Documentation

Linglib.Phenomena.Presupposition.Studies.White2014

White 2014: The Modalized Complement Analysis of forget #

@cite{white-2014}

White 2014 maintains a uniformly factive denotation for forget by positing a covert root modal in non-finite complements (the Modalized Complement Analysis, MCA). The MCA captures Williams' (1)/(4) and (2)/(5) data straightforwardly but was later shown to overgenerate: PRO-ing gerunds get a non-modal presupposition (cf. @cite{williams-2026}, §3.1.1) that the MCA cannot predict.

This file (a) verifies the MCA's coverage of finite-CP and plain-infinitive data, (b) records the gerund overprediction as a check against the empirical record in Studies/Williams2026.lean, and (c) confirms Fragment-level consistency: the English Fragment's split of forget into forget (negative implicative, infinitival) and forget_rog (factive, finite) aligns with the typology MCA + pre-existence together predict.

The pre-existence apparatus in Theories/Semantics/Attitudes/PreExistence.lean is post-2014 (Bondarenko 2019/2020, taken up by @cite{williams-2026}); where this file uses needsModalInsertion, treat that as the project-canonical upgrade of White's structural prediction, with the gerund case flagged as a known overprediction of the original MCA.

§1. MCA coverage of the canonical contrast #

White 2014 was designed to capture the finite-CP / plain-infinitive asymmetry uniformly. The MCA succeeds on these two cases.

§2. The gerund overprediction #

The PRO-ing gerund is non-finite but its presupposition is non-modal (Williams §3.1.1). The MCA's mcaPrediction is !isFinite, so it predicts modal insertion for the gerund — wrong.

§3. Pre-existence as the post-2014 fix #

The pre-existence-based needsModalInsertion correctly predicts non-modal for the gerund. This is post-2014 territory; the theorems here document the contrast between White's MCA and its successor.

§4. Fragment consistency #

The English Fragment splits forget into two VerbEntry records, one for the implicative use and one for the factive/rogative use. The split is a practical separation of entailment patterns, not a claim of lexical ambiguity (which Williams 2026 explicitly rejects). These theorems document that the Fragment's complement-type and factivity assignments align with the MCA's predictions, restricted to the cases MCA gets right.