Linguistic Typology PDF - Jae Jung Song (Oxford University Press)

OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Linguistic Typology OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi O X F OR D T E X T B O O K S IN L I N G UI S T I C S PUBLISHED The Grammar of Words Pragmatics An Introduction to Linguistic Morphology     by Yan Huang by Geert Booij Compositional Semantics A Practical Introduction to Phonetics An Introduction to the   Syntax/Semantics Interface by J. C. Catford by Pauline Jacobson An Introduction to Multilingualism The History of Languages Language in a Changing World An Introduction by Florian Coulmas by Tore Janson Meaning in Use The Lexicon An Introduction to Semantics An Introduction and Pragmatics by Elisabetta Ježek   A Functional Discourse by Alan Cruse Grammar for English Natural Language Syntax by Evelien Keizer by Peter W. Culicover Diachronic Syntax Principles and Parameters by Ian Roberts An Introduction to Syntactic Theory Linguistic Typology by Peter W. Culicover by Jae Jung Song A Semantic Approach to English Cognitive Grammar Grammar An Introduction by R. M. W. Dixon by John R. Taylor Semantic Analysis Linguistic Categorization A Practical Introduction   by Cliff Goddard by John R. Taylor IN PREPARATION Translation Cognitive Grammar Theory and Practice An Introduction by Kirsten Malmkjaer   by John R. Taylor Grammaticalization by Heiko Narrog and Bernd Heine Speech Acts and Sentence Types in English by Peter Siemund OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Linguistic Typology JAE JUNG SONG 1 OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 3 Great Clarendon Street, Oxford,  , United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Jae Jung Song  The moral rights of the author have been asserted First Edition published in  Impression:  All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press  Madison Avenue, New York, NY , United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number:  ISBN –––– (hbk.) –––– (pbk.) Printed and bound by CPI Group (UK) Ltd, Croydon,   Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work. OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Contents Foreword xiii Preface xv List of abbreviations xvii Genealogical afﬁliations and geographical locations of languages cited in the book xix 1. Linguistic typology: An introductory overview  1.1 Introduction  1.2 What is linguistic typology?  1.3 Linguistic typology: a short history  1.4 Concluding remarks  Study questions  Further reading  PART I Foundations: theory and method  2. Unity and diversity in the world’s languages  2.1 Introduction  2.2 The connection between diversity and unity  2.3 The Principle of Uniformitarianism: a methodological frame of reference  2.4 When and where similarities count  2.5 Types of language universals and universal preferences  2.6 Concluding remarks  Study questions  Further reading  3. Typological analysis  3.1 Introduction  3.2 ‘Comparing apples with apples’: cross-linguistic comparability  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 3.3 Comparative concepts vs descriptive categories  3.4 Concluding remarks  Study questions  Further reading  4. Linguistic typology and other theoretical approaches  4.1 Introduction  4.2 Conceptual differences between LT and GG  4.3 Methodological differences between LT and GG  4.4 Optimality Theory: a derivative of GG with a twist  4.5 Concluding remarks  Study questions  Further reading  5. Language samples and sampling methods  5.1 Introduction  5.2 Some recalcitrant issues in language sampling  5.3 Types of language sample  5.4 Biases in language sampling and how to avoid them  5.5 Independence of cases  5.6 Sampling procedures  5.6.1 Proportional representation in language sampling  5.6.2 Independence of cases in proportional representation  5.6.3 Having the best of both worlds: structural variation across and within phyla  5.7 Testing independence of cases at a non-genealogical level  5.8 Typological distribution over time: was it different then from what it is now?  5.9 Concluding remarks  Study questions  Further reading  6. Data collection: Sources, issues, and problems  6.1 Introduction  vi OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 6.2 Grammatical descriptions or grammars  6.3 Texts  6.4 Online typological databases  6.5 Native speaker elicitation: direct or indirect  6.6 Levels of measurement and coding  6.7 Concluding remarks  Study questions  Further reading  7. Typological asymmetry: Economy, iconicity, and frequency  7.1 Introduction  7.2 Typological asymmetry  7.2.1 Formal coding  7.2.2 Grammatical behaviour  7.3 Economy and iconicity (in competition)  7.4 Typological asymmetry = frequency asymmetry?: iconicity vs frequency  7.5 Concluding remarks  Study questions  Further reading  8. Categories, prototypes, and semantic maps  8.1 Introduction  8.2 Classical vs prototype approach to categorization  8.3 Prototype category as a network of similarities  8.4 Prototype in grammar  8.5 Semantic maps: ‘geography of grammatical meaning’  8.6 Concluding remarks  Study questions  Further reading  vii OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS PART II Empirical dimensions  9. Phonological typology  9.1 Introduction  9.2 Segmental typology  9.2.1 Consonants  9.2.2 Vowels  9.2.3 Consonant–vowel ratios  9.3 Syllabic typology  9.4 Prosodic typology  9.4.1 Tone  9.4.2 Stress  9.5 Concluding remarks  Study questions  Further reading  10. Basic word order  10.1 Introduction  10.2 Some basic word order patterns  10.3 Early word order research  10.3.1 Greenberg’s seminal work and other ‘derivative’ works  10.3.2 Bringing word order inconsistencies to order  10.3.3 Distribution of the six basic clausal word orders  10.4 The OV–VO typology and the Branching Direction Theory  10.4.1 Back to the OV–VO typology  10.4.2 Inadequacy of Head-Dependent Theory  10.4.3 Branching Direction Theory (BDT)  10.4.4 Further thoughts on BDT  10.5 Word order variation and processing efﬁciency  10.5.1 Basic assumptions of EIC Theory  10.5.2 Left–right asymmetry in word order  10.6 Structural complexity and efﬁciency  10.6.1 Processing principles and processing domains  viii OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 10.6.2 Word order and word order correlations in Hawkins’s extended theory  10.7 Areal word order typology  10.7.1 Areal distribution of the six clausal word orders  10.7.2 Areal distribution of OV and VO  10.7.3 Areal distribution of OV/VO and NPo/PrN  10.7.4 Areal distribution of OV/VO and RelN/NRel  10.8 Concluding remarks  Study questions  Further reading  11. Case alignment  11.1 Introduction  11.2 S-alignment types  11.2.1 Nominative–accusative type  11.2.2 Ergative–absolutive type  11.2.3 Tripartite type  11.2.4 Double oblique type  11.2.5 Neutral type  11.3 Variations on S-alignment  11.3.1 Split-ergative type  11.3.2 Active–stative type  11.3.3 Hierarchical type  11.4 Distribution of the S-alignment types  11.5 Explaining various case alignment types  11.5.1 The discriminatory view of case marking  11.5.2 The indexing view of case marking  11.5.3 The discriminatory view vs the indexing view  11.6 The Nominal Hierarchy and the split-ergative system  11.7 Case marking as an interaction between attention ﬂow and viewpoint  11.8 P-alignment types  11.9 Distribution of the P-alignment types  11.10 Variations on P-alignment  11.11 S-alignment and P-alignment in combination  ix OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 11.12 Case alignment and word order  11.13 Concluding remarks  Study questions  Further reading  12. Grammatical relations  12.1 Introduction 339 12.2 Agreement 348 12.3 Relativization  12.4 Noun phrase ellipsis under coreference 359 12.5 Hierarchical nature of grammatical relations 364 12.6 Concluding remarks 367 Study questions  Further reading  13. Valency-manipulating operations  13.1 Introduction  13.2 Change of argument structure with change of valency  13.2.1 Passive  13.2.2 Antipassive  13.2.3 Noun incorporation  13.2.4 Applicative  13.2.5 Causative  13.3 Change of argument structure without change of valency  13.4 Concluding remarks  Study questions  Further reading  14. Person marking  14.1 Introduction  14.2 Morphological form: variation and distribution  14.2.1 Person marking and case alignment  14.2.2 Person marking and grammatical relations  x OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 14.3 Paradigmatic structure: towards a typology  14.3.1 Grouping A: no inclusive/exclusive with split group marking  14.3.2 Grouping B: no inclusive/exclusive with homophonous group marking  14.3.3 Grouping C: inclusive/exclusive with split group marking  14.3.4 Grouping D: inclusive/exclusive with homophonous group marking  14.3.5 Structural dependencies in paradigmatic structure  14.4 Concluding remarks  Study questions  Further reading  15. Evidentiality marking  15.1 Introduction  15.2 Morphological form of evidentiality marking  15.3 Semantic parameters of evidentiality  15.3.1 Visual  15.3.2 Non-visual sensory  15.3.3 Inference and assumption  15.3.4 Hearsay and quotative  15.3.5 Order of preference in evidentials  15.4 Typology of evidentiality systems  15.4.1 Two-term systems  15.4.2 Three-term systems  15.4.3 Four-term systems  15.4.4 Languages with more than four evidentiality distinctions  15.4.5 Multiple evidentiality subsystems  15.5 Evidentiality and other grammatical categories  xi OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi CONTENTS 15.6 Concluding remarks  Study questions  Further reading  References  Author index  Language index  Subject index  xii OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Foreword Jae Jung Song died peacefully at home in Dunedin, New Zealand on  April  while this book was being prepared for publication. Born on  January  to Tae Hyun Song and Jin Sun Yoo, Jae was one of four siblings, including his brother, Jae Tag. He is survived by his son, Kee, daughter-in-law Coral, and his grandson, whose birth Jae was eagerly anticipating, but sadly missed. Jae received a BA Honours and PhD from Monash University, Melbourne, Australia. After several years teaching in Australia and Singapore, Jae joined the University of Otago in . Jae was instru- mental in re-establishing and expanding Otago’s Linguistics Pro- gramme to include a TESOL minor and a Graduate Diploma for Second Language Teachers. During his career he served on the editorial boards of the Web Journal of Formal, Computational and Cognitive Linguistics, the Australian Journal of Linguistics, the International Review of Korean Studies, Language and Linguistics Compass, and the New Zealand Journal of Asian Studies. He was also Associate Editor of the journal Linguistic Typology. This book is Jae’s tenth authored or edited book, capping a career that also included publication of thirty- two book chapters and forty journal articles. Although Jae is most widely known for his extensive contributions to linguistic typology, particularly in the area of causatives, he also published on South and North Korean language policy. Many of Jae’s colleagues are unaware that Jae was a survivor of polio, which he contracted as a child in South Korea. At the height of his career he was struck by post-polio syndrome, a virtually untreatable re-emergence of the effects of polio. In the ten years I worked with Jae, he progressed through the support of a cane, one crutch, two crutches, a mobility chair and crutches, and ﬁnally a motorized wheelchair. Jae was determined not to allow the syndrome to compromise his work, and maintained a usual workload alongside his impressive publication record. OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi FOREWORD Jae will be remembered by colleagues, students, and friends for his work in typology, but he will be missed for his unwavering determin- ation, acute intelligence, and ready wit. Anne Feryok University of Otago Department of English and Linguistics Dunedin, New Zealand xiv OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Preface Not many authors are fortunate enough to have an opportunity to write more than one introductory volume in the same ﬁeld. I am one of the fortunate ones. My ﬁrst introductory book, Linguistic Typology: Morph- ology and Syntax (Harlow, ), appeared sixteen years ago, and it was in great need of updating, addition, and revision, in view of a substan- tial amount of developments—theoretical and methodological—and new data drawn from previously as well as newly documented languages. Two introductory books have recently appeared, namely Viveka Velupillai’s An Introduction to Linguistic Typology (Amsterdam, ) and Edith Moravcsik’s Introducing Language Typology (Cam- bridge, ). These two are excellent introductions to the ﬁeld and continue to inspire and inform students and professional linguists alike. They are, however, written primarily with students with minimal prior knowledge of linguistics in mind, and there is a need for a text that is pitched at an advanced level. I envisage that readers at this higher level of study are capable of, and keen on, grappling with theoretical or methodological issues that have characterized most of the recent devel- opments in linguistic typology. I hope that my new book will meet that need. My earlier book was organized largely by topic area, which makes it different from William Croft’s Typology and Universals (Cambridge, ), which is organized by theoretical concept. The present book aims to strike a balance between these two different styles of organiza- tion, which I believe is the best way to introduce my target audience to the ﬁeld. I would like to thank John Davey (the former Commissioning Editor at Oxford University Press) for inviting me to write a proposal for this book, and his successor Julia Steer for her forbearance and support, especially when the writing of the book ran into difﬁculty, more than once, because of my personal circumstances. I would also like to men- tion a number of linguists whose writings have inﬂuenced, or contrib- uted to, my thinking not only about the various issues addressed in this book but also about many others. There are too many to list them here but the following eminent scholars deserve a special mention: Balthasar OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi PREFACE Bickel, Barry Blake, Guglielmo Cinque, Bernard Comrie, William Croft, Matthew Dryer, the late Joseph Greenberg, John Haiman, Martin Haspelmath, John Hawkins, Edith Moravcsik, Johanna Nichols, and, last but not least, the late Anna Siewierska. The writing of this book has also beneﬁted from the invaluable input of students in the typology course that I have taught for the last twenty-ﬁve years at the University of Otago. One of the advantages of teaching, as many teachers will agree, is that one always tries to ﬁnd a better way to explain things to students, and, hopefully, this advantage has made its way into the book. There are also a number of people who have contributed non-linguistically to the birth of this book. I am indebted to Jaetag Song and Kee Ho Song for their words of encouragement. I cannot forget to thank Eleni Witehira for her unfailing support, even in times of her own personal difﬁculties, and also Kun Yong Lee and Sunae Bang (Sushi Station, Dunedin) for often cooking something off the menu for me, despite their heavy workload. I also wish to express my gratitude to the following people who have never heard about linguistic typology and have kindly pro- vided me with non-work-related assistance and support, without which the writing of this book would have been near impossible: Sarah Andrews, Lisa Begg, Dr Rene Cescon, Dr Martin Dvoracek, Hannah Fleming, Linda Grady, Lynda Hurren, Toni Johnston, Sandi Lorincz, Miriam Mackay, Michelle Mielnik, Zena Pigden, Ange Price, Dr Markus Renner, Nic Rogan, and Tammy Waugh. Lastly, I would like to thank Lisa Marr for her assistance in checking the references and preparing the indices. xvi OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi List of abbreviations  ﬁrst person DIR.EV direct evidential  second person DISTR distributive  third person DO direct object ABL ablative DR direct ABS absolutive DU dual ACC accusative EMPH emphasis ALL allative EP epenthetic ANIM animate formative ANTIP antipassive ERG ergative AOR aorist EXCL exclusive APPL applicative EZ ezāfe ART article F feminine ASP aspect FIRST ﬁrst-hand ASSUM assumed FOC focus ATTRIB attributive FUT future AUX auxiliary GEN genitive BEN benefactive IMM immediate CAUS causative IMP imperative CLF classiﬁer INAN inanimate CLT clitic INCH inchoative CNTR contrast INCL inclusive COM comitative INCMP incompletive COMP complementizer IND indicative COMPL completive INDEF indeﬁnite COP copula INF inferred DAT dative INFV inﬁnitive DEC declarative INST instrumental DEF deﬁnite INT intentional DEM demonstrative INTR intransitive DET determiner marker OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi LIST OF ABBREVIATIONS INTRATERM intraterminal PL plural INV inverse POSS possessive IO indirect object PREP preposition IPFV imperfective PRES present IRLS irrealis PROCOMP procomplement LINK linker PRON pronoun LOC locative PROX proximate M masculine PRTV partitive N neuter PST past NARR narrative PV preverb NCL noun class Q question NEG negative QUOT quotative NF non-ﬁnal marker REC.P recent past NFUT non-future REF referential NOM nominative REFL reﬂexive NOMN nominalization/ REL relative nominalizer REP reportative NON.FIRST non-ﬁrst-hand RETRO retrospective NONVIS non-visual SBJ subject NR.PST near past SBJV subjunctive NUM number SD sudden discovery OBJ/O object SENS.EV sensory evidential OBL oblique SEQ sequence OBV obviative SG singular OPT optative SUB subordinating PART participle SUBR subordinator PASS passive TES testimonial PAUC paucal TNS tense PC particle of TOP topic concord TR transitive marker PEJ pejorative TRANSLOC translocative PERF perfect V verb PF phonological ﬁller VIS visual PFV perfective VOL volitional xviii OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Genealogical afﬁliations and geographical locations of languages cited in the book Each language name is followed by (X; Y: Z), where X is the name of its genus, Y the name of its language family, and Z the name of the country where it is spoken, e.g. Bayso (Eastern Cushitic; Afro-Asiatic: Ethiopia). Languages that do not belong to any known language family are identiﬁed as ‘isolates’, e.g. Korean (isolate: Korea). If there is no distinction between genus and family (i.e. a single-genus language family) or between language and genus (i.e. a single-language genus), only one genealogical name is given, e.g. Pirahã (Muran: Brazil) or Palauan (Austronesian: Palau). Genealogical and geographical infor- mation is omitted where it is deemed redundant, that is, given else- where in the relevant section or chapter. For further genealogical and geographical information on the world’s languages, the reader may like to visit: () http://wals.info/languoid; () http://glottolog.org/glottolog/language; and () https://www.ethnologue.com/browse/names. OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1 Linguistic typology An introductory overview 1.1 Introduction  1.2 What is linguistic typology?  1.3 Linguistic typology: a short history  1.4 Concluding remarks  1.1 Introduction This chapter provides a brief description of linguistic typology by explaining what it is, what it aims to achieve in its study of language, what kinds of question it raises about the nature of language, and how it does what it does. The description will give the reader a snapshot of linguistic typology before they embark on reading the rest of this book. This is followed by a historical overview of linguistic typology. It is important to learn how linguistic typology emerged and developed over time into what it is today, as the reader will need to have a background understanding of how linguistic typology has evolved conceptually into one of the most important theoretical approaches to the study of language. 1.2 What is linguistic typology? The reader may have come across the kind of restaurant menu to be described here (more commonly found in Chinese than other res- taurants): à la carte menu items listed under different main-ingredient OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY headings, e.g. beef, pork, chicken, duck, seafood, and vegetables. For instance, all the dishes containing beef as their main ingredient are grouped together under the heading of ‘beef ’; the dishes with seafood as their main ingredient are all listed together under the heading of ‘seafood’, and so on. In other words, à la carte dishes are put into different types according to the main ingredients used. The basic principle employed in this way of organizing the menu is typological in that the main ingredient (e.g. beef) is used as the criterion for classifying or typologizing a given dish (e.g. beef with ginger and spring onion) in a particular way (e.g. a type of beef dish). This menu can thus be thought of as a typology of the dishes contained therein. The dishes may involve secondary ingredients as well, and any one of these ingredients may alternatively be selected as the criterion for classifying the dishes (e.g. noodle dishes vs rice dishes). Different ingredients (i.e. typological properties), when chosen as the basis for typologizing, may give rise to different typologies of the same dishes. To wit, typology (e.g. a particular way of organizing the menu items) is the classiﬁcation of a domain (e.g. the menu). The same principle can be applied to the study of languages. Thus, the world’s languages can be put into different types according to a particular linguistic property. The output of this exercise will be a typology of languages or, speciﬁcally, a linguistic typology of languages, since the typological property used is a linguistic one. For instance, take adpositions, which are linguistic elements expressing the semantic relations between the verb and its related noun phrases, e.g. location, time, instrument. The English adposition in in () expresses the loca- tion of Rachel’s action: Rachel’s action took place in, as opposed to near, the garden. () English (Germanic; Indo-European: UK) Rachel washed her car in the garden. The adposition in () is known as a preposition, because it appears right before the noun phrase that it associates itself with. In languages such as Korean, adpositions (e.g. -eyse ‘in’) are placed after the noun phrase, as in (), in which case they are known as postpositions: () Korean (isolate: Korea) kiho-ka anpang-eyse thipi-lul po-ass-ta Keeho-NOM main.bedroom-LOC TV-ACC see-PST-IND ‘Keeho watched TV in the main bedroom.’  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 2 W H A T I S L I N G U I S TI C T Y P O L O G Y? Using the two types of adposition, it is now possible to classify languages into two different types: prepositional and postpositional languages.1 For instance, English is identiﬁed as a prepositional language, and Korean as a postpositional language. The use of adpositions as the basis of typologizing languages is a simple illustration of how to put languages into different types according to a particular typological property (or, simply put, how to typologize languages). Other typological properties may be selected as the basis for classifying languages. For instance, the grammatical roles within the clause, e.g. subject and object, may be expressed on the head or the non-head (aka the dependent) (for detailed discussion, see Chapter ). Note that in the clause, the verb is the head and its arguments (or the noun phrases that the verb associates itself with) are dependents. In Swahili, for example, the grammatical roles of the noun phrases are all indicated on the verb. () Swahili (Bantoid; Niger-Congo: Tanzania) Ahmed a-li-m-piga Badru Ahmed he-PST-him-hit Badru ‘Ahmed hit Badru.’ In (), the verb (or the head) contains overt marking (i.e. a- for the subject noun phrase and m- for the object noun phrase). In contrast, the noun phrases (or the dependents), Ahmed and Badru, do not have any marking that indicates their grammatical roles. This type of mark- ing is known as head-marking (i.e. the head is marked). In Pitta-Pitta, the overt marking of the grammatical roles of the noun phrases in the clause appears on the dependents (= the noun phrases) themselves, as in: () Pitta-Pitta (Central Pama-Nyungan; Pama-Nyungan: Australia) kan̩a-lu matjumpa-n̪ a pit̪i-ka man-ERG roo-ACC kill-PST ‘The man killed the kangaroo.’ Not unexpectedly, the marking type exempliﬁed in () is referred to as dependent marking (i.e. the dependents are marked). Thus, according 1 Less commonly found are languages with both prepositions and postpositions, and even much less commonly attested are languages with inpositions, which appear inside the noun phrase. For details on inpositions, see Dryer (c). For the sake of simplicity, these languages will be ignored here.  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY to the locus of grammatical-role marking in the clause, languages are classiﬁed into the head-marking or dependent-marking type (e.g. Nichols and Bickel a). Swahili is a head-marking language, whereas Pitta-Pitta is a dependent-marking language.2 While linguistic typology shares the same conceptual basis of classi- ﬁcation with the Chinese menu described at the beginning of this section, that’s where their similarity ends. Unlike the menu, linguistic typology goes well beyond the business of classifying its objects of inquiry (i.e. languages). In fact, the more fruitful part of linguistic typology begins only after the establishment of a typological classiﬁca- tion of languages. For concreteness, take adpositions again. Broadly speaking, the world’s languages can be classiﬁed into either the prep- ositional or the postpositional type—note that, if a language employs both prepositions and postpositions, it may still be possible to say that the language is either predominantly prepositional or predominantly postpositional. If the investigation stops at this point, doing linguistic typology will be rather bland, if not boring, although that investigation in itself may be a rewarding experience. Still, it is hardly the most exciting thing to point to the mere existence of prepositional languages and postpositional languages in the world without saying any more. This is why Joseph Greenberg (–), regarded as the father of modern linguistic typology, wrote: ‘The assignment of a language to a particular typological class becomes merely an incidental by-product and is not of great interest for its own sake’ (a: ). So, what can be done to make it intellectually stimulating and challenging? This is what needs to be done. One must make every effort to analyse or interpret a given typological investigation; one must attempt to account for what the typological investigation has produced, a linguistic typology of X, with a view to ﬁnding out about the nature of human language. To wit, one must make sense of the ‘incidental by-product’. But how does one make sense of the linguistic typology of adpositions, for instance? This particular question can be formulated in such a way as to ascertain what factors motivate the use of prepositions or postpositions. Asked differently, do languages choose randomly or systematically between prepositions and postpositions? This question, in turn, may lead to 2 For the sake of simplicity, languages with both head- and dependent-marking or with neither or with marking appearing elsewhere in the clause will be ignored. For details, see Nichols (), and Nichols and Bickel (a, b, c).  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 3 L I N G U I S T I C T Y P O L O G Y : A S H O R T H I S T O R Y further investigation of adpositions. Indeed, linguistic typologists have discovered that the use of prepositions or postpositions is not an inde- pendent phenomenon. There is a remarkably strong correlation between the types of adpositions and basic clausal word order (speciﬁcally the position of the verb in the clause) to the extent that one can conﬁdently state that verb-initial languages (i.e. verb appearing before subject and object in the clause) have prepositions, and verb-ﬁnal languages (i.e. verb appearing after subject and object in the clause) postpositions.3 In other words, what seem to be two logically independent properties (i.e. adpositions and basic word order) correlate with each other to the point of statistical signiﬁcance (for further discussion, see Chapter ). Imagine that linguistic typologists do not venture beyond the mere classiﬁcation of languages into the two adpositional types—that is, beyond the level of an incidental by-product. They will not be able to discover the correlations in question, let alone to ﬁnd themselves in a position to pose a far more intriguing, albeit challenging, question: why do these correlations exist in the ﬁrst place? That is, what is it that makes verb-initial and verb-ﬁnal languages opt for prepositions and postpositions, respectively? Other related questions may include: what is so special about the position of the verb that has a bearing on the choice between prepositions and postpositions? Is there something shared by the position of the verb and the type of adposition that it correlates with? Asking questions such as these is important because they may lead to an understanding of the nature of language. Further- more, such an understanding may provide useful insights into the way the human mind shapes language or, more generally, how the human mind works (see Chapter ). 1.3 Linguistic typology: a short history The prospect of being able to predict the presence of one property (e.g. the types of adposition) on the basis of the presence of another property (e.g. the position of the verb in the clause) is a very attractive one. First and foremost, this elevates linguistic typology to a different level of investigation: so-called implicational typology. Even better will be the 3 Verb-initial and verb-ﬁnal order have subsequently been generalized to VO and OV order, respectively. Thus, VO order prefers prepositions and OV order postpositions.  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY presence of one property (X) implying the presence of multiple prop- erties (Y, Z, etc.). For instance, there may be properties other than the use of prepositions or postpositions that verb-initial or verb-ﬁnal order correlates with. Indeed, verb-initial order is also known to prefer Noun– Genitive order (e.g. the residence of the Prime Minister) to Genitive–Noun order (e.g. the Prime Minister’s residence), whereas verb-ﬁnal order also has a strong tendency to co-occur with Genitive–Noun order (for further discussion, see Chapter ). Generally speaking, it is preferable to predict as much as possible on the basis of as little as possible. This situation is akin to generating multiple units of energy (e.g. electricity) by using one unit of energy (e.g. biofuel). One cannot fail to see an economic dimension at work here. The idea of using the presence of one property to draw inferences about that of other properties is not something that has recently come to the attention of linguistic typologists. It has actually been around for a very long time. Over a hundred years ago, the German scholar, Georg von der Gabelentz (–), was so excited about this idea that he expressed his thinking somewhat overenthusiastically: Aber welcher Gewinn wäre es auch, wenn wir einer Sprache auf den Kopf zusagen dürften: Du hast das und das Einzelmerkmal, folglich hast du die und die weiteren Eigenschaften und den und den Gesammtcharakter!—wenn wir, wie es kühne Botaniker wohl versucht haben, aus dem Lindenblatte den Lindenbaum construiren könnten. (Gabelentz : ) [But what an achievement would it be were we to be able to confront a language and say to it: ‘you have such and such a speciﬁc property and hence also such and such further properties and such and such an overall character’—were we able, as daring botanists have indeed tried, to construct the entire lime tree from its leaf. (translation by Shibatani and Bynon b: )] This statement, often cited in textbooks on linguistic typology, is probably the earliest articulation of linguistic typology, as is currently understood and practised. Indeed, ‘[i]t would be difﬁcult to formulate the research programme of linguistic typology more succinctly’ than Gabelentz did (Plank : ). Moreover, it was Gabelentz (: ) who coined the term ‘typology’ when he continued to write on the same page: ‘Dürfte man ein ungeborenes Kind taufen, ich würde den Namen Typologie wählen’ [If one were permitted to christen an unborn child, I would choose the name typology]. The term ‘typology’, usually in conjunction with a modifying expression ‘linguistic’, is now used to refer to the subject matter of the present book. Though the term  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 3 L I N G U I S T I C T Y P O L O G Y : A S H O R T H I S T O R Y ‘typology’ began its life with Gabelentz’s coinage at the turn of the twentieth century, linguistic typology itself has a much longer history, dating back to the eighteenth or even the seventeenth century. Needless to say, it is not possible to say exactly when linguistic typology began as a scholarly approach to the study of language. Like other scholarly approaches or scientiﬁc disciplines, linguistic typology underwent what Ramat (: ) aptly calls ‘the incubation phase’, during which scholars pondered over problems or issues in general terms without realizing what they were thinking or writing about would eventually contribute to the development of an innovative way of investigating their objects of inquiry. One early example from this incubation phase of linguistic typology is the French abbot Gabriel Girard, who proposed a distinction between ‘analogous’ and ‘transpositive’ languages in the mid-eighteenth century: analogous languages have what we now call subject–verb–object order, while transpositive languages have different or even free word order. Moreover, analogous languages are claimed to mirror the ‘natural’ order of the thought process (the agent [= subject] exists, (s)he then does something [= verb], and, as a consequence, the patient [= object] is affected by the agent’s action). By contrast, trans- positive languages do not reﬂect the ‘natural’ order of the thought process, displaying different or altered word orders, that is, word orders other than subject–verb–object. Moreover, analogous languages gener- ally lack inﬂected forms, while transpositive languages are rich in inﬂected forms. Thus, a correlation can be drawn between these two typological properties: clausal word order and inﬂectional morphology. This correlation makes sense because in transpositive languages inﬂec- tional morphology encodes the distinction between subject and object (cf. head- and dependent-marking), thereby allowing the word order to deviate from the ‘natural’ order of the thought process. By contrast, the lack of inﬂectional morphology in analogous languages makes it important that the word order in these languages reﬂect the ‘natural’ order of the thought process. While it was more speculative than empirical, Girard’s study can perhaps be regarded one of the earliest instances of implicational typology since the presence of one property was utilized to draw inferences about the presence of another property. Put differently, Girard’s work was more than a classiﬁcation of lan- guages because of its attempt to explore a structural principle ‘that was capable of deeply characterizing languages from the point of view of ’ the nature of human language (Ramat : ). For this reason, Ramat  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY (: ) goes so far as to say that the French abbot and his immediate followers can be considered the true founders of linguistic typology, although Grafﬁ (: ) is of the view that Gabelentz should be regarded as the originator of linguistic typology, not least for his coinage of the term ‘typology’ and succinct description of linguistic typology. The reader is referred to Ramat () for further examples of the incubation phase in the seventeenth and eighteenth centuries of linguistic typology. Regardless of who was its founder, linguistic typology has been shaped over the centuries by many scholars, some of whom had lived before Gabelentz’s christening of the discipline. In common with other disciplines, the manner in which linguistic typology was conceptualized and developed was inﬂuenced by the intellectual milieu of particular historical periods in which it found itself. In the seventeenth century, the dominant intellectual movement was that of Rationalism (also known as the Enlightenment). The main driver of this intellectual movement was reason, instead of faith or tradition. Recall that the abbot Girard’s typology was based on the natural order of the thought process. This order, in turn, was claimed, through logical reasoning (or rationalism), to be the human mind’s way of thinking (or the natural ﬂow of thought, as it were). Needless to say, this was highly speculative, since it was not backed by empirical evidence; rationalism alone would sufﬁce for scholars of this intellectual background. Note that the focus of linguistic typology of this historical period was on the human mind’s natural (read: universal) way of thinking (i.e. unity), not so much on how different languages might or might not reﬂect the human mind’s way of thinking (i.e. diversity). In the eighteenth and nineteenth centuries, the dominant intellectual milieu, in reaction largely to Rationalism, changed to Romanticism: emphasis was placed on human emotion or experience instead of reason. Not unexpectedly, language was also regarded as a human experience. For instance, speakers of different languages may have different experiences, which, in turn, may explain why they speak different languages in the ﬁrst place. Scholars of this intellectual orien- tation went so far as to believe that language possessed an ‘inner form’. The inner form, in turn, was thought to be a manifestation of the spirit of the people (Volksgeist) who spoke the language (Greenberg : ch. ). In the words of Wilhelm von Humbolt (Finck ; Lehmann’s (c: ) translation), ‘[t]he characteristic intellectual features and  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 3 L I N G U I S T I C T Y P O L O G Y : A S H O R T H I S T O R Y the linguistic structure of a people stand in such intimacy of fusion with each other that if the one were presented the other would have to be completely derivable from it’. To wit, ‘each language [was] a distinct revelation of the spirit (Geist)’ (Greenberg : ). The inner form was also assumed to be reﬂected in ‘variation in grammatical mechan- isms employed in relating lexical concepts to each other [or relational meaning]’ (Shibatani and Bynon b: ). This point of view led to the emergence of August von Schlegel’s morphological typology, in which three basic strategies in the encoding of relational meaning were recognized: inﬂectional, agglutinative, and isolating—Wilhelm von Humboldt later added a fourth, incorporating, to Schlegel’s tripar- tite classiﬁcation.4 In inﬂectional languages, a single morpheme bears more than one meaning or represents more than one grammatical category. In agglutinative languages, a word may consist of more than one morpheme and may not be impervious to the conventional morpheme-by-morpheme analysis. In isolating languages, there is an equivalence relationship between word and morpheme with the effect that a morpheme can be taken to be a word or vice versa. In incorp- orating languages, the verb and its argument(s) may readily combine to create compound words. The unit of analysis in the morphological typology was undoubtedly the word, the structure of which ‘was seized upon as in some sense central to the attempt to characterize the language as a whole’ (Greenberg : ) so that ‘the description of the entire grammatical system [could] be annexed to an exact descrip- tion of the structure of the word in every language’ (Lewy : ). Unlike the focus of Rationalism, that of Romanticism was laid squarely on linguistic diversity, and this interest in linguistic diversity was heightened by the discovery of ‘exotic’ languages outside the Old World through European imperial expansion (e.g. missionaries and traders). Unfortunately, the morphological typology came to be inter- preted in highly subjective, evaluative terms with the effect that inﬂec- tional and isolating languages were regarded as the most and the least developed languages on an evolutionary scale, respectively. (This value- laden view relates to the development of so-called ethnopsychology 4 Comrie (: ), on the other hand, adopts ‘fusional’ in lieu of ‘inﬂectional’ because ‘both [agglutinative] and fusional languages, as opposed to isolating languages, have inﬂection, and it is... misleading to use a term based on (in)ﬂection to refer to one only of these two types’.  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY (Völkerpsychologie), which emphasized the relationship between language and thought.) The nineteenth century witnessed the emergence of a profound intellectual paradigm, namely Darwinism. As is well known, this is a profound theory of biological evolution, proposing that all species of organisms emerge and develop through the natural selection of vari- ations that individuals inherit, thereby enhancing their ability to com- pete, survive, and reproduce. This theory had an enormous impact on nineteenth-century scholars’ view of language, as Bopp (: ) avers: Languages must be regarded as organic bodies [organische Naturkörper], formed in accordance with deﬁnite laws; bearing within themselves an internal principle of life, they develop and they gradually die out, after... they discard, mutilate or misuse... components or forms which were originally signiﬁcant but which have gradually become relatively superﬁcial appendages. (as translated by Sampson : ) Under this Darwinian view, languages behave like biological species. (Try to read Bopp’s remarks, with ‘languages’ replaced with the names of animals or plants.) Thus, just as biological species do, languages emerge, develop into different varieties, compete with other varieties or languages, and cease to exist. This apparent similarity was so remark- able that the linguist’s language families, languages, dialects, and idio- lects were thought to correspond to the biologist’s genera, species, varieties, and individuals, respectively (Sampson : ). This evo- lutionary view of languages stood in stark contrast with the preceding Romanticist view, which was more humanities-based than scientiﬁc in that the latter regarded language as a subjective human experience, not as an entity to be described objectively as part of the natural world (Sampson : ). The Darwinian view of languages gave rise to a historical approach to the study of languages; not unexpectedly, the focus of this approach was based on the historical development or evolution of languages and the genealogical relationships among lan- guages. In point of fact, this approach was strongly upheld at the time as the only natural one for the study of language and survived into the early twentieth century, as Meillet () went so far as to claim that ‘the only true and useful classiﬁcation [of languages] is genetic [i.e. genealogical]’ (Greenberg : ). The historical approach was also augmented by mechanistic physics, another important scientiﬁc paradigm of the nineteenth century, albeit not as inﬂuential on the study of language as Darwinism. In  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 3 L I N G U I S T I C T Y P O L O G Y : A S H O R T H I S T O R Y mechanistic physics, all natural phenomena could be understood by way of simple, deterministic laws of force and motion ‘so that all future states of the world could in principle be inferred from a complete knowledge of its present state’ (Sampson : ). (This is what Bopp had in mind when he mentioned ‘deﬁnite laws’ in his remarks cited above.) When extended to the study of languages, what this entails is that the history of language can also be described by simple, deter- ministic laws of sound changes. Scholars who fully embraced the two paradigms in their study of languages were so-called Neogrammarians (Junggrammatiker) in the latter half of the nineteenth century. They held the view that it was possible to ‘found a genuine science of laws [read: a scientiﬁc theory of language] based on rigorous methods and the discovery of sound law rested on historical comparison’ (Greenberg : ). Their primary objective was to discover sound laws that could account for the historical development of languages from their ancestors and their genealogical relationships. Thus, it comes as no surprise that unrelated languages fell outside the purview of Neogram- marians’ research by default, with the lack of genealogical relatedness leaving nothing to be taken care of by sound laws in the ﬁrst place. Indeed, ‘[these scholars] saw no point to the comparison of unrelated languages’ (Greenberg : ); linguistic diversity, one of the primary pursuits of linguistic typology, was not on their research agenda. Needless to say, the exclusion of unrelated languages would consider- ably underrepresent the structural diversity of the world’s languages. Unfortunately, it was also assumed widely—albeit mistakenly—that typologically related languages were genealogically related as well. The dominant typological classiﬁcation of this period was none other than the morphological typology mentioned earlier in relation to Romanticism; the morphological typology was, in fact, almost syn- onymous with linguistic typology during this period. So much so that anyone questioning or denying the connection between genealogical and typological relatedness was regarded as ‘“heretical” (Ketzerei) at the time’ (Greenberg : ). Thus, linguistic typology was claimed to be of little or no use, because it could be subsumed, as it were, under the historical approach. Moreover, linguistic typology was not thought to be scientiﬁc enough as it could not reduce the nature of language to principles akin to Neogrammarian sound laws. Thus, linguistic typ- ology came to be largely ignored in the latter half of the nineteenth century, falling by the wayside into near oblivion.  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY The ﬁrst half of the twentieth century witnessed the rise in linguistics of structuralism (Saussure ), which brought about two major changes in the study of language. First, the focus shifted from diachrony, i.e. linguistic changes over time, to synchrony, i.e. the state of a language at a speciﬁc moment in time. Second, as the consequence of the ﬁrst change, the historical approach, dominant in the latter half of the nineteenth century, found itself on the wane (for a useful discussion, see Sampson : –).5 These changes notwithstand- ing, the indifference to linguistic typology continued well into the ﬁrst half of the twentieth century. Nonetheless, a small number of linguists in Europe as well as in the US maintained their interest in linguistic typology—in particular those associated with the Prague School in Europe.6 The European typologists were focused on what Greenberg (: –) calls the generalizing goal of linguistic typology, that is, ‘the discovery of law-like generalizations in languages by drawing bounds to the concept “possible human language”’. Because of their anthropological—one may say, even Humboldtian—orientation, the few American typologists (especially Edward Sapir : ch. ), in contrast, were interested in discovering structural characteristics of individual languages or what Greenberg (: ) refers to as the indi- vidualizing goal of linguistic typology. The reader may recall from the earlier discussion that the generalizing goal is somewhat akin to what seventeenth-century Rationalist philosophers of language had in mind (i.e. similarity), while the individualizing goal can be related to Roman- ticist scholars’ emphasis on individual languages as revelations of the spirit of their speakers (i.e. diversity). While the European typologists were well aware—at least far more so than the American typologists— of the connection between these two goals, especially in the form of implicational typology, it was not until the s that serious attempts began to be made in linguistic typology in order to bring to the fore the linking of these two goals. The catalyst in this regard came when the Second World War forced some European linguists, 5 Greenberg (: ) reports that the ﬁrst occurrence of the word ‘typology’ in the linguistics literature was in the theses presented by the Prague linguists to the First Congress of Slavic Philologists held in . 6 The majority of American structural linguists (e.g. Leonard Bloomﬁeld) were not interested in linguistic typology because of its previous association with ethnopsychology (i.e. language in relation to thought), which, to their minds, was not suitable for empirical study. They were instead in favour of behaviouristic psychology, which deals with observ- able behaviour, not the unobservable in the mind.  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi 1. 4 C O N C L U D I N G R E M A R K S including the Russian linguist Roman Jakobson (–), to ﬂee or migrate to the US, where they (re)introduced linguistic typology, in particular implicational typology, to American linguists. However, linguistic typology went largely ignored, as it had been in Europe in the preceding century, until Joseph Greenberg (b) used the concept of implicational typology to investigate word order correlations, as brieﬂy illustrated earlier in this chapter, in a relatively large number of languages, thereby elevating implicational typology in particular and linguistic typology in general to new heights. In other words, it was Greenberg who married up the two goals of linguistic typology more successfully than his predecessors and showed how to do implicational typology. In doing so, he ‘opened up a whole ﬁeld of [linguistic– typological] research’ (Hawkins : ), revitalizing or resurrecting linguistic typology as a viable scientiﬁc approach to the study of language. 1.4 Concluding remarks In this chapter, a brief description of how linguistic typology is carried out has been provided with special reference to what kinds of research question are raised, and how those questions are answered. Also provided is a historical overview of linguistic typology with special emphasis on how it has evolved conceptually over the centuries. Study questions 1. Rapanui (an Oceanic language spoken on Easter Island) is a verb-initial language in that the verb appears before both the subject and the object, while Turkmen (a Turkic language spoken in Turkmenistan) is a verb-ﬁnal language in that the verb follows both the subject and the object. In §., the strong correlation between the verb position and the use of preposi- tions or postpositions was alluded to. According to that correlation, one may be able to predict that Rapanui and Turkmen use prepositions and postpositions, respectively. Check these predictions against the available data on Rapanui and Turkmen (e.g. Du Feu ; Clark ). 2. X and Y invested $, and $, in business, respectively. After a year in business, X and Y made $, and $, out of their investments,  OUP CORRECTED PROOF – FINAL, 20/11/2017, SPi LINGUISTIC TYPOLOGY respectively. Whose business do you think is more proﬁtable, and why? Now, imagine you have two possible situations in your implicational typ- ology research. In situation , you can state with conﬁdence that if a language has properties A, B, and C, it also has property D. In situation , you can state with an equal amount of conﬁdence that if a language has property E, it also has properties D, F, and G. Which implicational statement do you think is more valuable, and why? How would you characterize the common ‘principle’ underlying proﬁtability in business and value in impli- cational typology? 3. Morphological typology, the most popular way of classifying languages in the period of Romanticism, was also used to evaluate the world’s languages in terms of evolution. For instance, isolating languages were regarded as the most primitive or the least developed on the evolutionary scale. Even then, however, it was clear that Chinese ﬂew right in the face of the evolutionary interpretation of the morphological typology. Why do you think this was so? 4. Ramat (: –) points to diachronic dynamics of language: languages shift from Type X to Type Y while retaining some features of Type X. What kinds of problem do you think this poses for the Gabelentzian attempt to construct the entire language (‘the entire lime tree’) from one of its struc- tural properties (‘its leaf ’)? Further reading Grafﬁ, G. (). ‘The Pioneers of Linguistic Typology: From Gabelentz to Greenberg’, in J. J. Song (ed.), The Oxford Handbook of Linguistic Typology. Oxford: Oxford University Press, –. Greenberg, J. H. (). Language Typology: A Historical and Analytic Overview. The Hague: Mouton. Horne, K. M. (). Language Typology: th and th Century Views. Washington, DC: Georgetown University Press. Ramat, P. (). ‘ Typological Comparison: Towards a Historical Perspective’, in M. Shibatani and T. Bynon (eds.), Approaches to Language Typology. Oxford: Oxford University Press, –. Ramat, P. (). ‘ The (Early) History of Linguistic Typology’, in J. J. Song (ed.), The Oxford Handbook of Linguistic Typology. Oxford: Oxford University Press, –. Song, J. J. (). ‘Linguistic Typology’, in K. Malmkjær (ed.), Routledge Linguistics Encyclopedia, rd edn. Abingdon: Routledge, –.  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi Part I Foundations: theory and method OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 2 Unity and diversity in the world’s languages 2.1 Introduction  2.2 The connection between diversity and unity  2.3 The Principle of Uniformitarianism: a methodological frame of reference  2.4 When and where similarities count  2.5 Types of language universals and universal preferences  2.6 Concluding remarks  2.1 Introduction References were made to Rationalism and Romanticism in Chapter , where the history of linguistic typology was discussed. These two schools of thought dominated the intellectual landscape in early mod- ern Europe, the former in the seventeenth century and the latter in the eighteenth century and the ﬁrst half of the nineteenth century. Not unexpectedly, they both had an enormous impact, in their respective historical periods, on how language was conceptualized as well as on how the study of language was approached and conducted. From these intellectual traditions emerged two overarching, albeit seemingly contra- dicting, research themes in the study of language: unity and diversity of human language. It was the unity of language that Rationalism high- lighted, with particular reference to the human mind’s natural (read: universal) way of thinking. Within Romanticism, in contrast, scholars drew attention to the diversity of human language, with language now OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi UNITY AND DIVERSITY IN THE WORLD’S LANGUAGES conceptualized as a manifestation of the spirit of its speakers—that is, different languages as reﬂections of different experiences. As will be shown in this chapter, these two themes continue to play an important role in modern linguistic typology, from Jakobson and Greenberg to the present day. While they are equally important concepts in linguistic typology (and indeed in any study of language for that matter), the pendulum has over the decades swung between the two themes. It was Jakobson and Greenberg who perceived the connection in linguistic typology between diversity and unity, while recognizing the search for the unity of human language as a loftier research goal than the diversity of human language. For instance, Jakobson (: ) remarked: ‘Lin- guistic typology is an inference from the science of languages [read: diversity] for the science of language [read: unity].’ In other words, it is through the diversity of the world’s languages that one can arrive at the discovery of the unity of human language. Similarly, Greenberg (: –) explains how ‘law-like generalizations in languages’ (read: unity) can be discovered on the basis of typological classiﬁcations (read: diversity) ‘by drawing bounds to the concept “possible human lan- guage”’. So much so that a typological classiﬁcation is reduced to ‘an incidental by-product [that] is not of great interest for its own sake’ (Greenberg a: ). Put differently, diversity is what ‘provides material for establishing [unity]’ (Sanders : ; Mallinson and Blake : ). In the early stage of modern linguistic typology, there- fore, the focus was on the discovery of language universals, which place the bounds on what is possible in human language. This ‘subjugation’, as it were, of diversity to unity in linguistic typology, however, has recently been rethought to the point of diversity attracting an increas- ingly greater amount of attention, especially in view of the realization that absolute language universals (i.e. all languages have X or there are no languages that lack X) are few and far between, and that it is, in point of fact, unrealistic to ﬁnd them (e.g. Dryer ). So much so that linguistic typology should instead strive to discover what is prob- able, rather than what is possible, in human language. When the focus is placed on discovering what is possible vs impossible in language, the concept of language universal is based on the unity of human lan- guage (i.e. what is impossible in language will not be found in any known languages). In contrast, when the focus is shifted to what is probable in language, the concept of language universal must be attenuated to the effect that while X may be found in the majority of  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 2.1 I N T R O D U C T I O N the world’s languages, there may also exist languages that possess Y or Z instead of X. In other words, attention is drawn to diversity, albeit with an eye to unity, since the majority of the world’s languages may still fall into the type of X to the effect that one can still speak of a universal preference for X. This fundamental change in perspective, in turn, highlights the structural variation in the world’s languages (e.g. not only X, but also Y and Z). More recently, the pendulum has swung even further in favour of diversity (e.g. Bickel ; Evans and Levinson ). In view of the unrealistic goal of discovering absolute language universals and the focus on ‘what is probable in human language’, it is now even more important to ﬁnd out not only why the majority of the world’s languages have X (i.e. universal preferences), but also why a small number of languages may have Y or Z in opposition to the universally preferred X—or, as Bickel (: ) puts it, ‘what’s where why?’. This has motivated linguistic typologists to step outside the domain of language and explore historical, social, and cultural (i.e. non-linguistic) factors to account for the existence of languages that do not behave like the majority of the world’s languages. For instance, languages with Verb–Object order (e.g. English, as in (a), where kissed is V and the man is O) have a very strong tendency to place a Relative clause (enclosed in brackets) after the head Noun (underlined) that it modiﬁes (e.g. English, as in (b)), i.e. NRel order. () English (Germanic; Indo-European: UK) a. The woman kissed the man b. The woman [who kissed the man] is my sister. Until recently, indeed, it was widely believed or assumed that VO languages always had NRel order. Then it was discovered (Dryer , e; but cf. Mallinson and Blake : ) that there are a small number of VO languages with RelN order in mainland China and Taiwan (i.e. Mandarin Chinese, Bai, and Amis), as exempliﬁed in (). () Chinese (Sinitic; Sino-Tibetan: China) a. tāmen tōu zìxíngchē PL steal bicycle ‘They steal bicycles.’ b. [wǒ gěi nǐ de] shū SG give SG LINK book ‘the book that I gave you’  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi UNITY AND DIVERSITY IN THE WORLD’S LANGUAGES In other words, what was believed to be an absolute language universal (i.e. VO & NRel) has turned out to be slightly less than an absolute language universal. Note, however, that there is still a clear universal preference or a near language universal (Dryer , e): VO languages almost always have NRel order—.% of  VO languages sampled in Dryer (e) have NRel order indeed. Linguistic typolo- gists have proposed that the uncommon combination of VO & RelN, as opposed to the global preponderance of VO & NRel, was due to Chinese having been inﬂuenced by OV languages with RelN order, located to the north of China (e.g. Mongolian and Tungus) and, subsequently, Chinese, now equipped with RelN order, inﬂuencing other VO languages that it came into contact with (Hashimoto ; Dryer ). Clearly, this structural similarity between Chinese and the neighbouring northern languages has its origins in historical, social, and/or cultural circumstances that brought them together and should thus be treated as contact-mediated. 2.2 The connection between diversity and unity The area of linguistic typology dealing with diversity investigates the structural variation in the world’s languages (i.e. everything that is attested), whereas the area of linguistic typology concerned with unity focuses on the discovery of language universals (i.e. what is possible) and universal preferences (i.e. what is probable). Language universals impose constraints or limits on structural variation within human language, while universal preferences delineate the dominance of some structural types over others. Typological investigation, in con- trast, is concerned with classifying languages into different structural types. Thus, ‘it may seem to the uninitiated something of a contradic- tion in terms to handle these apparently quite distinct areas of inves- tigation together’ (Mallinson and Blake : ; Greenberg : ). But, as may be gleaned from the preceding discussion, the contradic- tion is more apparent than real. The search for the unity in human language, in fact, builds crucially on the structural diversity in human language. This is because in order to discover language universals or universal preferences, what linguistic typologists ﬁrst need is typo- logical classiﬁcation to work on. With languages classiﬁed into different types, linguistic typologists may be able to discern patterns or regularities  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 2.2 T H E C O N NE C T I O N B E T W E E N DI V E R S I T Y AN D U N I T Y in the distribution of the types, for example, with some types being signiﬁcantly more common than others, or with one (or more) of the logically possible types completely unattested or only marginally attested in the world’s languages. Put simply, ‘[i]n order to understand Language, it is essential to understand languages’ (Comrie : ). This close relationship between language universals and universal preferences on the one hand, and linguistic typology on the other can be demonstrated by the preponderance of subject-initial order in the world’s languages. According to Dryer (a), for instance, nearly .% of his , sample languages have subject-initial word order, i.e. Subject–Object–Verb or Subject–Verb–Object. If the world’s languages—or at least a signiﬁcant portion of them—had not been surveyed in terms of all possible word orders (that is, not only subject-initial but also verb-initial and object-initial), this strong ten- dency would never have been brought to light in the ﬁrst place. To put it differently, the typological classiﬁcation of the world’s languages in terms of the six word orders—(i) Subject–Object–Verb, (ii) Subject– Verb–Object, (iii) Verb–Subject–Object, (iv) Verb–Object–Subject, (v) Object–Verb–Subject, and (vi) Object–Subject–Verb—is a sine qua non for the signiﬁcant generalization to be made about human language, that is, the majority of the world’s languages have subject-initial word order. Imagine the prospect of discovering the universal preference in question by examining only one language or even a handful of languages! This may be too extreme an example but the point could not be made more strongly. Further demonstration of the fruitful interaction between unity and diversity comes from another strong linguistic preference (albeit of an implicational nature): the presence of verb-initial order implying that of prepositions (see §.). This implicational statement also entails what is not possible in human language, namely the absence of verb- initial languages with postpositions (but see §.). Thus, the implica- tional statement does not only sanction verb-initial languages with prepositions as possible human languages but it also rules out verb- initial languages with postpositions as impossible human languages. Moreover, by making no negative claims about non-verb-initial lan- guages, the implicational statement refers indirectly to the other two logical possibilities, namely non-verb-initial languages either with pre- positions or with postpositions. In order to arrive at the actual formu- lation of this implicational statement, however, one must ﬁrst ascertain  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi UNITY AND DIVERSITY IN THE WORLD’S LANGUAGES which of the four logical possibilities is attested or unattested in the world’s languages. That can be achieved only on the basis of an initial typological classiﬁcation of the languages of the world in terms of basic word order as well as the distribution of prepositions and postpositions. To wit, the search for the unity in human language is conducted on the basis of the structural diversity in human language. It is not possible to carry out the former without the latter. The interaction between unity and diversity also highlights one of the virtues of formulating language universals or universal preferences on the basis of typological classiﬁcation. Typological classiﬁcation naturally calls for data from a wide range of languages (see Chapter  on how languages are selected or sampled for this purpose). Only by working with such a wide range of languages is one able to minimize the risk of elevating some of the least common structural properties to the status of language universals. This risk is more real than some linguists may be willing to admit because, when deciding to work with a small number of familiar or well-known languages (for whatever reasons), one is likely to deal with structural properties which may not in any real sense be representative of the world’s languages.1 For instance, use of relative pronouns is very common in European lan- guages but it has turned out to be a cross-linguistically infrequent phenomenon (Comrie : ; Comrie ; Comrie and Kuteva c). Therefore, universal claims about, or universal theories of, relative clauses which are put forth on the basis of these European languages alone should immediately be suspect. 2.3 The Principle of Uniformitarianism: a methodological frame of reference Strictly speaking, to claim that something is a language universal or a universal preference is patently premature, if not meretricious, when one thinks about the current level of documentation among the world’s languages. There are reported to be nearly , languages in the world 1 For instance, Chomsky (: ) avers: ‘I have not hesitated to propose a general principle of linguistic structure on the basis of observation of a single language... Assuming that the genetically determined language faculty is a common human possession, we may conclude that a principle of language is universal if we are led to postulate it as a “precondition” for the acquisition of a single language.’  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 2.3 T H E P R I N C I P L E O F U N I F O R M I T A R I A NI S M (Ethnologue, ), but unfortunately, ‘[l]ess than % of these languages have decent descriptions (full grammars and dictionaries)’ (Evans and Levinson : ). (Incidentally, this highlights the urgent need to document languages before they disappear into oblivion (e.g. Nettle and Romaine ; Grenoble and Whaley ; Evans ).) What this means is that language universals or universal pref- erences proposed so far are all based on less than % of the world’s languages, even if we assume, for the sake of argument, that all docu- mented languages have been taken into consideration for every lan- guage universal or universal preference that has ever been proposed. Over % of the world’s languages, which await (proper) documenta- tion, remain to be brought into the fold of linguistic research. Put simply, proposed language universals or universal preferences should never be understood to be established facts about languages but should instead be taken to be nothing more than observations or hypotheses based on what has been documented about the world’s languages. In other words, proposed language universals or universal preferences are in need of veriﬁcation against further data as they become available. A case in point is one of the language universals proposed by Greenberg (b): ‘Languages with dominant VSO order are always prepos- itional’ (emphasis added). Subsequent research (e.g. Dryer : ), however, has revealed that this is not entirely correct. There are languages, albeit in small numbers, that have both VSO order and postpositions, e.g. Cora (Corachol; Uto-Aztecan: Mexico), Ewe (Kwa; Niger-Congo: Togo and Ghana), Finnish (Finnic; Uralic: Finland), Guajajara (Tupi-Guaraní; Tupian: Brazil), Kashmiri (Indic; Indo- European: Pakistan and India), Northern Tepehuan (Tepiman; Uto-Aztecan: Mexico), and Waray (Warayic; Gunwinyguan: Australia). When Greenberg proposed this universal—and his other universals for that matter—in the early s, he never intended them to be facts about the world’s languages, but rather generalizations based on the data available to him at the time. Put differently, Greenberg’s language universals were merely hypotheses or ‘summar[ies] of data observed to date’ (Dryer : ). In this respect—this is an import- ant thing to remember—linguistic investigation is no different from any other scientiﬁc investigation. Every scientiﬁc statement, whether in physics or genetics, is simply a summary of data observed to date, or a generalization or a hypothesis based on that summary. More to the point, no scientiﬁc investigation can examine every instance of what it  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi UNITY AND DIVERSITY IN THE WORLD’S LANGUAGES is investigating. Typically, a small sample of the population under investigation is drawn for purposes of formulating a generalization or a hypothesis, which will continue to be tested against further data as they are collected (see §. for sampling in linguistic typology). This important point should not be lost sight of when language universals or universal preferences are spoken of. To wit, the validity of language universals and linguistic preferences can only be strengthened or weak- ened by means of further empirical testing. In order to arrive at a proper understanding of the nature of human language, it is not sufﬁcient to study only living languages—even if it is possible to include all living languages in one’s investigation. Ideally, linguistic typologists must study extinct languages as well. Indeed, linguistic typologists often study not only currently spoken languages but also extinct languages. This may perhaps strike one as odd, if not surprising, because one may expect typological classiﬁcation to be concerned only with the currently spoken languages of the world. One may be inclined to think that language universals represent con- straints or limits on structural variation within human language as it is, not as it was (or for that matter as it will be). Why do linguistic typologists also include extinct languages in their investigation? The assumption underlying this inclusion is what is generally known as the Principle of Uniformitarianism in linguistics (see Lass (: –, : –) for discussion thereof in the context of historical linguis- tics).2 Basically, what it means is that human languages of the past—or of the future for that matter—are not essentially different in qualitative terms from those of the present. This principle claims, therefore, that the fundamental properties of human language have remained invari- able over time. There are believed to be no differences in evolutionary terms between languages of the past—as far back as one can go and claim the existence of human languages—and those spoken today. In other words, human language of today is at the same level of evolution as that of, say, , years ago. Thus, in order to come to grips with the full structural diversity of human language, researchers must also investigate not only living languages but also extinct ones. Imagine a possible situation 2 This principle was ﬁrst introduced into the study of language by Neogrammarians from the natural science thesis of Hutton and Lyell. Karl Brugmann is quoted as saying (Collinge : ): ‘[t]he psychological and physiological nature of man as speaker must have been essentially identical at all epochs.’  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi 2.3 T H E P R I N C I P L E O F U N I F O R M I T A R I A NI S M in which particular structural types unattested in living languages happen to have existed only in languages that are no longer spoken. The Principle of Uniformitarianism is, of course, something that has never been subjected to empirical veriﬁcation and cannot be put to the test for obvious reasons; one simply cannot go back in time and examine languages spoken, say, , years ago to see whether or not they were qualitatively the same as those of today. Nor is there any logical reason why the principle should be correct. Nonetheless it plays an important role in linguistic typology (and equally in historical linguistics). The primary aim of linguistic typology is to discover universal properties or preferences of human language. If human languages were spoken , years ago, then, these languages must also be included in any typological study, which is utterly impossible. To get an idea of the linguistic diversity in the past, one can refer to Evans and Levinson (: ), who suggest that  years ago, before European colonization began, there were probably twice as many languages as there are now, and to Pagel (: ), who claims that over half a million languages have ever been spoken on this planet, if humans began talking , years ago and languages evolved at a rate of one per  years. In the absence of the Principle of Uniformitarian- ism, then, no typological analysis will be possible or, more accurately, complete simply because it is impossible to ‘recover’ all unrecorded, extinct languages from oblivion. With the Principle of Uniformitarianism in place, however, linguistic typologists can examine languages spoken today and, if and where possible, extinct but documented languages as well and can still make statements or generalizations about the nature of human language. Similar comments can also be made of languages of the future. Since it is expected that they will also be human languages, any typological study must in principle include them as well, which is out of the question. But the Principle of Uniformitarianism also works in the opposite direction of time from the present, thereby allowing linguistic typologists to extend to languages of the future what uni- versal properties or preferences that they may have discovered on the basis of currently available data. While the complete structural diver- sity of human language will be impossible to capture, what is true of today’s human language can be assumed to be true of yesterday’s and tomorrow’s human language. After all, under the Principle of Uniformitarianism the nature of human language is assumed not to change qualitatively over time.  OUP CORRECTED PROOF – FINAL, 27/11/2017, SPi UNITY AND DIVERSITY IN THE WORLD’S LANGUAGES There are also practical reasons why the Principle of Uniformitar- ianism is adhered to in linguistic typology. Without this principle, languages must be seen to evolve constantly as time passes by. But if languages were evolving through time, and were conceived of as being at different stages of linguistic evolution, grammatical

Linguistic Typology PDF - Jae Jung Song (Oxford University Press)

Document Details

Tags

Related

Summary

Full Transcript