Dative External Possessors in Early English (Oxford Studies in Diachronic and Historical Linguistics) 9780198832263, 0198832265

Citation preview

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

Dative External Possessors in Early English

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

OXFORD STUDIES IN DIACHRONIC AND HISTORICAL LINGUISTICS General Editors Adam Ledgeway and Ian Roberts, University of Cambridge Advisory Editors Cynthia L. Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Cambridge      31 Grammaticalization from a Typological Perspective Edited by Heiko Narrog and Bernd Heine 32 Negation and Nonveridicality in the History of Greek Katerina Chatzopoulou 33 Indefinites between Latin and Romance Chiara Gianollo 34 Verb Second in Medieval Romance Sam Wolfe 35 Referential Null Subjects in Early English Kristian A. Rusten 36 Word Order and Parameter Change in Romanian A Comparative Romance Perspective Alexandru Nicolae 37 Cycles in Language Change Edited by Miriam Bouzouita, Anne Breitbarth, Lieven Danckaert, and Elisabeth Witzenhausen 38 Palatal Sound Change in the Romance Languages Diachronic and Synchronic Perspectives André Zampaulo 39 Dative External Possessors in Early English Cynthia L. Allen For a complete list of titles published and in preparation for the series, see pp. 285–8

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

Dative External Possessors in Early English CYNTHIA L. ALLEN

1

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3

Great Clarendon Street, Oxford, OX2 6DP, 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 © Cynthia L. Allen 2019 The moral rights of the author have been asserted First Edition published in 2019 Impression: 1 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 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019936324 ISBN 978–0–19–883226–3 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. 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, 3/9/2019, SPi

Contents Series preface Preface List of tables List of abbreviations A note on glossing

1. Introduction 1.1 1.2 1.3 1.4 1.5 1.6

Dative external possessors Theoretical issues Typological considerations Previous studies of external possession in early English Corpus and historical scope Organization of the book

2. Dative case in Old English: An overview 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Dative and instrumental Verbal objects Complements of adjectives and nouns Copulas, ‘impersonal’ constructions, and ‘extended existence’ Dative + PP with copulas Dative + to with other verbs ‘Free’ and ‘adverbial’ datives ‘Impersonal’ constructions with datives with lexical verbs Conclusion

3. Investigating dative external possessors in the history of English 3.1 3.2 3.3 3.4

Introduction The corpus Methodology Summary

4. Body and dative external possessors in Old English 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

Introduction Results: object possessa Results: subject possessa with lexical verbs Subject possessa with adjectival predicates Comparing subject and object possessa The question of Latin influence Direct arguments: conclusions Results: objects of prepositions (PObjs) Conclusions on DEPs of body possessa in OE

vii ix xiii xv xvii 1 3 4 12 14 22 23 25 25 27 31 32 36 37 38 43 44 45 45 47 50 56 58 58 59 71 81 84 88 93 94 99

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

vi



5. Early changes in English 5.1 5.2 5.3 5.4

Introduction A change from Germanic? A change within OE? Range and frequency of DEPs Conclusions on early changes

6. Mind and dative external possessors in Old English 6.1 6.2 6.3 6.4

Introduction Direct arguments PObjs Conclusions: body and mind

7. External possessors in Early Middle English 7.1 Introduction 7.2 The transition to Middle English 7.3 The pre-m1 period 7.4 Summary and comparison with late OE 7.5 M1 and beyond 7.6 The dative case in EME 7.7 Searching for ‘dative case’ and DEPs in Middle English 7.8 Direct arguments: body 7.9 Prepositional phrases 7.10 Results: mind 7.11 Conclusion: from OE to ME

8. Changes and explanations 8.1 8.2 8.3 8.4

Introduction Internal explanations External explanations Conclusions

101 101 101 109 120 121 121 122 134 148 150 150 151 151 162 164 165 175 177 183 193 195 199 199 200 214 222

9. Conclusion

229

Appendix A: Corpus Appendix B: Vocabulary lists Appendix C: Notes on searches

237 247 259

References Index

271 279

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

Series preface Modern diachronic linguistics has important contacts with other subdisciplines, notably first-language acquisition, learnability theory, computational linguistics, sociolinguistics, and the traditional philological study of texts. It is now recognized in the wider field that diachronic linguistics can make a novel contribution to linguistic theory, to historical linguistics, and arguably to cognitive science more widely. This series provides a forum for work in both diachronic and historical linguistics, including work on change in grammar, sound, and meaning within and across languages; synchronic studies of languages in the past; and descriptive histories of one or more languages. It is intended to reflect and encourage the links between these subjects and fields such as those mentioned above. The goal of the series is to publish high-quality monographs and collections of papers in diachronic linguistics generally, i.e. studies focusing on change in linguistic structure, and/or change in grammars, which are also intended to make a contribution to linguistic theory, by developing and adopting a current theoretical model, by raising wider questions concerning the nature of language change, or by developing theoretical connections with other areas of linguistics and cognitive science as listed above. There is no bias towards a particular language or language family, or towards a particular theoretical framework; work in all theoretical frameworks, and work based on the descriptive tradition of language typology, as well as quantitatively based work using theoretical ideas, also feature in the series. Adam Ledgeway and Ian Roberts University of Cambridge

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

Preface This book developed from a combination of my long-standing interest in the relationship between the loss of inflection and syntactic change in English and a more recent interest in what is known as the ‘Celtic Hypothesis’, that is, the hypothesis that speakers of Brythonic Celtic languages influenced Old English significantly as they shifted from their mother tongue to the language of their conquerors. Because so many proposals for explaining syntactic change in English by modern linguists have had to rely on an what seem to me inadequate empirical bases to support their hypotheses about how many changes came about, my own work has tended strongly towards strengthening the empirical base linguists can draw on to evaluate possible accounts for some syntactic changes that I have found particularly interesting, especially ones for which the deflexion or loss of inflection has widely been regarded as being the prime motivator of the change. The loss of dative external possessors is one such change, and is often presented as one of a number of changes related to deflexion that set Modern English apart from ‘more typical’ Germanic languages. A closer and more systematic examination of the facts shows that this account does not hold up. The loss of the dative/accusative category distinction in English is the leading traditional explanation for the change being studied in this book and one that has been incorporated into more modern treatments of case assignment. However, more recently, the hypothesis that English lost the dative external possessor construction because Brythonic Celtic, unlike most European languages, did not have it, has gained ground. While this idea has actually been around for a long time, it is only fairly recently that a re-evaluation of the survival of Celtic speakers after the Anglo-Saxon Invasions has combined with developments in the study of language contact phenomena to make the idea of a strong Celtic influence on the syntax of Old English more mainstream. In a long career, I have sometimes engaged with the question of foreign (Latin, French, or Scandinavian) influence on syntactic change in early English, but for the most part the question of contact phenomena have taken a back seat to documenting the syntax of Old and Early Middle English. While aware that written language is never just the same as spoken language, I have taken the view that while the written language is presumably more

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

x



conservative in its morphology and syntax than colloquial spoken language, nevertheless it probably reflects the relative timing of different changes reasonably well, even if the absolute timing of the changes was earlier in spoken language. Furthermore, the texts generally show the sort of systematicity in their morphosyntax that we would not expect of an artificial system that the scribes had learned formally, that is, a system that was natural to the scribe, even if it might be that of a more formal register. It seems to me, however, that the Celtic Hypothesis presents a serious challenge for the study of diachronic English syntax if we adopt the common assumption by proponents of that hypothesis that the spoken English of a majority of the population was suppressed in writing, which was controlled by a Germanic elite, until the Norman Invasion removed that control over the writing of English. If it is true that the syntax of the Early Middle English texts is not a development of the syntax underlying the language of Old English texts but rather the development of the syntax of a language that was not recorded, then our accounts of syntactic changes between these periods and proposed explanations for them are misguided—instead of looking at the history of a language in two periods, we would be looking at one language in one period and another one in a later period. The severity of the problem depends, of course, on how great the difference between spoken and written language was, and this is not something we can know. My own view is that the Early Middle English texts do not support the idea of a spoken language that burst out into writing once the shackles of the West Saxon Schriftsprache had been cast off. The idea of a Celticized ‘Brittonic’ English that was radically different from the Old English of speakers of Anglo-Saxon descent is at any rate purely hypothetical, and as discussed in Chapter 8 especially, I do not find the arguments that have been proffered for this view convincing. It is plausible enough that at least some Celts would speak an English that was very different from that of their conquerors, but it seems to me that the main reason for assuming that this Brittonic English was the spoken English of the majority of speakers and did not find its way into writing until Middle English is to provide an explanation for why some features of English that have been attributed to Celtic influence, for example, periphrastic do, do not make their appearance in the texts until the thirteenth century or later. The loss of dative external possessors is one of the changes that has sometimes been identified as something that may have taken place in speech but only became apparent in Early Middle English. I hope that I manage to convince the reader in this book that this view of a sudden disappearance of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi



xi

the construction from the texts, resulting from a new type of language, is incorrect. Internal possessors were always in variation with dative external possessors and were certainly not suppressed as Celticisms in writing. This does not mean, however, that Celtic language learners could not have had a role in the initial decline of the construction by narrowing its range and using internal possessors with a higher frequency than Anglo-Saxons. It does mean, though, that with this particular pair of constructions we can be confident that we are looking at a change to grammatical possibilities that were shared across the English-speaking community generally. It is a pleasure to thank those who have played a role in making this book possible. My thanks go to the Australian National University for appointing me an Emeritus Fellow on my retirement, facilitating my interaction with colleagues and giving me continued access to library and other facilities vital to my research. This book, as I believe is not unusual with monographs, started as an attempted journal paper, and I gratefully acknowledge helpful comments by Theo Vennemann and anonymous reviewers on drafts of that paper. It eventually became clear to me that the paper I was trying to write contained both too little and too much for a journal article, and these comments were helpful in identifying the matters that a paper could not give enough detail to and discussions that were unclear to the point of misunderstandings by the reviewers. My thanks also go to the anonymous reader who read the first draft of my book manuscript for OUP and caused substantial reorganization of some material. I hope that the book is the better for this feedback, but of course the usual statement about my own responsibility for any shortcomings apply. In her superb presentation of the grammar of the life of St. Juliana and other Dialect AB texts in her edition of Þe Liflade ant te passion of Seinte Juliene, S.R.T.O. d’Ardenne begins with an acknowledgement of the debt that all workers in the field of Middle English owe to each other and their predecessors. I would like here to make a similar acknowledgement to all scholars (including d’Ardenne) and researchers who have been so important to my own work. I would in particular like to thank the compilers of the electronic corpora that have provided the data in this book, and also the makers of the Helsinki Corpus, which led to the development of parsed corpora of earlier periods of English.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

List of tables

4.3 DEPs of body object possessa in LWS texts

48 61 62 62

4.4 IPs vs DEPs of body object possessa with affecting verbs in early OE texts

64

3.1 Text categories used in this book 4.1 DEPs of body object possessa in early OE texts 4.2 DEPs of body object possessa in General OE texts

4.5 IPs vs DEPs of body object possessa with affecting verbs in General OE texts

4.12 DEPs of body subject possessa in LWS texts, lexical verbs

65 66 66 68 70 71 72 73

4.13 IPs vs DEPs of body subject possessa with affecting verbs in early OE texts

76

4.6 IPs vs DEPs of body object possessa with affecting verbs in LWS texts 4.7 Number of DEPs of objects of verbs not fitting the affecting verb criteria 4.8 Overview of IPs vs DEPs of body object possessa with affecting verbs 4.9 IPs vs DEPs of object possessa in OE text types, DEP verbs 4.10 DEPs of body subject possessa in early OE texts, lexical verbs 4.11 DEPs of body subject possessa in General OE texts, lexical verbs

4.14 IPs vs DEPs of body subject possessa with affecting verbs in General OE texts 4.15 IPs vs DEPs of body subject possessa with affecting verbs in LWS texts 4.16 Summary of IP vs DEPs of body subject possessa with affecting verbs in OE text types

77 78

4.17 IPs vs DEPs of body subject possessa with DEP verbs in OE text types

78 80

4.18 IPs vs DEPs of body object and subject possessa with affecting verbs, all OE text types

84

4.19 Pronouns and nouns in IPs and DEPs of body subject possessa with affecting verbs in General OE texts 4.20 DEPs vs IPs of restricted body PObjs in selected OE texts, lexical verbs

87 98

5.1 IPs vs DEPs of direct argument body possessa with affecting verbs in texts of EWS and LWS composition

110

5.2 IPs vs DEPs of direct argument body possessa with affecting verbs in early and late OE prose texts

111

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

xiv

  

5.3 IPs and DEPs of direct argument body possessa with affecting verbs in OE poetic texts 5.4 DEPs vs IPs of restricted body PObjs with lexical verbs in OE poetry and selected OE prose text types 5.5 Comparison of DEPs in combinations of P + PObj 5.6 DEPs and IPs with on/to hand, EWS vs LWS 6.1 Datives and IPs of mind nominative possessa with copulas in OE poetry 6.2 IPs and DEPs of mind subject possessa in OE poetry, lexical verbs 6.3 IPs and DEPs with selected mind PObjs of lexical verbs in OE poetry 6.4 Copula dative constructions with selected mind PObjs in OE poetry 6.5 DEPs of mind PObjs with lexical verbs in OE prose 6.6 Dative possessors of selected mind PObjs in copular constructions in OE prose 7.1 DEPs (all) vs IPs (highly affecting verbs) with body possessa, m1 7.2 IPs and DEPs with selected body PObj possessa, 1150–1250 7.3 Comparison of OE and m1 body object possessa with affecting verbs 7.4 Comparison of OE and m1 body subject possessa, affecting verbs 8.1 IPs and DEPs in EME case-rich texts 8.2 Stages in the loss of the DEP in English A.1 OE prose texts used, by text type A.2 YCOE texts used for pre-m1 period A.3 Additional EME texts not included in the PPCME2 B.1 Body lemmas, OE B.2 Mind lemmas, Old English B.3 Body lemmas, m1 and m2 B.4 Mind lemmas, m1 and m2 B.5 Intransitive affecting/DEP verbs, OE B.6 Transitive affecting/DEP verbs, OE B.7 Affecting verbs, m1 and m2 B.8 Selected prepositions and body nouns used in OE body PObj searches B.9 Selected prepositions for mind PObj searches, OE B.10 Prepositions used in searches for PObj IP vs DEP comparisons in m1 and m2 C.1 Selected OE texts for IP and DEP comparisons for PObjs within three text types

113 114 115 117 125 126 137 140 141 144 177 185 195 195 207 226 238 243 245 248 250 250 252 253 254 256 258 258 258 266

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

List of abbreviations Abbreviations used in glosses ACC DAT DEF FEM FUT GEN IMP INST MASC NOM OBJ OPT PL SG SUBJ

accusative dative definite feminine future genitive imperative instrumental masculine nominative object case (for ME) optative (=subjunctive) plural singular subject

Abbreviations used in text ASC BL DEP DOE DP EME EP IO IP LAEME m1 m2 ME N NP OBJdat

Anglo-Saxon Chronicle Brook and Leslie’s edition of Laʒamon’s Brut dative external possessor Dictionary of Old English determiner phrase Early Middle English external possessor indirect object internal possessor Linguistic Atlas of Early Middle English First Middle English period in the electronic corpora (1150–1250) Second Middle English period in the electronic corpora (1250–1350) Middle English noun noun phrase dative object

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

xvi

  

OED PDE PObj Poss Det PPCME2 YCOE

Oxford English Dictionary Present Day English object of preposition possessive (followed by) determiner Penn-Helsinki Parsed Corpus of Middle English 2 York-Toronto-Helsinki Parsed Corpus of Old English Prose

OUP CORRECTED PROOF – FINAL, 13/9/2019, SPi

A note on glossing The glosses of the examples from earlier English and other languages are loosely based on the Leipzig Glossing Rules, available at https://www.eva. mpg.de/lingua/resources/glossing-rules.php. However, the glosses in this book are not intended to provide full morphological information. So while I have glossed every word in my Old English examples, I have generally provided only such morphological information as I believe will be useful to at least some readers in understanding the syntax of a sentence or a morphological point I am making, such as the ambiguity of some case inflections. The glosses do not usually provide a breakdown of the morphology of a verb; for example, I have glossed the plural past tense form forcurfon simply as ‘cut’ in an example where the fuller translation ‘the Romans cut Pope Leo’s tongue out’ gives as much information as the reader needs to understand the sentence, tense and number being irrelevant to the point being illustrated. I have treated the indicative mood as the default one, but have generally glossed subjunctives, since these often involve the use of a modal verb such as may in the Modern English translation. The OE subjunctive is an optative from an Indo-European point of view, and to avoid confusion with , which I have reserved for the grammatical relation of subject in the Middle English period in examples in which no distinctive nominative case marking is used, I have glossed subjunctives as . The emphasis on case marking in this book means that elements of noun phrases are more extensively glossed than are verbs, and some more detailed comments on this glossing are in order. I have not glossed gender except when I considered it important to the discussion, and case is only indicated regularly for possessors and possessa in the roles of subject or object. Since English still distinguishes number in nouns, I have generally used Modern English singular or plural forms in the words of the glosses, combining these with any specification for case, e.g. men is glossed ‘man:’ and mannum is ‘men:.’. I have treated the singular, being the most common, as the default, and have only indicated number in the glosses of plurals. In presenting examples cited by other people, I have adapted the glosses offered in the source when appropriate. There was a good deal of syncretism already in early Old English in inflections for case, number, and gender. Unless such syncretism raises the possibility of more than one morphosyntactic analysis, in which case there will

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

xviii

   

be a discussion in the text, I have not generally indicated the ambiguity. For example, the feminine pronoun hire was formally ambiguous between the genitive and the dative case in Old English, and this ambiguity makes the analysis of some sentences ambiguous between having a dative external possessor and an internal (genitive) possessor. In such examples, I have indicated the ambiguity. In some sentences, only one case or the other is a serious possibility, and I have simply glossed the pronoun as dative or genitive. A final note is necessary on the glossing of pronouns. I have generally used Modern English pronouns in the glosses; the gloss ‘his’ for his is adequate and generally easier for a reader to follow than one that breaks the pronoun down into third person, masculine, singular, and genitive. Complications arise from the fact that the third person pronouns his ‘his, its’, hire ‘her’, and heora ‘their’ were different from the first and second persons, which did not use the simple genitive form to express possession, where what are traditionally called ‘possessive adjectives’ were used. These possessives declined like strong adjectives, agreeing with the noun they modified in case, gender, and number; e.g. minum was a masculine or neuter dative singular or all-gender plural form of min ‘my’. I have not considered it necessary to convey this information in all the glosses, as when the possessive modifies the object of a preposition, which I have glossed for case only when that seemed helpful, usually glossing minum simply as ‘my’. The third person singular or plural reflexive sin, mainly found in poetry, was also an inflecting possessive and is treated here in the same way as the first and second person forms. On the other hand, direct objects often needed to be glossed to show their accusative case, and in this role I have for example glossed the direct object hige þinne as ‘mind: thy:’. I have not glossed possessives for gender or number, and so have not included the information that the possessive just cited is unambiguously masculine and singular also. With non-possessive forms of the pronouns I have also used Modern English forms, tagged for case only when the distinctions marked in the earlier and later stages of the language do not match. So, for example, I gloss hine as ‘him:’and him as either him: or ‘them:’, as the meaning of the sentence requires. In glossing Early Middle English sentences, I have glossed him as dative only when a clear dative/accusative category distinction exists in the text in question and is relevant to the discussion; otherwise him simply as ‘him’. Similarly, hire is glossed for case only when relevant to the discussion, e.g. when an example involves dative/genitive ambiguity. In a couple of Middle English examples, I have used the gloss you: when the Middle English form is clearly an object form but the accusative/dative distinction no longer exists.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1 Introduction

This book explores the loss of a construction in English which is widespread in European languages and was found in earlier English. The construction is illustrated for German in (1.1): (1.1) Die Mutter wusch dem Kind die Haare the mother washed the: child the: hairs: ‘the mother washed the child’s hair’ (Haspelmath 1999: 109, ex. 1) In this example, dem Kind is the possessor of die Haare, but instead of appearing within the NP containing the possessum, it is an NP at the sentence level, marked with dative case. In Old English (OE) texts, parallel examples are common: (1.2) Gif þu þæt þurhteon ne mæge scearpa him If you that carry-out not may scarify him: þa scancan the: legs: ‘If you can’t accomplish that, scarify his legs’ (colaece,Lch_II_[1]:4.4.13.489)1 This construction disappeared almost entirely in Middle English (ME), being infrequent already in such texts as are available c.1150, although examples like (1.3), from around 1180, indicate that the construction had some productivity after this date: (1.3) & all himm wærenn fet & þeos / Tobollenn & and all him were feet and thighs puffed.up and toblawenn. swollen ‘and his feet and thighs were all puffed up and swollen’ (CMORM,I,280.2293) ¹ Citations beginning with co, as for this example, are taken from the York corpora, and ones beginning with CM are from the Penn-Helsinki Parsed Corpus of Middle English 2. These corpora are briefly mentioned in section 1.3 and more fully described in Chapter 3 and Appendix A. I present the citations as given in the output of those corpora. Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2



In Present Day English (PDE), the only parallels of this construction are fixed expressions like look x in the eyes, look a gift horse in the mouth, and stare x in the face, where the possessum is the object of a preposition, rather than a subject or object. The demise in English of the possessive construction using a dative has been of considerable interest to linguists, and English has frequently been singled out as being unusual in European languages generally and the Germanic languages more specifically in not having this construction. The primary goal of this book is an empirical one: to provide data bearing on the question of when and how this construction died out as a productive possibility in English. Explanations that have been proposed for this development fall into two major categories: internal and external motivation. The leading possibility for an internal trigger is the loss of the dative/accusative distinction, which Ahlgren (1946) identified as the main cause of the change. The reasoning is that once the distinction had become completely levelled, it was not possible to distinguish an object as dative. While the internal motivation just outlined is the leading traditional explanation for the loss of the construction under examination, more recently a system-external explanation, language contact, has received serious attention. Language contact is a traditionally assumed trigger for the collapse of the case marking system and was therefore assumed by Ahlgren to play an indirect role in the disappearance of ‘sympathetic datives’; however, a more central role for language contact has recently been argued by Vennemann (2002), McWhorter (2002), Filppula et al. (2008), and Filppula (2010), among others. While McWhorter argues for the role of contact with Scandinavian speakers, the other sources point to the influence of Celtic speakers on English. All of these proposed explanations suffer from insufficient data for a proper evaluation of the assumptions that underlie them. Vennemann (2002: 212)² comments that his impression is that the ‘genitival’ construction is ‘relatively rare’ in earlier OE, but he notes the need for a more detailed investigation. In the years which have elapsed since the publication of Vennemann’s article, no substantial progress has been made in this regard, and it is the major goal of this book to provide the more detailed investigation that Vennemann called for. A secondary goal is to evaluate the hypotheses concerning the causes of ² Vennemann (2002) is reprinted without changes of content as Vennemann (2012). The citations given here are to the 2002 publication, but can easily be found in the 2012 reprint because it helpfully supplies the page numbers corresponding to the original.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.1   

3

the disappearance of the dative construction in English in light of the fresh data produced by this investigation. Linguists interested in investigating this development have had to rely on examples taken from the sources discussed in section 1.4. The present book presents the results of the first systematic investigation of the structures exemplified in (1.2) and (1.3) based on the syntactically parsed corpora now available for Old and Middle English, introduced in section 1.5 and described more thoroughly in Chapter 3 and Appendix A. The focus will be on the two major categories of inalienable possessions where the construction with a dative referring to the possessor was the mostly widely used in OE, namely words referring to bodies or body parts and those referring to the mind or spirit. I will refer to the former as body words and the latter as mind words. This investigation reveals that the explanations that have been suggested to date for the loss of the dative construction suffer from a common flaw: they focus on the Early Middle English (EME) period as the crucial period for this development. In this book I will argue that the decline of the use of the dative to refer to the possessor began much earlier, probably by the late ninth century and certainly within the OE period. This being so, explanations that depend on developments of the EME period, such as the decline of case marking or the demise of the what is known as the ‘West Saxon Standard’ or ‘West Saxon Schriftsprache’, must be reassessed. The following two sections provide background for this study by surveying relevant literature: first cross-linguistic and then specifically history of English.

1.1 Dative external possessors The possessors in the construction illustrated in examples (1.1) through (1.3) have gone under many names, including ‘sympathetic dative’ (or dativus sympatheticus), introduced by Havers (1911) as a label for datives that could be replaced by a genitive, and possessor dative. Because the possessor is external to the NP³ that contains the possessum, Vergnaud and Zubizaretta (1992) introduced the term ‘external possessor’, contrasting with the ‘internal possessor’, used in the child’s hair or his legs, in which the possessor is internal to the NP containing the possessa hair and legs. I will adopt this terminology

³ I will take no position on the existence of a determiner phrase (DP) and will use NP (noun phrase) for the phrases containing the possessa, ignoring the issue of whether NPs are complements to determiners inside DP, as is widely assumed in current syntactic theories.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4



here, using the abbreviation EP when the case of the possessor is not being specifically referred to, except when discussing the work of authors using different terminology. I will use the abbreviation IP for the internal possessor construction, which may also be referred to as the ‘genitive’, ‘genitival’, or ‘possessive’ construction by different scholars. In discussing IPs in this book, I will generally use ‘genitive’ to cover both nouns in the genitive case and possessive pronominal forms, although in some instances I will use a more specific term when it seems clearer. Cross-linguistically, a phrase in the dative case is by no means the only possibility for an EP, but as Haspelmath (1999) shows, it is the most typical EP in European languages. I will adopt Haspelmath’s term ‘dative external possessor’ (henceforth DEP) for these possessors in the dative case. A terminological difficulty arises for EME examples like (1.3) above. The problem is that the use of the term ‘dative’ is questionable here because although the text from which this example comes, the Ormulum, is from early in the EME period, it is from an area in which the loss of case marking was very advanced, and the text shows no hint of a dative/accusative distinction. Him must be regarded as a general object case form, like the modern him, rather than a specifically dative form, in the Ormulum.⁴ I will nevertheless continue to use the term DEP for examples like this because the use of the object pronoun can be considered a continuation of the OE DEP—the object-marked pronoun refers to the possessor of the feet and thighs, is not part of the normal case-frame of the intransitive participles tobolen and toblawen, and is clearly added to signal the adverse effect that the swelling had on the possessor of the body parts. In some modern syntactic treatments, him would be treated as having abstract dative case even though there is no longer a morphological dative/accusative distinction.

1.2 Theoretical issues External possessor constructions have attracted a great deal of attention among linguists for the theoretical questions they pose. Arguments based on data from a written corpus that is the result of chance rather than design are of limited utility in marshalling arguments supporting a particular formal treatment of DEPs. However, even though we cannot learn as much about the properties of the DEP in Old English as can be learned for living languages, where the intuitions of native speakers can be used in conjunction with written ⁴ For a detailed discussion, see Allen (1995: §5.4.1).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.2  

5

and spoken corpora, we can learn a good deal about the way the DEP was used in earlier English texts and how it disappeared. While this information will be limited in its ability to evaluate specific proposals about the demise of the DEP, it can often rule out general approaches, such as accounts that depend exclusively on the loss of the morphological dative/accusative distinction, a widelyheld approach both in informal and formal treatments. We can also get data that bears on the possible role of language contact in the demise of the DEP. In my research into the syntax of DEPs and related constructions in Old and Early Middle English, I have tried to be aware of questions of particular interest currently to formal syntacticians as well as ones that have been the focus of traditional grammarians. Although no particular formal analysis will be argued for, this empirical study is informed by the issues raised in the formal literature, which underpin some of the decisions made about the categories used in gathering the data, and I will make some suggestions about the broad outlines of possible analyses that the data suggest. In the following discussion, I briefly outline some of the major points of theoretical interest in external possessors, giving an overview of the sorts of approaches that have been taken to the analysis of DEPs in particular, without going into details. In EP constructions, an NP which behaves as an argument of the clause bears a semantic relationship of possessor to a noun within another NP in that clause. Payne and Barshi comment that these constructions ‘challenge the notion that clause-level syntax depends directly on the argument structure and subcategorization frame of individual verbs or verb stems’ (1999: 14–15). Landau (1999: 2) summarizes the ‘surface phenomenon’ of PDCs (Possessor Dative Constructions, an alternative name for DEPs): a dative phrase, syntactically behaving like a normal dative argument of the verb (by movement diagnostics and so on), is in fact associated with another argument in the sentence, interpreted as a possessor of that argument.

The argument-like syntax of the dative in this construction sets it apart from superficially similar ‘free datives’ such as the ‘ethical’ (or ‘ethic’) dative; these and other datives that are not the focus of this investigation are surveyed in Chapter 2. There is substantial disagreement on the best syntactic treatment of EPs. Landau (1999) outlines the two basic approaches. The first is exemplified by this clause in König’s (2001: 971) cross-linguistic definition of EP constructions: (iii) Despite being coded as a core argument, the possessor phrase is not licensed by the argument frame of the verb itself.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6



Payne and Barshi give essentially the same definition (1999: 3). By this approach, the dative element is not an argument of the verb, and its syntax is usually explained as being due to a process of raising from the NP containing the possessum. This approach is the basis for the older terms ‘possessor raising’, which is now normally replaced by the more neutral terms ‘external possessor’ or ‘possessor dative’. Landau (1999) is one of the fairly recent treatments taking this general approach. The second basic approach, as Landau explains, is that the possessor dative is an argument of the verb, carrying a thematic role such as . The appeal of such a treatment is that the construction is usually associated with a particular semantics. That is, a DEP is usually associated with some sort of effect on the possessor, which may be beneficial or adverse. Virtually everyone who has written on the subject of external possessors mentions that affectedness plays some sort of role in their use. For Haspelmath (1999: 111), the two most important characteristics of the ‘European EP prototype’ are the use of dative case for the possessor and the ‘strict affectedness condition, i.e. EPs are only possible if the possessor is thought of as being mentally affected by the described situation’.⁵ Vennemann, who assumes that OE was essentially like Modern German with regard to DEPs, states that ‘the dative is obligatory for affected possessors in German’ and explains that a DEP in German presents an action which happens to the possessor with respect to some body part, while an IP reports an event which happened to the body part (2002: 208). Krahe (1972: 88) similarly contrasts the emotive ‘dativus sympatheticus’ Er zerschmetterte der Schlange den Kopf ‘he shattered the snake:  the head:’ with the ‘ganz objektive’ (‘completely objective’) genitive Er zerschmetterte den Kopf der Schlange ‘he shattered the snake’s head’ (lit. head:  of the snake). Deal (2013) also comments on the fact that a DEP is very strange in German if the possessor is dead, and therefore incapable of feeling an effect, but points out that in some languages affectedness plays no role in external possession and comments that more research is needed into languages where this is the case. Vergnaud and Zubizaretta (1992: 595) treat the dative possessor in French as a complement of the verb which is ‘affected by the action or state referred

⁵ Haspelmath goes on to note that ‘the affectedness condition is not equally strong in all languages, and it has been conventionalized in various ways by different languages (1999: 113)’. See Wierzbicka (1979) for an insightful discussion of how different languages treat different situations as involving an effect and subtle semantic differences in the choice of construction used in different languages to express possession of body parts. Although a corpus search cannot give us a complete picture of what role affectedness played in early Old English, it is nevertheless informative.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.2  

7

to’, and Landau (1999: 3) notes that PDCs are not semantically equivalent to their IP counterparts because they all imply that the possessor is somehow affected by the action denoted by the verb [Landau’s emphasis].⁶ The notion of affectedness plays a central role in Lee-Schoenfeld’s analysis of possessor dative constructions in German, where it is observed that the German PDC gets better as the negative or positive effect is more obvious (2006: 108). The importance of affectedness in OE DEPs is discussed in Chapter 4. Lee-Schoenfeld (2006), making use of developments in the Minimalist approach to syntax, adopts an analysis in which the dative possessor both carries a sematic role and is raised. The basic idea is the possessor is generated with dative case in a position where that case cannot be licensed, and so the element raises to the position of the  argument of the verb. Deal (2013) similarly assumes that some sort of analysis in which a theta-role is assigned by the main verb is appropriate for German and other languages in which affectedness is part of an EP construction and adopts a case-driven analysis that prevents the possessor from staying in the same NP as the possessum. She argues, however, that in languages like Nez Perce, where affectedness does not play a role, EPs involve possessor raising rather than control. Objects of prepositions (PObjs) present some difficulties for the analysis of EPs, and the present investigation makes a systematic distinction between what I will refer to as ‘direct argument’ possessa, that is, subjects and objects, and PObj possessa. In their treatment comparing French and Modern English, Vergnaud and Zubizaretta (1992) note that English does not in general productively allow external possessors parallel to French, e.g. (1.4): (1.4) Le médecin leur a radographié l’estomac the doctor them: has x-rayed the.stomach ‘The doctor x-rayed their stomachs’ (Vergnaud and Zubizaretta 1992, ex. 4.a)7 However, Modern English does allow what they treat as external possessor of complements of prepositions in what they call the ‘PP construction’: (1.5) John tickled the children on the foot.

⁶ Landau points out problems with defining ‘affectedness’ in his section 5.2.3. However, he comments in footnote 1 that ‘unless otherwise mentioned, PDC is always associated with an affectedness implication for PD’. ⁷ I have modified the glossing that Vergnaud and Zubizaretta supply, in conformity with the Leipzig Glossing Rules. Note also that what they refer to as the ‘distributivity’ effect, by which a singular form is used in this example to refer to more than one stomach, is a general feature of the French construction, as Vergnaud and Zubizaretta (1992: 598) discuss.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8



Vergnaud and Zubizaretta’s analysis and their explanation for the grammaticality of the PP construction incorporates many theory-specific assumptions, and the details are not important here. However, their conclusion that ‘the inalienable phrase in the PP has a different analysis from the inalienable phrase in direct object position’(1992: 642) is highly relevant to this study, as is their observation about the entailment relationship involved in the PP construction. Although Vergnaud and Zubizaretta treat sentences like (1.5) as having EPs, not everyone does. König (2001) excludes sentences like (1.5) and his own parallel example, presented in (1.6), in his definition of EPs: (1.6) Ben punched Jim on the nose

(König 2001 ex.7a)

The basis for this exclusion is that the possessor (Jim in this example) is licensed by the argument frame of the verb. This is ruled out by clause (iii), presented above, in his definition of EPs. König notes that the omissible prepositional phrase ‘provides a further (metonymic) specification of the precise endpoint of the action expressed by the verb’. As discussed in more detail in Chapter 3, I similarly excluded such examples from my data collecting, while including non-thematic objects similar to those found in Modern English relics such as the answer is staring you in the face and I looked them in the eye. For our purposes, the most important fact about a Modern English PP construction like (1.5) and (1.6) is that the possessor is clearly a thematic argument, that is, one in the basic argument frame of the verb. This is evidenced by the fact that in the examples of (1.7) the possessor of the cheek cannot be left out, while the PP can be omitted and still result in a grammatical sentence: (1.7) a. John kissed the child on the cheek. b. John kissed the child. c. *John kissed on the cheek. (1.7a) entails (1.7b); if John kissed the child on the cheek, he kissed the child. In contrast, with the fixed expressions found in the (a) sentences of (1.8) and (1.9), the prepositional phrase is not omissible and there is no entailment relationship, as the ungrammaticality of the (b) sentences shows. (1.8) a. She looked him in the eye. b. *She looked him. (1.9) a. The answer was staring him in the face. b. *The answer was staring him.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.2  

9

In (1.7a), the child is a thematic object, that is, it is part of the verb’s argument frame. Linguists disagree on whether to count constructions in which the possessor is a thematic argument, as in (1.7a), as external possessor constructions. König notes that such examples in which the prepositional phrase can be omitted are different from uncontroversial EPs in lacking a semantic and syntactic dependence between the possessor phrase and the possessum phrase. While König’s overview of external possession does not use the term ‘implicit’ for the possessive relationship found in such examples with thematic objects understood to be the possessor, Haspelmath and König (2001: 579) present (1.10) as an example in which the possessor of the object of a preposition is ‘implicit’, the reference being controlled by the object: (1.10) Pierre a embrassé les enfants sur la joue Pierre has kissed the children on the cheek ‘Pierre kissed the children on the cheek’ An implicit possessor relationship is one that is not explicitly stated but understood. Whether or not we want to define external possessor constructions to include the PP construction, for this investigation it was important to include examples parallel to (1.8) and (1.9), with athematic objects, but to exclude ones like (1.5), (1.6), and (1.7a), with thematic objects. (1.5), (1.6), and (1.7a) have their OE parallel in examples like (1.11): (1.11) oþ hiene an cwene sceat þurh þæt þeoh until him: a woman shot through the thigh ‘until a woman shot him through the thigh’ (coorosiu,Or_3:7.64.28.1264) In this example, there is a clear entailment relationship between the sentence with and without the prepositional phrase, and hiene is clearly the object of the verb. Since the focus in this book is on documenting the decline and loss of EP constructions that became impossible in English, examples like this are excluded from the data of this study. In contrast, (1.12) has no productive parallel in Modern English: (1.12) ðonne hie him on ðæt nebb spætton when they him: in the face spat ‘when they spat in his face’ (cocura,CP:36.261.7.1700) It is examples like (1.12) that will be included in my statistics for PObjs. While I have used the term ‘implicit’ for the understood possessors of PObjs as in (1.10) and (1.11), the more common use of this term is for the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

10



understood possessors of direct objects. No overt expression of the possessive relationship is necessary in French in situations where it is obligatory in Modern English, as noted by Vergnaud and Zubizaretta: (1.13) Les enfants ont levé la main the children have raised the hand ‘the children raised their hands’ (Vergnaud and Zubizaretta ex. 1) The corresponding literal English translation *the children raised the hand is of course ungrammatical in the possessive meaning, but could only have a meaning that there was a detached hand that the children all raised, as Vergnaud and Zubizaretta point out. The English translation does not improve by substituting the plural hands, either. Vergnaud and Zubizaretta treat the construction of (1.13) as an EP, but like the prepositional objects just discussed, it is excluded by König’s (2001) definition of EPs. In Vergnaud and Zubizaretta’s analysis, the possessor relationship in (1.13) is not implicit, because they assume a binding relation between the possessum and the possessor, which is established via Predication. They seek to relate the difference between French and English to the fact that French, unlike English, allows expletive determiners (that is, determiners without a reference), while English does not. This is the same fact that they see as the key to the grammaticality of EPs in French. In this way, they make a typological prediction: languages that allow ‘implicit’ possessors will also allow EPs more generally. We find implicit possessors in OE, which is consistent with this correlation: (1.14) Hond up abræd Geata dryhten, hand: up raised Geats:. lord: ‘the lord of the Geats raised up his hand’ (cobeowul,80.2575.2108) Beowulf 2575 However, König and Haspelmath (1998: 579) note that while one might try to propose a correlation between implicit and external possessors, they find counterexamples to this proposed correlation and conclude that it cannot stand as a typological universal. While it would be interesting to see whether the decline of DEPs in English went hand in hand with the decline of implicit possessors, searches for implicit possessors would greatly complicate the already complex data gathering, since any search for them would have to encompass all sentences that mentioned a body part. My impression is that implicit possessors of objects are not very common in OE prose, and so such searches are unlikely to yield enough examples to make any convincing link

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.2  

11

between the loss of DEPs and the loss of the types of implicit possessors that are no longer possible in English. Furthermore, any linguist who is committed to an analysis of EPs that links them with implicit possessors will surely not be convinced by an apparent counterexample that comes from the history of English, but only by counterexamples that come from spoken languages where native speakers are available to give grammaticality judgements. For this reason, I have not made a systematic study of implicit possessors in my corpus, but I will make some observations about the appearance or otherwise of implicit possessors in the various texts. Lee-Schoenfeld (2006: 138), who presents a formal syntactic account of the possessive relationship obtaining with DEPS, differs from Vergnaud and Zubizaretta in considering that the possessive relationship in corresponding German sentences like (1.15) is what she refers to as ‘pragmatic’, that is, implicit rather than syntactic: (1.15) Er hebt die Hand He raises the hand ‘He raises his hand’

(Lee-Schoenfeld 2006 ex. 53)

Lee-Schoenfeld also mentions other situations that are relevant to the present study in which she considers that a pragmatic rather than a syntactic relationship of possession is involved because no possessor interpretation is obligatory (2006: 109). One is in the case of verbs that independently select a dative argument. Here, no possessive relationship need be involved, and when such a relationship is understood, no second dative is added. Similarly, a verb may be compatible with the addition of some other type of dative, such as an ‘ethical’ or ‘estimative’ dative. Datives of this sort are relevant to the comparison of DEPs and IPs in OE because they raise the issue of whether they should be included in the counts of DEPs. In keeping with my exclusion of sentences with thematic arguments in my counts of DEPs with PObjs, I follow LeeSchoenfeld in treating understood possessive relationships between such datives and a body or mind word as pragmatic rather than syntactic, and exclude these datives from my counts of DEPs. This is further discussed in Chapter 3, section 3.5. Scholten’s (2018) study of external possession in Dutch varieties takes a very different approach to those discussed so far and aims to provide a unified account of EP constructions cross-linguistically. For Scholten, the treatment of definite articles is key to a unified analysis. Scholten follows Giusti’s (2015) treatment of definiteness, and a major difference between Scholten’s account and all previous ones is that she assumes that all articles are expletive—a

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

12



definite interpretation of a DP is not due to the presence of a definite article but rather to the presence of a definite operator.⁸ Rather than assuming a movement analysis or one introducing an external possessor as a benefactive argument, Scholten assumes that the possessor and possessum are generated in separate nominal phrases (2018: 65). The possessum lacks its own referential index but gets it from the possessor, which does have such an index, through Agree. Scholten comments: ‘Because the definite article is not inextricably linked to a referential index, they have the possibility to receive their referential index from another element in the structure.’ A direct consequence of Scholten’s approach relevant to this study is the fact that if we follow it, the absence of DEPs in Present Day English cannot be due to a loss of expletive determiners somewhere along the line in the history of English. Apart from any theoretical considerations, Scholten’s study represents an important addition to the database that linguists have for working on external possession, especially in view of the fact that the existence of productive EP constructions in Dutch has been the matter of some debate, as discussed in section 8.2.1. Dutch has been cited as providing arguments both for and against a widely held language-internal explanation for the loss of DEPs in English, namely the collapse of the dative/accusative distinction. The discrepant views seem to have arisen from the rather limited acceptability of different EP constructions in Standard Dutch as well as variability among Dutch varieties. Scholten’s data furthermore indicates that any approach that makes a simple link between DEPs and implicit possessors must fail, because the broad generalization can be made that while most Dutch dialects allow implicit external possessors equivalent to (1.13) through (1.15), only those close to the German border allow external possession by an indirect object (i.e. a DEP) of a direct object body part.

1.3 Typological considerations The surveys of external possession constructions that have been carried out by typologists both provide parameters for categorizing the OE data and

⁸ In contrast, Le Bruyn (2014) argues that the French article in the DEP is not expletive but rather a ‘run-of-the-mill definite article’. Le Bruyn’s analysis will not be further discussed here, but for a discussion and criticism, see Scholten (2018: 53–7).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.3  

13

cross-linguistic data for assessing how DEPs in OE and their later loss fit into the typological picture. Typologists have discerned a number of hierarchies in patterns of external possession. To start with, EPs are especially associated with inalienable possession cross-linguistically, and Payne and Barshi (1999) place body parts at the top of the hierarchy of accessibility of possessa to external possession. Possessa relating to the mind or spirit are also high on this hierarchy, and it is for this reason that the present study focuses on words of these two categories, where we are likely to find the most examples of EPs. König (2001: 976) and Haspelmath (1999: 113–15) present universal syntactic relations hierarchies which show that direct objects are more favoured as the possessa of EPs than intransitive subjects are. Surface subjects can be divided into two major types: ones that are more like objects in being nonvolitional themes or patients rather than agents of the verb, and ones that are more agentive. The more object-like type of subject is widely known in the linguistic literature as an ‘unaccusative’ subject, and is illustrated by (1.16): (1.16) ac him ða eagan of his heafde ascuton but him: the: eyes: from his head out.shot ‘but his eyes shot out of his head’ (cobede,Bede_1:7.40.7.332) In these hierarchies, unaccusative subjects are in a higher position than volitional, active ones, which are known as ‘unergative’ subjects, and I have not found convincing examples of unergative subject possessa appearing with a DEP in OE; such an example would be something like him gazed the eyes on her beauty in a meaning ‘his eyes gazed on her beauty’. It is hardly surprising that unergative subjects should not be found with DEPs, since they are not typically affected. The same applies, but more so, to transitive subjects, which Haspelmath places at the bottom of his hierarchy. Haspelmath’s hierarchy, which is specifically for DEPs in European languages, differs from König’s by listing prepositional phrases at the top of the hierarchy. It is notable that many European languages that do not have EP constructions with subject and object possessa do allow them, to some extent at least, with PObj possessa. For example, Lødrup (2009a) shows that what he refers to as ‘possessor raising’ in Norwegian is limited to unergative verbs with PObj possessa: (1.17) Han tråkket henne på føttene he stepped her on feet: ‘He stepped on her feet’

(Lødrup 2009 ex. 13)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

14



In OE, examples of DEPs are easiest to find with PObjs, where no affectedness condition seems to hold, as discussed in detail in Chapter 4. It is notable that the apparent relics of DEPs in PDE, e.g. it’s staring you in the face and look me in the eye, all involve PObjs. One question that will be investigated in this book is whether as DEPs declined in English, their productive use remained longer with PObjs than with direct arguments generally and longest of all with direct objects or whether they disappear (except in fixed expressions) with PObjs, direct objects, and unaccusative subjects at pretty much the same time. Payne and Barshi suggest that ‘the particular hierarchy for a given construction most certainly depends on the diachronic origin of the EPC [External Possessor Construction] in question’ (1999: 11). They say that the DEP typical of the European languages developed from an extension of locative/goal or dative schema described by Heine (1997). Heine’s locative and goal schemas are of course distinct schemas, but the dative case is used for both locative and goal functions in European languages generally, although these concepts may also be conveyed by prepositions. Heine (1997: 154) comments that it is not unusual that in a given language two or more source schemas develop within the same language, and the same is true of attributive possession. While Heine mentions that the dative variant of the EP goes back to his goal schema, there seems to be no reason why some instances of DEPs might go back to a locative schema, and I suggest in Chapter 4 that a different origin as goal or locative might account for the fact that while DEPs with lexical verbs typically involve an affected possessor (a goal), DEPs with copulas do not necessarily indicate an affected possessor. While a human goal is likely to be affected by an action, copular constructions indicate location, and a human being who is a location is less likely to be affected.

1.4 Previous studies of external possession in early English Having surveyed recent research by linguists on DEPs in various languages, we turn now to a survey of previous work done specifically on these constructions in early English, mainly by traditional grammarians. This section focuses on research that establishes the empirical base available up to now for the study of the loss of the DEP in English. The theories attempting to account for the demise of this construction have been outlined in the introduction to this chapter, but detailed discussion is deferred until Chapter 8, when the data necessary for evaluating these explanations has been presented.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.4        

15

1.4.1 Havers (1911) Havers (1911) introduced the term dativus sympatheticus ‘sympathetic dative’ for the construction being investigated in this book. Havers’ use of this term was for the type of dative that could be used instead of a genitive, and he commented that it was more subjective and warmer than the genitive, which objectively states a possessive relationship (1911: 2). Havers’ study, which covers an impressive number of languages, encompasses all the early IndoEuropean families Although more than 100 years old, this work remains a valuable source of examples and insights about the use of these datives and comparisons with alternative expressions of possession. Havers distinguished six basic types of sympathetic dative and provided examples of each type from prose and poetical texts in the languages he investigated. Since Havers was surveying so many languages, he naturally made no attempt to cover all available texts from any period of English. His examples relevant to the study of DEPs come mainly from OE poetry, especially Beowulf and Cynewulf ’s works, with some examples from prose. His discussion of Middle and later English focuses on the use of alternatives to the old dative. Although Havers’ main concern was to investigate the possibility of interchange between the dativus sympatheticus and the ‘possessive construction’, that is, genitive case or a possessive pronoun, a secondary interest was to survey situations in which the dative case was a possible variant to some construction other than a possessive construction, for example, a locative construction. Havers’ taxonomy of examples used two dimensions: meaning and form. He categorized his examples into six major types on the basis of the type of possessive relationship, and within these types divided the examples according to whether the possessor is expressed by (a) a sympathetic dative, (b) a genitive, or (c) a locative expression. In discussing Havers’ study as background to the present one, I deal first with the three construction types, then with the semantic categories. Construction types (a) and (b) need no explanation, but (c) is alien to Modern English. In North Germanic languages, it has always been common for the possessor to be encoded as some sort of locative, as in this example: (1.18) hvé þar á Herkju hendr sviþnuþu how there on Herkja hands: burned ‘How Herkja’s hands burned there’ (Old Norse) (Gþr. III 10, 2 as cited in Havers 1911: 269)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

16



This construction is the normal EP construction in the modern Scandinavian languages.⁹ It is illustrated by Lødrup (2009b) for modern Norwegian in (1.19): (1.19) De måtte fjerne leveren på ham They must remove liver. on him ‘They had to remove his liver’ (Lødrup 2009b ex. 15) Although Havers identified a few examples that he treated as involving a locative construction in OE poetry, they are very different from the North Germanic examples. In those examples, as in the modern Scandinavian constructions, it is the possessor that is encoded as a locative. In the OE examples that Havers adduces, however, the possessor is in the accusative case and the possessum is in the dative/instrumental case: (1.20) hine þa heafde becearf him: then head:/ carved ‘then cut off his head’

Beowulf 1591

Example (1.20) is one of the examples of beceorfan using a dative of deprivation that is presented by the Dictionary of Old English (DOE) to illustrate the idiom that is discussed in section 2.2.1 of this book.¹⁰ It is this dative heafde that Havers is counting as a locative. Havers was not just considering how the possessor was expressed but also the possibilities of using locatives or datives in the expression of possession more broadly. In his discussion of North Germanic, Havers includes both examples of possessors and possessa in his locative category; example (1.20) is taken from that discussion. We can assume that if Havers had found similar examples to the North Germanic ones in OE, he would have included them, but he does not present any. This investigation, which included more texts than Havers’ investigation of OE, turned up one example with a body word that seems to be parallel to the Scandinavian locative possessors:

⁹ I will use Scandinavian in a narrow sense, as an abbreviation for the Germanic languages of Scandinavia. ¹⁰ An instrumental of the body part seems to have been particularly common in expressions of beheadings. It seems to be most common in the poetry, but occasionally occurs in the prose: (i)

se ilca, þe hine þy the same that him: the:  ‘the same one who cut off his head’

heafde beheow head: hewed (cogregdC,GDPref_and_4_[C]:24.293.25.4352)

The determiner is in a distinctive instrumental form. As mentioned in Chapter 2, nouns had no distinctive form for the instrumental, and are parsed as dative by the YCOE even when modified by an instrumental determiner.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.4        

17

(1.21) woldon æninga, ellenrofe, on þam hysebeorðre would: at.once brave:. on the: male.child: heafolan gescenan head: harm ‘They, brave ones, would at once harm the boy’s head’ (coandrea,34.1139.424) It does not seem that this locative construction was a usual way to express possession in OE. Havers did not distinguish between different grammatical roles of the possessum, a dimension that modern typologists have recognized as important. However, his decision to distinguish nouns and pronouns in his data yielded an interesting observation. Havers found that the dativus sympatheticus was the dominant construction with pronouns in the Germanic languages generally, while the genitive was more likely to be found with nouns. In Chapter 4, I test Havers’ observation against my own data and find some support for the idea that a nominal possessor was more likely to be expressed as an IP than a DEP, although the effect does not seem to be as strong as Havers suggested. Turning to Havers’ six semantic categories, we can start by noting that he was not limiting himself to expressions of possession, because although his main concern was to investigate the possibility of interchange between the dative and genitive constructions, a secondary interest was to survey situations in which the dative case was a possible variant to some construction other than a possessive one, for example, a locative construction. Therefore, not all his semantic categories relate to types of possession. His first category is man’s body or parts thereof. He considers this to be the central use of the dativus sympatheticus, a finding that is in line with that of more recent typological studies. His second category concerns words referring to the Seele, which overlaps with the English concepts of mind, soul, and spirit. It is these two categories, body and mind, that are the subject of the present investigation, a decision made on the basis that they are the two types of words that are most frequently involved in DEPs in OE. Havers’ third category is possession more broadly. His examples include (1.22): (1.22) þa he him of dyde isernbyrnan when he him: off put iron.corslet: ‘When he took off his iron corslet’

Beowulf 671

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

18



This is an amorphous and murky category, and the status of the dative as the expression of possession is less than certain, since the translation could be something like ‘when he took the iron corslet off him’; as Havers notes, the him is reflexive. Other examples involve the verb losian ‘to become lost (to)’: (1.23) a. gif him þæt rice losað if him: the: kingdom: loses ‘if the kingdom becomes lost to him/if he loses the kingdom’ Genesis A 434 b. ac him losode an sceap but him: lost one: sheep: ‘but a sheep was lost to him/he had lost a sheep’ (cocathom1,+ACHom_I,_24:372.25.4684) While rice and sceap can be considered a possession of the person who might lose it, the dative that occurs with the verb losian ‘to perish, become lost’ to convey the effect on a concerned person is not an external possessor. Him could be left out in the examples of (1.23), giving the meanings ‘the kingdom was lost’ and ‘one sheep became lost’ or ‘one sheep perished’. However, if we substituted a possessive his rice or his sceap here for the dative constructions, the meaning change would be more drastic than is usual with the exchange of an internal possessor for an external one, since it would mean that his kingdom or the sheep perishes, rather than the kingdom or sheep is lost to him. Such examples are not covered in this study. Havers only offers a few examples from this category and notes that all the examples he considers certain involve pronouns rather than nouns. Havers’ fourth category involves verbs of deprivation. The concept of possession is relevant here in that when someone is deprived of something, that thing can normally be thought of as a possession in a broad sense. However, Havers’ examples of the dative used with such verbs include ones in which the dative cannot really be thought of as a possessor: (1.24) nefne him witig god wyrd forstode if.not him: wise: God: fate: prevented ‘if the wise God had not prevented him (from accomplishing) that fate’ (cobeowul,33.1050.874) Beowulf 1056 Havers mentions that with some of these verbal types, the possessive construction is not an alternative, and also that some of his examples could be analysed as ‘ethical’ or other dative types. The examples that Havers presents

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.4        

19

from OE do not show a convincing alternation with internal possessors. It seems better to treat them simply as ditransitive verbs, in which the dative is a core argument in the normal argument frame of the verb. Havers’ fifth category involves nouns of human relationship such as ‘father’, ‘king’, ‘friend’, etc., as in (1.25): (1.25) Eadgilse wearð feasceaftum freond Eadgils: became destitute: friend: ‘He became a friend to the destitute Eadgils’ Beowulf 2392–3 Havers presents a few examples like this, but comments (p. 77) that the genitive is ‘auch ganz geläufig’ [also quite common/CLA]. My own observation is that DEPs are not very common, at least in the prose, and they all involve predicates of being or becoming, as far as I have been able to determine; I have noticed no examples equivalent to *They killed him the king as a variant to they killed his king. I have not made a systematic investigation of the use of datives with nouns of human relationship, but see the discussion of datives with the predicates of copulas in section 2.4.2. Havers’ final type is unlike the others in that it does not involve the interchange between the dative and the possessive construction but rather between the dative and locative expressions. Havers (p. 279) presents example (1.26) as an Old English instance of the dative used where a locative expression would have been an alternative: (1.26) his wæstm . . . þæt him com from weroda his harvest that him: came from hosts:. drihtne lord ‘his harvest that came to him from the lord of hosts’ Genesis A 255 Here the locative expression to him could have been used instead of the simple dative. This category has nothing to do with possession.

1.4.2 Ahlgren (1946) Ahlgren’s (1946) study, as his use of the expression ‘nouns of possession’ suggests, concerns inalienable possession. Ahlgren’s focus is on the historical use of the definite article in expressions of possession where Modern English would use a ‘possessive adjective’ instead, and a good deal of the study covers

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

20



what I am treating as implicit possessors, e.g. She had to submit and bow the head. The discussion of most relevance to this study is Ahlgren’s Chapter IV, which looks at constructions that have been for the most part lost along the way to Modern English. The ‘sympathetic dative’ is one of those constructions and therefore receives a good deal of attention. Since constructions lacking either a possessive or a definite article are also among those lost, Ahlgren also provides numerous examples of implicit possessors lacking a determiner, such as (1.14) above. From a contemporary linguist’s point of view, Ahlgren’s organization of his study by sections on examples that would be expressed differently in Modern English is frustrating because it makes it impossible to extract the system behind the examples. However, the study remains an important one because of its wealth of examples, and it makes some observations on the loss of the ‘sympathetic dative’ that need to be considered carefully. Ahlgren’s observations on the decline of this construction is discussed in detail in Chapter 8.

1.4.3 Visser (1963) Visser’s (1963) organization of surface construction according the number and type of elements occurring with verbs means that the discussion of possessive constructions is necessarily distributed through several sections, and the comprehensive historical sweep of the book from Old to Modern English precludes any discussion of the systems involved in any period. Working one’s way through the various sections related to DEPs nevertheless yields a good deal of useful data concerning constructions relevant to this study, both about OE and later periods. Visser furthermore provides some valuable examples of DEPs from EME texts that are not included in the PPCME2.

1.4.4 Mitchell (1985) Mitchell’s (1985: §§306–10) discussion of the ‘dative of possession’ in OE makes the observation (§306) that the ‘noun of possession’ (i.e. possessum) is most often governed by a preposition in OE. This important fact, which seems to have mostly gone unnoticed, is probably responsible for the widespread belief that DEPs were the most usual way to express inalienable possession in OE; DEPs are in fact easy to find with PObj possessa, but much less frequent with direct argument (subject and object) possessa, as discussed in Chapter 4.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.4        

21

Mitchell’s focus is naturally on the OE situation, but he does make some comments on the later fate of the DEP, saying that Ahlgren (1946) demonstrates that the dativus sympatheticus declines in frequency in the OE period and summarizing Ahlgren’s suggestions for this decline (1985: §310). Mitchell (§303) also observes that the use of a possessive (rather than a dative or some other means) with ‘nouns of possession’ was well established in OE.

1.4.5 Mustanoja (1960) Mustanoja treats the DEP under the heading ‘sympathetic dative’ in his survey of the uses of the different cases in ME (1960: 98–9). Mustanoja’s discussion of the sympathetic dative is naturally brief, since his book covers all of ME syntax, but he comments that the construction was common in OE, especially with pronouns, but is ‘comparatively infrequent in ME and loses ground steadily’. Mustanoja thus explicitly assigns the beginning of the decline of the DEP to the ME period.

1.4.6 Celtic Hypothesis scholars Summing up the findings for OE of the scholars discussed above, the use of a possessive/genitive was well established in OE, and also well established in much earlier Indo-European languages. Havers (1911) richly exemplifies variation between the dative and genitive, with little discernible difference in meaning, in Ancient Greek and other early Indo-European languages. However, some recent works by adherents of the ‘Celtic Hypothesis’, namely the hypothesis that OE grammar was significantly affected by contact with Celtic speakers in the OE period, may give the impression that the affected IP was a newcomer to OE syntax. The Celtic Hypothesis as it relates to the loss of the DEP in English has been outlined near the beginning of this chapter and is discussed in detail in Chapter 8. I will restrict myself here to some discussion of how the OE facts have been presented by two works that promote this hypothesis. Vennemann (2002: 208) states that ‘the Present-Day English construction is not Germanic, and it is not Indo-European’. He also refers to ‘the inherited construction’. Vennemann’s use of the term ‘construction’ here is confusing, because the most obvious interpretation of the word (to this reader at least) in this context is to refer to the IP, and so at first it seems that Vennemann is

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

22



making the surprising claim that IPs were not a feature of OE or Germanic. However, Vennemann is talking not about a specific construction for expressing possession but rather the fact that the use of IPs only in Modern English does not reflect the earlier situation in these languages. Vennemann later makes it clear that he is not suggesting that the IP was not used at all in OE, giving a few examples, but comments that ‘my impression is that it is comparatively rare in earlier Old English’ (2002: 212). His call for a more detailed investigation into ‘the rise of this alternative construction—by dialect, text type, and time’ is apt, but his use of ‘alternative construction’ here and ‘the inherited construction’ gives the impression that he is suggesting that the putative rareness of the IP in early OE is due to its recent appearance as an alternative to the DEP. Filppula (2008) draws heavily on Ahlgren’s and Mitchell’s accounts of the dativus sympatheticus in OE and recognizes that the use of internal possessors increased within the OE period, noting Mitchell’s comment that the Modern English patterns not using a dative were well established in OE (2008: 30–1). However, since he refers to the ‘innovative internal possessor type’ in OE, he seems to be suggesting that while internal possessors were used in the OE period, they were newcomers, and he explicitly states that ‘the external type was the prevailing one’ in earlier English, comparing this situation with ‘ModE’ [Modern English/CLA]. While Filppula notes (p. 32) that there appears to be a general consensus about a gradual decline of external possessors in English, his reference to the ‘change from external to internal possessors’ reinforces the impression that the loss of external possessors was a simple matter of replacement by internal ones, rather than a change from an ancient situation of variation to one in which the variation has disappeared. Filppula deserves credit for making the first use of electronic corpora to gather data on the use of DEPs in Middle English, using the Helsinki corpus to give more substance to impressionistic accounts of the scarcity of DEPs even in the early part of that period. However, his study only looks at three body words and seems not to have included any comparison of DEPs and IPs with these words, with the result that he has not actually established a drop in relative frequency of the DEP compared with the IP from earlier stages.

1.5 Corpus and historical scope While this study draws on the findings of earlier scholars, it offers a wealth of new data that has become available through the advent of syntactically parsed

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

1.6    

23

corpora. For OE, we have the York-Helsinki Parsed Corpus of Old English (Taylor et al. 2003, henceforth YCOE) and Pintzuk and Plug’s (2001) YorkHelsinki Parsed Corpus of Old English Poetry (York Poetry Corpus). For Middle English, we have the Penn-Helsinki Parsed Corpus of Middle English 2 (Anthony Kroch and Ann Taylor 2002, henceforth PPCME2). Because examples of DEPs become sporadic after the PPCME2’s m1 period (1150–1250), my systematic searches for DEPs extend only through the m2 period (1250–1350), although I have supplemented these searches with information gleaned from dictionaries and other sources, as detailed in Chapters 3 and 6 (for OE) and 7 (for EME).¹¹ It is important to realize that not all of the texts of these corpora are of equal value as sources of data. Some of the texts are from editions that are cobbled together from manuscripts emanating from very different periods, sometimes spanning as much as a century and a half. Others are from a single manuscript, but the manuscript is a copy of a text composed much earlier. With such texts, there is the possibility that scribes intervening between the time of composition of the original text and the final manuscript(s) may have revised or miscopied relevant sentences. For this reason, I have not included all the texts from these manuscripts. I have furthermore modified some YCOE texts by splitting the text into separate files according to the manuscript that contains the text, when these manuscripts belong to significantly different periods. These matters are detailed in Appendix A.

1.6 Organization of the book Chapters 2 and 3 are preliminaries to the investigation of DEPs in early English. Chapter 2 looks at some major uses of the dative case in OE. This discussion both informs the reader of the constructions employing the dative case that have been excluded from the systematic investigation of DEPs and sets the scene for a discussion in Chapter 7 about the loss of the DEP in Middle English. Chapter 3 offers an outline of my corpus and methodology, describing the corpora used for OE and the way the investigations were carried out. Chapter 4 is the heart of the book, as it provides an empirical base for any further work on the disappearance of DEPs in English. It presents the results

¹¹ The abbreviations for periods by electronic corpora are based on those of the Helsinki Corpus, e.g. ‘m1’ stands for the first period of Middle English, ‘o3’ stands for the third period of Old English, as explained in Kytö (1993).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

24



of the OE investigations into the body words, the words most frequently found with a DEP in the OE texts. Chapter 5 looks at some evidence that shows that the decline of DEPs of body possessa started early in English. First, I use comparisons of Gothic and OE to argue that the range of the DEP with body possessa seems to have been greater than what is found even in the earliest OE texts. Some comparisons of OE and Old Saxon are also made. Next, I use the data presented in Chapter 4 to show that the use of DEPs with these inalienable possessa was becoming more restricted in later OE, as far as we can tell from the texts. Chapter 6 details the results for the mind words, which are not as frequent with DEPs in OE and so provide less data but reinforce the evidence pointing to a reduction of frequency for this type of possessor. Chapter 7 describes the investigation carried out for EME and presents the results, which are crucial in evaluating hypotheses about how the DEP disappeared in that period. In Chapter 8, I examine theories of why the DEP died out in English in the light of the results of this study. My conclusions are reviewed in Chapter 9, where some avenues for future research are also pointed out. Appendix A gives further details on what texts were used in the study and how I organized them into types, while Appendix B lists the lemmas used in my searches and Appendix C gives a more detailed discussion of the methodology used in searching for examples that go beyond the general outlines given in Chapters 3, 6, and 7.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2 Dative case in Old English An overview

This chapter presents an overview of the dative case in Old English to give the reader a picture of how DEPs fit in with other uses of the dative case. This overview is not intended as a comprehensive discussion of all uses of the dative case, but it aims to cover the main uses that have relevance to the investigation of DEPs and also to assist in understanding the discussion of specific examples.¹ One of the purposes of this chapter is to discuss datives that might be construed as DEPs but have not been treated as such in this investigation. Additionally, this chapter gives the background for a discussion of how the near-loss of DEPs in EME ties in with a general decline in the use of dative case in English. Mitchell (1985: §§1345–79) provides discussion and examples of the uses of the dative case in OE and Visser (1963) provides rather idiosyncratically organized examples from both OE and ME. The discussion in this chapter relies fairly heavily on these two sources, but makes some additional observations on possible analyses and the relationship of various constructions to DEPs. Mitchell makes a distinction between what he calls the ‘dative proper’, which he says seems to have originally been used to indicate ‘personal involvement or interest’ with uses of the dative case that derive from the syncretism with the instrumental case (1985: §1345). As a preliminary to discussing the uses of the dative case, section 2.1 makes some brief remarks on the instrumental case in OE and its relationship with the dative case.

2.1 Dative and instrumental As Mitchell (1985: §1345) notes, the instrumental case was a syncretic one which combined forms and functions of other cases such as the ablative and locative. It had a marginal position in OE, never having a form distinct from ¹ For readers interested in the encoding of different thematic roles in the dative case, Maling’s (2001) study of three Germanic languages is a useful reference. Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

26

    :  

the dative in nouns of any gender or in feminine adjectives or determiners. The only distinctive forms were in masculine and neuter singular adjectives and singular determiners and the genderless interrogative hwy ‘why’. The demonstrative forms þon and þy were distinctively instrumental but did not distinguish masculine from neuter gender; as with other modifiers, there was no feminine form distinct from the dative. Consider example (2.1): (2.1) forþan ic hine sweorde swebban nelle because I him: sword: kill not.will ‘because I will not kill him with a sword’ (cobeowul,22.681.574, l. 600) Beowulf 679 Forþan illustrates the point that the traditional label instrumental for forms like the þan in forþan ‘reflects neither their origin nor their prevailing use’ (Campbell 1959: §708, n. 4). The noun sweorde is used in an instrumental sense, but I have glossed it as dative, as the YCOE does. This glossing is justified on the basis that the dative case was undoubtedly a living case category for nouns all through the OE period, in opposition with the accusative case, while nouns used in an instrumental sense never had a marking distinctive from the dative case. Example (2.2) exemplifies the difficulties involved in deciding how to deal with the instrumental case, where the glosses for case reflect the YCOE’s treatment: (2.2) þa geascode he þone cyning lytle then found he the:.. king:().. little: werode band: ‘then he discovered the king with a small retinue . . . ’ (cochronA-CC,ChronA_[Plummer]:755.10.514) The YCOE tags the form lytle as instrumental, since its dative form would be lytlum, but this is rather awkward because werode is tagged as dative. This treatment follows the principle of determining case labels on the basis of form—there is no distinct instrumental form for nouns—but sacrifices the principle of case agreement. Of course, the tagging used in this corpus is not intended as a morphological analysis, but only as a help to researchers in searching for forms and constructions. For our purposes, there is no need to make any decisions about how to treat the instrumental case, and we will not generally be looking at dative forms in instrumental or adverbial uses. Their main relevance to this investigation is the fact that their disappearance in ME

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.2  

27

is part of the general shift away from the use of case to express semantic relationships, but the discussion of the instrumental case should also be helpful in understanding the glossing of some examples. In this book, instrumental forms of determiners preceding an indeclinable particle to introduce subordinate clauses, e.g. the instrumental form þan in for þan þe, are glossed as such.

2.2 Verbal objects As is familiar from many languages, the dative case was used in OE on the objects of verbs and prepositions. No attempt will be made here to cover all the possible uses of the dative case as the case of objects here, but it is useful to note some of the more common uses of verbal objects.

2.2.1 Objects of ditransitive verbs The dative was the usual case of indirect objects, as in (2–3):² (2.3) Ðisne anweald forgeaf Crist þam this:.. power:(). gave Christ the:.. apostolum. & eallum biscopum apostles:(). and all:.. bishops:(). ‘Christ gave all the apostles and all bishops this power’ (cocathom1,ÆCHom_I,_16:309.75.2984) In (2.3), the dative case is clearly marked on both the determiner and the noun of the indirect object, and the accusative case is unambiguously marked on the fronted direct object. The ditransitive construction with a dative and an accusative object raises some difficulties for the discussion of DEPs when the dative object is understood to be the possessor of the accusative object. As mentioned in Chapter 1, Lee-Schoenfeld (2006) treats the possessive relationship with ditransitive verbs in German as being a pragmatic, rather than a syntactic one. I adopt this non-syntactic treatment. So for example some verbs of deprivation such as

² The term ‘indirect object’ is sometimes extended to the single objects of verbs in the dative case. I will be limiting this term to objects participating in a ditransitive construction and refer to dativemarked objects of monotransitive verbs as dative objects.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

28

    :  

ætbregdan ‘take away, take back’ could take a dative-accusative case frame as in (2.4): (2.4) Ic ætbrede him ða stænenan heortan. I remove them: the: stone: hearts: ‘I will take their hearts of stone away’ (cocathom2,+ACHom_II,_12.1:117.250.2546) In such examples, the accusative NP is necessarily possessed, both because a body part is inalienably possessed and because to have something taken away one must have it to start with. The dative can be thought of as having a ‘shared’ dative, since it is part of the case frame of the verb but is also interpreted as referring to the possessor, and I have accordingly not included examples like (2.4) in my counts of DEPs. With such datives, an overt genitive can be added to express possession, but another dative is never added. I assume that when no overt possessor is added, the possessor is understood to be the dative object due to the pragmatics of the situation, but an overt possessor can be added for clarity. Another dative-accusative construction that looks at first like a DEP involves the combination lætan blod, ‘let blood’, the medical procedure of bleeding a patient. The expression is especially common in Bald’s Leechbook and other medical texts. The patient does not have to be mentioned, as in (2.5a), but if he or she is, it is always with the dative case, as in (2.5b): (2.5) a. ac ærest mon sceal blod lætan but first one shall blood: let ‘but first, one shall let blood’ (colaece,Lch_II_[2]:1.2.3.2137) b. Læt him blod þus let him: blood: thus ‘Let blood from him thus’ (colaece,Lch_II_[2]:32.1.6.2736) There are good reasons not to treat the dative of (2.5b) as a DEP. If we did, we would have to treat this as a unique situation in which a DEP is the only possible expression of the possessor. It seems better to treat the combination blod lætan as a single lexical entry that optionally takes a dative object playing the role of source (or recipient). This is essentially the treatment that the DOE gives in its entry for blod in its sense A.2.b blod (for)lætan ‘to let blood, to bleed’. Compound nouns like blodlætere ‘blood-letter, blodlæs(wu) ‘blood-letting’ also indicate that letting blood was seen as a specific activity. This treatment of blod lætan is further supported by the fact that let x blood examples continue to be

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.2  

29

found frequently in the m3 period, considerably after the last clear examples of DEPs had disappeared. While the dative-accusative pattern was the most common one with ditransitive verbs, it was not the only one. In another common pattern, one of the objects is in the genitive case and the other either accusative or, less commonly, dative, as in (2.6): (2.6) and het hire ofteon ætes and wætes, and ordered her: deprive food: and water: ‘and ordered that food and water be withheld from her’ (coaelive,ÆLS[Agatha]:128.2091) Another case frame is possible, however, with some verbs of deprivation. In this frame the deprivee is in the accusative case, and the object of deprivation, rather than the human object, in the dative: (2.7) a. Hi þærrihte hine wædum bereafodon. They immediately him: clothes:(). deprived ‘they immediately stripped him of his clothes’ (cocathom1,ÆCHom_I,_29:425.206.5806) b. Þa het he hine heafde beceorfan then ordered he him: head:(). cut ‘Then he ordered him to be beheaded’ (cobede,Bede_1:7.38.2.309) This construction, called the ‘Direct object + ablative object’ by Visser (1963: §680), is particularly relevant to the study of possession in OE because the accusative and dative NPs can be regarded as being in a possessive relationship. In contrast to the DEP, however, it is the possessum, not the possessor, that is in the dative case in this construction. While the examples in (2.7) are parallel in the case marking of the NPs, there is some reason to treat them rather differently, and the YCOE does this by treating the dative heafde of (2.7b) as NP-DAT-ADT, i.e. an adjunct in the dative case, while in (2.7a) wædum is tagged simply as NP-DAT. Looking at beceorfan in the DOE, we see that the verb only appears thirteen times in the OE texts. Heafde beceorfan, which accounts for twelve of these occurrences, is treated as an idiom by this dictionary. The other example has feaxe becurfe ‘hair: cut-off ’. In contrast, the DOE notes ‘ca. 175 occurrences’ of bereafian, which can occur with either a dative or a genitive as the case of the derived thing, as well as other syntagms. There seems no reason to treat bereafian other than as a verb that has as one of its productive case frames an accusative object denoting the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

30

    :  

person and a dative object of deprivation. This is a pattern found with a number of verbs of deprivation. Beceorfan was apparently not a very productive verb, and the pattern with an accusative object of person and a dative object of what could be seen as a possessum is a very limited pattern, as discussed in section 1.4.1. Havers (1911: 275) lists examples with heafde becearf as his only illustrations of his category Ic (body part in a ‘locative’ construction) and points the reader to references documenting the origin of this dative as an instrumental. It seems likely that the use of the dative case derives historically from the instrumental case, with a meaning something like ‘with respect to the head’. It seems best to treat the dative found with this verb as a relic of an older instrumental pattern, limited to fixed expressions, although it is worth noting that heafde beceorfan is not limited to poetry or early prose, and the DOE’s feaxe becurfe example in fact comes from the late OE writer Ælfric, who also accounts for one of the heafde beceorfan examples. As will be discussed in Chapter 7, some EME texts show an extension of the genitive case to the object of deprivation with bikerven, the reflex of beceorfan. Examples like those in (2.7) will not be counted in the statistics presented in this book, since those statistics are restricted to dative, not accusative possessors.

2.2.2 Direct objects with monotransitive verbs The usual case of a direct object, i.e. the single object of a verb, was the accusative case, but a single object could also be in the dative case: (2.8) ond æt guðe forgrap Grendeles mægum and at battle destroyed Grendel: kinsmen:(). ‘and destroyed the Grendelkin in battle’ (cobeowul,73.2345.1918) Beowulf 2353 As with ditransitive verbs, there is a good deal of variation. Some verbs, such as derian ‘to harm’, occur mostly with dative objects, but the DOE notes a small number—fewer than ten—of instances of a clearly accusative form are found with this verb, mostly in translations.³ Also, the object of some verbs, e.g. helpan ‘to help’, could be in either the dative or the genitive case. The ³ It can also be noted that at least some of these examples are manuscripts that date to the early Middle English period, although the DOE treats them as OE because of their time of composition. In texts from such manuscripts, accusative case may be used for single objects of monotransitive verbs that normally governed a dative (or genitive) object in OE.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.3     

31

variability in the possible cases that a given monotransitive verb allowed for its object results in some difficulties for the study of DEPs when monotransitive verbs combine with prepositional phrases, as discussed in Appendix C, due to the syncretism of some case forms.

2.2.3 Reflexive objects Although the word self was used in OE, particularly in an emphatic meaning, simple personal pronouns could be used reflexively, either as true reflexives with transitive verbs, or as a ‘lexical reflexives’ with a number of intransitive verbs. Reflexive objects could be in the accusative (2.9a) or dative (2.9b) case: (2.9) a. and wende hine eft and went him: then ‘and then went from there’

þanon thence Genesis B 493

b. and gewende him þanon and went him: thence ‘and went from there’ (coaelive,ÆLS[Agnes]:381.1985) The use of personal pronouns reflexively continued after the collapse of the dative/accusative distinction and will not be further discussed in this book, but it is relevant to the discussion of some examples.

2.3 Complements of adjectives and nouns In Modern English, adjectives and nouns cannot take a bare NP as a complement; a prepositional phrase is needed in this function. In OE, NPs could take complements in the genitive or dative case as lexically selected by the head adjective. (2.10) ðeah þe he gecweme were Gode though that he pleasing were God: ‘even though he might be pleasing to God’ (coaelive,ÆLS[Pr_Moses]:240.2993) The complement could be separate from the adjective, as in (2.10). This example is one of the less common types in which the dative complement is to the right of the adjective; a position to the left is by far the more common, especially since the dative NP was most often a pronoun, where the leftward

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

32

    :  

position is the rule. The dative NP is frequently adjacent to the adjective, in which case it is not possible to be certain whether it is internal to the adjective phrase or external to it: (2.11) Þas þry hadas sindon Gode gecweme these three orders are God: pleasing ‘These three orders are pleasing to God’ (cocathom1,+ACHom_I,_9:255.203.1755) A reasonable analysis for these dative complements in a theoretical framework that uses movement is that the dative complement is generated within the adjective phrase and then can be moved out of it. The same effect can be achieved in a non-derivational framework like Lexical-Functional Grammar by allowing dative NPs to be generated in various positions in a clause with the information that identifies them as complements of an adjective in that clause. Difficulties of analysis arise when a dative complement to an adjective is understood to be the possessor of a body or mind word in a prepositional phrase, as discussed in sections 4.8 and 6.3.

2.4 Copulas, ‘impersonal’ constructions, and ‘extended existence’ 2.4.1 Simple clauses with copulas OE used a number of constructions involving a copula and a dative that are traditionally analysed as ‘impersonal’ or subjectless. One involving an adverb is presented in (2.12): (2.12) Wel bið þæm þe mot . . . Well is that: that may ‘It is well for the one that may . . . ’ (cobeowul,8.183.150) Beowulf 186 A natural translation into Modern English uses a ‘dummy’ or ‘expletive’ pronoun it, but this is not the only possible analysis. Thráinsson (2007: 160–1), in his treatment of similar-looking Icelandic examples, assumes that the dative experiencers in (2.13) are non-nominative subjects: (2.13) a. Mér er kalt me: is cold:. ‘I am cold’ (Thráinsson 2007: 160, example 4.34a)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.4 , ‘’   ‘ ’

b. þeim liður them: feels ‘They feel fine’

33

vel fine (Thráinsson 2007:159, example 4.31c)

The verbs in these examples do not agree with any element in the clause and are in the unmarked third person singular. Thráinsson notes that dative subjects are quite common in Icelandic. A similar analysis could be suggested for examples like (2.12). Another area in which a subjectless treatment might be suggested, and where a non-nominative subject analysis is less plausible, is in expressions of existence. Important differences exist between OE and later English in expressions of existence, and since some of them involve elements in the dative case, some discussion is needed here. Before introducing examples of this case used in expressions of existence, it is useful to note that þær ‘there’ was still a purely locative pronoun in OE, and is not used in ‘presentational’ sentences such as the later English there is a God.⁴ Linguists differ on assumptions about the presence of non-overt ‘expletive’ pronouns in presentational sentences. Of particular relevance to the discussion of possession in OE are sentences in which the presentation of the existence of something is combined with a dative: (2.14) Denum wearð . . . ealuscerwen. Danes:. became: terror:().. ‘There terror arose among the Danes’ (cobeowul,25.767.660) Beowulf 767–9 Such sentences are examples of what Heine (1997: 57) calls an ‘extended existence’ structure that encodes the general formula ‘X exists with reference to Y’.⁵ Here, it is doubtful whether we would want to speak of possession. At a stretch, this sentence (in which various phrases in the dative case that are in apposition to denum have been elided as not affecting the relevant structure) could possibly be translated as ‘the Danes had terror’, bringing out a possessive reading. One possible analysis of (2.14) is that ealuscerwen is the subject and denum a dative complement to wearð. However, given that adjectives could clearly take dative NPs as complements in OE, it is reasonable to suggest that NPs could also take such complements. By such an analysis, ealuscerwen denum would form a single NP, either by unification in a unification-based ⁴ The Dictionary of Old English presents three examples of ‘initial þær’ in A1.ii.b of its lemma beon ‘be’, but in all these examples the þær can be taken as indeterminate between having a locative and a purely presentational sense. Such uses undoubtedly served as a ‘bridge’ to later presentational there. ⁵ ‘Extended’ in involving two arguments, as opposed to ‘nuclear’ existence, with a single argument.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

34

    :  

framework or before movement of the dative NP in a framework that assumes such movement. This NP would play the role of subject, unless we analysed the sentence as having an unexpressed expletive subject. In (2.15), a meaning related to some sense of possession is clearer; had is much more natural than was in the translation: (2.15) Him wæs an fæder them: was one: father: ‘they had one father (in common)’

(coexodus,101.353.291)

However, it is possible here too to analyse him as a complement to an fæder. As Heine discusses at some length in the first chapter of his monograph, ‘possession’ is a difficult notion to define. The important point here is that extended existence ‘is a common source for predicative possession’ (1997: 58). One of the schemas that arises from expressions of existence is what Heine calls the Goal Schema, which he describes with the formula Y exists to/for X > X has, owns Y (1997: 59). The dative case is one common way of encoding the possessor, and in Latin, for example, a clear meaning of ownership can be conveyed using the dative case for the possessor. Heine demonstrates the Goal Schema in Latin with this example from Lyons (1967: 393): (2.16) Est Johanni liber Is John: book: ‘John has a/the book’ The present study focuses on attributive, rather than predicative possession, and makes no systematic study of examples in which the dative is used in predicative possession. It can be briefly mentioned, however, that examples parallel to (2.16), in which there is a clear meaning of ownership, seem not to be found in Old English. Example (2.14) is included in Visser’s (1963: §361) list of nouns that occur with the copulas beon and weorþan accompanied by a dative noun. These nouns typically refer to states of mind and emotions, as in (2.14), or situations, such as þearf ‘need’ and sib ‘peace’; none of them refers to the ownership of physical objects like books.

2.4.2 Dative + noun + noun While an analysis of the dative as an external possessor does not seem very suitable with the nouns discussed in the preceding section, the same is not true when there is a second NP in the nominative case. When the clause involves

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.4 , ‘’   ‘ ’

35

three NPs rather than two and involves a relationship between the two nominative NPs, as in the examples of (2.17), a translation into Modern English with a possessive is often natural: (2.17) a. þæt he Heardrede hlaford wære that he Heardred: lord: were: ‘that he should be Heardred’s lord’ (cobeowul,74.2373.1933) Beowulf 2375 b. wæs þæm hæftmece Hrunting was the:.. hilted.sword(). Hrunting nama. name: ‘the hilted sword’s name was Hrunting’ (cobeowul,45.1455.1201) Beowulf 1457 Such examples could be treated as instances of external possession. Adopting a ‘possessor raising’ analysis, we could say that in (2.17a) Heardrede is generated within the predicate nominal headed by hlaford and would surface as a genitive if nothing else happened, or would surface as a dative if it were raised to clause level, with no difference in meaning except perhaps a focus on affectedness with the dative. A non-movement analysis would relate the dative anaphorically to the possessor position within the predicate nominal. This analysis of these datives as DEPs seems at first to be supported by the fact, noted by Visser (1963: §351), that with many of these nouns a genitive is an alternative; we find both godes freond ‘God’s friend’, with a genitive, and gode freond, with a dative. Visser goes on to object to treating these as exact equivalents, citing the difference between my father and a father to me. As Visser also notes, the dative was often replaced in later English by the preposition to. In §333, where the construction with an adjective is discussed, he notes the replacement of the dative with a prepositional phrase, e.g. dear to him. These facts make an analysis with a dative complement to a noun natural. So in (2.17a) we can analyse he as the subject and hlaford as a predicate nominal, the noun taking Heardrede as a dative complement, which would surface as a dative whether or not it ended up in its ‘logical’ position within the predicate nominal or outside of it. Once again, a systematic investigation of the use of datives with words not referring to the body or the mind goes beyond the scope of this book, but it can be noted here that the use of datives with nama ‘name’, as in (2.17b), offers an interesting example of how the use of dative case became more restricted

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

36

    :  

within the OE period. Examples parallel to (2.17b) in consisting of a single clause with a copula seem to be limited to poetry. In prose all through the OE period, a genitive possessor is used rather than a dative. The dative remained common when it served as a relative pronoun, however: (2.18) Her Cerdic & Cynric ofslogon ænne Brettisc cyning, Here Cerdic and Cynric slew a British king þam was nama Natanleod. whom: was name: Nazaleod ‘in this year Cerdic and Cynric killed a British king whose name was Nazaleod’ (cochronA-1,ChronA_[Plummer]:508.1.173) By the Late West Saxon period, a genitive relative pronoun, which of course had always been an option, was the norm: (2.19) sum swiðe wundorlices lifes wer, þæs nama some very wonderful: life: man, who: name: wæs Seuerus. was Severus ‘a certain man of a very wonderful life, whose name was Severus’ (cogregdC,GD_1_[C]:12.87.29.1007)

2.5 Dative + PP with copulas In another common type of ‘extended existence’ pattern in OE with the copulas beon ‘be’ and weorþan ‘become’ and a few other presentational verbs, the predicate was expressed as a prepositional phrase headed by to rather than a simple NP: (2.20) a. þeah ðu þinum broðrum to banan wurde though thou thy: brother: to bane: became ‘although you were your brother’s slayer/killed your brother’ (cobeowul,20.583.495) Beowulf 587 b. noldon þæt se maðm wurde mannum to frofre. not.would that the treasure became men:. to comfort ‘they did not wish the treasure to become a comfort to men’ (coaelive,ÆLS_[Exalt_of_Cross]:4.5555) Visser 358 (1963: §357) lists a large number of nouns occurring with a dative in this construction, with examples spanning the entire OE period. As Visser notes, with these prepositional phrases the verb weorþan ‘become’ is more frequent than the verb beon ‘be’. This is not surprising given the core meaning

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.6  +    

37

of to as direction towards a situation or attainment of a position; the use of this preposition focuses on a final situation. Some nouns such as bana ‘killer’ are found both in the to and the simple dative NP construction. The analysis of this construction is not straightforward, but if we adopt the analysis suggested in section 2.4.2 of nouns taking dative complements, we can analyse the datives of (2.20a,b) as complements of cyninge and frofre, respectively. The predicative constructions of this section will not be discussed further in this book except in the discussion of the general decline in EME of constructions using nouns that had been in the dative case in OE.

2.6 Dative + to with other verbs We have seen how the core directional meaning of to was extended to state that a relationship existed or came into being. Related to the construction with a copula just discussed, to was widely used in OE in collocations explicitly indicating how such a relationship came into being. Here, no ‘possessor’ need be overtly expressed: (2.21) a. þa ða he wæs to cyninge gecoren. then when he was to king chosen ‘when he was chosen as king’ (coaelive,ÆLS_[Book_of_Kings]:2.3659) b. Ne mæg nan man hine sylfne to cynge make Not may no man him: self: to king: gedon ‘No man can make himself (a) king’ (cocathom1,ÆCHom_I,_14.1:294.111.2654) The to is often best translated with ‘as’ in Modern English, as in (2.21a), and sometimes best omitted, depending on the verb. In the examples of (2.21) ‘king’ can be treated as a secondary predicate. Visser (1963: §665) provides a list of verbs that appear with what he calls a ‘predicative adjunct’ with to or for. The relevance of this construction to this study is that when the ‘possessor’ was expressed, it could be, and with some nouns at least always was, expressed as a dative NP: (2.22) a. and geceas him to cynincge þone cenan and chose them: to king: the: valiant: Dauid David: ‘and chose the valiant David as their king’ (coaelive,ÆLS_[Book_of_Kings]:12.3668)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

38

    :  

b. and his agen swustor him geceas to wife and his own  sister: him: chose as wife ‘and chose his own sister as his wife’ (coaelive,ÆLS_[Chrysanthus]:103.7392) Adopting the analysis for the dative with these prepositional phrases suggested in section 2.5, the dative can be treated as the complement of nouns such as cyninge and wife. Further discussion of the dative + PP constructions is beyond the scope of this book, which is concerned with body and mind words, but it can be noted that they are different from the DEPs with PObjs studied here in a number of ways. Most importantly, the ‘possession’ is predicative rather than attributive. While the existence of a body part is presupposed in a DEP, in the examples of (2.22) the existence of a king or a wife is asserted, and there was no wife or king until the choosing took place. The construction seems to be lexically selected. The nominal in this construction never has a definite determiner in it, unlike the body part PObjs with DEPs discussed in Chapter 4. Another difference is that with the DEPs under study in this book, an IP is normally an alternative, but I have not found ceosan to his/heora cyninge/wife ‘choose as his/ their king/wife’.⁶ The dative seems to retain a benefactive or malefactive colouring. Note that when there is already a dative present because the sentence contains a ditransitive verb like sellan, no second dative appears: (2.23) þæt he his dohtor sealde Octauiane to wife that he his daughter: gave Octavian: to wife ‘that he gave his daughter to Octavian as (his) wife/for (his) wife’ (coorosiu,Or_5:13.129.10.2734) In (2.23), Octavian is the recipient, and no second dative is necessary to establish that by the gift he becomes the possessor of the wife. Indeed, a second dative with a ditransitive verb seems to be not only unnecessary but nonoccurring, as previously discussed in section 2.2.1.

2.7 ‘Free’ and ‘adverbial’ datives The datives that have been discussed so far are complements, either of a verb or of a noun or adjective. Nouns in the dative case were also used as adjuncts. ⁶ While the generalization that the dative is the only way of expressing the relationship with wife and cyninge seems to hold, further investigation may find that genitives are possible with some other nouns, as I have not attempted an exhaustive collection of examples other than with these two nouns.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.7 ‘’  ‘’ 

39

Mitchell (1985: §§1408–27) summarizes what he terms the ‘adverbial uses’ of the dative case, including to indicate location, time, etc. These uses will not be covered here. On the other hand, something needs to be said about what are often referred to as ‘free datives’ because these need to be distinguished from DEPs. As discussed in Chapter 1, there is disagreement about whether a DEP should be analysed as an argument of a verb. It is clear, however, that a DEP is an argument of something; it bears a thematic relationship to the possessum whether it is analysed as being generated within the NP containing the possessum or not. Free datives, on the other hand, are clearly not arguments, but merely add some sort of emotional colouring to a sentence. As far as we can tell from the texts, the range of free datives in OE was more restricted than in some early Germanic languages and in Modern German. LeeSchoenfeld (2006: 105) gives (2.24) as an example of a sentence that has what looks as if it might be a dative possessor but turns out to be a sentence dative: (2.24) Mein Bruder hat der Mami leider ihr Auto my brother has the: Mom: unfortunately her car: zu Schrott gefahren to scrap driven ‘Unfortunately my brother totalled Mom’s car’ (Lee-Schoenfeld ex. 6) In this sentence Mami is the owner of the car, but she gets dative case not because she is the owner, but rather because of the deleterious effect the totalling of her car had on her. Mami could also have been in the dative case if it had been someone else’s car that had been totalled but she was badly affected by this, so is a sense the fact that she is the possessor of the car is beside the point. Lee-Schoenfeld observes that with a real possessor dative, it is not grammatical to express the possessor again in the genitive. In OE, however, there are a number of examples that combine a dative and a genitive in what Ahlgren (1946) termed the ‘blended’ construction, as in (2.25): (2.25) a. Monegum men gescrincað his fet to his homme many:. man:. shrink his feet to his hams ‘Many a man’s feet shrink up towards his hams’ (colaece,Lch_II_[1]:26.1.2.812) b. Gif monnes eage him mon ofaslea if man: eye: him: one strikes If a man’s eye is struck out’ (colawaf,LawAf_1:71.203)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

40

    :  

The ‘blended’ construction will be discussed in the chapters that present the findings of the investigation. One possible treatment of this construction that combines the dative and the genitive case in OE is that the dative is a sentence dative combined with a normal IP, rather than a double expression of the possessive relationship. How exactly we would incorporate it into a formal treatment would of course depend on how we treat DEPs with no accompanying IP. Let us suppose that with a simple DEP we have the addition of an argument bearing a semantic role such as  to the basic argument structure of a verb, with some sort of formal mechanism linking this argument to the possessor of another argument. The blended construction could be treated as a normal IP with the addition of an adjunct dative that indicated that a person involved in the action was affected by that action. The difference would be that in this instance, there would be no formal linking of the dative to the possessor; rather, the coreference of the dative and the possessor would be inferred from the context. The dative NP referring to the affected person, being highly topical, is usually in the topical position at the front of the sentence, as in (2.25a), but it is also possible for the genitive to be placed in the topic position, as in (2.25b). The reason for this positioning in the latter example is probably that it is part of a list of compensations for injury to various body parts, so the body part is highly topical, but the dative is added to emphasize that this is not something that just happened to a person’s eye, but something that affected the person. The example is of interest in indicating that however we analyse the blended construction, it should not be seen as the simple addition of a genitive to reinforce the dative—if anything is to be seen as an addition in this example, it must be the dative. It is important to note, however, that sentence datives seem to have been more limited in OE than they are in modern German. One difficulty in studying sentence datives in OE is that in the absence of native speakers, we have limited information about the argument-like or other syntax of particular datives. A second difficulty in determining the range of sentence datives in OE is the fact that the terminology that scholars have used is not uniform. A case in point here is the ‘ethic’ or ‘ethical’ dative. There is a good deal of variation in the way the term ‘ethic’ is applied to datives. Lockwood (1968: 24) distinguishes the ethic dative from the dative of interest, where the dative marks a person directly involved in an action. He explains that with the ethic dative, the person’s involvement is more detached, but the dative conveys a pronounced emotional content. He illustrates the ethic dative in colloquial Modern German with the example du bist mir ein feiner Kerl

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.7 ‘’  ‘’ 

41

‘you’re a fine sort, I’m telling you’ (lit. you are me: a fine fellow), and also supplies some examples from Luther and Goethe, where the dative seems to mean something like ‘as far as I care’ or ‘as far as I am concerned’. LeeSchoenfeld (2006) gives this example to illustrate the ‘ethical’ dative in Modern German: (2.26) Schlaf mir jetzt schön ein, Kleines! sleep me: now nicely in little ‘Kindly fall asleep for me now, little one’ Lee-Schoenfeld 2006:105, ex. 7 This example is similar to Lockwood’s in the speaker’s lack of direct involvement in the action, but is rather different in that it does not just indicate some sort of attitude towards the proposition, but a keener interest, since the child’s falling asleep will benefit the speaker. Mitchell’s (1985: §1352) use of the term is similar to Lockwood’s. He quotes the definition given by the first and second editions of the Oxford English Dictionary: the dative when used to imply that a person, other than the subject or object, has an indirect interest in the fact stated Oxford English Dictionary, ethical, a. [Second Edition]

Mitchell says that neither he nor Visser (1963: 695) found any convincing examples in Old English of the ethic dative in the more restricted sense. However, the term is often used to cover a broader range, as Mitchell notes, and the OED’s definition of the ethic dative has changed since Mitchell’s book was published. In its most recent edition (Third Edition, 2014), the OED has removed the exclusion from its definition: a use of the dative case signifying that the person denoted has an interest in or is indirectly affected by the event Oxford English Dictionary, ethic, n. and adj. [Third Edition]

This broader and vaguer definition does not exclude an interpretation whereby a participant in the event might be in the dative case because of an interest in the event. Visser (1963: §320) indicates that the term ‘ethic dative’ has been used to cover examples such as him brekeþ þe sweore ‘his neck breaks’, in which the dative is treated as a DEP in this book. In such a use of the term ‘ethic dative’, the focus on affectedness and interest ignores the syntactic fact that the dative in such examples is not simply a ‘free’ adjunct,

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

42

    :  

as it is when it merely adds some emotional colouring, but must be interpreted as the possessor of the body part. These different definitions of the ethic dative make it difficult to determine how different Old English was from the other early Germanic languages. Also, not all grammars of these languages mention the ethic dative, although they mention some other uses of the dative. Grammars that do specifically mention the ethic dative tend to use the term in the broader rather than the narrower sense. Krahe (1972: 89) treats the dativus ethicus as a subtype of the dative of interest and gives (2.27a) as an example of the ethic dative in Gothic. This example can be compared with the West Saxon version of the same Bible verse, which uses a genitive: (2.27) a. nim þus bokos take thee: book:7 ‘Take your book’ b. nim þine feðere take thy: quill: ‘Take your pen’

Luke 16.6

Although the OE version of this Bible verse does not use an ethic dative, examples can be found which could be called ethic datives:⁸ (2.28) aris nu and hafa ðe Naboðes wineard: arise now and have thee: Naboth’s vineyard: ‘arise now and take Naboth’s vineyard for yourself ’ (coaelive,+ALS_[Book_of_Kings]:200.3811) Such examples do not fit under the narrower definition of the ethic dative, since the understood subject of the command is directly involved in the action. Van der Wal and Quak (1994: 103) say that the ethic dative was common in the old and middle continental West Germanic languages, especially in Old Saxon. However, the example they use to illustrate this dative, tho geng im thanan ‘then he went away’ lit. ‘then went him: away’, also does not fit under the narrower interpretation of the ethic dative. Similar examples are

⁷ The form bokos is formally plural, but was used to signify a collection of writing, i.e. a single book. Krahe translates the sentence into German as nimm dir den Brief (‘take thee: the: book:’). ⁸ Example (2.28) is cited under the sense IIB1.c. ‘to take, seize’ of habban used with an accusative direct object and a dative indirect object in the Dictionary of Old English. It does not appear that habban occurred freely in OE with an inherent reflexive. A final decision on the best analysis of (2.28) would require a thorough investigation of the uses of datives with habban.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

2.8 ‘’      

43

found in OE; an example was given above as (2.9b), where it was treated as a reflexive object. Such examples are not ‘free’ datives because they are lexically selected; not every verb can be used reflexively. Because of the lack of information about datives falling under the strictest definition of ethic dative, it is difficult to determine whether OE was already an odd man out in the Germanic languages because it is difficult to tell exactly what the situation was in the early sister languages and what can be reconstructed for Common Germanic. In this connection, it is interesting to note that Visser (1963: §695) says that it is difficult to trace expressions like he ate me up half a ham of bacon, used colloquially in some varieties of English, back to earlier English. He considers that the first convincing examples are from the sixteenth century; he gives examples from Shakespeare including she has me her quirks (Pericles IV, vi, 7). This is interesting because it is natural to assume that such examples are a continuation of a type of free dative in Old English, but this cannot be proven. An argument that datives were used in speech but not in writing is hardly convincing, since—apart from the fact that this is an unfalsifiable hypothesis—there is no reason why such datives would have been considered colloquial and therefore suppressed in writing in Anglo-Saxon times. Since Visser did not find indisputable examples from before the beginning of the sixteenth century, apart from an isolated Middle English example, expressions like he ate me up half a ham of bacon must be a fairly late innovation. If this use could arise in English after the period when English had lost the dative/accusative distinction, it would be just as natural for the ethic dative to arise in Gothic or German, a language with such a distinction, and the use of this dative in German cannot simply be assumed to be a continuation of Common Germanic. More research is needed here.

2.8 ‘Impersonal’ constructions with datives with lexical verbs In addition to the copular constructions, the dative case entered into many constructions in OE that are traditionally categorized as ‘impersonal’ or ‘subjectless’ with lexical verbs. Some constructions traditionally treated as impersonal have been discussed in section 2.5 as examples of ‘extended existence’. To round out our discussion of the uses of the dative case, it remains only to mention that a number of verbs are found in OE used in what are traditionally analysed as ‘subjectless’ constructions:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

44

    :  

(2.29) a. and don swa hine lyste and do: as him: desires: ‘and do as he desires’ (cocathom1,+ACHom_I,_10:264.178.1959) b. him ofhreow þæs mannes him: pitied the: man: ‘there was pity in him for the man/he pitied the man’ (cocathom1,+ACHom_I,_13:281.12.2357) In (2.29), the accusative hine in the (a) sentence and the dative him in the (b) sentence are traditionally analysed as objects in subjectless constructions. However, the subject-like syntactic characteristics of similar non-nominative NPs in parallel constructions in Icelandic were firmly established by Zaenen, Maling, and Thráinsson (1985), and Thráinsson (2007) covers a wide range of constructions that he treats as having non-nominative subjects in Icelandic. While arguments for non-nominative subjects are weaker for OE, I adopt them in Allen (1995) and will continue to assume them here. The analysis adopted has no bearing on the discussion of possession in this book, but some knowledge of the ‘impersonal’ uses of the dative and accusative is helpful in understanding some of the examples discussed. Furthermore, these constructions merit mention here as a prominent use of dative case in OE, particularly because they continued to flourish in the EME period, as briefly discussed in Chapter 7.

2.9 Conclusion A number of uses of the dative case have survived in another form into Modern English; for example, the prepositions to and for are used to perform many of the functions performed by the dative case in OE, and with ditransitive verbs, the first of two bare postverbal NPs continues to be used as the indirect object. The dative case was also used in functions that have not been replaced by prepositions or word order, and one of the questions addressed briefly in Chapter 7 and more extensively in Chapter 8 is the relationship between the loss of DEPs and the loss of other constructions using the dative case. Before that discussion, however, the uses of the dative discussed in this chapter will serve as background for the investigation of DEPs in OE.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3 Investigating dative external possessors in the history of English 3.1 Introduction The contemporary linguist has a great advantage that earlier scholars like Havers, Ahlgren, and Visser all lacked, which is access to electronic corpora. The use of these corpora makes it possible to carry out systematic investigations into a multitude of questions in a relatively short time rather than depend on impressionistic conclusions based on the sometimes haphazard collecting of examples of different types. Furthermore, current linguistic theory naturally raises questions for investigation which would not have occurred to these earlier scholars. Any presentation of data depends on the theoretical framework of the researcher, whether it is traditional grammar, generative syntax, or whatever. Of particular relevance here is the fact that earlier research into DEPs in the history of English has not made paid sufficient attention to what turns out to be a crucial distinction in the organization of the data, namely the grammatical relation of the possessum. Typologists have found that a hierarchy of these grammatical relations plays an important role in whether an EP is possible in a specific situation in a language that has EPs available, as discussed in section 1.3. Mitchell (1985: §306) is unusual in making a reference to the fact that DEPs are more common in OE when the ‘noun of possession’ (i.e. the possessum) is the object of a preposition than when it is the subject or object of a sentence, but since Mitchell’s monumental work covers all of OE syntax, it is unsurprising that he only made a bare comment on this fact. I will show that this general failure in early research into the syntax of DEPs in Old and Middle English to differentiate DEPs of possessa that play the role of object of preposition (PObj) from DEPs of subject and object possessa, as well as a lack of attention to some other crucial distinctions, has resulted in some important generalizations being missed. While parsed corpora represent a great advance, they are a mixed blessing. For a general discussion of some of the pitfalls to be avoided, see Allen (2016a). Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

46

   

For this study, the main problem with the electronic corpora is that they are based on editions of texts that reflect the concerns of the philologist but not of the linguist. A major concern of the philologist is to construct a ‘good’ text, and this may involve piecing together bits from different manuscripts, sometimes dating from a century or more apart. A simple example is provided by YCOE’s coaelhom, based on Pope (1967), which is discussed in detail in Appendix A. With a few YCOE texts of this sort, it was fairly straightforward to split the YCOE texts into two files and treat them as belonging to different periods, as detailed in that Appendix. The reader checking the YCOE will naturally not find the names I have given to the new files in the text information provided with that corpus, but any examples taken from these files will have citations provided by the YCOE. Without further study, it is not clear how faithfully a late copy reflects the language of the composer, and the language of the composer is generally the linguist’s primary concern. This is not to say that late copies cannot be very valuable linguistically; they certainly are, for example, in indicating linguistic changes that have taken place since the period of the original composition, and with sufficient study we can infer what old constructions were still acceptable (if perhaps rather old-fashioned) at the time of copying as opposed to ones a copyist felt he or she had to replace. Until such careful study of the texts in individual manuscripts is made, however, the best evidence comes from texts in manuscripts that are reasonably close to the time and dialect of composition of the texts. Of course, we can only work with what we have and make the best of evidence that is not of the strongest nature, but we can at least separate out the texts from manuscripts that give the strongest evidence for a given dialect and period from those that do not. We can always add in the results from manuscripts of mixed dialects and periods to increase our numbers if the results in those texts seem consistent with those from the ‘purer’ manuscripts. We can also be aware of the fact that a given manuscript may be the work of more than one scribe, and it is possible that linguistic variation in a manuscript may reflect variation within the community, that is, some people using one variant and some using the other, or variation within the individual, with a single person using two variants. With this awareness, we can check out the details of scribal hands in a given manuscript when we want to make certain that a single scribe is using two or more variants. In order to reduce the amount of ‘noise’ that could be introduced by insufficient attention to the nature of the manuscripts, the systematic investigations of this study exclude several texts included in the YCOE that I consider to be unreliable witnesses to OE usage. For a detailed discussion of why

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.2  

47

these texts are excluded, see Appendix A. However, I have at times discussed some examples from the texts excluded from systematic study to illustrate a point.

3.2 The corpus I defer discussion of the corpus used for EME to Chapter 7. The investigation into Old English relies primarily on CorpusSearch queries applied to selected texts of the YCOE and the York Poetry Corpus, introduced in section 1.5. However, the York Poetry Corpus, which for some poems covers only a selection, yields a disappointingly low number of possessed body parts playing the role of subject or object (in contrast with numerous body PObj possessa). I have augmented the examples from the poetry corpus with ones taken from my own examination of all of Judith (Griffith 1997 edition), which is not included in the York Poetry Corpus, and the lines of Andreas (Krapp 1932: 3–51), Christ (Krapp and Dobbie 1936), and Genesis (Krapp 1931: 1–87) that are not included in the electronic corpus. Examples taken from the two York corpora are cited as in those corpora, although in the case of examples from Beowulf I have added the line number, since this is not straightforwardly retrieved from the York Poetry Corpus citation. The output of searches on files in the York corpus containing collections of poems, e.g. the Exeter Book, does not identify individual poems, and in any example taken from such output, I have supplied the name of the poem containing the example. Examples that are presented from poetry not included in the electronic corpus can be recognized by their citation of the name of the poem and the line number in the edition used. While statistics are presented for the individual prose texts in tables in Chapter 4, a better picture of trends emerges when the individual texts grouped together on the basis of temporal and/or dialect similarities. The OE investigation reported here distinguished six categories of YCOE texts, or text types, summarized in Table 3.1. The poetic texts are not individuated and are treated together as one text type, although some tables distinguish figures from the York Poetry Corpus from ‘non-York’ poetry figures. The tables of Chapter 4 that present figures for the individual texts also present totals for the text types. The prose texts in each category are set out in Appendix A. This investigation used all the texts from the York Poetry Corpus, they are not set out in the Appendix, and the non-York poetic texts that were used were set out above.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

48

   

Table 3.1 Text categories used in this book Text category

Brief description

Poetry EWS 9thC(OE) Other Early

All OE poetry ‘Anchor’ EWS manuscript Probable 9thC composition, manuscript later but before xi 3rd quarter Composition no earlier than early tenth century, mixed or non-West Saxon dialect Mixed or unknown composition, manuscript earlier than xi 3rd quarter Late West Saxon composition, manuscript fairly close to time of composition Late West Saxon composition, significantly later manuscript

General OE LWS LWS(Late)

The prose categories are mostly, but not entirely, based on periods, and where possible, dialect. The dates used for OE manuscripts in this book are taken from Ker’s (1957) Catalogue and use his system of Roman numerals to indicate centuries. However, in assigning texts to periods, I have considered not only the date of the manuscript containing a given text but also the presumed date of composition. Let us turn now to a discussion of the text types, starting with the prose texts. In prose, the early period is associated especially with the writings associated with the court of King Alfred the Great (died 899) but naturally includes any earlier writings. The late OE period is especially associated with the late tenth and early eleventh-century West Saxon writings of Ælfric and Wulfstan. Assigning writings to the early or late period is often not a simple matter in practice, however, because of the frequent discrepancy between the time of composition of a text and the date of the existing manuscript(s) containing it. With some texts, on the other hand, the assumed date of composition of a text is very close to the date of the manuscript containing it. This is the case with the ‘EWS’ category in Table 3.1. This category comprises the texts found in the four manuscripts upon which our understanding of Early West Saxon is based. I will refer to these texts as the ‘anchor’ EWS texts, adopting a term used by McIntosh et al. (1986) and Laing (1993) for texts whose provenance is known and can serve as ‘anchors’ for identifying the dialect of texts of less certain origin. The EWS texts were composed mainly in a West Saxon dialect and are found in manuscripts not dating later than the early tenth century. For information on these manuscripts and texts relevant to this investigation, see Appendix A. Here, I note that only a portion of the A MS of the Anglo-Saxon Chronicle (ASC) serves as an ‘anchor’ EWS text. It is only examples from this

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.2  

49

portion of YCOE’s cochronA that are included in the EWS examples and figures in this book; examples from later annals in this manuscript are included in the figures for General OE. Other early prose texts are in other dialects or show considerable dialect mixing. The clearest picture of any changes from early to late OE will emerge if we keep the number of variables to a minimum, and so these texts get a separate category ‘Other Early’. These texts are found in manuscripts dating from no later than the early tenth century. The ‘9thC(OE)’ category is for texts which are believed to have been composed in the late ninth century but are only preserved in manuscripts from a substantially later time within the OE period. For example, the laws that are attributed to King Alfred are found in the same (Parker) manuscript as the A version of the Anglo-Saxon Chronicle, but Ker (1957: item 39, art. 2) assigns a date of the middle of the tenth century to this hand. I have separated these laws from EWS because of the possibility that intervening scribes may have made changes such as replacing external possessors with internal ones. Similarly, Bald’s Leechbook may have been composed in King Alfred’s late ninth-century court, according to Nokes (2004: 74), but is only preserved in a manuscript from about half a century later. While both these compositions are mainly West Saxon, Gregory’s Dialogues, translated from Latin by Bishop Wærferth of Worcester at King Alfred’s command sometime between the early 870s and early 890s (see Yerkes 1982: 9), is in a Mercian dialect. For a discussion concerning the versions of the Dialogues, see Appendix A. The General OE category is a mixed one. All texts in this category come from manuscripts from before the third quarter of the eleventh century. Some have an unknown date of composition and may have an early origin. Some of them, for example Cnut’s Laws, are certainly of late composition, but are included in the General OE category because of dialect mixture. Where appropriate, I have used these laws to illustrate the late retention of DEPs. Note that some examples taken from YCOE’s cochronA fall into the General OE category. The LWS category is for texts composed around the turn of the eleventh century in the West Saxon dialect that are found in manuscripts reasonably close to their time of composition, that is, no more than fifty years later. This category is dominated by the writings of one author, Ælfric, but it also contains some nonÆlfrician compositions. The final text type, LWS(Late), also contains texts of West Saxon origin and language, but are found in late manuscripts. They are texts mainly by Ælfric and Wulfstan. As it turns out, there is (unsurprisingly) no significant

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

50

   

difference to be found with regard to the treatment of possessed body parts in the two text types, and examples from the LWS(Late) category have the useful effect of adding some non-Ælfrician DEPs to the data, and it has been possible to add the two categories together in discussing LWS usage. In contrast to the prose, it is not practical to try to separate poetry into periods. Most extant poetry from OE is found in four fairly late codices, but at least some verse is traditionally assumed to be of considerably early composition. Fulk et al. (2003: 197) note that towards the end of the nineteenth century, there was a broad consensus about the relative dating of OE poetry, but since then a better understanding has arisen about the fact that this poetry generally shared certain conventions regardless of the time of composition. They point to the importance of Amos’ (1980) findings concerning the unreliability of linguistic tests for dating poetry. Scragg comments: There are no sure objective tests by which poetry can be dated, and no means of proving which of the surviving poems were composed orally and which in writing. All Old English poetry is of such uniformity in form and language that it is impossible to establish even a relative dating with any certainty. (Scragg 1991: 57)

Whatever the date of composition, it is clear that the poetry in general was archaic in at least some of its features. For example, Beowulf is found only in a manuscript of c.1000, but is remarkable in its noticeably reduced frequency in the use of definite determiners, according to Amos (1980). Taylor (2014: 277) assumes that this poem represents earlier syntax generally, as does Walkden (2014: 225). It is likely that other poems also enshrine at least some earlier syntax. Unfortunately, there are not enough relevant examples to compare the expression of inalienable possession in different poems, but we will see that taken as a whole, there are enough poetic examples to suggest that the use of DEPs in poetry probably represented early usage. In the presentation of the data in the tables of Chapters 4 and 6, the figures for poetry appear in the tables that include the earlier prose. However, the poetry results are discussed separately from the prose results.

3.3 Methodology This section outlines the methodology used in this study to investigate the questions surrounding the use of DEPs in written OE that have been discussed so far. The methodology used for the Middle English investigations is similar,

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.3 

51

but adapted to different circumstances as necessary, and is discussed in Chapter 7. The focus of the discussion in this chapter is on OE body words; special features of the methodology used with mind words are discussed in Chapter 6. A much more detailed account of the methodology for all the data gathering is given in Appendix C. As indicated by the divisions made in section 3.2, one goal of the searches is to address the question of whether the frequency in the use of DEPs decreased within the OE period. Two other basic questions within the text categories are these. First, how were DEPs used with body and mind words in the texts? Second, how does this use compare with the use of IPs with these words? The initial searches described here were aimed at these questions, but I also carried out some additional searches into questions that arose from my initial searches, and these will be described where relevant. In the review of previous research in section 1.4, the important point was made that previous discussions of DEPs in OE have mostly not distinguished between what I am calling ‘direct’ argument possessa (subjects and objects) and PObj possessa, obscuring some important generalizations and giving a distorted impression of how frequently DEPs were used in OE. This investigation deals with these two broad types separately. The methodology used for the direct argument possessa and prepositional object possessa differed in certain respects, and these differences will be detailed in the following subsections devoted to these two categories. Some particularly important constructions with datives that were not treated as DEPs in this investigation are discussed in Chapter 2, and for further details of how the searches dealt with various problems such as identifying dative possessors, dealing with case ambiguity, the fact that possessive phrases were sometimes discontinuous, etc., see Appendix C.

3.3.1 Direct argument possessa The foundation of the electronic searches was a list of (variant forms of) more than eighty words for body and body parts, compiled by consulting the Thesaurus of Old English (Roberts et al. 2000) as well as running lexicon searches on the texts in YCOE. The same procedure was followed for the mind words. The lists of body and mind lemmas are set out in Appendix B. The corpus searches fell into two basic types, those for IPs and those for DEPs. Separate searches were used to identify subject and object possessa. The following discussion only sketches the basic approach of the searches for body

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

52

   

words. The somewhat different searches I made for mind word are outlined in Chapter 6. The searches for DEPs specified a body word playing the role of subject or object (depending on which was being investigated) in a clause that also contains a dative element (including pronouns that were not parsed for case in the corpora but must be construed as dative). The examples were inspected for a possessive relationship between the dative and the subject or object. In this way, all DEPs were found, regardless of the nature of the verb. Searching appropriately for IPs of body words in order to compare their frequency with DEPs is more complicated than searching for DEPs, mainly because the IP was the default possessive construction in OE, while the DEP had a much more limited range. An initial investigation that collected examples of direct argument body parts showed that whenever a text contained a significant number of examples of body parts playing the role of direct arguments, the number of IPs was very large compared with DEPs. For example, there are seventy-five examples of IPs with body objects in my EWS texts, compared with only thirteen examples of DEPs with such objects. These raw numbers are not very illuminating, since, as discussed in Chapter 1, the use of EPs in languages which have them has always been found to be subject to limitations not imposed on IPs. If we want to assess the extent of the competition between IPs and DEPs, we can expect to learn the most from a comparison of examples of the two types in contexts where the more limited construction (the DEP) is found. One of the cross-linguistic findings concerning external possessors is that they are associated somehow with ‘affectedness’, although the exact nature of this affectedness can be hard to pin down and varies across languages that have these constructions. It is a reasonable hypothesis that DEPs would only be found when an effect on the possessor was being conveyed. The searches for all DEPs of direct arguments, which imposed no restrictions other than on the grammatical relation of the possessum, confirmed that DEPs were nearly completely restricted to verbs than could be expected to report an adverse effect, such as stingan ‘to stab’. Vennemann (2002) and others have assumed that a highly affected IP would at best be unusual in OE and that the DEP would be the norm for expressing such possessors. As Vennemann is careful to note, this is an impression rather than an established fact. An important goal of the investigation was to test this assumption. To do so, I drew up lists of ‘affecting verbs’ in order to compare the use of direct argument IPs and DEPs with such verbs. While the list of transitive verbs used in searches for objects, presented in

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.3 

53

Appendix B as Table B.5, was fairly straightforward, compiling the list of verbs to use in searches for subjects was more complicated. This list, presented in Table B.4, was compiled with a view to including only verb forms whose subjects were a priori likely candidates for a DEP. First, the counts of subject possessa are limited to intransitive constructions, including passives. Cross-linguistically, the subjects of transitive verbs are not likely to appear in external possessor constructions, and the searches for all DEPs in OE and EME, as expected, found no examples in which the possessum was the subject of a transitive verb. In the few examples in which a body part is the subject of a transitive verb, an IP is used: (3.1) swa þæt þæs mædenes fex befeng hi eall so that the:  maiden: hair: surrounded her all abutan sona about immediately ‘So that the maiden’s hair immediately completely enveloped her’ (coaelive,+ALS[Agnes]:144.1809) Counting examples of transitive subjects in the IP and DEP comparisons would skew the results towards a higher frequency of IPs by including a situation that was an ‘absolute knock-out’ for DEPs. In order to test the position that IPs were in fact common in situations where a DEP would be an alternative, it was necessary to avoid inflating the number of IPs by not counting, as far as possible, ones where only an IP could be used. The intransitive verb list presented in Table B.4 therefore does not include any verb which, as far as I could determine, was only transitive in OE. Excluding transitive verbs is not enough, since ‘subject of an intransitive verb’ is not a unified class semantically. The subject in a sentence like John ran, where John is in control of his actions, is different from the subject in John fell, where falling can be seen as something that happened to John, making John more like an object. Many linguistic treatments treat these differences as syntactic as well, treating ‘unaccusative’ subjects as originating in the same position as objects but ‘unergative’, more agentive subjects, originating in a subject position. Given that external possessors typically present an action as something that affects the whole person, not just the body part, it is no surprise that Haspelmath’s Syntactic Relations Hierarchy has direct objects at the top of the list for possessa that are most freely found with DEPs, followed by unaccusative subjects, then unergative one, with transitive subjects at the bottom of the list (1999: 113). It is also no surprise that the searches for DEPs, which had no specification for verb type, found no examples of DEPs

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

54

   

with unergative subjects. Accordingly, only unaccusative intransitive verbs appear on the list used to look for IPs of subjects. The searches for DEPs generally and the searches for affecting verbs answered the basic questions posed above and also shed light on questions regarding the range of DEPs in OE, such as whether they were limited to situations in which the possessor of a body part was normally presented as being affected by an action to that body part when a DEP was used. In addition, it is interesting to consider whether DEPs were associated with specific verbs rather than simply with affected possessors. A second, more limited list of verbs, namely verbs found appearing as the main verb in examples with DEPs, was used to investigate this question. These verbs are incorporated in Tables B.4 and B.5 as separate columns for the ‘affecting’ verbs, as explained in Appendix B. The copulas beon and wesan are not coded as verbs in the YCOE but as forms of BE, and they must be treated differently from other intransitive verbs. It is appropriate that the searches for DEPs should encompass these forms in order to capture the range of use of DEPs, which were sometimes used with copulas, as in (3.2): (3.2) & him bið micge and him: is urine: ‘and his urine is yellow’

geolu. yellow (colaece,Lch_II_[1]:42.1.2.1433)

Since the searches for DEPs did not specify that the clause must contain a verb, examples like (3.2) were picked up by the searches for DEPs of subject possessa. DEPs involving copulas are discussed in section 4.4, separately from the discussion of subject possessa with lexical verbs. No comparison of the frequency of DEPs and IPs with predicative adjectives or nouns is made in this study. Such a comparison would not be very illuminating; IPs are very clearly the dominant possessor type used with the subjects of predicative nouns and adjectives, and adding the BE forms to the list of external possessor verbs would complicate the data gathering considerably without much benefit. However, section 4.4 makes some observations about the use of DEPs with the copulas. The BE verbs are not included on the list of intransitive verbs, but the searches picked up examples with participles of ‘affecting’ transitive verbs used in passive constructions and also participles of intransitive verbs. These examples are included in the relevant statistics. To summarize, the searches for subject and object body parts in sentences which also contained a dative collected all examples of DEPs with direct arguments. The searches for IPs and DEPs with ‘affecting’ verbs gave a picture

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.3 

55

of the variation with such verbs, and the searches for IPs using the ‘external possessor’ verbs gave a way of examining the use of IPs with verbs that also appeared with DEPs.

3.3.2 Prepositional object possessa The fact that the majority of the examples of DEPs presented in the literature involve PObjs reflects the fact that DEPs with these possessa are much easier to find in OE than examples of DEPs with direct arguments. Comparing the use of DEPs and IPs with PObjs is fraught with difficulty, however. As discussed in Chapter 1, the analysis of external possessors with PObj possessa presents challenges even in Modern English, and Vergnaud and Zubizaretta (1992) concluded that the external possessors of PObjs had to be analysed differently from those with bare nominal phrases. One fundamental distinction relevant in any syntactic framework to investigating the history of DEPs in English is illustrated in (3.3): (3.3) a. She kissed John on the cheek b. She looked John in the eye As discussed in Chapter 1, there is an entailment relation in (3.3a); if she kissed John on the cheek, she kissed John. The sentence would be grammatical if locative phrase were omitted. In the (b) sentence, however, the locative phrase cannot be omitted. Furthermore, uses of an ‘athematic’ possessor, i.e. one not subcategorized for by the verb, as in (3.3b), are restricted in Modern English to a very small number of fixed phrases. Since one goal of this investigation is to trace how DEPs declined in English, the data presented in Chapter 4 exclude examples like (3.4a) where the DEP is thematic. Only examples like (3.4b), in which the dative is not part of the verb’s usual argument frame, are included: (3.4) a. oþ hiene an cwene sceat þurh þæt þeoh until him: a woman shot through the thigh ‘until a woman shot him through the thigh’ (coorosiu,Or_3:7.64.28.1264) b. ðonne hie him on ðæt nebb spætton. when they him: in the face spat ‘when they spat in his face’ (cocura,CP:36.261.7.1700) A possible problem here is that the valence of a verb in earlier English may be different from a Modern English equivalent. A truly thorough examination of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

56

   

dative external possessors in OE that separated thematic from non-thematic datives would require a thorough study of the valency of individual verbs and the entailment relations of pairs of sentences. Such a study faces difficulties and limitations that are too obvious to require discussion. For OE, however, matters are made easier by the distinction between the dative and accusative cases; we can count only the examples that are DEPS, i.e. the external possessor is in the dative case. Since some verbs could govern an object in the dative case (sometimes in variation with an accusative object), we may face some uncertainty about what to do about examples in which we may not be certain whether a dative phrase is part of the ordinary argument frame of a verb or there only to serve as the possessor, as discussed more thoroughly in Appendix C, but we can still get a reasonable picture of the range and frequency of DEPs with PObjs in OE. Apart from the general exclusion of examples in which the possessor of a PObj is thematic, some of the restrictions on the investigation of PObjs arise from the general principle that a meaningful comparison of the frequency of use of internal and external possessors in OE should exclude IPs in situations where a DEP would not be a possibility, which meant not including prepositions with meanings, such as instrumental, that were apparently incompatible with DEPs. For the lists of prepositions and body words used in the basic PObj searches, see Appendix B, and for further discussion of decisions made about the culling of the examples returned by the searches, see Appendix C.

3.4 Summary The corpus used for this research has been chosen to exclude very late texts that might introduce extra ‘noise’ into the data caused by possible scribal changes made in copying, and the division into text types is aimed at ensuring as clear a picture as possible of usage in specific dialects and periods within the OE period. The basic goals of the data gathering were to shed light on the range of the DEP with a large number of words for the body and body parts in the different text types and to compare the frequency of DEPs and IPs with these words, distinguishing three different grammatical roles of the possessum: subject, (direct) object, and object of preposition. In studying the range of the DEP with direct arguments, all examples containing forms of these body words appearing as a subject or object possessum and a DEP as possessor were collected, subject to the restrictions discussed above and in Appendix C. Two types of searches were used in

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

3.4 

57

comparing DEPs and IPs with these direct arguments. The first type looked for examples using a list of ‘affecting’ verbs, to compare the frequency of DEPs and IPs in the situations where a DEP could be expected to be most likely. The second compared the use of DEPs and IPs with a more restricted list of verbs, i.e. the verbs found in any example with DEPs of direct arguments. Besides being the first corpus-based investigation of DEPs in Old and Early Middle English, this investigation differed from earlier ones in distinguishing the grammatical relations of the possessa, and especially in distinguishing direct argument possessa from PObj possessa. Because of the considerations discussed above, the searches for all DEPs with PObj possessa were more restricted than those with the direct argument possessa, but gave a good basis for restricting the prepositions and body part words for subsequent searches. These further searches collected all examples of DEPs with these selected prepositions and body words in my entire corpus and compared the use of DEPs and IPs with these words in a more restricted corpus. This data collection deliberating skewed the results towards maximizing the number of DEPs in searching for contexts that could be expected to favour that construction. The only exception is that some examples arguably containing DEPs were excluded because of uncertainty about how to analyse them, in particular examples with an independent dative, governed by the verb or some other element that was coreferential with the possessor, which was never expressed by a second dative, although it could have an overt IP. The reason for maximizing the number of DEPs is that there is a widespread assumption that the DEP was the usual way of expressing inalienable possession in OE, at least when the possessor could be seen as affected. A challenge to this assumption will only be credible if it is based on data that reduces to a minimum the number of IPs where no DEP would be expected. The results discussed in Chapter 4 will demonstrate that there is no basis for this assumption, or for the position that the decline of DEPs started in the Middle English period.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4 Body and dative external possessors in Old English 4.1 Introduction This chapter presents and discusses the results from the searches described in Chapter 3. The findings for direct arguments start with objects (section 4.2) and subjects (section 4.3) of verbal predicates. In the presentation of the data in these two sections, the data for the individual texts is first presented in three tables to cover three groupings of text types. The first group can be thought of as ‘early’, comprising poetry plus texts of EWS composition found in manuscripts belonging either to the EWS period or later in the OE period. The second group is the large ‘General OE’ group. The third group contains texts of Late West Saxon composition. For details and discussion of these groupings, see Appendix A. Within each of these sections, I first give some illustrative examples and then present data on the raw numbers of DEPs of body possessa playing the role of object or subject, respectively. The presentation of the raw data for DEPs is followed by an examination of the comparative frequency of IPs and DEPs with the verbs on the ‘affecting’ list. While DEPs were pretty much restricted to adversely affected possessors within sentences with verbal predicates, this is not the case with copular sentences, and a separate section 4.4 is devoted to these. The results for subject and object body possessa are compared in section 4.5. Section 4.6 looks at the possibility of Latin influence affecting the results. My conclusions concerning the direct arguments are summarized in section 4.7. While the bulk of this chapter is devoted to the direct arguments, findings for the objects of prepositions are presented and discussed in section 4.8, followed by my conclusions about DEPs and IPs in OE in section 4.9.

Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

59

4.2 Results: object possessa 4.2.1 DEPs of object possessa Before discussing the results, I present some examples to illustrate object possessa with DEP possessors. Examples (4.1) through (4.3) illustrate the types covered by the the column headings of Tables 4.1 through 4.3. The leftmost column in these tables is for uncontroversial dative external possessors. By ‘uncontroversial’, I mean that the dative is not licensed by the verb as part of its normal argument frame: (4.1) þa heton þa consulas Hasterbale þæt heafod then ordered the consuls Hasterbal: the: head: ofaceorfan off.cut ‘Then the consuls ordered Hasterbal’s head to be cut off ’ (coorosiu,Or_4:10.105.34.2190)

As discussed in section 2.7, OE sometimes combined an internal possessor with a dative one in a construction that Ahlgren (1946) referred to as ‘blended’. The ‘Blended’ column in the tables refers to this construction, illustrated in (4.2): (4.2) Gif monnes eage him if man: eye: him: ‘If a man’s eye is struck out’

mon ofaslea one: strikes (colawaf,LawAf_1:71.203)

Finally, there are a few examples designated as ‘ambiguous’. Case distinctions were not always unambiguously marked in OE, even with pronouns. The ambiguity of relevance here is the ambiguity between the dative and genitive cases. This distinction was not marked with feminine nouns, for example, either on the noun itself or on a determiner or adjective modifying it. The singular feminine pronoun was also ambiguously dative or genitive. This means that in some examples we cannot be certain whether we are dealing with an internal or external possessor, as in (4.3): (4.3) and het bindan fet and honda and ordered bind feet:. and hands:. þære halgan fæmnan the:/ holy:/ woman:/ ‘And ordered the holy woman’s hands and feet to be bound’ (comargaT,LS_16_[MargaretCot.Tib._A.iii]:18.1.229)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

60

       

Here, þære halgan fæmnan is ambiguous between the dative and the genitive case. The YCOE parses it as genitive in this example, but a possible problem is that with a postnominal genitive we would expect a determiner preceding the head noun fet and honda. If we treat the phrase as dative, the fact that it comes after the body part instead of earlier in the sentence is a bit unusual, but there are some clear examples of DEPs appearing after the possessum: (4.4) a. Þa dyde Martinus on muð þam wodan Then put Martin in mouth the: mad: his agene fingras, his own fingers ‘Then Martin put his own fingers into the madman’s mouth’ (coaelive,+ALS_[Martin]:540.6315)

b. and gefeterode fet and honda and bound feet: and hands: sinum his: ‘and bound the feet and hands of his child’

bearnum child:

Genesis A 2903–4

It is worth noting, however, that in all the prose examples collected of unambiguous DEPs appearing late in the sentence, the possessum is the object of a preposition, as in (4.4a), which comes from a Late West Saxon text. Postposed DEPs of object possessa seem to be limited to poetic texts, as in (4.4b). It is likely, therefore, that the YCOE’s parsing of þære halgan fæmnan in (4.3) as genitive is correct, especially as it comes from a very late text, but I have counted it as ambiguous. To fill out the presentation of illustrative examples of object possessa, (4.5) is interesting because it illustrates what Vergnaud and Zubizaretta (1992: 598) call the ‘distributivity’ effect: (4.5) & þæm mæden cildum hie fortendun and the:. maiden chidren:. they burned.off þæt swiðre breost the:. right: . breast:. ‘and they burned off the right breasts of the female children’ (coorosiu,Or_1:10.29.32.581)

The ‘distributivity’ effect is that although we have more than one possessor, the possessum is presented in the singular; this is also found in French, as in example (1.4.) Vergnaud and Zubizaretta (1992: 619) argue that the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

61

distributive interpretation arises as a consequence of the binding relation between the dative argument of the verb and the variable in the direct object. Only a few examples illustrating this effect were turned up by this investigation, as were a couple of examples in which the effect does not seem to hold, i.e. a plural is used for plural possessors. König and Haspelmath (1998: 582) note that while a singular form is most frequently used in external and implicit possessor constructions in European languages, plurals are possible in others. This is a matter that I will not pursue further. Tables 4.1 through 4.3 present the findings of this investigation for object possessa for the three broad categories: early, general, and Late West Saxon.

Table 4.1 DEPs of body object possessa in early OE texts DEP

Blended

Ambig

Poetry York Non-York Total Poetry

5 3 8

0 0 0

0 0 0

Other Early Bede(early) codocu1 codocu2 codocu2 mart 3 (Sweet) mart 4 (Sisam) Other Early Total

1 0 0 0 1 0 2

1 0 0 0 0 0 1

0 0 0 0 0 0 0

EWS cochronA cocura cocuraC coorosiu coprefcura EWS Total

DEP 1 2 0 9 0 12

Blended 1 0 0 0 0 1

Ambig 0 0 0 0 0 0

9thC(OE) cogregdC cogregdH colaece colawaf colawafint Total 9thC(OE) Total Early

DEP 3 0 3 8 1 15 37

Blended 1 0 1 1 1 4 6

Ambig 0 0 0 0 0 0 0

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

62

        Table 4.2 DEPs of body object possessa in General OE texts

General OE coalex bede(xi) coapollo coblick cobyrhtf cochristoph codocu3 codocu3 codocu4 coeuphr coeust coherbar colacnu colaw1cn colaw2cn comargaT comart1 comart3 comarvel comary conicodD coquadru cosevensl cosolsat2 coverhom coverhomE coverhomL ASC(post-EWS) General OE Total

DEP

Blended

Ambiguous

1 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 5 0 0 0 1 0 0 1 0 0 0 11

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1

Table 4.3 DEPs of body object possessa in LWS texts LWS AEhomsA coaelive_mod cobenrul cocathom1 cocathom2 colaw6atr coprefcath1 coprefcath2 copreflives cotempo cowsgosp LWS Total

DEP

Blended

Ambiguous

0 2 0 0 1 0 0 0 0 0 1 4

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 (continued)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

63

Table 4.3 Continued LWS(Late) Aelfric_vincent AEhomsB cocanedgD cocanedgX cogenesiC coepigen coexodusP coinspolD coinspolX colsigewZ colwgeat colwsigeT colwsigeXa colwstan1 colwstan2 cootest coprefgen cowulf Total LWS(Late)

DEP 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 2 3

Blended 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Ambiguous 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

4.2.2 DEPs vs IPs of object possessa It can be seen immediately from Tables 4.1 through 4.3 that the number of DEPs with body object possessa is never very large. The small number of DEPs of Late West Saxon origin is particularly striking given the large size of this corpus. The raw number of DEPs is not terribly meaningful, however, until we compare it with the number of IPs, in particular, the IPs where a DEP might be expected as a possibility. The next set of tables compares IPs and DEPs found in searches that were limited to the affecting verb list of Appendix B. Comparing the first set of tables with the second, one striking fact is that when we compare the numbers in the DEP columns of the two groups of tables, we see that they are nearly identical—DEPs of object possessa were nearly completely restricted to sentences in which the main verb was on the list of verbs that had been put on the ‘affecting’ list because of their typical semantics, independently of whether they ever appeared with DEPs. This shows that some affectedness condition held for DEPs of objects in OE. As indicated in Table 4.7, only three examples belonging in the first set of tables do not go into the second. For each of these examples, an argument can be made that despite not fitting the criteria to fit into the ‘affecting’ column, the dative is used to indicate some

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

64

       

Table 4.4 IPs vs DEPs of body object possessa with affecting verbs in early OE texts Poetry

IP

DEP

Blended

Ambiguous

York Non-York Poetry Total

8 3 11

4 3 7

0 0 0

0 0 0

Other Early bede (early) codocu1 codocu2 codocu2 mart 3 (Sweet) mart 4 (Sisam) Other Early Total

0 0 0 0 0 0 0

1 0 0 0 1 0 2

1 0 0 0 0 0 1

0 0 0 0 0 0 0

EWS cochronA cocura cocuraC coorosiu coprefcura EWS Total

IP 1 5 0 3 0 9

DEP 1 2 0 9 0 12

Blended 1 0 0 0 0 1

Ambiguous 0 0 0 0 0 0

9thC(OE) cogregdC cogregdH colaece colawaf colawafint 9thC(OE) total Early Prose Total

IP 14 4 2 1 1 22 31

DEP 3 0 1 8 1 13 27

Blended 1 0 0 1 1 3 5

Ambiguous 0 0 0 0 0 0 0

sort of effect. The first example is from the York Poetry corpus. The verb niman in the example is not on the ‘affecting’ list for the searches because its central meaning of ‘take ’did not ordinarily imply affectedness, and a further reason why the example is not counted in Table 4.4 because the possessor is dead: (4.6) þar him hrefn nimeþ heafodsyne, where him: raven: takes eyes: ‘where a raven takes his eyes’ (coexeter,155.36.381) (The fortunes of men 36)

As discussed in Appendix C, the counts of IPs versus DEPs with affected possessors of direct arguments did not count dead possessors because these possessors, being incapable of being affected, are cross-linguistically unlikely to occur with DEPs. However, this line is part of a highly emotive description

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

65

Table 4.5 IPs vs DEPs of body object possessa with affecting verbs in General OE texts General OE

IP

DEP

Blended

Ambiguous

coalex bede(xi) coapollo coblick cobyrhtf cochristoph codocu3 codocu3 codocu4 coeuphr coeust coherbar colacnu colaw1cn colaw2cn comargaT comart1 comart3 comarvel comary conicodD coquadru cosevensl cosolsat2 coverhom coverhomE coverhomL ASC(post-EWS) Total General OE

0 1 2 5 0 1 0 0 0 1 2 1 0 0 2 6 0 6 1 4 0 0 2 0 6 3 1 1 45

1 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 5 0 0 0 1 0 0 1 0 0 0 11

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1

of the grief of a father whose son has been hanged. Because the son’s death was a punishment for crimes, the father cannot avenge him, and he has to watch the ravens despoiling his son’s body on the gallows. While the possessor of the body is not capable of feeling anything, the father surely feels agony on seeing his son’s body despoiled. Another possibility is that niman should be treated as a ditransitive verb, in which case the possessor would be implicit. There was no requirement that the dative object of a ditransitive verb be alive. With a ditransitive treatment, this example would not be included in the count of DEPs, since the dative would be part of the case frame of the verb. I have included the example as a DEP because I have not been able to find any examples of this verb used in a clearly ditransitive structure. However, the verb beniman ‘to deprive’ was used as a ditransitive verb with a dative of the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

66

       

Table 4.6 IPs vs DEPs of body object possessa with affecting verbs in LWS texts LWS

IP

DEP

Blended

Ambiguous

AEhomsA coaelive_mod cobenrul cocathom1 cocathom2 colaw6atr coprefcath1 coprefcath2 copreflives cotempo cowsgosp LWS Total

8 24 0 9 19 0 1 0 0 0 19 80

0 2 0 0 1 0 0 0 0 0 1 4

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

LWS(Late) Aelfric_vincent AEhomsB cocanedgD cocanedgX cogenesiC coepigen coexodusP coinspolD coinspolX colsigewZ colwgeat colwsigeT colwsigeXa colwstan1 colwstan2 cootest coprefgen cowulf Total LWS(Late)

IP 0 8 0 0 1 0 1 1 0 2 0 0 0 3 0 11 0 2 29

DEP 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 2 3

Blended 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Ambiguous 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 4.7 Number of DEPs of objects of verbs not fitting the affecting verb criteria Text coexeter colaece

DEP, Verb not on Affecting list 1 2

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

67

deprivee and an accusative of the thing taken away, and so it is not unlikely that this use was extended to the prefix-less niman here. The example is therefore not a very convincing example of a DEP with a dead possessor. The two prose examples come from Bald’s Leechbook: (4.7) a. ðonne sceal mon þam men mid drium then shall one the: person: with dry handum . . . þa handa & þa fet gnidan hands the: hands: and the: feet: rub swiðe & þyn smartly and squeeze ‘then one should briskly rub and squeeze the person’s hands and feet’ (colaece,Lch_II_[2]:3.1.7.2179) b. teoh him þa loccas pull him: the: locks: ‘pull his hair’ (colaece,Lch_II_[2]:16.2.6.2331)

Neither of the verbs teon or gnidan are on the affecting verbs list because an adverse effect is not the usual outcome of either pulling or rubbing. In these two examples, however, it is arguable that the composer of the sentence would have been considering that the action would have a painful effect on the patient—even though this pain might have been supposed to be beneficial to the patient in the long run. This is especially easy to argue for (4.7b), since pulling someone’s hair is bound to be painful. As for (4.7a), rubbing someone’s feet might be positively pleasurable, but the instruction to rub them ‘smartly’ suggests that the rubbing might cause pain. There is no way to be certain what was in the writer’s mind, however, and of course any conclusion that an adverse effect must be assumed because a DEP always signalled an adverse effect would be circular. It is interesting to note how few examples of expressed possessors, either DEPs or IPs, were in the Leechbook, as indicated in Table 4.1 and Table 4.4.¹ This might seem surprising, given that the subject matter of this text naturally

¹ Another example of a DEP with a body object possessum is to be found in the table of contents of the Leechbook, which is not included in the YCOE text: (i)

& gif him mon lim ofceorfan scyle and if him: one limb: off.cut shall ‘and if one should amputate his leg’ Bald’s Leechbook, ed. Cockayne (1961) table of contents item xxxv, p. 8, line 15

This example is not included in the statistics presented in this chapter.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

68

       

results in a very large number of body parts playing the role of object. The instructions concerning what is to be done usually do not mention the patient: (4.8) binde þæt heafod bind: the: head: ‘bind the head with it’

mid. with (colaece,Lch_II_[3]:1.1.1.3497)

When we discuss subjects in section 4.3, we will see that the patient is more frequently mentioned with subject possessa, which are typically found in examples describing symptoms. Tables 4.4 through 4.6 presents figures for individual texts. Table 4.8 summarizes these figures for the text types. Table 4.8 Overview of IPs vs DEPs of body object possessa with affecting verbs IP

DEP

Blended

Ambiguous

Poetry Total

11

7

0

0

Other Early EWS 9thC(OE) General OE LWS LWS(Late) Prose total

0 9 22 45 80 29 185

2 12 13 11 4 3 45

1 1 3 1 0 0 6

0 0 0 1 0 0 1

It can be seen from Table 4.8 that, setting aside the Other Early texts, which provide only two relevant examples, the only text type in which the number of DEPs with objects of affecting verbs exceeds the number of IPs is the EWS type, with twelve DEPs and nine IPs. A look at Table 4.4 shows that the EWS texts are by no means uniform in this respect; the high figure for DEPs in EWS is due to the frequent use of this construction in the Orosius. This text is rich in descriptions of beheadings and other drastic acts, and so it is no surprise to find that it is a rich source of DEPs. It is more of a surprise to find that even in the Orosius we find some IPs with verbs of drastic effect: (4.9) a. hie het gebindan, & . . . mid æxsum heora them: ordered bind and with axes their heafda: of aceorfan heads off cut ‘(He) ordered them to be bound, and . . . their heads to be cut off with axes’ (coorosiu,Or_2:3.40.18.766)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.2 :  

69

b. Seo cwen het þa ðæm cyninge þæt the queen ordered then the: king: the: heafod of aceorfan head: off cut ‘the queen then ordered the king’s head cut off ’ (coorosiu,Or_2:4.45.6.852) In the interests of making these examples completely clear, I will mention that the infinitive following the verb hatan ‘to order’ was used in OE in a ‘passive’ sense. That is, the recipient of the command was unexpressed, making a passive construction the best translation into Modern English. The accusative NPs heora heafda and þæt heafod get their accusative case as objects of the subordinate clause verbs. The variation in the Orosius illustrated in (4.9) is particularly important because the same scribe penned both examples in the so-called Lauderdale manuscript. This suggests variation in the language of the individual. Although the scribe was copying from an exemplar and we cannot be certain that the exemplar was not itself written by more than one scribe, the Lauderdale manuscript could not have been written sufficiently later than the time of the exemplar to appeal to layers of copying as resulting in usages from different periods. Whether the variation is to be ascribed to variation at the level of the individual or the level of the community, it is clear that the widespread view that IPs were unusual in OE is simply incorrect. We find IPs used with highly affected possessors in all the early text types, including poetry, which can plausibly be seen to reflect early syntax: (4.10)

a. þe þæt wif feoð.. and þin heafod tredeð thee the woman hates and thy: head: treads ‘the woman will hate you, will tread on your head’ Genesis A 912 b. ond ic sumra fet forbræc bealosearwum and I some:. feet: destroyed snares: ‘and I destroyed the feet of some with snares’ (cocynew,126.468.1356) (Juliana 472–3)

The example from Genesis A and another instance of a highly affected IP in lines 2491–2 of that poem are of particular interest because of the traditional assumption that the religious poems of the Junius manuscript are of early composition. A final comment to be made in our discussion of variation is that we find some variation with a given verb, even within the same text. An example

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

70

       

comes from Cnut’s laws, which are of course of composition very near the end of the OE period, but included in my General OE category because of dialect mixture. In these laws, there are only a few examples of body parts, but we find both IPs and DEPs in the prescriptions for punishment by mutilation: (4.11)

a. & ceorfan of his nosu & his earan & And cut off his nose: and his ears: and þa uferan lippan the: upper: lip: ‘and cut off his nose and his ears and the upper lip (colaw2cn,LawIICn:30.5.109) b. þæt man ceorfe him ða handa oððe that one: cut him: the: hands: or þa fet the: feet: ‘that either his hands or feet be cut off ’ (colaw2cn,LawIICn:30.4.107)

The discussion of variation leads us to the next comparison. It is interesting to ask whether DEPs are associated with particular verbs. Table 4.9 presents the figures for IPs with the objects of all verbs that occur with a DEP at least once in my corpus. The figures for DEPs are of course the same as the totals for text types in Table 4.1 through 4.3. Note that the proportion of IP tokens goes down as we limit ourselves to the DEP verbs rather than all affecting verbs; for example, in the General OE type we have forty-five IPs with affecting verbs in Table 4.5 but only thirty-five IPs with verbs that ever occur with a DEP in my data. It seems that some verbs were associated with DEPs. Note also that the overwhelming preference for IPs with affecting verbs in the LWS texts is drastically reduced when we look only at the verbs that occur in the corpus Table 4.9 IPs vs DEPs of object possessa in OE text types, DEP verbs IP

DEP

Blended

Ambiguous

Poetry Total

0

8

0

0

Other Early EWS 9thC(OE) General OE LWS LWS(Late) Prose total

0 9 9 35 24 12 89

2 12 15 11 4 3 47

1 1 3 1 0 0 6

0 0 0 1 0 0 1

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.3 :     

71

overall at least once with a DEP; the figures are eighty and twenty-four, respectively. It seems that DEPs were not only associated with negative effects but with particular verbs that were frequently used to describe negative effects.

4.3 Results: subject possessa with lexical verbs As discussed in section 3.3.1, comparing DEPs and IPs with subject possessa is somewhat more complicated than with object possessa because of the difference between agentive and non-agentive subjects and also because of the fact that with subjects we must deal with copulas as well as ordinary verbs. In this discussion, I will begin with the subjects of lexical verbs (intransitive and passive) before discussing DEPs with copulas.

4.3.1 DEPs of (verbal) subject possessa As with objects, the discussion of subject body part possessa begins with a look at DEPs. The number of DEPs with such possessa are set out for poetry and individual prose texts in Tables 4.10 through 4.12. Table 4.10 DEPs of body subject possessa in early OE texts, lexical verbs Earlier

DEP

Blended

Ambiguous

Poetry York Non-York Total Poetry

4 3 7

0 0 0

2 0 2

Other Early bede (early) codocu1 codocu2 codocu2 mart 3 (Sweet) mart (Sisam) Other Early Total

1 0 0 0 0 1 2

0 0 0 0 0 0 0

0 0 0 0 0 0 0

DEP 0 1 0 3 0 4

Blended 0 0 0 0 0 0

Ambiguous 0 0 0 0 0 0

EWS cochronA cocura cocuraC coorosiu coprefcura EWS Total

(continued)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

72

       

Table 4.10 Continued 9thC(OE) cogregdC cogregdH colaece colawaf colawafint Total 9thC(OE) Early Prose Total

DEP

Blended

Ambiguous

3 0 14 1 0 18 24

0 1 1 0 0 2 2

0 0 0 0 0 0 0

Table 4.11 DEPs of body subject possessa in General OE texts, lexical verbs General OE coalex bede(xi) coapollo coblick cobyrhtf cochristoph codocu3 codocu3 codocu4 coeuphr coeust coherbar colacnu colaw1cn colaw2cn comargaT comart1 comart3 comarvel comary conicodD coquadru cosevensl cosolsat2 coverhom coverhomE coverhomL ASC(post-EWS) General OE Total

DEP

Blended

Ambig

0 1 1 1 0 0 0 0 0 0 0 3 0 0 0 0 0 8 0 0 0 1 2 0 1 0 2 0 20

0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.3 :     

73

Table 4.12 DEPs of body subject possessa in LWS texts, lexical verbs LWS AEhomsA coaelive_mod cobenrul cocathom1 cocathom2 colaw6atr coprefcath1 coprefcath2 copreflives cotempo cowsgosp LWS Total LWS(Late) Aelfric_vincent AEhomsB cocanedgD cocanedgX cogenesiC coepigen coexodusP coinspolD coinspolX colsigewZ colwgeat colwsigeT colwsigeXa colwstan1 colwstan2 cootest coprefgen cowulf Total LWS(Late)

Dat

Blended

Ambiguous

0 5 0 0 0 0 0 0 0 0 0 5

0 1 0 0 0 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0

Dat 3 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 4

Blended 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 2

Ambiguous 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Before comparing DEPs with IPs, some remarks about the nature of individual examples of DEPs with body possessa subjects are in order. As expected, the subjects of verbal predicates are typically not in control of the action; the verbs are either passive, as in (4.12a), or ‘unaccusative’ intransitive verbs, in which the possessor is an undergoer rather than an agent, as in (4.12b, c). The subjects of these DEP sentences are similar to objects in being nonvolitional. However, an interesting difference between subject and object possessa emerges concerning the ‘strict affectedness’ constraint. With object possessa, we found no examples of a DEP used when the action resulted in a

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

74

       

clearly beneficial effect, and only a couple of examples involving gnidan ‘rub’ in which the effect was not clearly adverse. The majority of examples of subjects with DEPs or blended DEPs and IPs are similar to the object examples in involving adverse effects, usually quite drastic ones: (4.12)

a. Gif men sie se earm mid Honda If man: be: the: arm: with hand mid ealle of acorfen with all off cut ‘If a man’s arm with the hand and all below is cut off ’ (colawaf,LawAf_1:66.193) b. & him se maga micla þindeþ and him: the: stomach: greatly descends ‘and his stomach is greatly distended’ (colaece,Lch_II_[2]:51.1.2.3104) c. forðam þe him burston ut butu his eagan because that him: burst out both: his eyes: ‘because both his eyes burst out’ (coaelive,ÆLS_[Alban]:116.4074)

However, I have found a small number of examples of DEPs of subject possessa in which either no particular effect or a positively beneficial effect is described. Such examples in which the verb is a participle can arguably be treated as adjectives, as discussed in section 4.3.4. This leaves three examples with a simple main verb: (4.13)

a. oððæt him feax geweoxe until him: hair: grew ‘until his hair grew’ (cobede,Bede_4:1.254.30.2593) b. Þonne leohtað him se lichoma then lightens him: the: body: ‘Then his body lightens’ (coherbar,Lch_I_[Herb]:1.16.70) c. ne mæg him se lichoma batian not may him: the: body: heal ‘his body cannot heal’ (colaece,Lch_II_[2]:21.1.20.2437)

In (4.13a), no particular effect on the possessor is indicated, and in (4.13b) we have a clear example of a positive effect. Things are less clear in (4.13c);

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.3 :     

75

the verb is positive, but the overall message is negative. Note that the (a) and (c) examples come from texts composed in the early OE period, and while the (b) example comes from a later manuscript, the text is likely to be a copy of an earlier composition, suggesting that examples with a possessor that is not clearly affected negatively might be a feature only found fairly early. All the DEPs of subject possessa in the LWS and LWS(Late) texts refer to severely affected possessors, as in (4.12c), which describes the fate of an executioner who was attempting to behead a Christian saint. It is finally worth noting one example that illustrates the use of a verb that is not on the ‘affecting’ list for subjects because of its usual semantics, but the possessor of the body part is definitely affected: (4.14)

hwæþre him sio swiðre swaðe weardade however him: the: right: track: guarded hand on Hiorte hand: in Heorot ‘however his right hand remained behind (lit. “guarded the track”) in Heorot’ (cobeowul,65.2096.1708) Beowulf 2098–9

The verb weardian ‘guard, protect’ is transitive, and in that meaning, it has an agentive subject and would not be used with a DEP. However, the phrase swaðe weardian, which literally means ‘guard (the) track’, was a fixed expression, apparently limited to poetry, that meant ‘remain behind’. While the sentence does not get counted as one involving an ‘affecting’ verb, the possessor of the hand that was left behind, Grendel, was certainly affected, since he died from having his arm ripped off. Leaving aside occasional examples like (4.14), which are counted as DEPs but not as involving an ‘affecting’ verb, we can summarize the situation with subjects of lexical verbs by saying that while the DEPs of subject possessa usually referred to an adversely affected possessor, there is some evidence for the use of DEPs with unaffected or positively affected possessors of body part subject possessa, in earlier OE at least. This appears to make subjects different from objects, where the possessor of the body object always appears to be adversely affected. In addition, we see in section 4.3.2 that DEPs were not infrequent in neutral or positive descriptions in examples with a copula and an adjective. All in all, the strict affectedness condition seems to have been weaker, at the least, with possessa playing the role of subject than it is with objects.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

76

       

4.3.2 DEPs vs IPs of (verbal) subject possessa Turning now to a comparison of IPs and DEPs with subject possessa, Tables 4.13 through 4.15 present the figures for poetry and individual prose texts, while Table 4.16 aggregates the figures for each text type. As with object possessa, the possessor of a body word playing the role of subject is more likely to be expressed as an IP than a DEP in a large majority of texts. However, DEPs are more common with subjects than they are with objects, and one text, Bald’s Leechbook, goes strongly against the usual pattern. The discrepancy between subjects and objects is the subject of section 4.5. In addition to the poetic examples recorded in Table 4.10, the queries turned up two more possible examples of DEPs that are not included in these statistics:

Table 4.13 IPs vs DEPs of body subject possessa with affecting verbs in early OE texts Poetry

IP

Dat

Blended

Ambiguous

York Non-York Poetry Total

2 4 6

2 3 5

0 0 0

1 0 1

Other Early cobede (early) codocu1 codocu2 codocu2 mart 3 (Sweet) mart 4 (Sisam) Other Early Total

1 0 0 0 0 0 1

0 0 0 0 0 1 1

0 0 0 0 0 0 0

0 0 0 0 0 0 0

EWS cochronA cocura cocuraC coorosiu coprefcura EWS Total

IP 0 4 0 0 0 4

DEP 0 1 0 3 0 4

Blended 0 0 0 0 0 0

Ambiguous 0 0 0 0 0 0

9thC(OE) cogregdC cogregdH colaece colawaf colawafint 9thC(OE) total Early Prose Total

IP 3 3 1 1 0 8 13

Dat 3 0 13 1 0 17 22

Blended 0 1 1 0 0 2 2

Ambiguous 0 0 0 0 0 0 0

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.3 :     

77

Table 4.14 IPs vs DEPs of body subject possessa with affecting verbs in General OE texts General OE

IP

DEP

Blended

Ambiguous

coalex bede(xi) coapollo coblick cobyrhtf cochristoph codocu3 codocu3 codocu4 coeuphr coeust coherbar colacnu colaw1cn colaw2cn comargaT comart1 comart3 comarvel comary conicodD coquadru cosevensl cosolsat2 coverhom coverhomE coverhomL ASC((post-EWS) Total General OE

0 0 0 5 0 0 0 0 0 1 0 7 1 0 0 2 0 2 0 1 1 0 2 0 0 1 0 1 24

0 1 1 1 0 0 0 0 0 0 0 2 0 0 0 0 0 8 0 0 0 0 2 0 1 0 2 0 18

0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

(4.15)

a. wæs him gylp forod, beot forborsten, was them: boast: broken vow: failed and forbiged þrym, wlite gewemmed. and humbled glory: beauty: marred ‘Their boast was broken, their vow come to nothing, their glory humbled, their beauty disfigured’ (cogenesi,5.67.65) b. Næs hyre wloh ne hrægl, ne not.was her:/ hem: nor robe: nor feax ne fel fyre gemæled, hair: nor skin: fire: marked ‘Neither her hem (of her garment), nor her robe, nor her hair, nor her skin was marked by the fire’ (cocynew,130.590.1437) (Juliana 590–1)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

78

       

Table 4.15 IPs vs DEPs of body subject possessa with affecting verbs in LWS texts LWS

IP

Dat

Blended

Ambiguous

AEhomsA coaelive_mod cobenrul cocathom1 cocathom2 colaw6atr coprefcath1 coprefcath2 copreflives cotempo cowsgosp LWS Total

2 10 0 7 16 0 0 0 0 0 4 39

0 5 0 0 0 0 0 0 0 0 0 5

0 1 0 0 0 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0

LWS(Late) Aelfric_vincent AEhomsB cocanedgD cocanedgX cogenesiC coepigen coexodusP coinspolD coinspolX colsigewZ colwgeat colwsigeT colwsigeXa colwstan1 colwstan2 cootest coprefgen cowulf Total LWS(Late)

IP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 2

Dat 3 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 4

Blended 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 2

Ambig 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 4.16 Summary of IP vs DEPs of body subject possessa with affecting verbs in OE text types

Poetry Total EWS Other Early 9thC(OE) General OE LWS LWS(Late) OE Prose total

IP

DEP

Blended

Ambiguous

6

5

0

1

4 1 8 24 39 2 78

4 1 17 18 5 4 49

0 0 2 2 1 2 7

0 0 0 0 0 0 0

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.3 :     

79

In both examples in (4.15), the body word is part of a paratactic series involving participles, very common in the poetry. The relevant words, wlite in the (a) example and feax and fel in (b), are not the first element of the series and so are not clearly in the scope of the dative form at the beginning of the series. Whether we count these two examples as DEPs or not, a more important point is that IPs are not uncommon even with highly affected possessors. Importantly, such IPs can be found even in poetic texts, as in (4.16): (4.16)

ac sio hand gebarn modiges mannes but the: hand: burned brave: man: ‘But the hand of the brave man burned (i.e. was burnt)’ Beowulf ll.2697–8

The presence of this IP in Beowulf cannot be taken to represent late syntax or to be due to Latin influence, a topic discussed in section 4.6. There is one prose example that we might add to the LWS figures for DEPs of subject possessa: (4.17)

ne sceal eow beon forloren. an hær. of not shall you: be lost one: hair: of eowrum heafde your head ‘not one hair of your head will be lost to you’ (cocathom1,+ACHom_I,_16:311.122.3024)

There is no problem with the fact that eow is formally ambiguous between dative and accusative case, since it could only be dative here; the problem is that it seems likely that its presence is due to forloren. While forleosan is not a ditransitive verb and so the dative is not a typical indirect object, it seems to be an instance of the freer use of sentence datives that was possible in OE, the addition of a dative of interest. Even if (4.17) is counted as a DEP, it is very clear that the IP was by far the more usual way of expressing possession with direct argument possessa for Ælfric, even with very adversely affected possessors. We can finally note that a comparison of Tables 4.16 and 4.17 clearly shows that the occurrence of IPs of subject possessa overall in OE is greater than it is with verbs that ever occur with a DEP.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

80

       

Table 4.17 IPs vs DEPs of body subject possessa with DEP verbs in OE text types IP

DEP

Blended

Ambiguous

Poetry Total

0

7

0

2

EWS Other Early 9thC(OE) General OE LWS LWS(Late) Prose total

0 0 4 15 14 2 35

4 2 18 20 5 4 53

0 0 2 2 1 2 7

0 0 0 0 0 0 0

In other words, some verbs tend to occur more often with DEPs than with IPs, even taking into account the general favouring of DEPs with negatively affected possessors. An example of a situation in which a dative is the rule even in late OE is with expressions of people’s bowels falling out, typically when they went to the toilet. As indicated in Table 4.17, the LWS and LWS (Late) texts combined yielded a total of twelve examples in which a dative refers to the possessor of a subject body word, including three blended examples. All these examples are in texts originally composed by Ælfric (but in the case of the LWS(Late) examples, copied considerably later). Of these twelve examples, five are accounts of this horrific event.² In three of these five examples, a dative is used as the sole expression of the possessor (i.e. a DEP), as in (4.18a), and two examples have the ‘blended’ construction exemplified by (4.18b): (4.18)

a. ac him eode se innoð ut æt his forðgange. but him: went the: innards: out at his anus ‘but all his bowels fell out at his anus’ (coaelive,ÆLS_[Memory_of_Saints]:206.3445) b. swa þæt him aeode ut eall his innoð togædere so that him: went out all his innards: together ‘so that his bowels all fell out together’ (coaelhom,+AHom_10:159.1490)

² Ælfric’s mentions of this are usually in reports of God’s punishment to a persecutor of Christians, although at least one example involves the slitting open of a saint’s belly.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.4     

81

Ælfric, who used IPs as well as DEPs in discussions of drastic events like beheadings, did not leave us with any examples of IPs in accounts of people losing their bowels, nor, as far as I could find, did anyone else. With only five examples, we cannot be too dogmatic, but given Ælfric’s preference for IPs in general, the fact that he used a dative in all of these examples strongly suggests an association between particular negative events and the use of a dative possessor of the sort that is best encoded in the lexicon.

4.4 Subject possessa with adjectival predicates This section makes some observations on sentences in which the predicate with a body part subject is not a lexical verb but a copula meaning ‘be’ or ‘become’ combined with an adjectival predicate. No attempt will be made to compare the incidence of IPs and DEPs with such predicates. However, a few observations on the nature of the sentences in which a DEP is used are in order, because they seem to be of interest from a cross-linguistic point of view. In the previous section, we saw that when a possessor of a body part subject of a lexical verb was expressed as a DEP, that possessor was nearly always adversely affected by the action of the verb, although there are a very few exceptions in the texts. Matters are different in sentences with an adjectival predicate. The majority of examples do in fact report a situation likely to be adverse for the possessor, as in (4.19): (4.19)

& him bið micge and him: is urine: ‘and his urine is yellow’

geolu yellow (colaece,Lch_II_[1]:42.1.2.1433)

Here, the overly yellow urine is reported as a symptom of a disease. Such reports of symptoms using a DEP with the subject of an adjectival predicate are common in Bald’s Leechbook, although not as common as IPs in the description of symptoms: (4.20)

& his micgge bið blodread swilce hio blodig sie. and his urine: is blood-red as it bloody is ‘and his urine is blood-red as though it is bloody’ (colaece,Lch_II_[2]:17.1.13.2362)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

82

       

The use of DEPs in reporting states that are clearly not good for the possessor is not surprising, but it is surprising to find examples of pure description, where no effect on the possessor is obvious: (4.21)

a. wæs þæm deore all se hrycg was the: beast: all: the: back: acæglod swelce snoda studded as snood ‘The beast’s back was all studded with pegs like a snood’ (coalex,Alex:27.2.316) b. ðæm wæs seo onsyn sweartre þonne hrum whom: was the: face: blacker than soot ‘whose face was blacker than soot’ (comart3,Mart_5_[Kotzor]:Au25,A.9.1531) c. Is him þæt heafod hindan grene is him: the: head: behind green ‘his head is green behind’ (cophoeni,102.293.197)

Examples like those in (4.21) are of interest because Haspelmath (1999: 113) places stative verbs at the bottom of the hierarchy of types of situations where external possessor constructions are found. It is rather surprising to find DEPs in stative sentences in a language in which there is otherwise nearly a complete restriction on DEPs of subject possessa to reporting adverse effects. Payne and Barshi (1999: 13), however, note that Haspelmath’s hierarchy does not hold universally, and that stative verbs are actually the most likely to take an external possessor construction in some languages. Cross-linguistically, external possessor constructions take a number of forms. Payne and Barshi suggest that ‘the particular hierarchy for a given construction most certainly depends on the diachronic origin of the EPC [External Possessor Construction] in question’ (1999: 11). They note (p. 9) that the typical European DEP variant stems from an extension of locative/goal or dative schema described by Heine (1997) for predicative possession. While locatives and goals are both frequently expressed by dative case, as in OE, these schemas are of course different, and it seems possible that the DEPs with copulas derive historically from a locative schema, while those found with dynamic verbs come from a goal schema. Whatever the explanation for these DEPs of the subjects of predicative adjectives, it is worth noting that OE was not alone in the European languages

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.4     

83

in having such DEPs. Bolkestein (2001) in his analysis of dative possessors as experiencers in Latin, presents this example: (4.22)

rostra his et praelonga crura rubent beaks 3.. and long legs are.red ‘Their beaks and long legs are red’ (Plin. NH 10.129, as cited as Bolkestein 2001: ex. 22)

Bolkestein comments that he would have expected a genitive rather than a dative in this sentence but does not attempt an explanation. Whatever the source of DEPs in pure descriptions, they are also found with some frequency in Old Saxon, as discussed in section 5.2.2. The attestations of DEPs in pure descriptions is ‘lumpy’ in the OE texts. Most texts offer no examples, but there are three examples in the highly Latinate text translating a putative letter from Alexander the Great to Aristotle. (4.21c) is the only such example I have found in poetry. It is worth noting that this poem is a translation of a Latin poem, and it is possible that the use of datives in descriptions was part of a Latinate style, although this particular line does not translate any possessive construction in the corresponding Latin. Be that as it may, DEPs and blends in copular sentences are more likely to convey a negative situation than a pure description. I have found only two prose examples that I am counting as adjectival in which the state described is not only not negative but actually positive: (4.23)

a. Þa sona . . . geheortlice him wæron þa Then soon vigorously him: were the: limu cwiciende & fægre. limbs: quickening and fair ‘Soon then . . . his limbs were vigorously recovering feeling and fair’ (cogregdC,GDPref_and_4_[C]:37.317.15.4753) b. ac his lima beoð him ealle ansunde but his limbs: are him: all intact ‘But his limbs will all be sound’ (coaelhom,+AHom_11:320.1654)

Example (4.23a) describes a man’s revival from death after an angel has told him to go back to life and reconsider how he lived. The DOE lists the example presented here as in its lemma cwician under the sense ‘to become sensitive, recover feeling or warmth’. Cwiciende is the present participle of the verb cwician, and if it were not conjoined with an adjective, its treatment as a lexical verb would be straightforward. As it is, the YCOE parses cwiciende & fægre as

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

84

       

an adjective phrase. It is also from one of the more Latinate translations, and this particular sentence translates a Latin absolute construction, which may have promoted the use of the dative. Example (4.23b) is a ‘blended’ one. This investigation found no examples of DEPs or blends in copular sentences of pure description in the large LWS corpus, making this positive description all the more remarkable. It is attractive to analyse the dative of this solitary ‘blended’ Late West Saxon example as a benefactive complement to the adjective, ‘sound for him’, rather than an expression of the possessor. However these unusual examples are to be analysed, it seems that the ‘affectedness’ constraint on DEPs was more relaxed with subject possessa than with object possessa. It seems likely that this is due to the fact that datives were part of various copular constructions in which they were licensed by the copula, leading to their use occasionally as DEPs of subject possessa in constructions in which a form identical with a copula was used as an auxiliary verb. The final step here is a sporadic extension to using DEPs with the subjects of purely lexical verbs when no negative effect was involved.

4.5 Comparing subject and object possessa One fact that emerges from a comparison of the tables in sections 4.2 and 4.3.1 is that the frequency of DEPs compared with IPs with affecting verbs is higher for body part possessa in the subject role than in the object role. For convenience, I combine Tables 4.8 and 4.16 in one table, Table 4.18, which leaves out the ‘blended’ figures and adds the percentages (in parentheses when the total is less than ten).

Table 4.18 IPs vs DEPs of body object and subject possessa with affecting verbs, all OE text types Objects

IP

DEP

Total

%DEP

Subjects

IP

DEP

Total

%DEP

Poetry Total Other Early EWS 9thC(OE) General OE LWS LWS(Late) Prose total

11 0 9 22 45 80 29 185

7 2 12 13 11 4 3 45

18 2 21 35 56 84 32 230

39 (100) 57 37 20 5 9 20

Poetry Total Other Early EWS 9thC(OE) General OE LWS LWS(Late) Prose total

6 1 4 8 24 39 2 78

5 1 4 17 18 5 4 49

11 2 8 25 42 44 6 127

45 (50) (50) 68 43 11 (67) 39

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.5     

85

It is clear that the proportion of DEPs to IPs is higher in nearly all prose categories for subject possessa than for object possessa. The only exceptions are EWS, where the number of relevant subject possessa is too small to make the comparison very meaningful, and Other Early, where the low numbers of both object and subject possessa render any comparison meaningless. I have chosen one text type, General OE, to test for the statistical significance of this difference between subject and object possessa. This text type has a sufficiently large number of both IPs and DEPs to result in a large enough figure in each cell to make a chi-squared test reliable. This test shows that the difference between subject and object possessa is indeed statistically significant, with a chi-squared value 5.1434, giving a probability of only 0.02333 that the difference in the frequency of DEPs and IPs is due to chance.³ Small numbers of subject body part possessa make a chi-squared test unreliable for the other text types, but the overall picture seems to be that DEPs were more likely to be used with subject possessa than with object possessa. The main concern with comparing DEPs and IPs in this chapter is to establish that it is incorrect to think of DEPs as the usual way of expressing the possessor of a body part and also to present data that will be used in the discussion in Chapter 5 of the decline of the DEP already in the OE period. The choice of which expression of the possessor clearly involved many factors such as the verb used, and I will not attempt to explain the difference between DEPs with subject and object possessa. However, we can note that topicality is important in the use of DEPs. Citing Langacker’s (1993) statement that both topics and possessives function to name one entity as a reference point for talking about another entity, Payne and Barshi (1999: 8) comment that there is a functional motivation for expressing a possessor as a topic. We have seen that the DEP usually expressed affectedness, but also that affectedness did not always result in a DEP. It seems likely that one reason that combined with affectedness in favouring a DEP was to promote the possessor from a dependent of the body part to a clause-level constituent that reflected the topicality of the possessor better. The contrast between the very few examples of a possessor of an object body part in the Leechbook and the numerous examples of expressed possessors of subject body parts in that text is surely due to a difference in topicality. With objects, the focus is on the actions of the ‘leech’ or physician—what he is meant to do to the body part, rather than on the patient, so we get references to ‘the leg’ etc. as though the leg had no ³ For this and the other chi-squared tests in this book, I have used R (2017). This is a Pearson’s chi-squared test with Yates’ continuity correction.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

86

       

possessor, because the possessor doesn’t really matter in this context. With subjects, on the other hand, the focus is on the patient—in what way he or she is affected. We naturally get more expressed possessors, and they are highly topical. With intransitive verbs, including passive verbs, the possessor is the only human mentioned in the sentence, and so it is natural that this possessor should be elevated to the status of a core argument of the sentence. With an object, however, we are dealing with a transitive verb, and the subject of that verb is likely to be a topical human being. It is interesting to note that in more than half the examples in which the possessor of a body object is expressed as a DEP the subject of the sentence is definitely not topical, being either the indefinite pronoun man ‘one’ or otherwise having no specific referent: (4.24)

a. þæt him mon aslog þæt heafod of that him: one: struck the: head: off ‘that his head was cut off ’ (coorosiu,Or_6:34.152.22.3235) b. he him het þæt heafod of aceorfan he him: ordered the: head: off cut ‘he ordered his head to be cut off ’ (coorosiu,Or_5:12.127.29.2699)

Although the possessor is conceptually part of the object, he is the only human mentioned in the clause, and this, combined with the drastic effect of the action, is likely to favour his treatment as an element of the clause. I have noted that specific verbs are particularly likely to occur with DEPs. It is possible that one factor historically causing DEPs of objects to become associated with specific verbs is that those verbs were ones that tended to occur with unimportant subjects—the executioner is only occasionally mentioned in the report of an execution, for example (such as when his eyes fly out). This is speculative, and I leave explorations into the reasons why DEPs were more common with subjects than objects to further research. It is worth noting that topicality may be behind Havers’ (1901) observation that the ‘dativus sympatheticus’ was the dominant construction with pronouns in the Germanic languages generally, while the genitive was more likely to be found with nouns. In general, pronouns are more topical than nouns. Havers’ observation is somewhat impressionistic rather than based on a truly systematic collection of examples, and so I took a very limited look at the use of pronouns versus nouns in IPs and DEPs in my data. I have not attempted to count all the nouns and pronouns in the data, but limited the investigation into this question to one text type with a sufficiently large number of both IPs

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.5     

87

and DEPs to make a comparison meaningful, namely subject possessa in the General OE category. I present the results of this investigation in Table 4.19. Table 4.19 Pronouns and nouns in IPs and DEPs of body subject possessa with affecting verbs in General OE texts

Noun* Pronoun Total

IPs

DEPs

Total

% of Total

10 14 24

2 16 18

12 30 42

29 71 100

* In one of these examples, the dative is men: (ii) Gif men þæt heafod berste if person: the: head: burst: ‘If a person’s head should burst’ (coherbar,Lch_I_[Herb]:90.7.1437) I have treated men as a noun, but its status as a noun or pronoun is arguable. Van Bergen’s (2002) study of the properties of man focuses on this word in subject position, where she argues that man is a pronoun. Van Bergen notes that more than one scholar has stated that it is rarely found in pronominal use in any other function.

The strongest result indicated in Table 4.19 is that with both IPs and DEPs, the typical possessor is pronominal, accounting for a bit more than 71 per cent of all possessors. This is not really very surprising, because we are only looking at animate (mostly human) possessors, and human beings are likely to be topical. Only 53 per cent of pronouns in this text type are DEPS, so we cannot say with a great deal of confidence that if a possessor is a pronoun, there is a strong likelihood that a DEP will be used. However, there is some support for the idea that a nominal possessor is likely to be expressed as an IP rather than a DEP, even with highly affecting verbs, since this is true with 83 per cent of our nominal possessors, and this is a considerably larger percentage than the occurrence of IPs overall in the data, which is 57 per cent. The number of nominal possessors is not large, so this result is not a strong one, but it is suggestive, and it makes sense that nominal possessors, likely to be less topical than pronouns, might appear more often in the less marked construction (the IP), which reports the action to a body part without focusing as much on the effect on the possessor. It should be kept in mind that we are only looking at subjects of affecting verbs here, and only in one text type. Havers did not categorize his examples by the grammatical function of the possessa, and it is possible that an investigation of PObjs, where we find the most DEPs, would give stronger support for Havers’ observation.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

88

       

4.6 The question of Latin influence Since the majority of our examples from the prose texts are from texts translated from or at least based on Latin originals, we need to consider whether the use of DEPs and IPs in these texts might have been influenced by the Latin the scribes were translating. Ahlgren (1946: 211) expressed the belief that the use of the ‘possessive adjective’ (=IP) instead of the dativus sympatheticus ‘was most probably promoted [Ahlgren’s emphasis] by the influence of the language of the Church, which was largely modelled on Latin usages’. Ahlgren’s opinion received some cautious support more recently by Vennemann, who assumes that IPs were not very frequent in the texts, but considers that Latin influence might be one of the reasons for such examples of IPs as are to be found in OE (2002: 212). However, the examples that Ahlgren produces to show that the OE normally followed the Latin construction come entirely from biblical translations, especially from interlinear glosses. A more systematic comparison than this of the OE translations with the Latin originals is needed, keeping interlinear glosses separate from other translations. The following discussion mostly concerns freer translations rather than glosses. In looking at Latin influence, we need to realize that some translations from Latin stay more closely to the (assumed) Latin original than others. Interlinear glosses are particularly slavish translations that distort ordinary OE syntax in various ways, and although the glosses can be useful in exploring the limits of what a glossator was willing to do in imitation of Latin, I exclude them from the following discussion. Leaving such glosses aside, OE writers were especially likely to stay as close as possible to Latin syntax in translating the Vulgate Bible; see Taylor (2008) for an interesting case study of different translation effects in biblical versus other translations. However, they do not do real violence to normal OE grammar. They may stretch it a bit in places and when two variant expressions are possible for conveying the same situation, such as the DEP and the IP, and it is certainly reasonable to suggest that a translator might use the variant favoured in the Latin exemplar. Non-biblical translations may also stay very close to the Latin sources in both content and syntax, while some translations are very free indeed, such as the Orosius. In this translation, entire sections of the Latin text have been omitted and new observations inserted by the translator, along with substantial rewording of the sections where the essential story has been preserved.⁴ ⁴ For the Latin version of the Orosius, I have consulted Sweet (1885) and Migne (1857).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.6     

89

In such a translation we can expect that the translator will not be heavily influenced by the Latin syntax. In contrast, some translations stay close to its Latin source in both content and syntax, and these tend to use some constructions that are found only in translations, such as the ‘dative absolute’, which is modelled on the Latin ablative absolute. However, it is also possible for a translation to stay close to its source in content but not to show much influence in the syntax. This is the case with the Pastoral Care, a translation of Gregory the Great’s Cura Pastoralis (CP). Bately (1988: 126–7) stresses that despite the closeness of the translation in CP, it is not a slavish one.⁵ For example, she reports Potter’s (1931) finding that the dative absolute is used only once in this work, although there are 100 ablative absolutes in the Latin. This style of translating contrasts with the frequent use of dative absolutes in Gregory’s Dialogues and the translations of Bede’s Ecclesiastical History (1988: 133). We also need to be aware that a Latinate style often extends beyond simply direct translation of individual sentences; for example, in the more Latinate translations, a dative absolute may be used to translate a Latin sentence that does not have an ablative absolute as a model. The absolute construction seems to have been simply considered suitable for translations of Latin generally by some translators. Studies comparing the comparative frequency of constructions in individual texts can be very illuminating in detecting this more subtle Latin influence. Nevertheless, a study of translations of individual examples is also important in examining Latin influence. I have compared all my examples of both IPs and DEPs with direct argument possessa in Cura Pastoralis, Gregory’s Dialogues, and Orosius, and have made less systematic notes about translations from Latin for other texts. One finding of my comparisons is that Ahlgren’s assumption that language that was ‘largely modelled on Latin usages’ would not be language that eschewed DEPs. DEPs are not absent in the Latin originals. For example, in one of our earliest prose texts, we find (4.25):

⁵ The Pastoral Care is a good case to illustrate the fact that we are not always certain in our identification of the Latin text that a given translator was using. Clement (1985) argues that the recension of the Latin text that the translator (usually assumed to be King Alfred) used was not the one that forms the basis for the printed editions of the Latin text. That means that we cannot be certain that the Latin we are using as the basis for our comparison is what the translator actually had to work with. Nevertheless, it seems probable that any differences that might emerge in the syntax of a different Latin version would not materially change our conclusions regarding translation effects in the use of IPs and DEPs in the OE version.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

90

       

(4.25)

& him beoð þa sconcan forbrocenne ær and him: be: the: legs: fractured before þon he sy bebyrged that he be: buried ‘and let his legs be broken before he is buried’ Mart 4 (Sisam 1953: 220) Lat: nulla eum capiet sepultura nisi no: him: receive: grave: unless ei tibiae frangantur him: legs: are.broken ‘let no grave receive him before his legs are broken’

While it is not certain that the Latin presented in (4.25) is the source for the OE, the entry for for-brecan in the DOE produces this OE example with a cf. to Pass. Victor. 2.632.9, with the Latin of (4.25). Regardless of whether this is the exact source, it illustrates the use of DEPs in the Latin sources Anglo-Saxon writers of martyrologies used. Similarly, the DEP in (4.26) translates a Latin DEP: (4.26)

Ne forbinden ge na ðæm ðyrstendum oxum not bind ye not the: thirsting:t oxen: ðone muð the: mouth ‘Do not bind the mouth of the thirsty oxen’ (cocura,CP:16.105.7.690) Lat: non obturabis not bind:

os bovi trituranti mouth: ox: threshing:6

In these examples, the English follows the Latin, but it follows it in using a DEP. Other examples attest to the fact that divergences from the Latin were very common. An example of particular interest is presented in (4.27): (4.27)

& his eare of acearf and his ear: off cut ‘and cut his ear off ’ (cowsgosp,Mk_[WSCp]:14.47.3372) Lat: et and

amputavit illi auriculum cut.off him: ear:

The reason why this example is interesting is that it is apparently a case of translator substituting an IP for a DEP in the Latin. I say ‘apparently’, because ⁶ The fact that the translator has made a mistake in translating the Latin verb is irrelevant.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.6     

91

we cannot be certain of what version of the Vulgate the translator was using; the Latin of example (4.27) is that of the Nestle-Aland (1984) edition of the Vulgate New Testament. However, it seems unlikely that the dative was an oddity in versions of this verse, because the Lindisfarne Gospel has a dative here also, and the interlinear gloss, being a more slavish translation, follows the Latin in having a DEP, although it adds a determiner to the body part with ða earlipprica ‘the:. (external) ear:.’. Another rather surprising divergence from the Latin in biblical translations is this single instance of a DEP in the West Saxon gospels: (4.28)

& acerf him of þæt swyðre eare; and cut him: off the: right: ear: ‘and cut his right ear off ’ (cowsgosp,Jn_[WSCp]:18.10.7194) Lat: abscidit eius auriculam dextram cut his ear: right:

Assuming that the translator was working with this version of the Latin, he has substituted a DEP for a Latin IP. This departure from the Latin is notable in a biblical translation. As Tables 4.4 and 4.13 indicate, there is a striking difference within the Early West Saxon period in the relative frequency of IPs and DEPs in the Pastoral Care and Orosius, with both object and subject possessa. The greater frequency of DEPs in the Orosius compared with CP may be at least partially due to the fact that the translation of Orosius is a very free one, as mentioned earlier. In most of the examples of expressed possessors in the Old English Orosius, there is no corresponding mention of the possessor at all in the Latin, where implicit possessors are much more common, and so no question of preserving an IP in the Old English translation. That is the case, for example, with (4.9a) above, which lacks a corresponding Latin sentence. It is rather surprising that this translator, who favoured DEPs especially with beheadings and other amputations, chose to use an IP here, but the example illustrates the fact that IPs in translations from Latin did not by any means always translate an IP in the Latin. As already noted, CP sticks much more closely to the Latin text than the translation of Orosius does, and in this text a sentence with an explicit possessor in the English translation is more likely to correspond to a sentence in Latin which also has one. For example, the Latin corresponding to (4.29) differs from the English only in having a synthetic passive verb form and in word order.⁷ ⁷ For the Latin version of the Cura Pastoralis, I have used Westhoff (1860).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

92

       

(4.29)

Sien hira eagan aðistrode ðæt hi ne geseon be: their eyes: darkened that they not see:  ‘May their eyes be darkened so that they do not see’ (cocura,CP:1.29.8.118) Lat:

Obscurentur oculi eorum be.blinded eyes: their

We have also seen that the English followed the Latin in using a DEP in (4.26). However, even in this fairly close translation, the translator clearly was capable of substituting a DEP for an IP or unexpressed possessor in the Latin. One of the three DEPs with object parts in CP replaces an IP in the Latin: (4.30)

& him ðone stiðan suiran forbræce and them: the:.. stiff : neck: broke: ‘and (that he) broke their stiff necks’ (cocura,CP:33.229.3.1499) Lat: rigida colla victorum stiff necks: victor:.

The Latin has a genitive possessor here. It is worth noting also that example (4.30) seems to run counter to what may be a cross-linguistic tendency. In French, DEPs are not normally found when the possessum is in an NP containing an appositive adjective, according to Kayne (1975) and Vergnaud and Zubizaretta (1992). Vergnaud and Zubizaretta’s explanation for this restriction in French, if correct, would seem to predict that this restriction should hold in other languages with DEPs, and because of this I have kept track of the number of appositive adjectives that occur in my IP examples, in order to make sure that the data is not being skewed by counting IPs in a situation in which a DEP would not be possible. It turns out that the number of examples containing appositive adjectives is so small that they do not significantly change the statistics. Since even close translations could diverge from the Latin in using a DEP instead of an IP, there is no reason to think that DEPs were considered unsuitable for translating Latin. Clearly, translation effects did not prevent variation in the use of what were two native OE constructions. It nevertheless remains possible that translation effects did affect the frequency of DEPs in translations, given that these were less frequent than IPs in the Latin. So, for example, in Gregory’s Dialogues, IPs are considerably more common than DEPs, and in this text they usually translate an IP in the Latin. However, even in this more Latinate translation, we find a DEP substituting for an IP:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.7  :  (4.31)

93

& him þæt heafod syþþan of aceorf and him: the: head: afterwards off cut: ‘and then cut off his head’ (cogregdC,GDPref_and_3_[C]:13.198.3.2564) Latin:8 et and

tunc caput illius then head: his

amputa cut.off:

The frequency of IPs in the Dialogues may be affected by the Latin, but clearly Werferth was not averse to replacing a Latin possessive with a dative. We can also note that DEPs are more frequent in Alfred’s laws than they are in the Pastoral Care. The latter translated text is usually attributed to Alfred, and if these two texts do in fact come from the same person, some explanation is in order. I believe that the explanation has more to do with the subject matter of the two texts than the fact that one is translated and the other is not. The laws, with their discussions of injuries such as the loss of an eye and punishments such as cutting off a thief ’s nose, describe more drastic effects, where a DEP is more readily expected than in any of the CP examples. The very different subject matter of CP, which is a manual for priests, similarly provides fewer opportunities for expressing affected possessors of body parts than does the Orosius, in which descriptions of battles and decapitations are common. To sum up, the choice of an IP or a DEP may have been influenced in some individual sentences by translation effects. However, in the texts under consideration, any translation effect is to all appearances a matter of choosing one idiomatic option over another, rather than a distortion of English syntax or normal usage. The fact that IPs are found in native poetry confirms that both options were native. The subject/object asymmetry is also inconsistent with the idea that Latin was responsible for all IPs. Finally, some DEPs were used in the Vulgar Latin originals, a fact which suggests that scribes would have no reason to consider DEPs inappropriate in their translations.

4.7 Direct arguments: conclusions Two major conclusions emerge from the discussion so far in this chapter that are useful to summarize briefly at this point. First, with possessa playing the role of the two direct arguments at least, the widely held view that the DEP was the usual way of marking an affected possessor in the OE texts is demonstrably incorrect. The DEP must be ⁸ For the Latin version of the Dialogi, I have used Moricca (1924).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

94

       

considered a marked way of expressing a possessor. With the direct arguments, its use appears to have been to draw attention to the affectedness of the possessor. However, looking at the texts overall, the IP was the unmarked construction even with highly affected possessors, except with specific verbs. This result is not plausibly accounted for by appealing to Latin influence. The second conclusion is that the frequency of DEPs was significantly lower with body object possessa than with body subject possessa. This is an area where further research is likely to be fruitful. The reasons for the difference are likely to lie in discourse considerations. In Chapter 5, we look at another major finding, namely that DEPs of direct arguments appear to have become increasingly marked within the OE period, being limited to a few verbs of drastic effect by late OE. I will also show that the range of the DEP in early OE seems to have been more limited than its range in Common Germanic. Before moving on to this discussion, however, it remains to look at the situation with DEPs of PObjs in OE.

4.8 Results: objects of prepositions (PObjs) 4.8.1 Some comments on the data For the reasons outlined in 3.3.2 and discussed in detail in Appendix C, I did not count all examples of DEPs and IPs of PObjs with body words; such a count would be difficult and more importantly not terribly revealing because a search for all body PObj possessa on selected representative texts showed that most prepositions did not appear with DEPs. It should be kept in mind that unless otherwise stated, the data and discussion presented in this chapter are limited to examples collected by searches made for these selected words, which I will refer to as my ‘narrow’ investigation, as opposed to a broader investigation using all body words. In addition, as explained in Appendix C, various types of examples that might be treated as DEPs were systematically excluded from the tables of results, and some types of IPs were also systematically excluded in situations where a DEP seems not to have been a possibility, e.g. with instrumental senses of prepositions: (4.32)

witodlice Stephanus wæs to diacone gehadod æt truly, Stephen was to deaconhood consecrated at þæra apostola handum the:. apostle:. hands ‘Truly, Stephen was consecrated a deacon at the hands of the apostles’ (cocathom1,+ACHom_I,_3:201.81.538)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.8 :    ()

95

While the preposition æt is included on the list of prepositions used in the searches, its inclusion is due to its locative meaning. If instrumental examples were included, the disparity in the number of IPs versus DEPs would have been even larger than that reported here. Although I have made some exclusions of this sort, it should be noted that I made no attempt to identify all senses of prepositions where DEPs are not found, and the figures given here are not to be taken as a comparison of DEPs and IPs in situations where both were possible. My main purpose here is not to make a thorough study of variation of DEPs and IPs with PObjs but rather to make the point that internal possession in OE was not nearly as limited with inalienable possessions as it is often assumed to have been, even with PObjs, the type of possessa that occurred most frequently with DEPs. Judgements about these excluded examples means that the exact frequency of DEPs versus IPs would be rather different if different decisions were made about the excluded examples. The numbers of examples excluded are not large, however, and a change here would not modify the overall conclusion that DEPs were more freely used with PObjs than with direct arguments in OE or that they remained more common with PObjs in late OE than with direct arguments.

4.8.2 Range of the DEP with PObjs Before presenting and discussing the statistics derived from the investigations, I will make some general observations about the range of DEPs with PObjs. As noted in Chapter 1, Haspelmath puts prepositional phrases at the top of his syntactic relations hierarchy for DEPs in European languages (1999: 113). It is therefore no surprise to find that in OE, DEPs of PObjs were highly productive. There is no restriction on the syntactic type of verb that could be used in a sentence containing a DEP of a PObj. The (a), (b), (c), and (d) examples of (4.33) illustrate the use of these DEPs with agentive intransitive, non-agentive intransitive, copula, and transitive verbs, respectively: (4.33)

a. & hy crupon þæm mannum betuh þa þeoh9 And they crept the:. men:. between the thighs ‘and they (the fleas) crept between men’s thighs’ (coorosiu,Or_1:7.25.23.493)

⁹ Betuh ‘between’ is not one of my selected prepositions, but this example is included here to round out the range of verbs using a DEP of a PObj. Similarly, in (4.35), eaxl ‘shoulder’ is not one of my selected PObj nouns, but the example is included to illustrate the point.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

96

       

b. him stod sweflan lig of þam muðe him: issued sulphurous flame from the mouth ‘A sulphurous flame issued from his mouth’ (cocathom1,+ACHom_I,_5:221.135.1007) c. gif he sie men on cneowe if it: is: one: on knee ‘if it is on a person’s knee’ (colaece,Lch_II_[3]:59.1.1.4046) d. Þa dyde Martinus on muð þam wodan his Then put Martin in mouth the: mad: his agene fingras, own fingers ‘Then Martin put his own fingers into the madman’s mouth’ (coaelive,+ALS_[Martin]:540.6315) It is no surprise to find that DEPs would appear in expressions of locations like (4.33c), since the dative was widely used in expressions of location in relation to a person, as in (4.34): (4.34)

and him wæs ealne weg weste land on þæt steorbord and him: was all way waste land on the starboard ‘and there was wasteland on this starboard all the way’ (coorosiu,Or_1:1.14.21.239)

It is also not surprising to find a dative used in (4.35): (4.35)

him on eaxle wearð syndolh sweotol him: on shoulder became lasting.wound clear ‘a deadly wound became visible on his shoulder’ (cobeowul,26.815.695) Beowulf 816–17

Example (4.35) describes the coming into existence of the wound, and as discussed in Chapter 2, a dative was frequently used in existential sentences, meaning the dative could be treated as a ‘shared’ one. Alternatively, with a treatment of the dative as a DEP, the use of a dative here is unremarkable because the possessor is clearly adversely affected. It is perhaps more surprising to find that with PObjs, DEPs with dynamic verbs do not appear to have been subject to what Haspelmath terms the ‘strict affectedness condition’, which he says is one of the most important characteristics of the ‘European EP prototype’ (1999: 111). As we have seen in the preceding section, the affectedness condition generally held for direct argument possessa. With PObjs, however, we find datives that are either clearly or arguably to be analysed as DEPs where the possessor is not clearly affected,

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.8 :    ()

97

either adversely or beneficially. No clear effect is visible in examples (4.33b-d) or the commonly occurring expression feallan dat to fotum: (4.36)

& him to fotum feoll. and him: to feet: fell ‘and fell (deliberately) to his feet’

(cobede,Bede_1:7.38.18.318)

The person falling to someone’s feet in a gesture of supplication or submission is clearly trying to have an emotional effect (inciting pity) on the person to whose feet they are falling. Similarly, we might suggest that in (4.33b) the dative was used to highlight the effect that the appearance of the terrifying devil had on the person who saw it, rather than on the possessor of the mouth, the devil. If we extend ‘affectedness’ this far, though, it loses any real meaning. The DEP was clearly used in a wider semantic range with PObjs than with direct arguments, where we can usually objectively point to some physical change in the body, such as the loss of an eye. It appears to be true, however, that the possessor of a PObj was more usually clearly affected in some way, usually adversely, as in (4.33a). As with the direct objects, there was variation between IPs and DEPs in similar functions, even within a single text. For example, in Bald’s Leechbook a highly affected possessor could be expressed with either a DEP or an IP: (4.37)

a. & wið þon þe men acale þæt and with that: that one: totally.chill: the: fel of þam fotum, skin: of the feet ‘and against the situation that the skin on a person’s foot is destroyed by cold’ (colaece,Lch_II_[1]:30.1.1.857) b. manig man hæfþ micelne ece on his eagum, many man has great ache in his eyes ‘Many a man has great pain in his eyes’ (colaece,Lch_II_[1]:2.13.1.226)

Example (4.38) deserves a mention because of the appositive adjective in the PObj: (4.38)

Him weollon þa wurmas of ðam gewitnodan Him: welled then worms from the tormented lichaman, body ‘Worms then welled up from his tormented body’ (coaelive,+ALS_[Maccabees]:544.5187)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

98

       

Examples (4.38) and (4.30) are the only two examples with appositive adjectives in the possessa of DEPs in the examples I have collected, and so there are not enough examples to challenge Vergnaud and Zubizaretta’s cross-linguistic generalization that appositive adjectives are not possible in the possessa of external possessors.

4.8.3 PObjs: DEPs vs IPs Moving from general observations about the range of DEPs in PObjs, let us now compare the use of DEPs and IPs. Table 4.20 compares the number of DEPs and IPs found with PObjs with the selected prepositions and body parts in selected texts with verbal predicates. The reader is reminded that these searches for PObjs were restricted in various ways explained in Appendix C. The first column in Table 4.20 lists the names of the texts, and the second column indicates the category of each text. Table 4.20 DEPs vs IPs of restricted body PObjs in selected OE texts, lexical verbs Text Type York poetry Non-York poetry Poetry total

poetry poetry

bede (early) codocu2 cochronA cocura cocuraC coorosiu coprefcura colawaf colawafint colaece cogregdC (Bk I) ASC(post-EWS) coblick coeuphr coeust comart3 cosevensl coverhom coaelive cocathom1 cowsgosp Prose total

Other Early Other Early EWS EWS EWS EWS EWS 9thC(OE) 9thC(OE) 9thC(OE) 9thC(OE) General OE General OE General OE General OE General OE General OE General OE LWS LWS LWS

Internal

Dat

Blended

Ambiguous

Total

9 4 13

20 6 26

1 0 1

4 1 5

34 11 45

23 0 1 22 3 2 0 0 1 2 20 1 17 3 3 18 3 16 53 56 45 289

10 0 1 3 0 13 0 0 0 5 0 5 2 0 0 6 3 1 14 6 0 69

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

33 0 2 25 3 15 0 0 1 7 20 6 19 3 3 24 6 17 67 63 45 359

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

4.9        

99

It can be seen from Table 4.20 that there is considerable variation within a given text type, both concerning the number of PObj body part possessa found and the preference for IPs or DEPs. One striking fact that this table brings out is that in general there is a big difference between poetic and prose texts: while DEPs were more frequent in poetry, IPs were by far favoured in prose texts in which the number of possessed body part PObjs is sufficiently large overall to make a serious comparison. In most of the prose texts where the total number of examples is ten or more, the number of DEPs does not exceed the number of IPs. The exceptions are the Orosius and Bald’s Leechbook, where DEPs predominate. The numbers in the Leechbook are small enough that the difference between IPs and DEPs is not large enough to be considered significant, but I suggest in section 5.3.2 that the reason for the striking number of DEPs in the Orosius and the later chronicles has to do with the subject matter of these texts. Of the 289 prose examples of IPs in Tables 4.20, thirteen have appositive adjectives, and so probably would be precluded from having a DEP. This still leaves us with a huge difference in frequency between clear IPs and constructions with a dative (DEPs and blended) or ambiguous form. In OE prose, IPs were clearly the preferred possessor type for PObjs in sentences with verbal predicates.

4.9 Conclusions on DEPs of body possessa in OE Cross-linguistic studies have shown that in languages that have DEPs, the grammatical role of the possessum is important in determining whether a DEP can be used. In OE, separating examples of DEPs and IPs with direct argument possessa from PObj possessa results in findings that go beyond those produced by earlier studies into the ‘sympathetic dative’.¹⁰ The widely held assumption that DEPs were extremely frequent in OE turns out to be mainly due to the reduced restrictions on DEPs with PObjs compared with the direct arguments.¹¹ Since prepositional phrases were very common with inalienable possessa, an impression that DEPs were very common in early OE is produced. However, this high frequency does not apply to direct argument ¹⁰ However, Ahlgren (1946: 125) must be given credit for noticing that ‘[i]n prose, constructions without a preposition are but seldom met with’ although he does not provide any statistical evidence. ¹¹ So, for example, Ahlgren’s (1946: 129) statement that in Beowulf ‘constructions with the Dativus Sympatheticus are far more numerous than such with the poss. adj.’ is correct only when we lump PObj and direct argument possessa together.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

100

       

possessa, and even with PObjs, DEPs were in a minority in the prose compared with IPs. In the poetry, DEPs were certainly in a majority with PObj possessa, but IPs were by no means uncommon. Also, earlier studies were mainly concerned with showing differences between OE and Modern English and so naturally presented more examples of DEPs than IPs to illustrate the construction that has disappeared from the language. Cross-linguistic studies also tell us that DEPs are strongly associated with possessors who are affected, either beneficially or adversely. In OE, they are strongly associated with adversely affected possessors with direct arguments, but with PObjs and adjectival predicates no negative effect seems to have been necessary for the use of a DEP. Separating subjects from objects was fruitful because with this categorization of examples it becomes clear that DEPs occurred more frequently with subject possessa than with object possessa, for whatever reason. This is something that requires further investigation. The results of this study show that it is a mistake to assume that OE was simply like modern German. In OE, as apparently in the earlier Germanic languages generally, there was a good deal of variation between IPs and DEPs in accounts of similar actions. It is also misleading to refer to the DEP as the ‘inherited’ construction and the IP as the ‘innovative’ construction. It was not the IP that was innovative but rather the exclusion of the DEP from situations in which both had previously been used.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5 Early changes in English 5.1 Introduction This chapter looks at evidence for changes to the range and use of DEPs and IPs in English even before the earliest records of English, and further changes within the written OE period. In section 5.2, I show that the use of DEPs in even the earliest OE texts is less frequent than what is likely to have been the situation in Common Germanic. Section 5.3 presents clear evidence of a significant reduction in the use of DEPs in the texts within the OE period. These facts are highly relevant to the evaluation of theories about why and how the DEP was lost in English. The chapter focuses on body possessa, since these are the ones that occur most frequently in the texts with DEPs. Some observations about changes with mind words within the OE period will be made in Chapter 6.

5.2 A change from Germanic? Some recent literature gives the impression that affected IPs were not a feature of Common Germanic and were new to English at some stage. Filppula (2008: 30) refers to the ‘innovative internal possessor type’ in OE while Vennemann (2002: 208) speaks of the obligatory use of a DEP for affected possessors as ‘the inherited construction’; his later discussion makes it clear that by ‘construction’ we must understand ‘situation’. If DEPs were the only inherited construction used with affected possessors in OE, it would mean that the use of an IP for such possessors was an innovation of the early OE period. However, Havers (1911: 317) concludes that the Indo-European language families generally showed an interchange between the genitive and the ‘dativus sympatheticus’. For example, he shows (p. 1) that both constructions were used in Homeric Greek in what appear to be descriptions of the same situation, differing only in the attitude the writer is expressing towards the event. Havers’ discussion of the early Germanic languages shows that variation was the normal state in all of them. Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

102

   

In this section, I argue that it seems likely that there was a decline in the range of DEPs already between Common Germanic and early OE, prior to the decline within the OE period discussed in the next section. Although Havers’ (1911) survey as well as grammars of the individual early Germanic languages give us some information about the use of DEPs in those languages, systematic studies are lacking, and a comparison of all the earliest daughters of this language is beyond the scope of this book. However, in this section I take a limited look at Gothic and Old Saxon to see how similar or different the use of DEPs and IPs was in these two languages from the findings for OE.

5.2.1 Gothic I read the four gospels in Germanic looking for examples relevant to the use of IPs and DEPs with body possessa, using the Wulfila project (http://www. wulfila.be). This investigation was not systematic in the sense that I did not collect all relevant examples, but only some representative ones, and I compiled no figures. Since Gothic is the earliest Germanic language for which we have extensive records, it would be reasonable to assume that the syntax of Gothic reflects Common Germanic syntax reasonably well, all other things being equal. Of course, all other things are not actually equal. Our only extensive Gothic text is a translation from Greek of the major part of the gospels and St. Paul’s epistles. The study of Gothic syntax is greatly complicated by the fact that in some respects, such as the heavy use of participles, it is clearly affected by the Greek exemplar. However, most recent scholarship on Gothic syntax (e.g. Ferraresi 2005) has reached the conclusion that the Gothic text was no slavish translation; it is nothing like a word-for-word gloss.¹ The possibility of Greek influence on the Gothic text would render any attempt to compare the frequency of DEPs and IPs dubious, but an important fact that becomes clear from comparing the Gothic and the Greek is the fact that the ¹ As discussed by Ferraresi (2005: 2) and the electronic Wulfila Project (http://www.wulfila.be/), we do not know what Greek text Bishop Wulfila used. Nevertheless, the corpus is large enough that it is reasonable to expect that we should be able to get a good picture of consistent divergences from the Greek that was most likely to have been in Wulfila’s Greek original. The Greek presented by the Wulfila Project is that of the electronic version of the Nestle-Aland Novum Testamentum Graece (26th/27th edition). Although the discussion on this project’s website emphasizes that this version is clearly not Wulfila’s exemplar and is not reliable as the basis for a discussion of translation techniques, Ferraresi (2005: 3) comments that the syntax used in this critical edition usually coincides with that found in Streitberg’s (1919) attempted reconstruction of Wulfila’s original, the Greek text usually used by scholars. Whenever a comparison of ‘the Greek’ syntax is made with the Gothic text, I have checked the Greek on the Wulfila Project website against Robinson and Pierpont’s (2005) edition of the Byzantine text to make certain the two versions are the same in the relevant syntax.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.2    ? 103

Gothic treatment of possessed body parts frequently diverges from the Greek. It is therefore reasonable to expect that the Gothic translation can yield some useful information on this area of syntax in the Common Germanic period, even if the picture might be a bit distorted. The examples of (5.1) show that the Gothic does not slavishly follow the Greek original in the syntax of body parts. In Biblical Greek, the dative and genitive cases had not yet undergone the syncretism found in Modern Greek, and possessors of body parts could appear in the genitive case or as possessive ‘adjectives’/pronouns as well as in the dative case. In these examples, the Gothic uses a DEP while the (presumed) Greek original had a genitive: (5.1) a. sa izei uslauk augona that:.. who: .. opened eyes: þamma blindin the:.. blind:.. ‘he who opened the blind man’s eyes’ Gk. οὗτος ὁ ἀνοίξας τοὺς ὀφθαλμοὺς τοῦ this the having-opened the:. eyes . the: τυφλοῦ blind: John 11:37 b. iþ Seimon Paitrus . . . afmaimait imma auso Then Simon Peter severed him: ear: taihswo right: ‘Then Simon Peter . . . cut off his right ear’ Gk. καὶ ἀπέκοψεν αὐτοῦ τὸ ὠτάριον τὸ δεξιόν and cut-off his the ear: the: right: John 18:10 Harbert (2007: 166) notes that the DEP is not used in Gothic (or early Germanic languages generally) ‘in cases when only the meronym, not the holonym, is affected by the action’. That is, when the body part (the meronym) has been affected, but this does not affect the possessor (the holonym), an IP is the only construction which is found, as in ushafjands augona seina ‘raising eyes his’ at Luke 6:20. Note that the effect could be negative, as in (5.1b), but it could also be positive, as in (5.1a). This makes Gothic different from OE, where, as we have seen, the use of a DEP for a beneficial effect is nearly unknown, even in poetry, which is probably generally representative of an older use of DEPs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

104

   

While Harbert only indicates that the DEP is not found when there is no effect on the holonym, it would be easy to jump to the conclusion that this means that the IP was not used in Gothic with affected possessors. In fact, however, sentences with a highly affected IP are not rare: (5.2) usluknoda þan munþs is suns opened then mouth:. his immediately ‘his mouth was unlocked immediately’

Luke 1:64

Example (5.2) contrasts with (5.1a), where a dative is used for a similar beneficial effect on the possessor. The Gothic of (5.2) is like the Greek in using an IP, but as has been established by the examples of (5.1), the Gothic translator was not unwilling to depart from the Greek in its treatment of the possessors of body parts. It seems that an IP was an acceptable alternative in Gothic to a DEP even when the holonym was substantially affected. This does not mean that the translation was not affected in any way by the syntax of the original. Havers notes that the prenominal or postnominal position of a genitive in the Greek seems to affect the choice of an IP or DEP, with the postnominal genitive usually remaining genitive in the Gothic (1911: 257–67). So it is entirely possible that Greek influence resulted in more IPs than would have been natural in Gothic. The important point, however, is that both IPs and DEPs were grammatical, since the Gothic did not always follow the Greek in the use of an IP. It is worth noting that Harbert’s statement about DEPs not being used when only the meronym is affected by the action does not seem to hold for PObj body part possessa. In (5.3a) the Gothic has a DEP translating a possessive pronoun in the Greek in a sentence in which the possessor is not physically affected. This can be compared with (5.3b), where the Gothic follows the Greek in using a possessive: (5.3) a. gasaihvandei ina draus imma du seeing him: fell him: to Gk. ἰδοῦσα αὐτὸν ἔπεσεν αὐτοῦ πρὸς seeing him: fell his towards ‘seeing him, (she) fell at his feet’

fotum feet τοὺς πόδας the feet John 11:32

b. qimandei draus du fotum is coming fell to feet his ἐλθοῦσα προσέπεσεν πρὸς τοὺς πόδας αὐτοῦ coming fell towards the feet his ‘coming, (she) fell to his feet’ Mark 7:25

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.2    ? 105

Note that in the verse at John 11:32 this genitive possessor precedes the prepositional phrase and is discontinuous from the NP containing the possessum, although it is not in the dative case, while in the verse from Mark, the genitive follows and is presumably internal to the NP. The two examples thus follow Havers’ basic generalization about the translation strategy used in the Gothic. It appears that Gothic was similar to OE in using DEPs more freely with PObj possessa than with direct argument possessa. To sum up, it appears that the range of DEPs was wider in Gothic than OE, specifically in the free use in Gothic of DEPs for beneficially affected possessors. Assuming that Gothic is fairly representative of Common Germanic, it appears that the decline of the DEP began in English even before the first writings. However, Gothic was like OE in allowing variation between DEPs and IPs with highly affected possessors. This variation appears to be greater than what is allowed in Modern German, where the use of an IP instead of a DEP suggests a lack of effect, a clinical description. Gothic also seems to have been like OE in allowing DEPs for unaffected possessor of PObjs, e.g. in descriptions of people falling to someone’s feet. We cannot be certain of exactly what the situation was in Common Germanic, and further systematic study of early Germanic languages are needed here. However, it seems reasonable to conclude that while only IPs were used for unaffected possessors of direct argument possessa, IPs and DEPs were both possible with possessors who were affected, either adversely or beneficially.

5.2.2 Old Saxon The discussion of Old Saxon will be limited to the Heliand, an alliterative poem.² A look at Old Saxon is useful for the discussion of the decline of DEPs in early English for two reasons. First, unlike the Gothic texts, we have native poetry in the same Germanic tradition as OE poetry, and translation effects are not likely to be an issue in such poetry. Second, Old Saxon is more closely related to OE than Gothic is, and the Heliand is from roughly the same period as the earliest extensive English writings; Cathey (2002: 21) puts it at around 850. This genetic and temporal proximity makes any syntactic differences between Old Saxon and OE very interesting. In this instance, any big differences in the use of IPs and DEPs in the two languages bears on the ² For this investigation, I used Walkden’s (2015) HeliPad, which is based on Sievers’ (1878) edition of the manuscript known as C, which preserves 5968 lines.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

106

   

question of Celtic influence. Since the IP in English has been argued by adherents of the Celtic Hypothesis to be due to Celtic influence, and since the use of the DEP was more restricted in OE than in Gothic, if we found that Old Saxon showed a similar reduction in the use of DEPs as OE compared to Gothic, that would be evidence against Brythonic Celtic having triggered this restriction in OE, given that it is generally assumed by adherents of the Celtic Hypothesis that OE was more affected by Celtic influence than was Old Saxon. On the other hand, if we find that Old Saxon poetry used DEPs more freely than OE poetry did, this would strengthen the reconstruction we make of Common Germanic from Gothic and would also lend some support to the idea that English was unusual among the Germanic languages even at the time of its earliest writings. This support would not actually be terribly strong, since OE could have undergone a reduction in the use of the DEP for reasons other than Celtic influence, but it would at least be consistent with the Celtic Hypothesis. The number of relevant examples in the Heliand is not large, making an attempt at a statistical comparison with OE not worthwhile. It can be said, however, that DEPs are not difficult to find in the Heliand. One important difference Old Saxon shows with OE seems to be a freer use of datives to report beneficial effects. As with Gothic, a DEP is used to report Christ’s opening of the eyes of the blind: (5.4) a. that them blindon thuo bethion uurthun ogon that the: blind:. then both: were eyes: giopanod opened ‘that both of the blind men’s eyes were opened’ Heliand 3580–1 In contrast to the Gothic and the Old Saxon versions of this miracle, OE uses a possessive both here and in all other descriptions of Christ opening someone’s eyes or mouth: (5.5) And hyra eagan wæron ontynede and their eyes: were opened ‘and their eyes were opened’ (cowsgosp,Mt_[WSCp]:9.30.568) Although DEPs in accounts of beneficial effects are not entirely lacking in the OE examples of lexical verbs with subject possessa, they are rare, as seen in section 4.3. No clearly beneficial effect shows up in the data of object possessa, as discussed in section 4.2. It appears that DEPs were used to

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.2    ? 107

express beneficially affected possessors in Common Germanic, since they are found in both Gothic and Old Saxon, and their near non-use in OE is an innovation. Interestingly, a bit earlier in the report of the miracle just discussed, the Heliand poet uses a blended construction, combining a dative with a possessive: (5.6) sia badun ina helagna that hie im iro they asked the holy that he him: their: ogun opana gidedi eyes: open: did ‘they asked the holy one to open their eyes’ Heliand 3574–5 The existence of blended examples in this text must cause us to reject any idea that blending in OE was something that happened because DEPs were becoming unusual, since DEPs are abundant in this Old Saxon text. Another difference between the Heliand and OE generally is the frequent use in the Old Saxon text of datives with copulas in pure descriptions. Compare, for example, the DEP in Old Saxon in (5.7a) with the IP in the OE of (5.7b): (5.7) a. lik uuas im sconi body: was him: beautiful ‘his body was beautiful’

Heliand 199

b. cwæð þæt his lic wære leoht and scene said that his body was: light and shining ‘(he) said that his body was light and shining’ Genesis (B) 265 As noted in section 4.4, a DEP in pure descriptions was found in OE, but not frequently. The data contain only one example from the poetry, and the prose examples are particularly associated with fairly close translations from Latin. It appears that while DEPs in this construction in OE were grammatical, they were less favoured than in Old Saxon. The difference here is quantitative rather than qualitative. On the other hand, there are important similarities between Old Saxon and OE in the expression of possessors of body parts. As seen in section 4.4, DEPs in OE were not restricted to adverse effects with copulas, making OE similar to Old Saxon in that respect. It is also important that OE was similar to both Old Saxon and Gothic in that the IP was in variation with the DEP even with adversely affected possessors. Havers (1911: 295) notes an instance of this

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

108

   

variation in the same line of the Heliand, where we first have the possessive mina and then the dative mi: (5.8)

thar uuerthat mina hendi gibundan fathmos there become my: hands: bound arms: uuerthat mi thar gifastnod become me: there fastened ‘my hands will be bound there; my arms will be fettered’ Heliand 3526–7

I present another contrast in (5.9), with a dative in the first example and a possessive in the next line, both in a list of the characteristics of old age: (5.9) a. flesk is unc afallan flesh is us: fallen ‘our flesh has fallen away’ b. sind unca andbari odarlicron are our: appearance: altered ‘our appearance is altered’

Heliand 153

Heliand 154

In fact, with object possessa, even highly affected possessors are more frequently expressed as IPs than as DEPs in the Heliand, as in OE poetry. While the fact of this variation in Gothic might be attributed to Greek influence, the variation in the Heliand and preference for IPs with object possessa makes it even less likely that the appearance of IPs with highly affected possessors in OE is due to Latin influence.

5.2.3 Summary: Common Germanic and Old English The Gothic and Old Saxon facts presented here indicate that from what we can reasonably construct about DEPs in Common Germanic from these two languages, OE had undergone a substantial reduction in the use of DEPs by the time the earliest extensive texts were written. This reduction is mainly quantitative rather than qualitative. The closest thing to an absolute difference between OE and its close relative Old Saxon appears to be in the near complete absences of DEPs to express beneficially affected possessors in OE. As seen in Chapter 4, no examples of DEPs with beneficially affected possessors are to be found in the corpus studied here with object possessa, and with subject possessa such possessors are very few and limited to early texts. With subjects

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      109

of copulas and PObjs, however, OE is qualitatively similar to both Gothic and Old Saxon, but here also reports of adverse situations predominate in OE. The quantitative difference, impressionistically, seems large; DEPs of body possessa in copular constructions at least are considerably easier to find in the Heliand than in OE poetry of a similar length. It is especially interesting that OE had diverged from a close relative in this respect. However, the frequent use of blended possessors in Old Saxon complicates matters since it does not fit in simply with a simple picture of reduction in OE. The ramifications of this complexity for the idea of Celtic influence triggering the eventual loss of DEPs in English is discussed in section 8.3.3. In addition, the fact that blended examples are frequent in Old Saxon in a text where DEPs are so frequent is important because it goes against any idea that blending was a feature that only appeared when the DEP was on its way to replacement by IPs. Although it appears that OE started down the path towards lesser use of DEPs at an early stage, the importance of the fact that variation between DEPs and IPs is a feature of Gothic and Old Saxon as well as OE can hardly be overstated. What was unusual in OE was not such variation, but the frequency of use of IPs even with highly affected direct argument possessa with lexical verbs. A more systematic study of DEPs in early Germanic languages other than OE must be left to further research, but the facts established in this section indicate that the range of the DEP had contracted in OE by the time of the earliest writings. Having established the nature of this early reduction in DEPs, I move on to show a further reduction in the frequency of DEPs within the OE period.

5.3 A change within OE? Range and frequency of DEPs In section 5.2, I argued that DEPs in OE were probably more restricted than they were in Common Germanic. It appears that the decline of the DEP in English may have begun before we have written materials, and this reduction in the range of the construction continued in the OE period. A major finding of this investigation is that the disappearance of DEPs in the texts was not as sudden as has frequently been assumed. This section presents statistical evidence that the frequency of DEPs compared with IPs declined between earlier and later OE, looking first at direct arguments and then at PObjs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

110

   

5.3.1 Direct arguments This discussion of direct arguments is restricted to examples using verbs on the ‘affecting’ list. In order to control the variables of dialect and time as much as possible, we would get the clearest picture by restricting our examination of early and late periods to the same general dialect. Since most extant OE texts are written in a West Saxon dialect, that means a comparison of EWS and LWS. The figures for such a comparison are given in Table 5.1. ‘Blended’ examples are excluded from this table. Table 5.1 IPs vs DEPs of direct argument body possessa with affecting verbs in texts of EWS and LWS composition

Objects Subjects

EWS Comp. LWS Comp. EWS Comp. LWS Comp.

IPs

DEPs

13 109 6 41

22 7 18 9

Table 5.1 is based on figures presented in Chapter 4. The figures in the ‘EWS Comp.’ (Early West Saxon composition) cells combine the figures for the anchor EWS texts with those for three texts thought to be composed in the EWS period but found in manuscripts of up to fifty years later: colaece, colawaf, and colawafint. These totals are from Table 4.4 and Table 4.13 for object and subject possessa, respectively. The ‘LWS Comp.’ totals combine the totals given for LWS and LWS (late) figures given in Table 4.6 (object possessa) and Table 4.15 (subject possessa). The difference between the two periods is striking for both subjects and objects, with body possessa of both types much more likely to show up with DEPs in West Saxon texts from the early period than from the later period. However, it is unsafe to draw strong conclusions from this comparison about what was happening. For one thing, the LWS figures are dominated by texts from a single author, Ælfric. The disparity in the number of relevant examples from the two periods, including the much smaller numbers from the earlier period, also makes chi-squared tests unreliable for this data. For this reason, the text types are grouped into broad categories of ‘early’ and ‘late’. In Table 5.2, General OE texts are put in the ‘late’ category. Some of these texts

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      111

may actually be of fairly early composition but their date of composition is uncertain, and they are only found in manuscripts that are not of the earliest period. The figures for blended possessors are included in Table 5.2 for later discussion, but these are not included in the following discussion of chisquared tests. Chi-squared tests indicate that the difference in the periods is highly significant for both objects and subjects. For objects, chi-squared = 33.365, with a minuscule p-value of 6.648.e-09. For subjects, the chi-squared value is 10.0642 and the p-value is 0.001106. It must be kept in mind that this test only shows that the difference in the texts of the two groups is not likely to be due to chance; it does not actually tell us that it is the difference in the time of composition of the texts that is the source of the difference. Also, the works of Ælfric still dominate the data of the later period. The inclusion of the General OE texts does not eliminate this bias but it does lessen it. If our investigation of later texts were confined to the works of Ælfric, an apparent decline might have more to do with personal preferences than a diachronic development. However, the fact that the General OE texts show a lower frequency of DEPs than the texts categorized as early indicates that there was a general decline, even if Ælfric may have been at the extreme end of restricted DEP usage. Table 5.2 IPs vs DEPs of direct argument body possessa with affecting verbs in early and late OE prose texts

Objects

Subjects

Other Early EWS 9thC(OE) Early total General OE LWS LWS(Late) Late total Other Early EWS 9thC(OE) Early total General OE LWS LWS(Late) Late total

IPs

DEPs

Blended

0 9 22 31 45 80 29 154 1 4 8 13 24 39 2 65

2 12 13 27 11 4 3 18 1 4 17 22 18 5 4 27

1 1 3 5 1 0 0 1 0 0 2 2 2 1 2 5

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

112

   

The widespread belief of a rather sudden disappearance of DEPs in the EME period is based on impressions of what the texts contain, and this more systematic examination of these texts shows unambiguously that these impressions are wrong. The texts may not represent what was going on in the spoken language accurately, but they are all we have to work with, and the most obvious explanation for the difference between the earlier and later texts of the OE period is that there was indeed a general increase in the use of IPs as compared with DEPs within the OE period. It is interesting to observe that the ‘blended’ construction, combining a dative and an IP, does not show evidence of an increase as the IPs became more dominant, contrary to what one might have expected.³ It is true that Table 5.3 shows no examples of blending in the poetry at all, compared to a small number of blended examples in the prose texts, but the number of examples of direct argument body possessa is so small that the absence of blended examples in the much smaller poetry corpus does not carry much weight, and at any rate there is a blended poetic example with a PObj, as indicated in Table 5.4. More importantly, there is no discernible increase from early to later prose. While it is true that the raw number of blended possessors of body subjects is larger in the later period, as indicated in Table 5.2, we would expect a much larger increase due to the disparity in the size or the corpora of the two periods if blended constructions were becoming more frequent. With body objects, there is actually a decrease in raw numbers of blended possessa. We have furthermore seen in section 5.2.2 that blended possessors were common in Old Saxon, in copular constructions at least. It is finally worth having a look at whether there is any difference between poetry and early prose that might point to a decline in the use of DEPs in the OE period even before the time of the earliest prose. The reason why we might expect to find such a decline is that although the time of composition of most poetic texts is uncertain and may be late in some cases, it is reasonable to assume that the conventions of poetry, when they differ from prose, are likely to enshrine older syntax. The relevant figures for the early prose have already been given in Table 5.2. The figures for poetry given in Tables 4.4 and 4.13 are repeated in Table 5.3.

³ Ahlgren’s (1946: §130) comment that this blended construction was found during the process of the ‘poss. adj’ taking over the function of the ‘dativus sympatheticus’ does not actually state that English went through such a stage, but the view that the IP was a replacement of an older construction suggests it.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      113 Table 5.3 IPs and DEPs of direct argument body possessa with affecting verbs in OE poetic texts

Objects

Subjects

York Non-York Poetry total York Non-York Poetry total

IPs

DEPs

Blended

8 3 11 2 4 6

4 3 7 2 3 5

0 0 0 0 0 0

Comparing the figures of Table 5.3 for poetry with the figures given for prose in Table 5.2, we see that DEPs in fact occur with a lower frequency in poetry than in the early prose, contrary to expectations. Setting aside blended and ambiguous examples, DEPs occur in 47 per cent of the early prose examples with object possessa of affecting verbs, compared with only 39 per cent in the poetry. The possessor of a subject of an affecting verb is expressed as a DEP in 63 per cent of the prose examples, but only 45 per cent of the poetic examples. However, the poetic figures are too small to make this difference meaningful, and as noted in section 4.2.2, the inclusion or exclusion of a single text can significantly alter the percentages in a given text type. It is demonstrated in sections 4.2 and 4.3 that some verbs are especially likely to occur with DEPs, and it is probable that the fluctuation in the frequency of DEPs that we find in the early prose texts is at least partly due to differences in the frequency of use of such verbs in different texts. Idiosyncratic scribal preferences might also play a role in this variation. To sum up the comparison of poetry and early prose, no significant difference can be demonstrated because of the paucity of poetic examples. This comparison does have the useful result of adding to the evidence against Latin influence in the use of IPs presented in section 4.6. If we had found a reduced frequency of DEPs in the early prose, it might be argued that this was due not to a temporal difference but rather to the influence of Latin in the use of IPs, an influence that was more likely to be found in prose than in native poetry. The fact that DEPs are just as common, or more so, in some early prose at least, goes against a possible role of Latin influence on those texts.

5.3.2 Objects of prepositions The decline of DEPs within the OE period that is found with direct argument body possessa is also discernible with PObjs. Table 4.20 presented figures for

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

114

   

DEPs and IPs with PObjs in poetry and individual prose texts. In Table 5.4, these figures are presented by text type. Table 5.4 DEPs vs IPs of restricted body PObjs with lexical verbs in OE poetry and selected OE prose text types Text Category

IP

DEP

Blended

Ambiguous

Total

% IP

Poetry Other Early EWS 9thC(OE) General OE LWS

13 23 28 23 61 154

25 10 17 5 17 20

1 0 0 0 0 1

5 0 0 0 0 0

44 33 45 28 78 175

30 70 62 82 78 88

Table 5.4 indicates a decline in the frequency of DEPs of objects of prepositions within the OE period. As with direct arguments, we must remember that the bulk of the texts given for Late West Saxon are from a single author, Ælfric, and it is always possible that Ælfric was not representative of OE speakers/ writers in his preference for IPs. The fact that the frequency of DEPs in the LWS texts is not significantly lower than the frequency in the General OE category indicates that Ælfric was not unusual in his preference for IPs with the PObjs studied here, however. As with direct arguments, the decline of DEPs with PObjs seems to have started before the Middle English period. Comparing the decline in DEPs of PObjs with their decline with direct arguments, the differences are not striking. The ratio of IPs to DEPs of PObjs in Ælfric’s writings, for example, is smaller than with object possessa, but is about the same as with subjects, where DEPs were always more common than they were with objects. It can also be noted that as with the direct arguments, there is no increase in the ‘blended’ construction. The only ‘blended’ example turned up by the search with the restricted lists of prepositions and body words is (5.10): (5.10) Þærtoeacan him wexað untrumnyssa on his lichaman additionally him: increases illness in his body ‘In addition to that, illness increases in his body’ (cocathom1,ÆCHom_I,_4:210.122.749) The late appearance of this example, at a time when DEPs had declined substantially, suggests that analysis in terms of a dative being added to reinforce the possessive at a late stage is preferable to the idea of a possessive added to the dative.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      115

Statistics like those presented in Table 5.4 cannot tell us why DEPs of PObjs are still found so frequently in late OE writings, but a consideration of individual examples can suggest some reasons. Looking at the examples, we see that a small number of combinations account for a large proportion of these DEPs. OE used a number of idioms involving DEPs with  (=‘hand’) in meanings of submission or surrender; for example, Cyninge to handum gebigan is listed as a phrase meaning ‘to subject (something acc.) to the king’ in the DOE lemma gebigan, a word whose main sense is ‘to bend’. It is therefore not a surprise that DEPs are found in a high frequency in combinations of the prepositions on and to with  (and variant spellings and forms of this lemma). This is clearly shown by Table 5.5, which compares the number of DEPs found in different combinations in all EWS and selected LWS texts. The purpose of Table 5.5 is to make two comparisons. First, the leftmost and centre columns report the findings for DEPs of PObjs of all the body words versus the restricted list of body words used in the general investigation of PObjs, respectively. Both counts used the same restricted list of prepositions used for the PObj counts generally. These figures confirm that the small number of nouns chosen for the purposes of comparing DEPs with IPs generally account for almost all of the DEPs with the selected prepositions; the figures in the middle column are either the same or slightly lower than those in the leftmost column. The second comparison that can be made from this table is between the frequency of DEPs with the idiomatic expressions under discussion and PObjs generally. The final column gives the figures for a study that was limited to the prepositions on and to with the object . It turns out that these combinations form a substantial proportion of all DEPs with PObj possessa.

Table 5.5 Comparison of DEPs in combinations of P + PObj

EWS cochronA cocura cocuraC coorosiu coprefcura Total EWS LWS coaelive cocathom1 cowsgosp Total LWS

All body N

Selected N

On/to hand

1 4 0 13 0 18

1 3 0 13 0 17

1 0 0 12 0 13

17 6 0 23

14 6 0 20

4 2 0 6

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

116

   

This is particularly striking for Early West Saxon, where thirteen of the total of eighteen DEPs are of either   or  , e.g. (5.11): (5.11) a. þæt him se cyning self on hand eode that him: the: king: self in hand went ‘that the king himself surrendered to him’ (coorosiu,Or_4:10.104.29.2156) b. þy læs ðe eow on hand becume seo the: less that you: in hand come the: lease gesetnyss false: composition: ‘lest the false composition fall into your hands’ (cocathom1,+ACHom_I,_30:429.20.5895) In LWS, exemplified by (5.11b), these combinations account for a bit more than a quarter of the DEPs of PObjs. This apparent decline in the presence of these combinations in LWS cannot be taken as an indication of syntactic change during the OE period because it can be explained by differences in the subject matter of the texts of the two periods, and the domination of the Orosius in the EWS examples. Importantly, more than one idiom involving coming into someone’s hands referred to coming under someone’s control or surrendering. The DOE headword becuman lists a sense becuman on hand ‘to come into (someone’s) power’, and the headword gan ‘go’ gives a sense (III A.6.b) ‘submit to someone’ for the phrases on/under hand gan and gan to handa. The Orosius is largely concerned with descriptions of wars and conquest, with the result that references to surrenders and taking power over people are common. As Table 5.5 shows, twelve of the thirteen DEPs of PObjs in the Orosius with the selected prepositions involve on/to hand; these are all descriptions of surrenders or submissions. Our conclusion must be that in both the early and the late periods of West Saxon, IPs were the unmarked possessive construction for the body objects of at least the selected prepositions, but DEPs were found fairly frequently in certain idioms, although not restricted to them. Although DEPs were common in the idioms just discussed, it seems also that IPs were the unmarked construction in Late West Saxon even in combinations that were especially favourable to DEPs. While Table 5.5 presented figures only for DEPs, Table 5.6 compares the number of DEPs and IPs with /  in the two West Saxon periods.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      117 Table 5.6 DEPs and IPs with on/to hand, EWS vs LWS EWS cochronA cocura cocuraC coorosiu coprefcura Total

DEP

IP

LWS

DEP

IP

1 0 0 12 0 13

2 4 0 0 0 6

coaelive cocathom1 cowsgosp

4 2 0

8 11 9

6

28

We see that in the Early West Saxon period, DEPs were favoured over IPs with these combinations, but the selected Late West Saxon texts show a strong preference for IPs. It is interesting to see that variation is found in the writings of Ælfric even in the literal meaning of physical location in the hands: (5.12) a. and Abraham hæfde him on handa fyr and swurd; and Abraham had him: in hands: fire and sword ‘and Abraham had fire and a sword in his hands’ (cocathom2,+ACHom_II,_4:34.138.760) b. Þa genam se halga wer on his handa þæt lic, then took the holy man in his hands: the body ‘then the holy man took the body in his hands’ (coaelive,+ALS_[Martin]:1027.6635) However, IPs were the more common construction with this meaning for Ælfric, and I had to go to texts not covered by the above tables to find example (5.12a). With some other combinations, such as ascufan to handa/handum, for which the DOE gives ascufan (sense 2.c.iii) the translation ‘thrust/deliver (someone) into the grasp/hands of (someone gen/ dat)’, we find variation in the same text: (5.13) a. & neadunga þone witegan him to handum and unwillingly the prophet them: to hands: asceaf pushed ‘and unwillingly delivered the prophet to their hands’ (cocathom1,+ACHom_I,_37:504.204.7462) b. Gif he þonne nele þæt he beo . . . to deofles if he then not.will that he be . . . to devil: handa ascofen hands: pushed ‘If he then does not wish to be delivered into the devil’s hands’ (cocathom1,+ACHom_I,_19:331.172.3757)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

118

   

The examples of (5.13) are good evidence of variation in the language of a single person, Ælfric; the manuscript containing this text was from Ælfric’s own scriptorium, incorporating corrections in his handwriting (Clemoes 1997: 1), and so almost certainly reflected his language faithfully. Although variation was clearly possible, however, it is clear that DEPs were favoured in some combinations of prepositions with body objects as opposed to others. For example, that descriptions of shoving something into someone’s mouth seem to favour DEPs even at the LWS stage. It is beyond the scope of this book to investigate the question of what controlled the variation between internal and external possessors of PObjs, but it is clear enough that variation existed at the level of the individual, not just at the level of the community. It is worth noting here that despite the large size of the Ælfrician corpus, we do not always find the variation that we might expect with commonly occurring combinations. For example, I have not found any examples of DEPs in my broader study of Ælfric’s works in expressions meaning ‘fall at someone’s feet’. This is not likely to be a data gap, because I found forty-four examples in Ælfric’s works with an IP in the meaning of falling at someone’s feet or stretching out at someone’s feet in a gesture of submission: (5.14) þa feoll heo to hys fotum then fell she to his feet ‘Then she fell to his feet’

(coaelhom,+AHom_6:74.910)

It seems that Ælfric simply preferred the genitive in this construction (at least in writing). In contrast, my broader investigation indicated that Ælfric preferred the DEP in expressions meaning ‘come to someone’s attention through hearing’: (5.15) Þa com him to earan be Agathes drohtnunge Then came him: to ears about Agatha’s lifestyle ‘Then he heard about Agnes’ way of life’ (coaelive,ÆLS[Agatha]:7.2016) The fact that Ælfric shunned the DEP with ‘fall to x’s feet’ is rather sobering, because if we did not know that the expression ‘fell him to feet’ was still fairly common late in Middle English and only looked at the expression in Early West Saxon and Late West Saxon, we would get the false impression that the genitive had completely overtaken the dative in this combination by the end of the OE period. I have not in fact found a dative/genitive variation in my

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

5.3    ?      119

narrower investigation in any single text in the OE period, although I have found variation within the same text category, namely General OE: (5.16) a. & him to fotum feoll and him: to feet:. fell ‘and fell to his feet’ (cobede,Bede_1:7.38.18.318 &) b. and heo sona feoll to his fotum and she immediately fell to his feet:. ‘and she immediately fell to his feet’ (coeuphr,LS_7_[Euphr]:138.145) It seems likely that the use or not of a DEP with a PObj depended on a number of factors, including the combination of verb, preposition, and body word. In the preceding discussion in this section, I have focused on early and late West Saxon in order to look at possible diachronic changes in one dialect. It is appropriate to mention here that some of the General OE texts are actually of later OE composition than the LWS texts. In particular, late entries in the Anglo-Saxon Chronicles were obviously composed later than the LWS period of the early tenth century.⁴ When we look at the individual examples of DEPs in these annals, we see that the relatively high frequency of DEPs is due to the same factor that caused a high frequency in the Orosius—the Chronicle makes frequent reports of surrenders and also of property coming into someone’s control. An example that was composed considerably later than the LWS texts but still falling within the General OE period is given in (5.17): (5.17) and Scotta cyng eode him on hand and Scots’ king went him in hand ‘and the king of the Scots submitted to him’ (cochronD,ChronD_[Classen-Harm]:1031.1.1757) As we see in Chapter 7, DEPs of PObjs continue to be found in expressions of being in someone’s control in the First Continuation of the Peterborough Chronicle in entries that were composed in the first quarter of twelfth century. To summarize the findings on possessors of objects of prepositions in OE, a decline in the use of DEPs in favour of IPs is evident within the OE period. This is so even though the prepositions and nouns selected for comparison here favoured DEPs in that they were selected precisely because they seemed to occur with a comparatively high frequency of DEPs. Although the use of DEPs ⁴ For the treatment of examples in the chronicles other than the A manuscript and the later annals of the A manuscript, see Appendix A.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

120

   

remained more frequent in late OE with PObjs than with the direct arguments, as far as we can tell from the surviving texts, their use seems to have been restricted to a fairly narrow range of widely-occurring expressions even in the earliest prose. The outlines of an analysis that would account for this difference is suggested in Chapter 8.

5.4 Conclusions on early changes Although Ahlgren (1946: 131) states that ‘[t]he decline in the use of the Dativus Sympatheticus began in OE’, some more recent writings (discussed in Chapter 8) have assumed that DEPs disappeared suddenly in the Middle English period, and previous attempts at explaining the loss of the construction in English have focused mainly on the ME period. In this chapter, I have given statistical evidence backing up Ahlgren’s impressionistic observation about a decline within the OE period. The finding that DEPs were strongly associated with individual verbs and idiomatic phrases suggests that what was a productive construction in an earlier period was well on its way to becoming lexically governed by the late OE period. Developments in the ME period merely put an end to a construction that was already in decline before the ME changes that have been suggested as triggering the loss of this construction. The comparison of OE with Gothic and Old Saxon suggests that the more restricted use of the DEP began in English by the time of the appearance of the earliest substantial OE writings but also establishes that the use of IPs in variation with DEPs in reporting similar events and states was by no means an innovation in OE. This variation in early Germanic will be particularly important in Chapter 8 in the evaluation of hypotheses concerning the eventual loss of the DEP. This chapter and Chapter 4 have been limited to body possessa. Chapter 6 looks at DEPs with mind words in the OE period.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6 Mind and dative external possessors in Old English 6.1 Introduction Like words referring to bodies and part of bodies, words referring to the mind and spirit, being inalienable possessa, are high on the hierarchy as candidates for DEPs. This chapter looks at DEPs with selected mind words in OE. Some words referring to body parts, especially words for ‘heart’ and ‘bosom’, have metaphorical meanings related to the mind or emotions. Examples with those metaphorical meanings were excluded from my body counts but are included in the data for this chapter. Sometimes the line between the literal and the metaphorical is blurry, but I have followed the rule of treating these words as mind words when they seem to refer to the seat of emotions. The searches carried out with the mind words were broadly similar to those for body words. As with the body searches, the queries with DEPs or IPs as their target used a list of mind words, listed in Appendix B. Some differences with the body searches are explained in Appendix C. One general difference that can be mentioned here is that with the body words, a list of affecting verbs was used to compare IPs and DEPs. With the mind words, however, DEPs did not seem to be limited to negatively affected possessors and a similar list would not have been useful. A note on the glossing of examples in this chapter may be helpful to the reader. Translation between languages is never entirely exact, and individual OE mind words covered a range of meanings that make the best translations into English uncertain in many cases. One fact of particular relevance to syntactic analysis is that while the word mind in Modern English is a count noun, some OE words that are most naturally translated as ‘mind’ in some contexts have a wide range of meanings and are usually treated as mass nouns. For example, in (6.1) the noun modsefa, a masculine weak

Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

122

       

noun, is clearly in the singular, although the genitive plural hiora makes it clear that we are talking about more than one person. The verb form meahte is also singular; the plural would be meahton. Thus, hiora modsefa calls for singular agreement by the verb, just as their intelligence does in Modern English. (6.1) gif hiora modsefa meahte weorðan staðolfæst gereaht if their mind:. might: become stable ordered ‘if their disposition might be made stable’ (cometboe,170.11.96.232) Differences from Modern English of this sort are relevant to the grammatical analysis of some of the examples discussed in this chapter, and matters are complicated by such facts as the identical forms used for neuter nouns like mod for nominative and accusative, both in the singular and the plural. In my glossing, mod is identified as nominative or accusative on the basis of what would be expected in a given sentence when there is no disambiguating modifier to make this clear, but I have generally not indicated number. The chapter is organized as follows. It first looks at mind words as direct argument possessa in section 6.2, then PObjs in section 6.3. Within each of these sections, the findings for poetic texts are first presented and discussed, followed by the discussion of the prose. Section 6.4 summarizes the conclusions of the study, comparing the findings for OE with body and mind words.

6.2 Direct arguments 6.2.1 Poetry In Chapter 4, we saw that DEPs were less frequent generally with body objects than with body subjects in the poetry, as in the prose. The same seems to be true with mind words, but to a greater extent. There are only a very small number of DEPs with mind object possessa in the poetry. The search for mind words playing the role of object in the York Poetry Corpus yielded only one example in which the possessor was expressed by a DEP: (6.2) meotod him þæt mod gestaþelað ruler him: the: mind: strengthens ‘the lord strengthens his mind’ (coexeter,146.106.204) (Seafarer 108)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

123

I found three more examples in poems not included in the York selections: (6.3) a. Hæfde Abrahame metod mancynnes, Had Abraham: ruler mankind: Lothes, breost geblissad Lot: breast: gladdened ‘The lord of mankind had gladdened the breast kinsman’

mæge kinsman:

of Abraham, Lot’s Genesis A 2923–5

b. healtum ond hrofum hyge blissode lame: . and leprous:. mind: gladdened ‘He gladdened the minds of the lame and lepers’ Andreas 578 c. us ofslæpendum sawle abrugdon us: sleeping: souls: drew.out: ‘they drew out our souls as we slept’

Andreas 865

The mind object possessa offer a striking contrast with the body possessa in the use of DEPs with beneficially affected possessors. With the body words, such possessors are nearly completely lacking, both in poetry and prose, as discussed in Chapter 4. In contrast, they represent a substantial proportion of the small number of DEP mind possessors in the poetry, as an inspection of the examples of (6.2) and (6.3) makes clear. In contrast to the paucity of DEPs with mind object possessa, IPs are common in the poetry. A search for IPs with mind object possessa in the York Poetry Corpus yielded eighty-one possible examples before any culling of examples that were not intended targets of the search. I have not considered it worthwhile to sort through the examples, especially because it is clear that most of them are non-targets. However, there are enough examples of IPs that the paucity of DEPs is not likely to be due to any general lack of examples in which mind words play the role of direct object. In my reading of the poetry selections not included in the York Poetry Corpus, I recorded all examples of IPs as well as DEPs, and found eleven examples of IPs of mind objects compared with the three DEPs of (6.3). As just noted, no attempt was made to count all the valid examples of IPs in the York Poetry Corpus, but they are clearly not uncommon. Example (6.4) presents four examples of IPs of mind object possessa from the York Poetry Corpus: (6.4) a. hungor innan slat merewerges mod. hunger within rent sea-weary: mind: ‘hunger rent the sea-weary one’s mind within’ (coexeter,143.8.120) (Seafarer 8–9)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

124

       

b. Ic to dryhtne min mod staþelige I to lord: my: mind: fix ‘I fix my mind to the lord’ (cocynew,119.221.1201) (Juliana 221–2) c. ah þinne modsefan staðola wið strangum. but your: mind: fortify against strong:. ‘but fortify your mind against hard men!’ (coandrea,36.1209.497) d. herd hige þinne harden mind: thy: ‘harden your mind’

(coandrea,36.1212.502)

The small number of examples of DEPs makes it difficult to say much about variation between IPs and DEPs, but we can note that the (b) and (c) examples of (6.4) are very similar in meaning to (6.2). One difference is that with the examples with IPs, the owner of the mind is presented as having control over it, while in the example with the DEP the situation is presented as one in which God is controlling the mind. The agentivity of the possessor is likely to favour an IP, although there are too few examples to make a strong case. When we turn our attention to mind subject possessa, we get a very different picture. Here, datives that are co-referential to a possessor are not at all uncommon. As a comparison of Table 6.1 and Table 6.2 shows, a small majority of these examples involve a copula rather than a lexical verb, as in (6.5): (6.5) a. ne wæs him bleað hyge not was him: cowardly mind: ‘his mind was not cowardly’ (coandrea,9.230.167) b. him wæs sefa geomor: him: was mind: sad ‘his spirit was sad’ (cobeowul,81.2631.2159) Beowulf 2632 Examples like those of (6.5) raise a problem of analysis: the mind word is in the nominative case, and could certainly be analysed as the subject, with the dative analysed as a DEP. However, such examples could possibly be given an analysis of extended existence, an ‘impersonal’ structure with no nominative subject, in which the nominative mind word is a predicate nominal, with meanings like ‘there was no cowardly mind to him’ and ‘there was a sad mind to him’, with the dative being part of a possible case-frame of the copula.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

125

In one of the poetic examples, we find a combination of a dative and a genitive: (6.6) þa him eorla mod ortrywe wearð then them: noblemen:. mind: despairing became: ‘then the noblemen’s minds were despairing’ (coexodus,95.154.130) In this ‘blended’ example, as in examples with simple IPs, there is no doubt that the mind word is the subject. Because of the uncertain status of many of the nominative mind words with copulas, Table 6.1, which compares IPs and datives with copulas, uses ‘nominative’ rather than ‘subject’. This table makes it clear that with copulas, such datives were by far more common than IPs for referring to the possessors of nominative mind words in the poetry of the York corpus. The remaining poetry selections do not offer enough examples to make a comparison meaningful.

Table 6.1 Datives and IPs of mind nominative possessa with copulas in OE poetry

Poetry York Poetry Non-York Poetry Total

IP

Dat

Dat + Poss

Ambiguous

6 2 8

11 1 12

1 0 1

1 0 1

There is some reason to believe that the copular datives were associated with particular adjectives. So, for example, five of my examples with a dative are variants of (6.5b), with the same words but some variation of word order. I have not found any IPs with the combination geomor ‘sad’ and sefa ‘mind, spirit’, although I did find an example with hyge that expresses essentially the same idea: (6.7) Forþon is min hyge geomor, Therefore is my mind sad ‘Therefore, my mind is sad’ (coexeter,210.17.1022) (The Wife’s Lament 17) The DOE entry for geomor 1.d provides a sense ‘sad heart’ for geomor sefa and lists him bið/was geomor sefa as a phrase meaning ‘he will be sad/he/they were

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

126

       

sad’. While the DEP is certainly the rule with geomor sefa, (6.7) shows that an IP was possible in expressing a similar sentiment.¹ Two more examples of IPs in the mind subject of copulas are given in (6.8): (6.8) a. Mæg þin mod wesan bliðe on breostum May thy spirit become blithe in breast ‘May your spirit become blithe in your breast’ Genesis B 750 b. and þin modsefa mara wurde and thy thought greater became: ‘and your thought would become greater’

Genesis B 501

In contrast to copulas, the DEP is in a slight minority when we restrict the mind subjects to lexical verbs, including passive participles, even if we counted this ambiguous example as dative: (6.9) ða wæs þære fæmnan ferð geblissad, then was the:/ / soul gladdened ‘then the soul of the lady was gladdened’ (cocynew,121.287.1252) (Juliana 287) The York Poetry Corpus parses þære fæmnan as genitive, but both words could be dative forms. In the poetry, where no determiner was necessary in a DEP, a dative possessum could be juxtaposed directly with possessum noun, although I agree that a genitive interpretation is the more likely. The overall number of examples of mind subjects is lower than the number of mind objects; my searches for IPs in mind subjects with lexical verbs did not get a huge number of hits. This meant it was possible to count them all. The results are presented in Table 6.2. I include passive participles as lexical verbs. Table 6.2 IPs and DEPs of mind subject possessa in OE poetry, lexical verbs

Poetry York Poetry Non-York Poetry Total

IP

DEP

Blended

Ambig

8 5 13

4 5 9

0 2 2

1 0 1

¹ Havers did not sort his examples by grammatical relation but did sort them by whether the possessor was a possessive pronoun or a genitive noun. He comments that while a genitive (i.e. an IP) was dominant as an expression of the possessor in the semantic sphere of his category II (Seele, roughly equivalent to my mind words) when the possessor was a noun, he found no clear examples where a possessive pronoun replaces a sympathetic dative, although he thought that (6.14b) might be such an example (1911: 276). The examples of (6.8) show that some pronominal IPs are found in a wider poetic corpus than that investigated by Havers, at least with copulas.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

127

In eight of the nine examples of DEPs with mind subjects in Table 6.2, the subject in question is the subject of a passive verb, rather than an inherently intransitive verb with a non-agentive subject. The sole active example is given in (6.10): (6.10) Ne gemealt him se modsefa not melted him: the: spirit: ‘his spirit did not flag’ (cobeowul,81.2628.2153) Beowulf 2628 Two examples of the more common sentences with passives are given in (6.11): (6.11) a. Þa þam halgan wearð after gryrewhile gast then the: holy: became after terror.while spirit geblissod gladdened ‘then the saint’s spirit was gladdened after the time of terror’ Andreas 466 b. þær manegum wearð mod onlihted there many: became mind illuminated ‘there the minds of many became illuminated’ (cocynew,52.50.40) (Fates of the Apostles 52) In all the examples with DEPs, the possessor of the mind subject is clearly affected, but the effect can be a very positive one, as in (6.11a). In addition to the simple DEPs, the investigation turned up two examples with what looks like a ‘blended’ construction. These are presented in (6.12). (6.12) a. Ne læt þu þe þin mod asealcan Not let thou thee:/ thy spirit slacken ‘Do not let your spirit grow idle’ Genesis A 2168 b. oð þam þegne ongan his hyge hweorfan until the: thane: began his mind turn ‘until the thane’s mind began to change’ Genesis B 705-6 Although these two examples are entered as blended in Table 6.2, neither is straightforward. For one thing, complications arise from the fact that they both involve two verbs, and each of them presents at least one other problem. (6.12a), a blended analysis, treats þin mod as the subject of the infinitive asealcan, with the dative þe also referring to the possessor of mod in this example. However, other analyses are possible, including the treatment of þe as an inherently reflexive object of lætan, and the case of this pronoun is ambiguous between dative and accusative. In (6.12b), there is a possibility that

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

128

       

the auxiliary verb ongan might be responsible for the dative. It seems probable, however, that the dative is there because it is seen as a misfortune that happened to ‘the thane’; the passage is talking about a change of mind that happened because of someone else’s arguments, and furthermore ended in disaster. In the context of this example, Adam’s mind changes because of Eve’s representations, ending in the damnation of both of them. The verb hweorfan ‘turn, change’ of example (6.12b) raises the question of which intransitive verbs should be counted in a comparison of DEPs and IPs. The other two examples collected of this verb with a mind subject, presented in (6.13), have an IP: (6.13) a. Forþon nu min hyge hweorfeð ofer hreþerlocan, because now my mind departs beyond breast ‘because now my mind departs beyond my breast (i.e. escapes the confines of the body)’ (coexeter,145.58.160) (Seafarer 58) b. min modsefa mid mereflode ofer hwæles eþel my mind with sea-flood over whale’s domain hweorfeð wide journeys widely ‘my mind roams widely with the ocean, over the domain of the whales’ (coexeter,145.58.161) (Seafarer 59) These examples with hweorfan are included in my comparative counts of IPs and DEPs not only because of the dative in example (6.12b), but verbs with similar meanings of turning and changing are found in the small number of DEPs with mind subjects found in the larger prose corpus, as discussed in section 6.2.2. Ideally, if we are counting IPs to compare their frequency with DEPs, we want to exclude examples in which the subject is agentive, and with the mind words it is sometimes impossible to tell whether a writer thought of the mind or spirit as actively doing something or undergoing a process. For example, Havers (1911) only considers that (6.14b) is possibly as example of a possessive being used in a situation where a sympathetic dative would be a possibility, and his judgement that this is uncertain reflects an understanding that DEPs were not likely with agentive intransitive subjects.² The same objection

² The Dictionary of Old English indicates that only eight occurrences are found of the verb ahlyhhan in OE. When used with a person as subject, it could mean ‘to laugh’ or ‘to laugh at, to scorn’. The DOE entry indicates that it is found four times, all in poetry, with subjects referring to the heart, mind, or spirit with the meaning ‘exult’ (including the example in (6.14b)), and the fact that these examples all involve an IP is consistent with the view that a DEP would not be possible here. However, this number of examples is too small to declare that a dative would have been impossible.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

129

could be raised against (6.14a), an example that Havers does not discuss. For this reason, I have excluded these and some similar examples from the count of IPs in Table 6.2. (6.14) a. ac min hyge blissað but my mind rejoices ‘but my mind rejoices’ b. þa his mod ahlog; then his mind exulted ‘then his mind exulted’

Andreas 634

(cobeowul,24.730.616) Beowulf 730

The purpose of including the comparison of the numbers of IPs and DEPs in the poetry was to show that IPs were well established for the expression of possession of mind words in contexts in which cross-linguistic tendencies would lead us to expect that a DEP would not be impossible in a language that has the construction. The comparison is limited to situations where a DEP is most likely to have been a possibility, and this limitation involves a certain amount of uncertainly over which IPs to exclude. However, the exact number of examples of each type is not really important in making the point that DEPs were clearly in variation with IPs even in the poetry, both with inherently intransitive verbs with non-agentive subjects and in passive sentences.³ Some examples of IPs in contexts where a DEP would surely have been an appropriate alternative are given in (6.15): (6.15) a. Þa wæs modsefa myclum geblissod haliges on hreðre then was spirit: greatly gladdened holy: in breast ‘Then the holy man’s spirit was greatly gladdened in his breast’ Andreas 892 b. Ne læt þu þin ferhð wesan sorgum asæled not let you thy: spirit: be sorrows: fettered ‘do not let your spirit be fettered with sorrows’ Genesis A 2196 c. for hwan modsefa min ne gesweorce, for which mind my not darkens ‘for which my mind does not darken’ (coexeter,135.58.50) (Wanderer 59)

³ Example (6.15b) can be compared with (6.12a). It slightly weakens one of the possible alternative analyses of that example; if lætan could be used with an inherent reflexive, this was optional. However, such reflexives generally were optional.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

130

       

As with the DEPs, an IP could be used to describe a beneficial effect (6.15a) or a negative one (6.15b). This variation is an important point, because some recent publications by proponents of the Celtic Hypothesis have given the impression that DEPs were generally the normal expression of inalienable possession in the more ‘Germanic’ of the OE writings, with IPs being uncommon before the EME period and possibly due to Latin influence.

6.2.2 Prose We saw that with poetry, examples of DEPs with mind object possessa are infrequent. In the prose texts, I found no clear examples at all. This result is particularly solid because of the large size of the prose corpus. The closest I found to an indisputable example was (6.16), which turned up in my search for body object possessa: (6.16) & ablent ðæs modes eagan mid ðære costunga and blinds the: mind: eye: with the temptation ðæm folce the: people: ‘and blinds the mind’s eye of the people with the temptation’ (cocura,CP:18.129.13.880) I excluded this example from my examples of DEPs with body objects on the basis that the ‘eye’ in question is metaphorical. However, the fact that it would be an isolated example of a DEP with an object mind concept suggests that internal vision was treated more like a body part than as a mind concept. It is worth noting that the context of this example is a comparison of the mind’s eye with the blinding of the physical eye that occurs with dust. There is another example that might be treated as having a DEP with a mind word. This example was presented as example (2.4) and is repeated as (6.17) for convenience: (6.17) Ic ætbrede him ða stænenan heortan. I remove them:. the: stone: hearts: ‘I will remove their stone hearts’ (cocathom2,+ACHom_II,_12.1:117.250.2546) I have treated this example as using heorte as a mind word, but the line between the physical organ and its metaphorical extension as the seat of emotions is a blurry one. The biblical passage quoted here refers to taking

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

131

away hearts of stone and replacing them with hearts of flesh. Whether we treat this as a body or mind example, an important point is that ætbregdan was a ditransitive verb, with dative deprivee and an accusative of the deprived thing. The reasoning behind not treating the dative objects of ditransitive verbs as DEPs and regarding any possessive relationship as pragmatic rather than syntactic was discussed in section 2.2. In contrast to DEPs, IPs are easy to find with mind object possessa, both in early and late texts: (6.18) a. ðonne he ure heortan onlieht when he our hearts illuminates ‘when he illuminates our hearts’ (cocura,CP:48.369.14.2492) b. onlihte ure mod his godcundan leohte. illuminate: our mind his divine light ‘may his divine light illuminate our minds’ (cocathom2,+ACHom_II,_17:168.221.3730) With subject possessa, the contrast between prose and poetry is greater than it is with objects. DEPs are common with mind subject possessa in the poetry, as we saw, but most prose texts yielded no examples, either with lexical verbs or with copulas. It is not worthwhile to present large tables that consist mainly of zeros. Instead, since DEPs of this type are so infrequent, I present all my prose examples. In the EWS texts, I found no clear examples. As with objects, I found one example of a DEP with ‘mind’s eye’ in the subject role: (6.19) Ðonne ðam lareowum aðistriað ðæs modes eagan when the: teacher: darken: the: mind: eyes: ‘when the mind’s eyes of the teachers darken’ (cocura,CP:1.29.15.126) As with the object example, this example, being metaphorical, is excluded from the count of DEPs with body parts. As well as being a clear example of a DEP (whether it is counted with the body or mind words), it offers a good example of how free from Latin syntax the close translation of the Cura Pastoralis is. The Latin has no determiner, and no possessor mentioned, in an ablative absolute construction obscuratis ergo oculis ‘truly with darkened eyes’. Clearly, this translator who stuck close to the Latin in terms of content did not regard the DEP as inappropriate for translations from Latin, even though he did not use it very frequently.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

132

       

The contrast between the lack of any certain DEPs with mind subjects and IPs is very clear; I found thirty-six IPs of mind subjects in the EWS texts (verbal and non-verbal predicates combined) of which (6.20) is one: (6.20) & ðonne blissiað eowre heortan, and then rejoice your hearts ‘And then your heart will rejoice’ (cocura,CP:27.187.20.1253) I found no examples of DEPs with mind subject possessa in my Other Early texts. I found one example in the 9thC(OE) type: (6.21) þeah him sie gemynd oncyrred he biþ hal though him: be: mind: turned he is whole ‘Although his mind will be deranged, he will be hale’ (colaece,Lch_II_[3]:1.1.4.3506) The search of the selected General OE texts found two examples, both with verbs meaning ‘turn’: (6.22) a. Ða oncierde him seo gehygd to deofolgylde then turned him: the: mind: to idolatry ‘then his mind turned to idolatry’ (comart3,Mart_5_[Kotzor]:Ju18,A.9.993) b. oþ þæt him seo heorte eft to Criste gecerde, until that him: the: heart: again to Christ turned ‘until his heart turned again to Christ’ (comart3,Mart_5_[Kotzor]:Ju18,A.12.995) The searches of the large examples of DEPs of mind subjects in the large LWS and LWS(Late) corpus turned up no examples. It is perhaps not an accident that in all the examples of DEPs with mind words, the word is a body word extended metaphorically to the emotions or mind. Body words occurred more freely with DEPs than mind words in prose, and a body word extended to the mind might have been easier to use with a DEP. The fact that all the examples involve verbs with a meaning of turning further suggests a lexical restriction of the DEP with mind words. The near complete lack of DEPs with mind words in the prose cannot be attributed to any data gap caused by a dearth of these words with expressed possessors. I have already demonstrated this with the figure for IPs in the EWS texts. The bulk of OE prose is religious, giving a rich source of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.2  

133

sentences about people’s minds and spirits, and examples of IPs are easy to find in other texts also, including active and passive lexical verbs and adjectival predicates: (6.23) a. ond ðæs wisan monnes mod bið suiðe emn and the: wise: man: mind is very level ‘and the wise man’s mind is very level’ (cocura,CP:42.306.11.2050) b. Behealdað eow ðæt eowre heortan ne sin gehefegode beware you that your hearts not are: weighted mid oferæte with overeating ‘Beware that your hearts not be weighed down with overeating’ (cocura,CP:43.317.8.2123) c. þara manna heortan wæron gewended in the:. men: . hearts were turned into wilddeora reðnesse wild.animals:. savagery ‘the men’s hearts were turned into the savageness of wild animals’ (cogregdC,GDPref_and_3_[C]:11.195.4.2501) Counting all the valid IP examples to compare with the small number of DEPs with mind words in the prose would not illuminate matters further. Note finally that the total lack of examples of DEPs with the subjects of adjectival predicates makes both a sharp difference with the poetry, where such DEPs are common, and an important difference with body words, where such DEPs are not very common but can be found. It seems very probable that we are dealing with a diachronic development here, with poetry representing an earlier situation. Assuming that this is so, the loss of DEPs with the mind subjects of adjectival predicates is a reduction in the range of the DEP in an early period of OE.

6.2.3 Summary: mind direct arguments Datives that referred to the possessors of mind words were common only in the poetic texts, and only when the mind word played the role of subject. A small majority of the relevant examples involve a copula, and the status of the mind word as a subject or a predicate nominal is debatable. In copular

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

134

       

constructions, these datives are more common than IPs. DEPs with lexical verbs are not uncommon in the poetry, but they are in a minority compared with IPs, and there is clear variation between DEPs and IPs in similar contexts. The prose is strikingly different here, especially in the lack of DEPs with copulas. It is interesting to note that copular constructions continue to be used in expressions of extended existence in the prose late in the OE period: (6.24) swilce him eallum wære an heorte & an sawul as them: all: were: one heart and one soul ‘as though they had one heart and one soul’ (cocathom1,ÆCHom_I,_22:363.239.4512) Example (6.24), in which the verb habban would be an alternative to the copular construction, could be treated as a type of predicative possession, but the important point is that it expresses the existence of something—a unity of heart and soul in this example. This contrasts with the poetic him wæs geomor sefa ‘his mind was sad’ type, in which the existence of a mind is assumed, and the sentence is about the state of this mind. In the geomor sefa type, an IP is an alternative to the dative, since the possession is attributive rather than predicative. With mind objects, we find only a few examples of DEPs in the poetry, and no completely convincing examples in the prose. This is not due to any overall lack of examples of mind objects, since examples with IPs of such objects are easy to find in both the poetry and some prose texts. Although we of course cannot be certain that any written text represented the syntax found in everyday speech, it seems a reasonable hypothesis that the difference between the poetic and prose texts reflects a general reduction in the range of the dative case referring to the possessor of the mind words selected for study here.

6.3 PObjs As with the body words, the search for mind PObjs used a limited list of prepositions. This list is not quite the same as with the body words and can be found in Appendix B. Unlike with body words, there was no need to use a limited list of mind words with the searches for PObjs since the number of ‘hits’ in searches for DEPs was small enough that culling the examples in which a dative was not a DEP was not onerous. With the body words, the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

135

number PObjs in sentences with a dative, usually not a target DEP, is high, and I limited the list of texts searched in some text categories for PObjs. However, with the mind words, the number of PObjs in sentences with datives is considerably smaller, and all texts have been included in the searches for them, only excluding the H version of Gregory’s Dialogues. The subject matter of the Dialogues results in a large number of mind words, and many of the hits in both versions are non-target examples, such as examples with no expressed possessor, making the culling of examples timeconsuming and unrewarding. The target examples of the H version are usually repetitions of ones in the C version, meaning no significant data is lost by not using the H version. The investigation of PObjs is even more complicated with mind words than for body words because of the relatively large number of examples with copulas. With the body words, PObjs with copulas were infrequent, but this is not the case with mind words. With lexical verbs, I have excluded examples in which there is a ‘shared dative’, that is, the verb takes a dative in its caseframe and this dative happens to be coreferential with the possessor of a PObj. With copulas, this is not possible because it is generally impossible to say with confidence that a dative in a sentence has the sole function of expressing possession. We face similar problems here to those presented by mind words in the nominative case, discussed in section 6.2.1. With PObjs, however, we have a substantial number of sentences in which there is clearly no nominative subject and it is clear that the dative that is coreferential to the possessor is due to the copula. As discussed in Chapter 2, copulas were associated with datives in various constructions. A mind PObj could frequently be added to such a construction; for example, alongside (6.25a) we find (6.25b): (6.25) a. þeah he him leof wære though he them: dear were: ‘though he was dear to them’ (cobeowul,9.202.162) Beowulf 203 b. Criste wæron begen leofe on mode. Christ: were both dear in mind ‘they were both dear to Christ in his mind’ (coandrea,31.1016.323) In Modern English, the dative would be replaced with a prepositional phrase, clearly a complement of the adjective leof(e), but in OE the status of the dative is not so clear. Similarly, with constructions traditionally considered impersonal, as discussed in section 2.8, we find both (6.26a),

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

136

       

with only a dative to refer to the possessor, and (6.26b), with an added IP in the PObj: (6.26) a. Wearð me on hige leohte utan and innan became me: in mind light without and within ‘Things became light to me in my mind, inside and out’ Genesis B 676 b. Þe weorð on þinum breostum rum thee: becomes in thy breast wide ‘your mind will widen’ Genesis B 519 With nouns in expressions of ‘extended existence’ we get both (6.27a) and (6.27b): (6.27) a. Me is willa to ðam mycel on mode, me: is will to that great in mind ‘The desire it so great in my mind’ (codream,65.129.164) b. gif him aht wæs on his mode if him: something was on his mind ‘if something was on his mind’ (cogregdC,GDPref_and_4_[C]:19.289.5.4274) How much the treatment of the dative matters depends on the view of syntax that one adopts. In the traditional treatment of all these datives as indirect objects that would be categorized on semantic grounds as different types of ‘datives of interest’, it doesn’t really matter for the syntax. In a framework in which a DEP is treated as originating within the NP containing its possessum and is moved to the clause level, it does matter, since DEPs are syntactically different from datives generated at clause level as complements of copulas (or perhaps as the complement to an adjective or noun). I will assume that the dative is part of the case-frame of the copula, along with the adjective phrase containing the prepositional phrase. I suggest that when no IP is added, we have an implicit possessor of the PObj. As with verbs, implicit possessors of PObjs, e.g. (6.28), are in common in all texts: (6.28) Þa wearð se brema on mode bliðe then became the famous in mind blithe ‘then the famous one became blithe in his spirit’

Judith 57–8

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

137

Implicit possessors of mind PObjs reduce the number of examples with a dative that refers to the possessor, leaving us with too few examples to make attempts to sort them into types and make generalizations about when a genitive is likely to be added to the dative and when it is never added. It seems likely that some phrases such as leof(e) on mode were idioms, similar to young at heart, and did not freely admit a genitive. Difficulties in analysis are further complicated by the fact that some words, e.g. sar, could be either nouns or adjectives, and scholars disagree on what word class to assign weorce to—it is historically the dative case of a noun, used adverbially. In example (6.42a) in section 6.3.2, analysis is complicated by the fact that the etymologically dative weorce is conjoined with a noun in the nominative case. There are examples like (6.33a) in section 6.3.2 where the non-verbal predicate is an adverb, rume ‘widely’ in that example, alongside the adjective rum of (6.26b). It is therefore not at all clear that these datives, at least all of them, should be treated as external possessors. Nevertheless, it is worth looking at the co-occurrence of datives and possessives in these constructions, as well as the use of simple IPs, where the possessive relationship is very clear, with copulas. We shall see that there is a sharp difference between OE poetry and prose in the use of datives with copulas and mind PObjs. In what follows, I will not refer to the copular constructions with datives as DEPs, but will instead refer simply to datives and combinations of dative and genitive.

6.3.1 Poetry With the body words, DEPs are more common with PObjs than with direct arguments of lexical verbs all through the OE period, and the same is true with mind words. Although I have not compared IPs and DEPs with PObjs in the prose texts, I have done so for the poetry. The results for sentences with lexical verbs are presented in Table 6.3. Table 6.3 IPs and DEPs with selected mind PObjs of lexical verbs in OE poetry Text Category

IP

DEP

Blended

Ambiguous

Poetry York Poetry Non-York Poetry Total

6 9 15

16 6 22

0 0 0

0 1 1

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

138

       

It is clear that with the selected prepositions, DEPs were by no means uncommon with PObjs. The different results given for the majority construction in the York selections, where DEPs are heavily favoured, and the other poetry, where IPs predominate, are a reminder of how a larger corpus can change a picture significantly. The difference seems most likely to be due to a higher likelihood of a DEP with a given combination of verb and preposition and a coincidental higher frequency of such combinations in the selections in the electronic corpus, but I will not pursue this matter here. The important point is that IPs were by no means unusual even in the poetry. Although DEPs were frequent with mind PObjs, an examination of individual examples reveals that the combinations they occur in tend to be expressions that are listed as idioms in dictionaries. It should be noted that many examples might be counted as having a DEP or might be analysed differently. For example, consider (6.29): (6.29) ða gen ic Herode in hyge bisweop þæt then yet I Herod: in mind prevailed that ‘then, furthermore, I prevailed upon Herod’s mind that . . . ’ (cocynew,121.293.1259) (Julian 293–4) Since Herod is the owner of the hyge ‘mind’ in question, the dative Herode looks like a DEP, and I have counted it as such. But the DOE lists beswapan in hyge ‘prevail upon someone’s mind’ with a dative as an idiom. A possible alternative analysis is to treat beswapan in hyge as a unit that takes a dative object. Similarly, the same dictionary has a separate sense ‘to occur to (someone dat.)’ listed for beyrnan on mod(e): (6.30) a. Him on mod bearn þæt healreced hatan wolde him: in mind ran.into that hall.building order would ‘it occurred to him to order the building of a hall’ (cobeowul,5.67.55) Beowulf 67–8 Again, I have counted this example as a DEP, but beyrnan on mod(e) could alternatively be treated as a complex verb requiring a dative object. DEPs are not in variation with IPs with this idiom. Although DEPs appear to be lexically specified in constructions like the ones just discussed, IPs are not difficult to find with mind words as PObjs with a verbal predicate:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

(6.31) sibbe sawað on sefan manna peace sow: in hearts man:. ‘sow peace in the hearts of men’

139

(cochrist,16.485.320)

In contrast with the verbal predicates, datives seem to have been the rule with mind PObjs with copulas. Examples of DEPs with these non-verbal predicates were given in the beginning of section 6.3. The search of the York corpus found only two examples of IPs with a copula: (6.32) a. ðu eart dohtor min seo dyreste ond seo sweteste you are daughter my the dearest and the sweetest in sefan minum in mind my ‘you are my dearest daughter, the sweetest in my mind’ (cocynew,115.92.1113) (Juliana 93–4) b. gram wearð him se goda on his mode.4 angry became him: the: good: in his mind ‘the good one (i.e. God) had become angry with him in his mind’ Genesis B 302 In addition to these unambiguous IPs, I have come across three examples in which an ambiguous form could be either a dative or a genitive: (6.33) a. Þa wearð hyre rume on mode5 then became her:/ widely in mind ‘Then her spirit expanded’ b. Þa wæs Sarran sar in mode then was Sarah: / sore in mind ‘then Sarah was sorrowful in her heart’

Judith 97

Genesis A 2216

c. Hire wæs godes egsa mara in gemyndum her: / was God: fear greater in mind ‘The fear of God was greater in her mind’ (cocynew,114.38.1073) (Juliana 35–6) ⁴ The Dictionary of Old English lemma gram includes beon/weorþan gram ‘to be/become angry (with someone dat., wiþ and acc.)’ in its sense ‘angry, wrathful, hostile, fierce’. The dative in this example refers to Satan. ⁵ As mentioned above, rume is an adverb, while rum, found in (6.26b), is an adjective. Examples like (6.33a) are difficult to translate well into idiomatic Modern English, where adverbs can no longer serve as predicates, as they could in OE.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

140

       

Since possessive phrases could be discontinuous in OE, particularly in the poetry, an analysis of these examples as discontinuous IPs is not completely impossible. I believe the ambiguous forms in these examples are all probably to be taken as dative, however, because as far as I am aware, a possessive was not normally discontinuous from its possessum when that possessum was the object of a preposition. These examples of IPs with a copula are in a minority in the poetry, and because there are so few, no table is presented comparing IPs and datives with copulas for PObjs. Table 6.4 compares the figures for copulas with a PObj having only a dative to refer to the possessor versus ones with both a dative and an added IP. In the general discussion of copular PObj types at the beginning of this section, both types of examples were presented in (6.26) and (6.27). Table 6.4 shows that the (a) sentences in these examples, i.e. those with a dative alone, were more common than the (b) type, in which there is also a genitive. Nevertheless, examples of the (b) type are in substantial minority: Table 6.4 Copula dative constructions with selected mind PObjs in OE poetry

Poetry York Poetry Non-York Total

Dative alone

Added genitive

11 6 17

2 3 5

In sum, DEPs were not uncommon as possessors of PObjs in the poetry with verbal predicates. Implicit possessors were always a possibility with PObjs, and the datives used with copulas in various constructions could serve to identify the possessor of the mind, but the possessor might be made explicit by the use of a genitive. Copular constructions with IPs were not uncommon, in contrast with the uncommon ‘blended’ type with lexical verbs. Section 6.3.2 describes a very different situation with the copular constructions in the prose texts.

6.3.2 Prose As with body words, DEPs represented the possessor of PObj mind words more frequently than was true of the direct arguments. Table 6.5 presents the results for DEPs with the mind objects of the selected prepositions with lexical verbs in the texts selected for the collection of PObj examples.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

141

Table 6.5 DEPs of mind PObjs with lexical verbs in OE prose

Other Early EWS 9thC(OE)* General OE LWS LWS(Late) Prose totals

DEP

Blended

Ambiguous

5 1 14 8 8 2 38

0 0 0 0 0 0 0

0 0 0 0 0 0 0

* The searches excluded cogregdH, as discussed in the introduction to section 6.3.

As discussed in section 6.3.1, DEPs may be associated with particular combinations of verbs and prepositions in the poetry. This sort of association is definitely the case in the prose by the Late West Saxon period, as discussed shortly. It is therefore not worthwhile to compare the number of DEPs and IPs in the prose, but we can merely note here that the low number of DEPs in this large corpus is not due to a data gap, since IPs are not difficult to find in expressions in which a DEP would be an alternative. I offer a few examples of such IPs from different text types in (6.34): (6.34) a. ðonne ahefð he hine on his mode then raises he him: in his mind ‘then he raises himself up in his mind’ (cocura,CP:4.39.6.200) b. þæt he Godes ege hæbbe symle on his gemynde that he God: fear have: always in his mind ‘that he always have the fear of God in his mind’ (colaw1cn,LawICn:25.137) c. Þa getweode heora an on his mode then doubted 3. one in his mind ‘Then one of them doubted in his mind’ (comart3,Mart_5_[Kotzor]:Ma9,A.14.333) d. for ðan ðe he heold ða digelan ðrowunge on his for that: that he held the secret suffering in his mode mind ‘because he held the secret suffering in his mind’ (cocathom2,+ACHom_II,_42:315.146.7123)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

142

       

It is more illuminating to examine individual examples of DEPs in this case than simply to compare the number of them with the number of IPs. An inspection of the examples of DEPs suggests the conclusion that although DEPs continue to be found with PObjs all though the OE period, by the Late West Saxon period they appear to be pretty much restricted to fixed expressions. So, for example, of the eight instances of DEPs with PObjs in sentences with lexical verb predicates in the selected LWS texts, three are variants of beyrnan dat on mod(e), e.g. (6.35): (6.35) ac us bearn þis on mod. but us: ran this in mind ‘but this occurred to us’ (coaelive,+ALS[Peter’s_Chair]:232.2426) As discussed above, this was an idiom listed in the DOE, also found in the poetry. I have not found any examples of IPs in LWS texts with this combination of verb and PP, e.g. the hypothetical bearn on his mod, and the same is true of my other texts. The overall number of examples is small, so any conclusion is tentative, but it seems probable that no variation existed. A fourth DEP example in the LWS texts uses irnan without the be- prefix in a similar meaning: (6.36) swa hwylc idel swa him to geþance yrnð so which frivolity as them: to thought runs ‘whatever frivolity occurs to them’ (cobenrul,BenR:1.9.15.152) Three of the LWS DEPs involve the combination becuman on mod ‘come to mind’. The single example that did not have a meaning of ‘coming to mind’ is presented in (6.37): (6.37) þe hyra leahtrum geþeafige and him on gewil gange, that their sins allow: and them: on will go: ‘who will go along with their desire’ (cobenrul,BenR:64.119.5.1158) With becuman on mode at least, there is some variation with IPs: (6.38) Ne unlust on hire mod ne becom. nor bad.desire in her mind not came ‘Nor did bad desire come into her mind’ (cocathom2,+ACHom_II,_1:5.81.64)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

143

Of course, with a broader selection of prepositions, further examples of DEPs can be found. For example, the preposition of is not on the list of prepositions used in the searches because the initial searches, which included this preposition, threw up too many non-target examples, but this preposition can be found in some DEPs of mind PObjs in LWS: (6.39) þæt hit ne gange eow of gemynde that it not go: you: from mind ‘lest you forget about it’ (cocathom1,+ACHom_I,_15:299.3.2739) Gan x-dat of gemynde, ‘to go out someone-dat of the memory’, is of course the sort of expression that could easily be lexically encoded, although I have not found it listed as an idiom in the DOE. Further examples can also be found in texts not included in the systematic searches; for example, I noticed (6.40) in a General OE text not included in my texts selected for my PObj search: (6.40) Ac þa ðing þe me nu in gemynd cumað ærest but those things that me: now in mind come first ‘but those things that occur to me first’ (coalex,Alex:4.1.14) This is the only example of a DEP with the selected prepositions and mind words that I found in this text, and it adds to the evidence for the use of DEPs with expressions for coming into a person’s mind. It appears that DEPs with mind PObjs, with the selected prepositions at least, were pretty much limited to commonly recurring expressions even in the early prose. Of the five DEP examples in the Other Early category listed in Table 6.5, four of them are variants of coming in or to someone’s mind. The fact that the 9thC (OE) category has so many examples is due mainly to the repeated use of sentences like (6.41): (6.41) þonne seo untrumnes eft gehwearf me to gemyndum when the illness later turned me: to mind ‘when the illness later returned to my mind/when I later remembered the illness’ (cogregdC,GDPref_and_3_[C]:33.244.12.3453) The DOE sense B.7.d.i. for gehweorfan lists to gemyndum gehweorfan as meaning ‘to return to someone’s (dat.) thoughts/memory’, i.e. ‘to be recalled by someone’. Judging from the DOE’s discussion, this appears to be a Latininspired construction. As mentioned before, the Dialogues are on the more Latinate end of the translation spectrum.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

144

       

To sum up the use of DEPs with mind PObjs of verbal predicates, it seems that they were considerably more restricted in the prose than in the poetry, and if we assume that the poetry reflects earlier usage, this represents a diachronic change. It also represents a difference with the body words. We saw that the use of DEPs with those PObjs was particularly associated with certain verbs and prepositions, but still reasonably productive. Two differences with the poetry emerge when we look at the use of mind PObjs with copulas. The first difference is that in the poetry PObjs combined with copular constructions of various types using a dative were common. In the prose we continue to find some examples all through the OE period, but they are not common, with the selected prepositions at least. The number is smaller than for the poetry despite the much larger size of the prose corpus. The second difference is that the proportion of examples that combine a possessive with the dative is much greater in the prose than in the poetry. Because the majority of the prose texts offer no examples at all involving a mind PObj with a copular dative in the prose texts, Table 6.6 presents the aggregated figures for the text types rather than for individual texts. Table 6.6 Dative possessors of selected mind PObjs in copular constructions in OE prose Text type Other Early EWS 9thC(OE) General OE LWS LWS(Late) Total Prose

Dative alone

Dative + Genitive

0 0 1 4 1 1 7

0 2 1 4 1 0 8

Some examples using only a dative are presented in (6.42), while the examples of (6.43) illustrate the addition of a genitive: (6.42) a. & him wæs on mode myccle weorce and them: was in mind great: anguish: & mycel tweo and great doubt: ‘and they were in great anguish of mind and considerable doubt’6 (coblick,LS_25_[MichaelMor[BlHom_17]]:205.145.2619) ⁶ This is Kelly’s (2003) translation.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

145

b. for þam micclan ogan þe him on mode wæs. for the great dread that him: in mind was ‘for the great dread that was in his mind’ (cosevensl,LS_34_[SevenSleepers]:555.430) c. Eac wære þam earman leohtre on mode Also were: the: poor: lighter in mind ‘Also the poor man’s mind would be lighter’ (cocathom1,+ACHom_I,_23:366.51.4569) d. Wynsum us byð on mode, þæt we welwyllende beon. pleasant us is in mind that we benevolent be ‘it is pleasant to our minds that we be benevolent’ (colwgeat,+ALet_6_[Wulfgeat]:267.115) (6–43) a. Þa wæs him micel langung & sorh on heora then was them: great longing and sorrow in their heortan hearts ‘then there was so much longing and sorrow in their hearts’ (coblick,HomS_47_[BlHom_12]:135.82.1645) b. Wæs him ægweðer þæm eadigan were was him: both the: blessed: man: ge seo Godes lufu toðæs hat & toðæs both the: God: love: so hot and so beorht on his heortan . . . bright in his heart ‘The love of God was both so hot and so bright in his, the blessed man’s heart . . . ’ (coblick,LS_17.1_[MartinMor[BlHom_17]]:225.271.2905) c. Nis me nænig leoht ne nænigo byldo on not.is me: no light nor no courage in minum mode my mind ‘there is no light or courage in my mind’ (coverhom,HomU_7_[ScraggVerc_22]:62.2867) Example (6.27b) at the beginning of section 6.3. is a further prose example of a possessive combined with a dative. With the copular constructions, the reduction in the frequency of examples of datives with mind PObjs is due to a combination of factors, some of which

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

146

       

have nothing to do with the replacement of DEPs by IPs per se, but rather have to do with the use of datives with copulas. So, for example, in the poetry we find several examples in which a mind PObj is appended to an expression of ‘extended existence’ using a copular dative, as in (6.27). In the prose, this type of copular dative construction was still used with some frequency even at a late stage, as in (6.44): (6.44) ac ðe ne bið nan ærist to ðam ecan life but thee: not is no resurrection to the eternal life ‘but there will be no resurrection to the eternal life for you’ (coaelive,+ALS_[Maccabees]:142.4907) Although these expressions were still current in late OE, and indeed the expression ‘woe is me!’ extends into Modern English, the prose differs from the poetry in that the addition of a PP with a mind was unusual in the prose, although examples are not completely absent. Most of my prose examples come from General OE texts, but (6.45) is an unusual late OE example: (6.45) and him bið þonne wa on his awyrigedum mode and him: is then woe: in his accursed mind ‘and then it will be woe to him in his cursed mind’ (coaelive,+ALS_[Auguries]:166.3600) In the poetry, datives with non-verbal predicates were frequently used in descriptions of a person’s state of mind, but while these also occur in the late prose, as in (6.42c), in the prose generally we tend to find personal constructions in which the person is the subject and an IP is used in the PObj: (6.46) a. Þa wurdon hiora wif swa sarige on hiora mode then became their wives so sorrowful in their mind ‘then their wives became so sorrowful in their minds’ (coorosiu,Or_1:10.29.19.572) b. for ðan þy ðæt dioful bið on eowrum heortum for that: that: the devil is in your hearts ‘because the devil is in your heart’ (coverhom,HomU_7_[ScraggVerc_22]:108.2910) c. he wearð sarig ðearle on his mode, he became sorrowful greatly in his mind ‘he became very sorrowful in his mind’ (cosevensl,LS_34_[SevenSleepers]:337.254)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.3 

147

The net result of a combination of developments is an overall reduction in the number of datives that could be construed as the possessor of a mind PObj. At first glance, the increase in the use of a combination of a dative and a genitive in the copular constructions, as in the examples of (6.43), gives the impression of an increase in ‘blended’ constructions in the OE period. The similarity of these examples to the blended construction found sporadically with the body words through the OE period is only superficial, however. With the blended construction, the dative is not part of the ordinary case-frame of the verb. In contrast, the dative was an integral part of these copular constructions. In that regard, they are similar to ditransitive verbs, which are not included in this study of external possession, in having a ‘shared’ dative, that is, a dative that was selected by the predicate and could also be construed as the possessor of some noun in the clause. I have suggested that in this situation we have an implicit possessor, which could be made explicit by the addition of a genitive. I suggest a similar analysis for the combinations of a dative and a genitive with mind PObjs. The increase in the use of a possessive to augment a dative in the copular constructions can be seen as a decrease in the use of implicit possessors in PObjs in a period where the copular constructions using a dative still had some currency.

6.3.3 Conclusions: mind PObjs With the restricted list of prepositions used in this investigation at least, DEPs predominate overall in the poetry, but with verbal predicates they tend to occur in particular idioms, and IPs are common. The same preference for IPs with verbal predicates outside of particular idioms is found in the prose of all periods, but the common use of these idioms means that examples with DEPs continue to be found in prose throughout the OE period. Datives referring to a possessor of a mind PObj also predominate in copular constructions in the poetry, with IPs in a distinct minority. However, in these copular constructions involving a dative, a substantial minority has a genitive as well as the dative. In the prose, there are only a small number of examples of datives referring to the possessor of a mind PObj in any period, and in these examples, the combinations of a dative and a genitive are slightly more numerous than a dative by itself. Since examples are infrequent in any prose text, it is not possible to track a decline in DEPs of mind PObjs from early to late OE, but the increase in the proportion of IPs co-occurring with such

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

148

       

datives as are found with copulas suggests a diachronic change, with implicit possessors of PObjs becoming more restricted as the OE period wore on.

6.4 Conclusions: body and mind Before moving on to an examination of DEPs in EME, I summarize the main similarities and differences in my findings concerning DEPs with body versus mind words. With both body and mind words, it is with PObj possessa that we find the most examples of DEPs, and with both there is a general increase in overt expression of the possessor as an IP within the OE period. Clear differences between the two semantic spheres emerge when we look at direct argument possessa. In Chapter 4, it was shown that DEPs were less common with body object possessa than with body subject possessa, but they are found with both grammatical relations even in late OE. In contrast with the body words, DEPs are unusual, not just less common, with mind object possessa in the poetry, and there are no certain examples in the prose, even though the prose corpus is much larger than the poetic corpus. DEPs were common in the poetry with both body and mind subject possessa, but while DEPs remained a nontrivial proportion of expressed possessors with body subjects throughout the OE period in the prose, they suffered a drastic reduction with mind words in the prose. A final interesting difference between the body and mind DEPs is that with mind words there seems to be no limitation of DEPs to adversely affected possessors. When we combine the poetic and prose corpora, we get enough examples of direct argument body possessa with lexical verbs to be able to state with confidence that the use of a DEP with a possessor who was beneficially affected by the event described in a sentence was highly unusual. With mind possessa, on the other hand, we have a much smaller number of examples of DEPs, but most of them refer to a beneficially affected possessor. Overall, the limited range and frequency of DEPs with mind words suggests a decline from the Common Germanic period, although this is speculative in the absence of a systematic investigation of DEPs with this type of word in other early Germanic languages. Unlike with the body words, this decline does not seem to include a restriction to affected possessors in the small number of examples of mind words playing the role of a direct argument of a lexical verb. A similarity between the two types of inalienable possessa is that there is no good evidence to support the idea of Latin influence suppressing the use of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

6.4 :   

149

DEPs. For example, without further investigation, the difference between the frequent use of a dative with copulas with mind subject possessa in the poetry and its non-existence in that situation with the prose might suggest Latin influence as an explanation, since native poetry is generally free from such influence, unlike prose translations. However, this explanation does not bear closer scrutiny. As seen, some examples of datives are found with mind PObjs with copulas even in late prose, and we also have clear evidence that even close translations departed from Latin models in the matter of how possessors were expressed. The difference between prose and poetry is more likely to be caused by a diachronic development, a shrinking of the range of DEPs, than by any translation effects in the prose. This discussion of mind possessa concludes the examination of OE. Chapter 7 looks at Middle English and the disappearance of the DEP, except in some fixed expressions.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7 External possessors in Early Middle English 7.1 Introduction This chapter looks at the use of DEPs after the end of the OE period. I define the Early Middle English (EME) period as c.1100–c.1250. This temporal span is not quite the same as the m1 period of the Helsinki Corpus and PPCME2 (1150–1250). While the main focus of this chapter is on the m1 period, it also covers a few texts from what I will refer to as the ‘pre-m1’ period, that is, the period from the last half of the eleventh century through the first half of the twelfth. The chapter will also make some observations about the existence (or lack thereof ) of DEPs in the m2 period (1250–1350). The chapter is organized as follows. Section 7.2 begins the discussion of the transition from late OE to EME proper by introducing the pre-m1 period. Section 7.3 covers the corpus used from this period. Sections 7.3.3 and 7.3.4 compare the evidence from these pre-m1 texts for body and mind words, respectively. Section 7.4 summarizes these findings for the texts in manuscripts of the pre-m1 period and compares them with the findings for OE of Chapters 4 and 6. Section 7.5.1 sets out the corpus and methodology used for the m1 and m2 periods. DEPs cannot be studied in the EME period without setting them in the context of the preservation of the dative case in this period more generally, and so before discussing the results for DEPs, a brief overview of the state of case marking in EME is given in section 7.6, and section 7.7 explains how searching for dative case in the PPCME2 differs from searching for it in the YCOE. Section 7.8 is devoted to reporting and discussing the results for DEPs and IPs of body direct arguments. PObjs are the subject of a separate section, 7.9. While mind words in OE merited their own chapter, DEPs with words referring to these concepts are so reduced in the EME period that they can be discussed in one section, 7.10. My conclusions about DEPs in EME are presented in section 7.11. Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

151

7.2 The transition to Middle English The discussion in Chapters 4 through 6 ended with texts from before the third quarter of the eleventh century. This section makes some observations about the situation between that period and 1150, when the electronic corpora’s m1 period begins. The term ‘transitional’ is often applied to the period between LWS and the language that everyone would agree on calling ‘Early Middle English’, and this term is not inapt, because changes were taking place in English in the eleventh century that would result in a language that was very different from what is widely known as the West Saxon Standard, the language of most of the texts of the later OE period. However, I will avoid this term because it obscures the fact that many of these changes were simply the culmination of earlier processes. In particular, the long shift of English from a language that relied heavily on case marking to signal grammatical relations to one in which case marking played a more subordinate role began long before English was even written down, as did the syncretism of case markers that would eventually result in the final near-complete loss of case marking in English.¹ Furthermore, the term ‘transitional’ has overtones of a language in flux, which could make us lose sight of the fact that a language is in an important sense always transitional, as language learners construct internal grammars of their language and perhaps modify their language because of interaction with a wider speech community. It is particularly important not to appeal to the ‘transitional’ state of a language to dismiss counterexamples to the generalization we want to make about the language of a particular period, as though the syntax of one period was less of a rule-governed system than that of a more ‘settled’ period. For these reasons, I will adopt the term ‘pre-m1’ for the period roughly between the last half of the eleventh century and the middle of the twelfth.

7.3 The pre-m1 period 7.3.1 Texts and manuscripts Before turning to a consideration of DEPs and dative case generally in the pre-m1 period, it is necessary to have some discussion of the texts and ¹ See Hogg (1997) for a discussion of the fact that case marking on nouns and adjectives was more ambiguous in the OE period than is reflected in the standard paradigms presented, for example, in Campbell (1959). It is mainly the inflections on determiners that show the maintenance of case categories by the late OE period.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

152

     

the manuscripts containing them in this period. Treharne (2001) states that we have more than fifty manuscripts with at least some English in them from after the Norman Conquest of 1066 and before the burst of English writings in the late twelfth and early thirteenth centuries. Most, but not all, of these are copies of earlier material. In a number of publications, Treharne argues that these are not, for the most part, mere mechanical copies made by backward-looking antiquarian scribes, that there is clear evidence of updating of both the language and the contents for newer social, political, and religious settings, and the writings were meant to be used (Treharne 2000, 2001, 2012). She states: Such a reappraisal permits the categorical assertion that late West Saxon, the literary dialect of the later Anglo-Saxon era, as it developed into early Middle English, was still a living form of a standard written English, and that its uses were essential in the propagation of religious and didactic, legal and historical learning. (Treharne 2001: 404)

There is no doubt that at least some of the copied texts in manuscripts from what I am calling the pre-m1 period show updating in at least some linguistic features such as morphology and lexicon, and some of these texts exhibit modifications to the content of earlier versions. However, considerable caution is necessary in using them, because we usually cannot be certain whether later scribes have modified the syntax of the originals and whether the maintenance of the West Saxon Standard in some of these texts might have extended generally to the copying of constructions that were at least old-fashioned to the scribes that copied them.² I have tried to minimize the distortions that might result from copying by limiting the texts studied for the pre-m1 period to ones with a probable post-Conquest date of composition. A number of the texts included in the YCOE are from this period, but most of these are copies that certainly or probably originated in a much earlier period. Such copies can be valuable in various respects, especially when they suggest a change in the language, but I have excluded them from this investigation of the decline of the DEP. From among the YCOE texts, I have selected six as being of probable composition around the middle of the eleventh century or later. These texts, which were searched for all DEPs with body and soul words and IPs of direct arguments of affecting verbs, are set out in Appendix A, Table A.2. The date of ² For an informative discussion of scribal habit after the disintegration of the West Saxon Standard, see Liuzza (2000).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

153

the original composition of some of these texts is a matter of dispute, as discussed in Appendix A, but they were all at least probably composed later than the LWS period of Ælfric and Wulfstan. The discussion of pre-m1 language will also cover any relevant examples from the Anglo-Saxon Chronicle in annals falling within the period defined as pre-m1. These texts provide useful information about the expression of possessors of body and mind words in a period later than that represented by the LWS texts.

7.3.2 The dative case in pre-m1 Before looking at DEPs in the selected pre-m1 texts, it is useful to make some very brief observations about the dative case in them. These texts are all ones in which the old system of case-marking categories is intact, although there is increased syncretism because of late sound changes such as the neutralization of vowels in unstressed syllables. Of particular importance here is the retention of the distinction between the dative and the accusative as a category distinction, exemplified for example by the continued use of hine (or hyne) for accusative case and him for the dative. These texts provide clear evidence for the continued use of datives in old functions that disappeared in the ME period: (7.1) a. þonne wille ic hi habban me to wife then will I her:. have me: to wife ‘then I will have her as my wife’ (comargaC,LS_14_[MargaretCCCC_303]:5.6.51) b. for ic hæbbe minne Drihten me on fultume because I have my: Lord: me: in help ‘because I have my Lord as a help to me’ (comargaC,LS_14_[MargaretCCCC_303]:9.7.130) Example (7.1a) has an especially conservative feature in the use of hi for the feminine accusative singular. In general, the dative hire replaced hi in a given dialect before the historically dative him took over the old functions of the accusative hine. The form minne in (7.1b) illustrates the continued inflection of possessives in its use of the masculine accusative singular suffix -ne. The examples presented in sections 7.3.3 and 7.3.4 are similarly conservative in the retention of case-marking features such as the historically correct use of determiner forms se and him.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

154

     

7.3.3 Pre-m1 results: body I will first discuss the results for direct arguments, then those for PObjs. Since the texts show the continued use of the old dative functions exemplified in the examples in (7.1), it is no surprise to find some examples of DEPs in them. Disappointingly, the texts yield too few examples of body possessa, either DEPs or affected IPs, to make any comparative figures very useful; I found only seven examples of my direct argument body possessa with an overt possessor, not counting a small number of examples in which the possessor is not adversely affected and therefore not a good candidate for a DEP. I will therefore merely make some comments and present a few notable examples. Unlike some of the other pre-m1 texts, any chronicle annal from 1050 or later obviously presents no uncertainty about being a late composition and so it would be nice if we found several body direct argument examples with affected possessors in these annals to look at their syntax. However, these annals only turned up one relevant example with a direct argument, the sentence in cochronE of a DEP with an object presented as (7.2): (7.2) him het se cyng þa eagan utadon him: ordered the king the: eyes: out.put ‘the king ordered his eyes put out’ (cochronE,ChronE_[Plummer]:1096.7.3256) While (7.2) shows that DEPs of body objects were possible right up until the end of the eleventh century, this part of the chronicle tells us nothing about the comparative frequency of DEPs and IPs; I found no examples of an IP with a direct argument of an affecting verb in the pre-m1 part of this chronicle. Example (7.2) is the only clear instance of a DEP with a body direct argument in the selected pre-m1 texts of the YCOE. I found one more possible example, presented as (7.3): (7.3) and þæra eadigra fæmne þæt and the:/ blessed:/ woman:/ the: heafod of asloh. head: off struck ‘and struck the holy woman’s head off ’ (comargaC,LS_14_[MargaretCCCC_303]:22.11.351) Since accounts of saints’ sufferings were always rich sources of examples of affected possessors of bodies and body parts, it is unfortunate for our purposes

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

155

that this one is about a woman. The problem with example (7.3) is that both the feminine singular noun and its modifiers are ambiguous in form between dative and genitive case. The NP would be þære eadigre fæmne in standard West Saxon, but the -a endings are typical of the confusion of vowels in unstressed syllables in the EME period. This confusion does not affect the ability to determine the case, since the earlier e ending was also ambiguous. This NP is tagged as genitive in the YCOE, but I think it must be construed as dative because of the determiner þæt, which is unexpected following a prenominal genitive NP but completely typical of a DEP.³ On the other hand, despite having the same formal syncretism (7.4) is probably to be taken as a genitive: (7.4) and bæd þære fæmne fet and and ordered the:/ woman:/ feet: and handan tosomne gebindon hands: together bind ‘and ordered the woman’s feet and hands to be bound together’ (comargaC,LS_14_[MargaretCCCC_303]:18.1.289) Here, a genitive interpretation is the most likely because of the lack of a determiner before fet and handan. Things are not completely clear here, however, since as Crisma (2011) notes, coordinated NPs are ones where the requirements to mark definiteness are relaxed in OE as well as in some other languages. This would be especially the case with a commonly occurring phrase like ‘hands and feet’. Whether or not we count these two examples in MargaC as DEPs, the fact remains that the three examples containing clearly marked affected possessors have IPs:⁴ (7.5) a. min swyrd sceal þinne þone fægran lichaman eall to my sword shall thy: the: fair: body: all to styccan forcyrfan and þine lieman ealle tosindrian, pieces cut and thy limbs all sunder ‘My sword shall cut your fair body all to pieces, and tear apart your limbs’ (comargaC,LS_14_[MargaretCCCC_303]:7.13.104)

³ Although determiners were not usual after a nominal genitive in prenominal position, they were common after a pronominal possessor, but only when the possessum was modified by an adjective or was a substantival adjective, as established in Allen (2006). ⁴ The appositive adjective in the first of the two IPs in (7.5a) would presumably make a DEP impossible with that possessum.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

156

     

b. and þine ban ic sceal ealle forbærnan. and thy: bones: I shall all burn.up ‘and I shall burn up all your bones’ (comargaC,LS_14_[MargaretCCCC_303]:7.13.105) In addition to running my searches on the YCOE file, I read Clayton and Maginnis’ (1994) edition of this version of the Life of St. Margaret in looking at the state of case marking and other linguistic features of this late OE manuscript. I found other examples of highly affecting situations not captured by my list of affecting verbs. In all of them in which the possessor is expressed, it is expressed with an IP: (7.6) and heo willeð minne lichamen to sticcan gebringan and they will my: body: to pieces bring ‘and they want to reduce my body to pieces’ (comargaC,LS_14_[MargaretCCCC_303]:10.10.152) The overall impression, then, is that IPs are dominant in this text with direct arguments even with highly affected possessors. In contrast to the YCOE texts just discussed, the Worcester Fragments, although short, turn out to be a gold mine of DEPs in a very late manuscript. These fragments have ten examples of DEPs in a notable list of ailments of old age beginning with these sentences: (7.7) him deaueþ þa æren: him dimmeþ þa him: deafen the: ears: him: dim the: eiʒen eyes: ‘his ears grow deaf, his eyes dim’ Worcester Fragment B, l.17 Hall 1920 The next lines contain eight more DEPs of body subject possessa describing the decay of body and mind. These examples cannot of course be taken as evidence that DEPs were still a productive part of ordinary English around the turn of the thirteenth century, since these fragments are clearly a copy of much earlier English. They do of course show that DEPs with negatively affected possessors of direct arguments were still possible for some time after the Norman Conquest and they also suggest that the scribe known as the Tremulous Hand

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

157

of Worcester understood the structure and presumably found it tolerable, since he updated the OE in various respects in the large OE corpus that he copied. Before moving on to the discussion of PObjs, it is worth noting an interesting example of an implicit possessor that is no longer possible in Modern English that I noticed in one of the selected texts: (7.8) ac heng þæt heafod adun but hung the head down ‘but hung his head down’

(coleofri,Leof:85.91)

This example is notable because implicit possessors of this sort were uncommon in OE prose generally, as opposed to poetry. Unfortunately, the text containing this example, the Vision of Leofric, yields no examples of affected possessors of body possessa in any grammatical function or any DEPs of PObjs, and so offers no data on a possible connection between the use of DEPs and implicit possessors. In contrast with the nearly complete lack of DEPs with direct arguments in the small pre-m1 corpus, the data for verbal clauses in these texts contains eight DEPs of PObjs. However, with verbal predicates they are almost completely limited to the expressions involving control being in someone’s hands or coming into someone’s hands:⁵ (7.9) a. and eall þæt land þan kyninge to handan and all that: land: the:  king: to hands: ‘and subjected all that land to the king’ (cochronD,ChronD_[Classen-Harm]:1074.1.2413) b. and þam Romaniscan here on hand eodon and the: Roman: army: in hand went ‘and went into the hands of the Roman army’ (covinsal,VSal_1_[Cross]:17.1.119) I found only one example in the pre-m1 period that does not involve this sort of meaning: ⁵ As always, an IP was the only possibility when the noun was modified by ‘own’: (iii)

oððe on his agenre hand Or in his own hand ‘or had in his own control’

heold. held (cochronE,ChronE_[Plummer]:1100.17.3327)

There are several such examples in the late Peterborough Chronicle annals.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

158

     

(7.10) ær þan þe he wolde þa corona him on before that: that he would the crown him: on heafode settan, head set ‘before he would set the crown on his head’ (cochronD,ChronD_[Classen-Harm]:1066.70.2246) We see in section 7.9 that DEPs continue to be found, albeit at a low rate, in expressions for putting a crown on someone’s head in EME. It seems likely that at some point expressions like dat to fotum, him on heafode, etc. were idioms entered into the lexicon as units rather than due to a general availability of DEPS as a part of the grammar. As seen in Chapter 4, these idioms were the most commonly occurring uses of DEPs with PObjs in Late West Saxon and late OE generally. Clearly, these idioms were still used in compositions of the late eleventh century, as these examples from annals composed towards the end of that century illustrate: (7.11) a. & se cyng hit let þam men to handa þe and the king it let that: man: to hands that him eallra meast bead. him: all:. most offered ‘and the king made it over to that man who offered the most of all’ (cochronE,ChronE_[Plummer]:1086.26.2858) b. & þæs cynges castel gewinnan heom to handa. and the: king: castle win them: to hands ‘and to win the king’s castle into their control (i.e. to bring about the surrender of the king’s castle)’ (cochronE,ChronE_[Plummer]:1087.19.2992) The number of body PObjs with an expressed possessor in these texts is too small to make any comparisons with IPs or to say that DEPs were completely restricted to these expressions, but these examples add to the evidence that these idioms were still current in the latest stage of OE, and not just in Ælfric’s writings. Furthermore, example (7.9) illustrates how idiomatic the combination Verb + dat + on hand was at a late stage, since it translates the Latin posuerunt in manibus eorum, ‘placed in their hands’, which uses the genitive eorum. A look in more detail at the body PObjs in one particular text, comargaC, is instructive. The number of expressed possessors is admittedly too small in this text to give a convincing demonstration that the composer favoured IPs over DEPs with PObjs, but the few examples that are found are at least consistent

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

159

with this idea. The date of composition of this text is uncertain but is certainly later than the LWS period. A search using the restricted lists of prepositions and body nouns revealed no DEPs but four IPs of body PObjs.⁶ The absence of DEPs with PObjs in this text is more likely to be due to the subject matter, which did not involve surrenders, than to any disinclination or inability ever to use a PObj with a DEP, but it is notable that two of the sentences with IPs involve combination of verb and PObj that often used DEPs in OE: (7.12) a. And of his toþan leome ofstod and from his teeth light issued ‘and a glare emanated from his teeth’ (comargaC,LS_14_[MargaretCCCC_303]:12.6.184) b. And he þa Malcus to hire fotum gefyll and he then Malcus to her feet fell ‘He then, Malcus, fell to her feet’ (comargaC,LS_14_[MargaretCCCC_303]:22.6.343) Compare (7.12a) with the Ælfrician examples of (7.13): (7.13) a. Him stod stincende steam of þam muþe him: issued stinking breath: from the mouth ‘A stinking breath emanated from his mouth’ (cocathom1,+ACHom_I,_5:221.135.1007) b. him stod sweflan lig of þam muðe. him: issued sulphurous flame  from the mouth ‘a sulphurous flame issued from his mouth’ (cocathom1,+ACHom_I,_31:445.187.6233) As discussed in Chapter 4, Ælfric used DEPs with PObjs with some frequency, but most of the examples involve commonly occurring expressions. Since Ælfric used IPs more commonly than DEPs in general, it seems probable that his use of a DEP in this example reflects the fact that the construction was still common even at a late stage in horrifying descriptions of devils. Of course, an IP was always a possible variant in such descriptions, as it was in accounts of people falling to someone’s feet to beg for mercy, but the choice of an IP by the scribe who penned MargaC adds to the picture of a language in which IPs were dominant. In addition to these examples involving nouns on my

⁶ The form hire in (7.12b) is ambiguous, but it can only be interpreted as genitive in this position immediately following the preposition.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

160

     

restricted list of body nouns that I used in my searches for PObjs, I also found an example of an IP with a body noun (swire) not on the restricted list: (7.14) and hi hire swiðre fot upon his swire gesette and she her right foot upon his neck set ‘and she set her right foot on his neck’ (comargaC,LS_14_[MargaretCCCC_303]:14.9.221) In Chapter 4, it was noted that the strong preference in Late West Saxon for IPs might be exaggerated by the fact that the LWS corpus is dominated by the works of one author, Ælfric. The dominance of IPs in the admittedly small number of examples in MargaC is suggestive that Ælfric was not unusual in his preferences in this regard. This seems especially significant in view of the fact that in at least one respect, MargaC seems surprisingly conservative in its syntax. In particular, it is a bit surprising to find several examples of a possessive followed by a determiner in this late text. This Poss Det construction is found in the example given above as (7.5), and I present another example in (7.15): (7.15) for þinra þære mycele ara for thy: the: great mercy ‘for your great mercy’ (comargaC,LS_14_[MargaretCCCC_303]:19.13.323) This Poss Det construction is not directly relevant to our discussion of DEPs and IPs, but I think a mention of it not out of place here given the uncertainty about the date of composition of this saint’s life. The Det Poss construction seems to have been rather old-fashioned by the late eleventh century, as argued in Allen (2013), and the fact that several examples of it are to be found in this piece of writing is suggestive of an earlier, rather than a later, time of composition. The construction is still found in some other twelfth-century copies of OE writings but not found in original compositions that can be securely dated after a Peterborough Chronicle entry of 1093. It is certainly possible that this construction was copied by a scribe who would not have used it in his or her own speech. The same can of course be said for the datives of (7.14), but while the Poss Det construction was pretty clearly old-fashioned by the twelfth century, the datives must still have been pretty current, since some examples are still found in the case-impoverished texts of a century later, as is shown in section 7.8. The point of this digression about the Poss Det construction is that the scribe of MargaC seems to have generally been happy to use rather

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.3  -1 

161

old-fashioned constructions. Whether the composition was pre- or postConquest and the absence of DEPs is due to any updating of the language by a scribe or was there in the original manuscript, the text seems to reflect a language in which DEPs were little used despite a robust dative/accusative distinction. So far we have been looking at verbal predicates, which represent nearly all the small number of PObjs in the pre-m1 texts in YCOE, but is should be mentioned that The Avenging of the Saviour (covinsal) also provides an example of a DEP with a non-verbal predicate: (7.16) þæt ðæt adl, þe we hatað cancer, that the: disease: that we call cancer hym wæs on þam nebbe fram þam swyðran næsþyrle, him: was on the face from the right nostril oð hyt com to þam eage. until it came to the eye ‘that the disease that we call cancer was on his face from the right nostril until it reached the eye’ (covinsal,VSal_1_[Cross]:1.6.5) We have seen that examples like (7.16) were not very common in late OE, and so this example provides a bit of evidence that such constructions remained possible, even if infrequently used in the texts at a late period. Note that this is not a mere description, since it describes a serious affliction. Cross’s translation has ‘affected his nose’ (1996: 249).

7.3.4 Pre-m1 results: mind We saw in Chapter 5 that DEPs were much more restricted with mind possessa in OE prose generally than with body possessa, and in the LWS texts indisputable examples of DEPs of mind words with direct arguments of lexical predicates are not to be found. Examples of DEPs of PObjs, which had always been the most likely situation for a DEP, were nearly restricted to certain idioms in late OE. My findings for the selected texts, although limited, paint a similar picture for the pre-m1 period. The search for mind possessa as direct arguments with overtly expressed possessors in the pre-m1 texts turned up only seven examples, and given the LWS situation, it is no surprise to find that the possessor is an IP in all of them. I found twelve examples of PObjs in the pre-m1 texts, of which ten have an IP:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

162

     

(7.17) þa wearð he on his mode swiðe gedrefed. then became he in his mind very troubled ‘then he became very troubled in his mind’ (cochronE,ChronE_[Plummer]:1087.37.3003) The only one of these to have a simple DEP involves one of the same idioms used by Ælfric and other late writers: (7.18) becom him to gemynde his oðer scoh came him: to memory his other shoe: ‘He remembered his other shoe’ (coneot,LS_28_[Neot]:66.56) It is perhaps more of a surprise to find this example of a copula dative with an added IP: (7.19) Ac þonne þe ealre angsumest byð on þine mode but when thee: most.anxious is in your mind ‘but when your mind is the most anxious/but when it is the most anxious for you in your mind’ (coneot,LS_28_[Neot]:100.90) We saw in Chapter 6 that these copula datives with on mode are met with occasionally in late OE, but were already unusual, so this example adds to the evidence that although datives had become infrequent in such situations, in the texts at least, they were still a possibility. If this example were from an earlier composition, it would support Ahlgren’s (1946) idea of the ‘blended’ construction being an intermediate step between the dativus sympatheticus and what he calls the ‘poss. adj.’, i.e. the IP. In such a late text, however, this example is from a period when the IP is very well established as the dominant construction generally, and the dative seems more like an addition to the IP than the other way around. The interpretation of the increase in the number of ‘blended’ constructions in late texts is discussed a bit further in section 7.11.

7.4 Summary and comparison with late OE We saw in Chapters 5 and 6 that by the end of the period treated here as OE, IPs were the dominant construction for expressing affected possessors with both the body and mind words used in this study, with the DEP having lost ground during the OE period. With body direct arguments of verbal predicates, DEPs had become entirely restricted to highly affected possessors,

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.4      

163

and with adjectival predicates they were at least very unusual. The reduction of the frequency of DEPs with body possessa in late OE is particularly notable with PObjs. DEPs are still more frequent with PObjs than with direct arguments in the late OE texts, but the decline in frequency is more marked because PObjs had always occurred more freely with DEPs than direct arguments had. DEPs were unusual with direct argument mind possessa in prose texts all through OE, a fact which cannot be attributed to a data gap, since there are numerous examples of IPs with these words. DEPs were considerably more frequent with PObjs all through the OE period, but by the late OE period they were almost completely restricted to frequently-occurring expressions that could be listed in the lexicon. The fact that we only find a couple examples of mind PObjs in the pre-m1 texts gives the impression of a further decline from LWS, but it could well be due to the small size of the corpus and the subject matter, rather than to any change in grammatical possibilities. A small number of examples of the same expressions that used DEPs of PObjs are found in some later texts, as we see in section 7.10. As noted in Chapters 4 through 6, our information about late OE is dominated by the works of Ælfric and, to a lesser extent, Wulfstan. While the pre-m1 texts do not yield a great deal of information, they do provide useful examples from other late authors. The picture we get from the limited data of the pre-m1 period does not cause a reassessment of the findings based on the LWS texts of Chapters 4 and 5. The pre-m1 texts provide examples of highly affected possessor IPs. This is nothing new, but these few examples add to the late examples from texts not by Ælfric or Wulfstan. The example of a DEP in cochronE is evidence that the construction continued to be used in original compositions of the pre-m1 period. As in the earlier prose texts, DEPs of body direct arguments were limited to highly affected possessors. With body PObjs, we continue to find DEPs with the idioms that used them in the LWS texts, and the single example of a DEP with a copula is consistent with the earlier finding that such examples were rare but still possible at a late stage. The pre-m1 results for mind words afford no real surprises, the most notable fact being that the late appearance of a dative possessor in the copular construction in example (7.19) suggests that this construction was maintained fairly late in spite of a low frequency all through the OE period. All in all, the limited information provided by the pre-m1 texts is consistent with the findings of the much larger LWS corpus from an earlier period. Let us turn now to the period that everyone would agree in calling ‘Early Middle English’ (EME).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

164

     

7.5 M1 and beyond 7.5.1 Corpus Even if we extend EME into the m2 period, we have a smaller corpus to work with than with OE. The total word count for the PPCME’s texts for the m1 period is 217,388 and the m2 texts contain only 93999 words.⁷ In contrast, the YCOE contains 1.5 million words, although the corpus used in this study is smaller because it excludes some texts as too late to be considered good witnesses of OE. In looking at the disappearance of DEPs, it is important to keep the small size of the m1 and m2 corpus in mind, because a paucity of examples of DEPs might be a data gap—it could be that affected body possessors are simply not mentioned very often in the EME texts, whether with DEPs or IPs. Only a comparison of DEPs and IPs in the texts can rule such a data gap out. The data for this chapter come mainly from the PPCME2, but because of the small size of the m1 corpus I supplemented the investigation of the PPCME texts with some EME texts not included in the PPCME2. These are listed, along with their word counts, in Appendix A. As it turns out, they yielded only a very few relevant IPs and no DEPs. These texts are listed in the tables of this chapter only when they supply one of these examples. I have not systematically recorded examples in any other non-PPCME2 texts, but have noted some examples of interest, taken either from sources that discuss the ‘sympathetic dative’ in the history of English or from my own reading. The PPCME2 consists almost entirely of prose; the only verse that is included is a portion of the Ormulum, a late twelfth-century text which cannot be ignored in any study of EME for a number of features that make it a particularly good witness to the northeastern Midlands dialect in which it was written, including the fact that it is found in a holograph manuscript. The literature supplies a number of examples from poetry, especially Laʒamon’s Brut, and so the two versions of this poem, usually known as Brut (C) and Brut (O), after the manuscripts they are found in (British Library manuscripts Cotton Caligula A ix and Cotton Otho C xiii, respectively), are discussed in some detail in Appendix A. Here, it is sufficient to make some comments on how useful either of these versions is likely to be in reflecting the morphology and syntax of an ‘ideal speaker/hearer’ of any period or dialect.

⁷ The mx1 extension in the PPCME2 file names indicates that while a manuscript belongs to m1, the composition of the text was (or probably was) in the OE period. I have not included texts with this extension in my corpus.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.6     

165

The fact that these texts are copies, in addition to the generally conservative nature of poetic language, renders both versions less than ideal as a picture of the language of any area or time. When dealing with writings from this period, however, we must do the best we can with what we have, and although the problems with such a text must be kept in mind, there is much that it can teach us. Stanley says that the scribes who copied the Caligula manuscript ‘could not have thought that they were copying a poem written in the idiom of their own time’ as they copied the forms of their exemplars (Stanley 1968: 29). It seems likely that the Otho manuscript better reflects the language of the time and place of copying than does the Caligula. However, the language of Caligula is probably not a bad reflection of the morphological and syntactic options of English when overt case marking was declining but still very much a part of the language, always keeping in mind that we can expect the language of poetry to be rather archaic. The state of the dative case in these texts is discussed briefly in the next section. The examples presented in this book are taken from Brook and Leslie’s (1963, 1978) edition, henceforth abbreviated as BL.

7.6 The dative case in EME It is worth spending some time on the discussion of the case-marking systems of different texts because one popular theory about why the DEP was lost in English, discussed in Chapter 8, hinges on the loss of the dative/accusative distinction. With such an explanation, we would expect the presence or absence of DEPs to depend on the state of case marking in a given text. As we shall see, that is not the case. Here, I will give a brief discussion of the retention of a morphological dative/accusative distinction in some of the more important EME texts. For further discussion of the case systems of EME, see Allen (1995: Chapter 5), where the importance of distinguishing between dative morphology and the dative case as a grammatical category is emphasized.

7.6.1 Morphology and category The dative/accusative distinction died out as a morphological distinction in Middle English in different dialects at different times. As discussed in section 7.3.2, the category distinction was well preserved in all the texts studied in this investigation for the pre-m1 period, although syncretism of forms was more

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

166

     

pronounced than in the Late West Saxon of Ælfric. The category distinction was lost earliest in the north, where this development was already clearly in progress in OE texts from that area, and it is complete in the late twelfth- and early thirteenth-century texts of the Northeast Midlands and the northern part of the Southeast Midlands. In Kent, the distinction was still marked as late as 1340, in the Aʒenbite of Inwyt, a text in a holograph manuscript, possibly written by an old man (see Gradon 1979). The loss of case-marking categories in some dialects, as well as the variability across dialects, complicates the glossing of examples in the m1 period. The retention of a clear dative/accusative category distinction in some dialects in a period when others had lost this distinction complicates the question of how to gloss forms like him, a specifically dative form in OE but a general object form in some EME texts. Generally, I have glossed pronouns and other elements of NP without case information except when directly relevant, as discussed in the Note on Glossing at the beginning of this book. As mentioned in section 1.1, the loss of the dative/accusative distinction presents a problem of terminology. Even in texts written in dialects that maintained the most case marking in EME, we find numerous examples of bare NPs that have no clear case marking but occur in various functions that were previously typically marked with the dative case, as discussed in sections 7.8 and 7.9. By ‘bare NP’ I mean an NP (nominal or pronominal) that is not the object of a preposition. In texts in dialects that maintained a clear dative/accusative distinction in EME, there is of course no problem with identifying such an NP as dative if some element of it is in a dative form, although problems arise from syncretic forms. In texts written in dialects that have lost the dative case as a distinct morphological category, however, calling these bare NPs ‘dative’ requires a more abstract notion of case marking that is not directly connected with morphology. One way to avoid the terminological problem when discussing the loss of EPs (external possessors) in ME would be simply to use the term EP, whether the possessor is clearly marked dative or not. The problem with that approach is that the term EP generally covers external possessor of other types, in particular, ones that are expressed in locative prepositional phrases. Because such locative EPs are the subject of some discussion in this book, and also because the EPs that are the main subject of this chapter must be seen as continuations of the DEP, I will continue to use the term DEP for EPs that would have had dative marking in earlier English, even for examples from texts in which the morphological dative/accusative distinction has clearly been lost.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.6     

167

It is not a straightforward matter to sort out morphology and category distinctions for more than one reason. One difficulty is that we cannot simply equate historically dative forms with a dative category. In some texts, hine is restricted in the old accusative functions, principally direct objects, but the etymologically dative form him is sometimes found in these functions as well as in its historical functions. Vices and Virtues, written in a southeastern dialect c.1200, is an example of such a text. In Allen (1995), I argue that this is an instance of structured variation rather than case confusion—the variation only goes one way, with the dative forms expanding into the functions historically reserved for the accusative. I assume that the scribes who wrote such texts retained a category distinction between accusative and dative case, with hine being a specifically accusative form and him a form that was either dative or accusative. In this text, we even find the retention of different forms for dative and accusative in the singular feminine pronoun, a distinction that disappeared in other texts in which the distinction is still maintained in the masculine pronoun. I will use the term ‘case-rich’ for texts in which there is evidence for the retention of a dative/accusative category distinction. The system found in the case-rich texts can be opposed to those of ‘caseimpoverished’ texts such as the Ormulum, probably written in Lincolnshire c.1180 and the Second Continuation of the Peterborough Chronicle, written c.1154 in Peterborough.⁸ In both these texts, there is no trace of the dative/ accusative distinction, although there is some use of a case ending -e on the objects of prepositions that derives historically from the old dative case. It seems best not to treat this as dative case but rather as an optional prepositional case. In between the texts which clearly maintain a dative/accusative category distinction and those which clearly do not, we have texts like those of ‘Dialect AB’ of the West Midlands, where there are only some remnants of the dative/ accusative distinction in the masculine third person singular pronoun. The study of this dialect is complicated by the fact that with Dialect AB we are dealing with copied manuscripts, and it is very possible that a scribe who no longer commanded the dative/accusative category distinction was copying the work of an earlier one who did. I regard the texts of Dialect AB as

⁸ Note that the PPCME2 file cmpeterb.m1 contains both the First and Second (Final) Continuations of the Peterborough Chronicle. The First Continuation was written by the same scribe who copied the earlier annals, and this scribe attempted to imitate the case-marking system of the OE annals that he had copied, but clearly did not control the old system; see Allen (2007, 2008). For this reason, I excluded any examples coming from the First Continuation in this study.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

168

     

case-impoverished because it seems likely that the sporadic hine forms do not indicate a systematic distinction. It was indicated in section 7.5.1 that some observations would be made here about the retention of DEPs in the two versions of a text dating to the m2 period but not included in the PPCME2, Laȝamon’s Brut. Both manuscripts containing this poem were penned in dialects far enough to the south to have relatively conservative retention of case-marking distinctions at a fairly late date. Even the more modern language of Otho shows a systematic distinction between accusative and dative case, although syncretism of forms is much more advanced than in Caligula. Discussions of the language of Otho tend to focus on the replacement of the forms found in Caligula by newer forms, without distinguishing between case forms and case categories. However, a systematic study of the case categories of the Otho text shows that although the Otho scribe frequently uses the historically dative form him where the accusative form hine is used in Caligula, we still do find numerous instances of hine in Otho, where it is used ‘correctly’, that is, where it would have been used in OE, as is the occasional -ne suffix on masculine singular accusative adjectives. Uninflected determiners are also more frequent in Otho, but the accusative/dative distinction is still apparent in the determiners, and of particular interest here, determiners inflected for genitive case are still found, although the uninflected þe is more common. Given the Otho scribe’s modernizing tendencies, it seems reasonable to assume that his command of this case morphology is not a bad reflection of the linguistic situation in his community, and not simply the result of the retention of conservative forms from his exemplar.

7.6.2 Dealing with the syncretism As the morphological distinction between the two cases was lost, speakers came up with various ways of dealing with the new lack of this morphology. As is well known, a combination of the fixing of word order and the use of prepositions eventually replaced the dative case in its core function, namely marking an indirect object. What is less widely recognized is that for a surprisingly long period either the IO DO or DO IO order was possible. For a discussion of the fixing of the IO DO order when no preposition was used, see Allen (1995: section 9.6.3) and Allen (2006). From a functional point of view, the fact that people could tolerate this ambiguity in grammatical functions becomes less surprising when we realize that one object was

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.6     

169

normally human and the other object was normally not, making it easy to tell from context which object was direct and which indirect. In this subsection, I do not discuss the use of bare NPs or prepositional phrases in this core dative function but instead focus on the fate in EME of some of the other old dative functions surveyed in Chapter 2, as background to the discussion of the use of DEPs in sections 7.8 and 7.9. Most of the uses of the dative discussed in Chapter 2 for OE are found in EME even in case-impoverished texts in which the dative was not distinctively marked. Many of these functions are found only in a small number of examples, but they point to a grammar in which overt dative morphology was not essential for the use of a bare NP in the old dative functions. A construction that is not directly relevant to the study of DEPs but which is relevant to the general retention of bare NPs in old dative functions that continues to be found in m1 is the use of ‘modal’ be with an infinitive, a construction not covered in Chapter 2 because it is not directly relevant to the study of DEPs. In Modern English, be can be used ‘modally’ to express obligation and related concepts, as in I am to be there at 9:00 a.m. In OE, beon/wesan were used in a similar way, with the major difference that the dative case could be used to indicate the person upon whom the obligation fell:⁹ (7.20) Us is to smeagenne þæt word þe he cwæð us: is to ponder the word that he said ‘We must ponder the word that he said’ (cocathom1,ÆCHom_I,_18:322.153.3544) Visser refers to this type by the non-committal label ‘us is to donne hit’ and comments that is was common in OE but rarely met with in EME (1963: §373). A search of the PPCME2 found four examples in the m1 texts, and none in m2, and Visser gives a few later examples from texts not included in the PPCME2. Two of the m1 examples are presented in (7.21). Note that (7.21b) is from the completely case-impoverished Ormulum, and so himm is simply in the object case. The same is true of the me in (7.21a); although this AB text contains some remnants of accusative forms, it can be regarded as caseimpoverished.

⁹ This construction has been the subject of considerable debate about whether it is due to Latin influence and whether the infinitive should be treated as ‘active’ or ‘passive’. For summaries and commentaries on some of this debate, see Mitchell (1985: §§934–49) and Visser (1963: 367).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

170

     

(7.21) a. hwet me beo to donne what me is: to do: ‘what I am to do’ or ‘what I should do’ b. what himm wass to donne. what him was to do ‘what he should do’

(CMJULIA,108.204)

(CMORM,I,100.866)

Bare NPs continued to be used for a long time in a number of ‘impersonal’ constructions. Indeed, the inventory of verbs used ‘impersonally’ actually expanded in EME, with some verbs found used with a dative experiencer for the first time in early or even later Middle English. For discussions, see Allen (1995: section 6.3.1) and Miura (2014). Here, of most relevance is the retention of ‘dative’ uses with a copula. One of these uses was as a complement to an adjective: (7.22) a. Ah wel is hir þt luueð godd but well is her that loves God ‘But well it is for her who loves God’

(CMHALI,148.292)

In the m1 examples of the old dative functions, the bare NP was usually in the form of a pronoun, as in (7.22), since it most often referred to a topical person. These bare NPs were not susceptible to confusion with a nominative. However, in the m1 period there is a substantial number of examples involving an uninflected noun: (7.23) þt crist is swiðe icweme which Christ is very pleasing ‘which is very pleasing to Christ’

(CMJULIA,95.3)

Example (7.23) is particularly interesting because it is from the West Midlands dialect AB, in which the dative/accusative distinction had at best a tenuous existence. In (7.23), the noun crist is not marked for case, so if we want to say it is dative, this would be an abstract dative rather than a morphological one. Example (7.24), however, comes from a text, the Ormulum, written in a Northeast Midlands dialect in which the dative/accusative distinction disappeared early. In the Ormulum, there is no trace of this distinction in either the determiners, now not inflected at all for case, or the pronouns, where there is only evidence for a general object case. Examples like (7.24) show that even with a complete lack of case marking on a noun, godd ‘God’ in this example, bare NPs continued to be possible complements to adjectives:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.6     

(7.24) & godd ful cweme and God fully pleasing ‘and completely pleasing to God’

171

(CMORM,I,13.228)

The most straightforward account of such sentences is that particular adjectives were lexically licensed to appear with a bare complement. These bare complements alternated with prepositional complements, sometimes in the same text: (7.25) & in-wardliche bonen swa icweme to godd and inward prayers so pleasing to God ‘and inward (heartfelt) prayers, so pleasing to God’ (CMJULIA,112.268) Given that complements to bare adjectives could be found even in these caseimpoverished texts, it is not a surprise that similar examples are to be found in both versions of Laȝamon’s Brut with forms inflected for case. The bare complements were on their way to being ousted by prepositional ones, but were not uncommon in m1 and in later poetry. A few examples are found in the m2 Aʒenbite of Inwyt, e.g. (7.26a), but seem to be limited to pronouns: (7.26) a. Vor hi byeþ ous nyeduolle ine þise lyue dyadlich. for they are us needful in this life mortal ‘because they are needful to us in this mortal life’ (CMAYENBI,110.2125) b. Þyse þri þinges byeþ nyeduolle these three things are necessary þet in þe erþe wexeþ. that in the earth grow ‘These three things are needful to all the earth’

to alle þe þinges to all the things

things that grow in the (CMAYENBI,95.1842)

In (7.26a) the pronoun is used on its own, while in the (b) example the noun is preceded by a preposition. Prepositions are also found in this text with pronominal complements of adjectives in variation with bare NPs: (7.27) a. Hit is ynoʒ uor þe / þet . . . it is enough for you that . . . ‘It is enough for you that . . . ’

(CMAYENBI,104.2039)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

172

     

b. þet him ne is naʒt ynoʒ to onworþi ine his herte that him not is not enough to dishonour in his heart þe oþre the other ‘that it is not enough for him to dishonour the other in his heart’ (CMAYENBI,22.327) Prepositions were thus in the process of replacing bare NPs in this text in which pronouns maintained the dative/accusative distinction. The examples I have given here add some information to Maling’s (1983) discussion of the loss of transitive adjectives in English. The focus of Maling’s paper was not to trace the loss of transitive adjectives, that is, adjectives allowing bare NP complements, in English, but rather to trace the reanalysis of like and worth, formerly adjectives, as prepositions. However, she does comment that there is a correlation between transitive adjectives and morphological case marking in the Germanic languages, and that when case marking was lost in English, the oblique NP-complements to adjectives were typically replaced by prepositional phrases (1983: 254). The retention of NP complements to adjectives in the case-impoverished dialects of m1 show that this replacement was a gradual one, and Maling notes that near seems to be a relic of transitive adjectives in Modern English, since she finds that it passes all tests for being an adjective rather than a preposition. It appears that the loss of the dative/accusative distinction did not make the use of a bare NP complement with an adjective impossible, but it presumably gave an advantage to the competing use of a preposition, especially when the complement was nominal rather than pronominal. One type of construction with a copula that Visser (1963: §360) points to as declining sharply in EME is the one involving an inanimate noun, e.g. him wæs egesa ‘there was fear to/for them’, which was covered in Chapter 2 for OE in the discussion of ‘extended existence’. As noted there, it is not clear whether the noun should be treated as the subject or a predicate nominal. What is clear, however, is that the construction was lexically restricted to particular nouns that combined with a copula and a dative. Visser appears to be correct in his observation that the construction had nearly disappeared in EME. However, example (7.28) turned up in the PPCME search for DEPs with mind words:

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.6     

173

(7.28) La, Drihten, nis þe na gemynde þt min suster lætt Lo Lord not.is thee no concern that my sister lets me anen þenigen? me alone serve ‘Lo, Lord, is it no concern to you that my sister lets me do all the serving alone?’ (CMKENTHO,134.11) While (7.28) is not directly relevant to the study of DEPs, since gemynde is not used in the target meaning of ‘memory’ here and there is no expressed possessor, it is a good illustration of the use of an old construction at a late stage, in a translation made in the twelfth century. It is not particularly surprising to find such an example in a southern manuscript that retains the dative/accusative distinction. However, in the few examples Visser adduces in his §362 for Middle English, he does include one example from the case-impoverished Ormulum: (7.29) Mikell ned wass himm much need was him ‘He had great need’

Ormulum 906

Visser also produces a few examples with other nouns from later texts, but the construction was limited to a very few nouns in Middle English generally, the only commonly-used one being wa, which of course lived on into Modern English in the exclamation ‘woe is me!’ It seems that the loss of the dative/ accusative distinction did not make it impossible for copulas to participate in the old constructions that used a dative, but the loss of the dative case probably promoted the use of alternative constructions. In the case of nouns such as need, for example, the verb have was frequently used instead of a copula, similar to the PDE translation of (7.29). To sum up, the loss of dative/accusative morphology did not mean the immediate loss of the old dative functions, but it surely favoured the use of alternative constructions that already existed and made grammatical relationships clearer. The small size of the m1 corpus makes it difficult to evaluate how restricted these constructions had become by this period, but any decline in them does not seem to have been more marked in the case-impoverished dialects than in the case-rich ones. Before turning to DEPs in EME, it is worth noting a change directly relevant to the study of body possessa in EME, although it is in a construction

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

174

     

that I argued in section 2.2.1 should not be treated as a having a DEP. As discussed in that section, OE beceorfan ‘cut off ’ was a ditransitive verb taking an accusative object of the person from whom something was cut, with the part being cut (the head or the hair) in the dative case. In EME we find the dative heafde replaced by the genitive heafdes with the reflex of this verb: (7.30) Ich hit am, þet makede sein iuhan þe baptiste beon I it am that made St. John the Baptist be heafdes bicoruen head: cut ‘I am the one that caused St. John the Baptist to be beheaded’ (CMJULIA,110.235) (7.31) & he ham het euch fot heafdes bikeoruen. and he them ordered each foot head: cut.off ‘He ordered every one10 of them beheaded’ (CMJULIA,121.452) I found four such examples in dialect AB texts, and since these are the only examples of this meaning presented in the bikerven entry in the Middle English Dictionary, it is possible that this new use of the genitive case for the object of deprivation with this verb, integrating it into a more common ditransitive pattern for verbs of deprivation, was a specialty of this dialect. Although genitive objects were not common by this time, they still occurred with some verbs of deprivation such as binimen ‘deprive, despoil’; see, e.g., the lemma binimen, sense 2, in the Middle English Dictionary. This remodelling of a construction formerly using dative case in a function other than its core function of encoding a recipient or a beneficiary is an interesting instance of how English speakers managed to retain the use of a bare NP in an idiom while using the remaining case marking of their language, that is, the robust genitive case. It seems likely that the reason why the genitive case was used rather than a preposition is that the use of a bare NP continued the traditional idiom more closely. However, the idiom was apparently not very frequently used, and the verb died out. The replacement of the dative case with a genitive was only sporadic. With this background, we can now examine the use of DEPs in the m1 period and a bit beyond. ¹⁰ Euch fot was an idiom that meant ‘every one’.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.7   ‘ ’     

175

7.7 Searching for ‘dative case’ and DEPs in Middle English The EME investigation was broadly similar to the OE investigation in its methodology, with some adjustments due to the smaller corpus and differences in coding between the YCOE and the PPCME2. As for OE, the investigation used lists of forms for body and mind words found in EME and searched for dative possessors with these words. The EME lemmas are listed in Appendix B. A list of affecting verbs was used to compare IPs and DEPs with affected possessors of body direct argument possessa. As for OE, no list of affecting verbs was compiled for the mind words. The methodology used for PObjs is outlined in section 7.9. The loss of the dative as a morphological case presents challenges in collecting examples of external possessors in Middle English, and more generally in collecting examples of bare NPs in old dative functions. Unlike the YCOE, the PPCME2 does not code for case marking but only for grammatical function. This is the correct decision. In OE, it is safe to code an NP as dative if it has distinctive dative morphology or if it is ambiguous in case marking but is appearing in a position where a word that would be unambiguously marked would be marked in the dative. In ME, matters are quite different. In this period, dialects had different case-marking systems, and determining whether a given dialect had a dative/accusative distinction as a category distinction, even if it might be subject to a good deal of syncretism, requires careful study of texts in those dialects. There is also the theoretical matter of how abstract we are willing to go in calling a bare NP dative in case. For a brief explanation of how the PPCME2 codes elements that perform functions usually marked by dative case in OE, see Appendix C. Here, it needs to be mentioned that the PPCME2 uses a special tag, NP-DPS, for what it treats as dative possessors. The documentation to this corpus comments that a small number of these are still found in EME. A search for all NP-DPS in the m1 texts yielded twenty-one examples, all with PObjs, as in (7.32): (7.32) a. Forrþi chæs ure Laferrd Crist An maʒʒdenn himm therefore chose our Lord Christ a maiden him to moderr to mother ‘For that reason, our Lord Christ chose a virgin as his mother’ (CMORM,I,120.1044)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

176

     

b. & ʒet he dude mare vs to forbisne and yet he did more us as example ‘and he did yet more as an example to us’ (CMANCRIW-1,II.90.1084) c. Icc þatt tiss Ennglissh hafe set Ennglisshe men to I that this English have set English men to lare learning ‘I that have set this English for the learning of English men’ (CMORM,DED.L315.61) Although I did not count these examples as DEPs, for the reasons discussed in Chapter 2, they do show the continued use of an old dative function in the case-impoverished texts. Note the lack of any case marking in example (7.32c). This is not the only such example found in the PPCME2 sample of the severely case-impoverished Ormulum. Three examples of datives parsed as NP-DPS in the PPCME2 are included here in the count of DEPs of body possessa: (7.33) a. ant cwenene crune sette þe on heaued and queens:. crown set thee on head ‘and set a queens’ crown on your head’ (CMANCRIW-1,I.72.276) b. cruninge of þornes þe set him i þe heaued. crowning of thorns that sat him on the head ‘the crown of thorns that sat on his head’ (CMANCRIW-1,II.142.1909) c. Ðeos sæt wel þan Hælende æt foten & æt this sat well the: Saviour: at feet and at heafde head ‘this one truly sat at the Saviour’s feet and at his head’ (CMKENTHO,138.106) I also found other examples that I treated as DEPs, but which are not treated as dative possessors by the PPCME2, including (7.34): (7.34) ah blent ham þe ehnen. but blinds them the eyes ‘but blinds their eyes’

(CMSAWLES,171.75)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.8  : 

177

Examples like (7.34) seem parallel to the DEPs found in French, German, and other European languages.¹¹ It is important to note that the compilers of the PPCME2 (and other syntactically parsed corpora) are at pains to remind the reader that the parsing is not meant to be taken as a syntactic analysis but rather as a tool for users to find the examples they are seeking. Because examples like (7.34) seem to be parallel to examples that I have treated as DEPs for OE, differing only in lacking distinct dative case marking, I have treated examples like (7.34) as DEPs in EME. The special problems involved with PObjs are discussed in section 7.9.

7.8 Direct arguments: body 7.8.1 The m1 period The results for body possessa as direct arguments of affecting verbs in the m1 period are presented in Table 7.1. Table 7.1 DEPs (all) vs IPs (highly affecting verbs) with body possessa, m1 Affected Obj

IP DEP Blended Total Affected Subj

cmancriw-1.m1 4 cmancriw-2.m1 1 cmhali.m1 2 cmjulia.m1 2 cmkathe.m1 5 cmkentho.m1 1 cmlamb1.m1 0 cmmarga.m1 5 cmorm.po.m1 1 cmpeterb.m1 0 cmsawles.m1 0 cmvices1.m1 1 Total 22

2 0 0 2 2 0 0 1 0 0 2 0 9

1 0 0 0 0 0 0 0 1 0 0 0 2

7 1 2 4 7 1 0 6 2 0 2 1 33

IP DEP Blended Total

cmancriw-1.m1 1 cmancriw-2.m1 3 cmhali.m1 6 cmjulia.m1 2 cmkathe.m1 0 cmkentho.m1 0 cmlamb1.m1 0 cmmarga.m1 5 cmorm.po.m1 1 cmpeterb.m1 0 cmsawles.m1 0 cmvices1.m1 1 Total 19

0 0 0 0 0 0 0 0 1 1 0 0 2

0 0 1 0 0 0 0 0 0 0 0 0 1

1 3 7 2 0 0 0 5 2 1 0 1 22

As with the OE comparisons, the search for DEPs did not specify particular verbs, but the search for IPs specified affecting verbs. All the m1 DEP examples do in fact depict highly affected possessors, making it possible to present the results for the DEP and IP searches together in this table.

¹¹ The ham of this example is annotated as NP-OB2, the tag used for the indirect object of a ditransitive verb by the PPCME2, but an analysis of blinden as ditransitive does not seem right.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

178

     

The discussion of the examples begins with some remarks about sentences that cause problems of analysis. Example (7.35) could possibly be treated as a ditransitive construction, with þe cnapechild an indirect object: (7.35) Þatt daȝȝ þatt teȝȝ þe cnapechild Hiss shapp himm that day that they the boy.child his penis him ummbeshærenn circumcised ‘that day that they circumcised a boy’s penis’ (CMORM,I,145.1198) However, this text (in a passage not included in the PPCME2’s sample) provides clear evidence that it could be used as a monotransitive verb: (7.36) Þeȝȝ umbeshærenn þeȝȝre shapp they circumcised their penis ‘They circumcised their penis’

Ormulum 4084

Example (7.37) looks at first as though it should be included with the count of DEPs: (7.37) þt al ham is to-limet lið ba. & lire. that all them is dismembered limb both and flesh ‘That they are completely torn to pieces’ (CMHALI,142.221) The translation given in (7.37) is that provided for this sentence by Millett (1982) in her glossary lemma tolimet. It is the unambiguously non-nominative case of ham, annotated as NP-OB2, that might suggest its treatment as a DEP here. A DEP analysis would treat lið ba & lire as the subject rather than the NP-ADV that the PPCME2 assigns to it. The problem with this analysis is that is is not a plural form, as we would expect with a compound subject. The Middle English Dictionary’s lemma lith ‘joint, limb’ shows that this noun frequently shows up in conjunction with another body word in what can be considered fixed expressions; the combination lið & lire is very common often accompanied by some form of ‘both’. This suggests that the PPCME2’s treatment of the phrase as adverbial is a good one, and I have not included it as any sort of possessive.¹²

¹² If ham is not a DEP, what is it? One of the uses in the PPCME2 of OB2 is for the complement of an ‘impersonal’ construction, and that would be a traditional analysis of this clause and the analysis that the PPCME2 has assumed, with its indication that there is no overt subject. An analysis that I would favour is treating tolimet as a (participial) adjective and ham as a non-nominative subject of the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.8  : 

179

Leaving aside these questionable examples, there are a surprising number of clear DEPs of object body possessa. Because of the widespread perception that DEPs are hardly to be found in EME, it is worth presenting a number of clear examples: (7.38) a. Ich habbe iblend men & ibroken ham þe I have blinded men and broken them the schuldren. & te schonken shoulders and the legs ‘I have blinded men and broken their shoulders and legs’ (CMJULIA,114.303) b. & swipte hire of þt heaued. and swept her off the head ‘and swept her head off ’

(CMKATHE,52.533)

c. ah blent ham þe ehnen. þe þer beoð but blinds them the eyes that there are ‘but blinds the eyes of those that are there’ (CMSAWLES,171.75) d. þe freoteð ham ut to ehnen that eat them out the eyes ‘that eat out their eyes’

(CMSAWLES,172.78)

e. bint him swa euch lim þt he haueð wið isunged bind him so each limb that he has with sinned ‘bind so each of his limbs that he has sinned with’ (CMANCRIW-1,II.227.3280) f. & bed binden hire swa þe fet & te honden and ordered bind her so the feet and the hands ‘and ordered her feet and hands to be bound so’ (CMJULIA,121.456) Note that although the form hire in (7.38f) is ambiguously dative or genitive in form, it must be interpreted as dative here, because an IP would not have a determiner following the possessive. It is also worth noting that (7.38b) offers an example of how EME translations, like OE ones, could be independent of the Latin source in their grammar. An older edition of this text publishes the (assumed) Latin original as well as the Middle English text and indicates that the English is a translation of mox ille, insurgens, decollauit eam ‘there upon adjective, with the same treatment of lið ba & lire as that suggested by the PPCME2 parsing. On nonnominative subjects in EME, see Allen (1995, 1996).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

180

     

he, rising up, decapitated her’ (Einenkel 1884: 121). The translator could well have followed the Latin more closely by using the English verb bihev(e)den ‘behead’, but chose to use the DEP that was from OE times a more frequent alternative to describe a beheading. Turning to subject possessa, DEPs are indeed unusual. The total number of expressed possessors is not huge, but it is large enough to make it clear that DEPs were in a small minority. It is important to stress that the two clear examples come from texts in which the dative/accusative distinction has vanished without a trace: (7.39) a. & all himm wærenn fet & þeos Tobollenn & and all him were feet and thighs puffed.up and toblawenn. swollen ‘and his feet and thighs were all puffed up and swollen’ (CMORM,I,280.2293) b. & þrengde þe man þærinne ðat him bræcon and crushed the man therein that him broke: alle þe limes all the limbs:13 ‘and crushed the man in it so that all his limbs broke’ (CMPETERB,55.444) (1137.27) In addition to these clear DEPs, I found two examples that seem to continue the ‘blended’ construction of OE: (7.40) a. þine banes akeð þe. your bones ache you: ‘Your bones will ache’

(CMHALI,151.336)

b. & buffeteden him his deorewurðe muð and buffeted him his precious mouth ‘and buffeted his precious mouth’ (CMANCRIW-1,II.84.1018) I found a few IPs but no DEPs of either subject or object body possessa in the very small corpus of non-PPCME2 twelfth-century texts described in

¹³ There is no clear case marking on limes to distinguish the subject from the object, but I agree with the PPCME’s glossing of it as a subject here. If limes were the object, with the translation ‘until (they) broke all his bones’, we would have to assume an unexpressed subject in a tensed subordinate clause, which would be unexpected in this period.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.8  : 

181

Appendix A other than this possible example from an EME composition in the Cotton Vespasian A xxii manuscript: (7.41) and binde him hand and fett and bind him hands and feet ‘and bind him hand and feet’

An Bispel 231.15

This example could be treated as DEP, in which case ‘bind his hands and feet’ would be a good translation. However, we would expect the body nouns to be preceded by a determiner in a DEP, and hand and foot could be regarded as an adverbial NP similar to the biblical smote them hip and thigh. To sum up the situation with the direct arguments in the m1 period, the DEP seems to have retained more productivity than has often been assumed, although it was clearly a minority construction. As for OE, the search for DEPs did not distinguish verb types, but unlike in OE, there were no DEPs in the m1 output that did not involve a highly affecting verb. The DEP had become a marked construction for representing affected possessors.

7.8.2 The m2 period In the m2 period, the PPCME2 searches only yielded one example of a DEP with a direct argument, with a subject: (7.42) Þet heaued me akþ the head me aches ‘my head aches’

(CMAYENBI,51.900)

The use of a dative with this verb may have been promoted by the fact that the ‘impersonal’ me akþ ‘I ache’ was still used at this time. It should be noted that the m2 period is one in which prose is scanty and also liable to be affected by translation effects. Two of the texts are close translations from French and the other is a biblical translation, a genre which always presents the possibility of interference in the syntax. It is impossible to be certain whether the almost complete lack of DEPs in these texts is a reflection of common English usage at this time or not, especially given the fact that hardly any IPs of body possessa are to be found either, as discussed later in this section. More examples of DEPs are to be found in poetic texts from manuscripts on the boundary of the m1 and m2 periods. They are not rare in either the Caligula or Otho versions of Laȝamon’s Brut, which are in manuscripts from right around the beginning

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

182

     

of the m2 period. In some places, the two versions differ in what type of possessive is used. So, for example, at line 957 we have a DEP with a subject body possessum: (7.43) & breid Geomagog þat him þe rug for-berst and hugged Geomagog that him the back: broke ‘and hugged Geomagog so that his back broke’ Laʒamon’s Brut(C) BL 957 The Otho version has an IP here; breid Gemagog þat his rigge a two barst ‘hugged Geomagog so that his back broke in two’. A single example like this cannot be taken as evidence that the linguistically more progressive Otho version generally rejected this construction as old-fashioned, because there are other passages in which both versions use a DEP: (7.44) a. & sloh he him of þat hæued and struck he him off the head ‘and struck his head off ’ Laʒamon’s Brut(C) 1925 b. heo cærf him þene swure a-twa she cut him the: neck in.two ‘she cut his neck in half (i.e. beheaded him)’ Laʒamon’s Brut(C) 2002 Otho’s corresponding lines for (7.44a, b) are and swipte him of þat hefd and ȝeo carf him þane swere a-two. Note the retention of inflected accusative forms in both versions for the determiner of the body direct object swure/swere. Poetry is apt to be somewhat archaic in its language, and so it cannot be assumed that these DEPs represent the everyday speech of the time, but it would appear that DEPs were still at least understood around the middle of the thirteenth century. The fact that the modernizing Otho scribe was happy to use them also suggests that it was not terribly archaic. A complete reading of the poem The Owl and the Nightingale, another poem found in manuscripts of the second half of the thirteenth century, as described in Appendix A, turned up a dative possessor only in this ‘blended’ example: (7.45) Þet his nekke him that his neck him ‘so that his neck breaks!’

toberste break: (Owl and Nightingale 122, Atkins 1922)

In contrast, there is no dearth of IPs used for highly affected possessors of mind and body words in this poem, as is unsurprising at this late date.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.9  

183

The nearly complete absence of DEPs in the small m2 PPCME2 texts does not by itself indicate a significant change from the m1 period without a comparison of IPs in highly affecting situations. It turns out that with the direct arguments, the total number examples of expressed possessors of body possessa is so low that nothing can really be made of the fact that the search for DEPs only found one example with these verbs. The searches for IPs yielded only two examples with object possessa and five with subject possessa with the affecting verb list for the m2 period. I present one example of each in (7.46): (7.46) a. and harmeþ his bodi ‘and harms his body’ b. and my bones ben trubled. ‘and my bones are troubled’

(CMAYENBI,52.939) (CMEARLPS,34.1423)

In Section 7.9.2 we see, however, that when we add the data from PObj possessa, there can be no doubt that there was generally a significant change between the m1 and m2 periods. The PPCME2 searches did not yield any DEPs with direct argument possessa in the m3 period. The evidence suggests that the DEPs had a very tenuous existence at the beginning of the m2 period and had completely disappeared by the end of that period.

7.9 Prepositional phrases 7.9.1 Methodology One of the hypotheses mentioned in Chapter 1 concerning the loss of DEPs in Middle English was that they might have lingered longer with PObj possessa than with direct argument possessa. The present investigations into PObjs in Middle English were primarily geared towards testing this hypothesis. As with OE, most of the searches were restricted to a small number of prepositions to keep the data manageable; these are presented in Table B.10 in Appendix B. Because of the large size of the OE corpus, I restricted my searches for PObj possessa in that period to a small number of frequentlyoccurring body nouns in selected texts. The EME corpus is small enough that no text selection was necessary, and it became clear that most body nouns needed to be included in order to get sufficient data. Therefore, I restricted the prepositions as just described but searched for all body PObj possessa with the exception of hand and heart. The reason for the exclusion of these words was

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

184

     

the large number of non-target examples that searches with them returned; with hand a number of the examples were instrumental, and most of the heart examples were metaphorical and so come under the examination of mind rather than body words. However, I carried out a separate search for the prepositions on, in, and to with hand, since DEPs and IPs were in variation in expressions involving such combinations in OE, and the results of these are discussed separately, in section 7.9.3, after the results for the broader search are presented in section 7.9.2. I have separated my results for transitive and intransitive verbs. The restrictions just mentioned applied to the searches aimed at comparing DEPs and IPs. In order to capture all possible examples of DEPs, however, I made some searches that targeted all DEPs without restricting the prepositions. The first of these was a search in the m1 period for all sentences containing an OB2 plus any preposition with a body noun as an object. The second type of more general search was for NP-DPS, in both the m1 and the m2 periods, which searched for all NP-DPS, without restricting the search to particular nouns or prepositions. These searches both included heart and hand, which were excluded from the more restricted comparison-oriented searches.

7.9.2 Results As noted at the beginning of the discussion of PObjs, one hypothesis we might consider concerning the loss of DEPs in Middle English was that this was essentially a two-step process, with DEPs disappearing with direct object possessa earlier than with PObjs. Such a hypothesis seems, on the face of it, plausible for more than one reason. First, as already mentioned, some European languages, such as Norwegian, have the equivalent of DEPs with PObjs but not with subjects or objects. Second, what many linguists would treat as external possessors are still possible with thematic objects in English, e.g. she shot him through the thigh, and such use might make it possible for nonthematic objects as external possessors of PObjs. Finally, the majority of DEPs in OE involved PObjs, and it would not be surprising if some of the fixed expressions that were common in our OE texts survived well into Middle English. It is at least initially plausible that expressions like look/stare x in the face and look x in the eyes are relics of DEPs. It is suggested later in this section, however, that these expressions are probably not remnants of DEPs, that is, not relics of expressions using the dative case but rather remnants of transitive uses of the verbs look and stare.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.9  

185

The investigation of DEPs with PObjs in the m1 period of Middle English (1150–1250) failed to find any support for the idea of DEPs lingering longer with PObjs, at least as a productive construction as opposed to some fixed expressions. Table 7.2 IPs and DEPs with selected body PObj possessa, 1150–1250 Transitive

IP DEP Blended Total Intransitive

cmancriw-1.m1 17 cmancriw-2.m1 3 cmhali.m1 0 cmjulia.m1 2 cmkathe.m1 4 cmkentho.m1 2 cmlamb1.m1 2 cmmarga.m1 5 cmorm.po.m1 4 cmpeterb.m1 1 cmsawles.m1 2 cmvices1.m1 3 Vesp. A. xxii 0 Total 45

1 0 0 0 0 0 0 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0

18 3 0 2 4 2 2 5 4 1 2 3 0 46

IP DEP Blended Total

cmancriw-1.m1 10 cmancriw-2.m1 5 cmhali.m1 0 cmjulia.m1 5 cmkathe.m1 1 cmkentho.m1 0 cmlamb1.m1 1 cmmarga.m1 9 cmorm.po.m1 11 cmpeterb.m1 0 cmsawles.m1 0 cmvices1.m1 3 Vesp. A. xxii 2 Total 47

2 0 1 0 0 1 0 0 0 0 0 0 0 4

0 0 0 0 0 0 0 0 0 0 0 0 0 0

12 5 1 5 1 1 1 9 11 0 0 3 2 51

Table 7.2 indicates that if anything, DEPs were even less frequent with PObjs than with the direct arguments rather than the other way around. Twenty-one of the total of ninety-two examples with IPs should probably be discarded because the possessum is modified by an appositive adjective, which crosslinguistically tends to make DEPs impossible. This still leaves us with an enormous difference between IPs and DEPs. The fact that the searches did not include all prepositions means that we cannot make a categorical statement, but since these are the prepositions that were the most likely ones to occur with DEPs in OE, it seems clear enough that the DEP had become highly unusual with PObjs.¹⁴ It is also clear that the small number of DEPs cannot be attributed to a data gap caused by a possible infrequency of examples in which we might expect a DEP if one were possible. With direct arguments, the low number of examples of affected body possessors means that the small number of DEPs might possibly be attributed to such a data gap, but the higher

¹⁴ It may be worth noting also that my own extensive reading of EME texts, including ones not covered by the PPCME2, has done nothing to change my conclusion here.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

186

     

frequency of body PObjs (where affectedness had never been essential) rules out that argument for EME. The number of examples of DEPs of PObjs in the m1 period is so small that all examples can be presented. Three PObj examples that are tagged by the PPCME2 as NP-DPS have already been given above in (7.33) but are reproduced here for convenience, along with the two examples otherwise tagged which I consider to be DEPs of PObjs: (7.47) a. ant cwenene crune sette þe on heaued and queens:. crown set thee on head ‘and set a queens’ crown on your head’ (CMANCRIW-1,I.72.276) b. cruninge of þornes þe set him i þe heaued. crowning of thorns that sat him on the head ‘the crown of thorns that sat on his head’ (CMANCRIW-1,II.142.1909) c. þenne spit leccherie to scheome & to schendlac; then spits lecherie to shame and to disgrace meiðhad o þte nebbe. maidenhood in the face ‘Then lechery spits in maidenhood’s face, to shame and disgrace’ (CMHALI,139.167) d. Ðeos sæt wel þan Hælende æt foten & æt heafde This sat well the: Saviour: at feet and at head ‘This one sat truly at the Saviour’s feet and head’ (CMKENTHO,138.106) e. se muche se him eauer to muð kimeð. as much as him ever to mouth comes ‘as much as ever comes to his mouth’ (CMANCRIW-1,II.70.778) In three of the five examples presented, the DEP is in the form of a pronoun. In (7.47d), the DEP is a noun, but since this example comes from a text written in a case-rich dialect, the dative case of the noun is made clear from the form of the determiner. It therefore presents no challenge to the idea that DEPs only persisted for as long as the dative case was made clear. Example (7.47c), however, is of particular interest because the DEP meiðhad is not marked for case.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.9  

187

Restricting the prepositions was necessary for collecting data on how frequent IPs were in EME with PObjs, but because DEPs were so infrequent, it was possible to search for all body and mind DEPs with PObjs, using the searches described in section 7.9.1. Such a search was useful in confirming the infrequency of DEPs generally in EME. These searches revealed one more example of a DEP with a body PObj: (7.48) spit him amit þe beart to hoker & to scarn. spit him amid the beard to contempt and to scorn ‘Spit in his beard, to his contempt and scorn’ (CMANCRIW-1,II.215.3080) This search also found a ‘blended’ example, with the noun hand, which was not included in the more restrictive searches: (7.49) & himm wass sett inn hiss rihht hand An and him was set in his right hand a dere kineʒerrde; dear sceptre ‘and a precious sceptre was placed into his right hand’ (CMORM,I,284.2345) As with the direct arguments, this search produced no examples of DEPs with PObjs in the non-PPCME EME texts described in section 7.5.1, and the fact that several IPs with PObjs are to be found in the non-corpus texts adds to the conclusion DEPs were very infrequent with PObjs in the m1 period both in raw numbers and in comparison with IPs. It also seems clear that DEPs were associated with particular verbs at this time. Added together, the six examples presented in (7.47) and (7.48) and the one blended example (7.49) give a total of seven instances of a body PObj with an overt possessor that is not a simple IP. These seven examples all have one of four verbs: setten, sitten, spitten, and comen. A search for all m1 PObj body possessa with forms of these four found only seven examples of IPs, that is, the same number as examples with an element referring to the possessor that would have had dative marking in OE. I present two of these IPs in (7.50): (7.50) a. ant sette hire riht fot on his ruhe swire. and set her right foot on his shaggy neck ‘and set her right foot on his shaggy neck’ (CMMARGA,74.309)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

188

     

b. & sete iesu swucche swete sahen i mi muð and set Jesus such sweet sayings in my mouth to-marhen. tomorrow ‘And put, Jesus, such eloquent speech in my mouth tomorrow . . . ’ (CMKATHE,29.157) It seems, then, that with verbs that ever occurred with a DEP of a PObj, a DEP was about as equally likely as an IP, in these texts at least. It is worth noting also that in example (7.50a) the possessum is modified by an appositive adjective. As already discussed, studies of DEPs in other languages have found that they are not used when the possessum is modified by an appositive adjective, and so presumably an IP was the only option here. The searches of the PPCME2 texts of the m2 period found no examples of DEPs with PObjs. Unlike with the direct arguments, where the relevant number of examples makes the significance of the paucity of DEPs unclear, there can be no doubt that the lack of DEPs with PObj possessa is no accidental data gap; the m2 data contain no fewer than forty-nine IPs of PObjs (transitive and intransitive verbs combined). Although the PPCME2 m2 texts yielded no DEPs of PObjs, examples taken from the literature show that they persisted in some idioms up to the end of the Middle English period. These examples are quoted by Ahlgren (1946: §136) as late survivals of the dativus sympatheticus: (7.51) And sone fel him to þe fet and immediately fell him to the feet ‘and immediately fell to his feet’ (7.52) She falleth him to fote ‘She falls to his feet’

Havelock 615 Chaucer LGW 1314

The facts suggest that the disappearance of DEPs with PObjs began in late OE, when, as we have seen, they seem to have been fairly frequent but restricted to a fairly small number of combinations of verbs and prepositions. In the EME period, they became much more restricted but were apparently associated with particular verbs rather than only with specific idioms. By the beginning of the m2 period, the PObjs are only found in idioms.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.9  

189

7.9.3 On/in/to hand As discussed in section 7.9.1, hand was excluded from the general searches for IPs in PObjs but included in the search for DEPs. Hand presents some difficulties for a study of possession of body possessa beyond the fact that its frequent use in instrumental phrases results in a high the number of nontarget hits in a search for IPs. It has always been used metaphorically in a large number of idioms both with and without an expressed possessor, up to PDE in as have on hand and fall into the wrong hands, etc. Chapter 4 introduced expressions that have not persisted in English, including becuman to handa ‘come into possession’ and gan on hand ‘submit, surrender’. Because hand occurred in so many idioms with the prepositions on, in, and to in OE in the pre-m1 period, both with DEPs and IPs, I made a search specifically looking for DEPs and IPs with hand as the object of these prepositions. We saw in section 7.3.3 that DEPs as well as IPs continued to be used with hand in expressions of surrender and control in the pre-m1 period. Before looking at the question of the continued use of DEPs in such expressions in m1, we can observe that the old metaphorical meanings generally continued to be used with implicit possessors in the m1 period: (7.53) a. & hæfde hit þa on hande þre degas; and had it then in hand three days ‘and had it in his control for three days’ (CMPETERB,49.219) b. oc he wolde hauen baðe on hande, but he would have both in hand ‘but he would have both in his control’ (CMPETERB,49.212) We still have a an idiom with an implicit possessor expressing a different sort of control in sentences like we have the situation in hand. A metaphorical expression with an overt possessor that has survived into PDE is used in the problem’s in their hands now. Turning now to expressed possessors of hand as a PObj, the m1 search for DEPs with all prepositions described in section 7.9.1 turned up no examples of pure DEPs. The only example that could be construed as having a DEP is the ‘blended’ example that was presented as (7.49), in which the hand is literal, not

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

190

     

metaphorical. The search for on/in/to hand yielded two examples of IPs, one literal and one metaphorical: (7.56) a. and sette hes wel heiʒe on his fader swiðer hand, and set it very high on his father: right hand ‘and set it very high in his father’s right hand’ (CMVICES1,97.1152) b. & bitahte in his hond þe meske of and committed in his hands the honour of hire meiðhad her maidenhood ‘and committed into his hands the honoured state of her virginity’ (CMMARGA,56.18) I found no evidence for the use of DEPs either in literal uses or in idioms with on/in/to hand in the m2 PPCME2 texts or my non-PPCME2 m1 texts. In a pattern that has become familiar, however, it seems that the DEP continued to be used to some extent in poetry at a time when there is no evidence for it in prose texts. The Middle English Dictionary entry for hond(e) gives this example from The Owl and the Nightingale, a poem of EME composition discussed in section 7.8.2 and Appendix A: (7.57) Me þunch þat þu me gest an honde. me think that thou me goes in hand ‘It seems to me that you have delivered yourself into my control’ Owl and Nightengale 1651 My own examination did not turn up further examples in this poem, but other examples are to be found in Laʒamon’s Brut: (7.58) Alle heo eoden an honde þan kinge Gurmunde. All they went in hand the: king: Gurmund: ‘They all submitted to Gurmund’ Brut(C) 14454 As one might expect, these examples with a DEP in the Middle English Dictionary are outnumbered by the examples with an IP, even in the poetry. To sum up, the available prose texts of the m1 period provide no evidence that the DEP survived in EME in those metaphorical constructions with hand in which it had continued to appear with some, albeit limited, frequency in Late West Saxon. However, poetic texts show continued limited use of DEPs in idioms well into the ME period.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.9  

191

7.9.4 Survivals with look and stare It is natural to assume that expressions like he looked me in the eye and it’s staring you in the face were relics of DEPs from the period when they were productive, and that we would find a reasonable amount of evidence for these constructions in Middle English. It comes as a surprise, therefore, that the investigation did not find a single example of look or stare used with an external possessor in any period investigated, although both verbs date back to OE. I suggest that rather than being relics of DEPs, these expressions are relics of transitive uses of look and stare that developed in a later period. We can still use these verbs transitively in expressions like he looked me up and down and she stared him into silence, in which the object is not a possessor of a body possessum. In earlier English, both look and stare were used more widely, in different senses, in transitive constructions. So for example, the OED entry for look, v, sense 3b gives transitive examples with a meaning ‘take care of, watch over’, but the first examples are from the twelfth century, and examples do not seem to become common until the next century. Example (7.59) illustrates this sense: (7.59) þe helþe of þine bodye þou sselt loki. the health of thy body thou shalt look ‘You shall look after the health of your body’ (CMAYENBI,54.978) Of most relevance here is the fact that the OED only offers one OE example of its transitive sense 2a, ‘to direct one’s sight at; to look at’, and this example is a gloss from the Lindisfarne Gospels. Here, the verb is used as a translation of the Latin considerate and the syntax of the verb is likely to have been stretched somewhat to keep close to the Latin. The next examples listed by the OED are not until the fourteenth century, and this use does not seem to be common until later: (7.60) Þat no body look my wryghtyngys 1471 J. Paston in Paston Lett. & Papers (2004) I. 566 The OED’s sense 2b of look says that (7.61) is the first attested example of look a person in the face, which appeared earlier than look a person in the eye. (7.61) She dorste neuere after þat tym Loken Adam in þe face c1400 (▸1375) Canticum Creatione l. 378 in C. Horstmann Sammlung Altengl. Legenden (1878) 129 (MED)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

192

     

The OED’s first example of look a given horse in the mouth is from 1546. It seems likely that the use of the object of look as an external possessor is a development from the transitive uses rather than a direct retention from OE. The transitive uses of stare all seem to date from later than the EME period, and in fact the OED only gives a small number of examples of the verb in any meaning in OE. I have noticed an OE example, not mentioned by the OED, of starian used with a dative object indicating the focus of the gaze: (7.62) þæt hire an dæges15 eagum starede that her: on day: eyes: looked ‘that should look at her openly’

Beowulf 1935

(7.62) of course does not contain a DEP, since eigum is not the object of the gaze, but rather an instrumental use of eyes. This example, with its dative object hire, suggests that there might be more ‘semi-transitive’ (dative object) or transitive examples than the OED’s entry would suggest. Be that as it may, there is no evidence to support a direct descent of stare x in the eyes/face from OE. In its list of phrases involving the verb stare, the OED lists (7.63) as the first attested instance of stare (a person) in the face: (7.63) Whan my fader gyueth me mete She wolde theron that I were cheke And stareth me in the face. 1510–13 Mery Geste of Frere & Boye sig. A.iij., (OED stare, Phrases) As with look, the expressions with stare using an external possessor plausibly developed from other transitive uses of the verb that are attested before the external possessor is. The texts give no support at all for the idea that they are relics of DEPs. I have limited my investigation of relics in Modern English to these two verbs that are still used in idioms, but Ahlgren (1946: §§137–9) gives examples from earlier periods with some other verbs that are no longer grammatical, laugh someone in the face, used by Shakespeare and Marlowe. This extinct idiom contrasts with transitive uses in which the PObj is not a body word, e.g. laugh someone out of their bad mood. Ahlgren gives an example of transitive use of laugh in which the verb is not accompanied by a PP complement: (7.64) They’l laugh you, Sir, and find fault, and censure things . . . Villiers, The Rehearsal p. 39 (Ahlgren p. 218)

¹⁵ An dæges literally meant ‘by day’; here it means ‘openly’.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.10 : 

193

This play is from 1671. This example raises the possibility that like stare and look, the verb laugh could be used transitively for a period.¹⁶ If so, laugh someone in the face might not be a relic from an earlier period but an expression that arose from this transitive use. Further investigation of the origins of what Ahlgren quite reasonably assumed to be relics of the dativus sympatheticus of an earlier period to see whether they might be expressions that arose in Modern English would be useful.

7.10 Results: mind For the mind words, which are listed in Appendix B, there is no need to have a separate section for the direct arguments, for the simple reason that the investigation found no DEPs of the selected mind words in either the subject or object role in the m1 period. In contrast, it is not difficult to find IPs, including with affected possessors: (7.65) a. hit walde to swiðe hurten ower heorte it would too much hurt your heart ‘It would greatly hurt your heart’ (CMANCRIW-1,I.46.65) b. Þe þridde freotet his heorte of sar grome the third eats his heart from painful anger ‘The third, he eats his heart out of painful anger’ (CMANCRIW-1,II.174.2422) c. þe þonckes þe prokieð þin heorte. the thoughts that assail thy heart ‘the thoughts that assail your heart’

(CMHALI,133.73)

With PObjs of the same selected prepositions as used for OE, DEPs are unusual and apparently restricted to fixed expressions. The searches of the m1 texts of the PPCME2 and the non-PPCME2 texts only found one example of a DEP of a mind PObj:

¹⁶ I cannot find evidence for such a transitive use in the Oxford English Dictionary entry for laugh, v. The OED does document a transitive uses of the verb with a cognate object and objects that express the content of something uttered with a laugh (e.g. laughed his disdain), but it does not record example (7.64) or any other example in which a person is the direct object.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

194

     

(7.66) oðer ma oðer lees, as ow on heorte bereð, either more or less as you: in heart bears ‘either more or less, as the spirit moves you’ (CMANCRIW-1,I.64.215) The status of this example is somewhat uncertain. The headword beoren in Richard Dance’s glossary to Millett’s (2005) edition of the Ancrene Wisse, a different version of this text, lists as ~ ow/hire on heorte as having the meaning ‘as it carries you/her in your/her heart, i.e. as the spirit moves you/her’. The expression could be treated as transitive, with an unexpressed subject and an implicit possessor. The Middle English Dictionary entry for bēren (v. (1)) ‘bear’ gives a much later example with an expressed dummy subject and an IP in its sense 5b ~ in herte, meaning ‘seem (to s.b.) in his heart’: (7.67) & in his herte it him bar þat þei nere a-boute no good ‘and it seemed to him in his heart that they were not up to any good’ c1330 Otuel (Auch) 176 It is clear that with mind PObjs, DEPs were no longer productive. Nevertheless, their survival in fixed expressions is suggested by the single example caught in the search of m2 texts: (7.68) þanne him comþ a zorʒe to þe herte then him comes a sorrow to the heart ‘Then a sorrow comes to his heart’ (CMAYENBI,27.423) The Middle English Dictionary also gives evidence for the continuation of DEPs with the old expression cuman on mode ‘come to someone’s mind’ in the Brut: (7.69) Hit com him on mode he wolde of Engle þa it came him in mind he would of English the æðelæn tellen. nobles tell ‘It occurred to him that he would tell of the nobility of the English’ Brut(C) 6 The Otho manuscript also has the same DEP here, although the rest of the line is worded differently. We saw in Chapter 6 that the other situation in which datives continued to be found with mind PObjs in late OE was with copulas, often in combination with an IP. An example from the pre-m1 period was presented as (7.19), but

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.11 :    

195

none appear in the m1 texts. Given the combination of the apparent decline of copula datives with adjectives by the m1 period and the low frequency of mind PObj adjuncts to such constructions by late OE, the absence of examples in m1 is not surprising, although the smaller corpus size makes it impossible to be certain that this complete absence of a construction that was already uncommon at an earlier stage is not a data gap.

7.11 Conclusion: from OE to ME Let us now compare the findings for direct argument body possessa in OE and the m1 period. Tables 7.3 and 7.4 present the figures for objects and subjects, respectively. The figures do not give a picture of a sudden disappearance or even a steep decline of DEPs in the m1 period. Rather, the most noticeable decline is in the late OE period. Mustanoja’s (1960: 98–9) comment that there was a steady decline of the construction in Middle English is not borne out by a more systematic investigation, which finds that the gradual decline is more of a Table 7.3 Comparison of OE and m1 body object possessa with affecting verbs Text category

IP

DEP

Total

% DEP

Poetry EWS Other Early 9thC(OE) General OE LWS LWS(Late) m1

11 9 0 22 45 80 29 22

7 12 2 13 11 4 3 9

18 21 2 35 56 84 32 31

39 57 (100) 37 20 5 9 29

Table 7.4 Comparison of OE and m1 body subject possessa, affecting verbs Text category

IP

DEP

Total

% DEP

Poetry Other Early EWS 9thC(OE) General OE LWS LWS(Late) m1

6 1 4 8 24 39 2 19

5 1 4 17 18 5 4 2

11 2 8 25 42 44 6 21

45 (50) (50) 68 43 11 (67) 10

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

196

     

feature of Old than Early Middle English. The use of DEPs became increasingly restricted to highly affected possessors in the OE period, and it is clear that this restriction holds in the m1 period. DEPs seem to continue to have some productivity in the m1 period with direct arguments, but they are found with only a small number of verbs. When we turn to the objects of prepositions, the loss of DEPs is more dramatic. The widely held perception of a sudden disappearance of DEPs in EME is probably due to this sudden reduction in PObjs, the situation where DEPs had been most numerous in OE. Given that fixed expressions with PObjs such as stare x in the face and look x in the mouth look like remnants of DEPs, one might assume that DEPs lingered longer with PObjs than with the direct arguments. The facts presented in this chapter disprove that hypothesis, and the facts discussed in section 7.9.4 support the idea that these two expressions are not in fact survivals of the DEP but rather relics of an earlier transitive use of look and stare. With the mind words, the DEP was already very restricted with the direct arguments in all OE prose and seems to have disappeared by the m1 period. With PObjs, it was found reasonably frequently even in late OE prose, but almost entirely in expressions that were presumably entered into the lexicon. By the m1 period, even these fixed expressions had become difficult to find. The data from the m1 period do not supply evidence for a sudden disappearance of DEPs as a grammatical possibility in the earliest period of EME. Although the prose corpus from the m2 period is very small, the nearly complete lack of DEPs in this period suggests that the change to the grammar that made DEPs generally impossible to generate occurred sometime between the m1 and m2 periods. The DEP was always a marked way of expressing possession, more limited than the IP in being used only with inalienable possessions even in its broadest use. I suggest that the disappearance of the DEP began with a decline in the early OE period, when the range of body DEPS with direct arguments became restricted to situations in which the possessor was adversely affected. At this stage, DEPs can be analysed as dative objects that could generally be added to the usual argument structure of a verb to express affectedness of an argument of that verb. This dative was construed as the possessor of the subject or thematic object of the sentence, but only when the possessor could be seen as adversely affected. With PObjs, the addition of a dative had no such restriction. I suggest that with PObjs, the dative was a non-thematic argument linked to the possessor, rather than an argument with a semantic role. The difference, however it is to be captured formally, might stem from the different origins of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

7.11 :    

197

DEPs with the direct arguments and PObjs. With body concepts, this addition of a non-thematic object was a productive although increasingly specialized way to express possession. With mind words, the ability to add a non-thematic dative object as the possessor of a subject or object disappeared very early, being pretty much restricted to poetry, and with PObjs it is almost entirely restricted to fixed expressions all through the OE period. As OE wore on, the addition of both types of datives became lexically restricted to certain verbs and combinations of verbs and prepositions, rather than a generally available option. By the end of the m1 period, a change to the grammar seems to have taken place by which dative arguments could no longer be added even to specific lexical items. While the timing of this change does not coincide with the timing of the loss of the morphological dative/accusative distinction, it does fit well with the disappearance of bare NPs that would have been marked dative in various functions other than indirect object, where the dative is not an addition to the normal argument frame of a verb but rather a part of it. In the m1 period, for example, there are still examples tagged as NP-DPS in the PPCME2, e.g. the ones of (7.32), but no examples of NP-DPS are to be found in the m2 texts of that corpus. Bare NPs are still found in old dative functions in fairly late poetry such as the Brut and The Owl and the Nightingale, the same works in which DEPs are found late. The general disappearance of applied objects by m2 would account for the rather abrupt disappearance of DEPs with PObjs at the same time as the less sudden disappearance with direct arguments. Idioms such as look x in the eyes/ mouth and stare x in the face seem not to be survivors that are exceptions to this general disappearance but rather relics of later transitive constructions. If this is correct, then their survival into Modern English represents the very limited re-appearance of non-thematic objects in the language when the simple transitive use of these verbs disappeared. We saw in Chapter 4 that there is no increase in blended dative and IP constructions with body possessa during the OE period. However, there are a surprising number of blended examples in EME given the small total number of relevant examples involving body possessa. This suggests that the increase in IPs did not take place by the addition of IPs to DEPs. Rather, DEPs, which I am assuming to be datives that are explicitly linked to the possessor function within another NP, simply declined in the OE period. The blended examples of the m1 and m2 period are probably to be taken as augmentation of an IP by an NP by a dative argument that showed some empathy with the possessor, rather than the augmentation of a DEP with a possessor word at a time

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

198

     

when an unambiguous expression of the possessor had become overwhelmingly the norm. With an empirical base for evaluating hypotheses concerning the causes of the disappearance of the DEP in English now established, Chapter 8 examines the most common suggested explanations for this development against the evidence of the texts.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8 Changes and explanations 8.1 Introduction The explanations that have been offered to date for the loss of external possessors in English mostly share the assumption that the disappearance of the DEP, at least from the texts, was a rather sudden development of the EME period. I have given evidence in the preceding chapters that although the disappearance of the DEP does in fact date to that period, it was not an abrupt change, since the decline of the construction was well under way by in the OE period. In this chapter, I assess the explanations that have been proposed so far against the new evidence provided by the present investigation and offer my own observations. Suggested causes for the loss of DEPs fall into two broad types: internal and external. Internal explanations are ones that treat the change as one that is due to purely internal workings in the grammar. The leading possibility for an internal trigger is the loss of the dative/accusative distinction. This sort of explanation can be incorporated in a contemporary Generative Grammar analysis by assuming the loss in the grammar of a mechanism assigning dative case to the possessor. Some other internal considerations that have been proposed as important are discussed in section 8.2. External explanations link the loss of DEPs to contact with languages that lacked such constructions. Contact also plays an indirect role in internal explanations. For example, although Ahlgren (1946) assumed that the loss of the dative/accusative distinction was the main cause of the loss of the ‘sympathetic dative’, he also assumed that contact with Scandinavian speakers was instrumental in the loss of that distinction. This external influence on the loss of case marking fits well into contemporary Generative Grammar, where language learners are assumed to be the locus of syntactic change as they use Universal Grammar to construct an internalized grammar from the Primary Linguistic Data to which they are exposed. Contact with speakers of other languages and especially speakers who have learned English as a second language would change the language learners’ Primary Linguistic Data, including the evidence for the old case-marking system, causing the loss of this Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

200

  

system. Van Coetsem (2000) also emphasizes that internally induced change (such as changes trigged by the loss of inflection) can be induced by contact. In the discussion of external explanations in this chapter, however, the focus of the discussion of the role of contact will not be on this as a possible cause of the loss of the case-marking system but rather on contact explanations that see language contact as having a more direct effect on DEPs. The chapter is organized as follows. I discuss proposed internal explanations in section 8.2 and then external ones in section 8.3. In section 8.4 I summarize the discussion and propose my own view of the loss of the DEP.

8.2 Internal explanations 8.2.1 Loss of case marking One widely held assumption is that since the type of external possessors (EPs) typical of most European languages, i.e. DEPs, involve the dative case, the ability of a given language to have EPs depends on the existence of the dative as a case category in that language. By this assumption, the loss of the dative case in English would naturally cause the loss of the DEP. However, we saw in Chapter 7 that external possessors similar to the old DEPs but lacking distinctive dative case marking lasted in the texts after the dative had disappeared as a morphological case. One problem with assessing the role of case loss in the disappearance of the DEP is that we need to distinguish case morphology from cases as category distinctions. Some contemporary syntactic analyses assume abstract cases that are distinguished only by position, with no overt morphology to distinguish one case from another. Such radically abstract case is different from the sort of abstract case that must be assumed for OE, where, for example, we assume that some instances of the masculine singular noun cyning encode nominative case while others encode accusative because that morphological distinction clearly exists for other nouns, and we decide which case a given instance of cyning is by its role as subject or object and any agreement with modifiers such as determiners that marked case morphologically. Arguments for radically abstract case, on the other hand, must depend on very subtle arguments that are highly theory-dependent. Here, I will make the assumption that is shared by many syntacticians that the dative case continued as a morphological category in English as long as some element continued to show distinctive dative and accusative morphology. In English, the element that maintained this distinction the longest was the masculine third person singular pronoun.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

201

Hine was the old accusative form, and I will assume that the writers of texts who restricted this form to marking the direct object still had the dative/ accusative as distinct case categories. I argue in Allen (1995) that the encroachment of him into the old accusative territory of marking direct objects is not a sign of case confusion, as long as hine is restricted to its traditional use, but rather an indication that him now encodes both accusative and dative case, while hine encodes only accusative. In the following discussion I will not assume case categories that are never evident from morphology, and will consider that the dative/accusative distinction has been lost in a particular dialect when any systematic distinction between the two has disappeared. Ahlgren (1946: §131) identified the loss of the dative/accusative distinction as the main cause of the loss of the dativus sympatheticus. The reasoning is that once the distinction had become completely levelled, it was not possible to distinguish an object as dative without either a fixed position before a direct object or the addition of a preposition. The modern generative treatment of Ahlgren’s idea is that the possessor is no longer able to get dative case assigned/checked. Before looking at how Ahlgren’s hypothesis squares with the OE and EME facts, let us consider the general plausibility of the idea that the loss of a morphological dative/accusative distinction should render DEPs impossible. To start with, we need to consider whether there a correlation between languages that have a morphological dative/accusative distinction and languages that have a DEP. If such a distinction is found in all languages that have what looks like a DEP, that is, an EP in the form of a bare noun, as opposed to one introduced by a locative preposition, for example, then it seems that the loss of the distinction in English would indeed make the DEP impossible. The typological literature makes it clear that no simple distinction between the existence of DEPs and a morphologically distinct case in a language. First, it is clear that a language that formerly had a DEP as a possibility can lose the construction even while maintaining a clear dative/accusative morphological distinction. As has been pointed out by König and Haspelmath (1998: 583) and McWhorter (2002: 226), this is what has happened in Icelandic. Old Norse, the ancestor of the Scandinavian languages, had DEPs, as in (8.1): (8.1) a. hrafnar skulu þer . . . slíta sjónir ór ravens shall thee: tear eyes: out ‘ravens shall tear your eyes out’ (Fj. 45,1 as cited in Havers 1911: 268)

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

202

  

Although modern Icelandic differs from the continental Scandinavian languages in retaining a distinct dative case, it is similar to them in having lost the DEP and adopting the locative possessive construction found in the other Scandinavian languages.¹ This construction has already been illustrated for modern Norwegian in (1.19), and another example is given in (8.2a). An Icelandic example is presented in (8.2b): (8.2) a. Han ville slå nesen flat på NN he would beat nose flat on NN ‘he wanted to beat NN’s nose flat’ (Norwegian, Lødrup 2009b ex. 24) b. Han nuddaði á henni fætur-na he massaged to her: feet:. ‘he massaged her feet’ (Icelandic, König and Haspelmath (1998: 559, ex. 76b)) The locative construction is not innovative in the Scandinavian languages, since it was already found in Old Norse alongside the DEP: (8.3) hvé þar á Herkju hendr sviþnuþu how there on Herkja hands burned ‘How Herkja’s hands burned there’ (Gþr. III 10, 2 as cited in Havers 1911: 269) If we did not have Icelandic, we would probably assume that the choice of the locative construction was forced on the continental Scandinavian languages when they lost the dative/accusative distinction, but Icelandic shows that the DEP could simply be given up in favour of another way of expressing inalienable possession. Of course, the loss of the DEP in the Scandinavian languages is not an exact parallel with its loss in English, because the Scandinavian languages had the locative construction as an alternative, while English never developed one, despite its contact with Scandinavian languages. Icelandic does show, however, that a language that had a DEP could give it up for reasons other than the impossibility of maintaining the construction because of deflexion. It is clear that a robust dative/accusative distinction does not necessarily result in the maintenance of the DEP, but it does not follow that a DEP can be maintained without such a distinction. A morphologically distinct dative case might be a necessary but not sufficient condition for a language to ¹ Thráinsson (2007: 95–6) notes that the ‘possessive dative’ is not completely impossible in Icelandic, but is ‘quite formal, literary, or even poetic’ and is furthermore limited to the objects of prepositions.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

203

maintain a DEP. This leads us to the question of whether any languages have a DEP without a dative/accusative distinction. Two Germanic languages appear to fit this description. The first language is Norwegian. As we have seen, Norwegian makes use of a locative construction for possessed body parts, and it is a language that Haspelmath (1999) and König and Haspelmath (1998) place outside the area of Europe of languages with DEPs. However, Lødrup (2009a) shows that Norwegian has an EP construction in which the possessor is a bare NP.² This is limited to ‘unergative verbs’ (that is, intransitive verbs with an agentive subject) with a PObj possessum: (8.4) Han tråkket henne på he stepped her on ‘He stepped on her feet’

føttene feet: (Lødrup 2009a ex. 13)

Lødrup’s Lexical-Functional Grammar analysis of this construction is that unergative verbs can license a non-thematic object which is linked to the possessor role within a PP. With such an approach, no case is needed. The second language that lacks a morphological dative/accusative distinction but has been argued to have a bare EP (our DEP) is Dutch. McWhorter (2002: 226) argues against the traditional case loss explanation for the demise of EPs in English on the basis that Dutch, which underwent a similar loss of case distinctions, still has such EPs. However, Filppula (2008: 33) objects that Dutch is not a terribly good counterexample because the construction is quite limited in that language. It might therefore be taken as support for the hypothesis, rather than evidence against it, although some explanation would need to be given for the fact that some DEPs are possible in Dutch, as discussed by Vandeweghe (1987), who gives the examples in (8.5): (8.5) a. Anita wierp mij een vaas Anita threw me a vase ‘Anita threw a vase at my head’

naar het hoofd at the head (Vandeweghe 1987 example 3a)

b. Ze hebben hem de duimen geplet They have him the thumbs crushed ‘They have crushed his thumbs’ (Vandeweghe 1987 ex. 33b) ² It should be noted that Lødrup (2009a: 423–4, 429) explicitly rejects the analysis of sentences like (8.4) as involving dative external possessors on the basis that the object is non-thematic, while a DEP is ‘often’ assumed to have a semantic relation with the verb, such as benefactive or malefactive. However, I do not consider that Lødrup convincingly demonstrates that the Norwegian construction lacks the sort of ‘affectedness condition’ that is typically found with DEPs, and I consider that the construction shows the continuation of what was a dative possessor as a bare NP, limited to possessors of PObjs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

204

  

Vandeweghe reports that in Dutch it is only when the body part is part of an ‘adverbial complement’, as with the prepositional phrase in example (8.5a), that the construction is at all productive. However, he says that the construction is found to a limited extent where there is no such complement, as in (8.5b). Haspelmath (1999) places Dutch on his map inside the area of Europe that contains languages that have a DEP (1999: 116), but puts English outside this area. Haspelmath considers that Dutch is a language that has lost the dative/ accusative distinction but still has DEPs (1999: 125), in contrast with English, which has lost both this case distinction and DEPs.³ König and Haspelmath’s (1998: 590) map of EP constructions in Europe puts both Dutch and English inside the area with EPs with a dative possessor. König and Haspelmath’s decision to include English and Dutch in the languages which have the construction is presumably due to the fact that relics of the construction are still found in English, and the construction with prepositional phrases can be considered reasonably productive in Standard Dutch.⁴ Whether one counts Standard Dutch or English as having DEPs or not seems to depend on whether one counts a language as having DEPs when it is fairly restricted, not using the construction frequently with direct object possessa, but using it productively with PObj possessa. The problem of whether to treat Dutch as a language with a productive DEP is complicated by the fact that there are many varieties of Dutch. Fortunately for the examination of the link between DEPs and a morphological distinction between the dative and accusative cases, Scholten (2018) shows that while these varieties are the same in not having this morphological distinction, they are not uniform in their use of DEPs (or as Scholten prefers to call them, ‘indirect object possessors). In particular, there are dialects that allow DEPs with some productivity, not just in fixed expressions, with direct object possessa. Scholten says that example (8.6) is acceptable in some dialects: (8.6) Margriet het höm vanmörn d’handn mit zaalve Margriet has him this.morning the.hands with cream behandeld treated ‘Margriet treated his hands with cream this morning’ (Scholten 2018: 258, example 77a)

³ Haspelmath says that it might be suggested that this is possible because English is generally more advanced on a path away from a case-marking language, but he does not elaborate on this observation. ⁴ Although König and Haspelmath place English inside the DEP are on this map, they comment elsewhere that English is a language that has only internal possessors (1998: 579).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

205

Interestingly, the dialects that allow these DEPs are ones along the German border, and Scholten considers that the proximity to German, a language with a distinct dative case and highly productive DEPs, is responsible for the acceptability of examples like (8.6) in these varieties of Dutch. These dialects have no morphological dative case, but Scholten has to assume that they ‘have some remainings of the dative case, even though this is not visible in the nominal paradigm’ (2018: 268). In other words, an abstract dative case remains after the morphological case disappears. To sum up, it is clear enough that the loss of an overt dative/accusative distinction does not necessarily make a DEP impossible to maintain. There is no reason in principle why English speakers could not continue to use a DEP when the case-marking distinction disappeared, and given the reduction in the frequency of the construction in OE, it could have disappeared simply because it became less and less favoured. However, the fact that the loss of a dative/ accusative distinction is not inevitably a death blow to a DEP does not necessarily mean it has no effect. It does not seem unlikely that with the loss of the dative as a distinct case, a DEP becomes a ‘high maintenance’ construction which is harder for language learners to learn and is most easily maintained in contact with a language that has both a DEP and a morphological dative, as has apparently happened with Dutch dialects. It appears also that DEPs of PObjs are generally easier to maintain after a loss of dative case than they are with direct object possessa; as noted by Vandeweghe (1987) as well as Scholten (2018), the former are possible even in Standard Dutch, and we also have the limited use of bare NP possessors of PObj possessa with ‘unergative’ verbs in Norwegian. Having argued that the loss of the dative/accusative distinction in English did not necessitate the loss bare EPs which would have formerly been in the dative case, I turn now to the question of how good the fit is between the timing of the loss of this case and the loss of DEPs. Ahlgren (1946) refers several times to a gradual reduction in the frequency of DEPs in the OE period, and so it is a bit puzzling that he refers in §131 to ‘the rapid disappearance of the Dativus Sympatheticus’. It appears from the discussion that follows this statement that although he regarded the decline of the DEP as gradual, he saw the disappearance in ME as rather abrupt. He makes it clear that he sees the inflection of pronouns as particularly important, both because pronouns were more likely to be used in the dative construction than nouns and because pronouns were the last bastion of the dative/accusative distinction, and comments earlier in §131 that it seems reasonable to suppose that ‘its [the Dativus Sympatheticus’/CLA] final fate was sealed by the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

206

  

levelling of the accusative and dative cases of the personal pronouns under one form’. It was demonstrated in Chapter 7, however, that the prediction that DEPs would disappear as soon as the him/hine distinction did does not hold. We find clear examples of DEPs, that is, of bare nouns referring to the possessor, in some case-impoverished texts of the period. Importantly, some of these examples cannot be dismissed as fixed expressions. Furthermore, we do not find a correlation between dialects having levelled inflection and dialects favouring the IP most strongly. Ahlgren (1946: 135) notes that ‘it would seem natural that the disappearance of the Dativus Sympatheticus in different dialects should have run parallel with the weakened feeling for the correct uses of the dative and accusative case-forms’. He offers no evidence that this was in fact the case, and immediately after this statement changes the topic to the likely role of Latin influence, and so he seems to have left this as an open question. The nature of the EME texts that have been left to us do not make it easy to investigate this question. If we had martyrologies, which generally are rich in examples of possessors of body words being severely affected, from both case-rich and case-impoverished dialects, we could expect to get a very clear picture of whether those dialects differed in their preference for IPs even with highly affecting verbs. We do not, alas, and it is particularly unfortunate that the case-rich EME texts that we have only yield a small number of relevant examples. However, we can at least say that there is no evidence to suggest that DEPs remained longer in the case-rich texts than in the case-impoverished ones. This becomes clear when we extract the figures for the case-rich texts of Tables 7.1 and 7.2, which present the total number of DEPs in the m1 corpus, with direct argument and selected PObj possessa, respectively, without distinguishing the state of case marking in the texts. Table 8.1 presents the figures for the case-rich texts from Tables 7.1 and 7.2. With the direct argument possessa, Table 8.1 presents the number of IPs with affecting verbs and all DEPs (which unsurprisingly are limited to affecting verbs with these objects). For PObj possessa, Table 8.1 gives the figures for all examples, including copular constructions. The figures in this table differ from the figures given in Table 7.2 in that the figures for IPs in Table 7.2 include some examples in which the possessum is modified by an appositive adjective, but the figures in Table 8.1 exclude these. When we restrict our attention to the direct arguments in these case-rich texts, we have very little data to work with, but we can at least say that it gives no support to the hypothesis of a dialect difference correlating with the state of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

207

Table 8.1 IPs and DEPs in EME case-rich texts

Objects

Subjects

PObjs

Text

IP

DEP

Blended

cokentho cmlamb1 covices1 Vespasian A. xxii Total cmkentho cmlamb1 covices1 Vespasian A. xxii Total cmkentho cmlamb1 covices1 Vespasian A. xxii Total PObj Grand total

1 0 1 0 2 0 0 1 0 1 2 3 6 2 13 16

0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

the dative/accusative distinction. Furthermore, it is striking that the eleven DEPs recorded in Table 7.1 come from texts in which the distinction is at best tenuous. The conclusion must be that the DEP was greatly reduced even in the texts that had essentially the same case system as in the late ninth century. It is true that the comparison of case-rich and case-impoverished texts is not a balanced one because the case-impoverished texts form a much smaller corpus and only offer a very small number of relevant examples, too small to draw any conclusions about preferred constructions with direct argument possessa in those texts. We can at least say from these figures, however, that there is no evidence to support the hypothesis of a correlation between case loss and a preference for IPs. When we add the data for PObj possessa to the data for the direct arguments, the evidence against a correlation between the maintenance of the dative/accusative morphological distinction and the maintenance of DEPs in the m1 period becomes considerably stronger: the number of relevant examples from the case-rich texts rises to nineteen and all but three of these examples have IPs. As far as we can tell from the texts, the low rate of DEPs in the EME period was similar in different dialects. Ahlgren (1946) seems to suggest that the beginning of the decline in the DEPs in OE may have started with the loss of the dative/accusative distinction in the first and second pronouns, although he is not really explicit on this point. He mentions the fact that the distinction between accusative mec, þec,

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

208

  

and dative me, þe was retained in the poetry, but not in the prose, and also that the dativus sympatheticus was very frequent in the poetry. However, the idea that the growing preference for the IP in OE was due to an inability of the syncretic forms to make the ‘datival’ feeling clear does not fit well with the facts. To look at it one way, the writers of the EME case-rich texts seem to have been just as reluctant to use DEPs as writers in case-impoverished dialects, even though they could signal the dative/accusative distinction as well in pronouns as ninth-century prose writers could. Looking at it from the other way, writers in case-impoverished dialects did not seem to feel that using old dative forms that were not distinguishable from accusative ones prevented those forms from being used occasionally in DEPs, and not just in fixed expressions. It is clear that the loss of the dative/accusative distinction did not make DEPs impossible, any more than it has in eastern and northern Dutch dialects. This is not to deny that the increasing syncretism of case markers within the OE period and especially in EME could have played any role at all in the loss of DEPs. Although this development does not coincide with the loss of the dative/accusative distinction, it does fit well with the timing of the loss of other bare NPs used in the old dative functions, and constructions using these bare NPs may have become less favoured and more difficult to maintain with the loss of the dative/accusative distinction. By the m2 period, both had lost any sort of productivity, with the exception, of course, of the old bare NP in the indirect object function, which still exists in the present day. The long shift to a system in which grammatical relations were signalled primarily by means other than case marking, such as word order, began in OE, where there was already considerable syncretism of case categories in the earliest texts and where we already find some reliance on default word orders, for example the first NP a sequence of two NPs phrases that were ambiguously marked for nominative or accusative case was always interpreted as the (nominative) subject. Through the OE and m1 periods, speakers and writers continued to have grammars that allowed both DEPs and IPs as well as bare NPs in all the old dative functions, but by the last half of the twelfth century there was an increasing disfavouring of all but the indirect object functions for the old bare datives in the case-rich as well as the case-impoverished texts.⁵ The disfavouring of DEPs that had already begun in the OE period finished with their ⁵ It is important to note that the word order of the bare datives in indirect object function did not result in a fixing of the order of two objects immediately or even soon after the loss of the dative/ accusative morphological distinction, as Ahlgren, following most scholars, assumed. The fixing of this order only happened much later, as discussed in Allen (1995).

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

209

complete loss as a productive construction around the same time as the loss of the other old bare datives.

8.2.2 Loss of expletive determiners Vergnaud and Zubizaretta (1992) argue that only languages that allow expletive determiners have DEPs. If this is true, then the loss of the DEP in English must be linked to the loss of expletive determiners. The loss of expletive determiners must itself be linked to the loss of agreement features, since they assume that what licenses expletive determiners in French is agreement in person, number, and gender. Vergnaud and Zubizaretta’s suggested correlation between inflection on the determiner and DEPs is based on a very small sample of languages, and Haspelmath (1999: 124) notes that a simple correlation between agreement features on the determiner does not account for the typological facts, since there are languages like Basque and others that have a ‘French-type EP construct’ but a genderless definite article. On cross-linguistic grounds, then, it is not a surprise to find out that the predicted correlation between dialects that have inflected determiners and ones that have an indeclinable determiner does not hold up. For example, the definite determiner is completely indeclinable in the Ormulum and the Second Continuation of the Peterborough Chronicle, but we find DEPs in both of these, and it cannot be said that DEPs are rarer in these case-impoverished texts than in the case-rich ones, although the data are too limited to make meaningful comparisons. Even with the problem of a small number of examples, it is still clear enough that the DEP was not limited to fixed expressions in the texts with an indeclinable determiner. The following discussion will look at whether there is any clear evidence for a loss of expletive determiners playing a role in the loss of DEPs. The discussion begins with a consideration of whether OE had a definite article, a matter of some dispute in the literature. Crisma (2011) argues convincingly that a definite article was established in English by the end of the ninth century, that is, in King Alfred’s time, and also that this represented a change in the grammar underlying the prose texts from that of the poetic texts. In addition to her demonstration that marking of the definite/indefinite distinction was not optional in the grammar underlying the prose, Crisma assumes that OE prose allowed expletive articles generally and uses this assumption in an interesting suggestion concerning the use of such denotationally empty articles in DEPs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

210

  

Specifically, example (8.7) is one of the small number of examples in Crisma’s substantial corpus that appear not to conform to the generalization that a singular definite count noun must be marked for definiteness:⁶ (8.7) to þære stowe læded wæs, þær him mon sceolde to the place led was where him: one should heafud ofslean head: off.cut ‘and was led to the place where he was to be beheaded’ (Bede_5:17.456.5.4579) (Crisma 2011 ex. 17) Crisma’s suggestion is that the definite article is an expletive one with DEPs, but no expletive is needed to identify the referent in such cases. It is not required by the grammar of OE because its only function in these possessives would be to mark the noun as a count noun, but OE did not require marking of the count distinction. On cross-linguistic grounds, this explanation seems to be weakened by Scholten’s observation that the definite article appears to be a pervasive feature of external possessor construction cross-linguistically (2018: 38). If the OE could leave the definite article out more freely in an inalienable construction because the identity of the referent was clear, why would that not happen more frequently across languages that otherwise use a definite article? Looking specifically at OE, it turns out that (8.7) is unusual in lacking a definite determiner. Crisma’s study was not aimed at collecting body parts, and her corpus design, although it yielded ample evidence for the obligatory (or nearly so) marking of definiteness in the prose, did not provide enough examples of possessed body parts to compare the appearance or not of definite determiners in DEPs.⁷ The present study, on the other hand, resulted in a large enough pool of examples to give a clear picture of the use of the definite determiners with body possessa. This study supports Crisma’s only in finding a sharp difference between prose and poetry in the use of the definite determiner generally; in poetic examples definite body words were frequently not marked ⁶ Crisma found three other instances of inalienably possessed noun, not presented in her paper, for which a similar explanation could be offered. One of these is my example (8.8). I am indebted to Paola Crisma for generously supplying me in personal communication with all the examples she found of definite body nouns in her corpus that were not marked for definiteness. The majority of this small list of examples fall under other explanations she suggested for some recognizable types involving nouns that did not require definiteness marking generally. ⁷ Because she was looking at all common nouns, it was impractical for Crisma to attempt to look at the enormous number of examples that would have resulted from a search for all these nouns. She therefore restricted her corpus to nouns beginning with the letter h, after testing to make sure that the proportion of examples with and without a determiner were similar in this more limited corpus to the entire YCOE.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

211

as such in DEPs, making them similar to other definite nouns in general in poetry. In the prose, however, it is clear that DEPs with body nouns quite regularly used a definite determiner when that noun must be construed as definite. This fact goes against Crisma’s suggested explanation; it appears that the definite determiner could not generally be left out in DEPs because the definiteness of the body part was clear without it. Some other explanation is needed to explain the lack of a definite article in (8.7) and the similar (8.8): (8.8) þæt him man heafod of aceorfe buton oðrum that him: one head: off cut: without other witum punishments ‘that he should be beheaded without any other punishments’ (coblick,LS_32_[PeterandPaul[BlHom_15]]:189.335.2461) An obvious suggestion is that these examples involve object incorporation, since in both examples the object is adjacent to the verb (counting the particle as part of the verb). Such an explanation is especially appealing given the presence in OE of the verb beheafdian ‘behead’, by which decapitation could be described in a single word. While an object incorporation is plausible, independent evidence that heafod of slean and heafod of aceorfan combination demonstrating the properties of object incorporation would need to be adduced. However, it can be noted that an object incorporation analysis is also the most obvious one with the verb beheawan ‘hew, cut’: (8.9) and he wæs þær heafde beheawen and he was there head:/ hewn ‘and he was there beheaded’ (comart2,Mart_2.1_[Herzfeld-Kotzor]:De13,A.29.319) The DOE lists this example, from a manuscript of the third quarter of the eleventh century, under its sense 2.b. heafde beheawan ‘to deprive of one’s head, behead’ for the lemma be-hēawan. With this verb, in the active construction the possessor is in the accusative, rather than the dative case, resulting in the nominative case in the passive of (8.9). The possessum, rather than the possessor, is in the dative/instrumental case. For this reason, as discussed in Chapter 3, as well as the late date of the manuscript, this example was not covered in my investigation of DEPs. It seems likely that definite determiners were not as obligatory with nouns used as instrumental/locative phrases as they were with accusative objects. If that is so, then this use with beheawan might have caused the occasional lack of a determiner with other verbs of beheading.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

212

  

It must finally be noted that no object incorporation explanation will explain the only other example I have found in OE prose of DEPs that lack a definite determiner where I would expect to find one, that is, with a definite singular count noun: (8.10) Gif men sie maga asurod if one: be stomach: soured ‘If a person’s stomach is soured’ (colaece,Lch_II_[3]:69.1.1.4104) With a single example, a number of possible explanations could be offered, with no way of testing them. In the case of (8.10), one possibility is that the scribe who was copying the text saw the form sie and because of its similarity with the (feminine) determiner seo failed to add the similar-looking (masculine) determiner se. Scribal error is always a possibility when we have an example that runs counter to such a robust generalization as the obligatoriness of definite articles with definite nouns in OE prose. Whatever the explanation(s) for (8.7)–(8.10), it is pretty clear that no explanation in terms of the optionality of expletive articles in DEPs in OE will explain the fact that DEPs were like other NPs in OE prose in essentially requiring overt marking of a definite singular count noun. This of course is a separate issue to the hypothesis linking the loss of DEPs to the loss of expletive articles in English, to which we can now turn. This hypothesis runs into the same timing problem as the case loss hypothesis. Let us suppose for the sake of argument that OE did in fact allow expletive articles. If these expletive articles were what made DEPs possible, the disappearance of DEPs in ME should correlate temporally with the loss of overt agreement on definite articles with the nouns they modify. As we have seen, however, there is no good correlation, since DEPs are found in texts completely lacking in inflection of the definite article. Furthermore, they are infrequent in all texts, including the case-rich ones. In the case-rich texts, the determiners maintain agreement inflection on the definite article as well as the accusative/dative distinction. Under the expletive article hypothesis, we have to assume that expletive articles were lost in the case-rich dialects at the same time as in the case-impoverished ones. Even with that assumption, the continued appearance of some DEPs in the m1 appearance, at a rate that is too high to be accounted for as simple mistakes, is problematic within Vergnaud and Zubizaretta’s (1992) parameter-setting approach, in which a language either does or does not allow expletive determiners.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.2  

213

8.2.3 Increased configurationality of NP Van de Velde and Lamiroy (2016) argue that the loss of the DEP in English is due to an increase in NP configurationality. Specifically, they say there is a link between the rise of IPs and the introduction of definite articles, and that the more grammaticalized the determiner slot, the less common the DEP is. Their argument for such a link is not a convincing one. They have to admit that this is not a law but only a robust tendency, and that the fact that French has a DEP at all with its highly grammaticalized articles and the fact that German has a robust DEP with its undoubted NP configurationality are at odds with this tendency. Nevertheless, they maintain that their main claim stands because if we look within the same family, the correlation between a grammaticalized determiner slot and the less common DEP holds. To support this correlation they argue that French is more reduced in its use of DEPs than Romance languages with less grammaticalized determiners and German has more DEPs than English or Dutch because it has a less grammaticalized definite article than either of those languages. A generalization over a few languages in a family cannot be used to maintain a generalization that should hold crosslinguistically across families. To add to the counterexamples that Van de Velde and Lamiroy have to explain away somehow, Scholten’s (2018) study of Dutch dialects seems to indicate that while the dialects differ in the use of definite articles with inalienable possession, the differences do not seem to correlate with a difference in the acceptability of DEPs. Turning specifically to English, the correlation between a more grammaticalized definite article and a preference for DEPs does not hold up very well. Van de Velde and Lamiroy say that it is a moot point precisely when English developed a definite article. I do not consider this to be so, because I believe that Crisma’s (2011) study and my own work (Allen 2016b) has firmly established that OE prose had a definite article. Regardless of whether we want to date this development to that period or to the EME period, however, the decrease in DEPs in the OE period does not correlate with any increase in article-like behaviour of the determiner in that period, either generally or specifically with DEPs. It is true that there is a change from poetry to prose in the behavior of definite determiners, which are optional in the poetry but very seldom missing with singular count nouns in even the earliest prose, and this change coincides with a significant decrease of DEPs of some types. However, this decrease in specific DEPs types is not plausibly linked to what can certainly be treated as an increased grammaticalization of the definite

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

214

  

determiner. A problem here is that this decrease in DEPs is not even. We see it early with mind subject possessa, and later with body PObj possessa. DEPs with mind object possessa were already nearly gone in the poetry, contrasting with body object possessa (sections 6.2.1 and 4.2.) Van de Velde and Lamiroy also cite the ability of genitives to separate from their head nouns as a characteristic of less configurational NPs that goes along with EPs. A problem here is that although OE allowed such separation fairly freely, there can be no doubt that the NP was highly configurational by generally accepted criteria, with designated slots for determiners and modifiers. A highly configurational language may still have a considerable amount of discontinuous constituents.⁸ Furthermore, discontinuous possessor phrases are still highly productive in late OE, even though they might be reduced in frequency, at a time when the DEP was greatly reduced. The association of IPs with the definite article or greater configurationality within the NP does not give a satisfying explanation for the decrease in the DEP within the OE period. However, it does appear that a greater reliance on configurationality in general in signalling grammatical relations played a role in the loss of the DEP as a grammatical option at the end of its decline.

8.3 External explanations The language-internal explanations that have been offered for why the DEP became impossible in ME run into problems with the timing of the loss of this possessor type and furthermore cannot explain the beginning of the decline of the DEP early in the OE period. This brings us to external explanations, which offer more hope of explaining the beginning of the process of loss. Three languages (or language families) have been proposed as contact languages that caused or at least contributed to the loss of the DEP: Latin, Norse (a term covering the various Scandinavian dialects), and Brythonic Celtic.

8.3.1 Latin While Ahlgren (1946) considered the loss of the dative/accusative distinction the primary cause of the disappearance of the dativus sympatheticus, he also ⁸ Van de Velde and Lamiroy do not clearly specify what criteria would make a language configuration enough to discourage the use of DEPs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.3  

215

saw Latin as playing an important role because he assumed that the populace would imitate the high status and very familiar language of the Church, meaning they would favour IPs over DEPs: The disuse of the Dativus Sympatheticus in favour of the construction with the poss. adj. was most probably promoted [Ahlgren’s emphasis/author] by the influence of the language of the Church, which was largely modelled on Latin usages. (Ahlgren 1946: 211)

There are two aspects of possible Latin influence that must be kept separate here. The first is the possible influence of Latin on the frequency of IPs in the texts. As argued in Chapter 4, it is possible that some specific sentences in close translations from Latin may have been influenced by a Latin exemplar, but generally translations from Latin do not depend heavily on the Latin in the way they express possession. The second and more important aspect of possible Latin influence on English is the likelihood that Church Latin would have had a decisive influence in an overwhelmingly illiterate population increasing the use of IPs in speech to the point where language learners ceased to learn that the DEP was a possibility in their language. Such an influence seems implausible on general sociolinguistic grounds. Apart from the improbability that the language of church services would have had such an impact on everyday language (as opposed to language written in an elevated register), the prestige of the language of the Vulgate has not prevented the retention of the ‘vulgar’ DEP in the modern Romance languages, as noted by Filppula et al. (2008: 34) (2008: 34). We can safely reject Latin influence as a significant driver in the loss of the DEP. The idea that our data from the texts might be influenced by translation effects by Latin has, at least initially, more plausibility. In principle, it is possible that IPs would be more frequent in the texts than they were in everyday spoken language by at least the Germanic population due to Latin models being seen as appropriate for the written language. I have shown in Chapter 4 that DEPs do not seem to have been considered inappropriate in even fairly close translations from Latin. Ahlgren furthermore seems to have exaggerated the avoidance of DEPs in the Latin used in the Vulgate and other medieval religious writings. DEPs can be found in these Latin texts, even if they are less common than IPs, and some of the DEPs in translations are in fact translations of a DEP in the Latin. Furthermore, the lack of any overtly expressed possessor is extremely common in the Latin texts that were translated into OE, and the translators did not generally imitate the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

216

  

Latin in these situations, but usually added an overt possessor, either as an IP or a DEP. To sum up, the expression of possession in OE texts was highly independent of the Latin models. At most, there is a possibility that in some specific examples in which the Latin had an IP, an OE translator may have followed the Latin in using an IP when a DEP would have been an alternative. This would be a matter of affecting the frequency of a grammatical alternative in OE rather than changing our picture of the grammatical possibilities of OE. Any influence of Latin in the eventual disappearance of the DEP in English can be discounted.

8.3.2 Norse From a sociolinguistic point of view, a profound influence on English syntax by the Scandinavian dialects that came into contact with English is highly plausible. These dialects were not yet separate languages, and I will refer to them collectively as ‘Norse’. There was surely a large amount of intermarriage, exposing the children of mixed marriages to both English and some variety of Norse. The similarities the two Germanic languages might have encouraged a blending of the two; it would not have been difficult for one language to incorporate features of the other. When it comes to the loss of DEPs, however, Norse influence simply does not square with the evidence. The most complete argument for contact with Norse as the cause of the loss of DEPs in English is found in McWhorter (2002). To McWhorter, the DEP is just one of a number of Germanic features which fell victim to the reduction of complexity caused by the contact between Norse and English after the invasions that started in a serious way in the ninth century. An obvious possible objection to this hypothesis is the fact that the Scandinavians acquiring English had an EP construction with a locative at their disposal, as discussed in section 8.2.1, and might have been expected to import it into English. McWhorter counters this objection with Thomason and Kaufman’s (1988: 129 and passim) argument that when a population shifts to a new language, incomplete acquisition may result in reduction rather than outright transfer. Importantly, these effects often occur even where the languages in question have parallel cognate structures. Since IPs were a possibility in both languages even with highly affecting verbs, it is not impossible that speakers of both English and Scandinavian origin would tend to simplify communication by favouring one common alternative that both of them shared.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.3  

217

McWhorter is quite right in presenting this suggestion as plausible. There is no a priori reason to assume that speakers in contact would have imported the Norse locative EP into English or even maintained the DEP that the two languages had in contact. They might well have simplified the system by jettisoning all but the unmarked expression of possession, the IP. Nevertheless, the results of this study make it clear that the contact between English and Norse could at most only have sped up a process which was well under way. The decline of DEPs had already begun in West Saxon, the dialect least influenced by contact with Norse, at a time before such contact could have affected the language generally. Furthermore, if this contact played a decisive role, we would expect to find evidence of a difference between dialects in this respect, given that there was no standardization of English in the EME period which might have caused a writer to select a construction used in a dialect other than his own. Although a more systematic investigation of more and less Scandinavianized varieties of EME is needed, preliminary investigations suggest that no such evidence will be forthcoming. As we have seen, DEPs are found in the highly Scandinavianized Ormulum and in the Second Continuation of the Peterborough Chronicle, which also shows considerable Scandinavian influence. The Norse hypothesis also gives no explanation for the fact that the southern texts, which show minimal effects of contact with Norse, have only a small number of DEPs, just as with more Scaninavianized texts. At best, contact with Norse might have furnished the death blow to a construction that was already disappearing from the language.

8.3.3 Celtic The final contact scenario that has been implicated in the loss of the DEP involves Celtic. Like English, the modern insular Celtic languages are unusual in Europe in lacking DEPs. This similarity seems to have been noticed first as a possible area of Celtic influence in English by Pokorny (1927–30) and has prompted more recent scholars including Vennemann (2002), Filppula (2008: 30–40), and Hickey (2012) to present the sole use of the IP in Modern English as probably or at least plausibly due to Celtic influence. Before discussing various variants of the ‘Celtic Hypothesis’ as explanations for the loss of the DEP in English, let us first consider whether it is plausible more generally that Celtic speakers could have influenced OE syntax, before turning specifically to a consideration of DEPs and IPs.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

218

  

To start with, we can note that the traditional view that Celtic could not have affected Old English materially because of the paucity of Celtic loan words is considered no longer tenable by most scholars. The Celts were a subject population, and invaders normally do not freely borrow words from the languages of those they have conquered. Van Coetsem (2000) presents an approach to the study of contact phenomena which allows for more than one scenario in which language contact may bring about syntactic and morphological change. The scenario that is most likely to result in substantial grammatical change is one involving rapid language shift. If a group of speakers shifts quickly from their native language to the language of a socially and politically dominant group, they might ‘impose’ the most stable elements of their own language (i.e. grammar, as opposed to vocabulary) in making this while successfully acquiring the vocabulary of the new language. In that way, they affect the grammar of the language they shift to without affecting its vocabulary greatly. We can only speculate on how rapid any language shift from Celtic to English would have been after the Anglo-Saxon invasion, although we can assume that the speed of the shift would have been different in different parts of the country. This shift could only have had a serious effect on OE if the number of Celts remaining in England was large, compared to the AngloSaxon population. We will never be certain of what the contact situation was like in this period, but we can at least note that recent archaeological and genetic studies have cast very serious doubt on the traditional view that the Celts mostly fled or were slaughtered. The view that relations between the Anglo-Saxons and the Celts were uniformly hostile has also been discredited by findings such as those of Schiffels et al. (2016), who present evidence from ten graves in three cemetery sites near Cambridge ‘even in the early AngloSaxon period for a genetically mixed but culturally Anglo-Saxon community, in contrast to claims for strong segregation between newcomers and indigenous people’. They found individuals with ‘native’ (Celtic) and immigrant ancestry being buried in the same way in the same cemetery, as well as evidence of genetic intermingling in the early Anglo-Saxon period (fifth to sixth century). Interestingly, the richest grave they studied was of an individual of ‘native British ancestry’, while the one grave without any grave goods was genetically ‘foreign’. Of course, we cannot assume that the relations between the two groups of people were uniform across the country. Grimmer (2007) argues that the fact that Ine’s laws (promulgated between c.688 and c.693) specify a special (lower) wergild for Britons shows that there was a significant population of Celts in

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.3  

219

this kingdom of Wessex. This difference in wergilds in Ine’s laws is often cited as evidence for the lower status of Celts in Ine’s period at least. However, the fact that one of these lower wergilds specified is for a British thane, a rank second only to the king, shows that matters were not as simple as Celts being relegated to a low caste. Grimmer (2007: 112) suggests that Ine may have had a deliberate policy of encouraging assimilation of his British subjects into Anglo-Saxon society; such a policy could be useful in the border areas where the king would not want Britons to maintain overly strong ties with their Welsh neighbours. Quick assimilation would have obvious benefits for the Britons, who might avoid using distinctively Celtic words which identified them as British, and they could easily pick up the OE lexicon while retaining some grammatical habits. By the time King Alfred promulgated his law code towards the end of the ninth century, no distinction between Briton and Anglo-Saxon is evident in the laws (Grimmer 2001: 114). Fairly rapid language shift by a significant part of the Celtic population is therefore not implausible. More gradual shifts, such as a long period of bilingualism, could also be expected to have some effects on OE syntax. Although we can only guess at how Celtic speakers shifted to English after the invasion, we certainly cannot reject the general possibility of this shift affecting OE syntax. Turning to a specific examination of the Celtic Hypothesis as it relates to DEPs and IPs, we can start by pointing out that we cannot assume simply from the modern situation that the Brythonic Celtic languages and dialects that the Anglo-Saxons came into contact with in England were similar to the modern languages in lacking DEPs. Examples of IPs given from Middle Welsh that do not predate the OE texts are not very convincing, and there is a general lack of earlier Brythonic Celtic writings that would allow us to confirm that only IPs were used in these languages at the time when they came into contact with English. However, the fact that Breton lacks external possessor (EP) constructions (König and Haspelmath 1998 and Haspelmath 1999) lends support to the view that EPs were not a feature of Insular Celtic. It is reasonable to assume that the Celts of the period of the Anglo-Saxon invasion did not have DEPs in their languages, and so this is an area where in theory Celtic speakers might have imposed their syntactic habits in the process of language shift. What would the mechanics of Celtic influence in the use of DEPs and IPs have been? Vennemann (2002) includes the loss of EPs as one of the examples of ‘Celtic syntax’ that arose in Middle English. His impression is that the IP construction is ‘relatively rare’ in earlier OE with affected possessors (2002: 212). Vennemann argues that the language of the lower strata

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

220

  

of society was ‘Anglo-Saxon English lexically but Celtic English structurally’, or Brittonic English and IPs were part of this Brittonic English. Vennemann’s explanation for the putative infrequency of IPs for affected possessors in the texts is that writing was controlled by an Anglo-Saxon elite who, like Modern Germans, used DEPs for such possessors. His main explanation for the fact that some IPs are nevertheless undeniably found in the OE texts is that social mobility, especially in the clergy, would have caused some Briticisms to surface in writing. In Vennemann’s view, it was not until the Norman Invasion allowed the language of the lower strata to be used in writing that the IP became frequent. Vennemann’s assumption that Celtic features would have been suppressed in writing echoes comments made by Tolkien (1963), and Tristram (2004) goes so far as to call the OE situation diglossic, with a high written English and a low spoken one. A more recent proponent of the idea that Celticisms were suppressed in the written language is Schrijver (2014). The assumption that the OE texts did not represent the OE spoken by the majority or at least a large section of the population has been used as an explanation for a difficulty with the Celtic Hypothesis, which is that some features that have been attributed to Celtic influence sometimes do not show up in the texts until Middle English. A prime example of such a feature is the use of periphrastic do, which has similarities to Welsh and does not appear in English text until EME. If this construction was used in ‘Brittonic’ English, it would be stigmatized and suppressed in writing by the Anglo-Saxon elite, according to this view, until there was no longer an Anglo-Saxon elite controlling the language. Whatever the merits of this scenario, which is of course unprovable however plausible it might be, it is important to realize that IPs were not in the same position as periphrastic do. If periphrastic do was alien to the grammar that the Anglo-Saxons brought to England with them but was a feature of the English used by Celts who had learned English, it would indeed be likely to be stigmatized as a Celticism. However, there is no reason to believe that the IP in expressions of inalienable possession was stigmatized in and considered unsuitable for writing, even for writing down the most Germanic of genres, alliterative poetry, where we find IPs expressing highly affected possessors of body parts. The conclusion that variation between DEPs and IPs in reports of similar events was an integral part of the grammar that Anglo-Saxons brought with them is inescapable. Havers (1911) presents evidence concerning variation in the other earlier Indo-European languages even with highly affected possessors, a variation that is discussed in more detail in section 5.2. Given the

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.3  

221

variation that we must assume in OE at the time of contact with the Celtic population in England, it is not very plausible that the use of an IP even in a situation where an Anglo-Saxon speaker might have been more likely to choose a DEP would even have been noticed. The idea that IPs would be stigmatized and considered unsuitable for written language when a DEP was an alternative is furthermore incompatible with the notion (suggested by Ahlgren 1949 and considered possible by Vennemann 2002) that Latin models promoted the use of IPs in such situations. Finally, the idea of IPs with affected possessors only bursting out into the level of writing when the West Saxon Standard was smashed becomes utterly untenable when we realize that Ælfric, who corrected what he considered to be grammatical errors in some of the manuscripts containing his work and presumably kept tight control of his scriptorium, not only accepted IPs with highly affected possessors but generally favoured them. However, it is not necessary to assume that the IP was a Brittonic feature that was suppressed in writing to assume Celtic influence on the decline of the DEP in the OE period. Filppula (2008), following the account of Ahlgren (1946), accepts that the gradual replacement of DEPs by IPs had already begun in Old English, and correctly concludes that there is thus no timing problem for the Celtic Hypothesis to overcome in this instance. Filppula considers that it is plausible that ‘Welsh usages which rely on the internal possessor construction have triggered and promoted the change from external to internal patterns in English’ and does not exclude the possibility of ‘mutually reinforcing adstratal influences’ in the centuries following initial contact, leading to the unusualness of both languages compared with other European languages (2008: 39). Although I agree with Filppula about the lack of a timing problem for the Celtic Hypothesis as it regards IPs, his (2008: 30) statement that ‘[i]n earlier English, by contrast [with Modern English/CLA], the external type was the prevailing one’ is at best an oversimplification. In the OE texts studied here, DEPs were not in a majority as the expression of an affected possessor of a body object even in the poetry, and in the small number of prose texts in which they were a majority, the majority was only a small one. They could only be called ‘prevailing’ with specific verbs and idioms. This situation is by no means incompatible with the idea that Celtic speakers learning English were at least partly responsible for the initial reduced frequency of the DEP, which led to further decline in OE. We will never be sure what happened when British speakers learned English, but we can compare two basic scenarios for how these speakers might have influenced English. The first one is that Celtic

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

222

  

speakers learned English so poorly that they failed to learn the DEP at all, and only used IPs in their English. If there were enough such speakers in the community, the DEP might well have died out in speech fairly quickly after the Anglo-Saxon Invasion. Given that what we see in OE is a reduction in the range of DEPs, I think a second scenario is more likely. By this scenario, the Celtic language learners did learn the construction, but they modified its use. We could expect that Celtic speakers would have used a lower frequency generally, favouring the use of the construction found in their own language. They may well have picked up on the fact that with direct arguments it was a marked, special use construction, with the result that in their English they reserved it for the situations that were at the core of its use, namely when there was some sort of effect on the possessor that elicited empathy. Since the effect that was most likely to call for empathy with the possessor was a negative one, they would have restricted the range of DEPs with direct arguments to negatively affected possessors. If the English of these speakers affected the wider community, and made its way into writing, it could account for the absence of DEPs representing beneficially affected possessors in most of the texts. No such restriction is evident in copular constructions and with PObj possessa, but this does not present a difficulty. No effects, either negative or beneficial, seem to have been prominent in these constructions in the early Germanic languages discussed in section 5.2, so we can assume that they would not have been in the English that Celtic learners were confronted with either. To conclude the discussion of possible Celtic influence, the idea that Celtic learners of English simply imposed the Insular Celtic IP-only system on English is not plausible, but the idea that these language learners affected English by using DEPs in a more restricted way fits well with the observed fact of a more limited use of the DEP in English than in Gothic or Old Saxon, which can reasonably be assumed to represent the situation in Common Germanic. Of course, the fact that a hypothesis is plausible does not prove that it is correct.

8.4 Conclusions 8.4.1 The beginning of the decline: why? At some point in ME, presumably near the transition from the m1 to m2 periods (i.e. late thirteenth century) English grammar underwent a change that made it no longer possible to generate DEPs. Before this change to grammatical possibilities, however, there was a long decline of the use of

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.4 

223

DEPs, and before discussing approaches to capturing the grammatical change, it is useful to say a bit more about the triggering of this decline. Of all the proposed explanations that have been addressed in this chapter, the hypothesis that Celtic speakers learning OE changed it by restricting the range of the DEP and rendering it an even more marked construction, beginning a decline that became greater as the OE period wore on, fits the facts best. However, it is important to realize that when two variant ways of expressing possession exist in a language, speakers may simply start favouring one construction over another. This is especially likely to happen when one of the constructions is already marked—speakers may make it even more marked, leading to its eventual loss. It may well be that Celtic learners of OE initiated the move towards restricting the DEP, but this is unprovable and there is no reason why OE speakers might not have done this without any outside impetus. The fact that the poetry is already more restricted in its use of the DEP with direct arguments than Gothic suggests that the decline of the DEP was under way very early. An influx of new speakers of English may well have accelerated the decline. We will simply never know. With variant constructions, the loss of one construction, which is what happened in English, is one possibility. Another possibility is that the constructions might became more differentiated when they are used. I believe that this is what happened in German, for example. Vennemann (2002), in his comments about affectedness quoted in section 1.2, is describing a situation in contemporary German that is clearly different from Old High German as described by Havers (1911). Havers’ examples show that both the IP and the DEP were used to express possessors who were clearly affected, but for speakers like Vennemann at least, the DEP seems to have become the only choice with such possessors. It is often assumed, e.g. by Vennemann, that not only the DEP but the way it was used in Modern German are inheritances from Common Germanic. They are not; OE was not just like Modern German, and German has sharpened the distinction between the IP and the DEP in its history. This sharpening of the difference between the two constructions in German seems to have been a general development of the ‘nuclear’ Indo-European languages, while in the more peripheral European languages of Indo-European descent, the DEP declined or was lost. König and Haspelmath (1998) and Haspelmath (1999) show that EP constructions can be regarded as an areal feature of European languages, independent of genetic relationships, since non-Indo-European languages in the European linguistic area have adopted this typically IndoEuropean construction. As Haspelmath (1999: 116) comments, ‘the languages

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

224

  

displaying the dative EP construction form a contiguous area in the centre and south of Europe’. The European languages which lack DEPs are not in the nucleus of the European Sprachbund.⁹ It is unsurprising that a peripheral language like English should fail to participate in a general strengthening of a distinction between two constructions, particularly when the ‘all purpose’ variant (the IP) was already gaining ground on the ‘special purpose’ DEP in a period before this continental strengthening began. The study of Dutch varieties by Scholten (2018) adds to the plausibility of a strengthening effect caused by continued contact with other languages that have a DEP, since the dialects along the German border have generally maintained a more productive use of DEPs than other varieties, even though they have the same loss of a dative/accusative distinction. English, even more than the southern and western dialects of Dutch, lacked such a strengthening influence. Studies of why English is different from other Germanic languages have tended to start from the point of view that it is English that has changed, that it has lost its Germanic inheritance, but it needs to be kept in mind that the other Germanic languages have not stood still either, and it must always be questioned whether a feature that sets English apart from these related languages is not at least partly due to an innovation, in this instance the increasing obligatory nature of the DEP with affected possessors, that has spread in the other languages because of their proximity to each other rather than simply a change that took place in English.

8.4.2 Grammatical change: stages Accounts of the loss of the DEP in English that seek explanations solely in developments of the EME period, such as the loss of case marking or contact with Norse, are doomed to failure. The beginning of the end of this construction belongs early in the OE period, and accounts of this development must start in that period. However, at some point in the EME period, English grammar changed in some way that made the generation of sentences like *she slapped him the face and *spit him in the beard impossible, not just infrequent, outside of poetry. I am assuming a general approach by which the DEPs in clauses with verbal predicates are objects with the grammatical

⁹ We can dismiss Van de Velde and Lamiroy’s (2016) objection that these non-Indo-European languages use a blended sort of construction, combining an internal possessive with the external dative. It is the addition of the dative that suggests the influence of European contact.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.4 

225

relation that I will refer to as OBJdat (dative object) that are added to the normal valence of a verb. With the direct arguments, the semantics of the verb determine whether such an addition can be made, since the added OBJdat must normally refer to an adversely affected animate being, usually a person. I assume a formal association of this OBJdat with the possessor of the body or mind word. In addition, I assume the occasional addition of a ‘sentence dative’ that involves no syntactic relationship of possession. With non-verbal predicates, the dative is associated with the copula in a lexical frame, and no such semantic restrictions apply, but the dative generally adds a locative meaning. Where the dative is licensed independently like this, I assume that the understood possessive relationship is not syntactic but understood for pragmatic reasons. With the PObjs, I assume that the added OBJdat is nonthematic, again with no semantic restrictions. This is similar to Lødrup’s (2009b) analysis of Norwegian analysis of ‘possessor raising’ with ‘unergative’ verbs, discussed in section 8.4.1, in assuming that a verb with a PObj can license a non-thematic object. The major difference is that for OE I am assuming that the addition of the dative does not formally relate the possessor to the possessum. Futhermore, I am assuming that the added object has the role of OBJdat, rather than an OBJ, with the result that DEPs of PObjs are allowed to occur with transitive as well as intransitive verbs. I make the usual generative assumption that grammatical changes to the core syntax of a language normally happen in the acquisition of that language in childhood as a child constructs a grammar induced from the Primary Linguistic Data to which he or she is exposed, although of course changes in the frequency of the use of a particular grammatical option by an individual may occur later. I assume that smaller changes to a person’s syntax may take place in a more superficial way as that person is exposed to a wider linguistic experience in speech or writing, and these later changes might explain the appearance of the occasional example of an old construction, such as a DEP in the m2 period and especially in poetry. Before the total disappearance of DEPs except in a few fixed expressions, we need to assume some less obvious changes to the grammar of English that made certain types of DEPs unavailable. In Table 8.2 I set out the stages that the DEP seems to have gone through on the way to its final disappearance. In terms of changes to the grammar, we can propose the following steps, starting with the basic assumption that DEPs represent the addition of OBJdat to the normal valence of a verb, except in copular constructions, where the dative is associated with the copula. In the earliest stage, this added OBJdat was required to represent an adversely affected participant, but no such restriction was

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

226

  

Table 8.2 Stages in the loss of the DEP in English

1.

2.

3.

4.

5.

Body

Mind

Poetry probably represents earliest OE. With direct arguments, DEPs already nearly completely restricted to affected possessors, except with non-verbal predicates, where they occasionally are used in pure descriptions. The restriction is a bit weaker with subject possessa generally. IPs at least as common for affected possessors of direct arguments. DEPs not restricted with PObjs, and are dominant Early prose (late 9thC composition). DEPs still productive with negatively affected possessors of direct arguments. IPs now dominant with PObjs except in some commonly recurring expressions

Poetry probably represents earliest OE. DEPs predominate with non-verbal predicates, but IPs predominate with verbal predicates. DEPs not limited to negative effect with verbal predicates, but not common with object possessa. DEPs predominate with PObjs

Later prose. Significant decline of DEPs of direct arguments. Limited use of DEPs with PObjs, probably lexically restricted m1 Limited productivity of DEPs with affecting verbs. DEPs of PObjs limited to fixed expressions m2 DEPs no longer part of English grammar, still found in limited use in poetry

DEPs disappear with object possessa, uncommon with subject possessa of verbal predicates. DEPs with non-verbal predicates disappear. IPs now predominate with PObjs, with DEPs limited to commonly occurring expressions No change from early prose

DEPs with direct arguments disappear. DEPs apparently restricted to fixed expressions with PObjs DEPs no longer a part of English grammar, except with very sporadic use in fixed expressions, nearly completely limited to poetry

present with the athematic OBJdat that was interpreted as the possessor of a PObj. The grammatical possibilities were not quite the same with body and mind words, a difference that could be explained by assuming that the OBJdat could be either benefactive or malefactive when a mind word was involved. There was also a difference in the frequency of the IP and DEP variants with these two types of inalienable possessa. We might assume that particular words or classes of words were lexically marked to occur in the DEP constructions. In the second stage, with the body words there is only a change of frequency, but with the mind words it looks like there was a more substantial change making them unavailable with verbal predicates, the very small number of examples with subject possessa perhaps being adult add-ons to a scribe’s repertoire. In the third stage, there is a change in the frequency of use of grammatical options, but probably no real change to what those options were.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

8.4 

227

In the fourth stage, the m1 period, mind words disappear from the list of inalienable possessa that can be used with external possessors. With the body words we can treat the disappearance of most DEPs with PObjs as a change to the grammar that no longer allows non-thematic arguments to be freely added, since it now seems to be possible to list the fixed expressions that occur with such DEPs in the lexicon. There is apparently no grammatical change with the direct arguments, where DEPs do not seem to be limited to fixed expressions. It is at the fifth stage that we get the major grammatical changes that neither the non-thematic OBJdat nor malefactive OBJdat can be added to the valence of a verb. These two changes are related, since they are both instances of increased restrictions on the linkage between the core argument structure of verbs and the grammatical relations occurring with that verb. This change was presumably towards the end of the m1 period. The paucity of examples of IPs of direct arguments in the prose in the m2 period makes the finding of only one DEP and one ‘blended’ example in that period incapable of showing a significant change from the m1 period. However, the substantial number of IPs with PObjs in the PPCME2 m2 texts, compared with the complete absence of DEPs or blended examples, suggests that the grammatical change happened towards the beginning of the m2 period (c.1250), since DEPs are found in some restricted expressions with PObjs in the m1 period. It is notable that as far as we can tell from the spotty witnesses to dialects that are left to us from this period, this change was not directly linked to the loss of the dative/accusative distinction or to language contact. However, the change does correspond well to a more general change in all dialects of a significantly reduced use of bare NPs in what were old dative functions other than their use as indirect objects.¹⁰ Indirect objects could still occur either before or after the direct object at this time, so it would not be correct to say that indirect objects had to be linked to a particular position. Some bare noun phrases, that is, NPs not preceded by a preposition, were still possible as adjuncts, as they still are in Modern English in NPs like that night, big time, etc. However, bare NPs were no longer used to treat a participant like an argument of the verb when it was thematically an argument of some other element. Bare NPs also became more restricted around this time when they were complements to elements other than the verb, such as adjectives.

¹⁰ Lamiroy (2003: 271) similarly notes that in Dutch and French, it is not just in possessives that the dative has retreated but also for ‘non-lexical’ datives generally.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

228

  

Accounts of the loss of DEPs in English that depend entirely on a triggering event in ME, such as the loss of the accusative/dative distinction, contact with Norse, or the sweeping away of the West Saxon Standard, cannot account for the observed fact of the decline in DEPs in OE, and that fact that they seem to have been more restricted in OE than what can be plausibly reconstructed for Common Germanic. Such accounts predict a sudden loss of the DEP after a period of high productivity. The account that I have outlined treats the loss of the DEP as an abrupt change in the way that any change to the grammatical possibilities of a language must be abrupt, but this was preceded by gradual changes that meant that the final change to the grammar would hardly have caused a noticeable difference between the usage of the generation before the change and the one after it.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

9 Conclusion

The primary goal of this book was an empirical one: to provide a firm evidentiary basis for discussions of the loss of dative external possessors in English. Previous studies of the ‘sympathetic dative’ in the history of English have made some fine insights and observations but were not based on the systematic collection of examples in a wide corpus that has become possible with electronic corpora and failed to make sufficient distinction between the use in poetry versus prose. These previous studies were also not informed by recent cross-linguistic information provided by linguistic typologists that help in organizing the data in the most illuminating ways. The studies that provide the best collections of data, Havers (1911) and Ahlgren (1946), do not organize their examples according to the grammatical relation of the possessum, a parameter that typologists have found to be important in the availability of external possessors. More recent discussions by linguists of the loss of DEPs in English have not offered new data to the major studies by traditional grammarians. A crucial distinction not made before this study is between direct argument possessa and PObj possessa. This study has provided strong support for the view that the decline of DEPs in English began in the OE period. Ahlgren (1946) pointed to a decline in the use of what he called the dativus sympatheticus and what I have been calling the DEP within the OE period, and Mustanoja (1960) speaks of a general decline in ME, but these conclusions are impressionistic, not being based on any statistical comparison of datives versus genitives/possessives to express possession. In this book, Ahlgren’s impression of a decline within the OE period has been confirmed with solid data, although his proposed explanation for the major cause of the decline, the loss of the dative/accusative distinction, has not. The data show that the commonly held view that, with affected possessors, DEPs were the dominant construction, compared with IPs with inalienable possessa in OE, cannot be maintained. The present study does not support this view in general terms, and it also makes it clear that such a broad statement is at best an overgeneralization of a complex situation, even in the earliest period Dative External Possessors in Early English. First edition. Cynthia L. Allen. © Cynthia L. Allen 2019. First published 2019 by Oxford University Press.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

230



of OE, where DEPs are the most common. Havers’ (1911) study established that the use of the dativus sympatheticus was not uniform with inalienable possessa of different semantic fields, and this study has shown that DEPs were never common with mind words in OE except in copular constructions in the poetry and some PObj possessors throughout the OE period. In contrast, IPs were common with these words even in poetry. With body words used in subject or object function, DEPs were nearly completely limited to adversely affected possessors, but there is no such general restriction with the mind possessa. DEPs were always more common with PObj possessa than with subject or object possessa, and it is probably their frequency with these possessa, coupled with the relatively high frequency of DEPs more generally in poetry, that has given rise to the view that the DEP was the dominant construction in OE. IPs were by no means unusual in the poetry, however, even with highly affected possessors. The organization of the data according to the grammatical relation of the possessum has yielded the information that DEPs were generally more common representing possessors of subjects than of objects. This difference is likely to be due to pragmatic factors, but I have not attempted more than a couple of comments on this subject and further investigation of the factors accounting for this difference should be rewarding. The separating out of PObjs from the direct arguments has been particularly fruitful. It turns out that with subjects and especially objects, DEPs with lexical verbs were almost completely restricted to adversely affected possessors, an effect that becomes more noticeable within the OE period. With the direct arguments, the IP was the default possessive construction with both body subject and object possessa, with the DEP almost completely reserved for reporting a situation with a negative effect on the possessor. No such restriction held for PObjs, a fact that I suggested may be due to the typically locative semantics of the prepositions studied as opposed to the origins of DEPs as goals with the direct arguments. However the difference between PObjs and the direct arguments is to be explained, it is an important fact that IPs could be found in all situations where DEPs could be used, including in poetry, which can be reasonably considered to preserve some old syntactic features of the language. The distinctions made in this study in periods of texts, combined with the organization of examples by grammatical category of the possessa, provides a more nuanced account than has previously been possible in the broad-brush picture of a gradual decline of DEPs. The increasing restrictions on DEPs are clear even between poetry and the earliest prose, and the decrease of DEPs in one dialect, West Saxon, can be quantified. One rather surprising finding is

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi



231

that DEPs with PObjs participated in this decline within the OE period. DEPs were always more frequent with PObjs than with the direct arguments in OE. DEPs had always been associated with a small number of ‘affecting’ verbs with the direct arguments, but in earlier OE had occurred more freely with PObjs, although even with PObjs an IP had always been more common. One might expect, therefore, that DEPs continued to be used with PObjs longer than with direct arguments into ME, but that is not what this investigation found. With the restricted list of prepositions used in this study, at least, DEPs with PObjs were mostly restricted to a small number of expressions that could be encoded in the lexicon. It appears that whatever change to the grammar made the productive use of DEPs with direct argument possessa no longer possible encompassed PObj possessa as well. As well as a difference in the frequency of DEPs in poetry and prose, there appears to be a significant difference between the poetry and the prose in the structure of a direct argument NP containing a body possessum. In the poetry, no determiner was needed to give an NP a definite interpretation generally, and this is naturally also true in DEPs. Ahlgren (1946: §129) comments that the definite article ‘rarely preceded the noun of possession’ in the ‘oldest English’. In the prose even of the earliest period, however, a determiner seems to have been obligatory, or at least very nearly so, with a singular count noun body possessum in a direct argument position. This finding adds to the evidence provided by Crisma (2011) for the development of a definite article in English by the end of the ninth century, at a time when this article did not have a distinct form from the distal demonstrative. I only found a few apparent counterexamples in the prose of DEPs of singular count nouns in the subject or object role which lacked a determiner modifying that possessum. With DEPs of PObjs, on the other hand, a determiner might be used but was not obligatory. I have only briefly touched on this matter in section 8.2.2, because the main focus of this book is on the loss of the DEP rather than on the internal structure of the possessa of different types, but the use of the definite article in DEPs warrants further investigation. Further investigation is also needed into implicit possessors, in which the possessor is understood but not expressed by a dedicated possessive element. Although I have not made a systematic study of implicit possessors, they appear to be much more restricted in prose of all types, where examples of types that do not occur in Modern English are unusual, compared with the frequent appearance in the poetry of examples like Hond up abræd Geata dryhten, ‘the Lord of the Geats raised hand’. The fact that implicit possessors appear to have declined sharply at a time when DEPs were still robust seems

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

232



to go against an attempt to link the two constructions, and no obvious explanation occurs to me for why implicit possessors should have declined at about the same time that the definite article became obligatory in DEPs in the situations discussed above. One finding that may seem rather surprising is that the data do not support the notion that the ‘blended’ construction, combining a ‘sympathetic dative’ and a possessive, generally increased as the DEP declined. Ahlgren (1946: §130) does not explicitly state that the blended construction increased through OE, but his statement that we find blendings of the dativus sympatheticus and the ‘poss.adj.’ constructions during the process of the poss. adj. taking over the functions of the dativus sympatheticus certainly suggests that conclusion. The findings of the study show that with body words, the blended construction is infrequent in all periods, and no more frequent in prose than in poetry. It is only with mind PObjs in copular constructions that a combination of a dative and a genitive/possessive is at all common, and here it is in the poetry, where we find the most examples of the dative by itself, that we find the most examples of a combination of dative and genitive/possessive. If we treat the blendings as an addition of an IP to a DEP, these facts seem rather puzzling. When we consider that the dative in the copular constructions are part of the argument frame of the copula, however, the facts make sense, as discussed in Chapter 6. In looking for explanations for the loss of the DEP in English, we must start in the OE period. Whatever change took place in ME to English grammar that made it impossible to generate DEPs only came at the end of decreased usage of this grammatical option. The evidence suggests that although the decline of the DEP as used in prose was gradual overall, the final disappearance of the construction seems to have been fairly abrupt, as far as we can tell from the small and not terribly satisfactory corpus for the m2 period. A secondary goal of this research was to evaluate the explanations that have been offered for the demise of the DEP in English in the light of the evidence provided by the texts. This goal was addressed in Chapter 8, which looked at the most popular explanations for both the decline of the DEP in OE and its loss in ME. Proposed explanations can be broadly categorized as either structural or contact explanations, although the two types of explanations interact. Ahlgren (1946) saw the loss of the dative/accusative distinction as the main reason for the decline of the use of the DEP, and this remains the most widely held structural explanation. This explanation does not fare well in the face of both the reduction of DEPs even in the case-rich dialects maintaining this distinction and the retention, albeit at a low rate, of DEPs in case-impoverished

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi



233

texts of the m1 period. However, while there is no good correlation between the loss of this morphological distinction and the loss of DEPs, there is a good fit between the disappearance of the DEP as a grammatical option and the general narrowing of the range of dative functions by the m2 period, even in dialects in which the dative/accusative distinction clearly remained as a grammatical distinction. Both case-impoverished and case-rich dialects in EME moved to a system in which the old dative functions were pretty much restricted to the indirect objects of ditransitive verbs. I have suggested that the change that took place that finally killed off the DEP was a general change to the grammar, apparently in all dialects, removing the ability to add to the normal valence of a verb either a non-thematic object or an additional argument bearing a semantic role previously signalled by the dative case. A second contender for a structural explanation, the loss of expletive determiners in English, also offers no explanation for a decline in the DEP when this construction was still clearly a productive grammatical option and suffers from a similar problem in the timing of this development and the loss of inflection that is supposed to have triggered it. Non-structural explanations for the loss of the DEP see language contact as the primary cause of the change. Of these proposed explanations, the ones suggesting Latin influence and contact with Scandinavian speakers can be dismissed as not adequately accounting for the facts. Translations from Latin may have added somewhat to the number of IPs in some texts, but the expression of possession is an area where translations from Latin in general diverge markedly from their exemplars. It is not plausible that Church Latin would have influenced popular language in the way that Ahlgren (1946) suggested. The timing of the decline of the DEP rules out contact with Scandinavian speakers as a cause of that change, and the existence of DEPs even in highly Scandinavianized texts in the m1 period makes it unlikely that such contact even played a role in killing off a minority construction. Contact with Celtic speakers has some plausibility as an explanation for the increased use of IPs in OE compared with what we can reconstruct of Common Germanic. The version of the ‘Celtic Hypothesis’ as key in the loss of the DEP that best fits the facts is not one in which Celtic speakers shifting to English simply imposed their grammar of only genitives for marking a possessor on English, failing to learn the DEP at all. Such a scenario does not account for the gradual decline of the DEP in OE or the regularity of the way the DEP was restricted with the direct arguments of lexical verbs as opposed to its freer use in copular constructions and with PObj possessa. Rather, the version that works best is one in which the Celtic speakers learning English

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

234



learned the DEP but restricted its range, leading to a decreased use of the DEP and an increase in the IP. One important point regarding the Celtic Hypothesis is that if Celtic influence was the cause of the decline of the DEP, it was influence that showed up early in the OE texts. Advocates of the Celtic Hypothesis explain the fact that supposedly Celtic-influenced constructions like periphrastic do were part of the spoken language but were suppressed in writing by the Anglo-Saxon elite who controlled writing in OE, only appearing when the West Saxon Standard had been destroyed. As Filppula (2008) notes, however, in the case of the IP there is no timing problem to explain, since it is not the case that IPs only become established in the texts in the EME period. The IP is different from periphrastic do in that it was always a possibility in Germanic, and if it was the case that a higher frequency of IPs was characteristic of Celticinfluenced English, this is not something that it likely to have been stigmatized and suppressed in writing. There is no reason to think that OE texts were not a reasonably accurate reflection of the language in the use of DEPs, whatever the situation might have been with periphrastic do. While contact with Celtic languages is not implausible as at least playing a role in the decline of the DEP in OE, it is not provable either. I believe that the reduction in the use of the DEP is the sort of thing that actually does not require an explanation beyond the fact that when two constructions are in variation, that variation may remain stable, but it is also possible that speakers may start to favour one variant over another. It is important to realize that all the Germanic languages, not just English, have changed with respect to DEPs. The Scandinavian languages have for the most part lost the DEP, even Icelandic, in which the dative/accusative distinction is well maintained. It is true that these languages have retained an inalienable possessive construction using locative constructions, but the point remains that they have lost a construction found in all early Germanic languages. Modern German has changed too, sharpening the distinction between the DEP and the IP. Havers (1911) demonstrated that the two constructions were in variation in pretty much the same circumstances in earlier German, with the use of the DEP to add some sort of emotional colouring an option but not obligatory. In Modern German, however, there is less choice. Van de Velde and Lamiroy (2016) argue that the DEP is receding in present-day French in comparison with Spanish. The DEP, being in variation with the IP, is the sort of construction that can wax and wane for any number of reasons. It seems likely that the peripheral position of English and the Scandinavian languages was important in the loss of the DEP in those languages at a time

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi



235

when languages in a more central position of the European Sprachbund were strengthening the distinction between DEPs and IPs. The idea of contact as an influence in preserving DEPs gains strength from the fact that Dutch dialects in contact with German are the ones that have preserved the DEP best. Van de Velde and Lamiroy (2016) object to the idea that the peripheral geographic position of English is responsible for its loss of the DEP. They point out that while the DEP is losing ground in French, which is in the nuclear area of the European linguistic area, Spanish is not experiencing a similar development, despite being only in the core area. The flaw in this argument is that such a development in French in recent times says nothing about the spread of the DEP in earlier periods. There is no reason why the DEP might not recede in some language after an earlier period of expansion. While we know a good deal about the use of the DEP in the oldest European languages from Havers’ monumental study and we know a great deal about DEPs in the modern European languages from recent studies by linguists, we need more work on the earlier languages incorporating recent advances in the understanding of human languages. Most of all, we need studies of not just the earliest stages for which there is evidence but for periods in between those early stages and the modern period. Only with such studies will we understand, for example, how Norwegian lost the DEP generally but still has a limited sort of ‘possessor raising’ as described by Lødrup (2009a), or how the Scandinavian languages lost the DEP. More work needs to be done on nonstandard dialects also, as our present ideas about correlations between DEPs and other linguistic features as well as possible effects of language contact on DEPs are based mainly on a small number of standard dialects. Much work remains to be done concerning changes to the expression of possession in the European languages generally. I hope that the empirical study offered here will play a role in a deepening understanding of this area of diachronic English syntax, as well as adding a bit to our knowledge of the possibilities of syntactic change.

OUP CORRECTED PROOF – FINAL, 3/9/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

APPENDIX A

Corpus As discussed in Chapter 3, section 3.2, the investigation into Old English relied primarily on data gathered from queries run on selected texts of the YCOE and the York Poetry Corpus using CorpusSearch 2 (Randall et al. 2005–13) but was supplemented by some poetry not included in the York Poetry Corpus, as detailed in that section. The searches were applied to all files from the York Poetry Corpus, but the examination of prose excluded some of the YCOE files. First, a small number of texts that come from xi (3rd quarter) or later according to Ker (1957) were excluded from the systematic data collection for OE. These are texts that are probably of earlier composition of unknown time, such as comart2. On the other hand, the investigation included texts thought to be composed by Wulfstan that are in manuscripts dated by Ker as xi (3rd quarter) in the LWS(Late) category because we have a good idea of when the texts were composed and also because they form a valuable addition in somewhat lessening the dominance by the works of Ælfric in evaluating Late West Saxon syntax. The YCOE also includes some texts found in very late manuscripts. I consider that such manuscripts belong to the EME period rather than the OE period, although they may contain texts of OE composition. This means, for example, that the data do not include figures from two major ‘Alfredian’ texts, namely the Soliloquies of St Augustine and Boethius. These texts are of Early West Saxon composition, but the manuscript containing the Soliloquies is dated to the mid-twelfth century by Ker (1957: item 215), and large parts of Boethius are only found in a manuscript that its latest editors, Godden and Irvine (2009), place at the very end of the eleventh century or the beginning of the twelfth. These manuscripts are too far removed from the original composition to be reliable testaments to Early West Saxon syntax.i Parts of Boethius come from a much earlier (tenth-century) manuscript, but this was badly damaged by fire. Because it is not a simple matter to disentangle the parts from the two manuscripts, I have simply excluded this text from the systematic data collection. I have however run my most important searches on them to see if they yield any interesting results or add significantly to the amount of data. As it turns out, they contain very few examples of possessed body parts. Table A.1 sets out the YCOE files used in the investigation of OE. Note that some of the text names used by the YCOE include extensions that indicate the period assigned to the text by the Helsinki Corpus, e.g. o23, while texts not included in the Helsinki Corpus are not given such extensions in the YCOE.ii These extensions sometimes give a somewhat misleading indication of when the text was composed and are not included in the titles used for the texts in this book. In addition to the YCOE files, the investigation covered two

i It may seem overly fussy to exclude late copies, but occasionally, when we have both an earlier and a later copy of a text, it is possible to demonstrate that a scribe has changed the syntax in an important way, as discussed in Allen (1997). ii This means that two instances of codocu2 and codocu3 are listed in the table, because these are separate files with different temporal extensions in the YCOE.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

238

 

Table A.1 OE prose texts used, by text type EWS

LWS

cochronA cocura cocuraC coorosiu coprefcura

AEhomsA coaelive_mod cobenrul cocathom1 cocathom2 colaw6atr coprefcath1 coprefcath2 copreflives cotempo cowsgosp

9thC(OE)

Other_early

cogregdC cogregdH colaece colawaf colawafint

bede(early) codocu1 codocu2 codocu2 Mart 4 (Sisam) Mart 3 (Sweet)

General OE bede(xi) coalex coapollo coblick cobyrhtf cochristoph codocu3 codocu3 codocu4 coeuphr coeust coherbar colacnu colaw1cn colaw2cn

colaw5atr comargaT comart1 comart3 comarvel comary conicodD coquadru cosevensl cosolsat2 coverhom coverhomE coverhomL ASC(post-EWS)

LWS(Late) AEhomsB Aelfric_vincent cocanedgD cocanedgX coepigen coexodusP cogenesiC coinspolD coinspolX

colsigewZ colwgeat colwsigeT colwsigeXa colwstan1 colwstan2 cootest coprefgen cowulf

short prose texts not included in YCOE; they are listed in italics in Table A.1. These are both fragments of early martyrologies and are discussed under the Other Early heading below. The text categories are outlined in section 3.2. The following discussion explains the text categories in more detail and some of the manuscripts containing the texts, where this information may be useful.

A.1 EWS In her 1980 edition of the Orosius, Janet Bately states that our idea Early West Saxon is based on four manuscripts (1980: xxxix), of which the so-called ‘Lauderdale’ MS, the base manuscript for her edition, is one. The context of Bately’s remarks was a discussion of the language of the manuscripts containing the Orosius, and so she does not identify the other three manuscripts, but they are set out in Campbell (1959: §16). It is worth describing all the texts included in the EWS category. Besides the Lauderdale MS, Early West Saxon is identified with two manuscripts containing the Pastoral Care and part of the so-called Parker MS. The Parker MS, which

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

 

239

contains the oldest version of the Anglo-Saxon Chronicle (ASC), known as A, is Cambridge Corpus Christ College 173, ff. 1–56. It is dated s.ix/x–xi² by Ker (1957: item 39, art. 1). The entirety of Version A is contained in the YCOE file cochronA. It is the annals up to the year 924iii that are considered to be evidence for early West Saxon. The annals in this version of the ASC end with 1070, which takes it into the period that I have designated as ‘pre-m1’. In her examination of the phonology, as indicated by the spellings, of three different groups of hands—1 and 2, 3–6, and the post-Conquest hands— Bately says that ‘[t]he language of the sections attributed to hands 1 and 2 is with a few exceptions that considered typical of early West Saxon’ while ‘[s]ubsequent pre-Conquest hands have features that are typically late West Saxon’ and the post-Conquest hands are mostly similar to the ‘subsequent pre-Conquest hands’ but unsurprisingly show a greater frequency of typically late features and lack early features (1980: cxxxii, cxlii, cxlvi). Since we are comparing frequencies of variant possession types in early versus late West Saxon, we do not want to count any examples from the later hands as early West Saxon, and so I counted only the annals written in hands 1 and 2 in the statistics for cochron A in the EWS category. This is made easier by the fact that the YCOE has included information about the hand of each sentence in its citations. It turns out that examples of the constructions under consideration are so unusual in the ASC, especially in the later hands, which are sparser, that the exclusion of the annals after 924 does not materially affect the results. Examples that come from other entries in YCOE’s cochronA file are treated as General OE examples if they are from annals of before 1050, or pre-m1 examples if they are later, as discussed below. The ‘anchor’ EWS manuscripts include two manuscripts containing the OE translation of Gregory’s Cura Pastoralis (CP, called the Pastoral Care in English). The two manuscripts are Oxford, Bodleian Hatton 20 (Ker 1957: item 324) and British Library, Cotton Tiberius B.xi (Ker 1957 item 195). Both manuscripts belong to the end of the ninth or the early tenth century. The Tiberius manuscript was nearly completely destroyed by fire, but it was carefully copied by Junius before this event. The Hatton MS is the basis for the YCOE file cocura, which contains most of the CP. However, it is defective in part of one section, and this is replaced from Junius’ copy of the Tiberius manuscript in the file cocuraC. The YCOE compilers have separated out prefaces, and the preface to CP is contained in coprefcura. British Library, Additional MS 47967, usually known as the Lauderdale MS, is the base manuscript for Bately’s (1986) edition of the OE Orosius. Bately agrees with Ker’s date of x¹ for this manuscript (Ker 1957: item 133). Bately’s edition is the basis of YCOE’s file coorosiu. A complicating wrinkle is that the Lauderdale manuscript is defective, and in this place Bately has used British Library, Cotton Tiberius B.i, which Ker (1957: item 191) places in the first half of the eleventh century. As it happens, the replaced pages contain no relevant examples, making it unnecessary to worry about whether any examples should really not be counted as EWS.

A.2 9thC(OE) The general nature of this category, which comprises texts of known early composition found in manuscripts of a substantially later date but within the OE period, is discussed in iii Note that the numbering is slightly different in Bately’s (1986) edition. 924 is the date given in Plummer’s (1892) edition of the ASC, which is the edition used by the YCOE as the basis for its file cochronA. This annal is dated 900 in Bately’s (1986) edition.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

240

 

section 3.2. Here, it is necessary to make some remarks about the OE versions of Gregory’s Dialogues. It should be noted first of all that there are two distinct OE versions of the Dialogues. While the original version was a translation made by Bishop Wærferth of Worcester in King Alfred’s time (between the early 870s and early 890s), a revision was made later. None of the surviving manuscripts containing either the original version or the revised version is close to the time of Wærferth’s composition, although they fall within the OE period. The categorization into periods of the two texts included in the YCOE and used in this study, known as C and H, is rather problematic. The C manuscript (Corpus Christi College, Cambridge MS 322) contains the original version, but Ker (1957: item 60) dates it in the second half of the eleventh century. The H manuscript (Bodleian Library MS Hatton 76) is earlier, from the first quarter of the eleventh century according to Ker (1957: item 328.A), but it is a (partial) copy of the revision, which according to Yerkes (1982:10) was made sometime between 950 and 1050, that is, possibly as much as 175 years after the original composition. Thus, the inclusion of C in the 9thC(OE) category is open to dispute because of the lateness of the manuscript, while H’s status as a composition of the ninth century can be disputed because any parts of it that are not the same as C were obviously not composed in the ninth century. However, it seems likely that as far as the use of DEPs and IPs is concerned, both versions are reasonably accurate reflections of the earliest version. For one thing, when the two versions share sentences essentially conveying the same information, they generally have no clear difference in expression of the possessor, although they may differ in the choice of verb: & (A.1) a. & þæt he gedyde, þæt eall his andwlita awannode and that he caused that all his face became.livid and asweoll. swelled ‘and he brought it about that all his face became livid and swelled’ (cogregdC,GD_1_[C]:2.20.26.203) b. & oð þæt eall his andwlita wearð toswollen & and until that all his face became swollen and awannod. become.livid ‘and until his face became all swollen and livid’ (cogregdH,GD_1_[H]:2.20.25.176) Furthermore, an earlier manuscript containing a nearly complete version of the original composition, known as O (British Library Cotton Otho c. i), is available for comparison with C. Ker (1957: item 182) dates this manuscript at the beginning of the eleventh century. It is not included in the YCOE, as Hecht (1965) only printed the C and H versions, giving variant readings from O in footnotes. By comparing the O variants supplied by Hecht, it becomes clear that C and O do not differ significantly in their use of IPs and DEPS, and so it seems safe to treat C as reflecting ninth-century language, at least insofar as DEPs and IPs are concerned. The treatment of the revised version found in H presents a bit of a conundrum. At least some of the differences it has with the version in C might reflect diachronic differences, but of course they might just reflect differences in personal preferences. I considered simply not including examples from H, since a few of them are duplicates of sentences in C, but decided that it was useful to include the H examples as adding to our data on the use of IPs and DEPs with specific verbs, since the choice of verbs is often different even though the syntax is not. I have therefore included examples from H, but only ones that are not essentially duplicates of ones in C. H is an incomplete text, and

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

 

241

the number of relevant examples in it is small enough that including them does not substantially affect any statistics in this study.

A.3 Other Early The general nature of this text category is outlined in section 3.2. I have split the YCOE file cobede into two files, only one of which, designated ‘Bede(early)’, is included in my Other Early category. This is the portion contained in the early tenth century manuscript known as the Tanner manuscript. Because the Tanner manuscript is defective, missing material at the beginning and the end, Miller’s (1890) edition uses the much later Oxford, Corpus Christi College Manuscript 279. Ker (1957: item 354) dates this manuscript as s. xiin (beginning of the eleventh century), so I have included examples from it in my General OE category. Keeping the early portion separate from the later one is important given that one of the goals of this study is to determine whether changes in the frequency of DEPs took place within the OE period. Examples of IPs and DEPs from the earlier portion can also be seen as representing early prose use in a dialect other than West Saxon—Bately (1988: 98) states that Mercian authorship of the OE translation of Bede’s Historia Ecclesiastica Gentis Anglorum is now generally accepted. These examples from the translation of Bede are especially valuable because the nature of the remaining Other Early texts means that most of them do not yield examples of possessed body words. Because of the dearth of relevant examples in the Other Early texts covered by the YCOE, I added two short fragments of martyrologies to the YCOE files. These are listed in Table A.1 by the titles used in the DOE. Mart 3 (Sweet) is edited from British Library Manuscript Additional 23211 by Sweet (1885: 177–8). Ker (1957: item 127) dates this manuscript as s. ixex (end of the ninth century). Mart 4 (Sisam) is edited by Sisam (1953) from British Library Manuscript BL Add. MS. 40165, dated by Ker (1957: item 132) as s. ix/x. Martyrologies are particularly rich sources of highly affected body possessors, and so although these fragments are brief, they add a couple of valuable examples.

A.4 General OE The general nature of my General OE text type is briefly outlined in section 3.2. As indicated in the discussion of the Other Early category, I have treated examples from the later portion of YCOE’s cobede as General OE, designated bede(xi) in this book. We cannot be certain whether examples from the portion that I have assigned to General OE represent the language of either the period when the translation was first made or the period of the copying, or something in between. While this portion can therefore only be included with caution in comparisons of early and late use, examples of DEPs and IPs in it are nevertheless useful in adding to the picture of what was considered acceptable for writing around the beginning of the eleventh century. Further discussion of the treatment of ASC examples not included in my EWS category is necessary here. Besides the A version of the Anglo-Saxon Chronicle, the YCOE contains cochronC, cochronD, and cochronE. The C and D versions end in 1066 and 1079, respectively. The E version, known as the Peterborough Chronicle, is complicated and, since this version spans OE but is most important for the original material it contains from the twelfth century, is discussed in some detail in the m1 section of this Appendix. I will defer discussion of the use of examples from the E version that I have treated as OE

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

242

 

until after discussing the versions that end earlier. The different versions overlap in some places but are unique in others. In collecting my examples, I have counted what I consider to be the same example, that is, a description of the same event using the same verb, only once. As noted in the discussion of EWS, I have included examples from hands A-1 and A-2 found in cochronA as EWS examples. For later hands in this version and annals not appearing in this version but found in the C or D version, I have treated examples in annals earlier than 1050 as General OE, not presenting data for the individual versions, but treating them collectively as ASC (post-EWS). Examples from 1050 to the end of these chronicles are treated as pre-m1. Turning now to cochronE, we can note that the Peterborough Chronicle is found in a twelfth-century manuscript, and the annals referring to events in what I have treated as the OE period were copied from a borrowed source around 1122, when the part known as the First Continuation begins and the YCOE file ends. Most of the examples of possessed body words in this file are copies of sentences that are found in another version of the ASC and are not included in my counts. However, there are a couple of examples in the cochronE file in entries towards the end of my OE cut-off (1050) that do not appear in the other versions and show that DEPs were still being used in late compositions. I have accordingly counted unique examples in annals before 1050 as ASC(post-EWS) in the General OE category. Annals from 1050 to the end of the YCOE file (i.e. the beginning of the First Continuation) are treated as pre-m1.

A.5 LWS and LWS(Late) In my LWS category, in order to make the best comparisons with EWS texts, I have aimed at including only texts written in a fairly ‘pure’ West Saxon dialect in manuscripts not later than fifty years after they were composed. Such manuscripts are dominated by the works of Ælfric, and so I have added a category LWS(Late) to include copies made more than fifty years after their LWS origin. The texts thought to be composed by Wulfstan, who was writing around the same time as Ælfric and a bit later, go here, because many of them are found in late manuscripts. Some of the homilies of Bethurum’s (1957) collection are in fact found in earlier manuscripts and rightfully belong in the LWS category, but I have not sorted through Bethurum’s edition and extracted those homilies from the YCOE file. The LWS(Late) category also includes some works by Ælfric found only in late manuscripts. In Allen (1992), I identified works of Ælfric as belonging to an A or B group, with A being texts from manuscripts no later than fifty years from the assumed time of composition. Note that the YCOE file coaelhom is based on Pope’s (1967) supplementary edition of Ælfric’s homilies not included in the two large volumes of his Catholic Homilies. Pope’s edition uses manuscripts spanning more than a century and a half. We know from comparing some late copies of Ælfric’s works with extant earlier versions that copyists did not always simply copy his work but sometimes modified his syntax and also added material, which we could expect to reflect the language of the period of copying. Because of the possibility that the later copies might have substituted internal possessors for external ones in Ælfric’s original text, the YCOE file was split into two and put in two text categories for this investigation. The LWS homilies in this file are here designated AEhomsA and the LWS(Late) homilies are AEhomsB. As it turns out, examples of DEPs are so infrequent in the LWS texts that leaving the file unsplit would not have made a difference in this instance.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

 

243

A.6 Middle English A.6.1 Pre-m1 A small number of texts contained in the YCOE used in this investigation belong to what I refer to as the ‘pre-m1’ period, that is, later than 1050 but earlier than 1150, the beginning of what the electronic corpora refer to as the m1 period. As mentioned above, some examples from versions A–D of the Anglo-Saxon Chronicle fall into this category. In addition, there are six more texts in the YCOE that are from late manuscripts, but their probable date of composition is from around the middle of the eleventh century or later, making them sources of valuable information about the latest stage of OE. Two of these texts are in manuscripts from around the middle of the twelfth century, just within the m1 period, but are included because they contain material composed (or generally thought to be, in the case of St. Neot) at a date within the late OE or pre-m1 period. Information about the dates of the manuscripts containing the six YCOE files used for the pre-m1 period is given in Table A.2. The manuscript dates given in this table are those given by Ker (1957) and follow the widely-used system of Roman numerals for the century and superscripted Arabic numerals for first or second half of the century. Table A.2 YCOE texts used for pre-m1 period YCOE cochronE coleofri comargaC coneot

MS

Oxford, Bodlean Laud 636 Cambridge, Corpus Christi College 367 Cambridge, Corpus Christi College 303 London, British Library Cotton Vespasian D. xiv conicodA Cambridge, University Library Ii.2.11 covinsal Cambridge, University Library Ii.2.11

MS Date

Word Count

s. xii¹, xii med. xi² s. xii¹ xii med

n.a.* 1,017 4,196 2,003

xi (3rd quarter)-xii¹ xi (3rd quarter)

8,197 3,655

* Because this study excludes much of the YCOE file for the Peterborough Chronicle, as discussed below in this section, I have not given a word count for it. Note: for most of these texts, the YCOE has not added a temporal extension. However, the documentation for the corpus gives Ker’s (1957) dates, which I have confirmed by checking that catalogue. I have also sought information on dates from newer sources, as indicated in the discussion.

The Peterborough Chronicle (YCOE file cochronE) is a very valuable source, but using it requires care in distinguishing the copied annals from the annals in what are known as the First and Second Continuations. The language of the three different parts is very different, even though the first two parts were written by the same scribe. The part up to the beginning of the First Continuation, i.e. through the year 1121, is covered in the YCOE file cochronE. Some of this falls into the General OE category, as discussed above in this Appendix. The scribe who copied the earlier annals in this manuscript essentially maintained the DEPs and IPs of the exemplars he copied, and any examples from before 1051 are of interest here only in showing that a scribe living at the northern extremity of the South Midlands area was content to copy language very different from what he used

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

244

 

in his own additions, the First Continuation and the Interpolations.iv Examples from later than this up to the end of the YCOE file fall into the pre-m1 period. The DEP presented in (7.11), for example, shows that a particular idiom was still possible later than 1087 at least. The treatment of annals later than 1050 as pre-m1 examples ensures there is no possibility of a large gap between the time of composition of the annals and the time they were copied. The Vision of Leofric, YCOE’s coleofri, is one of the few texts in the manuscripts of the last part of the eleventh century whose composition we can date with some certainty. Stokes (2011), in his recent edition of this text, says it was probably composed shortly after Leofric’s death in 1057 and suggests a manuscript date as the end of the eleventh century or beginning of the twelfth. A scholarly controversy surrounds the composition of St. Neot (YCOE’s coneot). This text is in the British Library Manuscript Cotton Vespasian D. xiv. While this manuscript consists mostly of copied OE material, there are nine items in it that Clemoes (1997: 17) says were either certainly or probably composed in the twelfth century, and St. Neot is one of them. Godden’s (2010) arguments for a much earlier date, somewhere between 1014 and 1027, do not seem to have been widely accepted. Treharne, who is dismissive of Godden’s view, states that this text is a post-Conquest translation (2012: 159) and Younge’s study of this Life reviews the arguments for a post-Conquest date of composition and concludes that the beginning of the twelfth century is most plausible (2012: 368). The legend of St. Margaret contained in comargaC is certainly a copy, and according to the editors of this text, it is unclear how many removes the copy is from the original and there is little positive evidence for dating, but the language is consistent with a time of composition not long before the early twelfth century date of the manuscript (Clayton and Magennis 1994: 106). Treharne (2001) considers that in the process of copying this saint’s life, as well as the other saints’ lives found in this manuscript, was adapted with a twelfthcentury audience in mind. In Chapter 7, however, the discussion concerning examples like (7.15) suggests that the frequent use of an archaic construction in this text might suggest an earlier date of composition. The texts in conichodA and covinsal, found in the same manuscript, are translations from Latin. Cross (1996) suggests that the two texts may have been translated by the same individual. While the time of translation is not certain, Cross makes a case that Bishop Leofric had the Old English versions made during his episcopacy (c.1050–70). Neither of these texts appear to have been the original Old English version, but given that the texts are found in a manuscript of the third quarter of the eleventh century and assuming that the original translations were made in Leofric’s episcopacy, the copies are close enough to the originals to allay any concerns about copying effects. In addition to the six texts of found in the YCOE, I have included some examples from a text of 2579 words known as the Worcester Fragments. Although these fragments are found in a manuscript of a much later period, c.1200, they are clearly a copy, written in the West Saxon of a much earlier period. Hall (1920), who edited these fragments, gave his opinion that they were possibly composed around 1070, when most English prelates had been replaced by Normans. The fragments were certainly composed after the Norman Invasion, since they contain a famous complaint about how a people who once exported learning are now being taught by foreigners. iv It is generally agreed that the scribe who copied the OE material and wrote the First Continuation was also responsible for the Interpolations that were incorporated into the copied material (Clark 1970 n. 1 p. xlvi, Irvine 2004: xc).

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

 

245

A.7 M1 and beyond As discussed in section 7.3, I have supplemented the PPCME2 files for the m1 period with a few additional texts that are considered to be original compositions of the twelfth century. These are listed in Table A.3. Table A.3 Additional EME texts not included in the PPCME2 Text name

Date

Word Count

Cleopatra C.vi/Trinity sermon Poema Morale (L) BL Cotton Vespasian A. xxii EME pieces

C13b1 C12b2, comp. 1170–90? C12b2–C13al

1896 2656 2266

Dates such as C12b2 in Table A.3 are taken from A Linguistic Atlas of Early Middle English (LAEME) for the texts that are included in that atlas, which is still in progress, and from Laing (1993) for texts not yet included in LAEME. The numerals following C refer to the century, while ‘a’ and ‘b’ refer to the first and second halves of the century, respectively. When the manuscript can be dated more specifically within a quarter of a century, the numeral following the half-century indicates this. So for example, the date given for the sermon found in both the British Library, Cotton Cleopatra C vi and the Cambridge, Trinity College 43 (previously B. 1. 45) manuscripts indicates the third quarter of the thirteenth century. The Cleopatra sermon is edited by Dobson (1972: 110–11), with a facsimile of the Trinity version printed opposite it. There are several extant versions of the Poema Morale; the one used here is the one in London, Lambeth Palace Library MS 487. This is not the earliest manuscript containing this text, but I have chosen it for its more conservative case marking. It was published in Morris (1867–8: 159–83, odd pages). Laing (1992: 569) notes that the original of this poem is generally thought to have been composed around 1170–90. Both versions of the sermon and the seven versions of the Poema Morale are also available, with tagging, in LAEME. While the British Library MS Cotton Vespasian A. xxii mostly contains copies of OE writings, the two pieces examined here are of EME composition. Any data provided in this book from this manuscript come from these two texts. These were edited by Morris (1867–8: 231–43) as items XXV and XXVI and given the titles ‘An Bispel’ and ‘Induite Uos Armatura Dei’. They are not yet in LAEME. These additional texts unfortunately yielded no really useful information on the question of the retention of DEPs in the m1 period due to a paucity of examples of possessors of body words. Nevertheless, I mention these texts here because they might yield examples relevant to a reader’s research, and given the small EME corpus, every original composition from this period is valuable. In addition to the prose texts just discussed, I have used some examples from poetry. As discussed in section 7.5.1, Laʒamon’s Brut is found in two versions, in Cotton Caligula A. ix and Cotton Otho C xiii. The language of the Caligula version is in many respects more archaic than that of the Otho version, but recent scholarship has overturned the earlier assumption that the Caligula manuscript was older than the Otho one. LAEME contains descriptions by Margaret Laing which include discussions of the date and language of both manuscripts. These place the language of the Caligula manuscript in northwestern Worcestershire (South Midlands), with northwest Wiltshire (Southwest) for Otho, with a

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

246

 

date of C13b1, that is, in the second half of the thirteenth century, for both manuscripts. LAEME cites different scholars’ opinions, with a time around the middle of the century being dominant for both manuscripts. The manuscripts thus fall (just) into the m2 period, but it is likely that the poem was composed earlier, although the exact date remains uncertain. Elsweiler (2011: 5) says that ‘[c]urrent scholarship seems to favour a date of composition around 1205–1206’. The Otho version is not based on the Caligula text; rather, it is widely assumed that the Caligula version is removed from the poet’s original by at least two intermediate copies, and that the Otho version is a condensation of a version similar to the one upon which the Caligula manuscript is based (Elsweiler 2011: 3–4). The Brut is a lengthy text, and because of the uncertainty surrounding its date of composition, no attempt has been made for this investigation to study it systematically. The poem The Owl and the Nightingale is much shorter (1795 lines), and a complete examination for DEPs was feasible. This poem is found in two manuscripts of the second half of the thirteenth century, one of which is the Caligula manuscript also containing Laʒamon’s Brut. This line is essentially the same in both manuscripts. The date of composition of this poem is uncertain. It is usually assumed to be somewhere between 1189 and 1216, but Cartlidge (1996) argues for a much later date, probably not long before the dates of the manuscripts containing the poems. In addition to the texts discussed in this appendix, I have used a few examples cited in the Middle English Dictionary, supplying information about dates where necessary in the text.

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

APPENDIX B

Vocabulary lists The lists in this Appendix set out the lemmas used in the searches of the corpora and my additional readings. In compiling these lists, I have run lexicon queries on the electronic corpora and have endeavoured to include all variant forms of the Old and Middle English lemmas in the definition files that the queries for these searches called up, including forms that begin with the emendation symbol $ or an upper-case letter, although of course some unusual forms may have been missed. These variant forms are not listed in these tables, but in principle the forms of a lemma include all variants of that lemma in the lexicons run on the individual texts. For OE, in addition to the lexeme queries the Thesaurus of Old English (Roberts et al. 2000) was a valuable source of vocabulary for body and mind words. For the Old English citation forms, I have used the headword(s) used by the DOE where those are available at this time, i.e. where the word begins with A–I. For OE words not beginning with those letters, I have drawn on the Bosworth-Toller (1898) and Hall’s (1960) dictionaries, as well as the Thesaurus of Old English. For EME, I have used the lemmas provided by the Middle English Dictionary.

B.1 Nouns The tables in this section list the nouns that were selected for the broader body and mind searches. The more restricted list of nouns used in searches for body PObjs is given as Table B.8. In the tables of nouns, a Modern English lemma appears in the leftmost column, with one or more roughly equivalent OE/EME lemmas in the column on the right. The large number of lemmas in the righthand column in the tables for OE in some instances is due to the fact that there were many body words in OE that were used only in poetry. An OE lemma appears more than once in instances where there is no one-to-one correspondence between OE and Modern English, e.g. bearm in OE covered ‘lap’ as well as ‘bosom’. Note that the lemma gast ‘spirit’ is not on the list of mind words. The reason is that searches for it turned up a huge number of non-target examples, caught because they contained halig gast ‘Holy Ghost’ or gast agyfan ‘give up the ghost’, i.e. ‘die’. The preliminary searches for gast in the prose yielded no DEPs; for example an IP was always used in expressions meaning ‘give up one’s ghost’. Further general searches for gast therefore seemed unlikely to be fruitful. However, in the poetry, where halig gast is not frequent enough to be a problem, my reading of a poetic passage not included in the York Poetry corpus found a DEP with the expression gast geblissod ‘made joyful in spirit’, presented in Chapter 6 as example (6.11a). In order to count this example as a DEP with a mind word in the poetry, I searched the poetry and prose corpora for similar combinations to ensure that no examples of DEPs that would have been counted for the poetry were missed. Many fewer body and mind lemmas are found in the EME texts. The basic principles for the OE lemmas apply with these lemmas also.

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

248

  Table B.1 Body lemmas, OE Modern English Lemma

OE Lemma(s)

abdomen abdomen (lower) ankle arm back beard bile bladder belly blood body

wamb smælþearme ancleow(e) earm bæc, hrycg beard gealla blædre hrif blod, swat bancofa, banfæt, banhus, bansele, bodig, flæschama, hræw, lic, lichama, sawolhus ban, liþ inylfe æþm brægen andribb, bearm, breost, greada, hreþer, sceat, titt bæce spærlira, scanca ceace, hleor, wange bearm, bosm, breost, greada, sceat cinn, cinban midhrif eare earlæppa elnboga eage, heafodgimm bru bræw andwlita, ansien, hleor, neb(b), nebwlite meox finger liþ flanc flæsc, lira fot foranheafod, foreheafod gesceap toþrima feax hand, folm(e), mund winstre swiþra hafela, heafod heorte hoh, hela hype bæcþearmas, inylfe, innoþ, ropp, þearm ceace, ceafl, geaflas, geagl, goman, hraca liþ

bone bowels breath brain breast buttocks calf cheek chest chin diaphragm ear ear lobe elbow eye eyebrow eyelid face faeces finger finger (tip) flank flesh foot forehead genitals gum hair hand hand (left) hand (right) head heart heel hip intestines jaw(s) joint

(continued)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

  Table B.1 Continued kidney knee knee (back of) lap larynx leg leg (lower) limb lip liver loins lung mouth mucus mouth (inner) muscle nail neck nose nostril palm (of hand) penis pupils rectum rib scrotum septum shoulder shoulders side sinew skin skull sole spinal muscles spleen stomach tear thigh thigh (back of ) throat thumb toe tongue tooth trunk urine vertebra vein waist womb

ædre cneow hamm bearm, greada, sceat þrotbolle scanca, lim hohscanca, scanca ban, lim, liþ lippa, weler lifer lendenu lungen muþ hraca goma banloca nægl heals, hnecca, sweora nasu, neb(b) næsþyrl, nosþirl folm lim seon bæcþearm, utgang ribb codd næsgristle eaxl, sculdor gescyldru side seono fell, hyd heafodpanne, heafodban, brægenloca fotlæst, fotwelm, ile rægereose milte maga tear þeoh(scanca) hamm ceole, hraca, þrote þuma ta tunge toþ, tusc, tuxl bodig, buc micga banhring ædre, middelædr middel cwiþ(a), hrif

249

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

250

  Table B.2 Mind lemmas, Old English Modern English Lemma

OE Lemma

breast conscience heart soul mind memory spirit understanding thought will, intention wish, desire

breost, breostsefa, hreþer, hreþerloca ingehygd heorte, incofa sawol ferþloca, ferhþsefa, gemod, mod, modsefa gemynd ferhþ andgyt, gewit geþanc, geþoht, ingeþanc heorte, gewil, willa (ge)wil

Table B.3 Body lemmas, m1 and m2 Lexeme 1 abdomen ankle anus arm back beard bile bladder belly blood body bone bowels breath brain breast calf cheek chest ear eye eyelid face faeces finger flank flesh foot forehead genitals gum hair

maue ancle outgang arm bak berd galle bladdre beli blod bodi bon gut breth brain barm shank(e) cheke bosom ere eie eie-lid ansene mix finger flank(e) flesh fot forhed shap(e) gome her

Lexeme 2

Lexeme 3

Lexeme 4

rigge

wombe lich

lichame

innoth

bosom

brest

ble

face

titte

lire

(continued)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

  Table B.3 Continued hand hand (right) head heart heel hip intestines jaw(s) joint kidney knee lap larynx leg leg (lower) limb lip liver loins lung mouth mouth (inner) nail neck nose nostril penis pupils rib scrotum shoulder side sinew skin skull sole spleen stomach sweat tear testicle thigh throat thumb toe tongue tooth trunk urine vein waist womb

hande swither(e) hed herte hele hipe tharm cheke lith edre kne barm cod leg shank(e) lim lip(pe) liver(e) lend(e) longe mouth gome nail hals nose nostrille shape appel rib(be) cod axel side sineu fel scul(le) folm milt(e) maue swot ter(e) ston thigh throte thoume to tonge toth bodi migge edre middel innoth

lap(pe)

grede

nekke

swire

shulder

hide

wombe

womb(e)

251

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

252

  Table B.4 Mind lemmas, m1 and m2 EME Lemma

Modern English Lemma

brest herte iminde inihid inwit memorie mind(e) mod soul(e) understondinge

breast heart memory thought mind, consciousness, soul memory mind mind soul understanding

B.2 Verbs The two OE tables of this section, Table B.5 for intransitive verbs and Table B.6 for transitive ones, list the verb lemmas that were used in my comparisons of internal and external possessors of body subjects and objects, respectively. Infinitive forms with the ge- prefix appear in these lists only when the verb in question appears with it in an active sentence; that is, ge- forms which only appear in the data in passive participles are not included. Participle forms are included on the verb list for subjects, because the subjects of passive verbs, like objects, were prime candidates for external possession. Many verbs were used both intransitively and transitively and so are included in the intransitive verb list, but any transitive examples were excluded from the count of IPs with subjects. As discussed in Chapter 3, the examples of DEPs with body parts playing the role of subject in OE do not include unergative subjects, and so no unergative intransitive verbs appear on the intransitive verb list. The verb tables cover verbs that were either on the lists for affected possessors of body subjects or objects and/or the lists of verbs found with DEPs of subjects or objects. So for example in Table B.5, ‘1’ in the Aff_subj column indicates that forms of a given verb appear on the ‘affecting subject’ list in the definition file called up by the query searching for examples of affected possessors of body subjects. It does not indicate that I found any examples of any sort of expressed possessor of a body subject, only that it was on the list as identified as a verb that was likely to indicate an effect. A ‘0’ in this column does not indicate that the searches found no examples of the verb in a targeted possessor construction but rather that the verb is not one on the affecting subject list. This occurs with only three verbs, e.g. batian ‘get better’, which appears in one example with a DEP and so gets ‘1’ in the DEP_subj list but ‘0’ in the Aff_subj list. The ‘1’ in the DEP_subj column indicates that some form batian occurs on the ‘external possessor of subject’ list, i.e. the searches found at least one example with a DEP of a subject of batian. The table lists only inherently intransitive verbs, but the searches for body subjects also included participles of transitive verbs used in passive sentences. Since the purpose of the ‘affecting subject’ list was to identify IPs where DEPs were likely to be an alternative because the possessor was affected, nearly all the lemmas have a ‘1’ in the Aff_subj column. Table B.6 lists the lemmas of the transitive verb forms used in the definition files, following the same layout of Table B.4, but with these verbs any DEP is a DEP of a body object. Lemmas with the ge- prefix appear on this list only if I have found them in the active voice; ge- forms that I have found only in passive sentences are treated as intransitive.

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

  Table B.5 Intransitive affecting/DEP verbs, OE Verb abiterian ablindian abreoþan acalan acan adeadian adeafian ætspurnan afeallan agan (ut) ahlænsian amolsnian asceotan aslacian aslidan astifian asurian aswellan aþeostrian awannian batian berstan brosnian byrnan cinan clifian feallan floterian flowan forberstan forscrincan forsearian gan (ut) gebrosnian gebyrnan geopenian gescrincan geweaxan hangian hefigian hlænian iernan leohtian mistian oþspurnan rotian scimian sigan toberstan

Aff_subj

DEP_subj

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1

1 0 0 1 1 1 1 0 1 1 0 1 1 0 1 0 1 1 1 0 1 1 0 0 1 0 1 1 1 1 0 0 1 0 0 1 1 1 1 0 0 1 1 0 0 1 0 1 1

Meaning become bitter become blind fade, diminish be destroyed by cold ache become dead grow deaf strike against fall (of hair, eyes) go out grow lean, wither weaken shoot out weaken, droop slide stiffen, become unable to move turn sour swell up darken (inchoative) become livid get better break, burst out decay burn, be burnt be chapped, crack open cleave, stick together fall out flutter, overflow (of eyes) flow burst asunder shrivel up wither go out decay, become corrupt burn, be burnt open shrink grow hang become heavy become lean run become light grow dim, dazzled strike against putrefy grow misty fall out, leak burst asunder

(continued)

253

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

254

  Table B.5 Continued tofeallan toflowan togan toswellan þeostrian þindan þolian weallan windan

1 1 1 1 1 1 1 1 1

1 0 1 0 0 1 0 0 1

fall off, collapse gush out separate swell become dark swell, distend suffer swarm, seethe fly off, roll

Table B.6 Transitive affecting/DEP verbs, OE Verb abeatan ablendan abreoþan abreotan aceorfan adon ut ætspurnan ageotan aheardian aheawan ahefigian ahhacian alefan alefian asceotan aslean aslitan asniþan astingan aþeostrian awendan beatan belucan besmitan bindan blendan blodigian brecan brytan ceorfan cnucian cnyssan cwylm(i)an

Aff_obj

DEP_Obj

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0

Meaning knock out blind destroy destroy cut off put out strike against shed (blood) harden hew weigh down peck out injure, cripple injure, cripple shoot strike, beat cut off, destroy cut off stab, poke out make dim, obscure turn beat restrain defile bind blind make bloody break crush, grind cut strike beat, strike torture, torment

(continued)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

  Table B.6 Continued don ut eglian fæstnian fæþm(i)an forbærnan forbindan forbrecan forceorfan forniman forsittan forslean forspillan fortendan fortre(d)dan gebindan gebrecan gedrefan gefæstnian gefeterian gehefigian geniman gesceþþan geslean geswencan gewemman gnidan grapian hearmian hefigian hrinan mælan niman ofaceorfan ofaslean ofslean ofslitan ofsniþan ofsteppan ofstingan oftredan onhon oþdon pinian scearpian scieran slean sniþan stingan swencan swingan teon

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0

1 0 0 0 0 1 0 1 0 1 1 1 1 0 1 1 0 0 1 0 1 0 1 0 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1

put out ail, torment bind embrace (flame) burn up bind up shatter sever, cut off destroy obstruct strike, cut through destroy, maim sear, burn off trample, crush fetter break afflict, defile, trouble bind fetter oppress, burden seize injure strike afflict, oppress defile, besmirch, disfigure rub grasp harm make heavy; oppress touch, strike stain take, seize cut off cut off strike off bite cut off trample stab to death, pierce tread on hang, crucify put out (eyes) torture, torment scarify, score shear, hew strike, beat cut off stab, pierce afflict beat tug, pull

(continued)

255

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

256

  Table B.6 Continued teran tobeatan tobrecan tobregdan tobrysan toceorfan tocwysan togian tolucan toslitan toteran tredan þurhþyrlian þurhdrifan þurhfaran þurhferan þurhgan þurhiernan upahon wanian wendan ymbsniþan

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0 1 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0

tear beat severely shatter rend, tear apart bruise, shatter cut to pieces, cut off crush utterly drag tear apart, dislocate tear apart tear to pieces tread on pierce pierce pierce pierce pierce pierce hang up lessen, diminish turn (e.g. upside down) circumcise

Table B.7 lists the lemmas used in the searches for forms of affecting verbs in EME. Because of the smaller size of the EME corpus, the number of relevant verb lemmas is much smaller, and the verb lemmas for the m1 and m2 periods are combined here, without any distinction made between transitive and intransitive verbs or identification of verbs as occurring with DEPs or not.

Table B.7 Affecting verbs, m1 and m2 Word abuse, mistreat ache afflict ail badly swollen beat beat severely become dull become inflamed become lean bind blacken

Lemma 1 tuken aken derfen eilen forswollen beten tobeten dullen tendren lenen binden blaken

Lemma 2

Lemma 3

Lemma 4

forbeten

(continued)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

257

Table B.7 Continued blind, grow blind bloated break burn up burn, scorch burst cleave, cut into pieces cut cut to pieces destroy devour dig (into), prick dim drag eat explode grasp, take grow dark grow pale grow red grow weak hew to pieces hurt injure, defile injure, wound kill, cut off mar nail through overcome, overwhelm pierce prick rip, dismember rot separate shatter shear smart spit strike suffer swell up tear apart tear to pieces tear, lacerate torture, torment tremble trouble, mar, dim whirl to bits wound

ablinden toblowen breken forbrennen ontenden bresten tocleven kerven forkerven destroien devouren delven dusken louken eten tobresten nimen wannen bloken rudnen feblen forheuen hurten foulen greven ofslen merren thurghnailen overcomen thurghdriven priken torenden forroten totwemen tobreken sheren smerten spitten smiten suffren tobellen tolouken todelen renden pinen tremblen troublen towitheren wounden

blinden

bersten

fordon freten

forswolwen

taken

defoulen harmen

soilen

nailen

thirlen

thurghthirlen

tolimen roten

todrauen

forbreken

forbresten

slen

swippen

swellen todelen toronden

toteren

tormenten

thurghstingen

todrauen

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

258

 

B.3 Prepositions used in searches for PObjs Table B.8 lists the prepositions and body nouns used in the searches for body PObjs in OE. As with verbs and nouns, the lemmas are given in their citation forms for OE. The variant forms in the definition files included emended forms such as $o for on etc. Table B.9 lists the prepositions used in searches for mind PObjs. The same selection of preposition lemmas was used for these searches in OE and EME. The lemmas are given in their citation forms in OE. The definition files used different forms for the OE and EME searches, based on the lexicons for prepositions in the different periods, as spelling differences made necessary. In some texts, the preposition attaches to a following preposition, e.g. a-þet ‘on the’, and I have attempted to include such forms in my definition files for the prepositions used in these searches, as well as variants such as a, o, an, and on for ‘on’. Table B.8 Selected prepositions and body nouns used in OE body PObj searches Preposition

Meaning

Noun

Meaning

æt fram in into of ofer on to þurh

at from in into of, from over on to through

ban eage earm feax fot hand heafod heafodgimm heals hreþer lic lichama limu muþ

bone eye arm hair foot hand head eye neck breast body body limb mouth

Table B.9 Selected prepositions for mind PObj searches, OE OE lemma

Meaning

OE lemma

Meaning

fram in into

from in into

on to uppan

on to upon

Table B.10 Prepositions used in searches for PObj IP vs DEP comparisons in m1 and m2 æt in into

ofer on til

to þurh upon

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

APPENDIX C

Notes on searches In carrying out the searches on the electronic corpora, I used Beth Randall’s CorpusSearch2 program, downloadable from Sourceforge: https://sourceforge.net/projects/corpussearch/. This appendix does not present the queries used but is instead aimed at a discussion of the main problems that arose in searching for DEPs and IPs and explaining decisions made on which examples to include. The general approach to finding DEPs was to construct search queries looking for clauses containing both a body/mind word coded in a way that indicated it was a subject, object, or PObj, as required, combined with a possessor of that noun in the dative case. However, writing queries that target DEPs and IPs involves various complications, and a number of searches additional to the basic searches were required, as discussed below.

C.1 Identifying DEPs and IPs in the YCOE and the York Poetry Corpus: General C.1.1 Identifying dative case Dative possessors are not identified as such in these corpora, but dative case is identified as NP-DAT*, where the asterisk is a wild card that will catch both NP-DAT and any NP-DAT with a further specification. Some NPs that I judged should be treated as DEPs are tagged as NP-ADT* (an adjunct) or NP-RFL* (reflexive) in the corpora, rather than NP-DAT. My basic query searching for a DEP therefore searched for clauses containing any one of these elements plus a body (or mind) word tagged in a way that indicated that it played the grammatical relation targeted by the query. Separate queries were used for possessa subjects, objects, and PObjs. So for example the basic query looking for body subject possessa specified NP-DAT*, NP-ADT*, or NP-RFL* co-occurring in a clause with a body word in the nominative case. When references are made below about searches for datives in the York corpora, the combination NP-DAT*|NP-ADT*|NP-RFL* rather than just NP-DAT is to be understood. This basic search did not catch all relevant examples of DEPs due to various factors. For example, consider the two examples in (C.1): (C.1)

a. ac us bearn þis on mod. but us ran this in mind ‘but this occurred to us’ (coaelive,+ALS[Peter’s_Chair]:232.2426) b. Þa bearn me on mode . . . þæt . . . then ran me: in mind that ‘then it occurred to me . . . that . . . ’ (coprefcath1,+ACHom_I_[Pref ]:174.48.5)

I have treated both examples in (C.1) as DEPs of mind PObjs. In neither example is the pronoun tagged for case, because the forms me and us (as well as some other forms) were ambiguously dative or accusative in case, and the York corpora do not specify a case with ambiguous forms. The (a) example was caught by the basic query searching for such DEPs by the inclusion of NP-ADT as a possible dative, since us is parsed as (NP-ADT (PRO us))

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

260

 

in the corpus. The me of the (b) example, in contrast, is parsed as (NP (PRO me)) without any indication of grammatical function. To capture such examples in which no grammatical function is indicated by the parsing in the corpus but the PRO is to be taken as dative rather than accusative even though the morphology is ambiguous, I wrote additional queries substituting PRO for NP-DAT* etc. in order to find any examples of DEPs that the main query missed. With the output for the searches for direct arguments at least, it is usually easy to distinguish a PRO that is to be treated as a DEP from one that is to be construed as accusative. So for example, in (C.2) me could only be dative because the verb is intransitive, and I have counted it as a DEP with a subject possessum: (C.2)

Sint me leoðu are me: limbs: ‘My limbs are dislocated’

tolocen dislocated Andreas 1404

In the glosses in the examples presented in this book, I have treated the pronoun in examples like this as dative without further comment. There is also no problem when there is an accusative object in an example in addition to the morphologically ambiguous pronoun, since the latter could not also be accusative. Things are tricker when we are dealing with PObjs occurring in clauses with verbs that take some sort of object, and here I have been guided by sources such as dictionaries as well as Mitchell’s (1985: §1092) list of verbal rections. Another type of ambiguity concerns the syncretism of dative and genitive case found with strong feminine nouns and with weak nouns of all genders. An example is given as (4.3) in Chapter 4, repeated here as (C.3) for convenience. (C.3)

and het and ordered

bindan bind

fet feet:

and honda and hands:

þære the:/

halgan fæmnan holy:/ woman:/ ‘And ordered the holy woman’s hands and feet to be bound’ (comargaT,LS_16_[MargaretCot.Tib._A.iii]:18.1.229) In these cases, the York corpora generally assign either genitive or dative case to the ambiguous form, but I have treated them as ambiguous because either is actually a possibility in a position like this, as noted in section 4.2.1. DEPs of PObj possessa could similarly be positioned after the possessum: (C.4)

þa to fotum Loth þam then to feet: Lot: the:. ‘then Lot sank to the feet of the guests’

giestum hnah, guests:. sank (cogenesi,73.2441.471)

I have accordingly counted examples like (C.5) as ambiguous between having a DEP or an IP because frean could be either dative or genitive in its position: (C.5)

þe æt fotum sæt frean Scyldinga, that at feet sat lord:/ Scylings:. ‘who sat at the feet of the lord of the Scyldings’ (cobeowul,17.499.421) Beowulf 500

In some instances of technical ambiguity, only one case is plausible. So for example I had no hesitation in counting hire in (C.6) as an IP: (C.6)

ac hire swura næs but her: neck: not.was ‘But her neck wasn’t cut through’

þurhslagen through.struck (coaelive,+ALS[Ash_Wed]:235.2837)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

261

The reason is that by the late stage of this text, if hire were a DEP, the possessum swura would certainly have required a definite determiner, i.e. hire se swura.

C.1.2 Discontinuous possessors Searching for IPs would be simpler if a genitive were always adjacent to what it modified and therefore treated in the parsing as dominated by the same NP as the possessum. With the freer word order of OE, however, we find examples like (C.7) (in which the language imitates the situation being described by separating the body part from its owner): (C.7)

syðþan Æscheres on þam holmclife hafelan metton. after Æschere’s on the sea.cliff head: found ‘after they found Æschere’s head on the sea-cliff ’ (cobeowul,44.1417.1175) Beowulf 1430–1

Æscheres is not an external possessor but rather an IP, albeit a discontinuous one. Conceptually, it forms a unit with hafelan, and unlike a DEP bears no role in relation to the verb. Such discontinuous examples are not infrequent in the poetry but are rare in the prose. I wrote extra queries to look for discontinuous examples and applied them to both poetry and prose. As it turns out, these queries turned up no examples that needed to be included in my statistics—example (C.7) is an IP of a body object, but because metton is not on my affecting verbs list, this example does not get included in this investigation’s comparison of IPs and DEPs.

C.1.3 DEP or another type of dative? The difficulties caused by ‘shared’ datives, that is, when the referent of a dative that is part of the case-frame of a verb is also the possessor of the body or mind word, have been mentioned in section 2.2.1, where it is explained that such datives were excluded from the count of DEPs. Here, it can be mentioned that the potential for shared datives is greater with PObjs than with the direct arguments. For one thing, a body word playing the role of PObj might be combined with a DEP of a direct object, as in (C.8): (C.8)

þæte him mon aheow þa hond mid þy earme of that him: one hewed the: hand: with the arm of þæm lichoman; the body ‘that his hand with the arm was hewn from his body’ (cobede,Bede_3:4.166.12.1599)

In examples like this, I have counted the dative as the DEP of þa hond, and not counted the PObj þæm lichoman. Another situation in which ‘sharing’ of a dative could happen is when a prepositional phrase is added to a ditransitive construction with a dative indirect object, as in (C.9): (C.9)

ond hit ða swa heolfrig hyre and it then so gory her: ‘and gave it then, thus gory, into her hand’

on in

hond hand

ageaf gave Judith 130

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

262

 

In this example, hyre is ambiguous in form, but is surely to be construed as dative, as the indirect object of ageaf. I have not found examples in which a dative possessor appears in a clause when the dative is already governed by some other element, such as the verb. As discussed in section 2.2.1, examples like (C.9) can be treated as involving an implicit possessor, and I have not included them in my statistics. To keep down the number of IPs used in positions where DEPs were not possible, I have also excluded examples like (C.10) in my statistics, even though on þinre handa is a clear IP in a PObj: (C.10)

& ic hine giue ðe and I him: give thee: ‘and I give him to you in your hands’

on in

þinre handa; your hands (cogenesiC,Gen_[Ker]:42.36.279)

In some instances, it is not entirely clear whether a dative is shared because it is not clear whether it is governed by the verb of the clause: (C.11)

Hæbbe him ær on muþe ele oþþe buteran have him: earlier in mouth oil or butter ‘Let him have oil or butter in his mouth before’ (colaece,Lch_II_[1]:1.3.5.23)

I have counted the dative in (C.11) as a DEP, but it is possible that it is better regarded as some other sort of dative, with an implicit possessor. The implicit possessor analysis is made possible by the fact that the referent of the possessor is also the (unexpressed) referent of the subject. Furthermore, clear implicit possessors, with no overt mention of the possessor, either as a dative or a possessive, were very common with body PObjs generally. Such an analysis is made the more plausible by the fact that habban is sometimes found with what are typically treated as ‘indirect objects’ by dictionaries, as in example (C.12), which was discussed in Chapter 2 in the discussion of ‘ethic’ datives and is repeated here for convenience: (C.12)

aris nu and hafa ðe Naboðes wineard arise now and have thee: Naboth’s vineyard: ‘arise now and take Naboth’s vineyard for yourself ’ (coaelive,+ALS_[Book_of_Kings]:200.3811)

C.1.4 Conjoined possessa Conjoined possessa also required some general principles about what sorts of examples should be excluded. When two or more body words were the arguments of a single verb, the possessor of the first possessum could be either a DEP as in (C.13a) or an IP (C.13b,c). If the possessor was expressed by an IP in the first conjunct, the second and any other conjuncts could have another IP, as in (C.13b). In (C.13c) we have an interesting example in which two IPs are followed by an implicit possessive. (C.13)

a. ðonne sceal mon þam men mid drium handum . . . þa then shall one the: man: with dry hands the: handa & þa fet gnidan swiðe & þyn hands: and the: feet: rub greatly and squeeze ‘then one shall rub and squeeze the man’s hands and feet’ (colaece,Lch_II_[2]:3.1.7.2179)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

263

b. & symle on æfenne his heafod & his lic mid þy and always in evening his head: and his body: with that þwea & gnide. wash and rub ‘and always in the evening wash and rub his head with that’ (coquadru,Med_1.1_[de_Vriend]:7.10.272) c. & ceorfan of his nosu & his earan & þa uferan and cut off his nose: and his ears: and the: uferan lippan upper: lip: ‘and cut off this nose and his ears and the upper lip’ (colaw2cn,LawIICn:30.5.109) One might expect that a DEP for the first conjunct could be combined with an IP for the second, but I have not found any such examples. My procedure with these conjoined body words was to count only the first possessum. So in (C.13a), þam men is counted as a DEP of þa handa, but þa fet is not counted, being analysed as implicitly possessed. In (C.14b), two clear IPs are present, but following my practice of counting only the first possessum is such conjoined structures, only his heafod is included in the statistics. In (C.13c), I have only counted the first of the two IPs. The reason for excluding the second is that it appears that a DEP would not be possible in the second conjunct, and so counting the IP would not follow the general goal of showing how frequent DEPs were in situations where one might be expected as a possibility. When two or more verbs were conjoined, matters are different, because the associated subjects or objects are independent: (C.14)

Her Romane Leone þæm papan his tungon forcurfon & Here Romans Leo: the: pope: his tongue: cut and his eagan astungon, his eyes: stabbed ‘In this year the Romans cut out Pope Leo’s tongue and put out his eyes’ (cochronA-1,ChronA_[Plummer]:797.1.597)

I have treated þæm papan his tongue as a blended construction and his eagan as an IP. A DEP such as & him þa eagan astungon would plausibly have been an alternative here because of the second verb.

C.2 Treatment of ‘affecting’ verbs The counts of IPs with affecting verbs do not include all examples in which a verb identified as ‘affecting’ occurs. This is because many verbs are used in a wide range of senses. For example, don carries a wide range of meanings, including ‘do’ and ‘put’ and does not normally carry any meaning of affectedness. However, when this verb is combined with the particle ut, it means ‘put out’, and references to people’s eyes being put out are distressingly frequent in OE texts. When this violent action is reported, a DEP is usually used, even by Ælfric, who, as discussed in Chapter 4, strongly preferred IPs in general: (C.15)

and æfter worulddome dydon him ut þa eagan and after world.judgement put him: out the: eyes: ‘and following secular judgement, put out his eyes’ (coaelive,ÆLS_[Swithun]:265.4388)

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

264

 

Because of this specific meaning, I have included verbs like don on my list of affecting verbs, but I used an extra set of queries to look specifically for do in expressions meaning ‘put out’. It should also be noted that I found several examples of IPs in sentences where there was clearly a negative effect on the possessor which involved a verb that is not included on the ‘affecting’ list for some reason, such as because in its more common uses it did not convey such an effect. In keeping with the general approach of keeping down the number of IPs collected from sentences in which a DEP was not a likely alternative, I did not count such examples of IPs with verbs not on the list.

C.3 Special problems with PObjs C.3.1 OE Managing the data in searching for DEPs and comparing them with IPs was considerably more difficult with PObjs than with direct arguments. To begin with, the sheer number of examples that any search that would capture all DEPs of body PObjs produced would be very large, since this would require searching for all sentences that had both a dative of some sort and a prepositional phrase containing a body word. Most of the output would be non-target ‘hits’, since many of these datives would not be in a possessive relationship with the PObj. Instead of attempting to count all DEPs of PObjs in the whole corpus, I restricted the searches in a number of ways. First, I limited the investigation to a smaller number of texts than used in the investigation of direct arguments. Searches were carried out on three text categories: OE Poetry, EWS, and LWS. Because the LWS corpus is so much larger than the other two, I further limited my searches within this text category to three texts. Within these limitations, I carried out two types of searches for DEPs with PObj possessa. The first type was broad in looking for all body objects of all prepositions. I did not count the number of examples of DEPs in this corpus, but instead noted the prepositions and nouns most commonly used in DEPs. I used my observations from this broad investigation to compile a list of prepositions for use in the second, narrower type of inquiry, the comparison of DEPs and IPs with PObjs. So for example the list of prepositions did not include mid, which often expressed an instrument: (C.16)

and mid his halgum handum husel senode And with his holy hands Eucharist consecrated ‘and with his holy hands consecrated the Eucharist’ (coaelive,+ALS_[Maur]:64.1533)

DEPs were not found with the PObjs of instrumental prepositional phrases, but were common with locative and directional prepositions. The prepositions used in this restricted comparison of DEPs and IPs are listed in Appendix B. By narrowing down the prepositions to this list, the number of DEPs became manageable. However, when it came to comparing DEPs and IPs, the initial searches for IPs and PObjs using this restricted list still yielded a very large number of examples. For instance, even with the restriction the investigation of LWS to three texts when looking for body possessa as PObjs, the search yielded 439 potential examples. The large number results mostly from a combination of two facts. First, the overall number of

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

265

examples in which body possessa play the role of PObj is much larger than the number with direct arguments, and IPs were simply much more commonly used than DEPs with PObjs in these texts. Second, affectedness did not seem to play the same role in causing the selection of a DEP with PObjs as it did with the direct arguments. This means that unlike with the direct arguments, it was not useful to use the list of affecting verbs with searches for IPs with PObjs. Drawing up a list of verbs appearing with a DEP of a PObj for comparison with IPs with those verbs would also be of little use; DEPs appeared frequently in phrases like him gan on handa ‘come under his control’, but a search that included forms of gan ‘go’ would yield a large number of irrelevant examples in which a non-possessor dative occurred somewhere in a sentence that involved both this verb and a body PObj. Before moving on to further restrictions, it should be mentioned that I also excluded uses of the preposition of when it was within an NP as a complement: (C.17)

Ðæt feax ðonne on hira heafde getacnað ða uterran geðohtas, the hair then on their head betokens the outer thoughts ‘the hair on their head, then, signifies the outer thoughts’ (cocura,CP:18.139.16.947)

A DEP would not be possible in this position, even with an affected possessor. In order to reduce the number of examples, I reduced the body list to a small number of words that I noticed occurring commonly with DEPs of PObjs. This restricted list of body words (including variant forms) used by the searches for body PObjs is set out in Appendix B. While the number of potential IPs was greatly reduced by narrowing down the body words to this restricted list, the number of DEPs was not greatly affected. To be certain of this, I ran a further search using the restricted preposition list with all body words. This yielded only a few more examples of DEPs with PObj body words, e.g. (C.18), which has the word bearm, which is on the larger body list but not on the more restricted one: (C.18)

him to bearme cwom maðþumfæt him: to lap came treasure.vessel:nom ‘the precious vessel had come into his possession’ (cobeowul,75.2403.1958) Beowulf 2404–5

These searches for DEPs with all body word PObjs confirmed my impression, based on the earlier searches, that with the selected prepositions at least, the selected body words used for my DEP versus IP comparisons were indeed the body words most used in the texts as PObjs with a DEP. In other words, the higher number of IPs found with these body words was not due to a choice of words that were less likely to occur with DEPs as objects of these prepositions than were other body words. In searching for IPs, this combination of selected prepositions and restricted body words still gave a large number of examples of IPs to compare with DEPs but made a substantial reduction; the number of examples to work through in the three LWS texts was now 201. This meant it was possible to apply the queries searching for DEPs and IPs with this restricted combination to all text types. However, because of the large size of the corpora of LWS and General OE texts, I did not apply it to all texts within these text types, limiting the LWS corpus to the three texts used for the broader search described above and my General OE and 9thC (OE) searches to some selected texts. The selected texts of both these text types are laid out in Table C.1:

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

266

  Table C.1 Selected OE texts for IP and DEP comparisons for PObjs within three text types YCOE text

Text Type

coblick coeuphr coeust cogregC* comart3 cosevensl coverhom coaelive cocathom1 cowsgosp colawafint colaece

General OE General OE General OE General OE General OE General OE General OE LWS LWS LWS 9thC(OE) 9thC (OE)

The asterisk on cogregCo24 indicates that not all of this lengthy text was used for the body PObj searches; they were confined to Book I. The searches of the poetic texts carried out for the IP and DEP comparisons covered all the poetic texts used in all other investigations. The number of clauses that include both a dative element and a body PObj is large even with these restrictions, but successive searches of sorts that will not be detailed here reduced the number of hits to manageable numbers. Ambiguities of grammatical features are more problematic with PObjs than with the direct arguments. There is usually no difficulty recognizing a dative (as opposed to a formally identical accusative) with the direct argument possessa, but problems arise when a body PObj is used in the same clause with a verb that had an object with ambiguous case marking. Many verbs could take either an accusative or dative object, leaving the case marking of the possessor in (C.19) uncertain: (C.19)

Mid þy þe god me hafað gehæfted with that: that God me has fettered ‘Since God has fettered me by the neck’

be by

þam the

healse neck Genesis B 385

The form me is probably to be taken as accusative here, but if we wanted to make sure it could not reasonably be taken as a dative, we would have to check dictionaries or other sources for the valency of hæftan. This particular example can safely be excluded from our collection of DEPs at any rate because god me hafað gehæfted be þam healse surely entails god me hafað gehæfted. As discussed in Chapter 1, I have systematically excluded thematic datives as DEPs. The example points out the problem of determining what counts as a thematic argument with specific OE verbs, although in this particular example the decision is easy enough. In (C.20), ambiguity in a grammatical feature carried by a form combines with other factors to cause uncertainly about whether the sentence should be treated as having an externally possessed PObj: (C.20)

hateð him gewitan on þa winstran orders them: depart on the left ‘(He) orders them to depart on his left side’

hond hand Christ 1227

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

267

There is no doubt that him is dative in case, but it could be singular or plural. If him is taken as singular, this is an example of a dative external possessor, with the meaning ‘on his (Christ’s) left hand’. However, it is possible that him should be taken as plural. If so, it would not be the object of hatan, which normally governed the accusative case for its object, when that was expressed. The object of hatan did not need to be expressed; this verb could be followed directly by an infinitive, with the recipient of the order understood from the context or simply left unclear. Rather, him could be treated as a dative reflexive object of the verb gewitan, since a reflexive dative object was possible with this verb, as discussed in Chapter 2. This being so, him could refer to the damned souls whom Christ puts on his left hand. In that case, þa winstran hond would not have an overtly expressed possessor, a plausible analysis since implicit possessors were particularly common in poetry, especially in prepositional phrases. The count of DEPs of PObjs accordingly excludes this example as dubious; however, including it would not make a difference in the overall picture shown by Table 4.20 in Chapter 4, which clearly demonstrates that DEPs were the dominant construction with PObj body possessa in the poetry.

C.4 Culling the OE examples: further notes In deciding which examples returned by the queries should be excluded from the data, I generally followed the principle of maximizing the number of examples that could reasonably be treated as DEPs, and to minimize the number of IPs in situations where a DEP would not be expected to be a possibility, as far as this was possible on cross-linguistic grounds. The reason for this is that any attempt to demonstrate that DEPs were not as dominant in OE with inalienable possessors as is usually assumed would be unconvincing if the number of IPs was inflated by counting examples where no DEP would reasonably be expected in a European language where DEPs are widely used. Cross-linguistically, DEPs are not normally used when the possessor in question is inanimate. The counts of IPs therefore exclude inanimate possessors. A DEP is also not expected when the body in question is dead and the possessor is incapable of feeling an effect. Descriptions of things being done to dead bodies or parts of them are very common in the corpus, and so the treatment of such examples when the possessor of the body is expressed would affect the statistics substantially. When a possessor is expressed, it is nearly always with an IP, as in (C.21), in which the burning of a dead body in its grave is under discussion: (C.21)

& eac swa lange se lig bærnde his ban, and also so long the fire burned his bones ‘and also the fire burned his bones for so long’ (cogregdC,GDPref_and_4_[C]:33.309.10.4614)

One possible approach would be simply not to count examples where the possessor is dead. However, the searches did find one example of a DEP with a direct argument, discussed in section 4.2.1. This means that examples with dead possessors could not simply be ignored. Recall that the counts of DEPs with direct argument possessa were of two types: a simple count of all DEPs, where there is no comparison with IPs, and a second count in which DEPs and IPs are compared with a selected group of verbs, my affecting verbs. I treated dead possessors differently in the two counts. The simple count of all DEPs includes any examples with dead possessors, and the number of such examples is noted in the relevant sections. With the investigations comparing DEPs and IPs with affecting verbs, on the other hand, the statistics exclude all examples with a dead possessor, whether a DEP or an IP.

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

268

 

PObj possessa required a different approach from the direct arguments: with these, the affectedness condition did not seem to hold, and none of the queries specified affecting verbs. With PObjs, then, any example of a PObj with the selected words, both IPs and DEPs, was included in the counts, as long as the possessor was human, whether dead or alive. As it turns out, the investigation found only two examples of a DEP of a dead body PObj: (C.22)

a. þa gyt hie him asetton segen geldenne heah ofer then yet they him: set standard: golden: high over heafod head ‘then further they set a golden standard high over his head’ (cobeowul,4.47.40) Beowulf 47–8 b. gestodon him æt his lices heafdum stood: him: at his body: head ‘and (they) stood at the head of his body’ (codream,63.63.99)

There could be a difference here between poetry and prose, but it is impossible to be certain because of the paucity of relevant examples involving dead bodies in the poetry, with the target words used here, at least. More examples are to be found in the prose, and they all use an IP: (C.23)

& gong to cirican to þæs halgan Oswaldes lice, and went to church to the: holy: Oswald: body ‘and went to church to the holy Oswald’s body’ (cobede,Bede_3:10.186.26.1883)

The numbers are small here too, however, and so a clear difference between the poetry and the prose with regard to dead possessors cannot be established. The numbers are small enough in the prose that including the examples with dead bodies should not materially affect the statistics. A final type of exclusion made to reduce the number of IPs in situations where a DEP would not have been likely concerned examples in which the words on one of my lists were used in a sense not intended to be captured by my searches. Words for body parts often had a metaphorical meaning. In particular, certain body words such as heorte ‘heart’ could refer to someone’s nature, mind, or spirit: (C.24)

oþ þæt him seo heorte eft to Criste gecerde until that him: the: heart: again to Christ turned ‘until his heart turned again to Christ’ (comart3,Mart_5_[Kotzor]:Ju18,A.12.995)

I counted examples with this sort of metaphorical meaning in data for the mind words rather than the body words.

C.5 Infinitives My queries for OE only looked for possessors and possessa occurring in the same clause, encoded as an IP in the York corpora and the PPCME2. These queries captured examples involving a modal verb and a bare infinitive, because the infinitive after a modal is not parsed as being in a different IP from the main verb, but might have missed some examples of DEPs involving a to infinitive, which are treated as subordinate IPs by the

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

 

269

corpora. The queries may also have missed some examples of topicalized DEPs, since IP*, rather than CP*, was the boundary node used in the queries. However, my own extensive reading of texts makes me confident that such missed examples would be very few in OE, and trying to capture them would greatly increase the number of non-target hits thrown up by the queries. My experimentation with a number of versions of my queries left me confident that I had not missed a significant number of relevant examples.

C.6 Searching the PPCME2 As explained in section 7.7, the PPCME2 does not annotate NPs by case but by grammatical relation, necessitating a different set of queries for DEPs. The function that most frequently is used for the NPs that I treat as DEPs in this corpus is NP-OB2, but this is not the only possibility. In a small number of sentences, what I would count as a DEP is treated as NP-OB1. That is, the possessor rather than the body word is treated as a direct object: (C.25)

ha duluen me þe fet & þe they delved me the feet and the ‘They dug into my feet and hands’

honden. hands (CMANCRIW-1,II.215.3101)

In (C.25), the author is translating the Latin he has just presented, foderunt Manus meas & pedes meos, ‘they pricked/dug into my hands and my feet’, in which Manus meas & pedes meos are unambiguously accusative in case. It seems straightforward here to similarly analyse þe fet & þe honden as the direct object, and me as a DEP. It was possible to capture such examples by searching for body words dominated by NP-AD*, because in the example in which the assumed DEP is parsed as OB1, the body word is either NP-ADV (adverbial NP) or NP-ADT (adjunct NP). The explanatory material provided with the PPCME2 warns the reader that the coding for NP-OBJ1 and NP-OBJ2 is not always consistent. My searches for DEPs in that corpus therefore included a number of extra searches beyond my basic searches for DEPs to check for examples that might have been missed by searches that were written to capture them. I also read substantial stretches of texts included in the PPCME2 to make sure that they did not contain DEPs that my queries did not catch. The basic searches carried out for EME are sufficiently described in Chapter 7 to require little additional comment here. However, it can be mentioned that while it was noted in C5 that the searches for DEPs in the York corpora would not have caught examples in which a DEP was in one clause and the body or mind word was in another, which might have missed some examples involving infinitives, the EME corpus is small enough that the queries could be written to go across IP boundaries to capture as many examples of datives as possible. The list of prepositions used in collecting PObjs in EME consisted of the reflexes of the same prepositions used for OE except that of and from were excluded from the searches for body words in EME. The reason for the exclusion of the former was that in EME it resulted in a large number of hits involving ‘of genitives’: (C.26)

for hij herden alle þe words of for they heard all the words of ‘For they heard all the words of thy mouth’

þi mouþe; thy mouth (CMEARLPS,166.7329)

Although þi mouþe is certainly an internally possessed body PObj, it must be kept in mind that the reason for counting IPs is to compare them with DEPs, and this phrase is in a position where a DEP could not possibly have been used as an alternative; there are no

OUP CORRECTED PROOF – FINAL, 26/8/2019, SPi

270

 

examples equivalent to *they heard me the words of the mouth in OE. A large number of of genitives in the m2 cmearlps are of this type. Rather than make decisions about whether specific of genitives should be counted or not, I excluded of from my list of prepositions entirely. Including the of genitives in the count of IPs would certainly increase this number, especially in the m2 period, when they are very common, but exact numbers are not necessary to convince us that the DEP was by far a minority construction compared with the IP. The list excludes from because a preliminary search for IPs that included both of and from threw up a substantial number of examples of from that were instrumental or conveyed some other meaning that never used a DEP, meaning that the inclusion of this word complicated the data collection without yielding a significant benefit. For a list of the prepositions used for the ME searches, see Appendix B.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

References Ahlgren, Arthur (1946). On the Use of the Definite Article with ‘Nouns of Possession’ in English. Uppsala: Appelbergs Boktryckeriaktiebolag. Allen, Cynthia L. (1992). ‘Old English and the Syntactician: Some Remarks and a Syntactician’s Guide to the Works of Ælfric’, in Colman, Fran (ed.), Evidence for Old English (Edinburgh Studies in the English Language 2). Edinburgh: John Donald, 1–19. Allen, Cynthia L. (1995). Case Marking and Reanalysis: Grammatical Relations from Old to Early Modern English. Oxford and New York: Oxford University Press. Allen, Cynthia L. (1996). ‘A Change in Structural Case Marking in Early Middle English’, in Thráinsson, Höskuldur, Samuel David Epstein, and Steve Peter (eds.), Studies in Comparative Germanic Syntax. Volume II. Dordrecht, Boston and London: Kluwer Academic Publishers, 1–20. Allen, Cynthia L. (1997). ‘The Development of an “Impersonal” Verb in Middle English: The Case of Behoove’, in Fisiak, Jacek (ed.), Studies in Middle English Linguistics (Trends in Linguistics). Berlin and New York: Mouton de Gruyter, 1–21. Allen, Cynthia L. (2006). ‘Case Syncretism and Word Order Change’, in Ans van Kemenade and Bettelou Los (eds.), Handbook of the History of English. Oxford: Blackwell, 201–23. Allen, Cynthia L. (2007). ‘The Case of the Genitive in the Peterborough Continuations’, in Bergs, Alexander and Janne Skaffari (eds.), The Language of the Peterborough Chronicle (Studies in Medieval Language and Literature 20). Frankfurt am Main: Peter Lang, 77–92. Allen, Cynthia L. (2008). Genitives in Early English: Typology and Evidence. Oxford: Oxford University Press. Allen, Cynthia L. (2013). ‘The Poss(essive) Det(erminer) Construction in Early Middle English Writings’, in Wright, Laura and Richard Dance (eds.), The Use and Development of Middle English. Proceedings of the Sixth International Conference on Middle English, Cambridge 2008 (Studies in English Medieval Language and Literature 38). Frankfurt am Main: Peter Lang, 109–27. Allen, Cynthia L. (2016a). ‘Sifting through the Evidence: Principles and Pitfalls’, in Machan, Tim (ed.), Imagining Medieval English. Cambridge: Cambridge University Press, 166–87. Allen, Cynthia L. (2016b). ‘The Definite Article: What Happened with Þe?’, in Vikner, Sten, Henrik Jorgensen, and Elly van Gelderen (eds.), Let Us Have Articles Betwixt Us. Papers in Historical and Comparative Linguistics in Honour of Johanna L. Wood. Aarhus: Dept. of English, School of Communication & Culture, Aarhus University, 43–82. Amos, Ashley Crandell (1980). Linguistic Means of Determining the Dates of Old English Literary Texts. Cambridge, MA: Medieval Academy of America. Atkins, John William Hey (ed.), (1922). The Owl and the Nightingale. Cambridge: Cambridge University Press. Bately, Janet (ed.), (1980). The Old English Orosius (Early English Text Society Supplementary Series, 6). London: Oxford University Press. Bately, Janet (ed.), (1986). The Anglo-Saxon Chronicle: A Collaborative Edition. Volume 3: Ms A. Cambridge: D. S. Brewer. Bately, Janet M. (1988). ‘Old English Prose Before and During the Reign of Alfred’, Anglo-Saxon England 17: 93–138.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

272



Behaghel, Otto (1958). Heliand und Genesis. Herausgegeben von Otto Behaghel. 7 Auflage Bearbeitet von Walther Mitzka. 7 edn. (Altdeutsche Textbibliothek, 4). Tübingen: Max Niemeyer. Bethurum, Dorothy (ed.), (1957). The Homilies of Wulfstan. Oxford: Clarendon Press. Bolkestein, A. Machtelt (2001). ‘Possessors and Experiencers in Classical Latin’, in Baron, Irène, Michael Herslund, and Finn Sørensen (eds.), Dimensions of Possession (Typological Studies in Language v. 47). Amsterdam and Philadelphia: John Benjamins 269–83. Bosworth, Joseph and T. Northcote Toller (1898). An Anglo-Saxon Dictionary, Based on the Manuscript Collections of the Late Joseph Bosworth. Oxford: The Clarendon Press. Brook, G. L. and R. F. Leslie (eds.), (1963, 1978). Laʒamon: Brut (Early English Text Society, 250, 277). London: Oxford University Press. Campbell, Alistair (1959). Old English Grammar. Oxford: Clarendon Press. Cathey, James E. (2000). Old Saxon. (Languages of the World Materials, 252). München: Lincom Europa. Cartlidge, Neil (1996). ‘The Date of the Owl and the Nightingale’, Medium Ævum lxv: 230–47. Clark, Cecily (ed.), (1970). The Peterborough Chronicle 1070–1154. Oxford: Clarendon Press. Clayton, Mary and Hugh Magennis (eds.), (1994). The Old English Lives of St. Margaret (Cambridge Studies in Anglo-Saxon England, 9). Cambridge and New York: Cambridge University Press. Clement, Richard (1985). ‘King Alfred and the Latin Manuscripts of Gregory’s Regula Pastoralis’, Quidditas 6: 1–13. Clemoes, Peter (ed.), (1997). Ælfric’s Catholic Homilies: The First Series (Early English Text Society Supplementary Series, 17). Oxford: Oxford University Press. Cockayne, Thomas Oswald (ed.), (1961). Leechdoms, Wortcunning and Starcraft of Early England. London: The Holland Press. Coetsem, Frans van (2000). A General and Unified Theory of the Transmission Process in Language Contact. (Monographien Zur Sprachwissenschaft, 19). Heidelberg: Winter. Crisma, Paola (2011). ‘The Emergence of the Definite Article in English: A ContactInduced Change?’, in Sleeman, Antonia Petronella and Harry Perridon (eds.), The Noun Phrase in Romance and Germanic: Structure, Variation, and Change. Amsterdam and Philadelphia: John Benjamins, 175–92. Cross, James E. (ed.), (1996). Two Old English Apocrypha:The Gospel of Nichodemus and the Avenging of the Saviour (Cambridge Studies in Anglo-Saxon English, 19). Cambridge: Cambridge University Press. Deal, Amy Rose (2013). ‘Possessor Raising’, Linguistic Inquiry 44 (3): 391–432. Dictionary of Old English A-H Online (2018), edited by Angus Cameron, Ashley Crandell Amos, Antonette diPaolo Healey et al. Toronto: Dictionary of Old English Project. Dobson, E. J. (1972). The English Text of the Ancrene Riwle. (Early English Text Society, 267). London: Oxford University Press. Einenkel, Eugen (1884). Life of St Katherine. (Early English Text Society, 80). London: N. Trubner & Co. Elsweiler, Christine (2011). Laʒamon’s Brut Between Old English Heroic Poetry and Middle English Romance: A Study of the Lexical Fields ‘Hero’, ‘Warrior’ and ‘Knight’. (Münchener Universitätsschriften, 35). Munich: Peter Lang. Ferraresi, Gisella (2005). Word Order and Phrase Structure in Gothic. Leuven and Dudley, MA: Peeters. Filppula, Markku (2008). ‘The Linguistic Outcomes of the Early Contact’, in Filppula, Markku, Juhani Klemola, and Heli Paulasto (eds.), English and Celtic in Contact. New York: Routledge, 24–132.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi



273

Filppula, Markku (2010). ‘Contact and the Early History of English’, in Hickey, Raymond (ed.), The Handbook of Language Contact. Malden, MA: Wiley-Blackwell, 432–53. Filppula, Markku, Juhani Klemola, and Heli Paulasto (2008). English and Celtic in Contact. New York: Routledge. Fulk, R. D., Christopher M. Cain, and Rachel S. Anderson (2003). A History of Old English Literature. Malden, MA: Blackwell. Fulk, R. D. and Stefan Jurasinski (eds.), (2012). The Old English Canons of Theodore (Early English Text Society Supplementary Series, 25). Oxford: Oxford University Press. Giusti, Giuliana (2015). Nominal Syntax at the Interfaces: A Comparative Analysis of Languages with Articles. Newcastle upon Tyne: Cambridge Scholars Publishing. Godden, Malcolm (2010). ‘The Old English Life of St Neot and the Legends of King Alfred’, Anglo-Saxon England 39: 193–225. Godden, Malcolm and Susan Irvine (eds.), (2009). The Old English Boethius: An Edition of the Old English Versions of Boethius’s De Consolatione Philosophiae. 2 vols. Oxford and New York: Oxford University Press. Gradon, Pamela (1979). Dan Michel’s Aʒenbite of Inwyt. Voume II: Introduction, Notes and Glossary. (Early English Text Society, 278). Oxford: Oxford University Press. Griffith, Mark (ed.), (1997). Judith (Exeter Medieval English Texts and Studies). Exeter: University of Exeter Press. Grimmer, Martin (2007). ‘Britons in Early Wessex: The Evidence of the Law Code of Ine’, in Higham, Nick (ed.), Britons in Anglo-Saxon England (Publications of the Manchester Centre for Anglo-Saxon Studies 7). Manchester: Boydell Press, 102–14. Hall, J. R. Clark (1960). A Concise Anglo-Saxon Dictionary. With a Supplement by Herbert D. Meritt. 4th edn. Toronto: University of Toronto Press. Hall, Joseph (ed.), (1920). Selections from Early Middle English 1130–1250. Oxford: Clarendon Press. Harbert, Wayne (2007). The Germanic Languages. Cambridge: Cambridge University Press. Haspelmath, Martin (1999). ‘External Possession in a European Areal Perspective’, in Payne, Doris L. and Immanuel Barshi (eds.), External Possession (Typological Studies in Language, 39). Amsterdam and Philadelphia: John Benjamins, 109–35. Havers, Wilhelm (1911). Untersuchungen zur Kasussyntax der Indogermanischen Sprachen. Strassburg: Karl J. Trübner. Hecht, Hans (ed.), (1965). Bischof Waerferths von Worcester Übersetzung der Dialoge Gregors des Grossen über des Leben und die Wundertaten Italienischer Vater und über die Unsterblichkeit der Seelen. Darmstadt: Wissenschaftliche Buchgesellschaft. Heine, Bernd (1997). Possession: Cognitive Sources, Forces, and Grammaticalization. (Cambridge Studies in Linguistics, 83). Cambridge: Cambridge University Press. Hickey, Raymond (2012). ‘English and the Celtic Hypothesis’, in Nevalainen, Terttu and Elizabeth Closs Traugott (eds.), The Oxford Handbook of the History of English. Oxford: Oxford University Press, 497–507. Hogg, Richard (1997). ‘Some Remarks on Case Marking in Old English’, Transactions of the Philological Society 95 (1): 95–109. Irvine, Susan (ed.), (2004). The Anglo-Saxon Chronicle: A Collaborative Edition. Volume 7: MS E. Cambridge: D. S. Brewer. Kayne, Richard S. (1975). French Syntax: The Transformational Cycle. Cambridge, MA: MIT Press. Kelly, Richard J. (2003). The Blickling Homilies. Edition and Translation. London: Continuum. Ker, Neil R. (1957). Catalogue of Manuscripts Containing Anglo-Saxon. Oxford: Clarendon Press.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

274



König, Ekkehard (2001). ‘Internal and External Possessors’, in Haspelmath, Martin, et al. (eds.), Language Typology and Language Universals. Berlin and New York: Walter de Gruyter, 970–8. König, Ekkehard and Martin Haspelmath (1998). ‘Les Constructions à Possesseur Externe dans les Langues d’Europe’, in Feuillet, Jack (ed.), Actance et Valence dans les Langues de l’Europe. Berlin and New York: Mouton de Gruyter, 525–606. Krahe, Hans (1972). Grundzüge der vergleichenden Syntax der indogermanischen Sprachen (Innsbrucker Beträge Zur Sprachwissenschaft, 8). Innsbruck: Institut für Vergleichende Sprachwissenschaft der Universität Innsbruck. Krapp, George Philip (ed.), (1931). The Junius Manuscript (The Anglo-Saxon Poetic Records, 1). New York: Columbia University Press. Krapp, George Philip (ed.), (1932). The Vercelli Book (The Anglo-Saxon Poetic Records, 2). London: Routledge and Kegan Paul. Krapp, George Philip and Elliott Van Kirk Dobbie (eds.), (1936). The Exeter Book (AngloSaxon Poetic Records, 3). New York: Columbia University Press. Kroch, Anthony and Ann Taylor (2002). ‘The Penn-Helsinki Parsed Corpus of Middle English 2’ (PPCME2). Department of Linguistics, University of Pennsylvania. CD-ROM, 2nd edn (http://www.ling.upenn.edu/hist-corpora/). Kytö, Merja (1993). Manual to the Diachronic Part of the Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Laing, Margaret (1992). ‘A Linguistic Atlas of Early Middle English: The Value of Texts Surviving in More Than One Version’, in Rissanenen, Matti, et al. (eds.), History of Englishes: New Methods and Interpretations in Historical Linguistics (Topics in English Linguistics, 10). Berlin: Mouton de Gruyter, 566–81. Laing, Margaret (1993). Catalogue of Sources for a Linguistic Atlas of Early Medieval English. Cambridge and Rochester, NY: D.S. Brewer. Lamiroy, Béatrice (2003). ‘Grammaticalization and External Possessor Structures in Romance and Germanic Languages’, in Coene, Martine, Yves D’hulst, and Liliane Tasmowski (eds.), From NP to DP Vol. II: The Expression of Possession. Amsterdam: John Benjamins, 257–80. Landau, Idan (1999). ‘Possessor Raising and the Structure of VP’, Lingua 107: 1–37. Langacker, Ronald (1993). ‘Reference-Point Constructions’, Cognitive Linguistics 4: 1–38. Le Bruyn, Bert (2014). ‘Inalienable Possession: The Semantics of the Definite Determiner’, in Agular-Guevara, Ana, Bert Le Bruyn, and Joost Zwarts (eds.), Weak Referentiality. Amsterdam and Philadelphia: John Benjamins, 311–34. Le Saux, Françoise (1989). Laʒamon’s Brut: The Poem and its Sources. Cambridge: D.S. Brewer. Lee-Schoenfeld, Vera (2006). ‘German Possessor Datives: Raised and Affected’, Journal of Comparative Germanic Linguistics 9: 101–42. A Linguistic Atlas of Early Modern English 1150–1325 (2013–). Compiled by Margaret Laing. Version 3.2. Edinburgh: The University of Edinburgh. http://www.lel.ed.ac.uk/ ihd/laeme2/laeme2.html/. Liuzza, Roy Michael (2000). ‘Scribal Habit: The Evidence of the Old English Gospels’, in Swan, Mary and Elaine M. Treharne (eds.), Rewriting Old English in the Twelfth Century. Cambridge and New York: Cambridge University Press, 143–65. Lockwood, W. B. (1968). Historical German Syntax. Oxford: Clarendon Press. Lødrup, Helge (2009a). ‘Looking Possessor Raising in the Mouth: Norwegian Possessor Raising with Unergatives’, in Butt, Miriam and Tracy Holloway King (eds.), LFG2009, LFG09, 420–40. Lødrup, Helge (2009b). ‘External and Internal Possessors with Body Part Nouns: The Case of Norwegian’, SKY Journal of Linguistics 22: 221–50. (http://csli-publications. stanford.edu/).

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi



275

Lyons, John (1967). ‘A Note on Possessive, Existential and Locative Sentences’, Foundations of Language 3: 390–6. Maling, Joan (1983). ‘Transitive Adjectives: A Case of Categorial Reanalysis’, in Heny, Frank and Barry Richards (eds.), Linguistic Categories: Auxiliaries and Related Puzzles. Boston and Dordrecht, Holland: D. Reidel, 253–89. Maling, Joan (2001). ‘Dative: The Heterogeneity of the Mapping among Morphological Case, Grammatical Functions, and Thematic Roles’, Lingua 111 (4–7): 419–64. McIntosh, Angus, M. Samuels, Michael Benskin, Margaret Laing, and K. Williamson (1986). A Linguistic Atlas of Late Mediaeval English. Aberdeen: Aberdeen University Press. McWhorter, John (2002). ‘What Happened to English?’, Diachronica 19: 217–72. Middle English Dictionary (1956). Edited by Hans Kurath and Sherman Kuhn. Ann Arbor: University of Michigan Press. Free access to online version at https://quod.lib.umich. edu/m/med/. Migne, Jacques Paul (1857). Patrologiae Cursus Completus: Sive Bibliotheca Universalis, Integra, Uniformis, Commoda, Oeconomica, Omnium Ss.Patrum, Doctorum Scriptorumque Ecclesiasticorum, Vol. 31. Paris: Petit-Moutrouge. Miller, Thomas (ed.), (1890). The Old English Version of Bede’s Ecclesiastical History of the English People (Early English Text Society, 95–6 and 110–11). London: N. Trübner and Co. Millett, Bella (2005). Ancrene Wisse: A Corrected Edition of the Text in Cambridge, Corpus Christi College, MS 402, with Variants from Other Manuscripts. Drawing on the Uncompleted Edition by E.J. Dobson with a Glossary and Additional Notes by Richard Dance, 2 vols. (Early English Text Society, 325–6). Oxford: Oxford University Press. Millett, Bella (ed.), (1982). Hali Meiðhad (Early English Text Society, 284). Oxford: Oxford University Press. Mitchell, Bruce (1985). Old English Syntax. Oxford: Clarendon Press. Miura, Ayumi (2014). Middle English Verbs of Emotion and Impersonal Constructions: Verb Meaning and Syntax in Diachrony. Oxford: Oxford University Press. Moricca, Umberto (ed.), (1924). Gregorii Magni Dialogi. Rome: Instituto Storico Italiano. Morris, Richard (ed.), (1867–8). Old English Homilies of the 12th and 13th Centuries (Vol. I) (Early English Text Society, 29, 34). London: N. Trübner & Co. Mustanoja, Tauno F. (1960). A Middle English Syntax. Helsinki: Société Néophilologique. Nokes, Richard Scott (2004). ‘The Several Compilers of Bald’s Leechbook’, Anglo-Saxon England 33: 51–76. OED Online (2018). Oxford: Oxford University Press. Payne, Doris L. and Immanuel Barshi (1999). ‘External Possession: What Where, How, and Why’, in Payne, Doris L. and Immanuel Barshi (eds.), External Possession (Typological Studies in Language, 39). Amsterdam and Philadelphia: John Benjamins, 3–29. Pintzuk, Susan and Leendert Plug (2001). ‘The York-Helsinki Parsed Corpus of Old English Poetry’; York: University of York Department of Language and Linguistic Science. Available through Oxford Text Archive (http://www-users.york.ac.uk/~lang18/pcorpus.html). Plummer, Charles (ed.), (1892). Two of the Saxon Chronicles Parallel. Oxford: Clarendon Press. Pokorny, Julius (1927–30). ‘Das Nicht-Indogermanische Substrat im Irischen’, Zeitschrift für Celtische Philologie 16–18: 16: 95–144, 231–66, 363–94; 17: 73–88; 18: 233–48. Pope, John Collins (ed.), (1967). Homilies of Ælfric (Early English Text Society, 259). London: Oxford University Press. Potter, Simeon (1931). On the Relation of the Old English Bede to Werferth’s Gregory and to Alfred’s Translations, Mémoires de la Société Royale des Sciences de Bohême 1–76. R Core Team (2017). ‘R: A Language and Environment for Statistical Computing’; R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

276



Rissanen, Matti and Ossi Ihalainen (1991). Helsinki Corpus of English Texts: Diachronic and Dialectal. Helsinki: University of Helsinki. Roberts, Jane Annette, Christian Kay, and Lynne Grundy (2000). A Thesaurus of Old English, Two vols. 2nd edn. (Costerus Ns 131–2.) Amsterdam and Atlanta: Rodophi. Online version available at https://oldenglishthesaurus.arts.gla.ac.uk/. Robinson, Maurice and Willaim G. Pierpont (2005). The New Testament in the Original Greek: Byzantine Textform 2005. Southborough, MA: Chilton Book Publishing. Schiffels, Stephan, Wolfgang Haak, Pirita Paajanen, Bastien Llamas, Elizabeth Popescu, Louise Loe, Rachel Clarke, Alice Lyons, Richard Mortimer, Duncan Sayer, Chris TylerSmith, Alan Cooper, and Richard Durbin (2016). ‘Iron age and Anglo-Saxon Genomes from East England Reveal British Migration History’, Nature Communications, 7. doi: 10.1038/ncomms10408. Accessed 4 December 2017. Scholten, Jolien (2018). The Ins And Outs of External Possession: A Micro-Comparative Perspective. Utrecht: LOT. Schrijver, Peter (2014). Language Contact and the Origins of the Germanic Languages. (Routledge Studies in Linguistics, 13). New York and London: Routledge. Scragg, Donald G. (1991). ‘The Nature of Old English Verse’, in Godden, Malcolm and Michael Lapidge (eds.), The Cambridge Companion to Old English Literature. Cambridge: Cambridge University Press, 55–70. Sievers, Eduard (ed.), (1878). Heliand. Halle: Buchhandlung des Waisenhauses. Sisam, Celia (1953). ‘An Early Fragment of the Old English Martyrology’, The Review of English Studies, New Series 4 (15): 209–20. Stanley, Eric G. (1968). ‘The Date of Laʒamon’s Brut’, Notes and Queries 213: 85–8. Stanley, Eric (1969). ‘Laʒamon’s Antiquarian Sentiments’, Medium Ævum 38: 32–7. Stokes, Peter A. (2011). ‘The Vision of Leofric: Manuscript, Text and Context’, Review of English Studies 63: 529–50. Streitberg, Wilhelm August (ed.), (1919). Die Gotische Bibel. 2 vols. Heidelberg: Winter. Sweet, Henry (ed.), (1885). The Oldest English Texts (Early English Text Society, 83). London: Oxford University Press. Taylor, Ann (2008). ‘Contact Effects of Translation: Distinguishing Two Kinds of Influence in Old English’, Language Variation and Change 20: 341–65. Taylor, Ann (2014). ‘Old English Syntax’, in Ringe, Donald and Ann Taylor (eds.), The Development of Old English. Oxford: Oxford University Press. Taylor, Ann, Anthony Warner, Susan Pintzuk, and Frank Beths (2003). The York-TorontoHelsinki Parsed Corpus of Old English Prose (YCOE). Department of Language and Linguistic Science, University of York. Oxford Text Archive, 1st edn (http://www-users. york.ac.uk/~lang22/YcoeHome1.htm). Thomason, Sarah and Terrence Kaufman (1988). Language Contact, Creolization, and Genetic Linguistics. University of California Press. Thráinsson, Höskuldur (2007). The Syntax of Icelandic. Cambridge: Cambridge University Press. Tolkien, J. R. R. (1963). ‘English and Welsh’, in Lewis, Henry (ed.), Angles and Britons. The O’Donnell Lectures. Cardiff: University of Wales Press, 1–41. Treharne, Elaine M. (2000). ‘The Production and Script of Manuscripts Containing English Religious Texts in the First Half of the Twelfth Century’, in Swan, Mary and Elaine M. Treharne (eds.), Rewriting Old English in the Twelfth Century (Cambridge Studies in Anglo-Saxon England 30). Cambridge and New York: Cambridge University Press, 11–40.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi



277

Treharne, Elaine M. (2001). ‘English in the Post-Conquest Period’ [online text], in Pulsiano, Phillip and Elaine M. Treharne (eds.), A Companion to Anglo-Saxon Literature, Chapter 22. Oxford and Malden, MA: Blackwell Reference Online. doi:10.1111/b.9780631209041. 2001.00026.x. Accessed 10 November 2013. Treharne, Elaine M. (2012). Living through Conquest: The Politics of Early English, 1020–1220. Oxford: Oxford University Press. Tristram, Hildegard (2004). ‘Diglossia in Anglo-Saxon England, or What Was Spoken Old English Like’, Studia Anglica Posnaninsia 40: 87–110. Van Bergen, Linda (2002). Pronouns and Word Order in Old English, with Particular Reference to the Indefinite Pronoun Man. New York and London: Routledge. Van de Velde, Freek and Béatrice Lamiroy (2016). ‘External Possessors in West Germanic and Romance: Differential Speed and Drift Towards NP Configurationality’, in Cuyckens, Hubert, Lobke Ghesquière, and Daniël Van Olmen (eds.), Aspects of Grammaticalization: (Inter)Subjectification, Analogy and Unidirectionality, 353–99. Berlin: Mouton de Gruyter. Vandeweghe, Willy (1987). ‘The Possessive Dative in Dutch: Syntactic Reanalysis and Predicate Formation’, in Auwera, Johan van der and Louis Goossens (eds.), Ins and Outs of Predication. Dordrecht: Foris, 137–51. Vennemann, Theo (2012). ‘On the Rise of “Celtic” Syntax in Middle English’, in Noel Aziz Hanna, Patrizia (ed.), Germania Semitica (Trends in Linguistics: Studies and Monographs, 259). Berlin and Boston: Mouton de Gruyter, 147–77. Vennemann, Theo (2002). ‘On the Rise of “Celtic” Syntax in Middle English’, in Lucas, Peter J. and Angela M. Lucas (eds.), Middle English from Tongue to Text: Selected Papers from the Third International Conference on Middle English: Language and Text, Held at Dublin, Ireland, 1–4 July 1999 (Studies in English Medieval Language and Literature 4). Bern: Peter Lang, 205–34. Vergnaud, Jean-Roger and Maria Luisa Zubizaretta (1992). ‘The Definite Determiner and the Inalienable Construction in French and English’, Linguistic Inquiry 23: 595–652. Visser, F. Th. (1963). An Historical Syntax of the English Language. Leiden: Brill. Wal, Marijke J. van der and Aad Quak (1994). ‘Old and Middle Continental West Germanic’, in König, Ekkehard and Johan van der Auwera (eds.), The Germanic Languages. London and New York: Routledge, 72–109. Walkden, George (2014). Syntactic Reconstruction and Proto-Germanic. (Oxford Studies in Diachronic and Historical Linguistics, 12). Oxford: Oxford University Press. Walkden, George (2015). HeliPaD: The Heliand Parsed Database. Version 0.9. http://www. chlg.ac.uk/helipad/. Westhoff, Elbert Wilhelm (ed.), (1860). S. Gregorii Papae I Cognomento Magni De Pastorali Cura Liber. Monasterii Westphalorum: Aschendorff. Wierzbicka, Anna (1979). ‘Ethno-Syntax and the Philosophy of Language’, Studies in Language 3 (3): 313–83. Yerkes, David (1982). Syntax and Style in Old English. Binghamton, New York: Center for Medieval and Renaissance Studies. Younge, George (2012). ‘ “Those Were Good Days”: Representations of the Anglo-Saxon Past in the Old English Homily of Saint Neot’, Review of English Studies 63: 349–69. Zaenen, Annie, Joan Maling, and Höskuldur Thráinsson (1985). ‘Case and Grammatical Functions: The Icelandic Passive’, Natural Language and Linguistic Theory 3: 441–83.

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

Index adjectival and nominal predicates, see copula adjectives, appositive, and DEPs 92, 97–8, 99, 185, 188, 206 Ælfric 48, 110, 114, 159, 221, 242 limited use of DEPs 79, 80–1, 111, 117–18, 263 affectedness and DEPs/IPs: in Common Germanic (reconstructed) 101–3, 106, 107, 108, 233 cross-linguistically 6–7, 13, 52–3, 100 with direct argument possessa in Old English: assumption that DEPs were dominant in early Old English 57, 219–21, 229 beneficial effect or lack of effect 67, 74, 222 with body possessa 40, 63–4, 67–71, 74–9, 93–4, 100, 162, 196, 225–6, 229–31 subject possessa less constrained than object possessa by affectedness constraint 74, 84, 106, 108 mind possessa less constrained than body possessa by affectedness constraint 108, 121, 123, 127, 148 in Middle English 154, 155–6, 163–4, 177, 181–3, 193, 196, 206, 226 with PObj possessa in Old English 96–7, 100, 101 see also copula; EPs in languages other than English; grammatical relations and EPs ‘affecting’ verbs 52, 54, 57, 67, 252–7, 261, 263–4 Aʒenbite of Inwyt 166, 171 Ahlgren, Arthur 2, 19–20, 39, 59, 88, 99, 120, 188, 192, 193, 199, 201, 205–6, 207, 208, 214–15, 221, 229, 231, 232, 233 see also blended construction; deflexion Amos, Ashley Crandall 50

anchor manuscripts 48, 239 Ancrene Wisse 194 Anglo-Saxon Chronicle: Early West Saxon portion 48–9, 239 post Early West Saxon, MSS A-D 49, 154, 241–2 Peterborough Chronicle (cochronE) 154, 157 n5, 158, 160, 163, 167, 209, 217, 241–2, 243–4 animacy of possessor 6, 64, 67, 87, 267–8; see also affectedness arguments, thematic and non-thematic 8–9, 11, 55–6, 184, 196–7, 203 n2, 225–6, 233 Bald’s Leechbook: date of composition 49 DEPs vs IPs in 28, 67–8, 76, 81, 85, 97, 99 bare NP, definition of 166 Barshi, Immanuel, see Payne, Doris L. and Barshi, Immanuel Bately, Janet 89, 238, 239, 241 Bede’s Ecclesiastical History of the English People 89, 241 ‘blended’ possessive construction: Ahlgren’s definition and discussion 39, 112, 162, 232 in Middle English 180, 182, 187, 197, 227 in Old English 39–40, 112, 162, 232 in copular constructions 84, 125, 140, 147, 162 with direct argument possessa 59, 80, 112, 127, 140 with PObj possessa 112, 114 in Old Saxon 107, 109, 112 Bolkestein, A. Machtelt 83 Brittonic English 220–1 Campbell, Alistair 26, 151 n1, 238 Cartlidge, Neil 246

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

280



case: ambiguity in marking and distinguishing DEPs vs IPs: in Early Middle English 155, 179 in Old English 59–60, 126, 139–40, 261–2, 266 defining dative case in Early Middle English 4, 166–167, 175, 200–1 instrumental 16, 25–7, 30, 192, 211 variation in historically accusative and dative forms 167 see also deflexion case-rich vs case-impoverished Early Middle English dialects/texts: definition of 167–8 similar frequency of DEPs in both types 179–81, 206–8, 232–3 use of old dative functions in 169–73, 176, 208 Clement, Richard 89 n5 Clemoes, Peter 118, 244 configurationality 213–14 copula 34, 54 and adjectival and nominal predicates 81, 139, 225 and affectedness 14, 83, 225 in descriptions 82, 107 with DEPs expressing relationships 36–8 and DEPs in Middle English 162, 163, 170–3 mind possessa different from body possessa 124–6, 135–6 Old English compared with other early Germanic languages 109 Old English prose different from poetry 133–4 PObj possessa, special characteristics of 136–7, 140, 144–5, 194–5, 222, 230, 232, 233 see also blended construction; extended existence; impersonal constructions Crisma, Paola 155, 209–11, 213, 231 Cura Pastoralis (CP, Pastoral Care) 89, 91, 93, 131, 238, 239

Dance, Richard 194 dating systems for manuscripts and texts: Early Middle English 150 Old English 48 dative absolute 89 dative complements of adjectives and nouns 31–3, 35–7, 134, 136, 170–2, 227 dative objects of monotransitive verbs 30, 56 dative possessum 29 dativus sympatheticus, see sympathetic dative Deal, Amy Rose 6, 7 decline and disappearance of DEPs: in Middle English 120, 179–83, 184–90, 195–8, 199, 205–6, 208–9, 224–5, 226–7, 232 within /before the Old English period: (possible) change before written Old English 108–9, 133 attested change within Old English 109, 120, 205, 222–3, 226, 229, 232 with body direct argument possessa 110–13, 195–6 with mind possessa 146–8, 196, 197, 230 in (North Germanic) Scandinavian languages 201–2, 234–5 with PObj possessa 114–20, 147–8, 163, 196–7 definite article/determiner: cross-linguistic use in DEPs 11–12 in early English prose DEPs of direct arguments 155, 209–12, 231, 260–1 in Old English poetry 50, 126, 213 deflexion: as an explanation for the loss of DEPs 173, 201–3, 205–9, 227 generative approaches 201, 205 Ahlgren’s view 2, 19–20, 199, 201, 205–8 timing of 151, 153, 161, 165–8, 175, 200–1, 205–6, 207–8 DEPs (dative external possessors) with body words vs mind words 131, 132, 133, 148, 161, 226–7, 230 determiner phrase 3 n3

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

 Dialect AB 167, 169, 170, 174 Dictionary of Old English (DOE) text titles 241 entries for individual words: ascufan 117 ahlyhhan 128 n2 beceorfan 16, 29 becuman 116 beheawan 211 beon 33 n4 beswapan 138 beyrnan 138 blod (for)lætan 28 cwician 83 for-brecan 90 gan 116 gebigan 115 geomor 125–6 gram 139 n4 habban 42 n8 direct argument possessa, definition of 7, 51 discontinuous possessive phrase 105, 140, 214, 261 distributivity effect 7 n7, 60–1 ditransitive verb 261–2 with dative human object 27–29, 38, 65, 79, 131 expressing deprivation 16–17, 18–19, 38, 65, 131, 174, 177 with genitive object 29–30, 174 order of objects 168 Einenkel, Eugen 180 Elsweiler, Christine 246 EPs (external possessors, including DEPs): definitions of 3–6, 8–9 difficulties in identifying in early English 27–9, 55, 59–60, 135, 137 excluded constructions and relationships 9, 18–19, 28–9, 55, 57, 261–4; see also ditransitive verb in languages other than English 61 Basque 209 Breton 219 Dutch 11–12, 203–5, 208, 213, 224, 227 n10, 235 French 6, 7, 60, 92, 213, 227 n10, 235

281

German 1, 6, 7, 105, 205, 213, 223, 234 Gothic 102–6, 107, 109, 120, 223 Greek, Ancient 21, 101 Icelandic 201–2 Norwegian 16, 184, 202, 203, 205, 235 Old Norse 15, 201–2 Old Saxon 105–8; see also blended possessive construction in Old Saxon Spanish 234, 235 see also affectedness in DEPs vs IPs in Common Germanic; locative constructions in Modern English/Present Day English 8, 14, 184, 192–3 entailment and EPs 8–9, 55 external possessor, see EP ethic(al) datives 11, 18, 40–1, 42–3 expletive determiner 10–12, 209–12, 233 expletive subject 32, 33–4 extended existence, dative in 33–7, 96, 134, 136, 172–3 Ferraresi 102 Filppula, Markku 2, 22, 101, 203, 215, 217, 221, 234 Fulk, R. D. 50 genitive case: and possessive pronouns 4, 15, 17 on verbal objects 30, 174 see also pronominal vs nominal possessors Giusti, Giuliana 11 Godden, Malcolm 237, 244 Gregory’s Dialogues 49, 89, 92, 93, 135, 143, 240 Gradon, Pamela 166 grammatical relations and EPs: cross-linguistically 13–14, 45, 51, 82 diachronic origin of restrictions 14, 82, 196–7, 230 frequency of DEPs of object possessa vs subject possessa 73–5, 85–6, 94, 100, 106, 108, 113, 122, 148 PObj (object of preposition) possessa different from direct argument possessa:

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

282



grammatical relations and EPs: (cont.) in Gothic 104 in Middle English: frequency compared with direct argument body possessa 157–8, 185–6, 188, 196–7, 226 frequency compared with direct argument mind possessa 193, 196–7, 226 in Modern/ Present Day English 7–9, 14 in Old English: frequency compared with direct argument body possessa 20, 45, 230–1 frequency compared with direct argument mind possessa 137, 140–1, 148, 265 frequency in poetic vs prose texts 99–100 semantic range less restricted than with direct argument body possessa 95–7, 196, 231, 268 see also lexical association of DEPs with verbs/prepositions see also arguments, thematic and nonthematic; copula; subject possessa (intransitive verbs) Grimmer, Martin 218–19 Hall, Joseph 244 Harbert, Wayne 103–4 Haspelmath, Martin 1, 4, 6, 10, 13, 53, 61, 82, 95, 96, 201, 203, 204, 209, 219, 223 Havers, Wilhelm 3, 15–19, 21, 30, 86–7, 101, 104–5, 107, 126 n1, 128–9, 220, 223, 229, 230, 234, 235 Heliand, see Old Saxon Hickey, Raymond 217 Hogg, Richard 151 n1 idioms, see lexical association of DEPs with verbs and prepositions ‘impersonal’ constructions: in Icelandic 32–3 in Middle English 170, 178 n12, 181 in Old English 31–2, 43–4, 124, 135 implicit possessors 9, 136–7 in Dutch 12 in Latin 91

in Middle English 189 in Old English 10–11, 147–8, 157, 267 possible correlation with DEPs 10–11, 12, 231–2 analysis as a syntactic binding relationship 10 analysis as a pragmatic relationship 11, 27 inalienable possessa and external possession 3, 12, 13 Ine’s laws 218 instrumental case, see case internal possessor, see IP IP (internal possessor): default possessor in Old English 52, 94, 116, 230 definition of 3 excluded where DEP not likely 56, 128–30, 252, 262–4, 267–8 limited investigation of PObj possessa 183–4, 264–5, 269–70 see also affectedness Kaufman, Terrence 216 Ker, Neil R. 49, 237, 239, 240, 241, 243 Kayne, Richard 92 Kelly, Richard J. 144 n6 König, Ekkehard 5, 8, 9–10, 13, 61, 201, 202, 203, 204, 219, 223 Krahe, Hans 6, 42 LAEME (Linguistic Atlas of Early Middle English) 245–6 lætan blod, see EPs (excluded constructions and relationships) Laing, Margaret 48, 245 Lamiroy, Béatrice 213–14, 224 n9, 227 n10, 234, 235 Landau, Idan 5, 6, 7 Langacker, Ronald 85 language contact and syntactic change 5, 199–200, 205, 214, 216, 218, 235 and continental European languages 224 proposals for English: Brythonic Celtic 2, 21–2, 106, 214, 217–22, 233–4 Latin 88, 214–15, 233; see also translation effects Norse (Scandinavian) 2, 199, 202, 214, 216–17, 233

OUP CORRECTED PROOF – FINAL, 12/9/2019, SPi

 Latin, Medieval: DEPs in 83, 89–91, 93, 215–16 implicit possessors in 91, 92 see also translation effects laws, treatment of possessors in 70, 93 Laʒamon’s Brut 164, 168, 171, 181–2, 190, 194, 197, 245, 246 Le Bruyn, Bert 12 n8 Lee-Schoenfeld, Vera 7, 11, 27, 39, 41 lexical association of DEPs with specific words 120, 226–7 combinations of verb and PObj 119, 141 favoured verbs 70–1, 79–81, 113, 132, 187–8 idioms with PObjs: Old English 115–17, 138, 143, 191 Middle English 189–190 see also affectedness and DEPs/IPs; EPs in Modern/Present Day English Lexical-Functional Grammar 32, 203 Life of St Margaret 153, 154–6, 158–61, 243, 244 Life of St Neot 162, 243, 244 Liuzza, Roy 152 n2 locative constructions expressing possession 15–17, 55, 82, 166 in Old English 17, 30 in Scandinavian (North Germanic) languages 202–3, 216–17, 234 Lockwood, W. B. 40, 41 Ldrup, Helge 13, 16, 202, 203, 225, 235 Maling, Joan 25 n1, 44, 172 martyrologies 90, 206, 241 McIntosh, Angus 48 McWhorter, John 2, 201, 203, 216–17 metaphorical use of body words 121, 130, 132, 184, 189–90, 268 Middle English Dictionary as data source 174, 178, 190, 194 Millett, Bella 178, 194 Mitchell, Bruce 20–1, 22, 25, 39, 41, 45, 169 n9, 260 ‘modal’ be 169–70 Mustanoja, Tauno 21, 195, 229 nouns of possession, definition of 19 object incorporation 211–12 Ormulum 4, 164, 167, 169, 170, 176, 209, 217

283

Orosius 68–9, 88–9, 91, 93, 99, 116, 119, 238, 239 The Owl and the Nightingale 182, 190, 197, 246 The Oxford Dictionary of Old English as a data source 41, 191–3 Payne, Doris and Barshi, Immanuel 5, 6, 13, 14, 82, 85 Penn-Helsinki Parsed Corpus of Middle English (PPCME2) 23 coding of DEPs 175, 186, 269 coding differences with YCOE 175 corpus size 164 periphrastic do 220, 234 PObj (object of preposition), see grammatical relations and EPs poetry in early English: Early Middle English poetry as a data source 165, 181–2, 190 Old English: dates of composition 50 differences with prose: accusative and dative pronominal forms 208 determiners, see definite article/ determiner and early language 35–6, 50, 58, 69, 112, 133, 144, 149 frequency of DEPs 99, 130, 131, 133–4, 144, 230 implicit possessors 157 not subject to Latin influence 79, 113, 130, 149 see also texts and manuscripts Pokorny, Julius 217 Pope, John Collins 46, 242 Poss Det construction 160–1 possessive/possessor dative 3, 202 n1 possessor raising analyses 6, 7, 13, 35, 225, 235 Potter, Simeon 89 predicative possession 34, 38, 82, 134 pre-m1 period 151–3, 189 definition of 150, 151, 243–4 pronominal vs nominal possessors 17, 86–7, 205 Quak, Aan 42

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

284



reflexive verbs 31, 42 n8, 43, 127, 129 n3, 267 Roberts, Jane 51, 247 Schiffels, Stephan 218 Scholten, Jolien 11–12, 204–5, 210, 213, 224 Scragg, Donald G. 50 ‘shared’ dative 28, 96, 135, 147, 261–2 Shrijver, Peter 220 spoken vs written early English: Middle English 156, 160, 165, 181–2 Old English 134, 215, 234 see also language contact and syntactic change Stanley, Eric G. 165 stative verbs, DEPs with 82; see also copula subject possessa (intransitive verbs): passives and unaccusative verbs 13, 14, 53, 54, 73 unergative verbs 13, 53, 54, 252 subject possessa (transitive verbs) 53 sympathetic dative: definition of 3 types of (Havers) 15–19 Taylor, Ann 50, 88 texts and manuscripts: additions to parsed texts 47, 164, 181, 245 copied texts as sources of data for Early Middle English 152, 156–7, 160–1, 165, 167–8, 243 dialect mixing 46, 49, 110–11 exclusions from this study 46, 56, 88, 164 n7, 237, 266–7 grouping into types 47–50 as sources of data for Old English 23, 46, 69, 116, 237, 241 see also dating systems for manuscripts and texts; poetry; spoken vs written language A Thesaurus of Old English 51, 247 Thomason, Sarah 216 Thráinsson, Höskuldur 32, 33, 44, 202 n1 Tolkien, J. R. R. 220 topicality 40, 85–7 translation effects: on early English: Bible translations 88, 90–1, 181 French 181

Latin 191 Middle English 169 n9, 179–80 Old English 79, 83–4, 88–94, 113, 130–1, 143, 148–9 on Gothic 102, 104, 108 Treharne, Elaine 152, 244 Van Bergen, Linda 87 Van Coetsem, Frans 200, 218 Van de Velde, Freek 213–14, 224 n9, 234, 235 Vandeweghe, Willy 203–4, 205 variation in individual texts/authors 46, 68–9, 92, 98, 117–18, 171–2 see also Ælfric; Bald’s Leechbook; Orosius; EPs in languages other than English (Gothic, Old Saxon) Vennemann, Theo 2, 6, 21–2, 52, 88, 101, 217, 219–23 Vergnaud, Jean-Roger and Zubizaretta, Maria Luisa 3, 6–7, 8, 10, 55, 60–1, 92, 98, 209, 212 Vices and Virtues 167 Vision of Leofric 157, 244 Visser, F. Th. 20, 25, 29, 34, 35, 36, 37, 41, 43, 169, 172–3 Wal, Marijke van der 42 Walkden, George 50, 105 n2 West Saxon Standard 3, 151, 152, 221, 228, 234 Worcester Fragments 156, 244 Wulfstan 48, 49, 153, 163, 237, 242 Yerkes, David 49, 240 York-Toronto-Helsinki Parsed Corpus of Old English Poetry (York Poetry Corpus) 23, 47, 146, 259 York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE) 23 coding of case 16 n10, 26, 60, 155, 259–60 coding of copulas 54 corpus size 164 Early Middle English texts included 152 editions used 46, 237 treatment of dative types 29, 261–2 see also texts and manuscripts Zaenen, Annie 44 Zubizaretta, Maria Luisa, see Vergnaud, Jean-Roget and Zubizaretta, Maria Luisa

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

OXFORD STUDIES IN DIACHRONIC AND HISTORICAL LINGUISTICS General Editors Adam Ledgeway and Ian Roberts, University of Cambridge Advisory Editors Cynthia L. Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Cambridge  1 From Latin to Romance Morphosyntactic Typology and Change Adam Ledgeway 2 Parameter Theory and Linguistic Change Edited by Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar 3 Case in Semitic Roles, Relations, and Reconstruction Rebecca Hasselbach 4 The Boundaries of Pure Morphology Diachronic and Synchronic Perspectives Edited by Silvio Cruschina, Martin Maiden, and John Charles Smith 5 The History of Negation in the Languages of Europe and the Mediterranean Volume I: Case Studies Edited by David Willis, Christopher Lucas, and Anne Breitbarth 6 Constructionalization and Constructional Changes Elizabeth Traugott and Graeme Trousdale 7 Word Order in Old Italian Cecilia Poletto 8 Diachrony and Dialects Grammatical Change in the Dialects of Italy Edited by Paola Benincà, Adam Ledgeway, and Nigel Vincent

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

9 Discourse and Pragmatic Markers from Latin to the Romance Languages Edited by Chiara Ghezzi and Piera Molinelli 10 Vowel Length from Latin to Romance Michele Loporcaro 11 The Evolution of Functional Left Peripheries in Hungarian Syntax Edited by Katalin É. Kiss 12 Syntactic Reconstruction and Proto-Germanic George Walkden 13 The History of Low German Negation Anne Breitbarth 14 Arabic Indefinites, Interrogatives, and Negators A Linguistic History of Western Dialects David Wilmsen 15 Syntax over Time Lexical, Morphological, and Information-Structural Interactions Edited by Theresa Biberauer and George Walkden 16 Syllable and Segment in Latin Ranjan Sen 17 Participles in Rigvedic Sanskrit The Syntax and Semantics of Adjectival Verb Forms John J. Lowe 18 Verb Movement and Clause Structure in Old Romanian Virginia Hill and Gabriela Alboiu 19 The Syntax of Old Romanian Edited by Gabriela Pană Dindelegan 20 Grammaticalization and the Rise of Configurationality in Indo-Aryan Uta Reinöhl 21 The Rise and Fall of Ergativity in Aramaic Cycles of Alignment Change Eleanor Coghill

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

22 Portuguese Relative Clauses in Synchrony and Diachrony Adriana Cardoso 23 Micro-change and Macro-change in Diachronic Syntax Edited by Eric Mathieu and Robert Truswell 24 The Development of Latin Clause Structure A Study of the Extended Verb Phrase Lieven Danckaert 25 Transitive Nouns and Adjectives Evidence from Early Indo-Aryan John. J. Lowe 26 Quantitative Historical Linguistics A Corpus Framework Gard B. Jenset and Barbara McGillivray 27 Gender from Latin to Romance History, Geography, Typology Michele Loporcaro 28 Clause Structure and Word Order in the History of German Edited by Agnes Jäger, Gisella Ferraresi, and Helmut Weiß 29 Word Order Change Edited by Ana Maria Martins and Adriana Cardoso 30 Arabic Historical Dialectology Linguistic and Sociolinguistic Approaches Edited by Clive Holes 31 Grammaticalization from a Typological Perspective Edited by Heiko Narrog and Bernd Heine 32 Negation and Nonveridicality in the History of Greek Katerina Chatzopoulou 33 Indefinites between Latin and Romance Chiara Gianollo 34 Verb Second in Medieval Romance Sam Wolfe

OUP CORRECTED PROOF – FINAL, 29/8/2019, SPi

35 Referential Null Subjects in Early English Kristian A. Rusten 36 Word Order and Parameter Change in Romanian A Comparative Romance Perspective Alexandru Nicolae 37 Cycles in Language Change Edited by Miriam Bouzouita, Anne Breitbarth, Lieven Danckaert, and Elisabeth Witzenhausen 38 Palatal Sound Change in the Romance Languages Diachronic and Synchronic Perspectives André Zampaulo 39 Dative External Possessors in Early English Cynthia L. Allen   Redevelopment of Case Systems in Indo-Aryan Miriam Butt Classical Portuguese Grammar and History Charlotte Galves, Aroldo de Andrade, Christiane Namiuti, and Maria Clara Paixão de Sousa Morphological Borrowing Francesco Gardani Nominal Expressions and Language Change From Early Latin to Modern Romance Giuliana Giusti Syntactic Features and the Limits of Syntactic Change Edited by Jóhannes Gísli Jónsson and Thórhallur Eythórsson A Study in Grammatical Change The Modern Greek Weak Subject Pronoun τος and its Implications for Language Change and Structure Brian D. Joseph Reconstructing Pre-Islamic Arabic Dialects Alexander Magidow The History of Negation in the Languages of Europe and the Mediterranean Volume II: Patterns and Processes Edited by David Willis, Christopher Lucas, and Anne Breitbarth Variation and Change in Gallo-Romance Grammar Edited by Sam Wolfe and Martin Maiden