Semantics.Lexical.MeaningComponents #
@cite{levin-1993} @cite{beavers-koontz-garboden-2020}
Binary meaning-component features that define @cite{levin-1993} verb classes
diagnostically (via diathesis alternation participation), with the fuse
operator for componentwise composition.
Provenance #
Moved from Core/Lexical/VerbClass.lean in the cleanup that dissolved
Core/Lexical/. Lives at Theories/Semantics/Lexical/ (sibling of
LevinTheory.lean, LevinClassProfiles.lean, LevinClass.lean,
DiathesisAlternation.lean) because it encodes Levin's specific
diagnostic apparatus (CoS / contact / motion / causation as the 4
canonical features) — paper-anchored framework content, not consensus
substrate.
Framework commitment #
The 4-feature decomposition is @cite{levin-1993}'s diagnostic apparatus.
@cite{beavers-koontz-garboden-2020} argue these are SURFACE behaviors,
not root-level entailments — root-level structural features live in
Theories/Semantics/Lexical/Roots/RootFeatures.lean::RootEntailments
(state/manner/result/cause). The two carve-ups are NOT equivalent:
e.g., causation here is what diathesis alternations diagnose, while
B&KG's cause is a root entailment.
The instrumentSpec and mannerSpec features supplement the 4-feature
core for finer-grained subclass distinctions in Part II.
Note on fuse #
fuse a b is componentwise OR — the formaliser's design choice for
modeling how a construction augments a verb's inherent semantics.
Originally attributed to @cite{goldberg-1995} in source comments, but
Goldberg's actual constructional unification is not componentwise
boolean OR (it involves semantic frame unification with role-fusion
constraints, far more structured than disjunctive feature OR). The
substrate's fuse is a useful approximation but should not be cited
as Goldberg's operation directly.
Alternative frameworks not formalized at parallel substrate granularity #
The Levin-style alternation-diagnosed feature decomposition competes with other lexical-semantic frameworks worth formalizing as siblings:
- Generative Lexicon (@cite{pustejovsky-1995}): qualia structure (formal/constitutive/telic/agentive) as the primitive decomposition.
- Frame semantics (@cite{fillmore-1982}, @cite{fillmore-kay-oconnor-1988}): semantic frames as primitive, alternations as surface reflexes.
- Lexical Conceptual Structure (@cite{jackendoff-1996}): primitive predicates GO/STAY/CAUSE compose into LCS templates.
- Configurational lexical semantics (@cite{hale-keyser-1987}): verb meaning derives from syntactic configuration, not feature decomposition.
Binary meaning components that define @cite{levin-1993} verb classes.
These describe surface verb behavior, not root-level entailments.
@cite{beavers-koontz-garboden-2020} argue that surface CoS and causation
can come from either the template or the root; see RootEntailments
in Theories/Semantics/Lexical/Roots/RootFeatures.lean for the
root-level decomposition.
Diagnosed by participation in diathesis alternations:
changeOfState: middle alternation, causative/inchoative alternationcontact: body-part possessor ascension alternationmotion: conative alternation (with contact)causation: causative/inchoative alternation (with changeOfState)
The four canonical classes from Levin's Introduction:
- break = [+CoS, −contact, −motion, +causation]
- cut = [+CoS, +contact, +motion, +causation]
- hit = [−CoS, +contact, +motion, −causation]
- touch = [−CoS, +contact, −motion, −causation]
Additional binary features (from class descriptions in Part II):
instrumentSpec: verb specifies instrument/means (cut vs. break)mannerSpec: verb specifies manner of action
UNVERIFIED: Levin Part II page references for instrumentSpec/mannerSpec cited from memory.
- changeOfState : Bool
- contact : Bool
- motion : Bool
- causation : Bool
- instrumentSpec : Bool
- mannerSpec : Bool
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- Semantics.Lexical.MeaningComponents.break_ = { changeOfState := true, contact := false, motion := false, causation := true }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.cut = { changeOfState := true, contact := true, motion := true, causation := true, instrumentSpec := true }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.hit = { changeOfState := false, contact := true, motion := true, causation := false }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.touch = { changeOfState := false, contact := true, motion := false, causation := false }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.destroy = { changeOfState := true, contact := false, motion := false, causation := true }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.bend = { changeOfState := true, contact := false, motion := false, causation := true }
Instances For
Equations
- Semantics.Lexical.MeaningComponents.none = { changeOfState := false, contact := false, motion := false, causation := false }
Instances For
Componentwise OR. The formaliser's chosen approximation of construction-on-verb composition; not equivalent to Goldberg's actual constructional unification (see file docstring).
Equations
- One or more equations did not get rendered due to their size.