Using Subsumption Rather Than Equality in Functional Control

Abstract

In this paper I consider the typology of forward and backward control and raising structures, and argue that structure-sharing based on the relation of subsumption rather than equality provides a stronger linguistic basis for the typology (cf. Zaenen and Kaplan (2002, 2003)). I also consider the points of contact and divergence between structure-sharing and the ‘copy theory of movement’ that is prevalent in current minimalist syntax. 1. Functional Control in LFG Non-derivational syntactic theories such as LFG or HPSG traditionally analyze raising and control in terms of structure-sharing between (in the first instance) the subject of a predicate and the subject of a complement of that predicate, stated as properties of a lexical form. The same grammatical information simultaneously acts as the subject of the raising or control verb, and the subject within the infinitival complement. On the basis of this structure-sharing, such approaches would seem like prime candidates for extensions to insightful accounts of backward control and raising, phenomena that Polinsky and Potsdam (2002a, 2002b, 2006) (PP exactly the same grammatical information flows to each shared position (see (2)). (2) Equality: SUBJ = XCOMP SUBJ (information flows between both positions) Control and raising predicates in LFG are both subject to (2), and only differ in that the matrix subject is thematic in the case of control and non-thematic in the case of raising. The difference is represented in the lexical forms as shown in (3): thematic arguments appear within 〈 〉, non-thematic arguments appear outside. The f-structure in (3) is the same for both control and raising predicates. ∗I am grateful to Masha Polinsky and Eric Potsdam for help with examples and access to materials, and to the audience at LFG-06 in Konstanz for useful comments, in particular Ash Asudeh and Jonas Kuhn. Eric also provided a useful commentary on the original version of section 4.2. Parts of this material was presented at the Daegu Catholic University, Daegu, Korea, in July 2006, and I am grateful to Stan Dubinsky for comments on that occasion, some of which are incorporated here. (3) a. Control Predicate – thematic subject: (↑ PRED) = ‘c-pred 〈(↑ SUBJ), (↑ XCOMP)〉’ (↑ SUBJ) = (↑ XCOMP SUBJ) 

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 26 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.