To my embarrassment, I did not know what the pragmatics wastebasket was, so I did some googling.
The history of linguistics is a succession of scholars saying: X is what we will pay attention to, and Y is crap we can’t be bothered dealing with, because it’s too messy.
- 500 BC: morphology is all we deal with in grammar
- 100 AD: morphology, (rudimentary) syntax, and rhetoric is all we deal with in grammar, and why you would speak at all is philosophy, not grammar.
- 1850: language change is all we deal with, and what language has ended up as is boring
- 1920: the language system (langue) is what we deal with, and what comes out of people’s mouths (parole) is boring
- 1960: syntax is what we deal with, and semantics is the philosophers’ problem, not ours
- 1970: syntax and semantics is what we deal with, and pragmatics is a philosopher’s invention, not ours.
Now something changed in the 1960s into 1970s.
Sociologists started looking at what came out of people’s mouths, and not just their underlying model of language. That gave rise to sociolinguistics and discourse analysis.
Philosophers of language started looking at why people said things in the contexts they did. That gave rise to pragmatics.
Yehoshua Bar-Hillel was as formal a linguist as formal linguists could be. Machine translation people like to burn effigies of him, because Bar-Hillel wrote a report to the US military in the late ’60s, that the effort to date on machine translation was never going to pay off, thereby stopping all research in machine translation for the next 20 years. (He was right, btw: machine translation in practice has nothing to do with formal grammars, which was the route machine translation had been taking. But Chomsky got plenty of NATO funding out of machine translation, before Bar-Hillel’s report.)
Bar-Hillel did something very cool in 1971; especially cool for a formal linguist.
He wrote a little note in Linguistic Inquiry (the home journal of Evil Chomskian Formalists), saying something like this:
“We’ve been treating pragmatics as a waste-basket of random crap that we don’t bother to account for in language. Every so often, someone goes through the waste-basket of random crap, and picks out something they think they can account for in their new shiny formal syntactic–semantic theory.
Instead of treating pragmatics as a waste-basket, and cherrypicking it for bits to account for formally, why don’t we instead start taking pragmatics seriously, and account for the stuff in the “waste-basket” on its own terms?”
The pragmatics waste-basket is what linguists have since been getting away from. Instead of treating it as random crap, shoved into Generative Semantics if it will fit (which is what the fashion was in the late 1960s), pragmatics started being treated seriously as its own discipline, with its own way of explaining phenomena. Just as sociolinguistics and discourse analysis did.