Understanding Troll Writing as a Linguistic Phenomenon

Abstract

The current study yielded a number of important findings. We managed to build a neural network that achieved an accuracy score of 91 per cent in classifying troll and genuine tweets. By means of regression analysis, we identified a number of features that make a tweet more susceptible to correct labelling and found that they are inherently present in troll tweets as a special type of discourse. We hypothesised that those features are grounded in the sociolinguistic limitations of troll writing, which can be best described as a combination of two factors: speaking with a purpose and trying to mask the purpose of speaking. Next, we contended that the orthogonal nature of these factors must necessarily result in the skewed distribution of many different language parameters of troll messages. Having chosen as an example distribution of the topics and vocabulary associated with those topics, we showed some very pronounced distributional anomalies, thus confirming our prediction.

Collective language creativity as a trade-off between priming and antipriming

“…Surprising as it may seem, a new insight into the problem of language creativity came from an unexpected venue, namely, from studying discourse practices of internet trolls. Monakhov [ 24 , 25 ] showed that a number of features inherent in trolls’ tweets are grounded in the sociolinguistic limitations of this type of discourse, which, in essence, is an imitation, make-believe game. A troll wants to achieve their goal without being identified as trying to achieve it.…”