If anyone asks “What is truth?” and the answer is “That which corresponds to reality”, “truth” and “that which” aren’t and can’t be reality itself. It’s then only at best “ideas about” or “sentences about” reality. Of course this entails truth as mind-dependent rather than strictly “out there” as some property of reality. This naturally exposes that ideas and sentences about reality are subjective because we’re producing them, and, that these ideas and sentences are still objective, given they must relate to reality somehow. So, this brings us to “correspondence” and the idea that “corresponds” can only mean “descriptive of”. In that case, A or B (two different descriptions of reality) may or may not “correspond to reality”. The questions are then which best describes and which is more justified to take on.
Since the epistemological task is to say whether, or how A or B is assertable, justification is synonymous with truth, not correspondence.
That is to say, there is no property of A or B that we rely on to suggest correspondence. So at least in our asserting either using those two terms (truth, correspondence), the truth and correspondence of A or B only obtain via justification. Both A and B may seek to describe reality, however, only one may end up being a better description, or the best description. Saying then that A corresponds, or that B corresponds, is to say nothing about correspondence. It’s that A is better at being descriptive than B; or vice versa. Whether warrant for A or B is “one seems to do a better job than the other” or whether A or B is a rational conclusion, or whether A or B is concluded from some reliable methodology or is evidential, these are all about which has warrant, justification.
The only imaginable response to “That which corresponds to reality”, it at least seems to me, is the dialectic translation: “A descriptive idea about reality that warrants assertability”.