Functional Unification Grammar Revisited

In this paper, we show that one benefit of FUG, the ability to state global conslralnts on choice separately from syntactic rules, is difficult in generation systems based on augmented context free grammars (e.g., Def'mite Clause Cn'anmm~). They require that such constraints be expressed locally as part of syntactic rules and therefore, duplicated in the grammar. Finally, we discuss a reimplementation of lUg that achieves the similar levels of efficiency as Rubinoff's adaptation of MUMBLE, a detcrministc language generator. 1 I n t r o d u c t i o n Inefficiency of functional unification grammar (FUG, [5]) has prompted some effort to show that the same benefits offered by FUG can be achieved in other formalisms more efficiently [3; 14; 15; 16]. In this paper, we show that one benefit of FUG, the ability to conciselyl state global constraints on choice in generation, is difficult in other formalhms in which we have written generation systems. In particular, we show that a global constraint can be stated separately from syntactic rules in FUG, while in generation systems based on augmented context free ~g~nunars (e.g., Definite Clause Cn'amma~ (DCG, [13])) such consWaints must be expressed locally as part of syntactic rules and the~=for¢, duplicated in the grammar. Finally, we discuss a reimplementation of lUG in TAILOR [11; 12] that achieves the si.m/l~r leveLs of efficiency as Rubinoff's adaptation [16] of MUMBLE [7], a deterministc language generator. 1.1 Sta tement o f Const ra in ts Language generation can be viewed primarily as a problem of choice, requiring decisions about which syntactic structures best express intent. As a result, much research in language generanon has focused on identi~ing conswaints on choice, and it is important to be able to represent these constraints clearly and efficiently. In this paper, we compare the representation of constraints in FUG with their repn:sentation in a DCG generation system [3]. We are interested in representing functional constraints on syntactic sWacture where syntax does not fully restrict expression; that is, conswaints other than those coming from syntax. We look at the representation of two specific constraints on syntactic choice: focus of attention on the choice of sentence voice and focus of attention on the choice of simple versus complex sentences. We claim that, in a lUG, these constraints can be stated separately from rules dictating syntactic structure, thus leading to simplicity of the granunar since the constraints only need to be stated once. This is possible in FUG because of unification and the ability to build constituent structure in the grammar. In contrast, in a DCG, constraints must be stated as part of the individual grammar rules, resulting in duplication of a constraint for each syntactic rule to which it applies. 1.2 Passive/Active Cons t ra in t Focus of attention can determine whether the passive or active voice should be used in a sentence [8]. The constraint dictates that focused information should appear as surface subject in the sentence. In FUG, this can be represented by one pattern indicating that focus should occur f'u'st in the sentence as shown in Figu~ 1. This panern would occur in the sentence category of the grammar, since focus is a sentence constituent. This constraint is represented as part of an alternative so that other syntactic constraints can override it (e.g., if the goal were in focus but the verb could not be pmsivized, ~ constraint would not apply and an active sentence would be generated). The structure of active or passive would be indicated in the verb group as shown in Figure 2.1 The correct choice of active or passive is made through unification of the patterns: active voice is selected if the focus is on the protagonist (focus unifies with pro:) and passive if focus is on the goal or beneficiary Orocus unifies with goal or beheld. This representation has two desirable properties: the constraint can be stated simply and the construction of the resulting choice b expr=ssed separately from the constraint. ( a l t ( ( p a t t e r n ( f o c u s . . . ) ) ) ) Figure 1: Constraint on Passive/Active in FUG In the DCG, the unification of argument variables means a single rule can state that focus should occur first in the sentence. However, the rules specifying construction of the passive and active verb phrases must now depend on which role (protagonist, goal, or beneficiary) is in focus. This requires three separate rules, one of which will be chosen depending on which of the three other case roles is the same as the value for focus. The DCG v..presentation thus mixes information from the conswaint, focus of attention, with the passive/active construction, duplicating it over three tThis figure shows only the m'dm, of comtitmmu foe active and passive voice m~l does no¢ include odwr details of the co~au'ucdon.