Means of communication
« previous post | next post »
Bruce Eric Kaplan cartoon in the New Yorker of 22 September, p. 61: woman speaking on the phone, saying
You never write, you never call, you never fax, you never e-mail, you never text, you never page.
Six verbs here, all referring to means of communication. Such verbs have a certain amount of fame for English syntacticians, because some of them represent a clear island of regularity in what looks at first glance to be a sea of idiosyncrasy.
Some background. These verbs (representative of a larger set that includes (tele)phone, mail, telegraph, IMS, and others) can occur with one or the other of two arguments — A1, denoting the recipient of the communication, and A2, denoting the material communicated — or both. There are differences from verb to verb:
Write is compatible with A1, A2, or both: You never write me; You never write letters; You never write letters to me / write me letters. Phone is (for me) similar.
Call (by phone) is compatible only with A1: You never call me; *You never call your news; *You never call your news to me / call me your news. Page is (for me) similar, although you can google up a few instances of "page me your".
(There is some variation from speaker to speaker here.)
[Digression: a cute wrinkle with call is that the MANNER-of-communication verb call — in a class with shout, scream, whisper, etc. — is fine with both A1 and A2, in Kim called the answer to me in a loud voice, and for a great many speakers in Kim called me the answer in a loud voice as well.]
The WRITE, or ditransitive, class involves transfer of information (while the CALL, or simple transitive, class involves mere contact), so that the WRITE class falls in with a big class of verbs denoting (literal or figurative) transfer, of which give is the central member.
The verbs in the cartoon are also all "superficially intransitive" — they have no direct object — but they are all "understood transitively". In the context of the cartoon, they are all interpreted as having an A1 denoting the speaker: you never write me, you never call me, you never fax me, you never e-mail me, you never text me, you never page me. (When A1 or A2 or both go unexpressed, there's a range of interpretations for the unexpressed argument(s). That's an interesting topic in its own right, but not the one I'm aiming at here.)
For transfer verbs (of all sorts), there are (among others) two possible assignments of the arguments A1 and A2 to syntactic functions: letting DO stand for "direct object", PO for "prepositional object", and 2O for "second object", the two of interest here are
Prepositional dative (PDAT): V DO to PO (where DO denotes A2 and PO denotes A1), as in give the money to Kim
NP dative (NDAT): V DO 20 (where DO denotes A1 and 2O denotes A2), as in give Kim the money
[There's a tradition in transformational grammar of referring to NDAT as the "Dative Movement" construction, the idea being that PDAT is the basic construction and that NDAT is derived from it by moving PO to become DO, while demoting the DO of PDAT to 20. I don't accept some of the theoretical and analytical assumptions in this account, and prefer to discuss these constructions in a more neutral way.]
The possibility of PDAT, NDAT, or both varies from verb to verb (indeed, verb use to verb use) in a complex way. Transfer verbs generally can occur in the PDAT construction, though there are wrinkles in the pattern. (For example, though give is almost always fine in PDAT, the idiom give someone their due is NDAT-only: *We gave their due to them.)
The action comes mostly in the NDAT construction. There are many transfer verbs that don't work (for most speakers) in NDAT: donate (so close in meaning to give) is a standard example:
We gave/donated the money to Kim.
We gave/*donated Kim the money.
and transfer itself is unhappy in NDAT:
We transferred the money to Kim.
*We transferred Kim the money.
The literature on English verbs in NDAT is simply enormous. Some scholars have concluded that the whole thing truly is verb-by-verb: people just learn, one by one, which verbs allow NDAT. (And then we have a puzzle, which I've posted about here before: how do people know that it's ok to generalize some constructions, like PDAT, to lexical items they haven't heard before in this construction but have the right semantics, while failing to generalize in this way for other constructions, NDAT in particular? More generally, how do people learn restrictions on constructions?)
Other scholars have looked for factors that might predict, or at least favor, the possibility of NDAT for particular verbs, or that might act in concert to do so: semantics; phonology (in particular, number of syllables or number of feet; give and donate differ here); morphological composition (give is morphologically unitary, while donate looks like it has a suffix -ate); stratum of vocabulary (give strikes everyone as a "plain English" item, while donate seems Latinate); frequency (give is hugely more frequent than donate); level of formality (give is neutral, donate more formal).
The task of evaluating these hypotheses is daunting, but one thing has long been known to stand out as separate from all the messiness: means-of-communication transfer verbs are happy in NDAT as well as PDAT, and this seems to have nothing to do with their phonology, morphology, etc. — only with their semantics. In fact, as soon as a new means of communication arises and a verb appears for it (usually zero-derived from the noun denoting the means of communication — that is, the verb means 'use MEANS to communicate'), you don't have to hear NDAT examples for it to know that it can be used in the NDAT construction. So, not only
You should telegraph/fax/e-mail/IMS/text/Skype/… the answer to me.
but also
You should telegraph/fax/e-mail/IMS/text/Skype/… me the answer.
You get the NDAT use "for free" (and the PDAT use too, though that's no big surprise); just the noun can be enough. Meanwhile, donate and transfer and many many others resist NDAT (though there are a few grammatically adventurous people who go boldly into NDAT territory with some of these verbs).
Linguists love this sort of thing. They make nice puzzles. (You will, however, notice that I'm not offering an explanation. Please don't write to tell me that some of the NDAT examples just "sound right" and others don't; that's just restating the facts, and the puzzle.)
[Historical note: I have no idea who first noticed the generalization about means-of-communication verbs. I've seen it attributed to a long list of linguists, including me. Probably a number of people noticed it independently, and it diffused along several different pathways.]