Biases in use of positive and negative words across twenty natural languages

Abstract

Positive events are more common (more tokens), but negative events are more differentiated (more types). These observations and asymmetries about the world are consistent with a number of features or biases favouring positive adjectives that have been shown for English. Compared to their opposites, positive adjectives in English are more likely to be unmarked, negated into their opposite, define the entire negative to positive dimension, and occur first in conjunctions with their negative opposite. In this paper we document that these biases have considerable generality, appearing in all or almost all of 20 natural languages. The greater differentiation of negative states is illustrated here by the demonstration that five common nouns describing negative states in English (disgust, risk, sympathy, accident, murderer) have equivalents in most or all of the 20 languages surveyed, but the opposite of these nouns is not lexicalised in most of the 20 languages.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 107 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.