Large Scale Conlanging: An Areal Tendency

Consider the correlation between definiteness and direct objects. Granted, topics and subjects too have a similar and probably stronger correlation with definites, but let's disregard that for now. Objects, unlike subjects, tend to be more marked than subjects, which is what we want for this areal tendency.

Have object marking tend to become definiteness marking for subjects as well as objects over time. How this would work out for case-marking is obvious. The challenge might rather appear in considering how a language that mostly does mark for object and subject congruence on the verbs would develop within this tendency.