09.07.2015 Views

Practical Information - Generative Linguistics in the Old World

Practical Information - Generative Linguistics in the Old World

Practical Information - Generative Linguistics in the Old World

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

– Follow this procedure recursively for v. Ignore variables already met.This constructs a spann<strong>in</strong>g tree over variables <strong>in</strong> <strong>the</strong> semantic expression (2a). This can beapplied at every <strong>in</strong>cremental step or <strong>the</strong> tree can be grown dynamically.(2) (a) e ⇒ (b)2x 1 ∃x 2∀x 3∃x 2e2x 1e⇒ (c)t x2∀x 3We can permit 2 > ∀ > ∃ by declar<strong>in</strong>g that sisters can scope over each o<strong>the</strong>r. (As this ishappen<strong>in</strong>g <strong>in</strong> covert syntax, asymmetry is not required for e.g. a l<strong>in</strong>earization algorithm. Suchasymmetries can be <strong>in</strong>troduced <strong>in</strong> future work as needed without los<strong>in</strong>g <strong>the</strong> substance of thisproposal.) We can <strong>the</strong>n permit ∃ > 2, ∀ and ∀ > ∃ > 2 us<strong>in</strong>g a type of movement over this verylimited structure (2b,c). We <strong>the</strong>n propose <strong>the</strong> follow<strong>in</strong>g def<strong>in</strong>ition of VST-movement to restrict<strong>the</strong> possible movements of quantified variables:(3) VST-move: a variable v and its quantifier are permitted to move to e iff e is <strong>the</strong> immediateparent of v, or is <strong>the</strong> ancestor of v only via traces (t). A new copy of e is created as a parentto <strong>the</strong> orig<strong>in</strong>al e and v.We fur<strong>the</strong>r stipulate that no event node may have only traces as children, to prevent <strong>in</strong>f<strong>in</strong>itemovement. This excludes ∀ > 2 > ∃, because ∀x 3 cannot move as long as ∃x 2 is its parent.Event variables thus become analogous to CP <strong>in</strong> accounts with overt movement, and restrictionson covert movement are now def<strong>in</strong>ed <strong>in</strong> terms of proximity to <strong>the</strong> event. This gives us a pr<strong>in</strong>cipledway to understand <strong>the</strong> difference between <strong>the</strong>se sentences from Kallmeyer and Romero.(4) a. A student said you met every professor. (<strong>in</strong>verse scope forbidden)b. A student wants (you) to meet every professor. (<strong>in</strong>verse scope permitted)In <strong>the</strong> first example, <strong>in</strong>verse scope is forbidden by <strong>the</strong> presence of two fully-fledged events.Every professor would be represented <strong>in</strong> <strong>the</strong> VST as <strong>the</strong> child of a lower event. However, <strong>in</strong><strong>the</strong> control verb condition, <strong>the</strong> lower verb is not a fully-fledged event with a separate conjunct.This <strong>in</strong>sight is reflected <strong>in</strong> structures used by Asudeh and Toivonen (2012) who nest <strong>the</strong> lowerverb’s event <strong>in</strong>side <strong>the</strong> representation of <strong>the</strong> control verb. Consequently, <strong>the</strong> movement of <strong>the</strong>lower event would not be blocked.For wh-<strong>in</strong>-situ languages such as Ch<strong>in</strong>ese, quantifiers are not permitted to take <strong>in</strong>verse scope,but wh-items are. We can treat this as a parametric variation by build<strong>in</strong>g <strong>the</strong> VST based onquestion-bound variables ra<strong>the</strong>r than quantifier-bound ones. In conclusion, our VST analysisallows a pr<strong>in</strong>cipled unification of covert movement analyses for movementless formalisms thatdo not readily accommodate May-style generalizations.A. Asudeh and I. Toivonen. (2012) Copy rais<strong>in</strong>g and perception.L. Champollion. (2011) Quantification and negation <strong>in</strong> event semantics.A. Joshi, L. Kallmeyer, and M. Romero. (2007) Flexible composition <strong>in</strong> LTAG: quantifier scopeand <strong>in</strong>verse l<strong>in</strong>k<strong>in</strong>g.A. Koller, J. Niehren, and S. Thater. (2003) Bridg<strong>in</strong>g <strong>the</strong> gap between underspecification formalisms:hole semantics as dom<strong>in</strong>ance constra<strong>in</strong>ts.R. May. (1985) Logical form: its structure and derivation.T. Parsons. (1990) Events <strong>in</strong> <strong>the</strong> semantics of English.M. Romero and L. Kallmeyer. (2005) Scope and situation b<strong>in</strong>d<strong>in</strong>g <strong>in</strong> LTAG us<strong>in</strong>g semanticunification.E. Ruys and Y. W<strong>in</strong>ter. (2010) Scope ambiguities <strong>in</strong> formal syntax and semantics.2∀x 3e∃x 2e2x 1et x2t x3

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!