Syntax and Discourse in the Acquisition of Adjunct Control

A. N. Adler, 2006

for $19.95 x
or
+

Save $9.95 (24%) if bought together with Resumptive Chains in Restrictive Relatives, Appositives and Dislocation Structure

for $29.95 x
or
+

Save $9.95 (24%) if bought together with The Syntactic and Semantic Roots of Floating Quantification

for $29.95 x
or
+

Save $13.95 (23%) if bought together with Proceedings of the 8th Workshop on Altaic Formal Linguistics (WAFL8)

for $45.95 x
Show more deals...

This dissertation is a study of null subjects in adjunct clauses in English. The goal is twofold: to establish a comprehensive theory of control in adjuncts, and to utilize this theory to understand the adjunct control interpretations of children aged 3–6. The theoretical basis of this work is Landau (2000, 2001), who characterizes the complement control mechanism (1) as a syntactic Agree relation (Chomsky 1995, et seq.). I argue that the same mechanism governs control into low-attaching adjuncts (like before, after, while and without) as well (2). High-attaching adjuncts and gerund subjects, on the other hand, are subject to discourse-governed control rather than the syntactic Agree relation (3). I argue that the topic of the sentence is the controller in these cases.

(1)       a. Johni wants/tries PROi to open the presents.

            b. Johni tells Maryk PROk/*i to cut the cake.

            c. Johni promises Maryk PROi/*k to help with the dishes.

(2)       a. Suei talked to Johnk without PROi/*k mentioning the party.

            b. Johni hugged Maryk before/after/while PROi/*k cutting the cake.

(3)       a. PROi Being low on cash, Marki picked the cheaper gift.

            b. PROi Organizing the whole event tired Maryi out.

This theory makes certain predictions for acquisition. Barring other factors, control in verb complements and low-attaching adjuncts should be mastered at a similar age, since they are governed by the same mechanism. On the other hand, discourse-governed control as in (3) is subject to separate constraints, and may not develop along the same time course. However, this does not exactly fit the acquisition data: many researchers have observed that control in before and after adjuncts may be delayed until age 5, beyond the age at which children master complement control. What accounts for this discrepancy? Do children give adult-like responses on any adjunct control structures?

I present data from Truth Value Judgment Task experiments conducted with over 50 children showing that 3-year-olds are above 94% adult-like on adjuncts headed by without. This demonstrates that even young children are capable of implementing control in adjuncts, as predicted by the control theory I argue for.

I suggest that the non-adult interpretations in temporal adjuncts are due to a delay in proper adjunct attachment height, rather than a delay in control. It is argued that adjunct attachment height is subject to a gradual learning process rather than an innate constraint. The reasons why certain adjuncts might or might not be wrongly attached are discussed. The result of misattachment is that the Agree relation cannot hold; the null subject in this structure falls into the domain of discourse-governed control rather than syntactic control. This is the root of the overly liberal control judgments from children.

The misattachment of an adjunct clause should affect pronoun interpretation as well, since it changes the c-command relation between the two clauses. I present results from a TVJ task with 30 children demonstrating that there are correlations between non-adult control and non-adult Principle C violations in temporal adjuncts, both attributed to the misattachment of the adjunct clause. Furthermore, we find similarities between the range of interpretations that children give to PRO in temporal adjuncts and in true discourse control structures as in (3), as expected under this account.

Thesis Supervisor: Suzanne Flynn, Professor of Linguistics and Second Language Acquisition

1 Introduction 13
1.1 Innate and learned factors in language acquisition 13
1.2 Control theory 15
1.3 Issues in acquisition: control and adjunct structure 18
  1.3.1 Previous research 18
  1.3.2 My account: the relevance of adjunct attachment height 23
1.4 Overview of experimental studies 26
         
2 Control theory 30
2.1 Introduction 30
  2.1.1 The history of PRO 30
  2.1.2 The development of control theory 33
2.2 Landau’s account of control 36
  2.2.1 On the interpretation of PRO in OC and NOC 36
  2.2.2 PRO in OC: partial vs. exhaustive control 39
  2.2.3 Agree as the mechanism of control 42
  2.2.4 Further issues 47
  2.2.5 The nature of NOC 49
2.3 Alternative theories of obligatory control 51
  2.3.1 Kawasaki (1993) 52
  2.3.2 Williams (1992, 1994) 57
  2.3.3 Conclusion: Agree as the mechanism of control 63
2.4 Focus on adjuncts: a comprehensive account 65
  2.4.1 The basic patterns of adjunct control 65
  2.4.2 Phases and Agree in adjuncts 69
    2.4.2.1 The structure of nonfinite adjuncts 70
    2.4.2.2 The nominal analysis of gerunds 71
    2.4.2.3 Conclusions: structural control in adjuncts 81
  2.4.3 Discourse control 85
    2.4.3.1 NOC PRO as a logophor 86
    2.4.3.2 Properties of sentence topics 88
    2.4.3.3 Topic-oriented PRO 95
2.5 Final remarks 101
         
3 Adjunct control in child language 104
3.1 Background on the acquisition of control 108
3.2 The nominalization hypothesis 114
3.3 The adjunct misattachment hypothesis 119
3.4 My account: the interaction between adjunct attachment and control 128
  3.4.1 Introduction 128
  3.4.2 More on misanalysis: subordination vs. coordination 131
  3.4.3 A remaining issue: finite vs. nonfinite clauses 138
  3.4.4 The question of learnability 139
  3.4.5 Predictions of the account 145
    3.4.5.1 The effect of misanalysis on control 145
    3.4.5.2 Predictions beyond the domain of control 148
3.5 The acquisition of discourse control 152
  3.5.1 Review: discourse control in adult grammar 152
  3.5.2 Relevance to the adjunct misanalysis account 153
  3.5.3 Hypothesis on discourse delay 156
  3.5.4 Further discussion: non-licensed anaphora in child and adult grammar 160
3.6 Final remarks 163
         
4 Experimental studies 166
4.1 Introduction 166
4.2 Methodological design 166
4.3 Experiment 1: without, while, after 169
  4.3.1 Hypotheses and predictions 169
  4.3.2 Subjects 170
  4.3.3 Materials and procedures 170
  4.3.4 Results 173
  4.3.5 Discussion 174
4.4 Experiment 2: binding and control 176
  4.4.1 Hypotheses and predictions 176
  4.4.2 Subjects 177
  4.4.3 Materials and procedures 178
  4.4.4 Results and discussion 182
    4.4.4.1 Control structures 182
    4.4.4.2 Principle C and correlations with control 186
    4.4.4.3 Ambiguous vs. unambiguous pronouns 190
4.5 Experiment 3: discourse control 195
  4.5.1 Hypotheses and predictions 195
  4.5.2 Subjects 196
  4.5.3 Materials 196
  4.5.4 Results and discussion 198
4.6 Summary and conclusions 202
         
5 Conclusion 204
         
References 211
Appendix 1: Experiment 1 materials 220
Appendix 2: Experiment 2 materials 222
  A2.1 Control structures 216
  A2.2 Principle C structures 223
  A2.3 Ambiguous pronouns (fronted adjuncts) 227
Appendix 3: Experiment 3 materials 229