The cats and frogs are really raining out there.

If I do say so myself, I'm pretty good at picking up new languages. My chosen line of work was not selected at random. But pretty much the most interesting thing about learning new languages, for me, is analyzing patterns in the kinds of mistakes I make (at least insofar as I am able to do so -- after all, I know that I don't sound fluent, but I may not really know why.)

I've started keeping a list of non-nativisms that I happen to overhear people around me say. The first thing I realized is that non-nativisms are sort of like the old chestnut about pornography: pretty hard to define, but I know it when I see it. I know the conventional linguistic wisdom regarding native speaker performance errors -- who speaks in prescriptively well-formed, complete sentences all the time? -- which is why the linguist's notion of grammaticality has little to do with the precriptive grammarian's notion. For a linguist, a grammatical utterance is one that a native speaker would call grammatical.

Well, that's just great. Except try explaining that to someone trying to learn a foreign language, where what people really want is something concrete, some rule to follow. In linguistics we do a lot of work to generate models of grammatical systems. We are pretty darn good at models, actually. What we are not generally so good at is explanation. (Saying that some languages show agreement morphology while others do not because some languages require words to move to check off an agreement feature does not count as an explanation, even if you do use Greek letters in your description.)

But in fairness models can be tremendously useful. Especially for non-native speakers whose main interest is in identifying and successfully applying speech patterns, models can offer a framework for study. Give me a list of rules to show me where to put my indirect objects and it's pretty much smooth sailing.

Still and all, entertain the following conjecture: We can learn something pretty interesting about the parameters of the target language by looking at the patterns of mistakes that adults demonstrate when trying to learn it. We already know that we can infer quite a bit about the structure of the native language, but this is something a bit different. Something I'm thinking about, anyway.

In the meantime, check out some of these gems, said by people I know and recorded by me (not all of which shed light on the conjecture above, but all of which are awesome on their own terms):

  • The proof is in the batter.
  • It nailed the coffin.
  • You're like a donkey jumping ahead.
  • From my standpoint of view...
  • You're really freeping me out!
  • The cats and frogs are really raining out there.
  • Let's take a check for the rain.
  • I'm telling you, it was the straw on the camel's pack.

I've got tons. This stuff is really good. I'm trying to track the utterances by the speaker's native language of origin and some really interesting patterns are starting to emerge. I'll likely post an update here when I have some time to really dig into it -- and if you've got examples that are just too good to keep to yourself, I'd love to add them to my list.