New Word Grammar at work: How NWG structures the language

Organiser and chair: SUGAYAMA Kensei

This workshop is planned so as to introduce the audience into (New) Word Grammar (NWG) and to show that it is applicable to languages other than English, along with critique from different theoretical orientations. Word Grammar (WG) is a theory of language developed by Richard Hudson from the mid eighties. Hudson has recently produced a new version of WG, renamed here as NWG. The papers offered reflect the diversity of NWG, in the areas of language to which it has been applied and in the assumptions made and conclusions drawn. They also, however, reflect a high degree of convergence, particularly in the requirement for a more finely grained inventory of grammatical relations than the traditional one (subject, object, agent, theme, etc.).

WG and NWG: A contrastive introduction

SUGAYAMA Kensei

The organiser and chair, as the leader of the workshop, first speaks and summarises the NWG theory of language structure under the following headings: 1. A brief overview of the theory; 2. The cognitive network, 2.1 Language as part of a general network, 2.2 Labelled links; 3. Default inheritance; 4. The language network; 5. Syntax & Unrealised lexemes; and 6. Semantics.

Ideas and questions that will be discussed include the relationship between (N)WG grammatical structures and constructions (understood in terms of Construction Grammar, and more generally), the relationship between syntax and semantics and the relationship between language and usage. The first and second of these promise to allow considerable agreement, the third rather more controversy.

How NWG explains agglutination in Turkish: A case study

YOSHIMURA Taiki

In this presentation, I will show how NWG explains the word structure in Turkish, a language well-known for its agglutinative character. Although it seems quite difficult to draw clear-cut distinctions between derivation, inflection and cliticization, this does not raise any problem because these distinctions do not play any significant role in NWG; agglutination is explained by rich networks among concepts dealing with morphology, together with syntactic and semantic dependency networks. Cliticization in particular is easy to explain by combining morphological and syntactic networks. This type of analysis is indeed applicable to the interrogative clitic and the pronominal clitic, whose syntactic relations to the predicate resist the normal head-final tendency.

Dependency structure and phrase structure

MAEKAWA Takafumi

Most syntactic theories have adopted phrase structure as the basis for sentence structure. In a family of approaches known as ‘dependency grammars’, however, all relationships are word-based, and phrases do not have any syntactic status. It has been widely believed that dependency structure and phrase structure are merely notational variants (Gaifman 1965; Robinson 1970). In this paper, I will first argue that there are real differences between the two. I will then compare the Word Grammar dependency model with one of phrase-based frameworks, Head-Driven Phrase Structure Grammar. I will point out that there are some syntactic phenomena which require a phrase-based account. This suggests that phrase structure, rather than dependency structure, should be adopted as the basis for syntactic representation.

Closed-class semantics in NWG: A cognitive-based semantic network model

NAKANISHI Mitsukazu

This paper aims to show how recent research on closed-class semantics in Cognitive Grammar (CG) can contribute to the semantic notations in New Word Grammar (NWG). As for semantics, both CG and WG take the position that linguistic semantics is encyclopaedic in nature. However, there are fundamental differences in the semantic analysis between the two. While WG, by defining every concept in terms of its relationship to other concepts, gives a more formal description of linguistic semantics, CG provides a rather elaborated description of closed-class semantics. This paper demonstrates that a CG semantic description of closed-class items extends to formal semantic notations in WG. I take up the polysemous network of the be-ing construction in English as a case study.