Linguistics vs. linguistics.

There are some linguistics jobs in software. There are many, many more applied linguistics jobs in software. They aren't the same thing. One of those things about formal or computational linguists is that they -- we -- identify ourselves and ourselves exclusively as linguists. There is a healthy dose of snobbery involved in this identification of self and other, but by and large they -- we -- are okay with that. Applied linguists, by contrast, which set includes language policy experts, educational linguists, and some linguist developers, typically feel that a linguist is anyone with a well-informed interest in language. Their notion is much more inclusive.

I'm a formally trained theoretical and computational linguist who's been spending a lot of time in the applied linguistic space in the last couple of years. As our group delves more and more into what we're all calling linguistic services broadly construed, I've found it useful to think about the distinction. I find that the most helpful way to divide the space is to think about the kinds of principles that govern the system of language and the kinds of principles that govern the system of language use. Both of these systems are interesting and useful and may be related, but they very likely exploit different explanatory principles. And my grad school self wouldn't have identified either of these pieces as applied linguistics; where linguists have often criticized applied linguists is in saying that applied linguists don't work with systems at all.

But since I'm working in a range of language-y areas now I feel free to taxonify the world in whatever way I find helpful, so I'm sticking to it.

Within software, there's room to explore and use both kinds of principles. The principles governing language -- what I'll call linguistics -- are relevant to anyone developing wordbreakers, grammar checkers, machine translation, and so on. The principles governing language use -- applied linguistics -- are relevant to anyone working in human-computer interaction, language/standards policy, and so on. The second category is a lot more heterogenous than the first.