0% found this document useful (0 votes)
158 views12 pages

BAW Sympoly

Uploaded by

saidiaymane411
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
158 views12 pages

BAW Sympoly

Uploaded by

saidiaymane411
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

. . . . . . . . . . . . . . . . . . . . . . . . . .

Olympiad Training for Individual Study

Symmetric Polynomials
ft. Vieta’s polynomials

Evan Chen《陳誼廷》
3 December 2024


BAW-SYMPOLY


誼 se
陳 U
n《 al
h e rn
C nt e
n
a , I
Ev I S
y
B O T

OTIS, © Evan Chen, internal use only. Artwork contributed by Rohan Dhillon.

1
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§1 Lecture notes
A lot of the intuition you’ll get from this lecture is hard to state formally, but there are a
few results you should be aware of off the bat.

• Vieta: the coefficients of a polynomial are symmetric sums.

• Factoring theorem: again two things here.


(a) If P (x) is a polynomial with root r, then x − r divides P . Easy.
(b) If P (x, y, . . . ) is a polynomial vanishing whenever x = y (meaning P is
identically zero whenever x = y), then x − y divides P .

§1.1 Fundamental theorem of elementary symmetric polynomials


More involved is the fundamental theorem of symmetric polynomials: Let


P (x1 , . . . , xn ) be a polynomial which is symmetric. Consider the so-called Vieta ba-
sis:

誼 se
X
σ1 = xi

陳 U
i
X

《 al
σ2 = xi xj
i<j

n
e rn
X
σ3 = xi xj xk

h
i<j<k
..

C e
.

n nt
σn = x1 . . . xn

a , I
(These are just the formulas you get from Vieta formulas, without the pesky signs.)

Ev I S
Theorem 1.1 (Fundamental theorem of elementary symmetric polynomials)

y T
Any symmetric polynomial P (x1 , . . . , xn ) can be expressed uniquely as a polynomial

B O
in σ1 , . . . , σn . That is,

P (x1 , . . . , xn ) = Q(σ1 , . . . , σn )

for some unique polynomial Q.


Moreover, if P has integer/rational/real coefficients then so does Q.

As an example of the previous theorem, this means that it should possible to express
a3 + b3 + c3 + d3 in terms of the symmetric sums σ1 = a + b + c + d, σ2 = ab + bc + cd +
da + ac + bd, σ3 = abc + bcd + cda + dab, σ4 = abcd. And indeed it is:

a3 + b3 + c3 + d3 = (a + b + c + d)3 − 3(a + b + c + d)(ab + bc + cd + da + ac + bd)


+ 3(abc + bcd + cda + dab)
= σ13 − 3σ1 σ2 + 3σ3 .

This means that the Vieta basis is sort of the natural one in which one can “force” a
canonical representation of a symmetric polynomial.

2
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§1.2 Factoring multivariate polynomials


When you have a single variable polynomial P (t) with a root at t = r, then you know
that t − r is a factor of P , and depending on context you can sometimes use this to
extract factorizations of P .
A similar statement is true for multivariate polynomials given a linear relation like
P = 0 when a = b or P = 0 when a + b + c = 0. (Although we won’t use it, a slightly
more technical result is true for higher-degree relations; https://math.stackexchange.
com/a/463935/229197 if interested.) But when symmetric, you enjoy a lot of additional
symmetry which you can often exploit.
For example, consider the polynomial

P (a, b, c) = a3 + b3 + c3 − 3abc.


Notice that P has the property that it vanishes (i.e. equals zero) whenever a + b + c = 0.
This means that P should factor as the product of a + b + c and a degree-2 symmetric


polynomial. In other words, we can write

誼 se
P (a, b, c) = a3 + b3 + c3 − 3abc = (a + b + c) λ1 (a2 + b2 + c2 ) + λ2 (ab + bc + ca) .


陳 U
Now you can deduce the constants λ1 and λ2 by examining just a few particular coefficients.
For example, equating the a3 coefficient terms on each side gives

n《 al a3 = a · (λ1 a2 ) =⇒ λ1 = 1

h e rn
while equating the abc term gives

C t e
−3abc = λ2 (a · bc + b · ca + c · ab) =⇒ λ2 = −1.

n
a , I n
§1.3 Newton sums, classical form

v S
In the case where we want to express r1d + · · · + rnd in terms of the Vieta coefficients, the

E I
Newton Sums lets us do this explicitly (or at least recursively). This doesn’t come up

y T
too often but is nice to know.

B O
Theorem 1.2 (Newton sums)
Consider the polynomial

P (x) = an xn + an−1 xn−1 + · · · + a0

with complex roots r1 , r2 , . . . , rn .


For each d ≥ 0, define pd = r1d + r2d + · · · + rnd . Then

0 = an p1 + an−1
= an p2 + an−1 p1 + an−2 · 2
= an p3 + an−1 p2 + an−2 p1 + an−3 · 3
= an p4 + an−1 p3 + an−2 p2 + an−3 p1 + an−4 · 4
..
.

where we by convention let ai = 0 if i < 0.

3
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

To be concrete, consider the polynomial x3 + 7x2 + 14x + 8, say with roots u, v, w.


If we want to compute u3 + v 3 + w3 , then the Newton sums identities let us do this
recursively by
0 = 1 · p1 + 7 · 1 =⇒ p1 = −7
0 = 1 · p2 + 7 · p1 + 14 · 2 =⇒ p2 = 21
0 = 1 · p3 + 7 · p2 + 14 · p1 + 8 · 3 =⇒ p3 = −73.

§1.4 Appendix: Obscure forms of Newton sums (just for curiosity, not
actually useful AFAIK)
If you like determinants, you can use the following formula instead:


Theorem 1.3 (Newton’s identities in Cauchy form)


Let σi be as defined last section. For each d ≥ 0, we have

誼 se
 
σ1 1 0 0 0 ... 0 0
 2σ 2 σ1 1 0 0 ... 0 0

陳 U
 

 3σ 3 σ2 σ1 1 0 ... 0 0 
pd = det  4σ 4 σ3 σ2 σ1 1 ... 0 0

《 al
.
.. .. .. .. .. .. .. .. 

. . . . . . .

n

 
(d − 1)σd−1 σd−2 σd−3 σd−4 σd−5 . . . σ1 1

e rn
dσd σd−1 σd−2 σd−3 σd−4 . . . σ2 σ1

C h e
Note that this formula uses σi (symmetric sums) whereas the last one uses ai (coefficients);

t
for monic polynomials, these are related by σi = (−1)i an−i . So to redo the previous

n I n
example we would write

va ,
 
−7 1 0

E S
p3 = det  2 · 14 −7 1  = −73.

I
3 · (−8) 14 −7

y T
And if you like sums, or you just want the explicit polynomial:

B OTheorem 1.4 (Newton’s identities as an explicit sum)


Let σi be as defined last section. For each d > 0, we have

(−1)s
 
X s
pd = (−1)d · d · σ r1 σ r2 . . . σdrd
s r1 , . . . , rd 1 2
r1 +2r2 +···+drd =d
ri ≥0

where s = r1 + r2 + · · · + rd for brevity.

If you don’t like multinomial coefficients, you can also write out the factorials:
X (−1)s (s − 1)! r1 r2
pd = (−1)d · d · σ σ . . . σdrd .
r1 !r2 ! . . . rd ! 1 2
r1 +2r2 +···+drd =d
ri ≥0

You do not have to know either of these obscure forms as they are not used often.
(Heck, I didn’t even know either existed until after I graduated college.) They’re included
here as a reference, not as mandatory reading.

4
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§1.5 Examples
Factoring theorem example:

Example 1.5
Show that a4 + b4 ≥ a3 b + ab3 for a, b > 0.

Z68B4C1C Walkthrough. For the purposes of this example, assume you don’t know AM-GM or
Muirhead, etc. The goal is to show how to solve the problem with “bare hands”.
We will prove a4 + b4 − a3 b − ab3 ≥ 0.
(a) Noting that equality holds when a = b, what factor must divide the left-hand side?
(b) Imagine fixing b, and treating the left-hand side as a polynomial P (a). It has a


root at a = b. If you also know that P ≥ 0 everywhere, what kind of root must
that root be?


(c) Use this to factor the left-hand side completely. (There should be three factors.)

誼 se
The condition a, b > 0 isn’t actually used here but makes things simpler to think
about.

陳 U
Example 1.6 (AIME 2010)

n《 al
Compute the maximum possible value of a3 + b3 + c3 over all real numbers (a, b, c)

e rn
satisfying

h
a3 = abc + 2

C t e
b3 = abc + 6

n I n
c3 = abc + 20.

va ,
Walkthrough.

S
10AIME9

E I
(a) Express a3 + b3 + c3 in terms of abc. Thus it suffices to compute abc.

y T
(b) Find a way to get a quadratic equation in abc.

B O
(c) Solve for abc and use it to get the final answer.

Example 1.7 (Evan Chen, Fall 2015)


Let a, b, c be the distinct roots of the polynomial

P (x) = x3 − 10x2 + x − 2015.

The cubic polynomial Q(x) is monic and has distinct roots bc − a2 , ca − b2 , ab − c2 .


What is the sum of the coefficients of Q?

15OMOF12 Walkthrough. This can be done with brute force, by actually finding Q, but there is a
trick to it.
(a) Show the answer is given by Q(1) = (1 − bc + a2 )(1 − ca + b2 )(1 − ab + c2 ).
(b) Show that ab + bc + ca = 1.
(c) Prove that Q(1) = 2015000.

5
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

Example 1.8 (USAMO 1975/3)


If P (x) denotes a polynomial of degree n such that P (k) = k
k+1 for k = 0, 1, 2, . . . , n,
determine P (n + 1).

75AMO3 Walkthrough. The main idea is to define Q(x) = (x + 1)P (x) − x.


(a) Compute deg Q (in terms of n).

(b) Determine the roots of Q.

(c) Use (a) and (b) to establish the factorization, of Q up to a constant factor.

(d) Show that Q(−1) = 1 and use this to conclude that the leading coefficient of Q is
equal to


(−1)n+1
c= .


(n + 1)!

(e) Compute Q(n + 1).

誼 se
(f) Prove that P (n + 1) = n+1+(−1)n+1
.

陳 U
n+2

《 al
Example 1.9 (HMMT 2023 T2)

n
Prove there don’t exist pairwise distinct complex numbers a, b, c, and d such that

h e rn a3 − bcd = b3 − cda = c3 − dab = d3 − abc.

C t e
23HMMTT2 Walkthrough. There is a brute-force approach along the lines of taking

n
a , I n
a3 − b3 = bcd − cda

v S
and factoring out the common a − b.

y E I
(a) Use this idea to show that a2 + b2 = c2 + d2 .

T
B O
(b) Similarly, show that a2 + c2 = b2 + d2 , and a2 + d2 = b2 + c2 .
However, I think the following Vieta-based approach is more conceptually nice, since it
does not require any factoring and obviously generalizes to more variables.
(c) Let’s assume abcd 6= 0. By scaling, show that we may in fact assume abcd = 1.

(d) Conclude that we may define the number k by


1 1
k := a3 − = · · · = d3 − .
a d

(e) Let
P (X) = (X − a)(X − b)(X − c)(X − d).
Find the coefficients of P in terms of k.

(f) Derive a contradiction by noticing P (0) = −1.

(g) Weed out the edge case abcd = 0 we didn’t address earlier. This gives a complete
solution to the problem.

6
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§2 Practice Problems
Instructions: Solve [40♣]. If you have time, solve [52♣]. Problems with red weights are mandatory.

The Law speaks: you are cast out. You are un-dwarf. I AM
A WITNESS!

Angarthing in The Hammer of Thursagan,


from The Battle for Wesnoth

08AIME7
[2♣] Problem 1 (Added by Eric Wang). Let r, s, and t be the three roots of the
equation
8x3 + 1001x + 2008 = 0.


Compute (r + s)3 + (s + t)3 + (t + r)3 .
73AMO4


[2♣] Problem 2 (USAMO 1973/4). Determine all triples (x, y, z) of complex numbers
satisfying

誼 se
x + y + z = 3,

陳 U
x + y 2 + z 2 = 3,
2

《 al
x3 + y 3 + z 3 = 3.

n
96CAN1
[2♣] Problem 3 (Canada 1996, added by Haozhe Yang). If α, β, and γ are the roots of

e rn
x3 − x − 1 = 0, compute 1+α
1−α + 1−β + 1−γ .
1+β 1+γ

h
16HMNTGUTS27

C e
[2♣] Problem 4 (HMMT November 2016 Guts, added by Rohan Bodke). Let r1 , r2 , r3 ,

t
r4 be the complex roots of the polynomial x4 − 4x3 + 8x2 − 7x + 3. Calculate

n
a , I n
r12 r22 r32 r42

v
+ + + .
r22 + r32 + r42 r12 + r32 + r42 r12 + r22 + r42 r12 + r22 + r32

E I S
NIMO85
[2♣] Problem 5 (NIMO #8). Let x, y, z be complex numbers satisfying

y T
x2 + 5y = 10x

B O
y 2 + 5z = 10y
z 2 + 5x = 10z.

Find the sum of all possible values of z.


84AMO1
[2♣] Problem 6 (USAMO 1984/1). The product of two of the four roots of the quartic
equation x4 − 18x3 + kx2 + 200x − 1984 = 0 is −32. Determine k.
JASONMAO
[3♣] Required Problem 7. The cubic x3 − 7x2 + 3x + 2 has irrational roots r >
s > t. There exists a unique set of rational numbers A, B, and C, such that the cubic
x3 + Ax2 + Bx + C has r + s as a root. What is A + B + C?
20AIMEII11
[3♣] Problem 8 (AIME II 2020, added by Benjamin Wang-Tie). Let P (x) = x2 − 3x − 7,
and let Q(x) and R(x) be two quadratic polynomials also with the coefficient of x2 equal
to 1. David computes each of the three sums P + Q, P + R, and Q + R and is surprised
to find that each pair of these sums has a common root, and these three common roots
are distinct. If Q(0) = 2, compute R(0).

7
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

19AIME10
[2♣] Problem 9 (AIME I 2019, added by Joshua Im). For distinct complex numbers z1 ,
z2 , . . ., z673 , the polynomial

(x − z1 )3 (x − z2 )3 · · · (x − z673 )3

can be expressed as x2019 + 20x2018 + 19x2017 + g(x), where g(x) is a polynomial with
complex coefficients and with degree at most 2016. Compute
X
z j zk .
1≤j<k≤673
16AUT6
[3♣] Problem 10 (Austria 2016/6, added by Abdullahil Kafi). Let a, b, c be three
integers for which the sum


ab ac bc
+ +
c b a


is an integer. Prove that each of the three numbers

誼 se
ab ac bc
, ,
c b a

陳 U
is an integer.

《 al
ZEAC3666
[2♣] Problem 11. Factor the polynomial

n
e rn
a(b − c)3 + b(c − a)3 + c(a − b)3 .

h
H1883820

C e
[3♣] Problem 12. Let a, b, c be real numbers. Prove that

n nt
a3 + b3 + c3 = (a + b + c)3 if and only if a5 + b5 + c5 = (a + b + c)5 .

a , I
BMCQ54

v
[2♣] Problem 13. Let a, b, c, d be nonzero complex numbers such that

E I S
1 1 1 1
a+b+c+d=0 and + + + = 0.

y T
a b c d

B O
Prove that two of the four numbers have sum zero.
03AIMEII9
[3♣] Problem 14 (AIME II 2003, added by Lincoln Liu). Consider the polynomials
P (x) = x6 − x5 − x3 − x2 − x and Q(x) = x4 − x3 − x2 − 1. Given that z1 , z2 , z3 , and z4
are the roots of Q(x) = 0, find P (z1 ) + P (z2 ) + P (z3 ) + P (z4 ).
18CMIMCA9
[9♣] Problem 15 (CMIMC 2018 A9). Given the polynomial identity
2018
X
(x2 − 3x + 1)1009 = ak xk
k=0

calculate the remainder when a20 + a21 + · · · + a22018 is divided by 2017.


11SMTA7
[3♣] Required Problem 16 (Stanford Math Tournament 2011). Let P (x) be a polyno-
mial of degree 2011 such that P (1) = 0, P (2) = 1, P (4) = 2, . . . , and P (22011 ) = 2011.
Find the coefficient of x1 in P .
NOTVIETA
[5♣] Problem 17. Let a, b, c be integers with c 6= 0. Suppose the cubic polynomial

x3 + ax2 + bx + c has roots r ≤ s ≤ t. Show that rs + st + rt can be written as u ± v for
some rational numbers u and v.

8
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

H3233958
[2♣] Problem 18 (Mock ARML 2022, added by Shaheem Samsudeen). Let a, b, c
complex numbers with ab + bc + ca = 61 such that
1 1 1
+ + =5
b+c c+a a+b
a b c
+ + = 32.
b+c c+a a+b
Find the value of abc.
13SMTA9 √ √ √
[3♣]√Required
√ Problem
√ √19 (Stanford
√ √ Math Tournament 2013). Let a = − 3+ 5+ 7,
b = 3 − 5 + 7, c = 3 + 5 − 7. Compute


a4 b4 c4
+ + .
(a − b)(a − c) (b − c)(b − a) (c − a)(c − b)


Z5394300
[5♣] Problem 20 (Ritwin Narra). Fix an integer n 6= 1. Prove that if real numbers a,

誼 se
b, c, d satisfy
a + b + c + d = an + bn + cn + dn = 0

陳 U
then two of a, b, c, d sum to 0.

《 al
12BLACKMOP
[5♣] Required Problem 21 (Black MOP 2012). Let ABC be a triangle and let hA ,

n
hB , hC be the lengths of the altitudes from A, B, and C. Let a = BC, b = CA, c = AB.

e rn
Suppose that

C h e
p p p p p p
a + hB + b + hC + c + hA = a + hC + b + hA + c + hB .

n nt
Prove that triangle ABC is isosceles.

I
a ,
H2954369

v
[5♣] Problem 22 (Added by Jason Lee). Consider all complex numbers k for which

S
there exist complex numbers a, b, c, d and e satisfying

y E T I
a b
+
b c
=1

B O
b c
+ =2
c d
c d
+ =3
d e
d e
+ =4
e a
e a
+ = k.
a b
Find the sum of all possible values of k 4 .

Remark. For the previous problem by Jason Lee, avoid using a calculator — you can solve
it with rather little arithmetic if you set it up correctly.

ZC01F573
[5♣] Problem 23 (Prove the Newton sums in the reading). Suppose the complex-
coefficient polynomial

f (x) = an xn + an−1 xn−1 + · · · + a0

9
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

has complex roots z1 , z2 , . . . , zn . For each d ≥ 0, define pd = z1d + z2d + · · · + znd . Prove
that for every integer k ≥ 1 we have the identity

an pk + an−1 pk−1 + · · · + an−(k−1) p1 + k · an−k = 0.

where we by convention let ai = 0 if i < 0.


85LL19
[9♣] Required Problem 24 (Longlist 1985/19). Solve over R the system of simultaneous
equations
√ 1
x− − 2w + 3z = 1,
y
1
x+ − 4w2 − 9z 2 = 3,
y2


√ 1
x x − 3 − 8w3 + 27z 3 = −5,


y
1
x2 + 4 − 16w4 − 81z 4 = 15.

誼 se
y
Z40B5559
[9♣] Problem 25 (Z is a ring). Suppose that α and β are complex numbers and monic

陳 U
polynomials P, Q ∈ Z[x] satisfy P (α) = Q(β) = 0.

《 al
(a) Show that there is monic polynomial R ∈ Z[x] such that R(α + β) = 0.

n
e rn
(b) Show that there is a monic polynomial S ∈ Z[x] such that S(αβ) = 0.

h
20HMMTA9
[9♣] Problem 26 (HMMT 2020, added by Guanjie Lu). Let P (x) = x2020 + x + 2. Let

C t e
Q(x) be the monic polynomial of degree 2 whose roots are the pairwise products of
2020


n n
the roots of P (x). Let α satisfy P (α) = 4. Compute the sum of all possible values of

a , I
Q(α2 )2 .

v
[1♣] Mini Survey. Fill out feedback on the OTIS-WEB portal when submitting this

E I S
problem set. Any thoughts on problems (e.g. especially nice, instructive, easy, etc.) or
overall comments on the unit are welcome.

y T
In addition, if you have any suggestions for problems to add, or want to write hints for

B O
one problem you really liked, please do so in the ARCH system!

The maximum number of [♣] for this unit is [103♣], including the mini-survey.

10
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§3 Solutions to the walkthroughs


§3.1 Solution 1.5
Note that
LHS − RHS = (a − b)2 (a2 + ab + b2 ) ≥ 0.

§3.2 Solution 1.6, AIME 2010


Let m := abc. Multiply all three equations to get

m3 = (m + 2)(m + 6)(m + 20) = m3 + 28m2 + 92m + 240.

Hence


0 = 28m2 + 92 + 240 = 4(7m + 15)(m + 4).
So we either have m = −4 or m = 7 .Meanwhile, summing the given equations gives


−15

誼 se
 
151
a3 + b3 + c3 = 3m + 28 ∈ 16, .
7

陳 U
So in fact a3 + b3 + c3 can only take one of two possible values!
We claim that 1517 is indeed achievable, which will give the final answer. Indeed, one

《 al
can take

n
 
−1 3

e rn
5
(a, b, c) = √ ,√ ,√ .

h
3
7 37 37

e
So is the final answer.

C
151

t
7

n n
§3.3 Solution 1.7, Evan Chen, Fall 2015

va , I
The desired polynomial Q can be written as

E I S
Q(x) = (x − bc + a2 )(x − ca + b2 )(x − ab + c2 ).

y T
Hence, it suffices to compute Q(1).

B O
On the other hand, Vieta formula tells us that ab + bc + ca = 1. Hence

Q(1) = (1 − bc + a2 )(1 − ca + b2 )(1 − ab + c2 )


= (ab + ac + a2 )(bc + ba + b2 )(ca + cb + c2 )
= abc(a + b + c)3 = 2015 · 103 = 2015000.

§3.4 Solution 1.8, USAMO 1975/3


Consider the polynomial Q(x) = (x + 1)P (x) − x. It has degree n + 1 and has zeros at
x = 0, . . . , n. Hence we need to have

Q(x) = (x + 1)P (x) − x = cx(x − 1)(x − 2) . . . (x − n)

for some constant c. Plugging in x = −1 gives 1 = c · (−1)n+1 · (n + 1)!. Thus


Q(n + 1) = (−1)n+1 . Conclude

(n + 1) + Q(n + 1) n + 1 + (−1)n+1
P (n + 1) = = .
n+2 n+2

11
Evan Chen《陳誼廷》 (OTIS, updated 2024-12-03) Symmetric Polynomials

§3.5 Solution 1.9, HMMT 2023 T2


We proceed by contradiction, assuming the existence of pairwise distinct complex numbers
as above.

¶ Vieta-based approach. First, suppose one of the numbers, say d, is zero, and the
others are nonzero. Then we get a3 = b3 = c3 = −abc. By scaling, WLOG assume that
abc = −1. Then a, b, c are supposed to be distinct cube roots of unity, in which case
abc = 1, contradiction.
Otherwise, by scaling assume that abcd = −1. Then define
1 1 1 1
λ := a3 + = b3 + = c3 + = d3 +
a b c d


Then a, b, c, d are the complex roots of the polynomial


X 4 − λX + 1 = 0.

誼 se
But then abcd = 1 by Vieta, contradiction.

陳 U
¶ Factoring-based approach. Notice that we have

《 al
a3 − b3 = bcd − cda =⇒ a2 + ab + b2 = −cd.

n
e rn
Thus a2 + b2 = −(ab + cd) = c2 + d2 . In the same way we derive

h e
a2 + c2 = b2 + d2 , a2 + d2 = b2 + c2 .

C nt
This is enough to imply a2 = b2 = c2 = d2 , which is impossible.

n
va , I
E I S
y
B O T

12

You might also like