Situating Twitter Discourse in Relation to Spoken and Written Texts

Language use on Twitter is a recent and currently understudied phenomenon. The present study makes an empirical contribution towards understanding the similarities and differences between Twitter and offline registers. To this purpose, a lectometric analysis is conducted: on the basis of 236 linguistic features whose frequency information is measured in 7 309 spoken, written and Twitter texts, 10 dimensions of linguistic variation are developed. Within the space of variation circumscribed by these dimensions, Twitter is characterized not so much by radical difference as by gradual differentiation from other registers. The only dimension along which a clear distinction between Twitter and other registers can be seen is that of “colloquial markedness”. In general, language on Twitter can be described as involved and consciously informal, but not very much oriented towards either addressees or factual statements.