Problems from the Book Problem 19.

9
Let n N. Let w1 , w2 , ..., wn be n reals. Prove the inequality
!2
n X
n
n
X
X
ijwi wj
wi .

i
+
j

1
i=1 j=1
i=1
Solution by Darij Grinberg
Notations.

j
For any matrix A, we denote by A
the entry in the j-th column and the
i
i-th row of A. [This is usually denoted by Aij or by Ai,j .]


Let k be a field. Let u N and v N, and let ai,j be an element of k for every
(i, j) {1, 2, ..., u} {1, 2, ..., v} . Then, we denote by (ai,j )1jv
1iu the u v matrix
 
j
A which satisfies A
= ai,j for every (i, j) {1, 2, ..., u} {1, 2, ..., v} .
i
Let
 n N. Let t1, t2 , ..., tn be n objects. Let m {1, 2, ..., n}. Then, we let
t1 , t2 , ..., tc
m , ..., tn denote the (n 1)-tuple (t1 , t2 , ..., tm2 , tm1 , tm+1 , tm+2 , ..., tn )

ti , if i < m;
(that is, the (n 1)-tuple (s1 , s2 , ..., sn1 ) defined by si =
for
ti+1 , if i m
all i {1, 2, ..., n 1}).
Let R be a commutative ring with unity. Let a1 , a2 , ..., am be m elements of R.
Then, we define an element k (a1 , a2 , ..., am ) of R by
X
Y
k (a1 , a2 , ..., am ) =
a (i) .
S{1,2,...,m}; iS
|S|=k

This element k (a1 , a2 , ..., am ) is simply the k-th elementary symmetric polynomial evaluated at a1 , a2 , ..., am .
The Viete theorem states that
Y

(x a` ) =

`{1,2,...,m}

m
X

(1)k k (a1 , a2 , ..., am ) xmk

k=0

for every x R. If we choose some i {1, 2, ..., m} and apply this equality to
the m 1 elements a1 , a2 , ..., abi , ..., am in lieu of the m elements a1 , a2 , ..., am ,
then we obtain
Y
`{1,2,...,m}\{i}

(x a` ) =

m1
X

(1)k k (a1 , a2 , ..., abi , ..., am ) xm1k .

k=0

(1)

Theorem 1 (Sylvester). Let n N, and let A Rnn be a symmetric

n n matrix. Then, the matrix A is positive definite
if and only if every
  1jm !
j
m {1, 2, ..., n} satisfies det
A
> 0.
i
1im
For a proof of Theorem 1, see any book on symmetric or Hermitian matrices.
Theorem 2 (Cauchy determinant). Let k be a field. Let m N. Let
a1 , a2 , ..., am be m elements of k. Let b1 , b2 , ..., bm be m elements of k.
Assume that aj 6= bi for every (i, j) {1, 2, ..., m}2 . Then,
Q
((ai aj ) (bj bi ))

1jm ! (i,j){1,2,...,m}2 ;
1
i>j
Q
=
det
.
aj bi 1im
(aj bi )
(i,j){1,2,...,m}2

In the following, I attempt to give the most conceptual proof of Theorem 2. First
we recall a known fact we are not going to prove:
Theorem 3 (Vandermonde determinant). Let S be a commutative
ring with unity. Let m N. Let a1 , a2 , ..., am be m elements of S. Then,

Y

1jm 
det aj1
=
(ai aj ) .
i
1im
(i,j){1,2,...,m}2 ;
i>j

Besides, a trivial fact:

Lemma 4. Let S be a commutative ring with unity. Let a S. In the
ring S [X] (the polynomial ring over S in one indeterminate X), the element
X a is not a zero divisor.
And a consequence of this fact:
Lemma 5. Let R be a commutative ring with unity. Let m N. In the
ring R [X1 , X2 , ..., Xm ] (the polynomial
ring over R in m indeterminates X1 ,
Q
X2 , ..., Xm ), the element
(Xi Xj ) is not a zero divisor.
(i,j){1,2,...,m}2 ;
i>j

Proof of Lemma 5. We will first show that:

For any (i, j) {1, 2, ..., m}2 satisfying i > j, the element Xi Xj
of the ring R [X1 , X2 , ..., Xm ] is not a zero divisor.
(2)
h
i
ci , ..., Xm denote the sub-R-algebra of R [X1 , X2 , ..., Xm ]
Proof of (2). Let R X1 , X2 , ..., X
generated by the elements X1 , X2 , ..., Xi2 , Xi1 , Xi+1 , Xi+2
m eleh , ..., Xm (that is, the i
ci , ..., Xm [X]
ments X1 , X2 , ..., Xm except of Xi ). Consider the ring R X1 , X2 , ..., X
2

i
c
(this is the polynomial ring over the ring R X1 , X2 , ..., Xi , ..., Xm in one indeterminate
 h
i
ci , ..., Xm [X]
X). It is known that there exists an R-algebra isomorphism : R X1 , X2 , ..., X
R [X1 , X2 , ..., Xm ] such that (X) = Xi and (Xk ) = Xk for every k {1, 2, ..., m} \
= Xi Xj . Since X Xj is not a zero
{i}. Hence, (X Xj ) = (X) (Xj )
| {z }
| {z }
=Xi

=Xj , as

j{1,2,...,m}\{i}
 h
i
h
i
c
c
divisor in R X1 , X2 , ..., Xi , ..., Xm [X] (by Lemma 4, applied to S = R X1 , X2 , ..., Xi , ..., Xm
and a = Xj ), it thus follows that (X Xj ) = Xi Xj is not a zero divisor in
R [X1 , X2 , ..., Xm ] (since is an R-algebra isomorphism). This proves (2).
It is known that if we choose some elements of a ring such that each of these elements
is not a zero divisor, then the product
of these elements is not a zero divisor. Hence,
Q
(2) yields that the element
(Xi Xj ) of the ring R [X1 , X2 , ..., Xm ] is not
(i,j){1,2,...,m}2 ;
i>j

a zero divisor. This proves Lemma 5.

Now comes a rather useful fact:
Theorem 6. Let R be a commutative ring with unity. Let m N.
Consider the ring R [X1 , X2 , ..., Xm ] (the polynomial ring over R in m indeterminates X1 , X2 , ..., Xm ). Then,


1jm 
Y
mj
c
mj X1 , X2 , ..., Xi , ..., Xm
=
(Xi Xj ) .
det (1)
1im

Proof of Theorem 6. Let V =

1jm
Xij1 1im .

i {1, 2, ..., m} and j {1, 2, ..., m}, and



1jm 
det V = det Xij1 1im =

(i,j){1,2,...,m}2 ;
j>i


Then, V

j
i

= Xij1 for every

(Xi Xj )

(3)

(i,j){1,2,...,m} ;
i>j

(by Theorem 3, applied to S = R [X1 , X2 , ..., Xm ] and ai = Xi ).



1jm
ci , ..., Xm
Let W = (1)mj mj X1 , X2 , ..., X
. Then,
1im
 


j
ci , ..., Xm for every i {1, 2, ..., m} and j
W
= (1)mj mj X1 , X2 , ..., X
i
{1, 2, ..., m}.
For every i {1, 2, ..., m} and j {1, 2, ..., m}, we can apply (1) to x = Xj and
ak = Xk , and obtain
Y
`{1,2,...,m}\{i}

(Xj X` ) =

m1
X



ci , ..., Xm X m1k .
(1)k k X1 , X2 , ..., X
j

k=0

Now, for every i {1, 2, ..., m} and j {1, 2, ..., m}, we have

(4)

WV

m
X

j
i

m
X


k

W
=
i
k=1
| {z }
mk
ci ,...,Xm )
=(1)
mk (X1 ,X2 ,...,X

j
k

V
|2 {z3 }
k 5 k1
=V 4
=Xj
j



ci , ..., Xm X k1
(1)mk mk X1 , X2 , ..., X
j

k=1
m1
X



ci , ..., Xm X m1k
(1)k k X1 , X2 , ..., X
j

(here, we substituted k for m k in the sum)

k=0

(Xj X` )

(by (4)) .

(5)

`{1,2,...,m}\{i}

Thus, if j 6= i, then W V
Q
but the product


 

j
Q
j
T
= 0 (since W V
=
(Xj X` ) ,
i
i
`{1,2,...,m}\{i}
(Xj X` ) contains the factor Xj Xj = 0 and thus

`{1,2,...,m}\{i}

equals 0). Hence, the matrix W V T is diagonal. Therefore,

det W V

m
Y

WV

i=1

i
i

m
Y


=

(Xi X` )

i=1 `{1,2,...,m}\{i}

since (5), applied to j = i, yields W V

Y

(Xi Xj ) =

(i,`){1,2,...,m}2 ;
`6=i

T

i
i

`{1,2,...,m}\{i}

(Xi Xj ) =

(i,j){1,2,...,m}2 ;
j6=i

(Xi X` )
Y

(Xi Xj )

(i,j){1,2,...,m}2 ;
j>i

(Xi Xj )

(i,j){1,2,...,m}2 ;
i>j




2
since the
| j 6= i  is the union of the two disjoint
sets
 set (i, j) {1, 2,2 ..., m}

.
(i, j) {1, 2, ..., m} | j > i and (i, j) {1, 2, ..., m}2 | i > j

But on the other hand,

Y

det W V T = det W det V T = det W

(Xi Xj )
2

(i,j){1,2,...,m} ;
i>j

(since det V T = det V =

(Xi Xj )). Hence,

(i,j){1,2,...,m}2 ;
i>j

det W

(Xi Xj ) = det W V T =
2

(Xi Xj )
2

(i,j){1,2,...,m} ;
i>j

But since the element

Y
(i,j){1,2,...,m} ;
j>i

(Xi Xj ) .
2

(i,j){1,2,...,m} ;
i>j

(Xi Xj ) of the ring R [X1 , X2 , ..., Xm ] is not a

(i,j){1,2,...,m}2 ;
i>j

zero divisor (according to Lemma 5), this yields

Y
det W =
(Xi Xj ) .
(i,j){1,2,...,m}2 ;
j>i



1jm
ci , ..., Xm
Since W = (1)mj mj X1 , X2 , ..., X
, this becomes
1im

det



mj

(1)

ci , ..., Xm
mj X1 , X2 , ..., X

1jm 

1im

(Xi Xj ) .
2

(i,j){1,2,...,m} ;
j>i

Thus, Theorem 6 is proven.

Next, we show:
Theorem 7. Let R be a commutative ring with unity. Let m N. Let a1 ,
a2 , ..., am be m elements of R. Let b1 , b2 , ..., bm be m elements of R. Then,

1jm
Y
Y
=
det
(aj b` )
((ai aj ) (bj bi )) .
`{1,2,...,m}\{i}

(i,j){1,2,...,m}2 ;
i>j

1im

Proof of Theorem 7. Consider the ring R [X1 , X2 , ..., Xm ] (the polynomial ring over
R in m indeterminates X1 , X2 , ..., Xm ). 

1jm
j
Let Ve = aj1
. Then, Ve
for every i {1, 2, ..., m} and j
= aj1
i
i
1im
i
{1, 2, ..., m}.


1jm
mj
c
Let W = (1)
mj X1 , X2 , ..., Xi , ..., Xm
. Then,
1im
 


j
ci , ..., Xm for every i {1, 2, ..., m} and j
W
= (1)mj mj X1 , X2 , ..., X
i
{1, 2, ..., m}.
For every i {1, 2, ..., m} and j {1, 2, ..., m}, we can apply (1) to x = aj and
ak = Xk , and obtain
Y
`{1,2,...,m}\{i}

(aj X` ) =

m1
X



ci , ..., Xm am1k .
(1)k k X1 , X2 , ..., X
j

k=0

Now, for every i {1, 2, ..., m} and j {1, 2, ..., m}, we have

(6)

m
 j  X
=
WV
i

 

j
k
T
Ve
W
k
i
k=1
|2 {z3 }
| {z }
mk
c
=(1)
mk (X1 ,X2 ,...,Xi ,...,Xm )
k 5 k1
=Ve 4
=aj
j
m


X
ci , ..., Xm ak1
(1)mk mk X1 , X2 , ..., X
=
j

eT

k=1
m1
X



ci , ..., Xm am1k
(1)k k X1 , X2 , ..., X
j

(here, we substituted k for m k in the sum)

k=0

(aj X` )

(by (6)) .

`{1,2,...,m}\{i}

Hence,
1jm

W Ve T =

(aj X` )

`{1,2,...,m}\{i}

Thus,

.
1im

1jm



 
Y

T
e

det
(aj X` )
= det W det Ve T
= det W V

`{1,2,...,m}\{i}
1im
|
{z
}
=W Ve T

Y
2

(i,j){1,2,...,m} ;
j>i

(i,j){1,2,...,m} ;
i>j

1jm 

mj

ci , ..., Xm
(1)
mj X1 , X2 , ..., X
=
(Xi Xj )
since det W = det
1im

(i,j){1,2,...,m}2 ;

j>i
 



Q
j1 1jm

T
e
e
by Theorem 6 and det V
= det V = det ai 1im =
(ai aj )

(i,j){1,2,...,m}2 ;

i>j
by Theorem 3
Y
Y
(Xj Xi )
(ai aj )
(i,j){1,2,...,m}2 ;
i>j

(here, we renamed i and j as j and i in the first product)

Y
Y
(Xj Xi )
(ai aj )
(i,j){1,2,...,m}2 ;
i>j

(ai aj )
2



(j,i){1,2,...,m}2 ;
i>j

(Xi Xj )

(i,j){1,2,...,m}2 ;
i>j

((Xj Xi ) (ai aj )) =
2

(i,j){1,2,...,m} ;
i>j

((ai aj ) (Xj Xi )) .
2

(i,j){1,2,...,m} ;
i>j

Both sides of this identity are polynomials over the ring R in m indeterminates X1 , X2 ,
..., Xm . Evaluating these polynomials at X1 = b1 , X2 = b2 , ..., Xm = bm , we obtain

1jm
Y
Y
=
det
(aj b` )
((ai aj ) (bj bi )) .
`{1,2,...,m}\{i}

(i,j){1,2,...,m}2 ;
i>j

1im

This proves Theorem 7.

Proof of Theorem 2. For every i {1, 2, ..., m} and j {1, 2, ..., m} , we have
Q
(aj b` )
1
`{1,2,...,m}\{i}
Q
=
.
aj bi
(aj b` )
`{1,2,...,m}


Hence, the matrix

1
aj b i

!1jm

1jm
Q

is what we obtain if we take the matrix

(aj b` )
`{1,2,...,m}\{i}
Q
(aj b` ) for every j {1, 2, ..., m}. Therefore,

1im

and divide its j-th column by

`{1,2,...,m}


det

1
aj b i

1jm

!1jm
Q

det

(aj b` )

`{1,2,...,m}\{i}

1im

(aj b` )

1im
j{1,2,...,m} `{1,2,...,m}

det

!1jm
Q

(aj b` )

`{1,2,...,m}\{i}

1im

Q
(`,j){1,2,...,m}

(aj b` )
2

((ai aj ) (bj bi ))

(i,j){1,2,...,m}2 ;
i>j

Q
(i,j){1,2,...,m}

(aj bi )
2

(by Theorem 7). Thus, Theorem 2 is proven.

Theorem 8. Let n N. Let a1 , a2 , ..., an be n pairwise distinct reals. Let
c be a real such that ai + aj + c > 0 for every (i, j) {1, 2, ..., n}2 . Then,

1jn
1
the matrix
Rnn is positive definite.
ai + aj + c 1in
1jn
 
1
1
j
Proof of Theorem 8. Let A =
. Then, A
=
i
ai + aj + c 1in
ai + aj + c
for every i {1, 2, ..., m} and j {1, 2, ..., m}.
Thus, A Rnn is a symmetric n n matrix.
Define n reals b1 , b2 , ..., bn by bi = ai c for every i {1, 2, ..., n} . Then, aj 6= bi
for every (i, j) {1, 2, ..., n}2 (since aj bi = aj (ai c) = ai + aj + c > 0).


1im

Now, every m {1, 2, ..., n} satisfies

  1jm !

1jm !
1
j
det
A
= det
i
aj bi 1im
1im

 

1
1
1
j
since A
=
=
=
i
ai + aj + c
aj (ai c)
aj bi
Q
((ai aj ) (bj bi ))
=

(i,j){1,2,...,m}2 ;
i>j

by Theorem 2, since aj 6= bi for every (i, j) {1, 2, ..., m}2

(aj bi )

(i,j){1,2,...,m}2

(ai aj )2

Q
2

(i,j){1,2,...,m} ;
i>j

(ai + aj + c)

(i,j){1,2,...,m}2

since (ai aj ) (bj bi ) = (ai aj ) ((aj c) (ai c)) = (ai aj )2

and aj bi = aj (ai c) = ai + aj + c

>0
(since (ai aj )2 > 0 for every (i, j) {1, 2, ..., m}2 satisfying i > j (because a1 , a2 , ...,
an are pairwise distinct, so that ai 6= aj , thus ai aj 6= 0), and ai + aj + c > 0 for every
(i, j) {1, 2, ..., m}2 ).
Hence,
according
to Theorem 1, the symmetric matrix
definite. Since

1jn
 A is positive
1jn
1
1
A=
, this means that the matrix
is positive
ai + aj + c 1in
ai + aj + c 1in
definite. Thus, Theorem 8 is proven.
Corollary 9. Let n N. Let a1 , a2 , ..., an be n pairwise distinct reals.
Let c be a real such that ai + aj + c > 0 for every (i, j) {1, 2, ..., n}2 . Let
n P
n
P
vi vj
0 holds,
v1 , v2 , ..., vn be n reals. Then, the inequality
i=1 j=1 ai + aj + c
with equality if and only if v1 = v2 = ... = vn = 0.

v1
v2

Proof of Corollary 9. Define a vector v Rn by v =

... . Then,
vn

1jn
n X
n X
n
n
X
X
1
1
vi vj
T
v
v=
vi vj =
.
ai + aj + c 1in
a
+
a
+
c
a
+
a
+
c
i
j
i
j
i=1 j=1
i=1 j=1

(7)

Also, obviously,
v = 0 holds if and only if v1 = v2 = ... = vn = 0.
(8)

1jn
1
Now, since the matrix
Rnn is positive definite (by Theorem
ai + aj + c 1in

1jn
1
T
8), we have v
v 0, with equality if and only if v = 0. According
ai + aj + c 1in
8

n P
n
P

vi vj
0, with equality if and only if
i=1 j=1 ai + aj + c
v1 = v2 = ... = vn = 0. Thus, Corollary 9 is proven.

to (7) and (8), this means that

Corollary 10. Let n N. Let a1 , a2 , ..., an be n pairwise distinct reals.

Let c be a real such that ai + aj + c > 0 for every (i, j) {1, 2, ..., n}2 .
n P
n
P
ai aj wi wj

Let w1 , w2 , ..., wn be n reals. Then, the inequality

i=1 j=1 ai + aj + c
2
n
P
c
wi holds, with equality if and only if (c + a1 ) w1 = (c + a2 ) w2 =
i=1

... = (c + an ) wn = 0.
Proof of Corollary 10. Define n reals v1 , v2 , ..., vn by vi = (c + ai ) wi for every
i {1, 2, ..., n} .
Then,

!2
n X
n
n
X
X
ai aj wi wj
c
wi
a
+
a
+
c
i
j
i=1 j=1
i=1
!2
n X
n
n X
n
n X
n
n
X
X
X
X
ai aj wi wj
ai aj wi wj
=
=
+c
wi
+c
wi wj
a
+
a
+
c
a
+
a
+
c
i
j
i
j
i=1 j=1
i=1 j=1
i=1
i=1 j=1

!
2
n X
n
n
X
X
since
wi wj
=
wi
i=1

=
=

n X
n 
X
i=1 j=1
n X
n
X
i=1 j=1

i=1 j=1

ai aj wi wj
+ cwi wj
ai + aj + c

(c + ai ) (c + aj )
wi wj =
ai + aj + c

n X
n 
X

i=1 j=1
n
n
XX
i=1 j=1


ai aj
+ c wi wj
ai + aj + c
n

(c + ai ) wi (c + aj ) wj X X
vi vj
=
ai + aj + c
a + aj + c
i=1 j=1 i

(since (c + ai ) wi = vi and (c + aj ) wj = vj ). Hence,

!2
n
n X
n
n X
n
X
X
X
ai aj wi wj
vi vj
c
wi
holds if and only if
0.
a
+
a
+
c
a
+
a
+
c
i
j
i
j
i=1 j=1
i=1
i=1 j=1
(9)
Also, clearly,
v1 = v2 = ... = vn = 0 holds if and only if (c + a1 ) w1 = (c + a2 ) w2 = ... = (c + an ) wn = 0.
(10)
n P
n
P
vi vj
By Corollary 9, the inequality
0 holds, with equality if and only
i=1 j=1 ai + aj + c
n P
n
P
ai aj wi wj

if v1 = v2 = ... = vn = 0. According to (9) and (10), this means that

i=1 j=1 ai + aj + c
n
2
P
c
wi , with equality if and only if (c + a1 ) w1 = (c + a2 ) w2 = ... = (c + an ) wn =
i=1

0. Thus, Corollary 10 is proven.

The problem follows from Corollary 10 (applied to c = 1 and ai = i).
9