Head-driven phrase structure grammar
Citations
9,295 citations
3,345 citations
3,077 citations
2,835 citations
Cites background from "Head-driven phrase structure gramma..."
...Infrastructure for grammar development has a long history in unification-based (or constraint-based) grammar frameworks, from DCG (Pereira and Warren, 1980) to HPSG (Pollard and Sag, 1994)....
[...]
2,222 citations
Cites methods from "Head-driven phrase structure gramma..."
...Instead, the interactive alignment model is compatible with constraintbased grammar approaches in which syntax, semantics, and phonology form separate but equal parts of a multidimensional sign (Gazdar et al. 1985; Kaplan & Bresnan 1982; Pollard & Sag 1994)....
[...]
References
1,908 citations
1,856 citations
"Head-driven phrase structure gramma..." refers background or methods in this paper
...,As in the tradition of Gazdar et al. (1985), the analyses are detailed and made precise, so that a fair evaluation is possible....
[...]
...,As in the tradition of Gazdar et al. (1985), the analyses are detailed and made precise, so that a fair evaluation is possible. Moreover, the formalism used is similar enough to by-now traditional NLP grammar formalisms (in particular feature-based grammars) that readers without prior knowledge of HPSG implementations can have a reasonable idea of how to implement the analyses. As mentioned, the book carefully compares P&S's proposals to standard analyses within Principles and Parameters. Unfortunately, there is much less explicit comparison with work done within frameworks intellectually closer to HPSG, such as lexical-functional grammar (LFG) (Bresnan 1982). Finally, although the book focuses mainly on English syntax, P&S attempt to provide a cross-linguistic perspective in several chapters, in particular when dealing with agreement and relative clauses. To the novice reader, HPSG is a constraint-based grammatical formalism that belongs to the growing family of frameworks using feature structures as their basic data structure. In contrast to other feature-based frameworks, such as PATR-II (Shieber 1986, Shieber et al. 1983), LFG, or GPSG, HPSG does not rely on a context-free backbone; constituency is only one of the attributes of the linguistic object par excellence, the linguistic sign. It is on a par with other syntactic and semantic attributes. Moreover, HPSG is characterized by a systematic use of typing of feature structures (not unlike the use of templates in PATR-II). It is similar to DATR (Evans and Gazdar 1989a, 1989b) in this respect. HPSG uses a multiple-inheritance scheme over those types to cross-classify linguistic objects. Although they leave the question open, P&S de facto use a strict inheritance scheme in their analyses, rather than the default inheritance sometimes used in similar approaches (see Briscoe, de Paiva, and Copestake 1993). Overall, then, the formalism that HPSG uses, which owes a lot to King (1989), is very close to that discussed in detail in Carpenter (1992), for example. But for typing it uses a logic similar to that developed by Kasper and Rounds (1986) or Keller (1993). Typing in HPSG is used to factor out shared properties of linguistic objects, be they words, phrases, or anything else, into appropriate classes....
[...]
...Technically, the modeling of unbounded dependencies in HPSG (leaving aside the reformulation in chapter 9, for now) is similar to the classic SLASH percolation method used by Gazdar et al. (1985). There are three major differences between the two treatments....
[...]
...,As in the tradition of Gazdar et al. (1985), the analyses are detailed and made precise, so that a fair evaluation is possible. Moreover, the formalism used is similar enough to by-now traditional NLP grammar formalisms (in particular feature-based grammars) that readers without prior knowledge of HPSG implementations can have a reasonable idea of how to implement the analyses. As mentioned, the book carefully compares P&S's proposals to standard analyses within Principles and Parameters. Unfortunately, there is much less explicit comparison with work done within frameworks intellectually closer to HPSG, such as lexical-functional grammar (LFG) (Bresnan 1982). Finally, although the book focuses mainly on English syntax, P&S attempt to provide a cross-linguistic perspective in several chapters, in particular when dealing with agreement and relative clauses. To the novice reader, HPSG is a constraint-based grammatical formalism that belongs to the growing family of frameworks using feature structures as their basic data structure. In contrast to other feature-based frameworks, such as PATR-II (Shieber 1986, Shieber et al. 1983), LFG, or GPSG, HPSG does not rely on a context-free backbone; constituency is only one of the attributes of the linguistic object par excellence, the linguistic sign. It is on a par with other syntactic and semantic attributes. Moreover, HPSG is characterized by a systematic use of typing of feature structures (not unlike the use of templates in PATR-II). It is similar to DATR (Evans and Gazdar 1989a, 1989b) in this respect. HPSG uses a multiple-inheritance scheme over those types to cross-classify linguistic objects. Although they leave the question open, P&S de facto use a strict inheritance scheme in their analyses, rather than the default inheritance sometimes used in similar approaches (see Briscoe, de Paiva, and Copestake 1993). Overall, then, the formalism that HPSG uses, which owes a lot to King (1989), is very close to that discussed in detail in Carpenter (1992), for example....
[...]
...Technically, the modeling of unbounded dependencies in HPSG (leaving aside the reformulation in chapter 9, for now) is similar to the classic SLASH percolation method used by Gazdar et al. (1985)....
[...]
902 citations
849 citations