Special double-session in connection with the Rolf Schock Prize 2024: 

Craige Roberts (Ohio State/Barnard College, NY) & Barbara Partee (UMass, Amherst)

10.00-12.00 Craige Roberts (Ohio State/Barnard College, NY) Language and Logic: Ideas and Controversies in the History of Formal SemantiCompetence

Ideas and Controversies in the History of Formal Semantics

13.30-15.30 Barbara Partee (UMass, Amherst): Attitudes de re: Semantics and dynamic pragmatics

Map Venue

ABSTRACTS

Language and Logic: Ideas and Controversies in the History of Formal Semantics 

“Semantics” can mean quite different things in different contexts. There are many reasons to be interested in meaning, and it can be studied from many perspectives. The history of formal semantics over the last 60 years is a story of collaboration among linguists, logicians, and philosophers.

One central figure was the logician and philosopher Richard Montague, a student of Tarski’s, whose seminal works on the formal semantics of natural language date from the late 1960’s and early 1970’s. Starting in the 1970s, philosophers and especially linguists worked to integrate Montague’s semantics with Chomskyan syntax.

Substantive progress was rapid and has been successful, but there were foundational questions that have remained unresolved. While Chomsky and most linguists view linguistics as a cognitive science, the Fregean tradition in which Montague worked is strongly anti-psychologistic. Linguists who do formal semantics have generally also seen themselves as contributing to cognitive science, but this seems to require, at least implicitly, some rethinking of the traditional Chomskyan notion of “the competence of the native speaker”.

The history of these foundational issues in formal semantics, including the question of whether meanings are “in the head” suggests, following work of the philosophers Robert Stalnaker and Tyler Burge, that semantics, unlike syntax, needs a less “narrowly psychological” view of what it is to know a language.

Attitudes de re: Semantics and dynamic pragmatics

I argue that de re interpretations arise quite regularly as a result of two types of presuppositions associated with utterances wherein a definite NP (pronoun, definite or demonstrative description, or proper name) occurs in the complement of an attitude predicate or some other intensional context. The first (similar to that in Maier 2019) is an anaphoric presupposition triggered by the use of the definite NP (Heim 1982, 1983, 1992). Though the NP remains in situ in the embedded clause, non-local satisfaction of this presupposition results in truth conditional effects akin to those that would result from giving the NP semantic wide scope over the attitude: presuppositionally triggered wide pseudo-scope. In the resulting interpretation, the triggering definite’s contribution to semantic content is simply the res which is the actual denotation of the wider scope antecedent of the definite, yielding interpretation de re without syntactically unmotivated NP movement at LF (Keshet & Schwarz 2019) or structured propositions (von Stechow & Cresswell 1982). The second presupposition is a background condition entailed by any de re attitude attribution: Epistemically limited humans can never know a res directly or completely, but only under some guise or other with which they are acquainted. So an agent can only hold an attitude toward a res under some guise reflecting their acquaintance. This captures the acquaintance requirement of Kaplan (1986), but here arises as an ontological precondition on the existence of a de re attitude, a non-anaphoric presupposition (Roberts & Simons 2024). Thus, acquaintance under a guise is an entailment of the triggering attitude in conjunction with the wide pseudo-scope of the definite, arising from our world knowledge of the corresponding type of situation. To model it, there is no need to stipulate silent existential operators or concept generators at LF, avoiding Lederman’s (2021) problems for Percus & Sauerland (2003).

This account is straightforwardly realized using independently motivated tools in a dynamic pragmatics: presuppositionally triggered pseudo-scope for definites (including names), and the notion of presuppositional background content. Drawing on Aloní (2001), the notion of a guise is modeled as an individual concept in a conceptual cover, thereby inheriting her account’s advantages over previous accounts of the de re. In a de re attribution, the nature of the guise under which the agent is acquainted with the res is not semantically specified; the semantic content of the attitude report merely entails that there is some counterpart relation that maps the actual res to entities in the agent’s belief worlds which have the properties predicated of the embedded definite NP. But the nature of the entailed guise is often pragmatically evident from context, or even explicitly given by an appositive (Soames 2002). When this is the case, we can contextually enrich the meaning of the utterance by adding the presumption that the entailed guise of acquaintance is the particular contextually retrieved guise. Unlike with Aloní’s perspective shifting operator ℘ or Stalnaker’s (1979) diagonalization, on the present account we needn’t shift the NP’s semantic interpretation (Gluer & Pagin 2006, 2012); context merely pragmatically enriches that interpretation to make the presupposed guise more specific.

All this sheds new light on a number of classic problems pertaining to the de re (Frege 1892; Quine 1956,1961; Geach 1967, etc.). Moreover, with the addition of centered worlds (Lewis 1979), the theory straightforwardly extends to account for de se interpretations, while avoiding problems of the de se, pointed out by Ninan (2016), for the classical doctrine of propositions.

Rolf Schock Prize 2024