From Bill Keller’s op-ed piece “Falling In and Out of War” in the NYT on 3/19/12:
(1) Policy makers should – and President Obama mostly has – put a premium on appraising alternatives to war.
A real-life example of a phenomenon discussed by Geoff Pullum and me in a 1986 article “Phonological resolution of syntactic feature conflict” (Language 62, on-line here): the verb form put (boldfaced above) serves simultaneously as two different inflectional forms of the lexeme PUT — as the BSE complement of the modal auxiliary should and as the PSP complement of the perfect auxiliary has. For almost all verb lexemes in English, these two forms are distinct (compare PLACE, with BSE place and PSP placed), so that the sort of reduced coordination in the Keller example apparently wouldn’t be possible, since there’s no available form that’s both BSE and PSP. For a fully parallel coordination, the distinct verb forms would have to be supplied:
(2) Policy makers should place – and President Obama mostly has placed – a premium on appraising alternatives to war.
But for about two dozen verb lexemes, of which PUT is one, the BSE and PSP happen to be phonologically identical, so that the conflict between the two feature values can be “phonologically resolved”, and the reduced coordination is (exceptionally) possible.
(Details of the analysis in the paper, especially pp. 761-2.)
I bring this up because in 1986 Geoff and I used invented examples to illustrate phonological resolution, but now here’s one from the editorial pages of the Times, indeed in the writing of a former executive editor of the paper. It’s nice to be able to illustrate these things from such sources.
Now, a complexity, which we recognized in the 1986 paper. The main factual claim is that for almost all verb lexemes, no form can serve both as BSE and PSP:
(3) *Policy makers should – and President Obama mostly has – place a premium on appraising alternatives to war.
(4) %Policy makers should – and President Obama mostly has – placed a premium on appraising alternatives to war.
As far as I know, examples like (a) are unacceptable for everyone, but on (b) there is variation (which is why the example is marked with a %, indicating divided usage). For some speakers, a “principled resolution” of the feature conflict is possible, with the nearest conjunct (in this case, the one with perfect has) determining the form of the verb.
Principled resolution in favor of government by the nearest is very common indeed, to the point where I’m reluctant to treat things like (4) as simply ungrammatical (however much some usage critics insist that parallelism is an inviolable condition on coordination); see extended discussion here.
The immediate point, though, is that even people who reject examples like (4) accept examples like (1). It’s all a matter of having just the right verb: PUT, RUN, SET, SPREAD, etc., but not PLACE, JUMP, APPLY, or almost any other verb you could pick out of a hat.