Spatio-temporal variation of conversational utterances on Twitter

Area of Science:

• Computational Linguistics
• Sociolinguistics
• Natural Language Processing

Background:

• Language evolution is influenced by social norms and communication patterns.
• Previous research indicated a 200-year trend of decreasing utterance length in English fictional dialogues.
• Modern digital communication platforms offer new avenues for studying language change.

Purpose of the Study:

• To investigate recent trends in utterance length using a large dataset of Twitter conversations.
• To examine the relationship between utterance length and demographic factors in geographically tagged tweets.
• To explore potential explanations for observed changes in utterance length.

Main Methods:

• Analysis of 229 million Twitter utterances from September 2009 to December 2012.
• Statistical examination of utterance length trends over the three-year period.
• Correlation analysis between utterance length and state-level Black population percentage for US-based tweets.

Main Results:

• A significant shortening of utterance lengths was observed on Twitter over the 3-year study period.
• Geographically tagged US tweets showed an inverse relationship between utterance length and the percentage of the Black population in a state.
• The findings suggest that increasing use of jargon and coined words may contribute to shorter utterances.

Conclusions:

• Utterance length reduction is a recent phenomenon observable in contemporary digital communication.
• Socio-demographic factors, such as racial demographics, may correlate with language use patterns, including utterance length.
• The study supports the hypothesis that evolving communication norms, including jargon proliferation, drive changes in spoken and written language.