And then one may try to propose a reason within Universal Grammar that would justify forcing such a movement to
take place. But, it is evident that the analysis rapidly becomes forced.^83
Section 6.1 also mentioned a second difficulty for an analysis like (10): idioms like those in (12) cannot be shoehorned
into a lexical category.
(12) a.The jig is up.
b. That's the way the cookie crumbles.
c. The cat's got NP's tongue.These are complete sentences; if they were inserted under V, the underlying structure of the sentence would lack a
subject and hence be ill-formed.
Theoverwhelming generalizationabout idioms, in fact,is that theyhavethesyntaxof garden-varietyphrases. A theory
withstructures like(10) has topredictthatthisgarden-varietysyntax arises byobligatorilymovingeverything out from
under the V node into—coincidentally—its normal position.
But even this is not enough. There are a few idioms whose syntax is deviant:
(13) a.all of a sudden
b. by and large
c. Far be it fro mNP to VP. (usually NP isme)
d. How dare NPVP!(13a, b) appear in adverbial contexts, but their syntactic structure is not that of an adverbial, nor entirely that of any
other category. (13c, d) are complete sentences that cannot be stuffed into a V. Their syntax is at best a relic of some
earlier stage of English, but of course they are perfectly acceptable in modern speech.
The upshotis that,although idioms must be stored, theway they must be stored cannotbe reconciled witha theory in
which only individual words are inserted into sentences. The obvious solution is to admit that, for instance,kick the
bucketis a lexically stored VP. In the approach to the lexicon sketched in Chapter 5), such a solution is quite natural.
(14) illustrates with the idiomtake to task.
LEXICAL STORAGE VS. ONLINE CONSTRUCTION 169
(^83) Regrettably, suchstrategieshavebecome so commonplaceincertaincircles ofgenerativesyntaxthatmanylinguistsnolonger recognizetheir unnaturalness; indeed theymay
even think highly of the analysis on the grounds that it shows us more about the abstractness of Universal Grammar. Such a stance tends to distance syntactic theory from
psycholinguistic and developmental considerations; my overall program (see Part I) is to try to reduce that distance.