Browse Prior Art Database

Systemic Grammar and Functional Unification Grammar

IP.com Disclosure Number: IPCOM000128669D
Original Publication Date: 1987-Dec-31
Included in the Prior Art Database: 2005-Sep-16
Document File: 17 page(s) / 57K

Publishing Venue

Software Patent Institute

Related People

Robert Kasper: AUTHOR [+3]

Abstract

As this paper is being presented in the context of a workshop on systemic linguistics, it will assume some familiarity with systemic grammar. No familiarity with computer science or Functional Unification Grammar will be assumed. This section provides a historical perspective to what follows. Functional Unification Grammar (FUG) was first formulated by Kay (1979) as a creative method for experimentation in computational linguistics. It is a framework that represents a grammar in a form that is simultaneously readable by linguists and suitable for use by computer programs that generate or analyze text. Unlike many previous grammatical formalisms, which required a different representation for analysis than for generation, FUG is intended to be neutral in this respect.' In earlier work, FUG was also called Functional Grammar or Unification Grammar. The notation of FUG looks very different from that of Systemic Functional Gram-mar (SFG); but if we look beyond the differences of format, we find that FUG and SFG make many of the same assumptions about language and grammar. The similarities are due, in part, to the fact that when Martin Kay (1979) formulated FUG he was responding to many of Michael Halliday's ideas. Both FUG and SFG organize a grammatical description around feature choices, stress-ing the paradigmatic aspects of language structure. Structural realizations are stated as consequences of these feature choices. Both represent the function of constituents explicitly in addition to the position of constituents in a sequence. Both provide a means for multi-functional description, making possible the expression of textual and interpersonal roles in addition to ideational roles. Given these similarities, I was led to consider FUG as a promising tool for constructing an analysis program based on systemic grammar.2 A more careful investigation has shown that it is possible to represent most of a systemic grammar in FUG without any significant loss of clarity. Some natural extensions to the FUG framework handle the residue (see Section 4.3). Such rerepresentation of a grammar is motivated by Halliday's idea (1964) that the form of notation used to express a grammar should suit the task for which it is to be used. FUG notation is especially well suited to the task of computational processing. If we are able to represent SFG in FUG, then we have at our disposal the computational tools that have already been developed for FUG.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 6% of the total text.

Page 1 of 17

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

Systemic Grammar and Functional Unification Grammar

Robert Kasper USC / Information Sciences Institute

1 Background

As this paper is being presented in the context of a workshop on systemic linguistics, it will assume some familiarity with systemic grammar. No familiarity with computer science or Functional Unification Grammar will be assumed. This section provides a historical perspective to what follows. Functional Unification Grammar (FUG) was first formulated by Kay (1979) as a creative method for experimentation in computational linguistics. It is a framework that represents a grammar in a form that is simultaneously readable by linguists and suitable for use by computer programs that generate or analyze text. Unlike many previous grammatical formalisms, which required a different representation for analysis than for generation, FUG is intended to be neutral in this respect.' In earlier work, FUG was also called Functional Grammar or Unification Grammar. The notation of FUG looks very different from that of Systemic Functional Gram-mar (SFG); but if we look beyond the differences of format, we find that FUG and SFG make many of the same assumptions about language and grammar. The similarities are due, in part, to the fact that when Martin Kay (1979) formulated FUG he was responding to many of Michael Halliday's ideas. Both FUG and SFG organize a grammatical description around feature choices, stress-ing the paradigmatic aspects of language structure. Structural realizations are stated as consequences of these feature choices. Both represent the function of constituents explicitly in addition to the position of constituents in a sequence. Both provide a means for multi-functional description, making possible the expression of textual and interpersonal roles in addition to ideational roles. Given these similarities, I was led to consider FUG as a promising tool for constructing an analysis program based on systemic grammar.2 A more careful investigation has shown that it is possible to represent most of a systemic grammar in FUG without any significant loss of clarity. Some natural extensions to the FUG framework handle the residue (see Section 4.3). Such rerepresentation of a grammar is motivated by Halliday's idea (1964) that the form of notation used to express a grammar should suit the task for which it is to be used. FUG notation is especially well suited to the task of computational processing. If we are able to represent SFG in FUG, then we have at our disposal the computational tools that have already been developed for FUG.

'Any modifications to the generation grammar that are needed for the computational analysis program should be performed automatically by the program, as explained in (Kay, 1985). Bill Mann originally suggested that PUG might be an appropriate tool for this purpose.

The grammar being used for the study is Nigel (Mann & Matthiessen, 1985...