404 APPENDIX B. CLASSIFICATION OF REAL NUMBERS
where A(k) signifies the roots of Qk (x) ,{
α (k)1 , · · · ,α (k)mk
}. Thus, by the symmetric
polynomial theorem applied to the commutative ring Q [A(1) , · · · ,A(n−1)], the abovepolynomial is of the form
µ(r)
∑l=0
xµ(r)−l∑kl
Bkl (A(1) , · · · ,A(n−1))skl
11 · · ·s
kln
µ(r)
where the sk is one of the elementary symmetric polynomials in{α (n)1 , · · · ,α (n)mn
}and Bkl is symmetric in α (k)1 ,α (k)2 , · · · ,α (k)mk
for each k ≤ n−1 and
Bkl ∈Q [A(1) , · · · ,A(n−1)] .
Now do to Bkl what was just done to Bl featuring A(n−1) this time, and continue tilleventually you obtain for the coefficient of xµ(r)−l a large sum of rational numbers times aproduct of symmetric polynomials in A(1) ,A(2) , etc. By Theorem B.2.7 applied repeat-edly, beginning with A(1) and then to A(2) and so forth, one finds that the coefficient ofxµ(r)−l is a rational number and so the β (r) j for j ≤ µ (r) are algebraic numbers and rootsof a polynomial which has rational coefficients, namely the one in 2.17, hence roots of apolynomial with integer coefficients. Now 2.16 contradicts Corollary B.3.6. ■
Note this lemma is sufficient to prove Lindermann’s theorem that π is transcendental.Here is why. If π is algebraic, then so is iπ and so from this lemma, e0 +eiπ ̸= 0 but this isnot the case because eiπ =−1.
The next theorem is the main result, the Lindermann Weierstrass theorem. It replacesthe integers bi in the above lemma with algebraic numbers.
Theorem B.3.8 Suppose a(1) , · · · ,a(n) are nonzero algebraic numbers and sup-pose
α (1) , · · · ,α (n)
are distinct algebraic numbers. Then
a(1)eα(1)+a(2)eα(2)+ · · ·+a(n)eα(n) ̸= 0
Proof: Suppose a( j)≡ a( j)1 is a root of the polynomial
v jxm j + · · ·+u j
where v j,u j ̸= 0. Let the roots of this polynomial be a( j)1 , · · · ,a( j)m j. Suppose to the
contrary thata(1)1 eα(1)+a(2)1 eα(2)+ · · ·+a(n)1 eα(n) = 0
Then consider the big product
∏(i1,··· ,in)
ik∈{1,··· ,mk}
(a(1)i1 eα(1)+a(2)i2 eα(2)+ · · ·+a(n)in eα(n)
)(2.18)