I took this to mean, e.g., why is o-toluic acid more acidic than m-toluic acid or p-toluic acid? It's a phenomenon typically referred to as the ortho effect. So a couple of things are wrong: acidity can increase with resonance, but there are also other factors at play. Typically, the factor that most determines acidity is the stability of the conjugate base. So the more stable the deprotonated species is, the more acidic the protonated species will be. In the case of benzoic acid, the benzene ring pushes electron density onto the carboxyl group. The easiest way to see this is to compare the pKa's of formic acid (3.7) and benzoic acid (4.2), since formic acid is about as close as we can get to "nothing" as a substituent on the carboxyl group. In the bare benzoic acid (or in m- or p- substituted benzoic acids) the benzene ring and carboxyl group are aligned in the same plane. The benzene ring can then donate electrons onto the highly electronegative carboxyl group. This strengthens the bond between the carboxyl O and the acidic proton, rendering the compound less acidic (if the benzene were instead electron-withdrawing, the resulting compound would be more acidic). In the case of ortho-substituted benzoic acid, this planarity is broken by steric hindrance. The electron-donating effect of the benzene ring is reduced or absent, and the compound is therefore more acidic.
This is the standard textbook explanation of the ortho effect, and I'll be the first to admit that it seems a little wishy washy to me. For instance, in most cases meta-substituted benzoic acids are more acidic than para-substituted ones, which can't be explained by a steric effect (the order of acidity being o > m > p). However, we know that the phenomenon is relatively independent of whether the ortho group is electron donating or electron withdrawing, which seems to point away from an inductive effect. It seems like a good area for a research project.