Particle verb constructions — small-clause analysis #
@cite{dendikken-1995} @cite{baker-1988}
@cite{dendikken-1995} analyses verbal particles as ergative small-clause
heads. The book's primary focus is complex particle constructions
(make John out a liar, ch. 2), with applications to triadic constructions
and Dative Shift (ch. 3-4) and to applicatives and transitive causatives
(ch. 5). All four phenomena instantiate a SC-in-SC structural template
[VP V [SC1 Spec [XP Prt [SC2 NP {AP/NP/PP/VP}]]]] (book p. 269).
This file formalises only the simplex case (John looked up the information) of ch. 2.4, and only the surface SC after NP-movement
to SpecSC. It does not capture:
The SC-in-SC structure for complex PVCs (the book's main contribution). Encoding nested small clauses cleanly is pending substrate work on
SmallClauseitself; currentstructure SmallClausehas no recursion-friendly shape.Overt P-to-V head incorporation. @cite{dendikken-1995} (§2.4.3) argues that English does not feature overt particle incorporation. Two arguments (book p. 89, citing Emonds 1993:243, fn. 27): (i) V+Prt sequences like
brush offlack the English compound stress pattern found on N'brush offand V'baby sit; (ii) inflection attaches to V alone, e.g.Are your friends pushin' the car?(Vpush+-in'participle), not to a putative [V+Prt] complex. Right-Hand Head Rule also forbids right-adjoining the particle to V. What den Dikken calls "verb-particle reanalysis" is an LF operation (cosuperscripting / Government Transparency Corollary, @cite{baker-1988}), not formation of an overt complex head. Encoding LF reanalysis requires substrate distinct fromformComplexLI.A weight-based derivation rule (
pronoun → split/heavy → continuous). This is @cite{kayne-1985}'s account, which @cite{dendikken-1995} (§2.4.5) explicitly refutes (citing Diesing & Jelinek 1993). den Dikken's own account of the pronoun ban derives from LF reanalysis interacting with pronominal cliticization, again pending the LF reanalysis substrate.
The empirical pattern (pronoun_split = good, pronoun_continuous = bad,
heavy_continuous = good, heavy_split = bad) is recorded in
ParticleVerbs.Data as theory-neutral data; explanations from any
particular account (den Dikken, Kayne, Construction Grammar, processing)
are property of the corresponding study file.
Simplex PVC: surface SC after NP-movement #
For a simplex PVC John looked the information up (the V-NP-Prt order),
the surface structure has the DP raised to SpecSC and the particle as
the SC head. Note this is the output of NP-movement — den Dikken's
D-structure has the DP as the complement of the (ergative) particle,
with SpecSC empty. The current SmallClause shape forces a subject
field, so we record the post-movement form.
Equations
- One or more equations did not get rendered due to their size.
Instances For
All PVC small clauses have predicate category P.
Connection to canonical SC shape #
The two trivial unfolding theorems above are convenience aliases. The
load-bearing structural fact about pvToSmallClause is its tree shape
once converted to a SyntacticObject via SmallClause.toSO — a 2-leaf
binary tree (subject + predicate). This is the shape consumed by any
file that wants to compose PVC SCs into larger structures (e.g.
embedUnderV for the full [VP V [SC DP Prt]] analysis, as used by
HaddicanEtAl2026.pvc_sc). Stated as rfl over the canonical shape so
that downstream files can rewrite without unfolding pvToSmallClause.
Any PVC small clause is a 2-leaf binary tree (subject + predicate).
The predCat field of pvToSmallClause agrees with the
predicate.headCat reading — the well-formedness invariant
consumed by SmallClause.toSO_isSmallClause. Discharged via
the outerCat_leaf substrate simp on the mkLeafPhon-built
predicate (which is structurally a leaf SO).
The PVC small clause satisfies IsSmallClause — the companion
predicate over raw SyntacticObjects. Discharges via the
well-formedness invariant + the SC-side round-trip lemma.