Eliminative Parsing with Graded Constraints

Natural language parsing is conceived to be a procedure of disambiguation, which successively reduces an initially totally ambiguous structural representation towards a single interpretation. Graded constraints are used as means to express wellformedness conditions of different strength and to decide which partial structures are locally least preferred and, hence, can be deleted. This approach facilitates a higher degree of robustness of the analysis, allows to introduce resource adaptivity into the parsing procedure, and exhibits a high potential for parallelization of the computation. 1 I n t r o d u c t i o n Usually parsing is understood as a constructive process, which builds structural descriptions out of elementary building blocks. Alternatively, parsing can be considered a procedure of disambiguation which starts from a totally ambiguous structural representation containing all possible interpretations of a given input utterance. A combinatorial explosion is avoided by keeping ambiguity strictly local. Although particular readings can be extracted from this structure at every time point during disambiguation they are not maintained explicitly, and are not immediately available. Ambiguity is reduced successively towards a single interpretation by deleting locally least preferred partial structural descriptions from the set of solutions. This reductionistic behavior coins the term eliminatire parsing. The criteria which the deletion decisions are based on are formulated as compatibility constraints, thus parsing is considered a constraint satisfaction problem (CSP). Eliminative parsing by itself shows some interesting advantages: Fail soft behavior: A rudimentary robustness can be achieved by using procedures that leave the last local possibility untouched. More elaborated procedures taken fl'om the field of partial constraint satisfaction (PCSP) allow for even greater robustness (cf. Section 3). tural possibilities are maintained explicitly, the amount of disambiguation already done and the amount of the remaining effort are immediately available. Therefore, eliminative approaches lend themselves to the active control of the procedures in order to fulfill external resource limRations. Parallel izat ion: Eliminative parsing holds a high potential for parallelization because ambiguity is represented locally and all decisions are based on local information. Unfortunately even for sublanguages of fairly modest size in many cases no complete disambiguation can be achieved (Harper et al., 1995). This is mainly due to the crisp nature of classical constraints that do not allow to express the different strength of grammatical conditions: A constraint can only allow or forbid a given structural configuration and all constraints are of equal importance. To overcome this disadvantage gradings can be added to the constraints. Grades indicate how serious one considers a specific constraint violation and allow to express a range of different types of conditions including preferences, defaults, and strict restrictions. Parsing, then, is modelled as a partial constraint satisfaction problem with scores (Tsang, 1993) which can almost always be disambiguated towards a single solution if only the grammar provides enough evidence, which means that the CSP is overconstrained in the classical sense because at least preferential constraints are violated by the solution. We will give a more detailed introduction to constraint parsing in Section 2 and to the extension to graded constraints in Section 3. Section 4 presents algorithms for the solution of the previously defined parsing problem and the linguistic modeling for constraint parsing is finally described in Section 5. 2 P a r s i n g as C o n s t r a i n t S a t i s f a c t i o n While eliminative approaches are quite customary for part-of-speech disambiguation (Padrb, 1996) and underspecified structural representations (Karlsson,