Announcement of a talk at Stanford this coming Tuesday, with abstract:
Investigating a Syntactic Analysis: Revisiting Raising-to-Subject
Scott Grimm, Stanford UniversityVerbs such as “seem’” or “appear’” allow an alternation between infinitival and finite sentential complements. An example typical of those considered in the literature is given in (1) (Davies and Dubinsky 2004).
(1a) Barnett seemed to understand the formula.
(1b) It seemed that Barnett understood the formula.The standard treatment in generative syntax of this alternation is the raising-to-subject analysis [Subject-to-Subject Raising, or SSR], which is introduced to account for the observation that the surface subject of (1a) is taken to be the subject of the embedded verb (‘understand’), a relation that is established in different ways in different syntactic theories. In the instantiation which gives this phenomenon its name, the subject of the embedded verb “raises” to become the subject of the matrix verb.
This analysis assumes that (1a) and (1b) are meaning-equivalent and claims that the choice of infinitival or finite sentential complementation is a syntactic matter. Much recent work in syntax has shown that comparable alternations are influenced by a variety of semantic, pragmatic, and/or usage factors. For instance, the dative alternation has been shown to be affected by syntactic weight and animacy (Bresnan et al. 2007), as well as information structure constraints (Snyder 2003).
Based on an examination of corpus data, I show that the two constructions in (1a) and (1b) can also be distinguished once a broader array of examples are systematically evaluated. The two constructions manifest significant distributional asymmetries in terms of (i) their information structure preferences and (ii) whether they align with statements based on direct, i.e. experiential, evidence, or indirect evidence, e.g. inferences…
(There’s another alternative — like SSR but with a finite complement to the verb and with a resumptive pronoun in the complement — now known as Copy Raising:
(1c) Barnett seemed like he understood the formula.
Extensive discussion of Copy Raising in English and Swedish in an article by Ash Asudeh and Ida Toivonen, soon to appear in Natural Language & Linguistic Theory.)
On the principle of Meaning Equivalence, here’s a discussion from a posting of mine on for and because:
I’ve argued, again and again, on Language Log and elsewhere, that Meaning Equivalence rarely holds for variant expressions. I’ve called this the program of “unfree variation”: you might think that variants are equivalent in meaning, just “different ways of saying the same thing”, and in many circumstances the choice between them seems not to matter, but in some circumstances subtle (occasionally glaring) differences emerge; that is, most variation is at root not without meaning consequences, even when the syntax doesn’t determine the choice.
I attribute the program of unfree variation to the writings of Dwight Bolinger, who repeatedly cautioned linguists to look for differences in meaning (in a broad sense of meaning) rather than assuming that alternatives are “in free variation”. The posting linked to above mentions three more lexical cases (relativizer which vs. that, connective but vs. however, determiner much vs. a lot), from a considerable collection, and there are many many case of alternative argument structures, including the dative alternation Grimm alludes to (NP Dative vs. PP Dative; some discussion here). And of course alternations on a larger scale, like Grimm’s SSR case and the other “raisings” in English:
Subject-to-Object Raising (SOR):
(2a) We believe Kim to be a spy. [SOR]
(2b) We believe that Kim is a spy. [unraised variant]Object-to-Subject Raising (OSR), a.k.a. Tough-Movement:
(3a) Sandy is tough to convince. [OSR]
(3b) It is tough to convince Sandy. [unraised variant]
(A big list of some alternations in English, prepared for my 2007 Linguistic Institute course on “Choosing a Variant”, is available here.)
Leave a Reply