News & Events

Professor Li Yafei

Professor of Linguistics University of Wisconsin-Madison

Date:  6 May, 13 May, 27 May & 29 May 2009
Time:  4:30 – 6:30pm (time of the seminar on 29 May 2009 revised: 4:00 - 6:00pm)
Venue: B7603 (Lift 3, 7/F, Blue Zone), Academic Building, CityU

Abstract

Week 1: Facts vs. Theories （Wednesday, 6 May 2009）   Demonstrating that the mainstream theory (MT) of syntax has been gradually but steadily losing the empirical foundation that it claims to rest on, this lecture defends the thesis that the postulation of abstract mechanisms/structures must be preceded by robust proof that a more straightforward analysis is inadequate for the data at issue. MT is shown to deviate from empiricism in two (only apparently) opposite directions: insufficiently motivated proliferation of theoretical apparatuses, and arbitrarily imposed simplicity/minimalism on theory formulation. Examples representing the two deviating directions will be given and critically evaluated. Where feasible, alternative analyses more friendly to directly observable data are also articulated.   Week 2: X’-structure, Coordination, etc. （Wednesday, 13 May 2009）   After a review of the existing syntactic theories on coordination, this talk presents new data about non-constituent conjuncts, namely conjuncts whose components don’t form a constituent, that display different degrees of acceptability. It is argued that no theory based on treating a conjunct as a constituent is capable of dealing with the facts, and that the one which explains both gradational acceptability and other properties of coordination reported in the literature is also one that takes the structure of coordination to be in compliance with surface data, i.e., if a conjunct appears not to be a constituent, then it is not a constituent. This conclusion undermines the foundation of several popular ideas such as the double-VP structure, binary branching, and that asymmetric c-command is the sole relation determining syntactic relations. The empirical validity of these ideas is further questioned with both facts long ignored and a reinterpretation of certain core data for the mainstream X’-template.   Week 3: On Marginal Data （Wednesday, 27 May 2009）   The varying degrees of acceptability in coordination are but one example of an important aspect of modern syntax: a large proportion of the data used for theorizing is marginal – somewhere between perfect and hopelessly bad. Marginality is inevitable given the fact that the principles-and-parameters (PnP) syntax aims to investigate the boundaries of the human language faculty. What is missing in PnP is any caution at all in obtaining and analyzing marginal data, yet another manifestation of the field’s deviation from empiricism. The lecture starts by introducing a psychological experiment, published in 2006, on motivated reasoning, a technical term referring to “reasoning biased to produce emotionally preferable conclusions”. The implications of this experiment to syntactic data collection, especially when the degree of acceptability is at issue, are discussed, with the conclusion that the current practice in PnP constitutes the worst scenario for a field claimed to be empirically grounded. Examples from published works are provided to illustrate the influence of motivated reasoning. The way in which MT determines marginality is also challenged, again with actual examples found in the literature. The lecture ends with suggestions on possible measures to improve the reliability of syntactic data.   Week 4: From ACD to LF （Friday, 29 May 2009: 4:00 - 6:00pm）   LF is a component of the PnP model which, by definition, hosts mechanisms not directly observable, of which LF-movement is the most widely assumed and actively utilized. Antecedent-contained deletion (ACD) has been viewed among the core evidence for LF-movement despite well-known problems. This lecture takes a closer look at the reported examples of ACD and divides them into three classes: simple ACD, interactive ACD and apparent ACD. With the help of simple ACD, it is shown that the force of the phenomenon as evidence for LF-movement is based on the misconception that verbal ellipsis always relies on copying the whole antecedent VP for proper interpretation. The problem presumably raised by ADC, i.e., infinite regress, is only the result of this misconception. An alternative account without any movement, covert or overt, is proposed which not only incurs no infinite regress but also solves certain tenacious ACD-related problems with ease. The new theory is evaluated through interactive ACD, namely ACD mingled with binding and scopal elements. Lastly, some elaborate examples claimed to be ACD and serving as further evidence for LF-movement are shown to be unrelated to ACD at all, further casting doubt on the phenomenon’s relevance to LF-movement. A syntactic theory without LF-movement is briefly discussed which suggests a much more restricted approach to theory formulation: if no movement is heard, then no movement happens.

Speaker