|
On Optionality: A Brief History and a Case Study| title | On Optionality: A Brief History and a Case Study |
|---|
| start_date | 2022/11/09 |
|---|
| schedule | 10h30-13h - Heure de Tucson |
|---|
| online | yes |
|---|
| visio | https://arizona.zoom.us/j/83059854210 |
|---|
| location_info | En ligne |
|---|
| summary | Optionality of transformational operations was a central part of syntactic theory in the early years of generativegrammar. Optionality vs. obligatoriness was coded on particular transformations, as an arbitrary specification. These specifications played a major role in the analyses in Chomsky (1955). By the mid-1960's, some concerns were raised about this, based on the interaction between syntax and semantics, and Katz and Postal (1964), followed by Chomsky (1965), proposed, instead, that many of the previously optional transformations were actually obligatory, triggered by a marker that may or may not appear in the initial phrase marker. Other concerns began to be raised in the mid 1970's, particularly with respect to learnability issues, as suggested by Chomsky and Lasnik (1977). Given a transformational component with n transformations, where optional vs. obligatory is an available specification for each, all else equal there are 2n possible grammars. If n is large, as in classical transformational grammar, this creates a huge quantitative learnability problem. Further, if optional and obligatory are equally available as transformational specifications, if the learner’s hypothesis for a particular transformation is that it is optional while in the target grammar it is obligatory, in the absence of negative data the incorrect hypothesis will not be corrected, a ‘qualitative’ learnability problem. Among the solutions proposed were that all transformations are optional, the resulting over-generation handled by filters, as in Chomsky and Lasnik (1977) and Lasnik and Kupin (1977). This became the standard view in the GB era. Another sort of solution to the qualitative problem was to allow both obligatory and optional transformations, but with obligatory as the unmarked case, the initial guess. This was advocated by Dell (1981) and Berwick (1982). I will examine these developments and then turn to an interesting instance of optionality and some possible approaches to it.
The process I will explore is one raising an ECM subject into the higher clause, as extensively argued for by Postal (1974) and later by Lasnik and Saito (1991). In both of those works, the raising process was argued to be obligatory. A variety of older and newer arguments will be surveyed. Then I will present a set of arguments, some of them from Lasnik (2001), that for some speakers, the raising process is optional. I will conclude with possible ways of capturing this optionality and of characterizing the difference between the two classes of speakers. |
|---|
| responsibles | Piatelli-Palmarini, Chomsky, Bever |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2022/09/09 12:34 UTC |
| |
|