How character limit affects language usage in tweets

In November 2017 Twitter doubled the available character space from 140 to 280 characters. This provided an opportunity for researchers to investigate the linguistic effects of length constraints in online communication. We asked whether the character limit change (CLC) affected language usage in Dutch tweets and hypothesized that there would be a reduction in the need for character-conserving writing styles. Pre-CLC tweets were compared with post-CLC tweets. Three separate analyses were performed: (I) general analysis: the number of characters, words, and sentences per tweet, as well as the average word and sentence length. (II) Token analysis: the relative frequency of tokens and bigrams; (III) part-of-speech analysis: the grammatical structure of the sentences in tweets (i.e., adjectives, adverbs, articles, conjunctives, interjections, nouns, prepositions, pronouns, and verbs); pre-CLC tweets showed relatively more textisms, which are used to abbreviate and conserve character space. Consequently, they represent more informal language usage (e.g., internet slang); in turn, post-CLC tweets contained relatively more articles, conjunctions, and prepositions. The results show that online language producers adapt their texts to overcome limit constraints.