lingliblog

Psych Verbs Need a Mind: Compositional Denotations via Cognitive Situation Models

Fri, 27 Feb 2026 00:00:00 +0000

Psych verbs like frighten and concern differ in opacity, temporal structure, and causation — but nobody has derived these differences from a single compositional denotation grounded in cognitive architecture. We do, and opacity falls out by computation.

Nine Theories of Tense in One Type System

Sun, 15 Feb 2026 00:00:00 +0000

How Linglib formalizes nine competing theories of temporal reference and compares them against 22 phenomena.

Numerals, Knowledge, and Type-Shifting

Sat, 14 Feb 2026 00:00:00 +0000

Kennedy’s exact numeral semantics and Goodman & Stuhlmuller’s knowledge-sensitivity results look incompatible. Type-shifting resolves it.

Generalized Quantifiers as a Case Study in Cross-Module Verification

Tue, 10 Feb 2026 00:00:00 +0000

Tracing generalized quantifier theory from abstract properties through lexical entries to RSA predictions.

A Tour of Linglib

Mon, 09 Feb 2026 00:00:00 +0000

Introducing Linglib, a Lean 4 library for cumulative, machine-checkable formal linguistics.

A Map of Formal Linguistics

Mon, 01 Jan 0001 00:00:00 +0000

An interactive dependency graph of the mathematical foundations underlying LingLib — five independent roots, connected by bridges where the interesting linguistic theory lives.

Bibliography

Mon, 01 Jan 0001 00:00:00 +0000

Papers referenced, cited, or formalized in Linglib. Each entry links to the corresponding Lean source file.

Aarts, B. (1989). Verb-preposition constructions and small clauses in English. Journal of Linguistics, 25(2), 277–290.HaddicanEtAl2026

Roadmap

Mon, 01 Jan 0001 00:00:00 +0000

See the full roadmap on GitHub.

North Stars

For phenomenon P with behavioral data D, prove that theory T₁ predicts D and theory T₂ doesn’t (or both do, via different assumptions).

Priority Phenomena

Phenomenon	Why Interesting
Scope Freezing	Prove CCG and Minimalism predict same freezing via different mechanisms
RSA ≅ EXH Equivalence	Characterize when grammar/pragmatics debates are notational

Key Conjectures

rsa_fixed_point_unique: RSA iteration converges to a unique fixed point for α > 0
lexicon_refinement_monotone: Refining denotations can only strengthen pragmatic inferences
rsa_tropical_limit: α → ∞ recovers iterated best response (tropical semiring)

Short-term

Fill sorry proofs (scope freezing, CCG expressiveness)
Grounding proofs for lexical uncertainty, attitude embeddings
Scenario combinators (withUtterances, restrictWorlds)
Documentation: tutorials, replication guides

Long-term

Neural-symbolic emergence (Futrell & Hahn, RSA from LLM)
Speech acts formalization
Full Horn (1972) scale theory
Neo-Davidsonian event semantics