Proposition

Let RR be a commutative ring, and let a,b,cR[x]a,b,c \in R[x] be arbitrary. Then a(bc)=(ab)c.a(bc) = (ab)c.

Proof

Recall that the product of two elements of R[x]R[x] is given by

(n=0m1anxn)(n=0m2bnxn)=n=0m1+m2(k=0nakbnkindices add up to n)xn=n=0m1+m2(i+j=naibj)xn.\begin{align*}\left(\sum_{n=0}^{m_1}a_nx^n\right)\left(\sum_{n=0}^{m_2}b_nx^n\right) &= \sum_{n=0}^{m_1+m_2} \left(\sum_{k=0}^n \underbrace{a_kb_{n-k}}_{\text{indices add up to } n}\right)x^n \\ &= \sum_{n=0}^{m_1+m_2} \left(\sum_{i+j=n}a_ib_j\right)x^n.\end{align*}

Note that it is commutative, as neither the coefficients nor the bounds depend on the order due to commutativity of addition and multiplication in RR.

Using this, we can derive the following.

((n=0m1anxn)(n=0m2bnxn))(n=0m3cnxn)=(n=0m1+m2(k=0nakbnk)xn)(n=0m3cnxn)=n=0m1+m2+m3(l=0n(k=0lakblk)cnl)xn=n=0m1+m2+m3(l=0n(k=0lakblkcnlindices add up to n))xn=n=0m1+m2+m3(i+j+k=naibjck)xn\begin{align*} \left(\left(\sum_{n=0}^{m_1}a_nx^n\right)\left(\sum_{n=0}^{m_2}b_nx^n\right)\right)\left(\sum_{n=0}^{m_3}c_nx^n\right) &= \left(\sum_{n=0}^{m_1+m_2} \left(\sum_{k=0}^na_kb_{n-k}\right)x^n\right)\left(\sum_{n=0}^{m_3}c_nx^n\right) \\ &= \sum_{n=0}^{m_1+m_2+m_3} \left(\sum_{l=0}^n\left( \sum_{k=0}^la_kb_{l-k} \right)c_{n-l}\right)x^n\\ &= \sum_{n=0}^{m_1+m_2+m_3} \left(\sum_{l=0}^n\left( \sum_{k=0}^l \underbrace{a_kb_{l-k}c_{n-l}}_{\text{indices add up to } n} \right)\right)x^n\\ &= \sum_{n=0}^{m_1+m_2+m_3} \left(\sum_{i+j+k=n} a_ib_jc_k \right)x^n\\ \end{align*}

Now, note that a(bc)=(bc)aa(bc) = (bc)a. Applying the identity above to the product (bc)a(bc)a produces the same result as when we apply it to (ab)c(ab)c (modulo the order of multiplication of coefficients, but RR is commutative). Thus,

(bc)a=a(bc)=(ab)c.(bc)a=a(bc)=(ab)c.

\blacksquare

NOTE

Technically, we don’t need to assume commutativity of RR. We could do the same manipulation the second time and produce the same result. I don’t think this would contribute to any further insight, however.