Skip to content


LuisMorgadoCosta edited this page Aug 20, 2015 · 18 revisions

Schedule for LAD

The Linguistic Analysis Design event, to be held in connection with HPSG 2015 and DELPH-IN 2015 in Singapore, will consist of a set of three sessions devoted to collaborative design of linguistic analyses that are theoretically grounded, empirically justified, and amenable to implementation. In each of the sessions, a grammar developer will present a challenging linguistic phenomenon as yet lacking an adequate implementable analysis for that grammar, and invite participants, both theoreticians and grammar engineers, to work together on the design of candidate analyses. The location of this event is at Nanyang Technological University, in the HSS Conference Room on level 5 of the HSS building (HSS-05-57).

Sunday, 9 August
10:00-11:30 Session 1: Gapping in English -- One of the more challenging phenomena for an implementation in an HPSG grammar is gapping, found in English, Spanish, and other languages. In a sentence such as We gave Kim a book and Lee a laptop there is a notional coordination of two verb phrases, but with the omission of the head of the second VP. In a lexicalist framework, it is difficult to ensure that each of the conjoined headless sequences of complements and modifiers conforms to the subcategorization constraints of the overt head of the first conjunct phrase. It is similarly challenging to provide a compositional semantics of such sentences, where presumably one wants multiple occurrences of the predication introduced by the overt verb, one for each of the notional conjuncts in the gapping construction. The developers of the DELPH-IN grammars for English and Spanish have implemented one analysis of gapping, but at least for English, this analysis has fallen short both in coverage and in efficiency. Hence a better analysis is desired.
11:30-13:00 Lunch
13:00-14:30 Session 2: Reduplication in Chinese -- In the Chinese languages, the reduplicated forms are often used. The lexical types that can be reduplicated in the languages (e.g. Mandarin, Cantonese, etc.) include adjectives, verbs, and classifiers, and the reduplicated forms differ from each other syntactically and semantically. The reduplication of adjectives is an emphasis construction. For example, 干干净净 ganganjingjing ‘clean:AABB’ share almost the same interpretation as 很 干净 hen ganjing ‘very clean’ with an intensifier. If the adjective consists of two characters, there are four forms of reduplications, viz. AABB, ABAB, ABB, AAB. The reduplication of verbs involves a specific aspect: namely, tentative. Note that not all verbs can be necessarily reduplicated. This implies that the reduplication forms of verbs in Chinese needs to be treated like irregular inflectional forms in other languages. Classifiers can be reduplicated, and the reduplication imposes a semantic constraint on the head nouns. NPs with a reduplicated classifier may convey either universal quantifier meaning or distributive meaning. This is semantically similar to every in English (e.g. every student / every two days), but the reduplication of classifiers in Chinese has more to do with other grammatical structures. Similar issues arise with Indonesian.
14:30-15:00 Coffee break
15:00-16:30 Session 3: A-not-A construction in Chinese -- The Chinese languages, including Mandarin and Cantonese, employ two syntactic operations of yes-no questions. One uses the sentence final particle, such as吗ma in Mandarin (MA questions), and the other uses a sequential form of A-not-A (A-not-A questions), in which 不 bu "not" and 没 mei "not" are used as the negative operator in Mandarin. For instance, “张三 高 不 高 ?” conveys a meaning like “Is Zhangsan tall or not tall?” A-not-A questions in Chinese functions similarly to tag questions in other languages, but syntactically, semantically, and pragmatically all different from ordinary tag questions in other languages as well as MA questions in Chinese. There are several variations in the forms of A-not-A, and the implementation of the forms is apt to produce spurious overgeneration in terms of grammar engineering. In addition, it is possible to also repeat the object, A-O-not-A-O which very much complicates matters.
16:30-17:00 Concluding discussion