Language Faculty Science This book explores how we can aspire to accumulate knowledge about the language faculty in line with Feynman’s “The test of all knowledge is experiment.” The two pillars of the proposed methodology for language faculty science are the internalist approach advocated by Chomsky and what Feynman calls the “Guess-Compute-Compare” method.T aking the internalist approach, the book is concerned with the I-language of an individual speaker. Adopting the Guess-Compute-Compare method, it aims at deducing definite predictions and comparing them with ex- perimental results. It offers a conceptual articulation of how we deduce definite predictions about the judgments of an individual speaker on the basis of universal and language-particular hypotheses and how we ob- tain experimental results precisely in accordance with such predictions. In pursuit of rigorous testability and reproducibility, the experimental demonstration in the book is supplemented by the accompanying web- site, which provides the details of all the Experiments discussed in the book. HAJIME HOJI is Associate Professor of Linguistics and East Asian Languages and Cultures at the University of Southern California. Language Faculty Science Hajime Hoji University Printing House, Cambridge CB2 8BS, United Kingdom Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781107046764 © Hajime Hoji 2015 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2015 A catalog record for this publication is available from the British Library Library of Congress Cataloging in Publication data Hoji, Hajime. Language faculty science / Hajime Hoji. pages cm Includes bibliographical references. ISBN 978-1-107-04676-4 (hardback) 1. Linguistics, Experimental. 2. Language and languages – Research. I. Title. P128.E94H65 2015 410.72 – dc23 2015012674 ISBN 978-1-107-04676-4 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. Contents Preface viii List of abbreviations xi 1 Introduction 1 2 The fundamental schematic asymmetry 9 2.1 Introduction 9 2.2 Types of judgments and types of predictions 10 2.3 Working with schemata 11 2.4 The fundamental asymmetry 15 2.5 Summary 22 3 Deducing definite and testable predictions 24 3.1 Introduction 24 3.2 Universal hypotheses and language-particular hypotheses 25 3.3 The model of the Computational System 27 3.4 Specific hypotheses making reference to c-command at LF 32 3.5 Characterizing BVA(α, β) 36 3.6 The model of judgment-making 39 3.7 Dependency interpretation 42 3.8 Guess-Deduce-Compare 44 3.9 Summary 45 4 Obtaining definite and categorical experimental results 47 4.1 Introduction 47 4.2 Main-Hypotheses and Sub-Hypotheses 50 4.3 Main-Experiment and Sub-Experiments 53 4.4 Two other factors and Sub-Experiments 54 vv Contents 4.5 Informant classification 60 4.6 Summary 61 5 Experiments in language faculty science 65 5.1 Introduction 65 5.2 Experimental design 68 5.3 Conducting experiments 77 5.4 Interpreting results (1): Basics 81 5.5 Interpreting results (2): Informant classification 98 5.6 Significance of invoking a dependency interpretation 108 5.7 Summary 109 6 Illustration: Experiments in English 112 6.1 Introduction 112 6.2 Main-Experiment on the structural hypotheses on FD: EPSA [31]-4 114 6.3 Sub-Experiments 123 6.4 Across-occasion reproducibility and informant classification 137 6.5 okSchema-based predictions 151 6.6 Concluding remarks 167 6.7 Appendix to Chapter 6: On the purpose of informant classification 170 7 Illustration: Experiments in Japanese 182 7.1 Introduction 182 7.2 Effects of instructions: Experiments on the lexical hypotheses 185 7.3 Sub-Experiment and informant classification for EPSA [10]: EPSA [3]-7 198 7.4 Main-Experiments on the LF-c-command condition on FD 216 7.5 The internalist approach and bridging hypotheses 268 7.6 Concluding remarks 285 7.7 Appendix to Chapter 7: Further illustration of the role of bridging hypotheses 291 8 Summary and concluding remarks 310 8.1 Summary 310 8.2 Concluding remarks 313 vi Contents Appendix I: List of key concepts, hypotheses, schemata and examples repeatedly referred to in the text 319 Appendix II: The accompanying website 323 Glossary 326 Endnotes 339 References 376 Author index 382 Word index (English) 384 Word index (Japanese) 385 Subject index 386 vii Preface This book grew out of a concern addressed in my paper “Falsifiabil- ity and repeatability in generative grammar: A case study of anaphora and scope dependency in Japanese” (2003a) and the earlier paper “Null object and sloppy identity in Japanese” (1998a). The concern regards how we can evaluate proposals in our research so as to be able to secure and accumulate as much progress as possible. This concern is directly related to questions such as what our object of inquiry is, what counts as evidence for or against our hypotheses about our subject matter, wheth- er and how we can deduce definite predictions from our hypotheses, how such predictions can be put to rigorous empirical test, and how we should interpret the results of our experiments. The work subsequent to those papers led to the book manuscript A foundation of generative grammar as an empirical science (2009) and eventually to this book. In the meantime, the first template for the gen- eral design of the on-line experiments, to be discussed in this book, was developed in 2004. The template has undergone changes reflecting vari- ous stages of the methodological articulation that has led to the proposal to be laid out and illustrated in this book, and a number of on-line ex- periments have been conducted over the years. The empirical illustration of the proposed methodology for language faculty science in this book is based on some of those experiments, the details of which are available at the accompanying website (http://www.gges.org/hojiCUP/). Concern with empirical rigor in research in linguistics has resulted in a recent explosion of experimental research; see the references in Schütze and Sprouse 2013 and other papers in Podesva and Sharma 2013. Such research – to the extent that it deals with informants – typically deals with a group of informants and analyzes the distribution of their reac- tions by statistical methods developed for, and utilized in, fields such as agricultural, social, and behavioral sciences. Many of the experiments to be discussed below also involve a large number of informants, as large as 200. But, for us, such a multiple-informant experiment is none other than a collection of single-informant experiments. Our predictions are vviiiiii