Acceptable But Not Grammatical

Chomsky's discussion of acceptability does not include a consideration of the possibility that a sentence may be acceptable but not grammatical (1965, 11-14); indeed, the very idea seems absurd. Yet there are familiar acceptable idiomatic expressions that resist grammatical analysis. Idioms, by definition, have meanings that are not determined by their grammatical and lexical structures. One group of idioms includes sentences that "have a perfectly reasonable meaning if understood as nonidiomatic" (Chomsky 1980, 150):

       (1) Shake a leg.

Another group consists of fixed expressions which preserve grammatical forms that are no longer productive, such as the hortatory subjunctive.

       (2) Far be it from me.

       (3) God damn you.

Sentences like (3) alternate with shortened forms where an element is deleted:

       (4) Damn you.

(4) is comprehensible because of the existence of (3), with which it is in free variation. Yet there are cases where there is no such variation.

       (5) Blast you.

(5) is not frequently heard, but it is similar enough in form and meaning to (4) to permit the conclusion that (4) and (5) are struturally identical. Gregersen suggests that these are the realizations of an English construction of the form "verb you" (1977, 264).

Other idiomatic sentences are harder to parse. Consider the following:

       (6) Fuck you.

       (7) Screw you.

(4) and (5) differ from (6) and (7) in stress; the latter two examples are unusual in that the sentence stress falls on the object pronoun even when no contrast is implied. Moreover, (6) and (7) do not seem to be hortatory subjunctives since there is no way to recover the deleted subject noun phrase. In fact, there is no such noun phrase that can be used before (6) and (7) without sounding ludicrous. Nor can (6) and (7) be imperative sentences; if they were, reflexivization would have taken place, as in the following imperatives:

       (8) Fuck yourself.

       (9) Screw yourself.

Gregersen argues convincingly that the deleted subject in (6) was historically "the Devil" (1977, 265), but he is careful to add the following: "I do not claim that 'the Devil' still functions as subject; reanalysis has, I think, totally changed the construction" (266).

Thus, (6) and (7) have no apparent subject. Quang argues that (4) and (5) and (6) are not sentences but epithets, that epithets involve a lexical category of quasi-verbs, and that epithets are generated by the phrase-structure rule "Epithet - Quasi-verb NP" (1971, 7).

By dividing utterances into the two categories of sentences and eptithets, Quang is in effect stating that the grammar used to generate sentences does not work for epithets. In other words, epithets are not grammatical sentences but a separate category. Quang includes the following examples among the epithets:

       (10) Shit on you.

       (11) To hell with you.

Quang's distinction between "verb" and "quasi-verb" is quite useful in explaining the difference between (6) and (8); however, “quasi-verb” becomes too broad and vague a category when applied to (10) and (11). I believe that "shit" in (10) is a noun. (11) is part of a productive pattern that includes

       (12) To the gallows with you.

(10) appears to be a shortened form of

       (13) May there be shit on you.

Moreover, (10) and (11) are stressed like (4) and (5), not like (6) and (7). Although Quang's discussion is quite thorough, he does not mention (7), nor does he list mock euphemisms such as

       (14) ?Copulate you.

       (15) ?Intercourse you.

Native speakers who have never heard (14) or (15) nevertheless agree that they are stressed like (6) and (7). Quang admits that the stress of (6) is a problem he has not dealt with in his article (1971, 7). The existence of (7) and the marginal existence of (14) and (15) are evidence that the phrase-structure rule for epithets is too inclusive; it suggests unity where in fact there is diversity.

Dummy subjects exist in the deep structure of agentless passives, according to Chomsky (1965, 137). We might say that similar dummy subjects exist in the deep structure to the left of slang verbs meaning "to have sexual intercourse with" and that the dummy is deleted by a rule that shifts the sentence stress to the right. The inelegance of the preceeding explanation suggests that it is false. Perhaps it is better to conclude that (6) and (7) are ungrammatical because they have no subject, dummy or otherwise. In other words, they do not exist at the deep-structure level. "But," as Chomsky tells us, "idioms that appear only at the S-structure level are very rare; we can regard this possibility as excluded in principle, with such exceptions as should be expected in the case of idiomatic constructions" (1980, 152). If so, these "marginal exceptions" are acceptable but not grammatical. It is not surprising that these grammatically aberrant sentences exhibit aberrent stress as well.

Note: I am grateful to Franklin E. Horowitz, Thomas Wasow and William R. Leben for their suggestions and advice.


Chomsky, Noam. 1965. Aspects of the theory of syntax. MIT Press, Cambridge.

—.1980. Rules and Representations. Columbia University Press, New York.

Gregersen, Edgar A. 1977. “A Note on English Sexual Cursing,” Maledicta 1, 261-68.

Quang Phuc Dong [pseud.]. 1971. "English Sentences without Overt Grammatical Subject," Studies Out in Left Field: Defamatory Essays Presented to James D. McCawley on the Occasion of his 33rd or 34th Birthday. Arnold M Zwicky et al. (eds), 3-10. Linguistic Research, Inc. Edmonton and Champaign.

This article was first published in Maledicta, The International Journal of Verbal Aggression. Vol. IX, 1986 - 1987.

This article was also published in California Linguistic Notes 23:1 (Fall-Winter 1991).