Quantifier Scoping in the SRI Core Language Engine

QUANTIFIER SCOPING IN THE SRI CORE LANGUAGE ENGINE Douglas B. Moran Artificial Intelligence Center SKI International 333 Ravenswood Avenue Menlo Park, California 94025, USA ABSTRACT An algorithm for generating the possible quanti- fier scopings for a sentence, in order of preference, is outlined. The scoping assigned to a quantifier is determined by its interactions with other quan- tifiers, modals, negation, and certain syntactic- constituent boundaries. When a potential scoping is logically equivalent to another, the less preferred one is discarded. The relative scoping preferences of the individ- ual quantifiers are not embedded in the algorithm, but are specified by a set of rules. Many of the rules presented here have appeared in the linguis- tics literature and have been used in various natu- ral language processing systems. However, the co- ordination of these rules and the resulting coverage represents a significant contribution. Because ex- perimental data on human quantifier-scoping pref- erences are still fragmentary, we chose to design a system in which the set of preference rules could be easily modified and expanded.. The algorithm described has been implemented in Prolog as part of a larger natural language pro- cessing system. Extensions of this algorithm are in progress. INTRODUCTION One of the major sources of ambiguity in sen- tences results from the different scopes that can be assigned to the various quantified noun phrases in the sentence. Part of the problem in determining the preferred scopings of quantifiers is the number of factors involved. For example, consider these three sentences John visited every house on a street. (1) John visited every house on a square. (2) John visited every patient in a private room. (3) Each of these sentences has two quantifier scop- ings: in one, "every' has wider scope over "a," and while in the other, "a" has the wider scope. However, the readings that most people obtain for these sentences are quite different. In (1), the reading in which "a" has wider scope is highly preferred; in (3), the reading in which "everf has wider scope is highly preferred; in (2), the reading with wide-scope "everf is preferred, but wide- scope "a" is also acceptable. A plausible expla- nation for the difference between (1) and (2) is that, since the typical house is located on a street but not on a square, the default preference rep- resented by (2) is overridden by a conversational maxim of quantity--if "~ streeff has narrow scope, "on a street" would contribute too little informa- tion to justify its presence. A plausible explana- tion for the difference between (2) and (3) is based on the relationship among the components. The reading of (3) in which "a" is given wider scope is improbable because the domain of quantification for "every" would then be the single patient in the selected room--an infelicitous use of "every, ~ whereas there is no similar problem in (2) because there are normally multiple houses on a square. Similarly, in John visited a person on every committee. (4) John visited …

Content truncated to 500 words