Five ways of being new
« previous post | next post »
Geoff Pullum posted a little while ago on it's time (that) + a PST clause ("It's time (that) we left"), which he noted as an English construction that was new to him. [Correction 4/19/09: Geoff was actually commenting on the similar construction with a PRS clause.] One commenter, Stephen Jones, labeled this as an instance of the Recency Illusion, on the grounds that such expressions had been around for some time. But Geoff wasn't claiming that the construction was new in the language, only that it was (as far as he could tell) new to him; he replied:
My phrase "brand new fact about English syntax that I had no inkling of when I woke up this morning" meant "brand new in my experience". The post is explicitly about the personal experience of discovering things one didn't know before, not about the very difficult business of dating the start of an incipient change.
I read him as also suggesting (though not actually saying) that it hadn't been previously recognized by scholars (if it had, it would probably have come to his notice; but of course even distinguished scholars of English grammar miss things on occasion).
Linguists are delighted to come across phenomena that are new to them (new-1) and phenomena that are (relatively) new to a language (new-2) and phenomena that haven't been appreciated in the scholarly literature (new-3), but these are not the same thing.
On new-1 vs. new-3: the linguistic literature is full of observations that some bunch of facts cohere, in one language or across languages in significant ways. These are discoveries, some big, some small. The small discoveries are often language-particular and of the form: this pairing of form and meaning doesn't follow from already-described principles of the language. In the case at hand, what's been said so far about the way PST forms are used in English doesn't predict that "It's time that we left" can be used to convey something very close to (the forward-looking) "It's time for us to leave".
There are some tricky issues about how to talk about this case, and the (many) comments on Geoff's posting touch on some important ones but also get things mixed up in several different ways. I'll put off for another posting the task of clarifying fundamental concepts, which requires some crucial geekiness that will take up significant space and be strange to many readers.
But here I want to note that new-2 actually covers three importantly different notions of newness (so that there are five notions in all):
new-2a, with reference to when an expression or pattern (in any sense) first appeared in the language;
new-2b, with reference to when an expression or pattern (in the relevant sense) first appeared in the language;
new-2c, with reference to when an expression or pattern (in the relevant sense) "took off", became reasonably frequent.
An expression or pattern might have been around for quite some time, but not in the sense we're interested in. Lexicographers are plagued by such cases.
And an expression or pattern (in the relevant sense) might have been around for quite some time — even might have been innovated on different occasions — but have lain around, in small numbers "in the underbrush", "under the radar", until there's a fashion for it (for whatever reason; often, these things are inscrutable).
New-2c is the hot stuff. But all three notions of new-2 (which have to do with newness to the language) are different from new-1 and new-3 (which have to do with newness to certain observers of the language).