Vous êtes sur la page 1sur 5

AbelRuni theorem

Not to be confused with Abels theorem. equation might be soluble, with a special formula for each
equation.[4] However, this is not so, but this impossibil-
ity lies outside the scope of the AbelRuni theorem and
In algebra, the AbelRuni theorem (also known as
Abels impossibility theorem) states that there is no is part of the Galois theory.
algebraic solutionthat is, solution in radicalsto the
general polynomial equations of degree ve or higher with
arbitrary coecients. The theorem is named after Paolo 2 Proof
Runi, who made an incomplete proof in 1799,[1] and
Niels Henrik Abel, who provided a proof in 1824.[2][3] The following proof is based on Galois theory and it
is valid for any eld of characteristic 0. Historically,
Runi[1] and Abels proofs precede Galois theory. For
1 Interpretation a modern presentation of Abels proof see the article of
Rosen[5] or the books of Tignol[6] or Pesic.[7]
The theorem does not assert that some higher-degree One of the fundamental theorems of Galois theory states
polynomial equations have no solution. In fact, the op- that a polynomial P(x) F[x] is solvable by radicals over
posite is true: every non-constant polynomial equation F if and only if its splitting eld K over F has a solvable
in one unknown, with real or complex coecients, has Galois group,[8] so the proof of the AbelRuni theorem
at least one complex number as a solution (and thus, by comes down to computing the Galois group of the general
polynomial division, as many complex roots as its degree, polynomial of the fth degree.
counting repeated roots); this is the fundamental theorem
Consider ve indeterminates y1 , y2 , y3 , y4 , and y5 , let E
of algebra. These solutions can be computed to any de-
= Q(y1 , y2 , y3 , y4 , y5 ), and let
sired degree of accuracy using numerical methods such as
the NewtonRaphson method or the Laguerre method,
and in this way they are no dierent from solutions to P(x) = (x y1 )(x y2 )(x y3 )(x y4 )(x y5 )
polynomial equations of the second, third, or fourth de- E[x].
grees. It also does not assert that no higher-degree poly-
nomial equations can be solved in radicals: the equation Expanding P(x) out yields the elementary symmetric
xn - 1 = 0 can be solved in radicals for every positive inte- functions of the yi:
ger n, for example. The theorem only shows that there is
no general solution in radicals that applies to all equations s1 = y1 + y2 + y3 + y4 + y5 ,
of a given degree greater than 4.
The solution of any second-degree polynomial equation s2 = y1 y2 + y1 y3 + y1 y4 + y1 y5 + y2 y3 + y2
can be expressed in terms of its coecients, using only y4 + y2 y5 + y3 y4 + y3 y5 + y4 y5 ,
addition, subtraction, multiplication, division, and square
roots, in the familiar quadratic formula: the roots of the s3 = y1 y2 y3 + y1 y2 y4 + y1 y2 y5 + y1 y3 y4
equation ax2 + bx + c = 0 (with a 0) are + y1 y3 y5 + y1 y4 y5 + y2 y3 y4 + y2 y3 y5 +
y2 y4 y5 + y3 y4 y5 ,
b (b2 4ac)/2a.
s4 = y1 y2 y3 y4 + y1 y2 y3 y5 + y1 y2 y4 y5 +
Analogous formulas for third-degree equations and
y1 y3 y4 y5 + y2 y3 y4 y5 ,
fourth-degree equations (using square roots and cube
roots) have been known since the 16th century. What
the AbelRuni theorem says is that there is no similar s5 = y1 y2 y3 y4 y5 .
formula for general equations of fth degree or higher.
In principle, it could be that the equations of the fth The coecient of xn in P(x) is thus (1)5 n s n. Let
degree could be split in several types and, for each one F = Q(s1 , s2 , s3 , s4 , s5 ) be the eld obtained by adjoin-
of these types, there could be some algebraic solution ing the symmetric functions to the rationals. Then P(x)
valid within that type. Or, as Ian Stewart wrote, for all F[x]. Because the yi's are indeterminates, every permu-
that Abels methods could prove, every particular quintic tation in the symmetric group on 5 letters S 5 induces

1
2 3 HISTORY

a distinct automorphism on E that leaves Q xed and The proof remains valid if, instead of working with
permutes the elements yi. Since an arbitrary rearrange- ve indeterminates, one works with ve concrete
ment of the roots of the product form still produces the algebraically independent complex numbers, because, by
same polynomial, e.g. the same argument, Gal(E/F) = S 5 .

(x y3 )(x y1 )(x y2 )(x y5 )(x y4 )


3 History
is the same polynomial as
Around 1770, Joseph Louis Lagrange began the ground-
work that unied the many dierent tricks that had been
(x y1 )(x y2 )(x y3 )(x y4 )(x y5 ), used up to that point to solve equations, relating them
to the theory of groups of permutations, in the form of
the automorphisms also leave F xed, so they are ele- Lagrange resolvents.[9] This innovative work by Lagrange
ments of the Galois group Gal(E/F). Therefore, we have was a precursor to Galois theory, and its failure to develop
shown that S 5 Gal(E/F); however there could possibly solutions for equations of fth and higher degrees hinted
be automorphisms there that are not in S 5 . But, since the that such solutions might be impossible, but it did not pro-
Galois group of the splitting eld of a quintic polynomial vide conclusive proof. The rst person who conjectured
has at most 5! elements, and since E is a splitting eld of that the problem of solving quintics by radicals might be
P(x), it follows that Gal(E/F) is isomorphic to S 5 . Gener- impossible to solve was Carl Friedrich Gauss, who wrote
alizing this argument shows that the Galois group of every in 1798 in section 359 of his book Disquisitiones Arith-
general polynomial of degree n is isomorphic to Sn. meticae (which would be published only in 1801) that
there is little doubt that this problem does not so much
The only composition series of S 5 is S 5 A5 {e} (where
defy modern methods of analysis as that it proposes the
A5 is the alternating group on ve letters, also known
impossible. The next year, in his thesis, he wrote Af-
as the icosahedral group). However, the quotient group
ter the labors of many geometers left little hope of ever
A5 /{e} (isomorphic to A5 itself) is not abelian, and so S 5
arriving at the resolution of the general equation alge-
is not solvable, so it must be that the general polynomial
braically, it appears more and more likely that this res-
of the fth degree has no solution in radicals. Since the
olution is impossible and contradictory. And he added
rst nontrivial normal subgroup of the symmetric group
Perhaps it will not be so dicult to prove, with all rigor,
on n letters is always the alternating group on n letters,
the impossibility for the fth degree. I shall set forth my
and since the alternating groups on n letters for n 5
investigations of this at greater length in another place.
are always simple and non-abelian, and hence not solv-
Actually, Gauss published nothing else on this subject.[1]
able, it also says that the general polynomials of all de-
grees higher than the fth also have no solution in radi- The theorem was rst nearly proved by Paolo Runi in
cals. Q.E.D. 1799.[10] He sent his proof to several mathematicians to
get it acknowledged, amongst them Lagrange (who did
The above construction of the Galois group for a fth de-
not reply) and Augustin-Louis Cauchy, who sent him a
gree polynomial only applies to the general polynomial;
letter saying: Your memoir on the general solution of
specic polynomials of the fth degree may have dier-
equations is a work which I have always believed should
ent Galois groups with quite dierent properties, e.g. x5
be kept in mind by mathematicians and which, in my
1 has a splitting eld generated by a primitive 5th root of
opinion, proves conclusively the algebraic unsolvability of
unity, and hence its Galois group is abelian and the equa-
general equations of higher than fourth degree.[11] How-
tion itself solvable by radicals; moreover, the argument
ever, in general, Runis proof was not considered con-
does not provide any rational-valued quintic that has S 5
vincing. Abel wrote: The rst and, if I am not mistaken,
or A5 as its Galois group. However, since the result is on
the only one who, before me, has sought to prove the im-
the general polynomial, it does say that a general quin-
possibility of the algebraic solution of general equations
tic formula for the roots of a quintic using only a nite
is the mathematician Runi. But his memoir is so com-
combination of the arithmetic operations and radicals in
plicated that it is very dicult to determine the validity
terms of the coecients is impossible.
of his argument. It seems to me that his argument is not
The proof is not valid if applied to polynomials whose completely satisfying.[11][12]
degree is less than 5. Indeed:
The proof also, as it was discovered later, was incomplete.
Runi assumed that all radicals that he was dealing with
the group A4 is not simple, because the subgroup {e, could be expressed from the roots of the polynomial us-
(12)(34), (13)(24), (14)(23)} is a normal subgroup; ing eld operations alone; in modern terms, he assumed
that the radicals belonged to the splitting eld of the poly-
the groups A2 and A3 are simple, but since they are nomial. To see why this is really an extra assumption,
abelian too (A2 is the trivial group and A3 is the consider, for instance, the polynomial P(x) = x3 15x
cyclic group of order 3), that is not a problem. 20. According to Cardanos formula, one of its roots (all
3

age of 18) submitted to the Paris Academy of Sciences


a memoir on his theory of solvability by radicals, which
was ultimately rejected in 1831 as being too sketchy and
for giving a condition in terms of the roots of the equa-
tion instead of its coecients. Galois was aware of the
contributions of Runi and Abel, since he wrote It is a
common truth, today, that the general equation of degree
greater than 4 cannot be solved by radicals this truth
has become common (by hearsay) despite the fact that ge-
ometers have ignored the proofs of Abel and Runi
[1]
Galois then died in 1832 and his paper Mmoire sur
les conditions de resolubilit des quations par radicaux[15]
remained unpublished until 1846, when it was published
by Joseph Liouville accompanied by some of his own
explanations.[14] Prior to this publication, Liouville an-
nounced Galois result to the Academy in a speech he
gave on 4 July 1843.[4] A simplication of Abels proof
was published by Pierre Wantzel in 1845.[16] When he
published it, he was already aware of the contributions
by Galois and he mentions that, whereas Abels proof is
valid only for general polynomials, Galois approach can
be used to provide a concrete polynomial of degree 5
whose roots cannot be expressed in radicals from its co-
ecients.
In 1963, Vladimir Arnold discovered a topological proof
Teoria generale delle equazioni, 1799 of the AbelRuni theorem,[17][18] which served as a
starting point for topological Galois theory.[19]

of them, actually) can be expressed as the sum of a cube


root of 10 + 5i with a cube root of 10 5i. On the other 4 Notes
hand, since P(3) < 0, P(2) > 0, P(1) < 0, and P(5) >
0, the roots r1 , r2 , and r3 of P(x) are all real and there- [1] Ayoub, Raymond G. (1980), Paolo Runis contri-
fore the eld Q(r1 , r2 , r3 ) is a subeld of R. But then the butions to the quintic, Archive for History of Exact
numbers 10 5i cannot belong to Q(r1 , r2 , r3 ). While Sciences, 22 (3): 253277, doi:10.1007/BF00357046,
Cauchy either did not notice Runis assumption or felt JSTOR 41133596, MR 606270, Zbl 0471.01008
that it was a minor one, most historians believe that the [2] Abel, Niels Henrik (1881) [1824], Mmoire sur les qua-
proof was not complete until Abel proved the theorem tions algbriques, ou l'on dmontre l'impossibilit de la
on natural irrationalities, which asserts that the assump- rsolution de l'quation gnrale du cinquime degr"
tion holds in the case of general polynomials.[6][13] The (PDF), in Sylow, Ludwig; Lie, Sophus, uvres Compltes
AbelRuni theorem is thus generally credited to Abel, de Niels Henrik Abel (in French), I (2nd ed.), Grndahl &
who published a proof in just six pages in 1824.[2] How- Sn, pp. 2833
ever, this short number of pages was obtained at the cost [3] Abel, Niels Henrik (1881) [1826], Dmonstration de
of writing in a very terse style. This was due to the fact l'impossibilit de la rsolution algbrique des quations
that he had the proof printed at his own expenses and he gnrales qui passent le quatrime degr" (PDF), in
needed to save paper and money.[7] A more elaborated Sylow, Ludwig; Lie, Sophus, uvres Compltes de Niels
version of the proof would be published in 1826.[3] Henrik Abel (in French), I (2nd ed.), Grndahl & Sn, pp.
6687
Proving that the general quintic (and higher) equations
were unsolvable by radicals did not completely settle the [4] Stewart, Ian (2015), Historical Introduction, Galois
matter, because the AbelRuni theorem does not pro- Theory (4th ed.), CRC Press, ISBN 978-1-4822-4582-0
vide necessary and sucient conditions for saying pre- [5] Rosen, Michael I. (1995), Niels Hendrik Abel and
cisely which quintic (and higher) equations are unsolvable Equations of the Fifth Degree, American Mathemati-
by radicals. Abel was working on a complete characteri- cal Monthly, 102 (6): 495505, doi:10.2307/2974763,
zation when he died in 1829.[14] JSTOR 2974763, MR 1336636, Zbl 0836.01015
According to Nathan Jacobson, The proofs of Runi [6] Tignol, Jean-Pierre (2016), Runi and Abel on general
and of Abel [] were soon superseded by the crowning equations, Galois Theory of algebraic equations (2nd
achievement of this line of research: Galois discover- ed.), World Scientic, ISBN 978-981-4704-69-4, Zbl
ies in the theory of equations.[8] In 1830, Galois (at the 06534509
4 4 NOTES

[7] Pesic, Peter (2004), Abels proof. An essay on the sources


and meaning of mathematical unsolvability, MIT Press,
ISBN 0-262-66182-9, Zbl 1166.01010

[8] Jacobson, Nathan (2009), Galois Theory of equations,


Basic algebra, 1 (2nd ed.), Dover, ISBN 978-0-486-
47189-1

[9] Lagrange, Joseph-Louis (1869) [1771], Rexions sur la


rsolution algbrique des quations, in Serret, Joseph-
Alfred, uvres de Lagrange, III, Gauthier-Villars, pp.
205421

[10] Runi, Paolo (1799), Teoria generale delle equazioni, in


cui si dimostra impossibile la soluzione algebraica delle
equazioni generali di grado superiore al quarto (in Italian),
Stamperia di S. Tommaso d'Aquino

[11] Kiernan, B. Melvin (1971), The Development of


Galois Theory from Lagrange to Artin, Archive
for History of Exact Sciences, 8 (1/2): 40154,
doi:10.1007/BF00327219, JSTOR 41133337

[12] Abel, Niels Henrik (1881) [1828], Sur la resolution


algbriques des quations (PDF), in Sylow, Ludwig;
Lie, Sophus, uvres Compltes de Niels Henrik Abel (in
French), II (2nd ed.), Grndahl & Sn, pp. 217243

[13] Stewart, Ian (2015), The idea behind Galois theory,


Galois Theory (4th ed.), CRC Press, ISBN 978-1-4822-
4582-0

[14] Tignol, Jean-Pierre (2016), Galois, Galois Theory of al-


gebraic equations (2nd ed.), World Scientic, ISBN 978-
981-4704-69-4, Zbl 06534509

[15] Galois, variste (1846), Mmoire sur les conditions de


resolubilit des quations par radicaux (PDF), Journal de
Mathmatiques Pures et Appliques (in French), XI: 417
433

[16] Wantzel, Pierre (1845), Dmonstration de limpossibilit


de rsoudre toutes les quations algbriques avec des rad-
icaux, Nouvelles Annales de Mathmatiques (in French),
4: 5765

[17] Alekseev, V. B. (2004), Abels theorem in problems and


solutions. Based on the lectures of Professor V. I. Arnold,
Kluwer Academic Publishers, ISBN 1-4020-2186-0, Zbl
02146318

[18] Short proof of Abels theorem that 5th degree polynomial


equations cannot be solved on YouTube

[19] Khovanskii, Askold (2014), Topological Galois The-


ory: Solvability and Unsolvability of Equations in Finite
Terms, Springer Monographs in Mathematics, Springer-
Verlag, doi:10.1007/978-3-642-38871-2, ISBN 978-3-
642-38870-5
5

5 Text and image sources, contributors, and licenses


5.1 Text
AbelRuni theorem Source: https://en.wikipedia.org/wiki/Abel%E2%80%93Ruffini_theorem?oldid=780952093 Contributors: Axel-
Boldt, XJaM, Nealmcb, Michael Hardy, Dominus, Schneelocke, RodC, Charles Matthews, Dcoetzee, Fibonacci, Psychonaut, Marc Venot,
Giftlite, Nsh, Tbjablin, Anythingyouwant, Icairns, Mh, Edudobay, Ben Standeven, Geschichte, Andrewpmk, Gene Nygaard, Oleg Alexan-
drov, Simetrical, Rjwilmsi, Salix alba, FlaBot, VKokielov, Roboto de Ajvol, YurikBot, RobotE, Dmharvey, Crasshopper, LamilLerran,
RDBury, BeteNoir, Eskimbot, JCSantos, Nbarth, Colonies Chris, Javalenok, Zchenyu, Lambiam, BigDom, Mets501, Rlinnity, Saviour-
machine, CRGreathouse, Myasuda, Stormwyrm, Doctormatt, Paddles, The Transhumanist, JamesBWatson, Albmont, Kutu su~enwiki,
Sue Gardner, Fruits Monster, Inquam, VolkovBot, Mike4ty4, Quantpole, EverGreg, Wing gundam, JackSchmidt, Cheesefondue, Jdgilbey,
Thegeneralguy, He7d3r, Sandrobt, Marc van Leeuwen, MystBot, Addbot, Tide rolls, Arbitrarily0, Ettrig, Legobot, Luckas-bot, Yobot,
TaBOT-zerem, AnomieBOT, VanishedUser sdu9aya9fasdsopa, IRP, Xqbot, Raulshc, Lunae, Swordsmankirby, Trappist the monk, Duo-
duoduo, Skysmurf, WikitanvirBot, GoingBatty, Slawekb, ZroBot, Cf. Hay, Anita5192, Helpful Pixie Bot, BG19bot, Brad7777, Comfr,
Nigellwh, Stamptrader, Boazka, Denziloe, Loraof, Boazkat, WolfWalt, The4seasons, Svyatopolk and Anonymous: 57

5.2 Images
File:Ruffini_-_Teoria_generale_delle_equazioni,_1799_-_1366896.jpg Source: https://upload.wikimedia.org/wikipedia/commons/
c/ce/Ruffini_-_Teoria_generale_delle_equazioni%2C_1799_-_1366896.jpg License: Public domain Contributors: Available in the BEIC
digital library and uploaded in partnership with BEIC Foundation. Original artist: Runi, Paolo

5.3 Content license


Creative Commons Attribution-Share Alike 3.0