Montagues original goal was to characterize the notions of truth and validity. My impression after discussion with a semanticist is that current approaches in the field give a formalization and then argue for it on the basis of empirical evidence.

I am interested in cases/papers/seminal results where we have robust empirical evidence to prefer one theory T1 over another theory T2 where T1's notion of validity or truth is substantially different than that of T1's. I am also interested in the more general question of two formalizations F1, F2 of any linguistic phenomenon P where F1, F2 suggest substantially different conceptions of P, and empirical evidence favors one of them at least modestly robustly over the other.

If there are such results, does the formalization play a truly necessary role? Ie, perhaps P is characterized mathematically, and hence we must formalize, or perhaps the empirical evidence by itself must be viewed in the light of some reasonable assumptions and the formal approach makes the assumptions evident? If there are no such results, what is the motivation for formal approaches in linguistics? Ie, what do formal semanticians or synaticists write on their grants to get money?

Context: I have taken a course in formal semantics, the math seems good to me, and the work seems philosophically interesting to me, but I'm struggling to connect it with the broader paradigms in linguistics.

  • I asked a related question previously: What explanatory advantages does so-called "type theory" have?
    – curiousdannii
    Commented Mar 26 at 6:14
  • hi @curiousdannii thanks. This is indeed related, but still not quite my question. Indeed, perhaps pedagogy is different now. My course did not even bother to mention an alternative to type theory for formalizing semantics.
    – emesupap
    Commented Mar 26 at 23:09
  • however, the connection between categorical grammars and type theories is an interesting one (top answer on your linked post). I know type theoretical semantics are easier to implement, so perhaps this is the primary advantage (along with the naturality of using a typed semantics for natural language). If you wish to post a fleshed out answer regarding this, I would be thankful
    – emesupap
    Commented Mar 26 at 23:24
  • Well my perspective is that the whole field of "formal" linguistics is misguided and unprofitable. So I'm not the right person to ask...
    – curiousdannii
    Commented Mar 27 at 0:17
  • @curiousdannii I am also open to you sketching out why, with sources if possible
    – emesupap
    Commented Mar 27 at 3:49


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.