Twitter is a favourite forum of US President Donald Trump. He is known for using the 140 characters per tweet to talk up the things he likes, but more so to rant about things he doesn’t. At The Guardian, Steven Poole has produced “a handy guide to the topsy-turvy world of Trumpspeak” on the social media platform.
Acknowledging that the new POTUS has “made Twitter great again,” Poole gets under the surface of the Trump messages, examining mostly single-word items that tend to occur most frequently. Adjectives and nouns, often in the form of epigrammatic approvals or rebukes, are given a brief interpretation (e.g. “Great: under the permanent control of Donald Trump”), and then each is analysed in context and in further detail.
Items given the treatment are: bad, biased, deal, dishonest, dumb, enjoy, failing, fake news, great, horrible, over-rated, sad, smart, so, so-called, sorry, trouble, and ungrateful. The results are sometimes humorous accounts of what was comical material in the first place.
Poole does not disclose his methodology, but he may well be working from a linguistic corpus of the tweets. The mass of data compiled in a corpus allows analysts to adapt the granularity of scrutiny from a bird’s-eye view of (in this case, likely) thousands of items, down to particular phrases, words, or even parts of words. In doing so, they can abstract concise meanings for particular items (lexicographers also use corpora to assist in identifying new senses of words). As with the example of “great,” with Trump these are not necessarily what most people agree the word means.