Variation in the contextuality of language: an empirical measure

TLDR: Both the data and theoretical model suggest that contextuality decreases when unambiguous understanding becomes more important or more difficult to achieve, when the separation in space, time orbackground between the interlocutors increases, and when the speaker is male, introverted and/or academically educated.

Abstract: The context of a linguisticexpression is defined as everything outside theexpression itself that is necessary forunambiguous interpretation of the expression.As meaning can be conveyed either by theimplicit, shared context or by the explicitform of the expression, the degree ofcontext-dependence or ``contextuality'' ofcommunication will vary, depending on thesituation and preferences of the languageproducer. An empirical measure of thisvariation is proposed, the ``formality'' or``F-score'', based on the frequencies ofdifferent word classes. Nouns, adjectives,articles and prepositions are more frequent inlow-context or ``formal'' types of expression;pronouns, adverbs, verbs and interjections aremore frequent in high-context styles. Thismeasure adequately distinguishes differentgenres of language production using data forDutch, French, Italian, and English. Factoranalyses applied to data in 7 differentlanguages produce a similar factor as the mostimportant one. Both the data and thetheoretical model suggest that contextualitydecreases when unambiguous understandingbecomes more important or more difficult toachieve, when the separation in space, time orbackground between the interlocutors increases,and when the speaker is male, introvertedand/or academically educated.