From Insanely Jealous to Insanely Delicious: Computational Models for the Semantic Bleaching of English Intensifiers

Abstract

We introduce novel computational models for modeling semantic bleaching, a widespread category of change in which words become more abstract or lose elements of meaning, like the development of arrive from its earlier meaning 'become at shore.' We validate our methods on a widespread case of bleaching in English: de-adjectival adverbs that originate as manner adverbs (as in awfully behaved) and later become intensifying adverbs (as in awfully nice). Our methods formally quantify three reflexes of bleaching: decreasing similarity to the source meaning (e.g., awful), increasing similarity to a fully bleached prototype (e.g., very), and increasing productivity (e.g., the breadth of adjectives that an adverb modifies). We also test a new causal model and find evidence that bleaching is initially triggered in contexts such as conspicuously evident and insanely jealous, where an adverb premodifies a semantically similar adjective. These contexts provide a form of "bridging context" (Evans and Wilkins, 2000) that allow a manner adverb to be reinterpreted as an intensifying adverb similar to very.

Causal Effects of Linguistic Properties

“…Often, instead of the true linguistic property of interest we have access to a noisy measurement called the proxy label. Analysts usually infer these values from text with classifiers, lexicons, or topic models (Grimmer and Stewart, 2013;Lucas et al, 2015;Prabhakaran et al, 2016;Voigt et al, 2017;Luo et al, 2019;Lucy et al, 2020). When can we use such a proxy to recover the true effects of the latent linguistic property?…”

A Formidable Ability: Detecting Adjectival Extremeness with DSMs

“…Our approach is compatible with that of Luo et al (2019), but moreover explains why certain adverbs are more likely to modify similar adjec-…”

“…We begin with the novel hypothesis that an adjective's extremeness is a strong predictor of its future use in an intensifier adverbi.e., in phrases like staggeringly easy and monumentally wrong. Our second experiment then shows that our classification of extreme adjectives can improve over an existing computational approach (Luo et al, 2019) in a historical prediction task of emerging intensifier meanings.…”

“…Here we describe the two features we are investigating as predictive of the development of adverbs into intensifiers, our EXTREMENESS feature and the SIMADJMOD feature of Luo et al (2019). EXTREMENESS.…”

“…Rather than using the embedding of the adjectival basis of q for q t , as we did in Section 4 for data sparsity reasons, here we used the diachronic embeddings of the adverb q. In doing so, we mirror the operationalization of Luo et al (2019) for this feature exactly.…”