Williams 2026: The Presuppositions of forget #
@cite{williams-2026} @cite{kiparsky-kiparsky-1970} @cite{white-2014}
Empirical data and classification from @cite{williams-2026}, who argues that forget is canonically factive, independent of complement type, but the content of the factive presupposition varies by frame:
- Cognition reading (finite CP): non-modal — directly presupposes complement truth
- PRO-ing gerund: non-modal — patterns with finite CP, refuting an MCA-only account
- Psych-action reading (plain infinitive): modal — presupposes a modalized proposition (an obligation/plan held before the forgetting)
The cross-linguistic generalization is the SMINC (Selectivity of Modal Insertion in Non-finite Contexts): a covert modal heads the complement of forget iff the embedded lexical verb is a plain infinitive, and is banned elsewhere. Williams' Table 1 (§3.1.3) attests this pattern across English, Spanish, Italian, German, and Hungarian; this file records the English data points.
Scope #
Williams' opening contrast (his (1)/(2)) is that forgot to V both fails
to presuppose the lower event AND is implicatively-negative
(forgot to V → ¬V; @cite{karttunen-1971}, Karttunen 1971a, Horn 1972,
Jackendoff 1985). Williams (fn 1) brackets the implicative-negative entailment
to focus on presupposition; he does not refute it. The
implicativeNegative field below records it for cross-paper consistency
with the Karttunen tradition (formalized in
Phenomena/Complementation/Studies/Karttunen1971.lean and
Theories/Semantics/Causation/Implicative.lean).
A second presupposition Williams identifies for the psych-action reading, the necessity-and-sufficiency presupposition (his (5b), citing Karttunen 1971a — "if John stopped, it is because he remembered to do it") is left to future work in the paper and is not encoded here.
Naming #
The English Fragment splits forget into two VerbEntry records:
forget (negative implicative, infinitival) and forget_rog
(factive/rogative, finite). Williams argues these are one lexical item
with uniform factivity and a frame-driven presupposition split (the
pre-existence analysis is in Theories/Semantics/Attitudes/PreExistence.lean).
The Fragment split is a practical separation of entailment patterns;
White2014.lean makes the consistency claim formal.
Whether a factive presupposition has modal content.
Following @cite{williams-2026}: with finite CPs and gerunds, the presupposition is non-modal (event truth); with plain infinitives, it is modal (Williams' (5a): "John was supposed to V").
- nonModal : PresupContent
Directly presupposes complement truth (Williams (4a)).
- modal : PresupContent
Presupposes a modalized proposition: covert Mod over the embedded VP, yielding an obligation reading (Williams (5a)).
Instances For
Equations
- Phenomena.Presupposition.Studies.Williams2026.instDecidableEqPresupContent x✝ y✝ = if h : x✝.ctorIdx = y✝.ctorIdx then isTrue ⋯ else isFalse ⋯
Equations
- One or more equations did not get rendered due to their size.
Instances For
An empirical judgment about forget's presupposition in a given complement frame.
Every entry has a presupposition (uniform factivity per Williams).
The content field records the modal/non-modal split; the
implicativeNegative field records the orthogonal Karttunen
entailment that Williams (fn 1) brackets.
- frame : Semantics.Lexical.ComplementType
Complement frame being tested.
- sentence : String
Example sentence (Williams uses John throughout §1–§3.1).
- presupParaphrase : String
Paraphrase of what is presupposed.
- content : PresupContent
Modal vs. non-modal presupposition content.
- implicativeNegative : Bool
Negative implicative entailment (forgot to V → ¬V; Karttunen 1971a). Williams (fn 1) brackets this; recorded here for cross-paper consistency with @cite{karttunen-1971}.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Cognition reading — finite CP #
Williams (1)/(4): "John forgot that he stopped by the flower shop" presupposes that John stopped. This is the canonical factive reading recognized since @cite{kiparsky-kiparsky-1970}.
Equations
- One or more equations did not get rendered due to their size.
Instances For
English PRO-ing gerund #
Williams (7)/(9), §3.1.1: "John forgot stopping by the flower shop" patterns with the finite-CP case (non-modal presupposition) — a critical datum against the Modalized Complement Analysis, which would predict a modal presupposition for any non-finite complement. The gerund is non-finite but not modalized. Test: continuation "...but he didn't stop by the flower shop" is judged infelicitous, indicating the lower-event presupposition is present.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Psych-action reading — plain infinitive #
Williams (2)/(5), §1: "John forgot to stop by the flower shop" does not presuppose John stopped (and per the Karttunen tradition, entails he did not). What it presupposes (Williams (5a)) is the modalized proposition "John was supposed to stop by the flower shop". Williams' analysis (§4–§5): the plain infinitive's forward-oriented temporal profile violates the pre-existence presupposition, triggering covert Mod insertion that lets the obligation/plan be the target of the memory.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The English data points covered in Williams §1 and §3.1.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Williams' modal-presupposition split tracks Karttunen's implicative-negative entailment: the plain-infinitive frame is the locus of both. This is the cross-paper coherence claim — the two asymmetries identified in Williams' (1)/(2) (presupposition flip and implicative entailment flip) co-vary in the English data, even though Williams' analysis only addresses the presupposition side.