Segment Grammar: a Formalism for Incremental Sentence Generation

Abstract

Incremental sentence generation imposes special constraints on the representation of the grammar and the design of the formulator (the module which is responsible for constructing the syntactic and morphological structure). In the model of natural speech production presented here, a formalism called Segment Grammar is used for the representation of linguistic knowledge. We give a definition of this formalism and present a formulator design which relies on it. Next, we present an object-oriented implementation of Segment Grammar. Finally, we compare Segment Grammar with other formalisms.

Aiming at shorter dependencies: the role of agreement morphology

“…As we discuss later in more detail, availability-based accounts predict a general preference to place phrases that are short before those that are long, independently of the basic word order or other typological properties of the language (Arnold, Wasow, Losongco, & Ginstrom, 2000;de Smedt, 1994;Easy First principle in MacDonald, 2013;Stallings & MacDonald, 2011;Stallings, MacDonald, & O'Seaghda, 1998;Wasow, 1997aWasow, , 1997b. In contrast, in Hawkins' Performance Theory of Order and Constituency model (henceforth, PTOC), word order preferences depend on phrasal length and head direction of the language in question (whether dependents follow or precede heads in phrases).…”

“…Although most research on availability-based production has focused on the effects of conceptual accessibility (Bock, 1982;Bock & Warren, 1985;Ferreira, 1996;Ferreira & Dell, 2000, among others), several researchers have proposed a general preference to place short phrases before long ones (Arnold et al, 2000;de Smedt, 1994;Easy First principle in MacDonald, 2013;Stallings et al, 1998;Stallings & MacDonald, 2011;Wasow, 1997aWasow, , 1997b. The idea is that sentences such as (5b) are preferred over sentences such as (5a) below, because shorter phrases (to Bill) should be generally easier for speakers to retrieve from memory than longer phrases (the book she had been searching for since last Christmas) and hence can be assembled earlier.…”

Default inheritance in an object-oriented representation of linguistic categories

“…Originally propose.d by Kempen (1987), it has been further worked out and implemented by De Smedt 8z. Kempen (1991) SG has been implemented in CommonORBTT by uniformly representing all important grammaz units-syntactic segments, syntactic categories (phrases, words) and syntactic features-as structured objects. Inheritance allows the graznmar to be extended easily by creating segments as specializations or combinations of other ones.…”

Filtering errors and repairing linguistic anomalies for spoken dialogue systems

“…It relies on a set of elementary trees (defined in the lexicon) which have at least one terminal symbol on its frontier, called the anchor. Trees can be combined through two simple operations : substitution 4 and furcation (de Smedt and Kempen, 1990). Those operations are theoretically equivalent to Tree Adjoining Grammar operations.…”