Minimalist Passive as Smuggling #
@cite{collins-2005} @cite{chomsky-2001} @cite{legate-2003}Connects the empirical passive data in Phenomena.ArgumentStructure.Passive
to the smuggling theory in Theories.Syntax.Minimalism.Movement.Smuggling
and the Voice/phase infrastructure in Theories.Syntax.Minimalism.Core.Voice.
The Analysis #
@cite{collins-2005} argues that passive involves smuggling: PartP (= [Part V DP]) is the complement of v. In passive, PartP moves to Spec-VoiceP, carrying the object past the external argument in Spec-vP. The object can then raise to Spec-TP without violating Relativized Minimality (@cite{rizzi-1990}).
Key structural claims: #
- UTAH: external argument is in Spec-vP in both active and passive
- Feature dissociation: active v checks Case AND assigns θ; passive v assigns θ only — Case-checking dissociates to Voice/by
- Phase status: passive v is NOT v* (not a phase head) because Case-checking is what makes v* a strong phase
- PartP licensing: PartP is licensed by either (a) have c-selecting it or (b) movement to Spec-VoiceP
Derivation #
Active: [TP DP_obj [T' T [VoiceP Voice [vP DP_ext [v' v [VP V DP_obj]]]]]]
— v checks Case on DP_obj, DP_ext gets θ from v, DP_ext raises to Spec-TP
Passive: [TP DP_obj [T' T [VoiceP [PartP Part V t_obj] [Voice' by [vP DP_ext [v' v t_PartP]]]]]]
— PartP smuggles DP_obj past DP_ext to Spec-VoiceP
— DP_obj extracts from PartP to Spec-TP (no RM violation: DP_ext no longer c-commands)
— Voice/by checks Case on DP_ext (feature dissociation)
Passive Voice permits smuggling (not a phase head). @cite{collins-2005}: "neither the moved PartP nor an unaccusative vP are strong phases."
Passive Voice and anticausative Voice share the smuggling property. Both have phaseHead = false. The structural parallelism between passive and unaccusative is Collins' central insight and extends to @cite{storment-2026}'s QI analysis.
Active (agentive) Voice blocks smuggling — v* is a strong phase head. In active transitive, the object gets Case from v directly; no smuggling is needed or possible.
The active/passive distinction is exactly a dissociation of θ and Case. Active: Voice assigns θ, does not check Case (v does). Passive: Voice does not assign θ (v does), checks Case (by). These are complementary: no Voice head both assigns θ AND checks Case.
Phase status correlates with Case-checking remaining on v. When v checks Case (active), v is a phase head (v*). When Case dissociates to Voice (passive), v is not a phase head. This is the structural reason passive allows smuggling.
The passive derivation: passive Voice with PartP licenses smuggling. PartP moves to Spec-VoiceP, carrying the object past Spec-vP.
Without PartP, smuggling cannot occur — there is nothing to move.
The smuggling derivation predicts the c-command asymmetries in
Passive.cCommandData: after smuggling, the object is in Spec-TP
(c-commanding everything), while the external argument is in
Spec-vP (c-commanded by the object). The predictions follow from
the structural positions, not from stipulation.
Binding (ex. 10a-b): Spec-TP c-commands into by-phrase ✓ NPI licensing (ex. 10c-d): Spec-TP licenses NPI in by-phrase ✓ Scope (ex. 10): universal in Spec-TP > existential in by-phrase ✓
Passive, QI, and anticausative all share the same structural precondition for smuggling: non-phase Voice. @cite{collins-2005} passive, @cite{storment-2026} QI, and standard unaccusatives are all instances of complement extraction enabled by defective (non-phase) v.