Doctoral dissertation
defended August 29, 1996
at the University of Groningen.
The work reported here was carried out in the Department of Dutch of the University of Groningen, as part of the PIONIER project Reflections of Logical Patterns in Language Structure and Language Use. This project was jointly supported by the Dutch Organization for Scientific Research (NWO) and the University of Groningen.
Promotor — Prof. Dr. F. Zwarts
Co-promotor — Dr. J. Hoeksema
Copyright © 1996 by Sjoukje van der Wal
Mail comments to sng-redactie@hetnet.nl
Groningen Dissertations in Linguistics 17
Languages meet the communicative requirements of their speech communities; shaped by their speakers and hearers, they offer means to negotiate meaning in communicative situations. As some meanings regularly go together with a particular pragmatic intent, specific combinations of word meaning and sentence structure may in time become customary. This dissertation deals with an instance of such a settled pattern in language: polarity sensitivity. This term refers to the phenomenon that certain expressions may become sensitive to the polarity of the sentences in which they occur. Two categories of such expressions can be distinguished, depending on the direction of their sensitivity: positive and negative polarity items. In the present work, the focus is on negative polarity items (NPIs), words and expressions which more readily occur in negative contexts than in affirmative ones. For many of these expressions, the sensitivity to negative polarity is so strong that it must be cast in terms of licensing: sentences containing these expressions are incorrect when they are not related to negative polarity. The sensitivity then must be considered as an actual constraint in grammar.
This dissertation addresses the question as to how the restrictions on the distribution of NPIs are acquired by children. The focus in this investigation will be on Dutch, with a few short excursions into other languages. The data come from corpus studies of spontaneous speech and two experiments.
First, in Chapter 1, the multi-facet phenomenon of polarity sensitivity is presented. This chapter discusses the broad range of NPIs, the clustering of these expressions in specific semantic domains, as well as the different theories which have attempted to capture the restrictions governing their distribution. A point of specific relevance for acquisition matters is that NPIs are negative exceptions to the rules of grammar. This poses the no negative evidence problem: how do children, in the absence of evidence about what is not grammatical in the language they are acquiring, deal with the hazards of making too broad generalizations? The most advantageous acquisition strategy would seem to be a conservative one, in which children start off with a very narrow generalization which in time is to be widened on the basis of positive evidence available in the adult input.
Chapter 2 discusses spontaneous speech data about the onset of Dutch NPIs, at the end of the first year. The first-appearing NPIs, the verb hoeven (have to/need) and the adverb meer (anymore), occur in very simple sentences. Nevertheless, already at this young age, the crucial condition for grammaticality is met: these expressions occur only in combination with negation.
The spontaneous speech data presented in Chapter 3 show that this pattern of usage continues at a later age. At the same time, however, the licensed utterances are now accompanied by utterances which from an adult point of view are ungrammatical, as they contain no correct licenser. The question then arises as to whether such errors might result from overgeneralizing. In order to investigate this matter, an experiment with children is carried out which elicits indirect grammaticality judgments of both correct and incorrect utterances with the NPI hoeven. The response patterns make it clear that young children are sensitive to the restrictions on the distribution of hoeven; in general, they do not accept utterances in which hoeven is not accompanied by a licenser.
In the light of this finding, the patterns found in spontaneous speech are reconsidered in Chapter 4, in order to clear up the apparent incongruity between the experimental results and the naturalistic data. A close investigation of the spontaneous speech data shows that the acquisition of NPI licensing is indissolubly connected with the acquisition of negation, in such a way that children's developing knowledge of ways to express negative meaning functions as the forerunner for environments in which they use NPIs.
Chapter 5 pursues the developmental line further and investigates the question how children's early sensitivity to the restrictions on NPIs is adjusted and unfolded until it results in a licensing pattern which is in complete accordance with adult usage. Data from a written corpus and the results from a second grammaticality judgment experiment show that this is a protracted process which continues until the teenage years.
In retrospect, doing research on how children learn to use negative polarity items has been a challenge and, above all, a learning experience for myself. Friends and colleagues have, in various ways, made a contribution to the completion of this thesis and have made the four year period as an AIO one which I would not have missed for the world. I would like to take this opportunity to express my sincere gratitude to them all.
First and foremost, I want to thank the compilers of the child speech corpora, who carried out the time-consuming job of collecting and transcribing children's utterances and who were so generous in making these data available. Many thanks also to Brian MacWhinney, who distributed the corpora through CHILDES and who kept us posted via the info-CHILDES list. The data from German (the Clahsen, Stern, Wagner, and Wode corpora) and English (the Bloom, Brown, Clark, Kuczaj, MacWhinney, Sachs, Snow, and Suppes corpora) were all obtained from CHILDES, as were some of the Dutch data (the Utrecht, Van Kampen, and Wijnen corpora). In addition, several other Dutch corpora were made accessible to me, compiled by Gerard Bol, Jacqueline Frijn, Karin Heesters, Evelien Krikhaar, Folkert Kuiken, Paulien Rijkhoek, Kees Stevens, and Frank Wijnen, as well as diary notes from Jack Hoeksema, Charlotte Koster and Jacqueline van Kampen. Also, Jan Berenst has been more than generous in lending me his collection of elementary school children's essays.
On the basis of the spontaneous speech data, two experimental studies were carried out. Also in connection with these experiments, there are many people I owe thanks to. Teachers and group leaders have given a warm welcome at their schools and day care centers: MikMak and Dribbel in Haren, De Meerpaal, Pinokkio, Het Werkmancollege (thanks in particular to Jane Brinkers!), and De Eerste Openbare Montessorischool, all in Groningen, and Dukkie and Ebenhaëzer in Vriezenveen. A special thanks goes, of course, to the children and teenagers, ranging from two- to fourteen-year-olds, who participated in these experiments.
At this point, I would like to mention the invaluable contribution which has been made by Charlotte Koster. Soon after she joined the PIONIER project Reflections of Logical Patterns in Language Structure and Language Use, we were deeply involved in setting up and carrying out our Elicitation/Acting out experiment. I am most grateful to Charlotte for sharing her rich experience in linguistic experimentation, for her inspiring enthusiasm and a very pleasant collaboration throughout my AIO period. Also, I owe her many thanks for analyzing the experimental data and her determined effort with the statistics. Furthermore, she has been a thorough reader of papers and earlier versions of the manuscript and has made many detailed and useful comments. The final version of this thesis has profited immeasurably from her insights and suggestions.
In the same vein, I would like to thank the other members of the PIONIER project, to wit Jack Hoeksema - the project's supervisor and my co-promotor - Henny Klein, Hotze Rullmann, V¡ctor Sánchez-Valencia, Ton van der Wouden, and my promotor Frans Zwarts, for reading and commenting on earlier versions of the manuscript, and for being such fine and supportive colleagues. In addition, I would like to thank Hotze for his help, together with Charlotte, in setting up the paper-and-pencil grammaticality judgment experiment.
I am indebted to the Dutch Organization for Scientific Research (NWO) and the Faculty of Arts of the University of Groningen for financial support, which enabled me to attend summer schools in Buffalo (SIR 11-1084) and Lisbon, and to present papers at conferences in Albuquerque (SIR 11-1353), Bangor, Chicago, Freiburg, and Storrs. Also, I would like to thank the members of the thesis committee - Ger de Haan, Wolfgang Klein, and Anthony Sanford - for taking time to read the manuscript and to react so quickly, which made it possible to turn the long-planned provisional defence date at the end of August into a definite one.
Finally, I want to mention my sister Dieuwke and, once again, Charlotte. It means a lot to me that they will stand by me as paranymphs.
And ultimately, to save the best for last, I want to thank Gert for his love and support, and for being there always, even when he is far away.
On to next chapter: `The Structure and Use of Negative
Polarity Items'
Up
to general Table of Contents