Yesterday, Daniel Mahaffey wrote to ask about his friend's "unusual indirect object sentences". Thus after backing into a dog in a crowded kitchen, she said "I nearly stepped on me a dog".
Daniel reasons that this is analogous to the benefactive pronouns in standard written English phrases like "I wrote him a song", or in widespread vernacular examples like "I wrote me a song" (where the standard version would be "wrote myself a song").
But, as Daniel observes, a couple of things are different in this case. First, analogous examples in (what we might call) the standard vernacular would put the pronoun before rather than after an intransitive preposition: "I stacked me up some firewood", not "I stacked up me some firewood". Second, in "… stepped on me a dog", the pronoun me has a looser semantic connection to the verb than typical benefactives — it doesn't refer to a beneficiary, recipient, or purpose of the action, instead apparently adding just a vague sense of interest or involvement. This may be why "…stepped me on a dog" seems (if anything) even odder to me than "…stepped on me a dog".
But wait, there's more. Daniel notes that his friend (who is from SE Georgia) also says things like "I need to go look for me a dress" or "I'm going to the mall to shop for me a dress".. In these examples, the placement of the pronoun seems even more surprising, since for is a transitive preposition expressing an argument of look or shop — here "a dress" — and me is thus inside a prepositional phrase, not just on the wrong side of a particle.
For a recent discussion of constructions of this type, with a good survey of the literature, see Laurence R. Horn, "'I love me some him': The landscape of non-argument datives", in Bonami & Hofherr (eds.), Empirical Issues in Syntax and Semantics 7, 2008.
A familiar syntactic feature of dialectal (Southern and Appalachian) U.S. English is the optional occurrence of a nonsubcategorized “personal dative” pronominal in transitive clauses which obligatorily coindexes the subject but whose semantic contribution is ill-understood. As we shall see, this personal dative (PD) bears suggestive if not always straightforward relations to constructions in such languages as French, German, Walpiri,Hebrew, and Old English involving what have been variously termed “ethical”, “free”, “non-lexical”, or “affected” datives. Some of these datives are coreferential with the subject (e.g. Je me prends un petit café, lit. "I take me a little coffee") while others are non-coreferential (e.g. Ils lui ont tué son oiseau, lit. " They killed him his bird"); they typically invite benefactive and malefactive (adversative) understandings respectively.
On the helpful side, Larry argues that the southern American "personal datives" are not indirect objects, and indeed claims that "they are not arguments at all, but non-subcategorized pronouns". An analysis of this type might open up wider syntactic vistas, including even occurrence between a preposition and its object noun phrase. But Larry's list of the properties of PDs starts with two generalizations that rule out Daniel's friend's usage:
a. PD constructions always co-occur with a quantified (patient/theme) direct object.
b. PDs can’t be separated from the verb that precedes and case-marks them.
So if Daniel's description is accurate, and the cited sentences are not speech errors and not re-analyzable in some other way, then there's indeed something different here from the patterns described for the southern American personal dative by Larry and the series of past works that he cites. I believe that I've also heard things like Daniel's examples, and I'll check with my southern and south midlands consultants to see what they think.
[The section on the dative in Allen and Greenough's Latin Grammar observes that
In Latin the Dative has two classes of meanings:
- The Dative denotes an object not as caused by the action, or directly affected by it (like the Accusative), but as reciprocally sharing in the action or receiving it consciously or actively. […]
- The Dative is used to express the purpose of an action or that for which it serves. This construction is especially used with abstract expressions, or those implying an action.
Further down in the section, we find:
The Dative often depends, not on any particular word, but on the general meaning of the sentence (Dative of Reference).
The dative in this construction is often called the Dative of Advantage or Disadvantage, as denoting the person or thing for whose benefit or to whose prejudice the action is performed. […]
In this construction the meaning of the sentence is complete without the dative, which is not, as in the preceding constructions, closely connected with any single word. […]
The Dative of the Personal Pronouns is used to show a certain interest felt by the person indicated.
This construction is called the Ethical Dative. It is really a faded variety of the Dative of Reference.
This is not to say that English grammar is Latin grammar. But there's clearly an analogy in this case.]
[Update — two observed examples:
I have you some chocolate.
Context: response to a request for something hot to drink; the chocolate in question had been obtained, some time earlier, specifically for the requester.
Here's you a bowl of soup.
Context: serving something unrequested; the soup is for general consumption, but is offered outside of meal times.
Both examples seem roughly equivalent to a final "for you" — "I have some chocolate for you"; "Here's a bowl of soup for you".]