Problem to understand EA proof

I have problem to understand proof of error analysis for linear approximation. I Quotation my book “If f’’(t) exist for all t in an interval containing a and x, then there exists some point X between a and x such that the error E(x)=f(x)-L(x) in the linear approximation f(x)~L(x)=f(a)+f’(a)(x-a) satisfies the formula E(x)=f’’(X)/2 * (x-a)^2”. The formula is obtained by divide E(x) whit (x-a)^2 and then apply the Generalized Mean-Value Theorem twice. I understand everything except where the (x-a)^2 come from? E(x)/(x-a)^2 = [E(x)-E(a)]/[(x-a)^2-(a-a)^2] = E’(c)/2(c-a) = [f’(c)-f’(a)]/2(c-a) = 1/2 f’’(X). Noting that E(t)=f(t)-f(a)-f''(a)(t-a), E''(t)=f''(t)-f''(a) and E(a)=0.