-
2. The Syntactic Representation of Scope
- The MIT Press
- Chapter
- Additional Information
The Syntactic Representation of Scope The representation of quantifier scope is critical at multiple points in this monograph. We thus spell out here, in partially informal terms, our general assumptions in this area. Crucially, we take the representation of scope to be syntactic to exactly the same extent as the representation of phrase structure, word order, categories, and so on. We take the scope of quantifiers to be represented syntactically by the presence of DPs in clausal scope positions. For convenience, we follow May (1985, 1989) and assume that a scope position is a position adjoined to a clause [S DPi S]. In these cases, the embedded S contains a DP bound by DPi = [DP Di NPi] (in the scope position). S in effect provides the syntactic representation of an open sentence defining the argument of the semantic value of the quantificational DPi, where NPi denotes the restriction of the quantifier represented by Di. In other words, a quantificational DP will always have at least two distinct occurrences, one in a scope position and one in a nonscope position (an “argument” position in some approaches). There can of course be several DPs in scope positions of a single clause, yielding structures like [S DP1 [S DP2 [S DP3 S]]]. The key feature of representations of the type just outlined is that relative semantic scope is represented by relative height in syntactic structures. Since the relations between multiple DPs in scope positions are hierarchically represented , the fact that some quantifier DPs in a clause scope over others falls out from the representation. We naturally assume that there are principles relating syntactic scope to the corresponding semantics (see Heim and Kratzer 1998:chap. 7 for one such specification). In general, we need not concern ourselves with the details of such principles but we do spell them out in detail for polyadic quantifier cases in chapter 6. From the point of view of this monograph, it is mostly irrelevant whether the relation between a quantifier DP in scope position and the bound DP position in the associated S is due to quantifier lowering as in G. Lakoff 2 14 Chapter 2 1971:238–243, McCawley 1971:228–230, 1973:150, 294–295, and Seuren 1974a:106, 118–120, 1974b:192–195, 1996:301, 318–319, 1998:522–524; quantifier raising as in May 1985, 1989, and much later work (e.g., Heim and Kratzer 1998); multiattachment of the quantificational phrases in their argument and scope positions with scope positions picked out by a designated edge label, as would be natural in the Metagraph framework of Johnson and Postal 1980 and Postal 2010; or raising to the specifier of a dedicated functional projection (ScopeP), in the spirit of Rizzi 1997. The reason we need not focus on the (of course real and not unimportant) differences among these views is that in all the approaches cited, a key shared notion is that the height of a quantificational expression in a graph/tree indicates scope. For example, in Principles and Parameters work, the c-command domain of a quantificational expression is identified as its scope. That is, a structurally higher quantificational expression by definition scopes over a lower one. Given this background, in this monograph we represent (1a) as in (1b). Where necessary, we will provide a somewhat more elaborate representation. (1) a. Myron saw no student. b. [S [S Myron saw DP1]] In this structure, DP1 has two occurrences. One occurrence is in scope position and the other is in direct object position (on the definition of occurrence, see Collins and Stabler 2012). The notation coindexed with a distinct Xi is utilized throughout to represent distinct occurrences of the same constituent X, where the one in “” is unpronounced. As noted above, we take no position on whether structures like (1b) are formed by quantifier raising, formed by quantifier lowering, or simply nonmovement multiattachment structures. In particular, unlike May (1985, 1989), we do not assume that structures like (1b) are formed by covert movement, although such an approach would be consistent with our results. In the framework of Postal 2010, each member of the set of coindexed DPs would represent a single phrase, each X occurrence marking a separate arc sharing as head the single phrase represented by X. As is evident in (1), the assumptions just made must go hand in hand with principles determining which of the multiple occurrences of a DP are deleted, that is, which are pronounced and which not. One clear...