Theories which analyze probability in terms of beliefs or attitudes rather than anything in the world itself.
For one theory, associated mainly with Bruno De Finetti (1906-1985), the degree of probability of something is the degree of the speaker's belief, measured by his betting behavior, but subject to the constraint that his bets must be 'coherent'; that is, he must not bet in such a way as to lose whatever happens (sometimes called 'having a Dutch book made against one').
This constraint still leaves probabilities dependent on the vagaries of individual attitudes, unless we substitute those of 'the rational man' - but that takes us away from subjectivism. Others, notably Stephen Edelston Toulmin (1922-), offer a speech act theory whereby to call something probable is to assert it, though only tentatively.
This may well apply to some uses of 'probably', but hardly to all, and shares the objections to other speech act theories.
H E Kyburg, Probability and Inductive Logic (1970), ch. 6