A corpus-assisted discourse analysis of children’s and groomers’ talk in online grooming interactions

Harmful communication may not always be recognisable as such, especially when it is manipulative and deceptive and appears to be indistinguishable from innocuous communication. This is the case with online child sexual grooming, where talk from interactions between groomers and children may resemble that seen between friends or consenting adults chatting. However, recognising that online grooming may be taking place is not simply a matter of spotting tell-tale words or phrases. It requires engaging with ways that online grooming is discursive: involving groomers and children using language to perform particular functions as they pursue different goals through a dynamic exchange. We address this need in this study by providing the first ever complete account of online grooming discourse, one that identifies features not only of groomers’ talk but also of children’s, using collocates of the most frequent content words in a corpus of each. Comparing findings between the two highlights distinctiveness that helps make online grooming communication more identifiable. It also reveals strong similarity, perhaps reflecting groomers’ efforts to minimise perpetrator/victim contrast for deception purposes. An advantage of using a corpus-assisted discourse studies approach, as found in our study, is that it can uncover subtle, non-obvious patterns that may serve as indicators of online grooming despite such deception.