Head-Driven Phrase Structure Grammar

HPSG is a constraint-based, lexicalist approach to grammatical theory that seeks to model human languages as systems of constraints. Typed feature structures play a central role in this modeling. Some of the leading ideas of current work in HPSG are the following:

• Strict Lexicalism
  Word structure and phrase structure are governed by partly independent principles. Words and phrases are two kinds (subtypes) of sign.
• Concrete, surface-oriented structures
  `Abstract' structures (e.g. empty categories and functional projections) are avoided wherever possible, in favor of `minimal' grammatical structures.
• Geometric prediction
  The hierarchical organization of linguistic information plays a significant role in predicting the impossibility of certain kinds of linguistic phenomena.
• Locality of selection
  According the theory of valence articulated in Pollard and Sag (1994), lexical heads select only for the synsem objects (a kind of syntactico-semantic complex) of their complements, subjects, or specifiers. It follows that category selection, role assignment, case assignment, head agreement and semantic selection all obey a particular kind of locality determined by valence selection features. This is a kind of geometric prediction. Current work is exploring revisions of this architecture to accommodate limited selection of apparently non-local elements (e.g. subjects within saturated clauses and possessors within NPs).
• A distinction among types of agreement
  Agreement phenomena have been classified by Pollard and Sag (1994) as syntactic concord, anaphoric agreement, or pragmatic agreement. Their theory of indices predicts, inter alia, the absence of case agreement in anaphoric agreement. Recent work by Kathol, Wechsler, Zlatic, and Casillas has refined these proposals, extending the range of languages considerably and correcting mispredictions.
• Local encoding of unbounded dependencies
  Filler-gap phenomena and other long-distance dependencies are treated not via grammatical transformations, but rather in terms of certain feature specifications that are present throughout the `path' from filler to gap. This feature-based theory in essence predicts the existence of grammatical phenomena sensitive to such specifications, i.e. phenomena that occur only within the domain of unbounded dependency constructions. Precisely such phenomena have now been amply documented in such typologically diverse languages as Irish, Chamorro, Kikuyu and Thompson Salish (among others). This is another instance of geometric prediction.
• Lexical cross-classification
  Within HPSG, words are rich in information. Lexical information is not simply listed, however; rather it is organized in terms of multiple inheritance hierarchies and lexical rules that allow complex properties of words to be derived from the logic of the lexicon. Current research is developing extensions of hierarchical lexicons that allow lexical rules to be eliminated and linking patterns to be derived in a general fashion from semantic properties.
• Hierarchical cross-classification of grammatical constructions
  There are new proposals within HPSG to model constructions, as well as signs, in terms of feature structures. This allows constructions to be analyzed via multiple inheritance hierarchies. This in turn provides a way of modeling the fact that constructions cluster into groups with a `family resemblance' that corresponds to a constraint on a comon supertype. This strain of HPSG has thus coalesced with one conception of `Construction Grammar'.
• Obliqueness-based binding theory
  Generalizations about constraints on the binding of referentially dependent elements are stated in terms of relative obliqueness (o-command), rather than configurational superiority (c-command).
• Linearization theory
  Current work in HPSG is exploring modes of serialization which are not based on the model of traditional phrase structure grammar (where sentences are word strings defined derivatively in terms of phrase structure). This has implications for the treatment of discontinuous constituency, allowing even the introduction of levels of linear syntactic organization that are to some extent dissociated from the combinatorial relationships among the items serialized.