Thread by @brettrey3

A new English determinative is being born: "a couple" (e.g., a couple years, a couple things, a couple people). #linguistics #diachronic

It's clearly evolving out of the NP frame "a couple of" + plural NP and may be influenced by the determinative "a few". Even if you don't accept "a couple people", you might find "a couple hundred" works like a gateway drug.

Here's the frequency of "a couple of" + N sinc the 1990s. The first line is the number of tokens, the second is the size of the subcorpus in millions, and the third is the frequency per million words. Not much happening here.

Here's the frequency of "a couple" + N. Back in the early 90s, it was only about 1/13 as frequent as "a couple of", but today it's about 1/4 as frequent.

At about 22 times per million words, this is in the top 3,000 word families in English, quite frequent.

Predictably, it's most comon in the spoken subcorpus, followed by fiction, and still quite rare in the academic subcorpus.

Data from english-corpora.org/coca/ #corpus

You can even see it in the New York Times: "a couple species"
nytimes.com/2019/10/03/sci…

In this graph, you see the rise of "a couple hundred" in the Google Books corpus, relative to "a couple of hundred". This seems to start in the early 1900s. There are non-zero levels all the way back to 1800, but these are probably mostly spurious.

Notice that "a couple" is almost always followed by a plural NP, but that's not the case with "hundred" and "thousand".

Another factor in this may be "a couple more", where "a couple" is an NP functioning as modifier of "more".

"a dozen" may be leading our ears down this path as well, even though there are semantic differences, I think, between "dozen" and "couple" beyond the obvious 12 vs 2 difference.

This graph shows some of the most frequent nouns determined by "a couple". They seem to follow similar trajectories.

It's not common, but you can still find many exmples of "a couple" + Adj + plural N (e.g., a couple other things).

Geoff Pullum points out to me that, "a couple" could be merely an unusual NP functioning as a determiner, and not a determinative a la "a few" or "a little". #categoryfight
#corpuslinguistics

Arguing for the single-word analysis is the fact that you can hardly put a modifier between "a" and "couple". A few examples of "extra" (e.g., an extra couple feet) is about all you find.

There are also six instances of "mere" (e.g., a mere couple feet/lines/degrees/blocks) in the iWeb corpus, but we're talking 6 tokens in a corpus of 14 billion words.

Contrast this with "a few", which readily allows internal modification by "very", "select", "lucky", etc.

On the other hand, you get examples like "the next couple games", "the past couple games", or "every couple feet" and "the extra couple bucks".

This suggests there's something happening with "couple". There's no evidence that it has spawned an adjective (e.g., *they became couple, *very couple, *as couple). But I'm not sure what to do with it. Here are two possible analyses.

In arguing that "a few" and "few" are distinct determinatives, CGEL makes the semantic argument that they have distinct meanings. That approach doesn't seem possible here, since "couple" itself is not a determinative.

On balance, I'm in favour of the determinative analysis, but I admit it's got a couple problems.

Here's one with "a couple" as Det:NP

@threadreaderapp unroll