"

The Short Life of Evariste Galois

Legend has it that the young mathematician wrote down group theory

the night before he was fatally shot in a duel. More careful investigation

shows that Galois's remarkable ideas took somewhat longer to mature

by Tony Rothman

n the early morning hours of May years later the manuscripts he had left geni us, oppressed by the stupidity of his

I 30, 1832, the French mathematical

prodigy Evariste Galois, who was
behind for Chevalier were published by
the French mathematician Joseph Liou­
teachers, ignored by the mathematical
establishment and goaded by the events
then 20 years old, wrote to his friends ville, and the extraordinarily fecund of the times into political activities that
Napoleon Lebon and V. Delauney: branch of mathematics called group squandered his energies and eventually
"I have been provoked by two patri­ theory was born. cost him his life. Most remarkable of all,
ots . . . . It is impossible for me to refuse. I according to these accounts, is that
beg for your forgiveness for not having ew tales in the history of science throughout the political turmoil and
told you. But my adversaries have put F can equal the high romance of the even during his imprisonment Galois
me on my word of honor not to inform known facts about the life and death of continued to develop his mathematical
any patriot. Your task is simple: prove Galois. Yet because the facts of the sto­ ideas in his head and finally wrote them
that I am fighting against my will, hav­ ry are so compelling it is easy to read down the night before the duel. Bell's
ing exhausted all possible means of rec­ too much into Galois's letters, and it is description of the final night is worth
onciliation; say whether I am capable of tempting to sift through the events that quoting because it has probably given
lying even in the most trivial matters. led up to the duel for an explanatory the greatest impetus to the Galois myth:
Please remember me since fate did not thread that can match the melodrama "All night long he had spent the fleet­
give me enough of a life to be remem­ apparent in his writings. ing hours feverishly dashing off his sci­
bered by my country. It is known, for example, that at age entific last will and testament, writing
I die your friend, 17 Galois was instrumental in creating a against time to glean a few of the great
E. Galois" branch of mathematics that now pro­ things in his teeming mind before the
On the same night Galois also wrote vides insights into such diverse areas death which he saw could overtake him.
to his friend Auguste Chevalier: as arithmetic, crystallography, particle Time after time he broke off to scribble
"I have made some new discoveries in physics and the attainable positions of in the margin 'I have not time; I have not
analysis. The first concerns the theory of Rubik's cube. It is also a matter of rec­ time,' and passed on to the next franti­
equations, the others integral functions. ord that at the same age Galois failed cally scrawled outline. What he wrote in
"In the theory of equations I have in­ for a second time the mathematics ex­ those last desperate hours before the
vestigated the conditions for the solv­ amination for admission to the Ecole dawn will keep generations of mathe ­
ability of equations by radicals; this has Poly technique. He studied instead at the maticians busy for hundreds of years. "
given me the occasion to deepen this Ecole Normale in Paris, but by the time Recently, with the help of M arc Hen­
theory and describe all the transfor­ he was 19 he had been expelled from the neaux and Cecile DeWitt-Morrette of
mations possible on an equation even school and twice arrested and impris­ the University of Texas at Austin, I
though it is not solvable by radicals. oned for his political activities. Shortly have studied some of Galois's writings
All of this will be found here in three before the duel he had become involved and the later scholarly work on his life.
memoirs . . . . in an unhappy love affair, which in one Although it is clear from these materi­
" Make a public request of [Carl Gus­ of his last letters he seemed to link with als that the major events in Galois's life
tav Jacob] Jacobi or [Carl Friedrich] the duel itself. "I die," he wrote, "the have been known for some time, the re­
Gauss to give their opinions not as to victim of an infamous coquette and her constructions by Bell and others reveal
the truth but as to the importance of two dupes." more about the stereotypes of scientific
these theorems. After that, I hope some Unfortunately several of Galois's genius that appeal to the popular imagi­
men will find it profitable to sort out 20th-century biographers have not re­ nation than they do about Galois. The
this mess." sisted the temptation to arrange, inter­ true romance of Evariste Galois is a fas­
Galois's desperate state during the pret and embellish such facts. The story cinating story in its own right, and it
writing of these letters was fully war­ of Galois known to most people today bears telling on the 150th anniversary of
ranted in view of the subsequent events. is derived from popular accounts, such his death.
Shortly after sunrise on the morning he as those by the physicist Leopold In­

II. other contemporary documents the

completed the letters he left his room at feld and the astronomer Fred Hoyle. Apart from letters, official records and
the pension Sie ur Faultrier in Paris and The most influential version of the story
confronted a political activist named has been that of Eric Temple Bell, the principal source on the life of Galois is
Pescheux d'Herbinville in a duel of hon­ mathematician whose 1937 book Men an 1896 biography by Paul Dupuy, a
or on the banks of a nearby pond. There 0/ Mathematics is probably the best­ historian and the general superintendent
Galois was shot in the abdomen and known introduction to the lives of great of the Ecole Normale, the college Ga­
abandoned . A passerby found him and mathematicians. lois had attended 66 years earlier. Ac­
he was taken to the Hopital Cochin, In the popular retellings of the tale cording to Dupuy, Galois was born on
where he died the next day. Fourteen Galois is presented as a misunderstood October 25, 1811, in Bourg-la-Re ine, a

136

© 1982 SCIENTIFIC AMERICAN, INC

suburb of Paris. His father, Nicholas­
Gabriel Galois, supported N apoleon
and headed the town's liberal party; he
was elected mayor of Bourg-la-Reine in
1815 during the Hundred Days, Napo­
leon's first return from exile.
For the first 12 years of his life Eva­
riste was educated by his mother, Ade­
laide- M arie Demante Galois. She gave
her son a solid background in Greek and
Latin, and she passed on to him her
skepticism toward established religion.
It is unlikely, however, that the young
Galois was exposed to mathematics in
any more detail than the usual lessons
in arithmetic; a mathematics education
was not considered particularly impor­
tant at the time. There is no record of
previous mathematical talent on either
side of the family.
Galois's formal education began in
1823, when he was enrolled in the Col­
lege Royal de Louis-Ie-Grand, the Paris
preparatory school that was the alma
mater of Robespierre and Victor Hugo
(and is still operating today). At Louis­
Ie-Grand, Galois immediately began to
develop his political sensibilities. His
liberal, or antiroyalist, sympathies, ac­
quired from his parents, were in accord
with the political opinions of most of
the other students.
During Galois's first term, however,
relations between the students and a
newly appointed headmaster of the
school were badly strained. The stu­
dents suspected the headmaster of plan­
ning to return the school to Jesuit ad­
ministration; the Jesuits were leaders
of the right-wing backlash that followed
the N apoleonic era. The students staged
a minor rebellion: they refused to sing
at chapel, to recite in class or to toast
Louis XVIII at a school banquet. The
headmaster summarily expelled 40 stu­
dents he suspected of leading the insur­
rection. Although Galois was not ex­
pelled (and it is not known whether he
participated in the uprising), the arbi­
trary action of the headmaster undoubt­
edly helped to foster Galois's distrust
of authority.

here is little evidence that Galois

T was a poor student or that his intel­
lectual growth was stunted by inferior
teachers at Louis-Ie-Grand, as the pop­
ular accounts would have it. In his first
few years he won several prizes in Greek
and Latin and half a dozen honorable
mentions. The historian of science Rene
Taton calls his progress brilliant. Nev­
ertheless, during Galois's third year his
work in rhetoric was inadequate and he
had to repeat the year. Contrary to Bell's DRAWING OF GALOIS by David A. Jobnson depicts tbe mathematician at the age of 17,
statement that Galois's poor work in while he was a student at the College Royal de Louis-Ie-Grand. At the time Galois had studied
rhetoric was a result of his preoccupa­ mathematics for only two years, yet he had already published a paper on continued fractions
tion with algebra, it was only after this and had begun the investigations into the theory of equations that led him to consider an ab­
stract algebraic theory of sets of objects he called groups. Credit for the development of group
setback that Galois enrolled in his first
theory must also be given to several other mathematicians of the late 18th and early 19th centu­
course in mathematics. He was then 15. ries, notably Paulo Ruffini, Neils Henrik Abel and Joseph Louis Lagrange. The distinction of
The course, taught by Hippolyte Jean founding group theory, however, is usually accorded to Galois. Johnson's drawing is based on
Vernier, awakened Galois's genius for the two known renderings of Galois. One was done when Galois was 15 and the other was
mathematics. He raced through the usu- completed from memory by Galois's brother Alfred in 1848, 16 years after Evariste's death.

137

© 1982 SCIENTIFIC AMERICAN, INC

 4$\
\'

138

© 1982 SCIENTIFIC AMERICAN, INC

al texts and went straight for the mas­ most difficult problems in mathematics,
ters of the day, devouring Adrien Marie a problem that had confounded mathe­
Legendre's work on geometry and the· maticians for more than a century.
memoirs of Joseph Louis Lagrange: The In 1829 the central question for the
Resolution 0/ Algebraic Equations, The theory of equations was: Under what
Theory 0/ Analytic Functions and Lessons conditions can an equation be solved?
on the Calculus 0/ Functions. It was un­ More precisely, what was sought was a
doubtedly from Lagrange that Galois method of solving an equation having a
first learned the theory of equations, to single variable x whose coefficients are
which he was to make fundamental con­ all rational numbers and whose highest­
tributions over the next four years. Ver­ power term is xn. The method was to be
nier seems to have appreciated his stu­ a general one that could be applied to all
dent's talents: his remarks in Galois's such equations, and it was to rely on
trimester reports carry high praise such only the four elementary operations of
as "zeal and success" and "zeal and arithmetic (addition, subtraction, multi­ IDEA OF A GROUP can be illustrated by
progress very marked." plication and division) and the extrac­ considering the group S(3), which is the group
With his discovery of mathematics tion of roots. If the solutions or roots of of permutations of three objects. An element
of S(3) operates on the objects by rearranging
Galois's personality underwent a strik­ an equation can be obtained from the
them. The permutation (123) moves the ob­
ing change. He began to neglect his oth­ coefficients solely by these operations, ject in the first square into the second, the ob­
er courses and aroused the hostility of the equation is said to be solvable by ject in the second square into the third and
his teachers in the humanities. His rhet­ radicals. the object in the third square into the first.
oric teachers called him "dissipated" on From a historical perspective it was Because there are six possible arrangements
the trimester reports, and the words natural to expect that solving an equa­ of three objects there are six elements in S(3).
"withdrawn," "bizarre" and "original" tion of the nth degree would call for no
appear on his evaluations. Even Ver­ operations more elaborate than the ex­
nier, although not seeking to cool Ga­ traction of nth roots. The solution to the
lois's passion for mathematics, urged general quadratic, or second-degree, theory with applications far outside the
him to work more systematically. Ga­ equation ax2 + bx + c =0, known to theory of equations, now called the the­
lois did not follow the advice: he decid­ the Babylonians, requires the extraction ory of groups.
ed to take the entrance examination for of the square root of a function of the
the Ecole Polytechnique a year early coefficients, namely b2 - 4ac. Hence the alois submitted his first papers on
and without the usual preparatory cour­ general quadratic equation is solvable G what was to become group theory
ses in mathematics. Evidently lacking by radicals. Similarly, the general solu­ to the French Academy of Sciences on
some of the basics, he failed. tion to the cubic equation, devised by M ay 25 and J une 1, 1829, near the end
the Italian mathematicians Scipione dal of his final year at Louis-Ie-Grand. Less
alois considered the failure an injus­ Ferro and Niccolo Fontana, or Tarta­ than two months later he was to take
G tice, and it hardened his attitude glia, in the early 1500's, requires tak­ the entrance examination to the Ecole
toward authority. Nevertheless, he con­ ing cube roots of functions of the co­ Poly technique for the second time, but
tinued to progress rapidly in mathemat­ efficients. The solution to the general meanwhile events in his life took an un­
ics and enrolled in the more advanced fourth-degree equation, first achieved fortunate turn. On J uly 2, a few weeks
course at Louis-Ie-Grand taught by by the Italian mathematician Lodovico before the examination, Evariste's fa­
Louis-Paul-Emile Richard, a distin­ Ferrari at about the same time, requires ther suffocated himself in his Paris
guished instructor. Richard immediate­ the extraction of fourth roots. apartment. The Jesuit priest of Bourg­
ly recognized Galois's abilities and By the time of Galois, however, near­ la-Reine had forged Mayor Galois's
called for his admission to the Ecole ly 300 years of effort had not yielded a name to a number of malicious epi­
Poly technique without examination. Al­ solution by radicals to general equations grams directed at Galois's own rela­
though the recommendation was not fol­ of the fifth degree or higher. A number tives. The senior Galois could not bear
lowed, Richard's encouragement pro­ of mathematicians had come to suspect the scandal. The entrance examination
duced spectacular results. In March, that such general solutions are not possi­ therefore took place under the worst­
1829, while Galois was still a student, ble, even though in special cases, such possible circumstances. Furthermore,
his first paper was published. It was ti­ as the equation x7 - 2 =0, the solution Evariste apparently declined to follow
tled "Proof of a Theorem on Periodic can be found by radicals. (In this in­ the examiner's suggestions for exposi­
Continued Fractions" and appeared in stance one solution is 7 V2.) Galois pro­ tion and was failed for a second and
Annales de mathematiques pures et ap­ vided definitive criteria for determining final time. The two disasters crystallized
pliquees of Joseph Diaz Gergonne. whether or not the solutions to a given his hatred for the conservative hierar­
The paper, however, was a minor equation can be found by radicals. Per­ chy then ruling France.
aside. Galois had already turned to the haps even more remarkable than Ga­ Forced to consider the less prestigious
theory of equations, the topic he had lois's findings in the theory of equations Ecole Normale (then called .the Ecole
first explored in the works of Lagrange. were the methods he developed to study Preparatoire), Galois took the baccalau­
At age 17 he was taking on one of the the problem. His investigations led to a reate examinations required for entry in
November, 1829. This time he passed
on the basis of an exceptional score in
mathematics, and he was given universi­
MARGINAL NOTE on one of the papers left behind by Galois on the morning of his duel is ty status at about the time his first papers
the most famous of the documents cited in support of the legend that Galois wrote down his on group theory were to be presented to
ideas on group theory in a single night. The note reads: "There is something to complete in this the Academy of Sciences. The papers,
demonstration. I do not have the time (author's note)." ("II y a quelque chose a completer however, did not receive a hearing.
dans cette demonstration. Je n'ai pas Ie temps (Note de I'A.).") According to the familiar ac­
When the papers were received by the
count of the life 9f Galois by Eric Temple Bell, the phrase "I do not have the time" is written
frequently in the manuscripts. Actually the page reproduced here is the only place the phrase
Academy, Augustin Louis Cauchy, then
appears. The rapid handwriting of the note contrasts sharply with the careful and deliberate the most eminent mathematician in
hand of the body of the text, which suggests that Galois did not write the paper the night before France and a staunch supporter of the
the duel but only edited it. Indeed, the paper had already been submitted to the Academy of conservative restoration, was appointed
Sciences and returned to Galois by Simeon Denis Poisson with suggestions for reworking it. referee. Cauchy had already investigat-

139

© 1982 SCIENTIFIC AMERICAN, INC

 ed permutation theory, a forerunner of
group theory, and he later wrote exten­
sively on group theory itself. Although
legend has it that Cauchy lost, forgot or
discarded Galois's manuscripts, it is far
more credible that Cauchy recognized
their importance and handled them with
care. Indeed, a letter discovered in 1971
by Taton in the archives of the Academy
makes it clear that on January 18, 1830,
Cauchy had planned to give Galois's re­
sults a full hearing before the Acade­
my. Cauchy wrote: "I was supposed to
present today to the Academy ...a re­
port on the work of the young Galois.
... Am indisposed at home. I regret
not being able to attend today's session
and I should like you to schedule me
for the following session for the ...in­
dicated subjects."
The following week, however, when
(12) • (123) = (13) Cauchy read a paper of his own to the
Academy, he did not present Galois's
"MULTIPLICATION" of one element in S(3) by another element is carried out by determin­
work.Why this happened remains a sub­
ing the arrangement of objects that results from the operation of the first permutation and
ject of speculation. Taton conjectures
then applying the second permutation to this arrangement. The single permutation that brings
that Cauchy urged Galois to expand his
about the same rearrangement is called the product of the two permutations. In general the
multiplication of groups is not commutative: the product of two elements depends on the se­ work and submit it for the Academy's
quence they are applied. Thus (12) * (123) is equal to (13) but (123) * (12) is equal to (23). Grand Prize in mathematics. Although
Taton's conjecture cannot yet be docu­
mented, Galois did submit an entry for
SECOND ELEMENT the prize in February, the month before
the deadline. The entry was sent to Jean
Baptiste Joseph Fourier, the mathemati­
(1) (123) (132) (12) (13) (23) cian who invented what is now called
Fourier analysis, in his capacity as Per­
petual Secretary to the Academy. Fou­
rier died in May, however, and Galois's

(1) (1) (123) (132) (12) (13) (23)

manuscript could not be found among
Fourier's effects. Galois later attributed
his bad luck to malicious intent on the
part of the Academy, accusing the prize
committee of rejecting his paper out-of­

(123) (123) (132) (1) (23) (12) (13)

hand because his name was Galois and
because he was still a student. The Ga­
I­
lois legend passes on such accusations at
Z face value, but there can be little doubt
w
:2 that Galois's attitude toward authority
(132) (132) (1) (123) (13) (23) (12)
w
-'
W was becoming somewhat paranoid.
I­
(fJ

I
II: n spite of the setbacks Galois re­
u::
mained a productive mathematician

(12) (12) (13) (23) (1) (123) (132)

and began to publish in Baron de Fe­
russac's Bulletin des sciences mathe­
matiques. astronomiques, physiques et chi­
miques, a far less conspicuous forum
than the meetings of the Academy. His

(13) (13) (23) (12) (132) (1) (123)

articles make it clear that in 1830 he had
progressed beyond all others in the
search for the conditions that determine
the solvability of equations, although he
did not yet have the complete answer in
(23) (23) (12) (13) (123) (132) (1) hand. By January, 1831, however, he
had reached a conclusion, which he sub­
mitted to the Academy in a new mem­
oir, written at the request of the mathe­
MULTIPLICATION TABLE for the six permutations of three objects provideS a verification matician Simeon Denis Poisson. The pa­
that the permutations satisfy the properties of a group. For every pair of permutations a and bthe per is the most important of Galois's
table shows that their product a * b is itself a permutation. There is an identity element, namely
works, and its existence more than a
the element (1), with the property that a * (1) is equal to a. For every element a there exists an
year before the duel makes nonsense of
element called the inverse of a, or a-I, with the property that a * a-I is equal to (1). The in­
verse of (123), for example, is (132). Finally, the associative law, which states that for any per­ the story that all Galois's work on the
mutationsa, bandcthe producta * (b * c) is equal to (a * b) * c, can be checked in the table. The per­ theory of groups was written down in a
mutations in color form a subset of the six permutations. Their multiplication table, also in col­ single night.
or, shows that they too form a group. Such a group within a group is called a proper subgroup. In order to understand Galois's work

140

© 1982 SCIENTIFIC AMERICAN, INC

 So contrary to anything else approach the problem from the
HILPING PRONATORS you may have heard, pronation is ground up. Right under the
GITIACK not an unnatural act. Everybody 5penco® sockliner we put a soft
does it. Some more than others. orthotic to cradle the heel and bring
ON THIIR FliT. From our own survey and rock and roll under control.
It's beginning to sound like the those of others, it appears that about After a few more innovative
only reason runners get hurt nowa­ 12 percent of the population are touches, we wound up with a shoe
days is because they have one likely to suffer because they pro­ that proved effective in reducing
unspeakable fault. They pronate. nate too much. About 8 percent, pronation an average of 22 percent.
Depending on whom you talk because they pronate too little.
to, the wages or pronation include All very interesting. But if
everything from stress fractures high-speed cameras are required to
and AchiHes tendinitis to crippling see what's really going on, how
knee injuries. can you tell if you even have a
Not much fun. And not sur­ problem? Much less what it is?
prising that a lot of people are out You can't for sure. But you
there trying to put a stop to it. can get a good idea the next time TIME (MS) 50 100
But is pronation really such a you take a shower. Make two sets
Two different shoe models tested by nine runners.
ba habit? W hat, in fact, is it? of footprints. One while standing, After 54 trials, shoe A exhibits an average of 4.9·
one sitting. A certain change in less maximum pronation than shoe B.
shape is normal, but too little If you need it, we've got it.
change means your feet are rigid. But make sure. Because as bad as
Worry less about pronation and pronation can be, we suspect it's
more about adequate cushioning. getting a bum rap. Many of the
The kind you'll find in our Colum­ problems laid at its feet may
bia, Aurora and Terra TIC. actually be the result of overtrain­
If, on the other hand, the sec­ ing. Especially when the difficulty
Supinated Neutral Pronated ond set is much fatter than the first, doesn't recur in the same part of
your feet are flexible. And likely to the body every time.
Computer digitized film data show that most runners' find pronation irresistible. We're not the only ones with
feet first contact the ground in a supinated position If that's the case, there are an opinion, however. There are
then roll through a neutral into a pronated position.
things to look for in a shoe. Like a· several excellent works on the sub­
Well, unseen by the naked stiff heel counter. Make sure it fits ject, including Harry F. Hlavac's
eye, as your feet are running for­ snug, otherwise it won't be effec­ The Foot Book.
ward, they are also rolling from one tive. Normally, you're also better off So listen to the experts. Listen
side to the other. Landing first on with a firm midsole. And good to your doctor. But more important,
the outer edge, moving to the neu­ arch support. listen to your body. It has a vested
tral plane, then continuing inward Those are just a few of the interest in your well being. And if
to what is called a pronated properties we designed into the something isn't working, your body
position. Nike Centurion and Internationalist. is sure to ret you know.
Frankly, they have good rea­ But when it came to the But please.
son to behave this way. If they Equator, we went even further. Don't make it
didn't pronate, the shock of hitting Recognizing that the base of the cry to be
the ground would be consider ­ heel bone is like a small billiard ball heard.
ably worse. that loves to roll, we decided to

© 1982 SCIENTIFIC AMERICAN, INC

it is unprofitable to study the ongl­ as inadequate, although the statement that makes them extremely difficult to
nal papers. Poisson did his best to un­ being proved could be shown to be follow, and they are not free from error.
derstand the 183 1 manuscript, but he true through a result proved by La­ With the benefit of a century and a half
finally recommended that the Acade­ grange. According to Peter Neumann of of clarification, however, it is now possi­
my reject it, encouraging Galois to ex­ the University of Oxford, the criticism ble to set forth the essentials of the theo­
pand and clarify the exposition. Poisson is completely accurate. Galois's argu­ ry in an accessible form. To this end I
also criticized one of Galois's proofs ments are presented in a concise form have had the assistance of the astro-

PERMUJ'ATIONS IN S(3) can invariably be written as the product called a normal subgroup of S(3) if for any element h in A(3) and any
of permutations that interchange only two objects. If a permutatiou elementg in S(3), the elementg* h *g-l is also an element of A(3). To
can be written as a product of au even number of such interchanges, prove thatA(3) is a normal subgroup of S(3) supposeg is an even per­
it is called an even permutation; otherwise it is an odd permutation. mutation. Then g* h *g-l is the product of three even permutations,
If an even permutation (colored circles) is multiplied by an even per­ which is also an even permutation and so is a member of A(3). If g is
mutation (colored arrows), the product is an even permutation; if an an odd permutation,g* h *g-l is the product of an odd by an even by
even pern.utation is multiplied by an odd permutation (black arrows), an odd permutation, which is again an even permutation. Hence A (3)
the product is odd. Similarly, if an odd permutation (black circles) is is a normal subgroup. By a similar argument it can be shown that for
multiplied by an even permutation, the product is odd, whereas if an any number n, A(n) is a normal subgroup of S(n). The number of ele­
odd permutation is multiplied by an odd permutation, the product is ments in a subgroup must divide the number of elements in the parent
even. The even permutations form a subgroup, namely the subgroup group without remainder. Because A(n) bas half as many elements
printed in color in the bottom illustration on page 140. This subgroup as S(n), A(n) includes the maximum number of elements that a prop­
is called the alternating group, or A(3). A subgroup such as A(3) is er subgroup of S(n) can have. A(n) is the maximal normal subgroup.

142

© 1982 SCIENTIFIC AMERICAN, INC

 'When curiosity flourishes,
worlds can be changed.
Why? How?What if?Young �eople question.Tal<ing
j oy ill the search for solutions. Their worlds abound
with endless possibilities. So, too, it is with scientists.
�,>;�' •

Whose laboratories are as illnitless as the universe. -t;

Whose ideas shape worlds. To interest j t- young minds

ill the wonders of science, P hillips Petroleum has made
20ssible a film series called "The Search for Solutions�
Stimulating films aired on PBS and seen :�I" by over
� .�
. ��\'
' two million students
per month. They
t capture the
excitement of
dlScove!y
,�!
· f"\-

And the ''C

discoverer. 1(,

To teach.To
encourag e.
But most of
all, to illterest.
chilcllil<e '-''-''-'" '-'>"-J '-'

. . ill the right

hands can help turn
darl<ness into light. m

© 1982 SCIENTIFIC AMERICAN, INC

EQUATIONS OF DEGREE 3: THERE EXIST a, b, e AND d SUCH THAT
!II THE GALOIS GROUP OF THE EQUATION
ax3 + bx2 + ex + d 0 =

ax3 + bx2 + ex + d =
0 IS S(3).
THE MAXIMAL NORMAL SUBGROUP OF S(3) IS A(3).
THE MAXIMAL NORMAL SUBGROUP OF A(3) IS I.
[S(3)/A(3)1 =3!/3 613 2.
= =

[A(3) II] =3/1 = 3.

BECAUSE 2 AND 3 ARE PRIME, S(3) IS SOLVABLE.
BECAUSE ITS GALOIS GROUP IS SOLVABLE,
ax3 + bx2 + ex + d 0 IS SOLVABLE ALSO.
=

EQUATIONS OF DEGREE 5: THERE EXIST a, b, e, d, e AND f SUCH THAT

THE GALOIS GROUP OF THE EQUATION
axs + bx' + ex3 + dx2 + ex + f = 0 251
(I 43 ax s + bx' + ex3 + dx2 + ex + f 0 IS S(5).
=

521 THE MAXIMAL NORMAL SUBGROUP OF S(5) IS A(5).

( I 43
341 THE MAXIMAL NORMAL SUBGROUP OF A(5) IS I.
231 (I 25
(I 45 [S(5)/A(5)J= 5! 1(5!/2) =
120160 2. =

321 4 31 [A(5)lll = (5!/2)/1 6011

= 60.
=
(I 45 (I 25
4 51 BECAUSE 60 IS NOT PRIME, S(5) IS NOT SOLVABLE.
341 (I 3 2 31
(I 52 (15 BECAUSE ITS GALOIS GROUP IS NOT SOLVABLE,
431 541 axs + bx' + ex3 + dx2 + ex + f
(1541 0 IS NOT SOLVABLE EITHER.
=

(I 52 (I 32
241 2 51
(I 53 (I 34 (24
31
(135
\
421 521
(I 5 3 (I 34 (25
31
(145
1
A(5) 241 31
31 41 , 2
( I 5 42 (I 3 5 (25 (23
41 (
(151
321 421 41 41
(I 54 (I 3 5 ( 35 (2:3
51
(15 1'(2
351 51 !II
(I 42 (I 23
4 1 (341
(245 (151
'
531 541 45 1 '1(3
41
(I 42 (I 23 (34
51
(2 31'( (12
35 1 ' 3
( 51 (351
(I 24 (132
1
(241 (121'
(I 24
5 31 1 '(3
41 (451
(142 (251 (121'
1 (241
(152 (12
31 (131'
31 1 1 251
(14 (124 (131'
1 (451
(125 (131'
41 ( 231
(13 (141'
(251
((41'
1
.(:35
(141

SOLUTION OF EQUATIONS was the problem for which Galois solvable if it generates a series of maximal normal subgroups whose
developed group theory. A general method of solution would employ composition factors (determined from the number of elements in the
only addition, subtraction, multiplication, division and the extraction parent group and the subgroups) are prime numbers. The composition
of roots, and it could be applied to any equation of degree n, where factors generated by S(3) and its series of maximal normal subgroups
11 is the highest power to which a variable is raised. Galois proved are all prime numbers. It follows that all third-degree equations are
that no such method exists when n is equal to 5 or more. Every equa­ solvable. When n is equal to 5 or more, however, it can be proved that
tion of degree 11 can be associated with the group S(II) or with some the maximal normal subgroup of A(n) is the identity group I, which
subgroup of S(n); the group associated with an equatioD is now called includes only the identity element. Since A (11) is the maximal normal
the Galois group of the equation. Galois showed that an equation subgroup of S(n), the composition factors of S(II) when 11 is equal to
can be solved by the methods of arithmetic and root extraction only 5 or more are not all prime numbers. Hence there are equations of
if its Galois group is what he defined as a solvable group. A group is degree 5 or greater that cannot be solved by the allowed methods.

144

© 1982 SCIENTIFIC AMERICAN, INC

physicist Adrian C. Ottewill of Oxford.
What is a group? On its deepest lev­
el group theory concerns the symme­
tties inherent in any system. Imagine a
snowflake whose points or vertexes are
equally spaced at angles of 60 degrees.
If the snowflake is rotated about an
axis through its center by 60 degrees
or any integral multiple of 60 degrees,
its basic pattern remains unchanged,
even though any particular vertex may
change its position. An operation that
leaves a pattern invariant in this sense is
called a symmetry operation.
If two rotations by integral multiples
of 60 degrees are carried out in se­
quence, the snowflake remains invari­
ant, and the position assumed by the
vertexes is one that could have been
reached by a single operation. For ex­
ample, a 60-degree counterclockwise
rotation followed by a 240-degree
clockwise rotation is equivalent to a
180-degree clockwise rotation. In gener­
al, if R(n) denotes a rotation through
60n degrees, and if the result of per­
forming first one such operation and
then another is written R(n) • R(m), then
for all integers n and m the expression
R(n) • R(m) is equal to R(n + m). Math­
ematically the equivalence states that
the "product" of two symmetry opera­
tions is also a symmetry operation.
There are three other important prop­
erties of the rotations of a snowflake.
First, a rotation through zero degrees, or
R(O), always leaves the pattern invari­
ant, since it does nothing. The product
of any rotation R(n) and R(O) is R(n), so
that R(O) plays much the same role in
rotations as the number 1 plays in ordi­ If you'd like to know some other unusual things about Lynchburg, drop us a line.
nary multiplication. R(O) is therefore
called the identity rotation. Second, a
rotation R(n) followed by a rotation
WHEN GOOD FRIENDS GET
in the opposite direction by the same TOGETHER in downtown Lynchburg,
amount, which can be denoted R( - n),
returns the pattern to its starting point. you'll never see a glass of Jack Daniel's.
Thus the product R(n) • R( - n) is equiv­
alent to R(O). The rotation R( - n) is
called the inverse of the rotation R(n).
The county where we make our whiskey is
Third, the expression R(m)' R(n) • R(P)
is unambiguous, because [R(m)' R(n)]
dry. (It voted that way in 1909.) So when
• R(P) is equivalent to R(m)' [R(n) • folks have a friendly chat, it's usually over
R(P)]. This is a formal property of the
operation " by means of which two ro­ ice cream or soda. Of course,
tations are combined, called the asso­
ciative property. we hope the law isn't
as binding in your home­
CHARCOAL
he four properties that hold for com­
T binations of snowflake rotations are
characteristic of any set of symmetry op­ town. And that, at your MELLOWED
erations on any system; they are called
the group properties. The system need
next friendly get-together, 6
not be a geometric pattern such as a
a glass of Jack Daniel's DROP
snowflake. For example, an equation is
also a system whose symmetries can be will be somewhere In 6
described by the group properties. In ab­ BY DROP
stract terms a group consists of elements the picture.
or symmetry operations a, b, c and so

bining any two elements. The elements

forth, and a rule denoted by • for com­ Tennessee Whiskey· 90 Proof· Distilled and Bottled by Jack Daniel Distillery
Lem Motlow, Prop., Inc, Route 1, Lynchburg (Pop. 361), Tennessee 37352
of the group and the rule • are assumed
Placed in the National Register of Historic Places by the United States Government.
to satisfy the closure criterion, which

145

© 1982 SCIENTIFIC AMERICAN, INC

There's a lot worth saving in this country.
Today more Americans who
value the best of yesterday are
working to extend the life of a
special legacy.
Saving and using old buildings,
warehouses, depots, ships, urban
waterfront areas, neighborhoods and
Main Streets is more than just a senti­
mental gesture. Economically it makes
good sense to restore and revitalize quality
structures. Preservation saves energy and
materials at a time when our resources are
We can also appreciate the artistry and
design of these many old structures, which are as
much a part of our unique culture as the art and
music we have given the world.
The National Trust for Historic Preser­
vation-in cities, towns and neighborhoods­
helping to keep our architectural heritage alive
for us and for our children.
Help preserve what's worth saving in
your community. Contact the National Trust,
P.O. Box 2800, Washington, D.C. 20013.

National ust for

Historic Preservation
Preservation builds the nation

© 1982 SCIENTIFIC AMERICAN, INC

states that for any elements a and b in plied and generates the arrangement
the group, a • b is also an element of the bishop-knight-rook. This final arrange­
group. The group must include an iden­ ment could have been reached in one
tity element 1, which is defined so that step from the initial arrangement by ap­
for any element a in the group, a • 1 is plying the permutation (13), which in­
equal to a. Moreover, for every element terchanges the object on square 1 with
a there must be some inverse element the object on square 3. Thus the result of
a-1 with the property that a * a-1 is the permutation (12) followed by (123)
equal to 1. Finally, the elements of the generates the same arrangement of ob­
TEXAS INSTRUMENTS
group and the operation are assumed to jects as the permutation (13) does. Sym­

ADIANCED PROIRAMIIAIIU CALCUlATORS

satisfy the associative property, which bolically, then, (12) * (123) =(13).
states that (a • b) • c is equal to a • (b * c).
The theory of groups is one of the
most fruitful areas of mathematical re­
he number of permutations or ar­
T rangements of n objects is equal to n �$170
search; Bell is correct when he writes factorial, written nL The factorial of a
that it will keep mathematicians b usy number n is the product of all the whole
for hundreds of years. One of the most numbers from 1 to n inclusive; 5!, for
important recent achievements in group example, is equal to 1 X 2 X 3 X 4 X 5,
theory was a proof announced at a meet­ or 120. Hence the number of elements in
ing of the American M athematical So­ S(n), the permutation group for n ob­
ciety in January, 1981, by Daniel Gor­ jects, is nL The number of elements in a n - 59 PROGRAMMAIII CAI.CUlATOR
• Over 175 functions and operations. 12-charecter
enstein of Rutgers University. Goren­ group is called the "order" of the group. LED display 0 Up to 960 program steps or up to 100
stein showed that a list of 26 groups S(3), the permutation group for three memories. Includes master library.

called sporadic finite simple groups is a objects, includes the 3! (or six) permuta­ n - sac PROGRAMMAIlI CALCULATOR
complete list. In a sense this finding im­ tions (1), (12), (13), (23), (123) and (132). o Over 170 functions end operations 0 Up to 480
program steps or up to 60 memories 0 Complete
plies that the components or building Here (1) is the identity permutation that programming editing features. Constant memory
blocks of any group with a finite num­ leaves any arrangement of objects un­ • Includes master library

CALL TOLL FREE 800-611·1169

ber of elements have now been exhaus­ changed.
tively classified. It turns out that certain subsets of the EXCEPT illinois, Alaska, Hawaii
set of elements in a group can by them­

I\.. that satisfies the properties of a

Another set of non-numerical elements selves satisfy all the properties of a
group, in which case they are said to
group is the group of permutations on a form a subgroup. If the number of ele­
fixed number of objects. The permuted ments in the subgroup is less than the
objects might be chess pieces, for exam­ number of elements in the parent group,
ple, or letters of the alphabet. It is essen­ the subgroup is called a proper sub­
tial to recognize, however, that the ele­ group. For example, it is easy to verify
ments of the group are neither the chess that [(1), (12)] is a group, and so it is a
pieces nor the letters but rather the func­ proper subgroup of S(3).
tions that generate the various permu­ For any proper subgroup H of a
tations. To find the "product" of two group G a number called the composi­
elements a and b of the group (that is, tion factor can be defined: it is the or­
to find a * b) one finds the result of the der of the parent group divided by the
first permutation on the set of objects order of the subgroup and is generally
and applies the second permutation to written [G/ H]. The composition factor
this result. of the subgroup [(1), (12)] with respect
Suppose three chess pieces are ar­ to the group S(3) is 6/2 or 3. Accord­
ranged so that a rook is on a square la­ ing to an elementary theorem of group
beled 1, a knight is on a square labeled 2 theory, which I shall not prove here, the
and a bishop is on a square labeled 3. order of any subgroup must exactly
One element of the permutation group divide the order of its parent group, so
for these objects can be written (12); it that the composition factor is invaria­
takes the object on square 1 and moves bly a whole number.
it to square 2 and takes the object on Galois introduced three critical con­
square 2 and moves it to square 1. The cepts whose interrelations enabled him
effect of the element (12) on the arrange­ to prove that there is no general method
ment rook-knight-bishop is to exchange for solving an equation of the fifth de­
the rook and the knight, generating the gree or higher when all solutions must
arrangement knight-rook-bishop. If the be found by radicals. First Galois noted
operation is then done again, it exchang­ that every equation can be associated
es the pieces on the same squares, re­ with a group of permutations. Such a
creating the arrangement rook-knight­ group is a representation of the symme­
bishop. Thus the group element (12) is try properties of the equation; it is now
its own inverse. called the Galois group .
Another group element, designated To appreciate the properties of the
(123), moves the object on square 1 to Galois group, consider any third-degree
square 2, the object on square 2 to equation whose coefficients are rational
square 3 and the object on square 3 numbers. It can be proved that such an
to square 1. Suppose the initial ar­ equation has three roots, although the
rangement rook-knight-bishop is again proof does not reveal whether the roots
transformed by the element (12), giv­ can be found by radicals. If the roots are
ing rise to the arrangement knight-rook­ designated u, v and w, one can form poly­
bishop. Now the element (123) is ap- nomial functions of them, such as u - v

147

© 1982 SCIENTIFIC AMERICAN, INC

"INFAMOUS COQUETTE" whom Galois blames for his troubles the initials "s" and "E" into a single cursive diagram. From letters
in a letter written the night before the duel was probably the same and other manuscripts it is clear that Galois's angry epithet is his
woman whose name appears frequently in the margins of Galois's reaction to an unhappy love affair with a woman he had met only
papers. In the manuscript re produced above the name "Stephanie" a few months before the duel. She has been identified as Stepha­
can be read under the name "Evariste," and Galois has also combined nie-FeIicie Poterin du Motel, the daughter of a Parisian physician.

or uv + w - l . Any such function can pie without knowing the values of the subgroup of S(3), for instance. is [( 1 ).
be converted into a related function by roots of the equation. For Galois's pur­ ( 1 23 ), ( 1 3 2)]. In turn the maximal nor­
permuting the roots u, v and w. For poses, however, the calculation was not mal subgroup of [( 1 ). ( 1 23 ). ( 1 3 2)] is
example. the permutation ( 1 2) inter­ necessary. All he needed to show was [( 1 )]. The composition factor defined for
changes u and v and so converts the that there are invariably equations of S(3) and its subgroup [( 1 ). ( 1 2 3 ). ( 1 3 2)]
function u - v into the function v - u. degree n whose Galois group is the larg­ is 6 / 3 or 2, and the composition factor
Many functions of the roots are changed est-possible group of permutations of for the group [( 1), ( 1 23 ). ( 1 3 2)] and its
in value by such a permutation, but the roots, namely S(n). subgroup [( 1 )] is 3 / l or 3. Since both 2
some of them are not. For example, and 3 are prime numbers. S(3) is a solv­
the function u + v + w is not changed he second concept introduced by able group.
in value by any permutation of u, v and T Galois was the concept of a normal The term solvable group is well j usti­
w. Since the group S(3) includes all pos­ subgroup. A subgroup H of a group G is fied by Galois's theory: he was able to
sible permutations of u, v and w, u + normal in G if and only if the following show that an equation is solvable by
v + w is said to be invariant under S(3). condition is satisfied: When one "multi­ radicals if and only if the Galois group
It is possible to show that the value of plies" any element h of the subgroup H of the equation is a solvable group. In
u + v + w is a rational number for any on the left by any element g of the parent order to prove that equations of the fifth
third-degree equation with rational co­ group G, and then "multiplies" the prod­ degree or higher cannot in general be
efficients. Other polynomial functions uct on the right by g - l (the inverse ele­ solved by radicals. Galois had to show
of the roots may be rational for some ment of g), the result is an element of that there are equations of this kind for
equations and irrational for others, de­ the subgroup H. Symbolically, if H is which the Galois group is not a solvable
pending on the coefficients of the equa­ normal in G, there is an element h in H I one. As it happens the group S(n) is not a
tion. If the value of such a function is such that h' =g . h • g- l. For example, solvable group when n is equal to 5 or
rational, there exists some group of per­ one can verify that [( 1), ( 1 2 3 ), ( 1 3 2)] is more [see illustrations on pages 142
mutations of u, v and w that do not a normal subgroup of S(3) [see illustra­ and 144]. Since for all such values of n
change the value of the function. The tion on page 142]. there are equations of degree n for which
Galois group of an equation is the larg­ If a finite group G has any normal S(n) is the Galois group, the general
est group of permutations that meet this subgroups at all, there must be one sub­ equation of the fifth degree or higher is
requirement for every rational-valued group whose order is the largest of all not solvable.
polynomial function of the roots. In the normal subgroups of G: it is called
other words, for any polynomial func­ the maximal normal subgroup of G: y the time Galois's work on group
tion of the roots that has a rational val­ Similarly, a maximal normal subgroup B theory was nearly completed. the
ue, every permutation in the Galois may in turn have a maximal normal sub­ events of his life had become markedly
group leaves the value of the function group of its own. and the sequence of political. In J uly. 1 83 0. the republican
unchanged. When a permutation of the maximal normal subgroups continues opponents of the restored monarchy
roots does not change the value of any until the smallest normal subgroup pos­ took to the streets and the revolution
rational-valued polynomial function of sible is reached. Any group G therefore forced the Bourbon king Charles X into
the roots, the roots are indistinguishable generates a sequence of maximal nor­ exile. While the students of the left-wing
Ecole Poly technique played an active
beled G, H. L J,
for that permutation. Hence the larger mal subgroups. If the sequence is la­
the number of elements in the Galois then .a series of maxi­
. . . • role in the fighting. Galois and his fel­
group is, the more permutations there mal normal composition factors can be lows at the Ecole Prt!paratoire were
are for which the roots are indistinguish­ defined: [GI H]. [HI 1]. [1/ J] and so on. locked up in the school by its director.
able. For this reason the Galois group is The third important concept of Ga­ Inc'e nsed. Galois tried to scale the walls;
a powerful way of representing the sym­ lois's theory is the concept of a solvable he failed and thereby missed the brief
metry properties of an equation. group. Galois calls a group solvable if revolution.
Calculating the Galois group for a every one of the maximal normal com­ Although the Bourbon abdication
given equation is generally difficult, al­ position factors generated by the group seemed a great victory for the republi­
though it can always be done in princi- is a prime number. The maximal normal cans. it proved a short-lived one. Louis

148

© 1982 SCIENTIFIC AMERICAN, INC

Philippe was placed on the throne, to have cO!ll p romised." After a fellow pris­ She later married a language professor.
the disappointment of Galois and like­ oner was shot, it seems Galois accused It is also unlikely that the man who
minded liberals. In the months follow­ the prison superintendent of arranging killed Galois was in the pay of an anti­
ing the revolution Galois joined repub­ the shooting. Galois was subsequently republican plot, in spite of the assertion
lican societies, met republican leaders confined to the dungeon, perhaps as a of Galois's brother Alfred that Evariste
(notably Fran�ois Vincent Raspail) and result of the accusation. was murdered. According to Dumas,
probably took part in the riots and dem­ In all this turmoil, however, the worst Galois's adversary was Pescheux d'Her­
onstrations that were racking Paris. He blow was the rejection of Galois's 1 83 1 binville, not a political enemy but an
joined the Artillery of the N ational paper. In the scathing preface to his ardent republican. Indeed, d'Herbinville
Guard, a branch of the militia made up memoirs, which he wrote while he was was one of the 1 9 officers of the Artillery
almost entirely of republicans. In De­ in prison, he stated: "I tell no one that I Guard whose acquittal was the occasion
cember his break with the Ecole Prepa­ owe anything of value in my work to his of Galois's defiant toast to the king.
ratoire became official. He wrote a let­ advice or encouragement. I do not say Moreover, when agents of the crown
ter that called the director of the school so because it would be a lie." were exposed during the revolution
a traitor for his actions during the J uly of 1 848, d'Herbinville was not among
revolution; not surprisingly, Galois he end of Galois's life has always them. A summary of an article recently
was expelled. T had a particular fascination for the­ sent to me by Taton indicates that the
The impression of Galois that one orists. Biographers have been unwilling duel was between friends and unfolded
gains from the events of this period is to accept at face value the implication of as a kind of Russian roulette in which
not that of a victim of circumstances, as his own words, namely that the duel was only one pistol was charged.
legend would have it. Instead he appears a result of a personal quarrel. Instead his Galois's mathematical writings the
to have been a hothead whose extreme biographers have looked for prostitutes, night before the duel were actually con­
actions consistently got him into trou­ agents provocateurs and political oppo­ fined to making editorial corrections on
ble. A letter by the mathematician So­ nents to account for his death. There is two manuscripts and to summarizing
phie Germain implies that Galois regu­ no evidence to support any of these con­ the contents of these and one other pa­
larly attended sessions of the Academy jectures. per in a long letter to Chevalier. The first
of Sciences and habitually insulted the In the middle of March, 1 83 2 , Galois paper was the one rejected by Poisson;
speakers. After his expulsion from the was transferred from Sainte- Pelagie to the second was a fragmentary version
Ecole Preparatoire he moved to his the nursing home Sieur Faultrier be­ of an article that had already been pub­
mother's house in Paris but proved to be cause of a cholera epidemic in Paris. lished in Ferussac's Bulletin. The third
so difficult to live with that she fled. There he apparently met the "infamous has not been found and its content is
The climactic event of the turbulent coquette." The involvement was brief, known only from the summary in the
spring of 1 83 1 took place on May 9 dur­ . but it is absurd to suggest that the girl letter; it apparently concerned integrals
ing a republican banquet celebrating the was a prostitute or a conspirator who of general algebraic functions.
acquittal of 1 9 artillery officers who had helped to arrange his assassination. The What of the famous words "I do not
been accused of plotting to overthrow epithet "infamous coquette" has been have the time" that Galois is supposed
the government. According to the mem­ associated with the words "quelque co­ to have written repeatedly in frustration
oirs of Alexandre Dumas (pere), Galois quette de bas etage" and so taken as con­ at being unable to complete his work?
stood to propose a toast. "To Louis Phi­ firming the prostitute story. According The phrase does appear, in the margin
lippe!" he said, raising a glass and a dag­ to the account by Raspail, however, the of the first memoir, but only once. Ap­
ger at the same time. For this provoca­ latter phrase was spoken by Galois a pended to it in parentheses is the com­
tive act he was arrested the next day and year before the duel; it may even have ment "Author's note."
held for more than a month in the pris­ been Raspail's own invention. More­
on at Sainte-Pelagie. over, on May 25, six days before his do not believe that the facts about the
At the ensuing trial Galois's defense death, Galois alludes to a broken love I life of Evariste Galois as I have pre­
claimed that the toast had been "To affair in a letter to his friend, Auguste sented them diminish his stature as a
Louis Philippe, if he betrays" but that Chevalier : "How can I console myself mathematician in the slightest degree.
"if he betrays" had been drowned out in when in one month I have exhausted the Many manuscript fragments indicate he
the uproar. Whether the j urors believed greatest source of happiness a man can carried on his mathematical investiga­
the defense or were moved by Galois's have, when I have exhausted it without tions not only while in prison but also up
youth (he was then 1 9) is not known, but happiness, without hope, when I am cer­ to the time of his death. That he could
they acquitted him in minutes. Never­ tain it is drained for life? " work productively through such turbu­
theless, on Bastille Day, J uly 1 4, 1 83 1 , Who was the woman? Two fragmen­ lent times is testimony to the extraordi­
less than a month after his acquittal, Ga­ tary letters were written to Galois in nary fertility of his imagination. Quite
lois was arrested again, this time for the weeks before the duel, suggesting a apart from the circumstances under
illegally wearing the uniform of the Ar­ personal quarrel in which Galois was which the work was done, there is no
tillery Guard. The guard had been dis­ more of a participant than he admitted. question that Galois developed one of
banded as a threat to the crown and Ga­ The first letter begins: "Please let us the most original ideas in the history
lois's gesture was therefore an act of de­ break up this affair. I do not have the of mathematics.
fiance. This time he spent eight months wit to follow a correspondence of this His reputation is not served, however,
in Sainte-Pelagie. nature but I will try to have enough to nor is the history of science, by a legend
The prison term was devastating : Ga­ converse with you as I did before any­ that insists a scientific genius must be
lois was alternately despondent and rag­ thing happened . . . . " The second letter above reproach in his personal life, or
ing. Raspail, who was serving a sentence is similar in tone, and both of them that any contemporary who does not
at the same time, later recalled that Ga­ bear the signature "Stephanie D." In the appreciate his genius is either a fool,
lois once, while drunk, had to be re­ Galois manuscripts Carlos Alberto In­ an assassin or a prostitute. The notion
strained from an attempt at suicide. Lat­ fantozzi of the University of the Repub­ that genius is not tolerated by medioc­
er, according to Raspail, Galois confid­ lic in Uruguay has managed to read a rity is too old a platitude to be adopted
ed a chilling vision of his demise : "I shall name that Galois had erased: Stephanie uncritically as accurate history. From
die in a duel for the sake of some worth­ Dumotel. Further detective work by In­ this point of view a genius would have
less girl [quelque coquette de bas etage]. fantozzi shows she was Stephanie-Feli­ to be recognized as such even when
Why? Because she will invite me to cie Poterin du Motel, the daughter of a standing at a banquet table with a dag­
avenge her honor, which another will resident physician at the Sieur Faultrier. ger in his hand.

149

© 1982 SCIENTIFIC AMERICAN, INC