There Is No Logical Negation Here, But There Are Alternatives: Modeling Conversational Negation with Distributional Semantics

Abstract

Logical negation is a challenge for distributional semantics, because predicates and their negations tend to occur in very similar contexts, and consequently their distributional vectors are very similar. Indeed, it is not even clear what properties a “negated” distributional vector should possess. However, when linguistic negation is considered in its actual discourse usage, it often performs a role that is quite different from straightforward logical negation. If someone states, in the middle of a conversation, that “This is not a dog,” the negation strongly suggests a restricted set of alternative predicates that might hold true of the object being talked about. In particular, other canids and middle-sized mammals are plausible alternatives, birds are less likely, skyscrapers and other large buildings virtually impossible. Conversational negation acts like a graded similarity function, of the sort that distributional semantics might be good at capturing. In this article, we introduce a large data set of alternative plausibility ratings for conversationally negated nominal predicates, and we show that simple similarity in distributional semantic space provides an excellent fit to subject data. On the one hand, this fills a gap in the literature on conversational negation, proposing distributional semantics as the right tool to make explicit predictions about potential alternatives of negated predicates. On the other hand, the results suggest that negation, when addressed from a broader pragmatic perspective, far from being a nuisance, is an ideal application domain for distributional semantic methods.

Learning to Negate Adjectives with Bilinear Models

“…It is natural to finish the statement That's not a table, it's a ... with desk or chair, but not pickle. Kruszewski et al (2016) find that near-est neighbors in a vector space are a good approximation for human judgements about alternatives. We hypothesize that a set of alternatives can stand in for the semantic domain.…”

Linguistically-Informed Specificity and Semantic Plausibility for Dialogue Generation

“…Researchers have noticed that distributional semantics largely fail to capture semantic plausibility, especially in terms of discrete properties (e.g., negation) (Kruszewski et al, 2016) and physical properties . Kruszewski et al (2016) created a dataset building on synthetically generated sentences for negation plausibility.…”

“…Average scores are reported. In addition, we evaluate plausibility by asking judges whether they think the given response sentence without the prompt can reasonably be uttered, following instructions from Kruszewski et al (2016). The percentage of plausible ratings are reported.…”

A proof-theoretic approach to scope ambiguity in compositional vector space models

“…As an example, allowing a non-linearity in lexical lambda terms or as a syntactic mechanism means the copying of material which is not possible with 1 Though there is work on simulating negation in a tensor based setting [7], it is not clear what negation really means in a distributional setting. For instance, an alternative view is to treat distributional negation as conversational [11]. 2 For instance, the composition and type-raising combinators one finds in Combinatorial Categorial Grammar [21] don't easily translate into a standard category, and the Displacement Calculus of [18] subsumes it's structural rules in the rules of the system [22].…”