Documentation

Linglib.Phenomena.Causation.Studies.MartinRoseNichols2025

Thick vs Thin Causative Verb Data #

@cite{martin-rose-nichols-2025} @cite{embick-2009}

Corpus survey data from Table 3: 37 English causative verbs classified by four binary properties:

  1. alternating: Participates in the causative/anticausative alternation
  2. thick: Encodes manner of causing (subject restriction on abstract causes)
  3. ASR: Compatible with strong adjectival resultatives (break open)
  4. omissionSubjects: Compatible with omission/quality-denoting subjects

Key Findings (§4.3) #

A single verb entry from Table 3, extending a Fragment VerbEntry. The Levin class, verb form, root profile, etc. are all inherited from the Fragment entry — only the @cite{martin-rose-nichols-2025} annotations are new.

Instances For
    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      Verb form (convenience accessor).

      Equations
      Instances For

        Table 3 data (representative subset) #

        We include all 13 thick verbs and a representative set of thin verbs covering the key patterns. Numbers in comments refer to Table 3 rows.

        Equations
        Instances For
          Equations
          Instances For
            Equations
            Instances For
              Equations
              Instances For
                Equations
                Instances For
                  Equations
                  Instances For
                    Equations
                    Instances For
                      Equations
                      Instances For
                        Equations
                        Instances For
                          Equations
                          Instances For
                            Equations
                            • One or more equations did not get rendered due to their size.
                            Instances For
                              Equations
                              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
                                    Equations
                                    Instances For
                                      Equations
                                      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
                                            Equations
                                            Instances For
                                              Equations
                                              Instances For
                                                Equations
                                                Instances For
                                                  Equations
                                                  Instances For
                                                    Equations
                                                    Instances For
                                                      Equations
                                                      Instances For

                                                        Per-datum verification theorems #

                                                        Bridge to ThickThinClass #

                                                        Verify that the data entries' classifications match the theory.

                                                        Thick manner verbs are ASR-compatible per the theory.

                                                        Thick state verbs are NOT ASR-compatible per the theory.

                                                        Thin verbs are NOT ASR-compatible per the theory.

                                                        Bridge to @cite{levin-1993} classes #

                                                        The thick/thin distinction cross-cuts Levin classes: verbs in the same general domain (change of state, causation) can be thick or thin. The difference is whether the verb specifies manner of causing.

                                                        Break (thick) and destroy (thin) both have CoS + causation in their Levin meaning components. The thick/thin split is orthogonal to the basic meaning component profile.

                                                        Thick manner verbs belong to Levin classes that predict the causative alternation.

                                                        Cut (thick) is in a class that predicts conative and BPPA alternations. Unlike break, cut does NOT participate in causative/inchoative because instrument specification blocks the inchoative.

                                                        Destroy (thin) is also predicted to participate in causative alternation by its meaning components, but empirically it does not alternate. This shows the limits of meaning-component prediction.

                                                        Kill (thin, murder class) is predicted to participate in causative alternation but empirically does not alternate.

                                                        All ThickThin verb entries (for aggregate bridge theorems).

                                                        Equations
                                                        • One or more equations did not get rendered due to their size.
                                                        Instances For
                                                          theorem Phenomena.Causation.Compare.thick_mostly_alternate_bridge :
                                                          have thickVerbs := List.filter (fun (x : ThickThin.ThickThinEntry) => x.thick) ThickThin.allEntries; have altThick := List.filter (fun (x : ThickThin.ThickThinEntry) => x.alternating) thickVerbs; altThick.length * 100 / thickVerbs.length 70

                                                          Most thick verbs alternate (have both Voice variants).

                                                          Alternating thick verbs: the transitive form has agentive Voice, the anticausative has non-thematic Voice. Example: break.

                                                          • "John broke the vase" = Voice_AG + vDO + vCAUSE + vGO + vBE
                                                          • "The vase broke" = Voice_∅ + vCAUSE + vGO + vBE

                                                          Non-alternating thick verbs (cut) only have the agentive Voice form.