Martin, Rose and Nichols (2025): Burning facts: thick and thin causatives #

This file formalizes the division of English lexical causatives into thick and thin verbs. A thick causative such as burn, break or bury conveys a way of bringing the change about; a thin causative such as kill, destroy or change names the result state alone. Rose, Sievers and Nichols distinguish two concepts of causation, production, the transfer of a conserved physical quantity from cause to effect, and dependence, the counterfactual dependence of the effect on the cause. Production entails dependence, and only a physical entity produces, so an absence, a fact or a degree is at most a dependence cause. A thick causative used transitively in its physical sense conveys production and so rejects a subject with abstract reference (the lack of sunscreen burned her skin) that a thin causative accepts (the lack of water killed my plants). Martin, Rose and Nichols derive this production constraint from the competition between the lexical causative and periphrastic cause: the manner information makes production the salient reading, the lexical form takes the more specific meaning and leaves dependence to cause, the division of pragmatic labour that weak bidirectional optimality theory yields. The constraint lapses where the competition does, in the abstract sense and in the anticausative with a causer phrase, and it is the positive form of Wolff's directness constraint: a subject that is only a necessary condition for the transference cause (the flat tire cracked the axle) fails a thick verb but not a thin one.

A corpus survey of thirty-seven frequent causative verbs records whether each alternates, is thick, occurs with an adjectival strong resultative (break open), and occurs with an omission or quality subject in its concrete sense. The thick verbs are, except for bury, Embick's causative manner verbs, the result verbs found with strong resultatives, and the thin verbs found with strong resultatives (set, trigger, turn) lexicalize no result state of their own. Bury is thick through a result state that reveals the process producing it rather than through an event predicate.

Main definitions #

CauseConcepts — production and dependence over a domain of causal relata, with the postulates that production entails dependence and that only physical relata produce.
Reading, Reading.holds — the two readings of CAUSE and their truth conditions.
Form, tableau — the competition between the lexical causative, periphrastic cause and the anticausative with a causer phrase, as a bidirectional optimality tableau.
Sample — the thirty-seven verbs of the survey, Sample.entry their fragment entries, and the columns alternating, thick, strongResultative, resultless, omissionSubject.
causativeMannerVerbs, thickState — Embick's class and the thick verbs outside it.

Main results #

production_constraint — with production promoted, the lexical form conveys production and cause dependence; lexical_dependence_of_not_promoted, anticausative_readings — the dependence reading survives where nothing promotes production.
not_holds_of_not_physical — under the production constraint no subject with abstract reference satisfies the lexical causative; sunburn and axle are the paper's contrasts.
causativeMannerVerbs_subset_thick, thickState_eq, strongResultative_sdiff_thick — the correlation between thickness and strong resultatives.
thin_without_omission, thick_inter_omissionSubject — omission subjects in the survey.
alternating_iff_hasCausativeAlternation — the alternation column agrees with the fragment's diatheses.

Implementation notes #

The proportions the paper reports (twelve of thirteen thick verbs with a strong resultative, ten of thirteen thick and twelve of twenty-four thin verbs alternating) are left to prose; the theorems state the correlations with their named exceptions. Table 3 lists melt among the thick verbs found with an omission subject, though the discussion names only burn, lift and lock; the table is followed. The survey's annotators judged cool non-alternating, against the fragment, which gives it an intransitive with the patient as subject.

References #

[MRN25]
[RSN21] — the two concepts of causation and the omission contrasts.
[Emb09] — causative manner verbs and strong resultatives, as the paper reports them.
[Wol03] — the directness constraint, as the paper quotes it.
[Blu00], [Hor84] — weak bidirectional optimality and the division of labour.

Martin, Rose and Nichols (2025): Burning facts: thick and thin causatives #

Main definitions #

Main results #

Implementation notes #

References #

Two concepts of causation #

The competition between covert and overt CAUSE #

Witnesses #

The corpus survey #