Decomposing Manner Modifiers

Abstract

This thesis is concerned with the categorical identity and internal structure of manner modifiers (i.e., manner adverbials, MAs). Starting from the observation that MAs do not correspond to a uniform syntactic object, it is argued that their categorical identity is uniformly determined by functional structure. Building on the idea that the syntax of anaphoric dependencies and that of manner modification share some abstract formal properties, manner modifiers are shown to be subject to the so-called the IDI constraint (Inability to Distinguish Indistinguishables), which states that the computational system of human language (C-HL) cannot distinguish two predicates merged in the same local domain unless their linguistic environment allows them to be formally distinguishable (i.e., identifiable as different occurrences). This formal distinguishability can be achieved through licensing strategies such as protection, in which a bare predicate (e.g., quick) is transformed into a more complex form (e.g., quick-ly). Manner modifiers are thus analyzed as consisting of (i) a predicate and (ii) a protector — a functional element that ensures formal distinctness from the (verbal) predicate. To capture this systematically, it is proposed that protectors are uniformly functional heads (PFP hypothesis: Protectors as Functional Projections). Adopting the Exo-Skeletal Model, data from Modern Hebrew support a typology of protectors yielding the tripartition: (i) adverbial PPs (P⁰; locative adpositions), (ii) αPs (α⁰; bare adjectives), and (iii) DPs (D⁰; phrasal idioms). Under the approach developed in this thesis, the categorical identities of MAs become a reflex of the functional structure associated with the relevant protector.