tones in the scales of Aristides, and the obvious way to divide the ratio $:4 is into 9:8 and 30:9. Working on these principles, then, we obtain the subjoined list of seales to which a modorn pianoforte could be tuned without any difficulty: Lydian... 36235 Dorian... 9:8 Phrygian, 9:8 Tonian 28:27 Mixolydian ... 28:27 ‘Syntonolydian 28:27 10:9 (or 9:82). 28:27 36:35 45:32 ‘The interval of the spondeiasmos remains to be discussed from the point ‘of view of these Pythagorean ratios. The nearest ratio to three-dieses is 32:11, such as we find in the Syntonon chromatic or the Homalon diatonie of Ptolemy. As we saw previously, the spondeiasmas was the first interval in the upper tetrachord of the scale of Olympus, and that fact leads us to look to the Homalon diatonic rather than to the Syntonon chromatic for our identi cation, Indeed, it seems to have been a prevalent view that the intecval was a diatonic one, since Aristoxenus expresses disagreement. His grounds, how- ever, are not very cogent. He says it cannot be diatonic, because the interval is a diesis less than a tone; but Aristoxenus had his own ideas about what constituted the diatonic genus, and it is noticeable that he does not suggest any alternative. His second objection is contemptible, for the spondeiasmos ‘was professedly not equal to @ tone; to imagine what would happen if336 THE MUSICAL SCALES OF PLATO'S REPUBLIC is just avoiding the point at issue. In short, there was nothing very abstruse in the scale of Olympus. He was using a Dorian scale (E, F,G, A, B, C, D, E). In the lower tetrachord he omitted the Lichanus (G), so thiat it was not immediately clear whether the scale was diatonic, chromatic, or enbarmonic. In the upper tetrachord (B-E) his first interval was sfondciasmos, which indicated that he was employing a certain (unusual) type of Diatonic, It would seem, therefore, that his sequence of intervals was: 8:27 9:7 9:8 rasax (ato) ‘The last interval, taking the range of the scale of Olympus up to E, is probable but conjectural. The relation of this scale to those of Aristides now occasions no difficulty; it is a diatonic form of one of the Harmoniai (the Dorian) which Aristides gave in the enharmonic form Since most students of Plato have not the leisure to worry over the abstcuser points of musical theory and acoustics, it may be useful to sum up here the position which this article tries to establish + There is reliable evidence for the actual musical scales to which Plato makes reference. They were of a semi-primitive nature, and lacked that homogeneity which a fally developed musical system would show. They differed from each other by'the varying sequence of the larger and smaller intervals of which they were composed. ‘The extant list of scales shows them in the enharmonic form, but parallel chromatic and diatonic forms may be conjectured, and in one instance demonstrated. Finally, the precise intonation of the scales can be reconstructed with the aid of the musical ratios of Arcbytas. J. F. Mountrorp. ‘Tae Unovarstry, Rommmenes, 2 Te the Syotonon chromatic of Ptolemy there seat the “ekbele" fs an Interval 76, which saay poasbly repre