-
3. Double Object Structures
- The MIT Press
- Chapter
- Additional Information
3 Double Object Structures 3.1 Basics I begin to address more directly the complex issue of double object (ditransitive) structures illustrated by (3.1a) and (3.2a) and the common (though not invariably possible) alternations of the first object with prepositionally marked ones, as in (3.1b) and (3.2b). (3.1) a. Melissa sent my brother a message. b. Melissa sent a message to my brother. (3.2) a. Claudia bought Jack a toy truck. b. Claudia bought a toy truck for Jack. These two patterns were distinguished by Fillmore (1965) mostly on grounds of di¤ering passivization possibilities, a matter I return to in chapter 7. Hereafter, I call the pattern in (3.1) Class A ditransitives, that in (3.2) Class B ditransitives. Barrel A literature mostly analyzes such English constructions without appeal to a notion of indirect object, though the term is sometimes used informally. This reality is consistent with the theoretical fact, touched on in chapter 2, that while Chomsky (1965) proposed configurational definitions of both the subject and direct object relations, he neither there nor anywhere since proposed any definition for an indirect object relation.1 But consideration of other NLs, even those related to English like French, German, and Spanish, leaves little doubt that there is a notion of indirect object (3 object) of the same order of reality in these NLs as the direct object (2 object) relation. In the NLs just cited, one could hardly fail to appeal to 3 object, specifically with respect to analogs of (3.1) and (3.2). So trying to justify a distinction between 2 object and 3 object in French, German, Spanish, or a host of other NLs would, internal to the grammatical traditions of these NLs, be absurd. One should see the ‘‘missing’’ notion of 3 object in Barrel A as a function of three facts: (1) the overwhelming dominance in the development of Barrel A ideas of work on English, (2) the fact that the extension of the notion of 3 object in English is much less clear than in NLs like French, and (3) the fact that Barrel A ideas o¤er no natural reconstruction of the 3 object relation. Point (2) will be explicated in what follows. That Barrel A work avoids a notion of 3 object in the analysis of English ditransitive structures is also remarkable for other reasons. First, traditional grammarians of English of course utilized this concept (see, e.g., Curme 1931; Jespersen 1927 [1961]; Sweet 1891). Second, as argued below, considerable evidence supports recognizing a 3 object relation to represent the first object in English ditransitives. Although the vast literature on English constructions like (3.1) and (3.2) has yielded a generally accepted analysis of neither the double object cases nor their (a)/(b) alternations, two general types of analysis subsume the majority of treatments. In the most common, ditransitives like (3.1a) and (3.2a) are taken to involve the occurrence as a superficial 2 object of an underlying oblique phrase represented by the respective PPs in (3.1b) and (3.2b); often this oblique phrase is taken to be a 3 object. This is the so-called dative shift analysis, discussed in works like Akmajian and Heny 1975; Culicover 1976; Emonds 1972a, 1972b, 1976, 1993; Emonds and Whitney 2006; Fillmore 1965; Green 1974; Jackendo¤ 1977; Jackendo¤ and Culicover 1971; Langendoen, Kalish-Landon, and Dore 1976; Larson 1988, 352–354; Marantz 1993, 117; McCawley 1998; Oehrle 1976; Postal 1971; Ross 1967 [1986]; Smaby and Baldi 1981; Soames and Perlmutter 1979; and Wexler, Culicover, and Hamburger 1975; to cite some of many. The other approach treats the first object of a ditransitive as a PP with an invisible P. I discuss this analysis type after providing more factual substance for the notion English 3 object. Embarrassingly, I myself long accepted a dative shift–style analysis (see Perlmutter and Postal 1983; Postal 1971, 1982). A relational variant of it called 3-to-2 advancement formed a building block of early RG accounts of clause structure. Statement (3.3) provides a principles-and-parameters version. (3.3) Ouhalla 1994, 174–175 ‘‘58a Mary gave John a book Starting with the former hypothesis, if the two NPs in (58) both receive their Case from the verb, we should expect both of them to 76 Chapter 3 [35.171.22.220] Project MUSE (2024-03-28 19:43 GMT) behave like direct objects of...