I’m interested in the effect of language use on linguistic structure. This field had been abhorred by mainstream linguistics ever since the time of de Saussure and the Structuralists, and rightly so: In the absence of proper resources or methodology, it was a bedlam of oddball theories linking the loss of suffixes to the decline of civilisation or explaining away sound changes by suggesting that the speakers had moved to a mountainous area (where they couldn’t open their mouths that wide due to the cold) or had learned how to cook meat (so their jaw muscles got weaker).
The adoption of rigorous statistical methods and the increase of available computing power, however, have allowed linguists to draw meaningful links between use and structure in language, not so much in the domain of altitude and culinary habits, but rather by investigating how social patterns, individual biases, and the networks of our interactions shape individual linguistic behaviour on the one hand, and language as a complex, adaptive system on the other.
This strand of, let’s call it, experimental quantitative sociolinguistics (with a great deal of psychology thrown in for good measure) has something of an unsound reputation as a science that is not much more than obsessive bean counting, only of interest to (a) sociologists who would find it remarkable that you meet more posh-sounding people in a posh department store or (b) historical linguists who appreciate the fact that words rarely used by the Anglo-Saxons eventually died out. The truth of the matter is, however, that inviting external influences into our linguistic descriptions is a great way of integrating our knowledge of human cultural and biological evolution, cognition, social dynamics, and acoustic and articulatory phonetics into a cogent theory of how humans speak — and think — together.