I was reading through a brief response by Erin Brenner to Bill Walsh’s contention that the try and X construction should be opposed. (You know, like “I’ll try and write a new post sometime soon.”) Walsh’s basic point:

“‘I have to enforce this peeve,’ [Walsh] said. ‘You try to do something. To try and do something is to (a) try to do it, and (b) do it, which is not the intended meaning of the phrase.’

And Brenner’s:

The problem I have with Walsh’s reasoning is that try and is an idiom. There’s no point in trying to make sense of an idiom’s grammar; an idiom has its own unique (‘peculiar,’ says the American Heritage Dictionary) grammar. It doesn’t have to make literal sense.”

I agree with Brenner here. Sure, try and X doesn’t seem to make much sense.* But it doesn’t matter if it makes sense; if we’re trying to study language, we don’t get to say “I don’t understand this data” and throw it away. We’re stuck with the fact that people say and write try and X (the OED even offers an example from Paradise Regained, and Google Books has one from 1603) and it feels natural to most people.

When Walsh says that “To try and do something is to (a) try to do it, and (b) do it”, it’s clear what he’s getting at, but he’s wrong because that’s not what it means. What it means is what people use it to mean, and people overwhelmingly use it to mean (approximately) “try to do something”. That’s how language works; if everyone thinks a construction means X, then it means X.

It’s a similar problem with could care less, which people exasperatedly complain should be couldn’t care less. Of course it “should”. But everyone understands could care less to mean what it’s used to mean (with the possible exception of non-native English speakers and obstinate native speakers). And whenever most everyone agrees on what something means, whether it be a word or a phrase or an idiom, that’s right, no matter how illogical it seems.

That might sound weird. But if we’re going to treat language as something to be studied, as a science, then that ties our hands a bit. Quantum mechanics is a hot mess, and it sure would have been easier if it were Newtonian physics all the way down. But physicists don’t get to say, “nah, that’s crazy, let’s just keep using Newtonian models.”** Taxonomists don’t get to say “nope, platypuses are too strange, we just won’t classify them.” And so on.


Thankfully, taxonomists don’t have to classify Psyduck.

You can have an unassailable argument for why we shouldn’t be able to get the meaning out of a word or phrase or construction, but if everyone understands it, your argument is wrong. This is an essential fact of language. There are rules in language, but if the language itself breaks them, then it’s a shortcoming of the rule, not of the language.

So what can we say about try and? We can try to put together an explanation for how the unexpected meaning arose, looking at possible ancestries for it, possible analogical routes that might have spurred it. We can classify where and when and how it’s used (it’s generally informal, for instance). But when it comes time to figure out how it makes sense, it could well end up that all we can do is throw up our hands, call it an idiom, and move on. After all, what’s really interesting about language (at least for linguists) is the higher-level stuff like phonology or syntax or computational psycholinguistics; idioms are just the charming baubles that catch our eyes.

Of course, none of this means that one can’t be against an idiom — only that its supposed illogic is one of the weakest reasons to oppose it. I don’t have a problem with Walsh correcting try and in situations where it’s inappropriate or likely to cause confusion (e.g., formal writing or writing directed at an international audience). I do the same with non-literal literally, not because it’s confusing or incomprehensible or uneducated or new — it’s not — but because it feels cheap and hyperbolic to me, especially when used regularly. But these are stylistic choices, not grammatical ones. They aren’t returning logic to language.

*: It makes a little bit more sense when you think of the construction as an analogue of come and X or go and X, and realize that and in this situation is indicating dependence between the attempt and the action rather than simultaneity. The seeming noncompositionality of the construction may be in part due to language change, as the MWDEU notes that various related constructions (e.g., begin and) were common in the past, and thus when try and emerged, the dependent sense of and may have been more productive. In fact, the MWDEU hypothesizes that try and may predate try to.

**: Of course, they can do this when they’re staying at macroscopic scales, where quantum effects are undetectable — and thank God for that, or I’d’ve never survived college physics classes.