Theory and Experiment in Syntax

Grant Goodall

• Publisher: Routledge
• Year: 2021
• Print ISBN: 9780367749019
• eBook ISBN: 9781000516517
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book reflects on key questions of enduring interest on the nature of syntax, bringing together Grant Goodall's previous publications and new work exploring how syntactic representations are structured and the affordances of experimental techniques in studying them.

The volume sheds light on central issues in the theory of syntax while also elucidating the methods of data collection which inform them. Featuring Goodall's previous studies of linguistic phenomena in English, Spanish, and Chinese, and complemented by a new introduction and material specific to this volume, the book is divided into four sections around fundamental strands of syntactic theory. The four parts explore the dimensionality of syntactic representations; the relationship between syntactic structure and predicate-argument structure; interactions between subjects and wh-phrases in questions; and more detailed investigations of wh-dependencies but from a more overtly experimental perspective. Taken together, the volume reinforces the connections between these different aspects of syntax by highlighting their respective roles in defining what syntactic objects look like and how the grammar operates on them.

This book will be a valuable resource for scholars in linguistics, particularly those with an interest in syntax, psycholinguistics, and Romance linguistics.

Part I
Three-dimensional syntax

1 Coordination

Extract from Chapter 2 of Parallel structures in syntax: Coordination, causatives, and restructuring

DOI: 10.4324/9781003160144-3

1. Extraction

In this section, we will examine a class of phenomena first discovered by Ross (1967): wh-movement within coordinate structures. Our primary concern will be the Coordinate Structure Constraint (CSC) and the Across-theBoard (ATB) exceptions to this constraint, the two major generalizations noted by Ross. The facts in this domain are relatively complex, but the relevant facts are directly derivable from the formulation of the union of phrase markers we have been using thus far.

1.1 Coordinate Structure Constraint

Ross (1967) formulated the Coordinate Structure Constraint (CSC) as given in (166).

1. (166) The Coordinate Structure Constraint
   1. In a coordinate structure, no conjunct may be moved, nor may any element contained in a conjunct be moved out of that conjunct.

The first stipulation, that no conjunct may be moved, accounts for the ungrammaticality of sentences like those in (167).

1. (167)
   1. *What did George eat tuna fish and __?
   2. *Who did the mailman see __ and his wife?
   3. *I know a book which the senator wrote an article and __.

One of the conjuncts has been moved in each of these cases, clearly violating the CSC. Ross pointed out that this first part of the CSC is derivable from the A-over-A condition,¹ assuming that the two conjuncts form a constituent, as in (168).

1. (168)

   

Grosu (1973) and Pesetsky (1982) have argued that the inclusion of this case with the rest of the CSC is a false generalization, and they choose to let the A-over-A condition alone bar sentences like those in (167). However, we shall see that, in our system, these sentences will be ruled out independently of the A-over-A condition.

The second part of the CSC forbids movement out of a conjunct. This can be seen in (169).

1. (169)
   1. *What did Mary cook the pie and Jane eat __?
   2. *Who does Tom like __ and Larry hate primates?
   3. *This is the man who it was raining and Bill shot __.

Again, these examples straightforwardly violate the CSC.

An explanation for these facts may be found, I will show, in the system we have developed thus far. Consider the component sentences of the structure underlying (169a):

1. (170)
   1. What did Mary cook the pie?
   2. What did Jane eat t?

Recall that each of these must be syntactically well formed. (170b) is grammatical; the wh-phrase what receives a θ-role (and case) by means of its trace. This is not so in (170a): there the wh-phrase has no corresponding trace, and, in addition to constituting an instance of vacuous quantification, it receives no θ-role. Because (170a) is ill formed, it then follows that (169a) is as well. We may thus account for the ungrammaticality of (169a) and the other sentences in (169) and (167).

Notice that one cannot avoid this result by positing a different source for (169a), as in (171).

1. (171)
   1. Mary cooked the pie
   2. What did Jane eat?

This may have the right meaning, but it cannot be linearized as (169a). A linearization of (171) would have to count Mary cooked the pie and What did Jane eat? as conjuncts. This is not the case in (169a).

A similar account may be given of the ungrammaticality of sentences like those in (172):

1. (172) *How loudly is John sick and moaning __?

Here again, one of the component sentences, (l73a), is ill formed:

1. (173)
   1. How loudly is John sick?
   2. How loudly is John moaning t?

Sick does not subcategorize for a manner adverb, and thus, by the Projection Principle, there is no trace in (173a). This means that the quantifier phrase how loudly has no corresponding variable and that (173a), hence, violates the ban on vacuous quantification. Because (173a) is ungrammatical, so is (172).

Ross noticed that the CSC does not affect the operation of passive, as seen in the now-familiar example shown in (174).²

1. (174) John hunted tigers and __ was killed by snakes.

This sentence is generable, as in (175):

1. (175)

   

As we saw there, the θ-criterion is satisfied in this case.

We have seen, then, that the CSC effects follow directly from our formulation of the union of phrase markers. This is not the first time that the CSC has been derived from purportedly independent properties of coordination (see, e.g., Schachter 1977; George 1980; Gazdar 1981; Pesetsky 1982; Anderson 1983; Gazdar et al. 1985; Steedman 1985), but it is nonetheless significant that we have also been able to do it here.

1.2 ATB exceptions to the CSC

Despite the impressive generality of the CSC, there is one systematic class of exceptions to it. Consider sentences like the following:

1. (176)
   1. Which man __ ran the race and __ won the prize?
   2. Which film did the critics hate __ and the audience love __?
   3. How loudly is Mary screaming __ and John moaning __?

The CSC prohibits movement out of a conjunct, but in each of these cases there has been movement out of both conjuncts. While one might expect that this movement, what Ross calls “Across-the-Board” movement, would be an extreme violation of the CSC and consequently ungrammatical, this is not the case. In order to handle the grammaticality of sentences like those in (176), Ross adds the following condition to the CSC:

1. (177) unless the same element is moved out of all the conjuncts

I will show here that just as we do not need to stipulate the CSC, we similarly do not need to stipulate this exception to it. Both fall out from the union of phrase markers.

Let us consider the source of the sentences in (176). The component sentences involved are as in (178).

1. (178)
   1. Which man t ran the race?
      Which man t won the prize?
   2. Which film did the critics hate t?
      Which film did the audience love t?
   3. How loudly is Mary screaming t?
      How loudly is John moaning t?

Unlike real CSC violations, such as in (170), here both component sentences are well formed, because in each case the wh-phrase is coindexed with a trace and thus quantifies nonvacuously.

Thus, the grammaticality of sentences like (176) is unproblematic, given the union of phrase markers approach. The ATB exceptions to the CSC turn out to be completely regular, rather than exceptional.

1.3 Extraction in Palauan

The analysis given previously for English applies equally well to the Austronesian language Palauan, in which extraction phenomena appear, superficially at least, to be quite different.

As discussed in Georgopoulos (1983, 1985), Palauan obeys the CSC:

1. (179) *[a delak [a uleker er ngak [el kmo ngngera [a sensei a milskak a buk] mother asked P me COMP what teacher gave book
   me [a Toki a ulterur __ er ngak]]]]
   and Toki sold P me
   ‘My mother asked me what the teacher gave me a book and Toki sold me _. ’

In both English and Palauan, the sentence ‘My mother asked me what the teacher gave me a book’ is ungrammatical, and hence the sentence in (179) is ungrammatical. Neither language permits vacuous quantification.

Sentences in which both component sentences are grammatical are themselves grammatical:

1. (180) [a delak a uleker [el kmo ngngera [lulterur __ a Toki el me er ngak] me mother asked COMP what sold Toki come P me and
   [a Droteo ulterur __ el mo er a Toiu]]]
   Droteo sold go P Toiu
   ‘My mother asked what Toki sold __ to me and Droteo sold __ to Toiu.’

This, of course, is an instance of ATB extraction, in th...

Table of contents

• Cover
• Half Title Page
• Series Page
• Title Page
• Copyright Page
• Dedication Page
• Contents
• List of Tables
• List of Figures
• Introduction: Five themes in the study of syntax
• Part I Three-dimensional syntax
• Part II Syntax and argument structure
• Part III The syntax of subjects and wh-dependencies
• Part IV Constraints on wh-dependencies
• Index