Grammaticalization at Work: Studies of Long-term Developments in English [Reprint 2011 ed.] 9783110810745, 9783110156317

Citation preview

Grammaticalization at Work

w DE

G

Topics in English Linguistics 24

Editor

Herman Wekker

Mouton de Gruyter Berlin · New York

Grammaticalization at Work Studies of Long-term Developments in English Edited by

Matti Rissanen Merja Kytö Kirsi Heikkonen

Mouton de Gruyter Berlin · New York

1997

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin.

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication

Data

Grammaticalization at work : studies of long-term developments in English / edited by Matti Rissanen, Merja Kytö, Kirsi Heikkonen. p. cm. - (Topics in English linguistics ; 24) Includes bibliographical references (p. ) and index. ISBN 3-11-015631-8 (cloth ; alk. paper) 1. English language - Grammar, Historical. I. Rissanen, Matti. II. Kytö, Merja. III. Heikkonen, Kirsi, 1 9 6 2 IV. Series. PE1075.G65 1997 425'.09-dc20 97-25408 CIP

Die Deutsche Bibliothek

— Cataloging-in-Publication

Data

Grammaticalization at work : studies of long term developments in English / ed. by Matti Rissanen ... — Berlin ; New York : Mouton de Gruyter, 1997 (Topics in English linguistics ; 24) ISBN 3-11-015631-8 D B N : 95.102826.X© SG: 52

© Copyright 1997 by Walter de Gruyter & Co., D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any f o r m or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Printing: Ratzlow-Druck, Berlin. Binding: Lüderitz & Bauer, Berlin. Printed in Germany.

Preface

This book is one of three volumes reporting the results of the project 'English in transition: Change through variation', carried out in the English Department of the University of Helsinki. The first volume, Early English in the computer age: Explorations through the Helsinki Corpus (ed. by Matti Rissanen, Merja Kytö and Minna Palander-Collin, Mouton de Gruyter, 1993) is now followed by two volumes, Grammaticalization at work: Studies of long-term developments in English and English in transition: Corpus-based studies in linguistic variation and genre styles. Both these volumes approach change in English from the angle of linguistic variation. The articles deal with processes of change in morphology, syntax and lexis, and pay special attention to the role played by textual and discourse factors across the centuries. From the methodological point of view, diachronic variation analysis and the multi-feature approach aiming at the identification of co-occurrence patterns in genres are the main frameworks adopted. The aim of the present volume is to give new insights into the development of adverbs and indefinite pronouns and in the means of reflexivization, in relation to various grammaticalization processes. The English in transition volume sheds light on the development of some central verb constructions (with BE and HAVE), expository apposition, and genre-specific features of expressions of affect and attitude in text. All the studies in these volumes are based on the Helsinki Corpus of English Texts; supplementary material has been drawn from other corpora and concordances and from primary texts outside the corpora. The 'English in transition' project was initiated in 1990 as a continuation of an earlier project which produced the Helsinki Corpus. The core team of both projects has been the same, consisting of the authors and editors of the volumes. The editors would like to express their special thanks to all research assistants of the projects and particularly to Arja Nurmi and Päivi Koivisto-Alanko for their excellent work in producing these volumes.

vi

Preface

We are most grateful to the Academy of Finland for funding our project for four years. We are indebted to the University of Helsinki for giving us research premises, and to the English Department for up-to-date technical facilities, travel grants and other support. Our thanks are due to Mrs Leena Sadeniemi for giving us expert advice in computer technology and training us to use programs. Finally, we would like to thank the editors of Mouton de Gruyter for accepting the two volumes now published in their Topics in English Linguistics series.

Helsinki, June 1996

M.R.

M.K.

K.H.

Contents

Introduction Matti Rissanen

Indefinite pronouns with singular human reference Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

17

The pronominalization of one Matti Rissanen

87

The processes of adverb derivation in Late Middle and Early Modern English Terttu Nevalainen

145

There compounds in the history of English Arne Österman

191

The development of reflexive strategies in English Kirsti Peitsara

277

A medieval case of grammaticalization, methinks Minna Ρalander-Collin

371

Bibliography Index of subjects

405 429

Introduction Matti

1.

Rissanen1

Introductory

The articles of the present volume deal with the formation of adverbs and indefinite and reflexive pronouns, with particular focus on grammaticalization. Studies of grammaticalization are often based on rather meagre empirical evidence and they tend to concentrate on individual linguistic elements. We hope that the present volume, which reflects the goals and principles of our research project, will add new dimensions to these studies. First of all, our studies take a long-term diachronic perspective, from Old or Middle English to the end of the 17th century or even beyond, always with reference to the Old English background of the developments. Most of them systematically look at long-term changes in a linguistic category or subcategory as a whole: adverbs, indefinite pronouns with personal reference, expression of reflexivity (see section 3 for our variationist approach). Studies dealing with the grammaticalization of individual items (METHINKS, ONE) link their discussion with the more general reorganization of the structure of English. By adding a more comprehensive diachronic approach to the study of grammaticalization, we hope to be able to place the operations of the various processes of grammaticalization in a wider perspective and in this way open up new insights into the history of English. We do not only trace the development of those grammaticalized items which survive in Present-day English, but we also attempt to analyse the factors which lead to the diminishing popularity and obsolescence of some of the forms and usages in past centuries. Particularly, we aim at illustrating the interde-

2

Matti Rissanen

pendence of structural systems in a diachronic perspective, not forgetting semantic and pragmatic considerations. Secondly, our studies are based on systematically collected empirical evidence derived from a large body of computer-readable texts. The improved possibilities for data retrieval and analysis offered by large corpora have encouraged us to apply a method of combining quantitative and qualitative analysis which, we hope, will offer a clearer and more 'lifelike' picture of developments which have taken place centuries — or even more than a millennium — ago. From a more heuristic point of view, careful analysis of corpus evidence calls attention to research problems that might not be discovered by less systematic data handling. The use of electronic corpora facilitates data retrieval but it also makes demands on our analytic tools. Our model of grammar would have to cope with the fuzziest realizations of the categories we are interested in, both changes in progress within these categories and the vagueness inherent in ordinary-language data (see further section 3).

2.

On grammaticalization

Of all the Germanic languages, English has undergone the most radical structural and lexical changes during its existence. A detailed survey of these changes is beyond the scope of this introductory note. It is obvious, however, that the levelling and loss of the inflectional endings, traceable even at the earliest stages of Old English and accelerating in Late Old English and Early Middle English, resulted in a gradual rearrangement of the grammatical system of English, which was finally established in the Modern period. The questions of causality and of the interdependence of the changes involved in this rearrangement have been a topic of lively scholarly debate in recent years, and very far-reaching conclusions should be avoided, especially as the diachrony of the various structural systems has not yet been adequately researched. It seems, however, that the cooccurrence of the following developments with the collapse of the system of endings can hardly be a coincidence:

Introduction

3

1.

Establishment of the S - V - 0 order, with the reorganization of the arguments of the verb

2.

Development of the auxiliary system

3.

Development of new pronominal forms, particularly those serving as the head of the noun phrase

4.

Development of new adverbial forms and functions

5.

Development of the prepositional system

6.

Development of new links between the elements of the sentence, both at clause and phrase level

All these developments are in one way or another connected with the concept of grammaticalization. According to the simple definition by Hopper—Traugott (1993: xv, 2), grammaticalization is the process in which lexical items and constructions come to serve grammatical functions, and once grammatical ized, continue to develop new grammatical functions. Grammaticalization is, of course, inseparable from change of meaning; one basic tendency to be noted here is when 'meanings based in the external or internal described situation become meanings based in the textual and metalinguistic situation' (Traugott 1989: 35). The direction of the semantic change can in most cases be conveniently described as 'bleaching'. Hopper—Traugott (1993: 87-93) warn, however, against too straightforward an association of grammaticalization with loss of meaning and point out that traces of the original lexical meanings of the grammaticalized forms often remain. They prefer the expressions 'pragmatic enrichment', 'strengthening', etc. to 'bleaching' or 'fading' and point out that grammaticalization results in the development of new uses as often as in the loss of old ones. In grammaticalization, the speaker's subjective attitude toward the proposition is of importance (see Traugott 1989: 35 and Traugott—König 1991). Subjectification as such, seems, however, too strong a factor in the analysis of grammaticalization; it is possible that the speaker's subjective attitude is, to some extent, present in most changes of meaning. For this reason, the more general starting-point of pragmatic inferencing, i.e. the role of the speaker and hearer negotiating meaning in communication

4

Matti

Rissanen

situations, and the dichotomy 'expressiveness/routinization' (Hopper— Traugott 1993: 63ff.) offers a less problematic starting-point. As mentioned above, all the articles in this volume are concerned with grammaticalization, although this process has been treated as an approach for the analysis and understanding of the changes rather than as a linguistic phenomenon to be studied in its own right. The types of grammaticalization can be divided into 'primary' and 'secondary', on the basis of whether or not they involve radical changes in the word class or structural properties of the items. The changes discussed in the present volume mainly belong to the category of primary grammaticalization: they deal with developments of pronouns and adverbs from other parts of speech. Secondary grammaticalization can be seen in subsequent developments of the pronominal or adverbial forms discussed.

3.

Study of change and variation in language

The last two decades have seen a rapid increase and methodological development in the studies of variation in language. The starting-point of these studies is "orderly heterogeneity", i.e. variability which is not random but affected by linguistic and extralinguistic factors (Weinreich— Labov—Herzog 1968; Samuels 1972; Labov 1994). Language can be seen as meaning potential which is realized in choices between forms and expressions "meaning the same thing" (Halliday 1973: 51; see also Halliday 1978). In diachronic studies, the variationist approach gives us a good opportunity to observe the various stages of change: embedding, transition and actuation (Weinreich et al. 1968: 102; cf. also Milroy—Milroy 1985: 341-343; Milroy 1992: 20-21). We can trace not only the birth and death of variant expressions, but perhaps more interestingly, their changing frequencies within a variant field at subsequent periods of time, and we can analyse the changes taking place in the intricate mesh of the factors cooccurring with the variants. The insights derived from recent trends in sociolinguistics have added a new angle to the discussion of these factors, both in view of the language of individual speakers and of larger speech communities (cf. e.g. Romaine 1982; Nevalainen—Raumolin-Brunberg 1996).

Introduction

5

The emphasis placed on variation has enhanced the role played by the text in the study of change. The distribution of variant forms in texts representing different genres or text types, as defined by extralinguistic factors (topic, the author's background, discourse situation and participant relationship, etc.) offers us a way to analyse register variation of the past. A comparison of the linguistic features in texts representing different distances from spoken language is also practically our only way to form hypotheses of the structures and vocabulary typical of the speech of past centuries. The same method of textual comparison must be used in attempts to describe the relationship of the standard(s) to regional or social dialects. The long time-span and the wealth of primary data set specific demands for the grammatical models used in all diachronic studies. The model should make it possible to compare changing grammatical phenomena across time, and, at the same time, be comprehensive enough to provide researchers with analytical tools for a very wide range of morphosyntactic issues. Furthermore, the main aim in diachronic studies is less often to develop new theories of grammar but, rather, to provide descriptions of interesting linguistic phenomena in the history of the language in a form that will offer a basis for further studies using different theoretical models. The model which has proved most useful in this type of research is a structurally oriented one, such as A comprehensive grammar of the English language by Quirk et al. (1985) for Present-day English. This starting-point has been employed by the authors of the present volume in different ways and to different degrees. We could characterize our grammatical orientation as eclectic: we have made use of various approaches, from traditional grammar to semantic, pragmatic and textual theories. This approach has provided a way of categorizing the variant expressions from the diachronic point of view in a way which maintains the main lines of development, while simultaneously making possible an analysis with a relatively high level of detail and even accounting for problematic borderline cases.

6

Matti

Rissanen

4.

Computerized corpora

In recent years, computerized corpora have played a decisive role in diachronic studies of variation. They have radically shortened the time needed for collecting evidence of the occurrence of variant expressions and made possible a detailed comparative analysis of the linguistic and textbased factors affecting the choice of the variants. The computer has enabled scholars to study even such topics as have earlier been avoided for the sheer amount of material collecting involved and for the uncertainty as to the relevance of the results. The Helsinki Corpus of English Texts, which forms the basis of the studies reported in this volume, was the first large computerized corpus to cover the time-span of several periods in the history of English. With its 1.5 million words — c. 400 samples of texts dating from the 8th to the 18th century — it gives reliable indicators concerning the structural and lexical developments of English for over a millennium. The results are, however, in many cases, only diagnostic; they must, and fortunately can, be supplemented from other corpora, concordances and primary texts. In the future, the usefulness of the Helsinki Corpus will be further increased by the addition of word-class tagging and syntactic bracketing to the text samples. 2 Each text sample of the Helsinki Corpus is equipped with a battery of parameter values containing information on the text and its author, if known. In Old and Early Middle English, this information concentrates mainly on the date and dialect of the text and on a fairly loose description of the genre. In Late Middle and Early Modern English, the genre selection is more extensive than in the earlier periods, with samples from drama texts, private letters, law court records, diaries, prose fiction, etc. In these periods, sociolinguistic information is given on the authors of the texts and, in the case of letters, on the relationship existing between the writer and the receiver. A large and structured corpus, equipped with textual parameter codings, has made it possible for the authors of the present volume to combine qualitative and quantitative analysis of changes through variation. We have been able to discuss the process of change with reference to different genres or text types within a long time-span, mostly from Early Middle to Modern English, i.e. from the time of the radical reorganization of the structure of the language to the period of its gradual establishment. The parameter coding and broad textual basis has also allowed us to ob-

Introduction

7

serve the role played by dialectal distribution and the differences between prose and verse texts in the survey of our Old and Middle English variant forms. In the discussion of late Middle and Early Modern evidence, special emphasis has been placed on the distributions shown by speech-based texts, or texts showing a relatively high degree of orality.

5.

Pronominalization

In pronominalization, the source concepts can either be lexical nouns as is the case with MAN, BODY and THING, or represent other, semantically less definite categories, such as the numeral ONE and the adjectival/ pronominal SELF. As Heine et al. (1991: 3 3 - 3 6 ) and Hopper—Traugott (1993: 97) point out, source concepts are often derived from the basic vocabulary, and even from words belonging to the level of superordinate categories in Roschian terms, as with the lexemes meaning 'person' and 'thing'. Hopper—Traugott's comment (1993: 97) on the possibility of grammaticalization of formerly fairly specific terms after their generalization shows its validity in the development of BODY, which first meant 'the material frame of man', then 'personal being, individual' {OED, sv. body, I, III), the latter meaning becoming the source of the present-day indefinite pronouns in -BODY. On the other hand, the numeral ONE is commonly used in the creation of indefinite pronouns in different languages, which stresses its importance as part of the basic vocabulary along with the plurality quantifier MANY (Lehmann 1982: 52 and Heine et al. 1991: 33). The same is true of the word indicating SELF, which easily becomes part of compound reflexives in various languages. The lexical and grammatical items which lend themselves to pronominalization typically function as noun phrases or noun phrase heads. There is a resemblance between regular noun phrases and pronouns, which makes it difficult to draw a borderline between them. Quirk et al. (1985: 335-336), who in terms of a prototype framework classify pronouns as central or peripheral, distinguish them from nouns according to different criteria, which as a whole apply to the central pronouns only. Semantically the meaning of pronouns is general and undetermined so that their interpretation depends to a large extent on what information is supplied by the context. Syntactically, most pronouns incorporate their own determiner. Morphologically, some pronouns have characteristics which noun

8

Matti

Rissanen

phrases do not have, such as a contrast between subjective and objective cases, a contrast between first, second and third persons, a distinction between personal and non-personal as well as between masculine and feminine gender, and morphologically unrelated number forms. The prototypical core formed by elements such as personal pronouns seems relatively stable, while grammaticalization, increasing the number of items or replacing old ones, can be expected in the periphery, for instance among the indefinite and reflexive pronouns (see RaumolinBrunberg 1994). On the other hand, pronominalization with nominal elements as sources can be characterized as weak grammaticalization or, as Heine et al. (1991: 44) put it, 'structure-preserving abstraction', a process which does not radically affect the categorical status of the linguistic structures concerned. The character of pronominalization as a weak type of grammaticalization becomes evident if this process is compared with the metaphorically and metonymically conditioned grammaticalization processes appearing for instance in the development of connectives and different verbal systems (cf. e.g. Traugott—Heine 1991; Hopper—Traugott 1993). Cognitive and pragmatic concerns are, nevertheless, important in pronominalization. Subjectification (see e.g. Traugott—König 1991) plays a central role in the development of ONE, and a search for new, more expressive emphatic forms for the universal, negative or reflexive pronouns corresponding to Present-day English EVERYONE/EVERYBODY, NO ONE/NOBODY and MYSELF, YOURSELF, etc., has led to repeated grammaticalizations. Despite the criticism directed at the idea of semantic bleaching (see e.g. Hopper—Traugott 1993: 87-93), it appears to be a conspicuous phenomenon in the process where noun phrases become pronouns, since semantic emptiness is one of the main characteristics of pronouns. Pronominalization often involves change at different linguistic levels. Among indefinites and reflexives, compounding easily leads to phonological changes, such as the stress placement in the development from ANY BODY to ANYBODY, or HIM SELF(UM) to HIMSELF. Phonological erosion can also take place, e.g. in Old English jEFRE ,/ELC, which has developed into EVERY. Structure-preserving grammaticalization does not involve major syntactic changes, but, as mentioned above, semantic fading is obvious, and in historical research it seems easier to tackle this topic than the elusive stress placement. Three articles in the present volume deal with the development of the forms of pronouns. Helena Raumolin-Brunberg and Leena Kahlas-Tarkka discuss the development of the indefinite pronouns with singular human

Introduction

9

reference from Old English to the end of the 17th century. Their study focuses on the gradual development and repeated grammaticalization of the complex forms of these pronouns during their long history. This development seems to have been connected with the tendency to mark the head of a noun phrase with a distinctive element, after the loss of the morphological gender and number markers of the adjective (Fischer 1992: 222). The same tendency probably influenced the development of the pronominal uses of ONE, as argued in Matti Rissanen's article. The development of the pronominal ONE also improved the cohesion of the sentence and helped in phrase boundary marking. Kirsti Peitsara's article on the expression of reflexivity describes another compound pronoun development, the reanalysis of the combination of the personal pronoun and the optional emphasizing element SELF as reflexive pronoun. The factors influencing the rise of the reflexive pronouns are, however, different from those of the compound indefinites: the rise of compound forms may have been supported by the heavy functional load of the oblique forms of the personal pronouns (in addition to 'ordinary' pro-forms, they were used as the so-called possessive or sympathetic dative, as dative of interest, etc.). The study of reflexive pronouns also links pronominalization with verb complementation and the variation of the pronominalized forms with zero forms or with middle voice.

6.

Adverbialization

The term adverbialization is used here collectively to refer to the various processes of adverb formation in the history of English. Insofar as adverbs can be considered a more grammaticalized category than the other open-class categories, adverbialization is also a process of grammaticalization. This view is morphologically justified by the derived status of adverbs, and the fact that few of them can be inflected (Traugott 1988: 132-134). Most adverbs in the world's languages are indeed derived from other word classes, notably from adjectives, nouns and verbs. Despite their status as a derived category, adverbs must nonetheless be considered open-class lexical items: their number can be freely augmented. As with other open-class categories, more grammaticalized subsystems may arise within the adverbial category as a result of secondary grammaticalization,

10 Matti Rissanen that is, further functional specialization and semantic change. This process produces, for instance, deictic compound adverbs (HEREAFTER, THEREUPON, WHEREIN), which themselves consist of closed-class pronominal adverbs (HERE, THERE, WHERE) combined with a preposition (or, from the Old English point of view, prepositional adverb). It also yields focusing adverbs (MERELY, JUST, EVEN), many of which arise from prior intensifier or focusing adverb homomorphs. Neither adverb class can be augmented at will nor can their members be inflected (see Nevalainen 1991: 10-18). Adverbs can be divided into complex, compound and simple on the basis of their morphological make-up. The vast majority of Present-day English adverbs are morphologically complex. Most of them are formed by adding the suffix -LY to an adjectival base. The process typically produces adverbs of manner ('in an X way', where X corresponds to the adjectival base: RAPID > RAPIDLY). Other adverbial suffixes include -WARD(S) and -WISE (see e.g. Quirk et al. 1985: 438-439, Huddleston 1984: 332-334). Suffixation is by far the most common means by which adverbs were also formed in the past. Swan (1988) suggests that -LY adverbs provide one of the most important sources for the formation of different kinds of sentence adverb in English. They include various evaluative and modal subclasses and arise from verb-modifiers and intensifiers; compare, for instance, they denied it frankly ('in a frank manner') and frankly, they denied it ('speaking with candour'). These sentence adverbs result, in our terms, from a process of secondary grammaticalization, which does not alter their category status as adverbs but brings about a change in their function. As they incorporate the speaker's comment on the proposition that they are associated with, sentence adverbs are more subjective and hence more grammatical ized than their verb-modifying homomorphs. Sentence adverbs may also arise through primary grammaticalization, category shift, as in the case of you know, I think and the by now archaic methinks. The process involves routinization and idiomatization of these expressions, which as a result come to be stored and used as simple units (see Hopper— Traugott 1993: 65, 201-203; Palander-Collin 1996). Unlike some other adverb classes, the English sentence adverbial category remains relatively open-ended. As a means of primary adverbialization, zero-derivation is a much more marginal process in Present-day standard English than suffixation. Most lexicalized zero-adverbs are derived from adjectives. Diachronic study of zero-derivation suggests that in the past functional links between ad-

Introduction 11 jectives and adverbs were perhaps closer than in today's standard language (Nevalainen 1994), although functional overlapping between the two classes is still common. This is typically the case of subject-related modifiers, where it is not always easy to determine whether the modifier is to be classified as an adjective or an adverb (the sun shines bright, he fell flat). Functionally, zero-derived adverbs range from these weakly codified subgroups to some more specialized and grammaticalized ones, such as process and focusing adverbs and intensifiers (CLOSE, FAST, DIRECT, RIGHT, SLOW, WIDE; Quirk et al. 1985: 405-407). Compared with the high productivity of the process of suffixation, compound adverbs are not very numerous in Present-day English. Some of them are morphologically opaque historical relics, such as the negative operator NOT, which goes back to the Old English noun phrase NAWIHT. It is questionable whether compounding with WHERE, THERE and HERE is still productive today. Many compound adverbs have become polyfunctional as a result of secondary grammaticalization. Many have also reduced their functional load in the course of time, including the conjuncts HOWEVER and THEREFORE (see Finell 1992 and Österman, below). In Present-day standard English few adverbs inflect for comparison. Those that do are often formally identical with their adjective homomorphs and share their comparative and superlative inflections (for instance, FAST/-ER/-EST, HARD/-ER/-EST). Some -LY adverbs also have inflectional comparatives and superlatives in standard English {it's easier said than done 'more easily'; speak clearer 'more clearly'; Quirk et al. 1985: 406). Most adverbs in -LY were compared periphrastically (MORE/ MOST X-LY) in earlier English, too. In terms of morphological fixedness compound adverbs constitute the more grammaticalized end of the adverb category. These adverbs typically do not allow any kind of comparison, inflectional or otherwise. The three studies by Terttu Nevalainen, Aune Österman and Minna Palander-Collin explore these various processes of adverbialization in the history of English. In terms of productivity and grammaticalization they cover the two extremes of the adverb category from the relatively closed class of compounding at one end to the rather open-ended class of -LY suffixation at the other, with zero-derivation retreating towards the less productive end with time. Nevalainen concentrates on the gradual increase in the derivations with the -LY ending at the cost of zero-derivation in Late Middle and Early Modern English. She also calls attention, among other things, to the questions of the continuum between 'adjectiveness'

12

Matti Rissanen

and 'adverbness' and the role played by the syntactic function of the adverb in the selection of the type of derivation. Ö s t e r m a n ' s article traces the increasing popularity of the adverbs formed with THERE, such as THEREFORE, THEREIN, THEREABOUTS, in Middle English and their rapid decline in the Early Modern English period. The study by Palander-Collin on METHINKS focuses on the primary process of adverbialization which grammaticalizes subject-verb collocations as sentence adverbs. Adverbialization of verb phrases such as PLEASE, PRITHEE and YOU KNOW has been attested from the Middle English period onwards and is not uncommon today (see, e.g., Erman 1987). In the course of time these phrases typically become reanalysed as sentence elements with various adverbial functions in speaker-hearer interaction denoting, for instance, politeness or the speaker's attitude to the proposition. In her article Palander-Collin studies the grammaticalization of METHINKS as a sentence adverb expressing the speaker's point of view.

Notes 1.

2.

This introduction is based on the ideas and suggestions of all the authors of the present volume. Section 5 (pronominalization) was compiled by Helena Raumolin-Brunberg and section 6 (adverbialization) by Terttu Nevalainen. The authors have done their best to make their articles readable without detailed knowledge of the structure and conventions of the Helsinki Corpus. It is obvious, however, that familiarity with the Corpus will make the appreciation and estimation of the results of the articles easier. For information on the Helsinki Corpus, see Rissanen et al. (1993) and Kytö (1996). When citing the examples from the Helsinki Corpus, we follow the typographical conventions explained in Kytö (1996). For convenience, the references to the Helsinki Corpus source texts are listed in the Bibliography at the end of the present volume. The Helsinki Corpus of Older Scots (Meurman-Solin 1995) has decisively improved our possibilities for studying regional variation in late Middle and Early Modern English. At the English Department of the University of Helsinki, there are three other projects in progress which will, among other results, produce new diachronic corpora: the Corpus of Early English Correspondence (see Nevalainen—Raumolin-Brunberg 1996), the Corpus of Early English Medical Writing (see Taavitsainen—Pahta 1997) and the Corpus of Early American English (see Kytö 1993). The Penn-Helsinki Parsed Corpus of Middle English contains a syntactically coded version of the Middle English prose texts included in the Helsinki Corpus. A new ver-

Introduction

13

sion of this corpus, with a more many-sided linguistic coding and additional prose texts, is in preparation. The Brooklyn-Geneva-Amsterdam-Helsinki Parsed Corpus of Old English will provide a linguistically coded and glossed version of the Old English section of the Helsinki Corpus. Kytö et al. (1994), Kytö—Rissanen (1996) and Hickey et al. (1997) give information on other historical corpora of English, either completed or under preparation.

References Erman, Britt 1987 Pragmatic expressions in English: A study of you know, you see and I mean in face-to-face conversation. Stockholm: Almqvist and Wiksell International. Finell, Anne 1992 "The repertoire of topic changers in personal, intimate letters: A diachronic study of Osborne and Woolf', in: Matti Rissanen—Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.), History of Englishes: New methods and interpretations in historical linguistics. (Topics in English Linguistics 10.) Berlin: Mouton de Gruyter, 720735. Fischer, Olga 1992 "Syntax", in: Norman Blake (ed.), The Cambridge history of the English language. Vol. 2. Cambridge: Cambridge University Press, 207408. Halliday, Μ. A. K. 1973 Explorations in the functions of language. London: Edward Arnold. [Reprinted 1974], 1978 Language as social semiotic. The social interpretation of language and meaning. London: Edward Arnold. Heine, Bernd—Ulrike Claudi—Friederike Hünnemeyer 1991 Grammaticalization: A conceptual framework. Chicago: University of Chicago Press. Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Hickey, Raymond—Merja Kytö—Ian Lancashire—Matti Rissanen (eds.) 1997 Tracing the trail of time. Proceedings from the Second Diachronic Corpora Workshop. Amsterdam and Atlanta, GA: Rodopi. Hopper, Paul J.—Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press.

14

Matti Rissanen

Huddleston, Rodney 1984 Introduction to the grammar of English. Cambridge: Cambridge University Press. Kytö, Merja 1993 "Early American English", in: Matti Rissanen—Merja Kytö—Minna Palander-Collin (eds.), 83-91. Kytö, Merja (comp.) 1996 Manual to the diachronic part of the Helsinki Corpus of English Texts: Coding conventions and lists of source texts. (Third edition.) Helsinki: Department of English, University of Helsinki. Kytö, Merja—Matti Rissanen 1996 "English historical corpora: Report on developments in 1995", ICAME Journal 20: 117-132. Labov, William 1994 Principles of linguistic change. Vol. 1. (Language in Society 20.) Oxford: Blackwell. Lehmann, Christian 1982 Thoughts on grammaticalization: A programmatic sketch. (Arbeiten der Kölner Universalien-Projekts 1.) Cologne: Universität zu Köln, Institut für Sprachwissenschaft. Meurman-Solin, Anneli 1995 "A new tool: The Helsinki Corpus of Older Scots (1450-1700)", ICAME Journal 19: 49-62. Milroy, James 1992 Linguistic variation and change: On the historical sociolinguistics of English. (Language in Society 19.) Oxford: Blackwell. Milroy, James—Lesley Milroy 1985 "Linguistic change, social network and speaker innovation", Journal of Linguistics 21: 339-384. Nevalainen, Terttu 1991 BUT, ONLY, JUST: Focusing adverbial change in Modern English 1500-1900. (Mömoires de la Soci6t6 N6ophilologique de Helsinki 51.) Helsinki: Societe Nöophilologique. 1994 "Diachronic issues in English adverb derivation", in: Udo Fries— Gunnel Tottie—Peter Schneider (eds.), Creating and using English language corpora. Amsterdam: Rodopi, 139-147. Nevalainen, Terttu—Helena Raumolin-Brunberg (eds.) 1996 Sociolinguistics and language history: Studies based on the Corpus of Early English Correspondence. Amsterdam: Rodopi. Palander-Collin, Minna 1996 "The rise and fall of METHINKS", in: Terttu Nevalainen—Helena Raumolin-Brunberg (eds.), 131-149.

Introduction

15

Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman. Raumolin-Brunberg, Helena 1994 "Prototype categories and variation studies", in: Francisco Fernändez —Miguel Fuster—Juan Jos0 Calvo (eds.), English historical linguistics 1992. (Current Issues in Linguistic Theory 113.) Amsterdam: John Benjamins, 287-303. Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.) 1993 Early English in the computer age: Explorations through the Helsinki Corpus. (Topics in English Linguistics 11.) Berlin: Mouton de Gruyter. Romaine, Suzanne 1982 Socio-historical linguistics, its status and methodology. (Cambridge Studies in Linguistics 34.) Cambridge: Cambridge University Press. Samuels, M. L. 1972 Linguistic evolution, with special reference to English. Cambridge: Cambridge University Press. Swan, Toril 1988 "The development of sentence adverbs in English", Studia Linguistica 42/1: 1-17. Taavitsainen, Irma—Päivi Pahta 1997 "Corpus of Early English Medical Writing 1375-1750", ICAME Journal 2\: 71-78. Traugott, Elizabeth Closs 1988 "Is internal semantic-pragmatic reconstruction possible?", in: Caroline Duncan-Rose—Theo Vennemann (eds.), On language, rhetorica, phonologica, syntactical A festschrift for Robert P. Stockwell. London: Routledge, 128—144. 1989 "On the rise of epistemic meaning in English: An example of subjectification in semantic change", Language 65: 31-55. Traugott, Elizabeth Closs—Bernd Heine (eds.) 1991 Approaches to grammaticalization. 2 vols. Amsterdam: Benjamins. Traugott, Elizabeth Closs—Ekkehard König 1991 "The semantics-pragmatics of grammaticalization revisited", in: Elizabeth Closs Traugott—Bernd Heine (eds.), 1: 189-218. Weinreich, Uriel—William Labov—Marvin I. Herzog 1968 "Empirical foundations for a theory of language change", in: W. P. Lehmann—Yakov Malkiel (eds.), Directions for historical linguistics: A symposium. Austin: University of Texas Press, 95-195.

Indefinite pronouns with singular human reference Helena

Raumolin-Brunberg

and Leena

Kahlas-Tarkka

1. Introduction 1 The object of this investigation is a subgroup of pronouns, i.e. items that in traditional terms are called indefinite pronouns with singular human reference, e.g. SOMEONE, ANYBODY, EVERYONE and NOBODY. The traditional characterization is not necessarily accurate, since there are serious doubts about the indefiniteness of some of the items (for further discussion, see 2.2 below). Nevertheless, since it is difficult to invent a more pertinent name for this subcategory of pronouns, we will use the wellestablished term 'indefinite pronouns'. The basic data for this study were retrieved from the Helsinki Corpus of English Texts (see Kytö 1996; Rissanen et al. 1993), supplemented by other texts, e.g. the Toronto (Healey—Venezky 1980), Shakespeare (Wells —Taylor [eds.] 1989), and Century corpora (Milic 1990) and a Presentday English Bible translation {Revised English Bible, REB, 1989). The full inventory of the pronouns discussed is presented in Table 1. The repertoire and its subdivision follow Quirk et al.'s discussion of indefinite pronouns (1985: 376-392). As time wears on, the four paradigms, assertive 'someone', nonassertive 'anyone', universal 'everyone' and negative 'no one' become more or less symmetrical, all of them having compound variants (ending in MAN, ONE or BODY) and simple items, which in later times became similar to the corresponding determiners, except for NONE (determiner NO). The simple pronouns are called OF-pronouns by Quirk et al. (1985: 379), since in Present-day English they occur in partitive phrases. Semantically they are quantifiers, although this study does not adhere to the logic-based formal tradition from which this term has been derived.

18

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Table 1. Indefinite pronouns with singular human reference. Assertive 'someone'

Non-assertive 'anyone'

Negative 'no one'

Universal 'everyone'

ANY ANYONE ANYBODY

NONE NO ONE NOBODY

EACH EVERYONE EVERYBODY

ANY ANY MAN ANY ONE ANY BODY

NONE NO MAN

EACH/EVERY EACH/EVERY MAN EACH/EVERY ONE EVERY BODY

ANY ANY MAN

NONE NO MAN

Present-day English Simple -ONE -BODY

SOMEONE SOMEBODY

Early Modern English (1500-1700) Simple -MAN -ONE -BODY

SOME SOME MAN SOME ONE SOME BODY

NO BODY

Middle English (1150-1500) Simple -MAN -ONE

SOME SOME MAN

-BODY

SOME BODY

EACH/EVERY EACH/EVERY MAN EACH/EVERY ONE

NO BODY

Old English (-1150) Simple

SUM

-MAN

SUM MAN

JENIG

NyENIG NAN

/ENIG MAN

NAN MAN

GEHWILC/GEHWA yEGHWILCAEGHWA /ELC y£GPER//EGHW/EI>ER -MAN

T h e time span of this study is very long, from Old English to the beginning of the eighteenth century. The differences between the data available from different subperiods are considerable, affecting many aspects of the study, from the n u m b e r of occurrences to the possibilities for textual comparisons. Although our main purpose has been to offer a picture of the longitudinal development of the indefinite pronouns, w e have had to be relatively cautious in our analyses of the earliest periods in the history of the English language.

Indefinite pronouns with singular human reference

2.

19

The background and theoretical framework

2.1. Method and borderlines The research was carried out within the variationist framework, the items being chosen to form paradigms in which the members were interchangeable in several, if not all, environments. This procedure made it possible to perform quantitative comparisons along with qualitative ones. Both linguistic and extralinguistic factors constraining the choice of variants could be tackled by this method. Comparisons were carried out not only within the paradigms but between them as well. As Raumolin-Brunberg (1994a) points out, delimitation of paradigms based on syntactic categories is often problematic because of their prototypical nature. Protototypical and peripheral items may be semantically so different from each other that it is difficult to find underlying constants which would allow items to be regarded as variants of the same set. There were particular problems in this study both at type and token level. At the type level it was difficult to decide which items to include in the paradigms. The study includes only pronouns containing the indefinite element SOME, ANY, EVERY, EACH and NO. Types like WHOEVER (Old English SWA HWA SWA) and corresponding relative constructions such as SE DE in Old English and later HE THAT/WHO, as well as reciprocal pronouns have been excluded or touched upon only superficially. The Old English interrogatives HWA or HWILC, also used as indefinites, have been excluded, as they became extinct at quite an early stage, and only those pronouns showing continuity from Old English to Early Modem English have been included. Apart from traditional pronouns, we included such lexical items as later developed into pronouns, e.g. BODY when it was preceded by SOME, ANY, EVERY and NO. These forms may be called prepronouns at the stages preceding prominalization. The phrases with MAN were particularly problematic. We decided to select the forms with MAN which correspond to Present-day English compound pronouns, e.g. ANY MAN, NO MAN, etc. but not the item A MAN, which Poutsma ([1966]: 1205) includes in the group of quasi-indefinite pronouns. In the discussion of the Bible translations it became evident that A MAN was an alternative to the pronouns in the assertive 'someone' series. We felt, however, that A MAN had too much of its semantic content left to be included among pronouns. It seems to parallel the presentday use of A PERSON in reference to indefinite human beings.

20

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

As far as tokens are concerned, analysing which occurrences of SOME, ANY and NONE were singular and which plural was problematic, since only the former were to be included. Also, as regards the tokens of MAN, instances of full semantic content, such as 'male human being', 'human being as opposed to God', or 'someone's servant' had to be excluded (see also Raumolin-Brunberg 1994b). There is no way of avoiding some degree of subjectivity in deciding where the borderlines lie between full nouns and pronouns. Therefore we would like to point out that the quantification in this study is not to be taken to represent any 'absolute truth'. Other scholars might make different decisions in the interpretation of the less clear cases. We believe that the figures nevertheless give a good picture of the general developments in the long history of the pronouns in question. Examples (1) and (2) illustrate the problem of number. The occurrences of SOME in example (1) represent the singular on the basis of the verb form,2 while example (2), not included in the corpus, is indeterminate, since neither the verb form nor pronouns or other contextual features reveal the number. Example (3) is a passage of Wyclifs New Testament. While the form with MAN is included in the corpus as it was later translated by the pronoun NO ONE, it is difficult to say whether any man in example (4) or sumum menn, celces monnes and sumes monnes in (5) refer to a person in general or a male human being only. These examples are excluded from the present corpus. (1)

What say you to this? that the witches haue their spirits, some hath one, some hath more, as two, three, foure, or fiue, some in one likenesse, and some in another, as like cattes, weasils, toads, or mise,... (EModE2 Gifford, A Dialogue Concerning Witches and Witchcraftes B4V)3

(2)

And on the Thursday the said Lordes fearing the people, imagined howe to escape out of theyr handes, and caused some of their seruauntes to sette fire on certaine houses in the Towne, thinking that... (EModE2 Stow, The Chronicles of England 546)

(3)

But we knowun this man, of whennus he is; but whanne Crist schal come, no man woot of whennus he is. (ME3 Wycliffe and Purvey, The New Testament John 7 27)

Indefinite pronouns with singular human reference

21

(4)

... wher it is wel knowen by alle manere of euydences that they könne shewe for hem self or any man for hem that they ben ire tenent3 and fre holderes to the kyng in chief... (ME3, Petitions in A Book of London English 1384-1425 201)

(5)

Bisna f>e be sumum menn, for {Dam J)e celces monnes lif biö sumes monnes lar. (OE3/4 Diets of Cato 10) 'set someone as an example for yourself, because everyone's life is someone else's source of learning'

2.2. Indefiniteness The discussion of definiteness has a long tradition, both in philosophy and linguistics (see e.g. Donnellan [1971]; Searle 1969: 26-27; Lyons 1977: 174-192; Hawkins 1978; Givon 1984: 398-412; Reuland—ter Meulen [eds.] 1987; Chesterman 1991). Indefiniteness has raised far less interest; mostly it has represented the binary opposition of definiteness. In later studies definiteness has been considered to be of a composite nature and a continuum, and there are scholars who have claimed that indefiniteness in fact is a more complex phenomenon than its apparent opposite (Givon 1984: 431; de Jong 1987: 271; Chesterman 1991: 40). The traditional characterization of definiteness and indefiniteness could read as follows: a definite NP has a referent which is assumed by the speaker to be unambiguously identifiable by the hearer; an indefinite NP has a referent which is assumed by the speaker not to be unambiguously identifiable by the hearer (Chesterman 1991: 10). In other words, definite NPs represent familiarity, indefinite ones new information. Although these principles work in a large number of cases, they have their problems, especially in the treatment of nonreferential and uncountable nominals. In many analyses, the concepts of definiteness, specificity and referentiality have not been kept apart. Definiteness has been defined as "reference to a specific individual or class" (Lyons 1977: 185) or in terms of having "a particular individual in mind" (Ihalainen 1974: 29). The problem here is that both referentiality and specificity form continuums which intersect that of definiteness; in other words, both definite and indefinite NPs can be specific and nonspecific, referential and nonreferential (see e.g. Ihalainen 1974: 81; Huddleston 1984: 254).

22

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

(6)

The murderer of Smith must be insane.

(7)

John is the acme of courtesy.

(8)

John is a good man.

(9)

In came a young man.

Examples (6)-(9) illustrate the phenomenon. Example (6) from Donnellan ([1971]: 103-104) and Chesterman (1991: 11) has two readings since either the speaker has a specific individual in mind or not. In the former case the NP is both referential and specific, in the latter nonreferential and nonspecific. Examples (7)-(8) from Chesterman (1991: 12) show that subject complements which are nonreferential, can be either definite or indefinite. Example (9), in turn has a referential and specific indefinite noun phrase as subject. Hawkins (1978) introduces a pragmatically-oriented theory of definiteness, which he calls location theory. Its main components are shared sets and inclusion versus exclusion. In the case of definiteness the speaker instructs the hearer to locate the referent(s) in some shared set of objects (which are broadly defined) and refers to the totality of the objects or mass within the set (=inclusion). As regards indefiniteness, the speaker is uncommitted about the shared set, but refers to a proper subset, i.e. notall, of the potential referents (=exclusion; Hawkins 1978: 167-187). In other words, in the case of exclusion there should be at least one other referent that could potentially be referred to by the same expression. This approach bears a resemblance to Givon's (1984: 434—435) claim that new arguments are introduced into the universe of discourse as one member of many within the type. Hawkins (1991) develops these ideas further in relation to recent approaches to pragmatics. The article revises concepts and terminology, but the basic arguments of location theory appear to hold. Without going too deep into the discussion of the theory of sets, Figure 1 presents a simple model to illustrate the approach used in this study (see Suojanen 1977: 27 and Kahlas-Tarkka 1987: 89). It goes without saying that the size of the set can vary a great deal from two to all people in the world. In text the set may be expressed explicitly, either grammatically, e.g. by an OF-phrase (example (2), some of their seruauntes), or lexically (example (19), The boys). It can also be inferred from the context, so that in example (28) the set consists of those people who were around, among whom some body should call the speaker's wife.

Indefinite pronouns with singular human reference

'someone'

ο

ο

ο

ο

ο

ο

'anyone'

λ

//»\ / if η . Ο,Φ ·\0 11' κ ι /I' II» 11' ιι ι ' ιιΟ ι' ιι ι' ιι ',0 ι

ι

ο

ο

'everyone'

ό

ι ι

b

'no one'

Ο

Ο

Ο

ο

ο

ο

ο

ο

Figure 1. Indefinite pronouns in a framework of sets.

24

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

Although its approach to the nature of sets is different, 4 it is interesting to look at Figure 1 in terms of the location theory. While the assertive and nonassertive pronouns clearly fulfil the condition [+indefinite], because there are other referents that could be referred to by the use of the same pronoun, the case is different with the universal paradigm. The universal pronouns behave exactly the way definite expressions do in location theory, referring to all referents within the set.5 Suojanen's (1977: 32) apt name for this class of pronouns is inclusive pronouns, or pronouns of totality. Jespersen (1933: 184-187) distinguishes a nonindefinite category of pronouns, those of totality, which includes two subgroups, positive and negative pronouns (ALL, BOTH, EVERY, EACH, and NO, NONE, NEITHER). It is also interesting to observe that Safir (1987: 71) points out how EVERYONE behaves like definite NPs in existential there-clauses, while SOMEONE is grouped with indefinites. Pesetsky (1987: 103) also shows some differences between EVERY and the indefinites. Figure 2 shows that restrictive modification creates subsets. The set in example (10) consists of all the people who overheard certain people talk, and one of them is picked out. The set in (11) comprises those people who were in the church. (10)

The silly old man willingly paid his penny before hand, and was going ouer; but some that ouer-heard their talk, hindered his journey and laughed at the jest, yet pitied his simplicitie, and sets him in the right way. (EModE2 Armin, A Nest of Ninnies 42)

(11)

Dr. Bolton preached and that you may know that Mr. Hannington is of no ordinary esteeme, I will assure thee hee had such a presence with him as though a stranger to every one in the Church, hee was ushered in from his standing amongst severall gentlemen and seated next to the Bishop of London himselfe ... (EModE3 Henry Oxinden, The Oxinden and Peyton Letters 277)

Indefinite pronouns with singular human reference

'some that ouer-heard their talk'

the people that ouerheard their talk

'every one in the church'

the people in the church

Figure 2. Indefinite pronouns in a framework of subsets.

26

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

2.3. Specificity It seems that the common denominator for these pronouns is the possibility of nonspecific reference. In fact it appears to be specificity, not definiteness that is tested by analysing whether the speaker has a particular or specific individual in mind. Quirk et al. (1985: 391) argue that the basic difference between the assertive 'someone' and nonassertive 'anyone' series is that the former is specific, though unspecified, and the latter nonspecific. It is, however, not difficult to find examples of the nonspecific use of SOMEONE (examples (12)—(13) to be compared with (10); see also Lyons 1977: 189). (12)

I found a wallet in the bus. Someone must have left it behind, [-specific]

(13)

And there should be two or three roomes made a little remote from the dwelling house, to which Scholars may be removed and kept apart, in case they be sick, and have some body there to look to them. (EModE3 Hoole, A New Discovery of the Old Art of Teaching Schoole 226) [-specific]

There are languages where this distinction has been grammaticalized, e.g. Russian has two different pronouns for SOMEONE: kto-nibud [-specific] and kto-to [+specific]. In Finnish the pronouns eras (standard) and yksi (colloquial) tend to be used with specific reference, while joku is nonspecific (see also Rissanen 1987). The distinction between specific and nonspecific reference may, in broad terms, be expressed as follows: if the speaker has a specific individual in mind but the hearer does not know the identity, the reference is specific; if neither the speaker nor the hearer knows the identity of the individual in question, the reference is nonspecific. 6

2.4. Genericness Genericness is usually attributed to NPs referring to whole species or subspecies, although their referentiality has been questioned by some scholars who argue that only specific nouns refer (cf. section 2.5, below). At a more general level, genericness concerns propositions which are not bound to specific spatio-temporal situations (Davison 1981: 361).

Indefinite pronouns with singular human reference

27

For many scholars genericness is a property of noun phrases, for others (e.g. Lyons 1977: 194) of whole propositions. In this study our focus is the pronominal noun phrase, although the context inevitably plays an important part in the analysis and semantic interpretation. According to Suojanen (1977: 42), the fact that someone does something often, usually, or always gives rise to different degrees of genericness. What is important is that genericness does not necessarily mean that whole species or subspecies are always involved; the class or type can be relatively limited instead. Restrictive modification creates subsets here too (Chesterman 1991: 80). As Givon (1984: 265) points out, specific individuals may indulge in generic activities in generic times and places (example (14)), and generic groups may together indulge in the same activity at the same specific time and place (example (15)). (14)

John always smokes wherever he goes.

(15)

Everybody held their breath as the news came in.

Generic nominals can be either definite singular, indefinite plural or indefinite singular (examples (16)—(18) from Ihalainen 1974: 51). The choice between them is prompted by several syntactic and semantic factors, but there are also a large number of contexts where they are interchangeable (Ihalainen 1974: 51-52; Chesterman 1991: 78). (16)

The beaver builds dams.

(17)

Beavers build dams.

(18)

A beaver builds dams.

As regards indefinite pronouns, it is the nonassertive, universal and negative pronouns that tend to occur with generic reference, while the assertive paradigm is usually nongeneric. The pronouns referring to all people in the world may be given a generic interpretation (example (3)). The genericness of these pronouns, of course, resembles that of indefinite singular NPs (example (18)), and they cannot be used with verbs denoting the existence of species, etc.

28

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

2.5. Reference Referentiality, like so many other linguistic phenomena, can be seen as a cline (Givon 1984: 390, 430; Chesterman 1991: 188). The following discussion takes up only some aspects of this extensive topic. While many of the central pronouns are mostly used with textual reference, it is seldom the case with the indefinite pronouns, in particular the compound pronouns (Suojanen 1977: 19-20). If anaphoric relations appear at all, the question is usually not one of coreference, but whether the pronoun has the set as antecedent (example (19)). The use of the pro-forms is an exception, however (example (20)), for a similar use of ONE, see Rissanen's type 'substitute', this volume). (19)

The boys rushed in. Everyone was dirty.

(20)

So that he bore ill will to no Person, nor hated any upon personal accounts. (EModE3 Burnet, The Life and Death of John Earl of Rochester 144)

According to Givon (1984: 390) and Ihalainen (1974: 67-71), among others, referentiality is a property of specific NPs only. If this analysis were to be applied here, very few pronouns could be considered referential and Figure 1 would be impossible. In an earlier study, Givon (1978: 293) claims that a referential NP "involves, roughly, the speaker's intent to 'refer to' or 'mean' a nominal expression to have non-empty references, i.e. to 'exist' within a particular universe of discourse." If we look at the matter from the angle we did in Figure 1, it becomes obvious that the existence of the sets and members can be presupposed. The smaller the set is, the more obvious is its existence. The result is that we regard as referential even pronouns which do not select a specific individual (see Suojanen 1977: 27). However, as for negative forms the analysis is problematic. It is of course difficult to claim that the subjects in (21) and (22) are referential, but we can still claim that the sets exist. The existence of a small definite set as in (22) may make it easier to accept this view of referentiality. (21)

No one saw him.

(22)

None of them saw him.

Indefinite pronouns with singular human reference

29

There is a clear connection between referentiality and syntactic function. According to Givon (1990: 900), the only clausal participant types that are very likely to be nonreferential are subject complements and expressions of manner or instrument. It is important to notice that the pronouns under examination rarely occur in these inherently nonreferential functions. Givon (1984: 391) also points out that propositional modality has a connection with referentiality. Under the scope of FACT modality (pre-supposition and realis assertion), nominal arguments can only be referential. Under the scope of NON-FACT modality (irrealis assertion, negative assertion), they can also be nonreferential.

2.6. Conclusion To conclude we present the continuum for indefinite pronouns proposed by Davison (1981). The criterion used has been a scale of indefiniteness, based on the degree to which the word picks out referents from the universe of discourse. "Negative indefinites, which pick out no referent, would be at one extreme, while generics, which pick out indiscriminate referents of a given subclass, would be at the other. In between would be non-specific indefinites which pick out a referent without further identification, and specific indefinites, for which further identification could be supplied" (Davison 1981: 359). Given the similarities between generic and universal pronouns, it may not be inappropriate to classify the universal pronouns with the generics at the same end of the scale (for the similarities and differences, see e.g. Vendler 1967 and Jackendoff 1972).

3.

The four paradigms of indefinite pronouns

3.1. Overall characteristics Two types of the variants under scrutiny, the simple pronouns, including words like EVERY, etymologically a compound, and the compounds with MAN persist throughout the Helsinki Corpus data. The indefinite pronouns in -BODY and -ONE are newcomers, introduced in Middle English.

30

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

Both establish themselves in Early Modern English, although the pronoun with ONE finds its way into the negative paradigm relatively late. Table 2 gives an overall picture of the trends in the usage throughout the Helsinki Corpus material. Some figures in the table call for special attention. It is striking, but it is difficult to explain why there seems to be such a great difference between Old and Middle English. For SOME and ANY in particular, there is a considerable decrease from Old to Middle English, whereas NONE represents a contrary development. No obvious explanations are offered by the material, apart from the fact that it is not always comparable, especially in the earlier subperiods. Relatively high frequencies may be found in some individual texts; out of the 46 instances of SOME in OE3 as many as 23 are found in two poems, Christ and The Fortunes of Men, and a third of the 120 examples of ANY in OE3 appear in Wulfstan's Homilies. The fact that some texts cause peaks in the frequencies may be somewhat misleading for general conclusions. The same genres are not represented throughout, and the dialectal background is somewhat different. No obvious linear continuum can be pointed out in the available material. On the other hand, there is an obvious rise in the popularity of the -BODY compounds in later subperiods and as obvious a decrease in the -MAN compounds. As regards the syntax of the indefinite pronouns, it is obvious that they are freely modified by postmodifiers such as relative clauses, prepositional phrases other than OF-phrases, adverbs, BUT-phrases, adjectives and participles. The use of the partitive OF-phrase or the genitive plural form in Old English as a postmodifier (henceforth both usually termed OF-phrases) forms a dividing line between the two. Only the simple pronouns and compounds with ONE (available from ME1 onwards) can appear in the partitive OF/genitive structure. The other two of the pronoun types, compounds in -MAN and -BODY, often occur as unmodified subjects and objects, with reference to human beings in general.

Indefinite pronouns with singular human reference

oo Ν oo (S — co — co

f « Ν m co ττ

r^ in » CS —' W-l

so © oo os d d « cs

ο os ο os Κ ri t^ ri

oo os — »ri 00 CS '—' COSO (N

Tt V) 't so CO CO

CO Tj- OS >n cs

CO CO CS CS CO so

Tf OS SO cs •n

oo

ο CO CO SO e d8lytable noys of mynstralsy, And of alkyn swet tones of musyke, Pat til any mans hert mught like; (ME3 The Pricke of Conscience 249) 'of the delightful sound of musical entertainment, and of all kinds of swe8t ίοηεβ of music that must please anyone's heart;'

(40)

It is agreed by the whole Bench, that the Procurer and the Adherent be deemed alwayes Traytors, when as a trayterous Acte was committed by anye one of the same Conspiracies.. (EModEl State Trials, The Trial of Sir Nicholas Throckmorton I 75.C2)

Indefinite pronouns with singular human reference (41)

39

But you Blockhead, I ask you whether you did see any body else? (EModE3 State Trials, The Trial of Lady Alice Lisle IV 113.C2)

The frequency of the 'anyone' series is many times the assertive paradigm. Apart from OE1 with only one instance, the lowest frequencies in the Helsinki Corpus are found in ME1 and ME2, 11 and 10 instances per 100,000 running words respectively. During the later Middle English and Early Modern period the numbers grow, reaching 68 occurrences in subperiod EModE3. Tesch's present-day data (1990: 61-69, 80, 85-86) contain 33 occurrences per 100,000 running words (see 3.2.1 above). It is interesting that, unlike the corresponding subperiod of the Helsinki Corpus, the Shakespeare Corpus includes more instances of ANYONE than ANYBODY (c. 10% versus 5%). It is also noteworthy that although Shakespeare in general makes more use of the compounds in -ONE and -BODY than the EModE2 subsection of the Helsinki Corpus, he prefers ANY and ANY MAN (c. 40% each). In the Century Corpus ANYONE dominates (51 instances against 13 of ANYBODY), and ANY and ANY MAN still appear a few times.

3.3.2. Syntactic and semantic properties The variants may be divided into two groups: ANY MAN and ANY BODY on the one hand and ANY and ANY ONE on the other. The former group shares the following characteristics: preference for unmodified structure, use in the genitive case (example (39)), and non-appearance with partitive OF/genitive structures. ANY ONE and ANY are in turn found modified by partitive OF-phrases or corresponding plural nouns in the genitive case (examples (34), (35), (36) and (40)). More than half of the occurrences of ANY occur with a partitive OF-phrase in all periods after OE4, except for the EModE3 subperiod, although it should be kept in mind that the frequencies for the early subperiods are very low. All items except ANY ONE are found to be modified by relative clauses (example (42)). The variants can also be followed by ELSE, although this postmodifier seems to become frequent only in Early Modern English (example (41)), gradually replacing OTHER (not included in the corpus). Other prepositional phrases and adverbs also occur as post-modifiers (example (43)).

40

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

(42)

... I shewed vnto them, that my purpose was not to put any faulte eyther in the acte or any man that made it, or in the othe or any man that sware it... (EModEl The Correspondence of Sir Thomas More 502)

(43)

Therefore if any heere haue fauoured this occupation before, let him now submit his thoughts vnto Gods thoughts... (EModE2 Smith, Sermons B1V)

As in the assertive paradigm, most instances represent the functions of subject and object, here however in clauses that are not affirmative declarative main clauses. It is worth mentioning though that the first occurrences of ANY ONE (EModEl and EModE2) are syntactic agents in passive sentences (example (40)). Although all the variants make use of the different ways of expressing the set, some preferences can be discerned. The simple form favours syntactic means, in particular the OF-phrase; contextual expression also appears, but from ME3 to EModEl ANY never refers to people in general. ANY MAN, in turn, seldom uses syntactic means (in OE1-ME4 no instances), but is mostly chosen for reference to human beings in general. The newcomers ANY BODY and ANY ONE also favour general reference. The factor giving rise to the name of this paradigm is present in the syntactic context of all its four members, all occurring in nonassertive contexts (cf. Quirk et al. 1985: 784-785). These contexts in the corpus, which to some extent resemble Givon's (1984: 391-397) modal category NON-FACT, include the following constructions: questions (examples (41) and (45)), conditional clauses (examples (38), (43) and (44)), negative contexts (examples (33), (42) and (52)), comparative and superlative contexts (examples (46)-(48)), and a variety of half-negative words and phrases (examples (49)—(51)). (44)

ah 3 if eni mon hit muste isean. he mahte iseon ane berninde glede J?et hine al for-bernaö J^urut to cole. (MEX/1 Lambeth Homilies 3 27) 'and if anyone could see it, he could see a burning coal that burns all of it to charcoal.'

Indefinite pronouns with singular human reference

41

(45)

He saede fiaet he aet sumum cirre wolde fandian hu longe fjast land nor|)ryhte laege, hwseöer cenig mon be noröan {jasm westenne bude. (OE2 Ohthere and Wulfstan MS L 17) 'He said that he once wanted to find out how far north the land extends, or whether anyone lives north of the wilderness.'

(46)

... for in Jjat is more vertue and good and norture Jjan any man may J)enke or seye. (ME4 The Book of Vices and Virtues 111)

(47)

Per η is non betere anonder sunne, Pat eni man of telle cunne. (MEX/2 King Horn 26)

(48)

Ungrateful, ungrateful adds to the Load which is between Man and Man, and is the basest Crime that any one can be guilty of. (EModE3 State Trials, The Trial of Lady Alice Lisle IV 122.C2)

(49)

God forbid any Man should accuse himself upon his Oath. (EModE2 State Trials, The Trial of Sir Walter Raleigh I 213.CI)

(50)

For it would almost seem incredible to any one, and... (EModE3 Preston, Boethius 177)

(51)

... seile hire forseid fissh and vitailles to the comunes of the same Citee, to lordes, and to alle other, thedir repeiryng freliche withoute destourbaunce or lettyng of eny man in priue or apiert... (ME3 Proclamations 32) '...making his way freely without anyone disturbing or hindering him in private or open places'

(52)

ForÖam ne maeg cenig man Godes mihta ne his msröa geminsian ne aenig öing awyrdan. (OE3 Wulfstan's Homilies 8c 178) 'Therefore no one can diminish God's power or destroy anything.'

The nonassertive contexts do not form the only environments for the 'anyone' series. A distinction has been made between two kinds of ANY, one of which is usually employed in nonassertive contexts, while the other is

42

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

not (e.g. Huddieston 1984: 423-328). We will modify Sahlin's (1979) Convention of calling them Any II and Any III by identifying them as Anyl and Any2 and assume that the principles which are valid for the determiner are also valid for independent pronominal use.10 According to Sahlin (1979: 20-22), our Anyl is the indefinite non-assertive unlimited quantifier, whose sense can characteristically be brought out by the addition of phrases such as 'at all' or 'worth mentioning' (example (41)). Any2, on the other hand, may be paraphrased by 'no matter who', 'whoever you like'. It mostly occurs in assertive clauses but can also be found in nonassertive contexts. Any2 with its generic meaning is really the sense presented in Figure 1, section 2.2, above. We would, however, wish to point out that in many cases it is difficult to make a distinction between the two senses in written texts. Apparently we should rather posit a cline where Anyl occupies one end and Any2 the other. The fuzziness of the context often prevents us from making unambiguous interpretations. Nevertheless, it seems to us that both senses have been available from Old English onwards (examples (53)—(56) of Any2). (53)

Ga to f)ine feder burinesse oöer J)er eni of J)ine cunne liö in. and esca hine hwet he habbe bijeten mid his wohe domas. (MEX/1 Lambeth Homilies 3 35) 'Go to your father's grave or where any of your family lies and ask him what he has gained from his perverse judgements.'

(54)

But from all put together that I hear from other people, I am likely to pass as well as anybody. (EModE3 Pepys, The Diary of Samuel Pepys 315)

(55)

If any body should ask me, if your were the same Mr. Oates that was at St. Omers; I should say, I heard so; and it would be very good Evidence, unless some one else were produc'd. (EModE3 State Trials, The Trial of Titus Oates IV 75.C2)

(56)

If ony man schal entre bi me, he schal be sauyd... (ME3 Wycliffe and Purvey, The New Testament John 10 9)

That intonation plays an important part in the distinction between Anyl and Any2 is exemplified in (55), where the pronoun can have two different readings, indicated in spoken language by different intonations. In this

Indefinite pronouns with singular human reference

43

case we might paraphrase them either "if anybody at all should ask me" or "if anybody, no matter who, should ask me". Example (56) raises an interesting point, when we compare the corpus translations with that of the Revised English Bible, where John 10:9 reads "I am the door; anyone who comes into the fold through me, will be safe". The corpus includes several instances where the four early translations, West Saxon, Wyclif, Tyndale and the Authorized Version, have a conditional clause with ANY MAN as subject, while the Present-day English translations begin with a main clause with a generic ANYONE modified by a relative clause as subject (see further Raumolin-Brunberg— Kahlas-Tarkka 1995). A closer look at the data indicates that any member of the nonassertive paradigm as subject of an affirmative declarative main clause (example (57)) is in fact found only once, and even here the subject is not clauseinitial. The Shakespeare and Century corpora both have one instance of clause-initial subject (examples (58)-(59)). (57)

And so much the freer is any man from the same, as neerest... (EModE2 Elizabeth I, Queen Elizabeth's Englishings ofBoethius, Plutarch, &c 93)

(58)

Mercutio: Any man that can write may answer a letter. (Romeo and Juliet 2 4 10)

(59)

Any man may know that I mean tragedy. (The Gentleman's Journal, 1691)

It would be interesting to compare the Early Modern English frequencies with Present-day English, but no corresponding data seems to be available. Tesch's quantitative study of Present-day English (1990) does not give separate information on main clauses, but examples of this usage are included for instance in Sahlin's study (1979: 127-128), which is based on the Brown Corpus. The data of this study point to an interesting change that appears to have taken place in Late Modern English.11

S.S.3. 'Someone' versus 'anyone' Although it might seem that the 'someone' and 'anyone' paradigms are in complementary distribution in Present-day English, this is not the case.

44

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Instances of the 'someone' series can be found in nonassertive contexts and it is generally held that it is not only the clause type that triggers the choice between 'someone' and 'anyone' (e.g. R. Lakoff 1969; Sahlin 1979: 139-141; Quirk et al. 1985: 83-84). According to Lakoff (1969: 608), semantic notions such as presupposition, speaker's and hearer's beliefs and previous discourse play an important role in the choice of the variant in nonassertive contexts. 'Some' represents a positive orientation, 'any' neutral or negative. On the other hand, the differences between the 'some' and 'any' paradigms may be very subtle, and in some cases either form is acceptable (Sahlin 1979: 140-141). The infrequency of data of the 'someone' paradigm makes it difficult to draw conclusions from the usage in the corpus. In Old English poetry, where SUM is relatively frequent, the use may even have stylistic effect as in (60) and (61), which are both one of a group of similar structures following each other (see also Rissanen 1986). As mentioned in section 3.2.2, the members of this paradigm appear in assertive clauses except for once in a conditional clause (example (62)). (60)

Sum sceal mid hearpan aet his hlafordes fotum sittan, (OEX/3 The Fortunes of Men 80-81) 'One will sit with his harp at his lord's feet,'

(61)

... Sum masg styled sweord, waspen gewyrcan. (OEX/3 Christ 679-680) 'Somebody can make steel swords and weapons.'

(62)

..., but if some should come vnto me in that necessitie and extremitie which I can imagine, and aske; may I borrowe money of these Vsurers... (EModE2 Smith, Sermons E7R)

(63)

And what wyltow seyn of this: yif that a man hadde al forlorn his syghte, and hadde foryeten that he evere sawhe, and wende that no thing ne faylede hym of perfeccioun of mankynde; now we that myghten sen the same thinges — wolde we nat wene that he were blynd? (ME3 Chaucer, Boethius 448.C2)

Indefinite pronouns with singular human reference

45

The clause here represents a hypothetical condition that could be paraphrased by 'supposing that' or 'assuming that'. In Present-day English pronouns of the 'someone' series are often used in such cases (Sahlin 1979: 151-152). Our impression is, however, that in times past the form A MAN was favoured in this type of conditional (example (63), not included in the corpus). As mentioned in section 3.2.1, an increase in the frequency of the assertive paradigm appears to have have taken place after 1700. It is possible that the present subtle variation between 'some' and 'any' came about only in the late modern period, when obviously the 'someone' series found new areas of usage to account for the increase. Some evidence is found in the Bible translations; where the translations in the Helsinki Corpus have ANY MAN or A MAN in nonassertive contexts, the Revised English Bible has SOMEONE. This is particularly true of questions, of which John 4:31-33 serves as an example: (64)

And in the meane while his disciples prayed him sayinge: Master, eate. He sayde vnto them: I have meate to eate, that ye knowe not of. Then sayd the disciples bitwene them selves: hath eny man brought him meate? (EModEl Tyndale, The New Testament) 'MEANWHILE the disciples were urging him, "Rabbi, have something to eat." But he said, "I have food to eat of which you know nothing." At this the disciples said one to another, "Can someone have brought him food?'" {Revised English Bible)

The Revised English Bible implies that the disciples expect an affirmative answer, in other words that Jesus has received food from someone without their knowing of it. This interpretation is not offered by the earlier translations.

3.4.

The universal paradigm: 'everyone'

The universal paradigm consists of pronouns with distributive meaning, since they pick out the members of a set singly (Quirk et al. 1985: 382).

46

Helena Raumolin-Brunberg

and Leena Kahlas-Tarkka

3.4.1. The variants Table 5a. The universal paradigm: 'everyone'. (Absolute figures). OE1 OE2 OE3

OE4

ME1 ME2 ME3 ME4 El 18 5 6 4 21 6 0

8 9 24 7 11 4 0

20 25 6 30 6 8 0

60

63

95

EACH EVERY EACH MAN EVERY MAN EACH ONE EVERY ONE EVERY BODY

0

28

139

25

0

22

47

34

0

0

0

0

0

0

0

0

17 1 28 4 14 0 0

Totals

0

50

186

59

64

E2

E3

9 11 0 34 1 4 1

5 2 6 14 1 23 0

14 1 0 5 0 19 6

60

51

45

Table 5b. The universal paradigm: 'everyone'. (Percentages). OE1 OE2 OE3 EACH EVERY EACH MAN EVERY MAN EACH ONE EVERY ONE EVERY BODY

OE4

ME1 ME2 ME3 ME4 El 27 1 44 6 22 0 0

0

56

75

42

0

44

25

58

0

0

0

0

0

0

0

0

30 8 10 7 35 10 0

13 14 38 11 18 6 0

21 26 6 32 6 8 0

15 18 0 57 2 7 2

E2 10 4 12 28 2 45 0

E3 31 2 0 11 0 42 13

A s Tables 1, 2, 5a and 5b show, the number of the variants is larger in the universal paradigm than the other series. At the early stages of the language the n u m b e r of variants is particularly great, all the variants included in this study i.e. ^ELC, /EGWHILC, ^ G H W A , £Gi>ER, ,EGHW,EI>ER, GEHWILC and GEHWA appearing throughout the Old English period. For convenience' sake, we have classified all the predecessors of the later simple items EACH and EVERY as simple, even in Old English, although their f o r m s m a y reveal a c o m p o u n d origin. Examples of the different variants are given as (65)—(81).

Indefinite pronouns with singular human reference

47

(65)

cymö to demende eallum mancynne on {)am micclan daege, celcum be his daedum (OE3/4 JElfric's Letter to Sigeweard 56) 'He will come to judge all mankind on that great day, each one according to his deeds'

(66)

Pa het se gerefa heora celcum gebindan leades bloman on heora swyran (OE2/3 Martyrology 20) 'Then the judge asked that a mass of iron be bound on their necks'

(67)

How eche of thyse thre Ihu~s rehercyd dyde helpe. & saue ye people in his tyme /... (ME4 Fitzjames, Sermo Die Lüne A2V)

(68)

And Jjaet is witodlice ful soö, gelyfe öe wylle, an tima cymö ure ceghwylcum f>aet us waere leofre })onne eal J^ast we on worulde wiöasfitan us lasfaö, (OE3 Wulf start's Homilies 13 230) 'And that is certainly true, believe it who will, that a time will come to each of us when we would prefer that worldly life be left behind us'

(69)

ond ic haslan masg ceghwylcne anra, £>ara {)e him biö egesa to me (OEX/3 The Dream of the Rood 64) 'and I can save everyone who fears me'

(70)

frumcyn feora, faederasöelo gehwces (OEX/3 Exodus 361) 'the origin of human beings, the ancestry of each'

(71)

And so hi seghen {so J?et bi f e Morghen waren icomen. {)et hi waren last icume. Hedden here euerich ane peny; (ME2 Kentish Sermons 220) 'And so those who had come in the morning said that they were the last ones. Each of them had a penny.'

(72)

... the seid ordynance to endure and be available to ev~y of the said p~sons as longe as he abideth... (EModEl The Statutes of the Realm III 26)

48

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

(73)

for J)u eart celces marines fultum, Jje on {>e geleafaö (OE3/4 Α Passion of St Margaret 175) 'because you are the helper of everyone who believes in you'

(74)

For hy biswikejj euchan mon Pat mest bi-leueö hem ouppon. (MEX/2 The Thrush and the Nightingale 101) 'Because he deludes everyone who has most faith in him.'

(75)

öa gesinhiwan mon sceal manian, & eac gehwelcne mon, Öaet hie no lass ne ne geÖencen whast oöre men him forberaö & geöafiaö (OE2 Alfred's Cura Pastoralis 397) 'one must admonish married people and everyone else as well not to think less of what other people tolerate and approve'

(76)

fjset masst öearf is, £>at ceghwelc mon his aÖ & his wed wasrlice healde (OE2 Laws (Alfred) 46) 'what is most needful is that everyone keep his oaths and promises faithfully'

(77)

And Philip seyde to Crist })at loues of two hundret pens suffisede not to hem, Jjat eche man take a luytelwhat. (ME3 Wycliffite Sermons 1 412)

(78)

That was the true light, which lighteth euery man that commeth into the world. (EModE2 The Authorized Version John 1 9)

(79)

Lord, to J>e is mercy, for £>ou shalt 3elde to ichon efter his Werkes (ME2 The Earliest Complete English Prose Psalter 72)

(80)

Pat echone shuld hys owne charge bere, And J)at he shulde warne hyt none, But 3yue hyt ήιφβ to euerychone; (ME2/3 Mannyng, Handlyng Synne 1317) 'That everyone should take care of his own responsibility and he should not blame anyone for it but pass it on to everyone'

(81)

..., I mean the place of Chief Justice of Chester, for which I think every body does believe himself qualified. (EModE3 John Somers, Original Letters (Illustrative of English History) 329)

Indefinite pronouns with singular human reference

49

As regards the frequency of this paradigm, there is a declining development from OE2 where there is a peak of 84 instances per 100,000 running words. With the exception of OE4 (only 6 instances) the decrease is relatively gradual, reaching 45 in ME4, when the distribution between the different variants is also relatively even. The frequency then decreases to 26 occurrences per 100,000 words in EModE3. Although it is difficult to account for this decrease, we might suspect that there has been a stylistic change towards more explicit referentiality and cohesion. Unlike the more central pronouns indefinite pronouns seldom function as cohesive elements (see section 2.5, above). Tables 5a and 5b show the frequencies throughout the corpus so that all the variants of EVERY and EACH are treated as one in Old English and only from ME1 onwards as two different types. After the peak in OE3 the simple pronouns EACH and EVERY are quite common in ME4. In Early Modern English both of them are on the decrease until EModE3, when EACH suddenly reaches its highest frequency but EVERY appears only once. On the other hand, the compounds with the modifying part EACH practically disappear during the early modern period, and EVERY becomes the only variant used in compounds (EVERY MAN, EVERY ONE, EVERY BODY). Apparently there was never a pronoun or prepronoun combining EACH and BODY. In EModE2 EVERYONE overtook the dominating role from EVERY MAN. The Shakespeare Corpus differs from the corresponding Helsinki Corpus subperiod EModE2 especially in the use of EACH, which in the Shakespeare Corpus accounts for almost 45% of the cases (ca 10% in EModE2). A considerable difference can also be attested in the use of EVERY ONE, where the percentage in the Shakespeare Corpus is 21% and in the Helsinki Corpus EModE2 45%. Shakespeare does not use the variant EVERY BODY at all. The Century Corpus, on the other hand, has this new variant in ca 11% of the cases, which is exactly the same percentage as in the latest subsection of the Helsinki Corpus.

3.4.2. Syntactic and semantic properties The variants of the universal series can also be divided into two groups, the simple items together with the compounds in -ONE on the one hand and the compounds with MAN and BODY on the other. It is to be noted that the function of ONE (an/anra) in the Old English phrases is not a pronominal one, but is rather used for emphasis and individualization.

50

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Therefore no instances of ONE in OE1-OE4 are included. (For the function of ONE, Rissanen 1967 and this volume, and Kahlas-Tarkka 1987). The borderline goes with the partitive OF-phrase. EACH, EVERY, EACH ONE, and EVERY ONE are all employed as heads of OF-phrases (see examples (66), (67), (69) and (72)), while EACH MAN, EVERY MAN and EVERY BODY never found in these phrases. It is, however, important to notice that all variants also appear as independent unmodified pronouns, although EVERY loses this potential in Early Modern English (example (82)). EVERY is also the only variant never modified by a relative clause. (82)

.../the kynge dyde do ordeyne so moche mete / that euerych fonde ynough /... (ME4 Caxton, The History of Reynard the Fox 51)

(83)

Jx> {)e kyng of fraunce y-herde |)is, anon assemblede he is dousse pers eueruchon (ME2 Historical Poems 11) ...'each of his twelve peers'

(84)

And the thinges eek that men wenen ne haven none soules, ne desire thei nat, iche of hem, by semblable resoun to kepyn that that is hirs (ME3 Chaucer, Boethius 435.C2) 'Do not those things that people believe have no soul desire, each of them, for a similar reason, to keep that which is theirs'

(85)

I xall convycte J^em, I hope, euerychon. (ME4 Mankind 167)

(86)

...and the people departed, talkyng diuersly of the matter euery man as his fantasye gaue hym. (EModEl More, The History of King Richard III 80)

Not unexpectedly, most of the variants function as subjects and objects. Both EACH and EACH ONE, which were freely used as subjects in Middle English and earlier, begin to be rare in this function in non-reciprocal use in Early Modern English. On the other hand, EVERY ONE is not found as a subject, except in apposition, until Early Modern English. This compound is not only quantitatively the most expansive of the variants, but it also readily adopts new modification patterns and functions during the early modern period.

Indefinite pronouns with singular human reference

51

There is a syntactic characteristic peculiar to the universal paradigm, viz. the use of apposition. As examples (83)—(86) show, the pronouns appear in apposition to plural subjects and objects, emphasizing the distributive reference to every member of the set. Instances of this appositive usage is found for all variants except for EVERY BODY, and they are particularly frequent with EACH ONE. As time wears on, specialization occurs to the extent that during the last subperiod of the Helsinki Corpus only EACH is found as an appositive. It is also worth mentioning here that the pronouns of the universal paradigm do not occur in existential THERE-sentences. This is one of the properties bringing the indefiniteness of the universal pronouns into question (Safir 1987; see 2.2, above). As regards the expression of the set, in the earlier subperiods there are no great or obvious differences between the behaviour of the variants. As a result of later specification, the MAN compounds came to refer to all people in general (example (87)), which is relatively common even in Old English, while the simple items use grammatical means, such as the OFphrase (examples (67), (68) and (72)), and the subject and object associated with apposition (examples (83)-(86), dousse pers, they, pem, and the people, respectively). It is also worth pointing out that only the simple pronouns EACH and EVERY are found to refer to two persons, the latter only in Middle English (examples (88)-(89)). (87)

Lette euerie man do his owne busines, and folow his callying. (EModEl Latimer, Sermons 29)

(88)

..., and bring yee before the LORD euery man his censer, two hundred and fifty censers, thou also and Aaron, each of you his censer. (EModE2 The Authorized Version Numbers 16 17)

(89)

Also to the twey Clerkes that bene in the chirch euerich of hem

ij.s. (ME3 Testaments and Wills 209) Finally, let us take a look at a pragmatic property of universal pronouns, i.e. the dichotomy between strict and loose application (see Labov 1985). Strict application conforms to the logical tradition, where all members of a set are picked out exhaustively. Application is loose when the members of a set are referred to, but not necessarily exhaustively. Loose interpretation is common in everyday language for emphasis, intensification, exaggeration, etc. According to Labov (1985: 193), EACH in Present-day

52

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

English is almost entirely limited to formal language and to the strict interpretation, while other universal quantifiers can be found in loose application. Although historical corpora can provide evidence of informal language only to a limited extent, it was possible to trace examples of loose application in the Helsinki Corpus. All instances of EVERY BODY in subperiod EModE3 represent loose interpretation (example (90)) and there are also examples of EVERY ONE from the last two subsections of Early Modern English (example (91)). As regards EACH, it is not found in this function either as a simple indefinite or the modifying element of a compound after ME3, even though commonly so in earlier subperiods (example (92)). An excellent example of EVERY MAN in loose interpretation dates back to the same subperiod (example (93)). In the earliest subperiods all variants appear in instances of loose application, as examples (94) and (95) illustrate. (90)

... I mean the place of Chief Justice of Chester, for which I think every body does believe himself qualified. (EModE3 John Somers, Original Letters (Illustrative of English History) 329)

(91)

Symon at firste beinge the yongeste apprentice of four, was put to all the worste, and being little and smalle of stature and yong of yers, every on did triumphe over him ... (EModE2 Forman, The Autobiography and Personal Diary of Dr. Simon Forman 6)

(92)

..., the architriclyn clepith the spouse, and seith to hym, Ech man settith first good wiyn, and whanne men ben fulfillid, thanne that that is worse... (ME3 Wycliffe and Purvey, The New Testament John 2 10) 'the governor calls the bridegroom, and says to him that every man at the beginning serves good wine, and then, when men are satisfied, that which is worse..'

(93)

But when y was ryche, alle men me honourede and worsshepede; and now euery man hath of me scorne and despite! (ME3 The Brut or the Chronicles of England 19)

Indefinite pronouns with singular human reference

53

(94)

& wenö J)aet celcum men sie swa swa him si, & celcum men öince swa swa him joined. (OE2 Alfred's Boethius 122) 'and he thinks that the everybody's circumstances are like his, and things seem to everyone the same as they seem to him'

(95)

Hio on sybbe forlet secan gehwylcne agenne eard, (OEX/3 Elene 598-599) 'She let everyone seek his own country in peace.'

3.4.3. Reciprocal

pronouns

Although some instances of reciprocal use in Old and Middle English appear within the nonassertive and negative paradigms, it is the universal series where grammatical ization of one of the present-day variants, EACH OTHER, occurs. This reciprocal compound pronoun developed by subperiod EModE3, at which time none of the other universal variants under examination appear in reciprocal use any longer. As examples (96)—(101) indicate, the development took place from a discontinued structure towards one unit. The clearest evidence of unity, use of both elements within one prepositional phrase, is not attested in the corpus before EModE3 (example (101)), even though there are examples in which the elements immediately follow each other (97)-(98). The unity also means that the reciprocal pronoun cannot function as a subject as before. The last subjects in the corpus are found in EModEl (example (99)). Only EACH is used in reference to two persons (example (100)). (96) Than they swore hit sholde be done, and [so] passed forth sir Launcelot, and ecchone of the bretherne halpe other as well as they myght. (ME4 Malory, Morte Darthur 200) (97)

...ac J?aet hy symle gearowe syn, and geworhtum beacne hy butan elcunge arisende caflice gehwyle operne forestasppe and to öam Godes weorce efste... (OE3 The Benedictine Rule 47) '...but they should always be ready, and having seen a sign rise without delay and each should go quickly before the other and hasten to God's service...'

54

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

(98)

And utan understandan J)aet nis nan rihtra dom f^onne ure celc odrum beode {jaet we willan Jjast man us beode. (OE3 Wulfstan's Homilies 13 228) 'And let us understand that there is no more righteous judgement that we would like people to give us than what we offer each other'

(99)

Then the taylour lokyd about & saw the mylner ron~e away & the sexten a nother way & hard the preest cry help: had wend it had bene the co-stable w1 a gret co~pany cryeng for help to take hym & to bryng hym to pryson for stelyng of ye shepe wherfore he threw downe the shepe & ran away a nother way as faste as he coud/ & so euery man was afferd of other wythout cause. (EModEl A Hundred Mery Talys 37)

(100)

A courtyer & a frere happenyd to mete togyder in a fery bote & in co~munycacyon betwene them fell at wordys angry & dyspleasyd eche with other/ & fought & strogled togyder/ (EModEl A Hundred Mery Talys 126)

(101)

Yea, with the prophaned I hold it is not Lawful, but dost thou conceive that Saints can play with each other? (EModE3 Penny Merriments 148)

3.5.

The negative paradigm: 'no one'

The negative paradigm consists of pronouns with an incorporated negative element. They are used to indicate that no member of a given set is singled out.

3.5.1. The variants The Helsinki Corpus data comprises only three variants, since the compound with ONE does not occur in the corpus. Apart from ME2, NO MAN is the most popular variant until EModE3, when NO BODY takes the lead (Tables 6a and 6b). The increase in the use of NO BODY is indeed astonishing: from 5% in EModE2 to 58% in EModE3.

Indefinite pronouns with singular human reference

55

Table 6a. The negative paradigm: 'no one'. (Absolute figures). OE1

OE2 OE3

OE4

ME1 ME2 ME3 ME4 El

E2

E3

NONE NO MAN NO BODY

1 1 0

14 26 0

36 64 0

5 28 0

28 53 0

61 12 0

31 71 1

46 68 2

16 62 0

20 38 3

11 6 23

Totals

2

40

100

33

81

73

103

116

78

61

40

E2

E3

Table 6b. The negative paradigm: 'no one'. (Percentages).

NONE NO MAN NO BODY

OE1

OE2 OE3

50 50 0

35 65 0

36 64 0

OE4

ME1 ME2 ME3 ME4 El

15 85 0

35 65 0

84 16 0

30 69 1

40 58 2

20 80 0

33 62 5

27 15 58

Examples (102)—(116) illustrate the variants in the Helsinki Corpus. It is noteworthy that the compounds in (114) and (115) are spelled as one word. (102)

jjast ure ncenig wiste hwasr hiora asni cwom. (OE2/3 Alexander's Letter 16) 'that none of us knew where any one of them came from'

(103)

beleaf f)asr nan butan an munec he waes gehaten Leofwine (OE4 Chronicle 207) 'there was no one else left except a monk whose name was Leofwine'

(104)

ne fwer nan ne swaeltasö, forfjam öe j^asr ne byö nan acenned (MEX/1 Bodley Homilies 12 128) 'no one dies there, as no one is born there'

(105)

Ne he ne sy gebletsod fram nanum jjara, J^e hine met (OE3 The Benedictine Rule 50) 'He is not blessed by any one of those who meet with him'

56

Helena Raumolin-Brunberg

and Leena Kahlas-Tarkka

(106)

Criblin ne schal nan of ow for luue ne for hure (ME1 Ancrene Wisse 216) 'None of you shall engage herself in needlework for love or for hire'

(107)

öat nis non bute öo öe godes wille luuieö and foljiö (ME1 Vices and Virtues 1 15) 'that there is no one except those who love and follow god's will'

(108)

Non ne ssel zigge uader min. bote jje ilke jjet ys his zone be kende wyj>oute gynnynge (ME2 Ayenbite of Inwyt 1 101) 'No one should say my father except he who is his son begotten without conception'

(109)

As jje tueie willames & richard . & gileberd atten ende . & god nolde non of horn . (ME2 Robert of Gloucester 729) '...as the two Williams and Richard and Gilbert at the end, and God did not want any of them'

(110)

A FRERE ther was, a wantowne and a merye, A lymytour, a ful solempne man. In alle the ordres foure is noon that kan So muchel of daliaunce and fair langage. (ME3 Chaucer, The Canterbury Tales. The General Prologue 208-211)

(111)

GeÖence he, Jiaet he nanum men ne deme J)aet he nolde öast he him demde (OE2 Laws (Ine) 44) 'He should consider that he does not judge anyone whom he does not want to judge him.'

(112)

WiÖ wifes flewsan, genim ]x>ne camb J)e heo ana hyre heafod mid cemde, & ncenig man aer mid cemde ne asflter cembe. (OE2/3 Medicina de Quadrupedibus 11) 'As to a woman's hair, take the comb that she alone and no one else has combed her hair with before or after.'

Indefinite pronouns with singular human reference

57

(113)

In is chaumbre he sat priueliche : at is boke fill stille: No man ne dorste come him neijh : to letten him of is wille. (ME2 South English Legendary 442-443) 'He sat in his room on his own, quiet at his book. No one dared go near him to hinder his will.'

(114)

Naman that heres this sail thynk to be perfit withouten lange trauaile and gret besynes in saule. (ME2/4 Rolle, The Psalter 22) 'No one who hears this shall think that he is perfect without working long and hard trials within his soul.'

(115)

Hee said nobody had said anything agt mee. (EModE3 Henry Oxinden, The Oxinden and Peyton Letters 280)

(116)

No, my good lords, it is not that offends; It is not that that hath incensed the Duke. It is because no one should sway but he, No one but he should be about the King (1592, Henry VI, Part I, lines 35-38)

Although the Helsinki Corpus does not contain any instances of NO ONE, it had been introduced to the language by EModE2, as Shakespeare uses it six times (example (116)). In the Century Corpus, representing one hundred years of English from 1680 onwards, NO ONE is used quite frequently, accounting for ca 15% of all instances of the negative set. Shakespeare's language also differs from the corresponding Early Modern English period EModE2 to the extent that NONE and NO BODY are more frequent in his writings than in the Helsinki Corpus. One reason for this discrepancy could be the inclusion in EModE2 of Bible extracts, where the compounds with MAN are strongly favoured. The abrupt increase in the proportion of the BODY variants in EModE3 could also find some explanation in the different set-ups between the subsections EModE2 and EModE3, the former with Bible excerpts, the latter without (for further discussion, see 4.1). On the whole, it is interesting to notice that Shakespeare's use of the BODY variants appears to support the pattern of their introduction presented in Figure 4, below: NO BODY was introduced first, and has the highest number of occurrences in Shakespeare's texts (24), SOME BODY has 8 tokens, ANY BODY 4 and EVERY BODY no tokens at all.

58

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

In Old English two variants of NONE appear, nan and ncenig. The last instances of ncenig are to be found in OE3. Both are used for phrases for NONE as well as for NO MAN, even though ncenig only covers less than 20% of all recorded examples of the negative paradigm in Old English. No obvious difference between the two variants can be detected, even though it may be assumed that ncenig may have at least originally been more emphatic. Both variants may even appear in immediate contact with each other, as in (117). (117)

Ne beo ncenig man her on worldrice on his gelohte to modig, ne on his lichoman to Strang, ne ηφβ to georn, ne bealwes to beald, ne bregda full, ne inwit to leof, ne wrohtas to webgenne, ne searo to renigenne. Ne f>earf Jjaes nan man wenan Jjaet his lichama mote ο{φβ masge f>a synbyrjjenna on eorjjscrafe gebetan. (OE2/3 The Blickling Homilies 10 109) 'No one in this world should be too proud in his thoughts, nor too strong in his body, nor to eager for hatred, nor too ready for mischief, not full of fear, not deceit too dear, nor contrive slander or to arrange treachery. No one should think that either his body or the burden of sin of his race can be amended in the grave.'

The relative frequency of the negative paradigm gradually decreases from ME1 to EModE3. In the subperiod 1150-1250 there were 75 negative pronouns per 100,000 running words. By 1640-1710 the frequency dropped to 23 instances per 100,000 words. This decrease can be explained by the tendency towards more analytic structures (for further discussion, see 3.5.3).

3.5.2. Syntactic and semantic properties The complementary distribution familiar to us from the three paradigms discussed above emerges here as well: while the simple pronoun is found with partitive OF-phrases (examples (102), (105H106), (109) and (118)), the compounds in -MAN and -BODY are never used in these constructions. (118)

Dyd not Moses geve you a lawe, and yet none of you kepeth the lawe? (EModEl Tyndale, The New Testament John 7 19)

Indefinite pronouns with singular human reference

59

(119)

...to come downe hymself and toke on hym oure mortallyte, gave us a newe lawe, wold suffre none but hymselfe to be oure mayster... (ME4 In Die Innocencium 9)

(120)

Also, he is not aqueyntyd wyth no body but wyth Wekys, and Wekys ad told hym J>at he wold... (ME4 Clement Paston, Paston Letters 200)

NO MAN and NO BODY mostly appear as unmodified subjects, while NONE prefers modification. The exclusive BUT-phrase offers a popular modification pattern (examples (103), (107H108) and (119M120)), occurring with all the three pronouns. Negative pronouns are found in existential THERE-sentences in specific appositive structures, too (example (121)). This type of expression resembles that which was found frequently in the universal set, there being a plural noun expressing the set and a singular pronoun singling out every member of the set as relevant. Interesting parallel cases are found in ME3, where NONE is found separated from a singular noun, possibly for rhyme (see also Rissanen, this volume). In example (122) it is reasonable to argue that the word noon is a determiner, but this argument is hardly valid in (123), where the indefinite article a precedes the noun phrase bettre preest. These structures are only found in Middle English. (121)

— good Lord, seyp he, make vs saue, for seyntes per is none as per were. (ME3/4 Middle English Sermons (MS. Royal) 251) 'good Lord, he says, save us, because there are no saints as there used to be.'

(122)

In al the parisshe wif ne was ther noon That to the offrynge bifore hire sholde goon; (ME3 Chaucer, The Canterbury Tales. The General Prologue 449-450)

(123)

What so he were, of heigh or lough estat, Hym wolde he snybben sharply for the nonys. A bettre preest i trowe that nowher noon ys. (ME3 Chaucer, The Canterbury Tales. The General Prologue 522-524) 'No matter, high or low, he would chastise him there and then. I do not think that there is a better priest anywhere.'

60

(124)

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

What man that mihte wepne bere of alle he wolde non forbere; (ME3 Gower, Confessio Amantis I 178) 'He could not stand anyone who could carry weapons.'

Discontinuity is a relatively common phenomenon among pronoun-headed noun phrases, and the negative indefinites make use of this pattern more than the others (example (124)). The set is expressed by various means as in the other series. Syntactic means are preferred by NONE, while the MAN compounds usually refer to human beings in general.

3.5.3. 'No one' versus ΝEG + 'anyone' An intriguing issue in pronoun use is the choice between a pronoun with an incorporated negative element (synthetic negation) and a separate negator plus a nonassertive pronoun (analytic negation, see e.g. Tottie 1981, 1983, 1991). Table 7 compares the negative paradigm with the negativecontext instances of the nonassertive pronouns from ME3 onwards. The table only gives a rough idea of the development, since exclusion of the preverbal occurrences of the negative pronouns is the only measure taken to enhance the interchangeability of the data. In the earlier subperiods this kind of comparison is less relevant and is complicated by the more flexible word-order and the frequent use of multiple negation. In Old English multiple negation is almost a rule, and the order of the elements of the negative phrase is very flexible in the early subperiods (examples (104)(109), (111) and (113)). On the whole, it seems that synthetic negation is the most frequent structure in Old and Early Middle English, which may also partly explain the relatively low frequencies of ANY in the early periods, except in OE3. NEG + 'anyone' does appear in Old English, and a striking example in Wulfstan's homilies might be mentioned, in which sentence initial Ne (cefre) cenig man appears as often as 25 times and can be most probably taken as a conscious stylistic choice for emphatic expression (examples (125H126)).

Indefinite pronouns with singular human reference

o\ Ή es in o\ Ν

S £ •si

NO

r- r»i Ο Ο fN fN ΓΟ fN

•ο-o N ο ΓΛ

§§

m fN Ο

oo r^ ο

© oN fN Tt

NO Ο

oo oo TT Tf TT CN

ο rm

w T3 Ο2 s οin w

tN r-

ο ο ο

«

Tt

^

fN

OO fN m no

in

©

fN fN

oo -Η Ο Ο fN

OS fN

Ο Ο mι

η 't η Tf m

>n 0\

ci rm no

>n

W 2 S 2

t Tf fN fN m

cn NO

— fN ο

r~ ο m ΓΛ NO

•M-

-> -H

in ro

m on ο

ο «TΝ T ψ—ι W "ι

no

NO

a

ο Η

ο

ΓΛ

NO

No eowige. (OE3 Wulfstan's Homilies 10c 204) 'No one should arrest another person too easily, nor deprive him of his freedom unlawfully.'

(126)

Ne cefre cenig man idelnesse lufige ealles to gelome. Ne cefre cenig man unnyt lof & idel gylp lufige to swyöe. Ne cefre cenig man ofermetta lufie, ac aefre hy ascunie. (OE3 Wulfstan's Homilies 10c 204) 'No one should love idleness too much. No one should love useless praise or meaningless boasting. No one should love pride, but should always avoid it.'

(127)

& ne beo cenig man aeniges teames wyröe... (OE3 Laws (Eleventh Century) 326) 'No one is worthy of claiming his property in court...'

Despite some surprises, for instance the extremely low proportion of analytic structures in ME4 (5%), the overall tendency turns out as expected, with the synthetic negation decreasing in Early Modern English, although the development is not linear. A comparison of the figures of Table 7 with those of Tables 4a/4b and 6a/6b reveals roughly the same proportions between the different variants in each paradigm, except for EModE2, where the simple pronouns NONE and ANY appear more frequently in Table 7 than Tables 4a/4b and 6a/6b. The data is too limited for tracing the lexical diffusion of analytic negation (cf. Tottie 1991). Nevertheless, one interesting phenomenon is worth mentioning in this context; while negative pronouns abound in existential THERE-sentences in Middle English (example (128)), they become rarer in Early Modern English. The first instances of existential THERE-sentences with analytic negation are found in EModE2 (example (129)). (128)

And he encresyd so mervelously in that scyence that al Oxford had grete wondyr of hym for his grete connyng, for there was none lyke hym in all Oxford. (ME4 The Life of St. Edmund (in Middle English Religious Prose) 168)

Indefinite pronouns with singular human reference (129)

4.

63

..., ther was not any in the house that had the measels but they. (EModE2 Joan Everard, Barrington Family Letters 115)

Textual variation

It is obvious that in a study like this, with a wide chronological coverage and a text corpus which necessarily represents heterogeneous text material, it is very difficult to find direct linear continuity in pronominal uses. This continuity obviously exists in the language, but the limited material available from the early phases of English, particularly the obvious lack of material from the transition period between Old and Middle English, and the selective nature of the corpus may distort reality. The early material available represents different dialect areas, which makes a comparative study even more difficult, although it obviously also enables us to find dialectal peculiarities in language use. There are some aspects in the uses of the indefinites under scrutiny which particularly invite a study of textual variation.

4.1. Dialectal variation and genre-specific priorities In Old English in particular, the great variety of the EVERY pronouns is an interesting detail which invites speculation as to its cause. It has been suggested tentatively elsewhere that ^EGHWILC might have been particularly favoured by Anglian (Kahlas-Tarkka 1987: 128). This assumption is based on the fact that £iGHWILC is frequent in the Anglian versions of the gospels {Li and Ru) as well as in the Blickling Homilies and Martyrology which both are generally considered to contain many Anglian elements. This does not, however, mean that /EGHWILC would be exclusively Anglian. From similar textual comparison it can also be suggested that fiLLC is favoured in the West-Saxon dialect, and this of course would also explain the continuity of 7ELC in later times.

64

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Some dialectal peculiarities in the choice of the EVERY variants can also be detected in the Early Middle English period (cf. Kahlas-Tarkka 1987: 156-160 and 1994: 307-309). The differences seem to be restricted to the choice of the words themselves rather than to any syntactic peculiarities. Thus ech(e) can be characterized as a typically southern, more restrictively an East Midland form, as well as ilk which is also found in the northern areas. Uch(e) is particularly favoured in the West Midland dialect. A comparatively rare form euch(e) occupies a belt running south of Hereford and Worcester and is frequently attested in Sawles Warde, Hali Meidhad, Ancrene Wisse and the Katherine Group. Compounds like eauer-euch also appear. In the East Midland dialect the prevailing originally compound forms are euery, eueri, every. Southern and Kentish texts tend to be conservative in their uses, showing frequent instances of Old English celc. In the later subperiods dialectal differences seem to be levelled out, obviously due to the rise of the Standard as well. The compilation principles of the Early Modern English section of the Helsinki Corpus included avoidance of clearly dialectal texts. Hence it cannot provide data for the study of dialectal usage (see Nevalainen—Raumolin-Brunberg 1989 and Raumolin-Brunberg—Nevalainen 1990). Bolinger's claim (1976) about the particularly frequent use of the compounds in -BODY in the Scottish dialect does not find support in the historical development evidenced in the Helsinki Corpus of Older Scots (see Meurman-Solin 1993). In addition to local dialects, some genre-specific priorities can be detected in the present material. Poetry and prose obviously differ in style and have different demands, due to alliteration in Old English or rhyme in Middle English, which even effects the choice of variants when such are available. All extracts from the Bible translations favour the compounds with MAN. No occurrences of the BODY indefinites can be found, while a few instances of the simple series and EVERY ONE appear. This preference for MAN is also reflected in other religious texts, such as religious treatises, homilies and sermons (cf. Raumolin-Brunberg 1994b). The translations of Boethius during the various subperiods (OE2, ME3, EModEl-3) also prefer the MAN compounds. The use of this variant appears to be connected with frequent generic reference to human beings and human life in general, a characteristic of religious and philosophical writing.

Indefinite pronouns with singular human reference

65

The use of the simple pronoun EVERY, on the other hand, is strongly favoured in legal documents and The Statutes of the Realm. In Early Modern English it is only found in legal texts, official correspondence and trial proceedings. As was mentioned in section 3.4, EACH gradually gave way to EVERY as the determiner morpheme of the compound. The Wycliffite Bible still preferred EACH in this context, and it was also found in contemporary homilies and late Middle English sermons. At a more individual level, it has been pointed out earlier that a few idiolectal features can be discerned even in the pronominal uses. A comparison between ^ l f r i c and Wulfstan reveals that both favour JELC, but in addition jElfric frequently uses GEHWILC as a determiner (gehwilc man), whereas Wulfstan prefers GEHWILC modified by a genitive plural form {manna gehwilc). ^GHWILC, on the other hand is almost absent from -iElfric's writings, whereas it is frequent in Wulfstan (cf. e.g. KahlasTarkka 1987: 122).

4.2. Oral versus non-oral genres Although it is evident that historical corpora like the Helsinki Corpus cannot include any specimens of spoken language, it is possible to argue that, in relative terms, some of the texts resemble the typical oral mode of expression better than others (e.g. Biber 1988). This is the view adopted in many studies following the tradition introduced by e.g. Tannen (1982) and Chafe (1982). Despite the general usefulness of the distinction between oral and literate modes of writing, it is difficult to assess on external criteria which genres could be characterized as more oral than others. For this study we decided to classify as oral the following genres from Late Middle English onwards: fiction, drama and private letters. This choice was supplemented by diaries, autobiographies and trial proceedings in Early Modern English. They represent texts with passages of direct speech, which were often written for private purposes or entertainment. The Helsinki Corpus does not provide relevant data for this approach from before 1350.12

66

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Indefinite pronouns with singular human reference

67

Figure 3 and Table 8 give the percentages of the occurrences of the indefinite pronouns in the genres characterized as oral. The percentages stated below the bars indicate the proportion of the oral texts from all texts in each subsection of the corpus. During the Middle English period 3, 1350-1420, the share of the oral texts was therefore very low, only 9%, while during the last subsection of Early Modern English, the oral texts comprise 40% of all data. A horizontal line has been drawn over the bars of each subperiod to mark the proportion of the oral texts. The purpose of the graph is to show that the pronouns whose bars exceed the line are favoured in oral texts and those pronouns whose bars do not reach the line are given priority in the non-oral texts. For instance, in EModEl ( 1 5 0 0 1570) and EModE2 (1570-1640) the simple pronouns, the first bar on the left for each period, both have 44% and the proportion of oral texts is 36% and 35%, respectively. This means that the simple pronouns are slightly more frequent in the oral texts than in the non-oral ones. Table 8. Indefinite pronouns with singular human reference in oral genres. Oral genres Ν %

All genres Ν

9.1%

ME3 1350-1420 Simple -MAN -ONE -BODY Totals

Proportion of oral genres

8 5 1 0

14 3 7 0

59 143 15 1

14

6

218 22.0%

ME4 1420-1500 Simple -MAN -ONE -BODY

27 37 9 3

24 29 64 100

110 128 14 3

Totals

76

30

255

68

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

Table 8. Continued. Oral genres Ν %

All genres Ν

EModEl 1500-1570

Proportion of oral genres 35.5%

Simple -MAN -ONE -BODY

25 39 3 3

44 28 50 100

57 142 6 3

Totals

70

34

208

EModE2 1570-1640

35.2%

Simple -MAN -ONE -BODY

25 13 5 6

44 14 19 100

57 95 27 6

Totals

49

26

185

EModE3 1640-1710

40.0%

Simple MAN -ONE -BODY

20 5 8 47

51 15 24 85

39 33 34 56

Totals

80

49

162

Although Figure 3 has some problematic features, such as the high overall proportion of the indefinite pronouns in the oral texts in the last Middle English subsection (ME4), the graph clearly indicates that there is one pronoun type consistently favoured in the oral texts, viz. the compounds in -BODY. All instances except the very first one in ME313 occur in the texts characterized here as oral until the last subperiod of Early Modern English, and even there the proportion of oral texts is very high (85%). The usage clearly points to a colloquial origin for the BODY-compounds, a fact that is still reflected today in comments about its being the less

Indefinite pronouns with singular human reference

69

elegant or less literary of the two present-day compound variants (see e.g. Raumolin-Brunberg 1994b). Another interesting observation is that the share of the ONE compounds is also high in the oral genres for the subperiods ME4 (1420-1500) and EModEl (1500-1570), but decreases later. This too might point to an earlier preference for these forms in colloquial language until they were adopted by the literate varieties. However, as regards all quantitative information in this connection, one caveat must be kept in mind; before subperiod EModE3 (1640-1710) the absolute figures for the new compounds are low, BODY from 3 to 6 occurrences per subperiod, ONE from 1 to 9, and the instances of ONE are practically restricted to EACH ONE and EVERY ONE alone (see Table 8). Furthermore, Figure 3 shows that the simple pronouns are fairly frequent in the oral genres in Early Modern English, while the compounds in -MAN tend to be favoured in the more formal types of writing.

5.

Grammaticalization

As mentioned in the Introduction to this volume (section 5, above), pronominalization represents a relatively weak process of grammaticalization. The source concepts for the indefinite pronouns whose development is examined here are either nouns, e.g. MAN and BODY, or grammatical items that become more grammatical, e.g. ÄsLC and ONE. The nominal character of the elements is preserved during the process, and hence the development can be called structure-preserving or weak. The grammaticalization of indefinite pronouns offers evidence for unidirectional processes, the degree and extent of which can vary a lot.

5.1. Old English GEHWA, ^EGIIWILC or others, and Early Middle English EVERY from MFRE MLC The etymology of all the five OE words for EVERY and EACH shows that, even though we have treated these pronouns as simplexes, they are really complex forms, compounds, and can thus be regarded as instances of grammaticalization. The Old English -HWA pronouns go back to a Proto IE stem used for indefinite and interrogative pronouns (see especially

70

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

Hermann 1940-41, and the brief summary in Kahlas-Tarkka 1987: 13-17), whereas the -HWILC words are compounds of this stem and the Gmc noun *lika 'body, form'. Only a few of the simple forms of the pronouns were originally used in the sense of 'everyone'. It is the intensifying *ga and *aiw that mark the distinction between 'everyone' and other indefinite pronouns, and it seems that these particles were mostly used to form compounds of words meaning 'everyone'. The meaning of *ga is not totally clear, and the origin of its intensifying character is unknown, even though it seems to imply "togetherness". Most probably, however, it is the older of the two prefixes, and the introduction of *aiw 'always, ever' can best be explained by the weakening force of *ga "durch zu häufigen Gebrauch" (Hermann 194(M1: 189-190). The Old English pronouns were formed with a (aiw) or ge or both of them. For the very common pronoun celc two different etymological developments can be given: *alalika > *alika 'all + body', further emphasized by *aiw (a) would give celc, or from *aiwo(n) galiko-z (OE a gelic) 'ever alike'. The noun meaning 'body, form' would thus only lie behind the word in the first development. If a reconstructed development like that of Gmc 'every' evidenced by etymology can be seen as a case of grammaticalization, a further step in such a process was taken in the formation of Early Middle English EVERY from JEFRE JELC. In Old English ^ELC is by far the most frequent variant of the simple universal pronouns. It is probably this frequency that explains the rise of i£FRE ^ELC in Late Old English and the gradual expansion of this phrase. It is, however, still infrequent in Late Old and Early Middle English, but this again can possibly be explained by the limited material available from the transition period. The adverb J E F R E was occasionally used in phrases with additional intensifying adverbs (cefre and a 'ever and ay') or in combination with pronouns and/or adverbs with generalizing or intensifying force like hwa cefre 'whoever' cefre swa hwcer swa 'wherever' (Toronto DOE s.v. cefre). JEfre could also occasionally be found combined with celc. The different stages in the development towards a compound word are illustrated in examples (130) and (131). (130)

ac f>a he stod Jjaer swa hreowlice ana tomiddes eallum J)am folce, hit cefre be him micele swiöor celc man sprasc geond J>a byrig (OE3 Mfric's Lives of Saints 34 639) 'but when he stood there so miserably, alone in the midst of those people, everyone in the whole city spoke much about him'

Indefinite pronouns with singular human reference (131)

71

\>a westre sona Jsas landes, for ceuric man sone raeuede ojjer {>e mihte. (OE4 Chronicle MS Ε 1135.8) 'There was treason in the country, because everyone who could, immediately robbed the others.'

The DOE records five occurrences of AsFRE AiLC (s.v. cefr-celc) as one word from manuscripts of the twelfth century or later. The development of /EFRE Λ-LC into one word for EVERY was relatively slow, and Early Middle English spellings like ceuere-celch, ceuercelch or euer-ech still seem to indicate the writer's awareness of the compound nature of the pronoun. It is much less obvious in e.g. the frequent euerich forms in Ancrene Riwle. It seems thus that the slow grammaticalization process in which cefre lost its independent adverbial function was completed during the 13th century.

5.2. Compound pronouns with MAN, BODY, and ONE Although the indefinite compounds with MAN date back to the earliest Old English texts, owing to morphological similarity, we found it more appropriate to discuss them together with the more recent compound pronouns in -ONE and -BODY than dealing with them along with the other Old English developments. Characteristic of these items is a relatively weak grammaticalization process in which compounding leads to transparent end results (cf. Croft 1990: 231). As members of the basic vocabulary, the lexical items MAN and BODY offer themselves as natural sources for the grammaticalization of pronouns with personal reference (see e.g. Heine et al. 1991: 33-36 and Hopper —Traugott 1993: 97). According to the OED (s.v. man, 1.1), the prevailing meaning of the noun MAN in Old English was 'human being, (irrespective of sex and age)'. The meaning of BODY was 'a human being of either sex, an individual' (OED, s.v. body, III) in examples between 1297 and 1833 (see also section 5 in the Introduction, above). MAN underwent a semantic change later on, and 'male human being' became its main meaning. BODY also lost the sense relevant in this study. ONE, for its part, is a grammatical item which is very common in the creation of indefinite pronouns in different languages, English being no exception (Lehmann 1982: 52; Rissanen 1967 and this volume).

72

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

The increasing employment of the compound pronouns was associated with a Late Middle and Early Modern English typological change in the structure of the noun phrase. This development made a separate headword practically obligatory, and the indefinite pronouns came to comply with this full-NP pattern, with the exception of the pronouns modified by partitive OF-phrases. It is important to stress, however, that this development was slow, and in Early Modern English the use of the unmodified simple pronouns was not uncommon (cf. Raumolin-Brunberg 1994b). Standard historical grammars do not include the compounds with MAN among pronouns. Some deviations from this practice occur, however, as Jespersen (1914: 445-446) for instance mentions them along with the BODY compounds, and the Swiss scholars Fröhlich (1951), Meier (1953) and Jud-Schmid (1962) regard the forms with MAN as pronouns. Poutsma (1916: 1305) refers to MAN as a prop-word along with ONE, BODY and THING, among others. It is difficult to find criteria for establishing the pronounhood of the MAN compounds. It is just as problematic to define when the regular NPs consisting of a determiner and the noun BODY became pronouns. Franz (1939: 318) claims that the forms with BODY were not yet felt to be real compound pronouns in Shakespeare's times. It is evident that no morphosyntactic criteria are available (see e.g. Raumolin-Brunberg 1994b), but we can posit a continuum with the regular noun phrase at one end and the pronouns at the other. The indefinite pronouns are placed close to the regular noun phrase end, since they do not fulfil the conditions of prototypical pronouns (see section 5 in the Introduction, above). Semantically, however, the indefinite pronouns share the characteristics of central pronouns, viz. generality of meaning and undeterminedness (Quirk et al. 1985: 335-336). Evidence for compounding can also be found in stress placement, but this is a phenomenon difficult to trace in written texts. As far as the forms with MAN are concerned, the stress placement in the Present-day English relic-set phrases no man's land or everyman's library indicates that they are compounds. Support for their pronominal character can be found in the Bible translations, where the Revised English Bible (1989) consistently substitutes the pronouns with ONE for the MAN compounds (example (132)). The Bible also gives us an example where reference to a male person is excluded on contextual grounds (example (132)).

Indefinite pronouns with singular human reference

73

(132)

The New Testament, John 6 65 f>aet nan man ne masg cuman to me... (0E3, West-Saxon Gospels) that no man may come to me,... (ME3, Wycliffe) that no man can come vnto me,... (EModEl, Tyndale) that no man can come vnto me,... (EModE2, AV) that no one can come to me,.. (PE, REB)

(133)

When Iesus had lyfte vp him selfe agayne, and sawe no man but the woman, he sayde vnto hyr. (EModE 1 Tyndale, The New Testament John 8 10)

There is no way of knowing which routes into the language the compounds with MAN used, since these pronouns can be traced even in the earliest Old English texts. However, it is possible to look at their use and gradual disappearance from the language, which in turn may shed some light on the emergence of new the members of the paradigms during Middle and Early Modern English. It would seem that the semantic development of the noun MAN and the other pronominal uses of MAN played a role in the diminishing employment of the MAN compounds in Early Modern English. According to the OED, the dominant present-day meaning of MAN, 'human being of male sex' (s.v. man, II.4) existed in Old English in parallel with the meaning 'human being', and it gained more currency as time wore on. This development was possibly supported by the emergence of unambiguous synonyms meaning 'person', such as BODY and PERSON. Perhaps the semantic shift made MAN a less appropriate element for the indefinite pronouns referring to human beings in general and not only male persons. On the other hand, the narrowing of the scope of MAN in other pronominal uses, in particular its disappearance as the generic indefinite pronoun, may have influenced the gradual shift from the MAN compounds to those in -BODY and -ONE (see Rissanen, forthcoming). The increase of the use of ONE in other pronominal functions (see Rissanen, this volume) may have supported the steady spreading of the indefinites in -ONE. The first prepronouns with BODY appear in the fourteenth century (NO BODY), but the real expansion of these variants took place during the last of the Early Modern English subperiods, when the BODY compounds gained a frequency that surpassed the use of the other variants. During this period they had also spread to non-oral genres, as Figure 3, above, indicates.

74

Helena Raumolin-Brunberg and Leena Kahlas-Tarkka

The introduction of the ONE compounds to all series of indefinite pronouns took much more time than that of the ones in -BODY. In Old English ONE was often combined with the pronouns of the universal set for emphasis or individualization, and in these phrases it is difficult to see any signs of ONE having a pronominal function (example (134); see also Rissanen 1967; Kahlas-Tarkka 1987). The first instances of what we interpret as instances of EACH ONE stem from ME1 (1150-1250; examples (135) and (136)). In (135) metrical demands may have affected the choice of the structure, but (136) is a clear instance of running prose, and the use of euchan seems genuine, especially as there are six instances in all in the same text. ANY ONE and SOME ONE appear in the Helsinki Corpus in Early Modern English, and NO ONE is not found at all, but its existence in the language around 1600 is known from Shakespeare's texts. (134)

...

ond ic haslan maeg ceghwylcne anra, J>ara f)e him bi£> egesa to me... (OEX/3 The Dream of the Rood 64) 'and I can save eveiy one of those who fear me'

(135)

Patt illc an shollde witenn wel Whillc lott himm shollde re33senn, (ME1 The Ormulum I 14) 'that everyone should know well which part should be given to him'

(136)

fie feoröe suster rihtwisnesse ... demeö euchan his dom efter his rihte. (ME1 Sawles Warde 168) 'The fourth sister righteousness judges everyone according to what he deserves'

It is interesting to compare the routes by which the introduction of the compounds with BODY and ONE took place. Figure 4 shows that both entered the language through the pronouns of totality (see section 2.2, above). The forms with ONE are introduced in the universal paradigm, from which they gradually spread to the other series. Although the Helsinki Corpus has no evidence of its use in the negative paradigm, we know from Shakespeare that it was also available. The BODY compounds, on the other hand, arose in the negative paradigm and also spread to the other sets of variants.

Indefinite pronouns with singular human reference

3ο

ο S VO CΓ-

o ο Γ- Tt-

ο ο ο ^ οin in γ«S W— —

Tt Ο Ο [Ü © Ν «Ο Ο

ο ω ω ω

>«

Q Ο 03

ω ^ S Ω >· o g Q a Η Ο Ο CQ § s ο ο Ο Ο οο &ο ζ ζ

C I ο. ο "Β > ο Ω Κ a

.00

76

Helena Raumolin-Brunberg and Lee na Kahlas-Tarkka

The introduction of new forms in the two paradigms expressing totality is not surprising and complies with Davison's classification of the indefinites given in section 2.6. There is evidence from different languages that negators tend to lose their expressive power, and new elements and structures come to be added for this purpose (e.g. French ne...per sonne, pas, etc). The universal paradigm in English, as we have seen from the early developments, seems to accommodate a constant need for emphatic expression and it becomes a locus for repeated grammaticalizations (see 5.1, above). Nevertheless, Figure 4 raises a question which is difficult to answer. It seems that the other two series of the variants, the simple pronouns and the pronouns with MAN, were becoming inappropriate; the simple pronouns because of their functional load as determiner and singular and plural use, the compounds with MAN for reasons of semantic change. Hence they were gradually giving way to more appropriate pronouns. The question still remains as to why should both types of variants, the compounds in -BODY and those in -ONE, develop in all the four paradigms? Would not the language have functioned well enough with different items in different pronouns, as it did in Late Middle and Early Modern English? The latter situation is also the case with the other Germanic languages, for example German JEMAND, NIEMAND, JEDER-MANN/EIN JEDER, Swedish NÄGON, INGEN, VAR OCH EN/ENVAR and the Fenno-Ugric Finnish: JOKU/ERÄS, KUKAAN, JOKAINEN. Most present-day grammars claim that the pairs of compounds with ONE and BODY are identical in meaning (Quirk et al. 1985: 376-377; Jespersen 1914: 444). Bolinger (1976; 1977) nevertheless argues that there is a subtle difference in meaning between the series, ONE and its compounds being marked for closeness to the speaker and individualization, whereas -BODY is unmarked in these senses. Bolinger's (1976) test with undergraduates as subjects shows, however, that in most of the uses both alternatives were possible, and it is obvious that intuition plays an important role in decision-making. Hence it seems impossible to posit a semantic difference between the compounds which would explain their existence. As described above (section 3.1), there is a syntactic difference between the forms as well, the forms with BODY not appearing in partitive structures. Nevertheless, it would seem that so many overlapping syntactic uses remain that syntax can hardly be used to explain the existence of both series. Although textual constraints exerted and still exert influence on the choice of the compound variant, the development has rather

Indefinite pronouns with singular human reference

77

been towards convergent usage than the opposite. It appears that textual criteria also fail to account for the existence of the two parallel series. It is difficult to find reasons other than analogy, the search for symmetry or pressure for paradigmatic cohesion, as Lehmann (1982: 132-137) puts it. In other words, what we see is a development towards a tight and symmetrical repertoire of the indefinite pronouns (cf. Croft 1990: 234). What this development shows is at least that in terms of psychological reality, the group of pronouns discussed here appears to form one integrated category. The speakers of English created the double series for all paradigms apparently without paying any attention to their different degrees of indefiniteness (cf. sections 1-2).

6.

Concluding remarks

The above study covers a long time span and thus clearly shows how radical changes have taken place in one closed class of words in English, namely indefinite pronouns with singular human reference. The development clearly supports the traditional division of English into three main periods. In Old English there are two main syntactic types, the simple pronouns and those with MAN, but a great variety of words within the indefinite paradigms especially for the universal series. In Middle English, the rise of pronominal ONE changes the picture at the same time as the number of variants is somewhat reduced, apart from pure spelling variants. The last stage is reached in Early Modern English where BODY establishes itself along with ONE among the compound pronouns. Syntactic details like OF-phrases taking the place of earlier genitive plural constructions, analytic negation becoming more frequent in Early Modern English and EVERY becoming only attributive are all well illustrated by the present corpus. It has been noted above that the fact should never be ignored that a selective corpus may give a misleading picture of the linguistic reality. Keeping this in mind it seems, however, that the corpus material used for this study is representative enough to give an idea of the central lines of the development of the indefinite paradigm.

78

Helena Raumolin-Brunberg

and Leena

Kahlas-Tarkka

It has been interesting to notice that language does not always seem to aim at economy, the rise and continuing use of the compounds with BODY and ONE being one example. The search for symmetry may at least partly explain this, even though it seems to remain a question without a conclusive answer.

Notes 1.

2.

3.

4.

5.

The division of labour has been as follows: Leena Kahlas-Tarkka is responsible for the Old English and Early Middle English data (700-1350) and Helena Raumolin-Brunberg for Late Middle and Early Modern English (1350-1710). Raumolin-Brunberg has written sections 1, 2, 4.2 and 5.2 and Kahlas-Tarkka 4.1, 5.1 and 6. Section 3 represents joint work. Both coauthors subscribe to the general views adopted in the article. We would like thank Andrew Chesterman and Erkki Ahlström for reading section 2 of the paper. We also hope that our gratitude could somehow reach Professor Ossi Ihalainen, who commented on the same section. It appears to have been one of the last pieces of scholarly work he came to comment on before his untimely death in September 1993. Verb inflection does not always provide us with valid criteria. In Middle English in particular, the suffixes -TH and -S were used both in the singular and the plural. Text-internal comparisons have been required for reliable results. The Helsinki Corpus is divided into the following subperiods: Old English: OE1 -850, OE2 850-950, OE3 950-1050, OE4 1050-1150 Middle English: ME1 1150-1250, ME2 1250-1350, ME3 1350-1420, ME4 1420-1500 Early Modern English: EModEl 1500-1570, EModE2 1570-1640, EModE3 1640-1710 Hawkins (1978: 103-179) lists the conditions for shared sets, and it would seem that e.g. 'all people in the world' could not make a set in his framework. Chesterman (1991: 57), however, extends the discussion in this direction, when he argues that in Ί have a head' the noun head has the indefinite article because of exclusion, i.e. the relevant set is 'all heads' and hence there are other heads elsewhere in the world. Here it is important to bear in mind that reference to all members is not always to be taken literally but pragmatically 'more or less' all (Chesterman 1991: 66). This application is called by Labov (1985) 'loose application', which can be contrasted with the logic-based 'strict application'.

Indefinite pronouns with singular human reference 6.

79

This working definition is not unproblematic since knowing a person's identity varies according to situation. Sometimes it may, for instance, be sufficient to recognize someone's face without knowing the name; in other cases the name is very important. The corpus includes a good example of the problem: "So when we came to my Lady Lisle's, on the Tuesday night, somebody took the Horses, I cannot tell who, if I were to die; the two went in; and after..." (EModE3 State Trials, The Trial of Lady Alice Lisle, IV 121.C1)

The witness here has a specific person in mind, but his identity still seems to be unknown to him. This case may be interpreted as indeterminate as regards specificity. 7. The figures for EACH/EVERY include those for Old English GEHWILC, GEHWA, iEGHWILC, yEGHWA, JELC, £GPER and ^GHW^PER. 8. For the earliest quotations in the OED, see Raumolin-Brunberg (1994b). Rissanen's example (86) of SOMEONE in Wyclif s translation of Mark 9:38 (this volume) antedates the ones in the Helsinki Corpus. 9. Although we have interpreted this instance to represent the singular, the case is not as straightforward as it would seem on the basis of the verb form was. In Fox's autobiography the form was can also be found with plural nouns. 10. Any I is the indefinite nonassertive article in Sahlin's system and not relevant here (Sahlin 1979: 19). 11. The developments described in this section give rise to some speculations about the diffusion of the pronouns of the 'anyone' series into affirmative declarative main clauses. It seems possible that the introduction of these pronouns has followed a pattern similar to that of WH-relative pronouns, (the Accessibility Hierarchy; Keenan—Comrie 1977), according to which subjects are the last to accommodate the new grammatical item. It might have been that affirmative subordinate clauses, in particular THAT-clauses and relative clauses came first to employ pronouns with an ANY element in all syntactic functions, and this usage spread into nonsubjects in main clauses and to subjects later on. Nevertheless, this must remain conjectural until a larger corpus is studied. A counter-argument to the above might be that the matrix clause structure with a clause-initial subject followed by a relative clause is perceptually awkward and hence rather avoided. Nevertheless, parallel cases with HE THAT, EVERY/EACH MAN THAT and WHO(SO)EVER as matrix clause subjects abound in the early Bible translations, and consequently this explanation does not seem valid here (see Raumolin-Brunberg—Kahlas-Tarkka 1995).

80 12.

13.

Helena

Raumolin-Brunberg

and Leena

Kahlas-Tarkka

The list of texts characterized as oral in this study differs from that used in Raumolin-Brunberg (1994b), where a rough division into two types of writing, oral versus literate, was made. General biographies and travel stories are not included among the oral texts in the present study, although they were in Raumolin-Brunberg (1994b). This example interestingly juxtaposes NO MAN and NO BODY (prepronoun), perhaps in search of emphasis by lexical variation: ...J)u schost greue no-man·, and certayn, J>at howte be lißt e n o u j to Jje, for JJU myßt grue no-body, Jjey J)e woldest, bote 3 it Jni smyte hym wit Jjy tounge. (ME3 Aelred of Rievaulx's De Institutione Inclusarum 36) 'you should not terrify anyone, and certainly that ought to be easy enough for you, because you could not frighten anyone, although you would like to, without smiting him with your tongue'

References Biber, Douglas 1988 Variation across speech and writing. Cambridge: Cambridge University Press. Bolinger, Dwight L. 1976 "The /«-group: One and its compounds", in: P. A. Reich (ed.), The SecondLACUS Forum 1975. Columbia, S. C.: Hornbeam Press, 2 2 9 237. 1977 Meaning and form. London: Longman. Chafe, Wallace 1982 "Integration and involvement in speaking, writing and oral literature", in: Deborah Tannen (ed.), Spoken and written language: Exploring orality and literacy. (Advances in Discourse Processes IX.) Norwood, New Jersey: Ablex, 35-53. Chesterman, Andrew 1991 On definiteness. A study with special reference to English and Finnish. Cambridge: Cambridge University Press. Croft, William 1990 Typology and universals. Cambridge: Cambridge University Press. Davison, Alice 1981 "Contextual effects on 'generic' indefinites: Cross-linguistic arguments for pragmatic factors", in: David Sankoff—Henrietta Cedergren (eds.), 359-368.

Indefinite pronouns with singular human reference

81

de Jong, Francisca 1987 "The compositional nature of (in)definiteness", in: Eric J. Reuland —Alice G. B. ter Meulen (eds.), 270-285. Donnellan, Keith 1966 "Reference and definite descriptions", in: D. D. Steinberg—L. A. [1971] Jakobovits (eds.), Semantics. An interdisciplinary reader in philosophy, linguistics and psychology. Cambridge: Cambridge University Press, 100-114. Fernandez, Francisco—Miguel Fuster—Juan Jose Calvo (eds.) 1994 English historical linguistics 1992. (Current Issues in Linguistic Theory 113.) Amsterdam/Philadelphia: John Benjamins. Franz, Wilhelm 1939 Die Sprache Shakespeares in Vers und Prosa. Halle: Max Niemeyer. Fröhlich, Jürg 1951 Der indefinite Agens im Altenglischen unter besonderer. Berücksichtigung des Wortes man. (Schweizer Anglistische Arbeiten 15.) Bern: A. Francke AG Verlag. Givon, Talmy 1978 "Definiteness and referentiality", in: Joseph Greenberg—Charles Ferguson—Edith Moravcsik (eds.), Universals of human language, Vol. IV. Syntax. Stanford: Stanford University Press, 292-330. 1984 Syntax. A functional typological introduction. I. Amsterdam—Philadelphia: John Benjamins. 1990 Syntax. A functional typological introduction. II. Amsterdam—Philadelphia: John Benjamins. Hawkins, John A. 1978 Definiteness and indefiniteness. A study on reference and grammaticality prediction. London: Croom Helm. 1991 "On (in)definite articles: Implicatures and (un)grammaticality prediction", Journal of Linguistics 27: 405-442. Healey, Antonette diPaolo—Richard Venezky 1980 A microfiche concordance to Old English. (Publications of the Dictionary of Old English 1.) Toronto: The Pontifical Institute of Mediaeval Studies. Heine, Bernd—Ulrike Claudi—Friederike Hünnemeyer 1991 Grammaticalization. A conceptual framework. Chicago—London: University of Chicago Press. Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki.

82

Helena Raumolin-Brunberg

and Leena Kahlas-Tarkka

Hermann, E. 1940-1 'Jeder einzelne in den germanischen Sprachen', Nachrichten aus der neueren Philologie und Literaturgeschichte 3. (Nachrichten von der Akademie der Wissenschaften zu Göttingen, Philologisch-historische Klasse, N. F., Fachgruppe IV), 173-206. Hopper, Paul J.—Elizabeth Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Huddieston, Rodney 1984 Introduction to the grammar of English. Cambridge: Cambridge University Press. Ihalainen, Ossi 1974 On modality of English noun phrases with special reference to genericity. Ph.D. Thesis. University of Wisconsin, Madison. Jackendoff, Ray S. 1972 "Any vs. every", Linguistic Inquiry 3: 119-120. Jespersen, Otto 1914 A Modern English grammar on historical principles. Part II. Syntax. [1933] Heidelberg: Carl Winters Universitätsbuchhandlung. Jud-Schmid, Elisabeth 1962 Der indefinite Agens von Chaucer bis Shakespeare. Die Wörter und Wendungen für "man". (Schweizer Anglistische Arbeiten 39.) Bern: A. Francke AG. Kahlas-Tarkka, Leena 1987 The uses and shades of meaning for every and each in Old English. (Memoires de la Societ6 N6ophilologique de Helsinki 46.) Helsinki: Societe Neophilogique. 1993 "Toward the Modern English dichotomy between every and each", in: Matti Rissanen—Merja Kytö—Minna Palander-Collin (eds.), 201218. 1994 "What does the jungle of Middle English manuscripts tell us? On ME 'every' and 'each' with special reference to their many variants", in: Francisco Fernändez—Miguel Fuster—Juan Jose Calvo (eds.), 305315. Keenan, Edward L.—Bernard Comrie 1977 "Noun phrase accessibility and universal grammar", Linguistic Inquiry 8: 63-99. Kytö, Merja (comp.) 1996 Manual to the diachronic part of the Helsinki Corpus of English Texts: Coding conventions and lists of source texts. (3rd edition.) Helsinki: Department of English, University of Helsinki.

Indefinite pronouns with singular human reference

83

Labov, William 1985 "The several logics of quantification", in: Mary Niepokuj—Mary Van Clay—Vassiliki Nikiforidou—Deborah Feder (eds.), Proceedings of the Eleventh Annual Meeting of the Berkeley Linguistic Society. Berkeley, 175-195. Lakoff, Robin 1969 "Some reasons why there can't be any some — any rules in English", Language 45: 608-615. Lehmann, Christian 1982 Thoughts on grammaticalization. A programmatic sketch. (Arbeiten der Kölner Universalien-Projekts. Vol. I.) Cologne: Universität zu Köln, Institut für Sprachwissenschaft. Lyons, John 1977 Semantics. Vols. I—II. Cambridge: Cambridge University Press. Meier, Hans Heinrich 1953 Der indefinite Agens im Mittelenglischen (1050-1350). Wörter und Wendungen für "man". (Schweizer Anglistische Arbeiten 34.) Bern: A. Francke AG Verlag. Meurman-Solin, Anneli 1993 Variation and change in early Scottish prose. Studies based on the Helsinki Corpus of Older Scots. (Annales Academiae Scientiarum Fennicae. Dissertationes Humanarum Litterarum 65.) Helsinki: Suomalainen Tiedeakatemia. Milid, Louis T. 1990 "A new historical corpus", ICAME Journal 14: 26-39. Nevalainen, Terttu 1991 BUT, ONLY, JUST. Focusing adverbial change in Modern English, 1500-1900. (Mömoires de la Societe Neophilologique de Helsinki 51.) Helsinki: Societe Ndophilologique. Nevalainen, Terttu—Helena Raumolin-Brunberg 1989 "A corpus of Early Modern Standard English in a socio-historical perspective", Neuphilologische Mitteilungen 90/1: 67-111. Pesetsky, David 1987 "fFA-in-situ: Movement and unselective binding", in: Eric J. Reuland —Alice G. B. ter Meulen (eds.), 98-129. Poutsma, Hendrik 1916 A grammar of Late Modern English. Part II. The parts of speech. [1966] Section I B. Pronouns and numerals. Groningen: Noordhoff. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London—New York: Longman.

84

Helena Raumolin-Brunberg

and Leena Kahlas-Tarkka

Raumolin-Brunberg, Helena 1994a "Prototype categories and variation studies", in: Francisco Fernändez —Miguel Fuster—Juan Jos£ Calvo (eds.), 287-303. 1994b "The development of the compound pronouns in -one and -body in Early Modern English", in: Dieter Kastovsky (ed.), Studies in Early Modern English. Berlin: Mouton de Gruyter, 301-324. Raumolin-Brunberg, Helena—Leena Kahlas-Tarkka 1994 " Whoever and anyone who: What was there before? Observations on Biblical English", The New Courant 3: 181-193. Helsinki: Department of English, University of Helsinki. Raumolin-Brunberg, Helena—Terttu Nevalainen 1990 "Dialectal features in a corpus of Early Modern Standard English?", in: Graham Caie et al. (eds.), Proceedings from the Fourth Nordic Conference for English Studies. Copenhagen: Department of English, University of Copenhagen, 119-131. Reuland, Eric J.—Alice G. B. ter Meulen (eds.) 1987 The representation of (in)definiteness. (Current Studies in Linguistics Series 14.) Cambridge, Massachusetts—London, England: The MIT Press. The Revised English Bible, with the Apocrypha 1989 Oxford University Press and Cambridge University Press. Rissanen, Matti 1967 The uses of one in Old and Early Middle English. (Memoires de la Societe Neophilologique de Helsinki 31.) Helsinki: Societe N6ophilologique. 1986 "'Sum' in Old English poetry", in: Phyllis R. Brown—Georgia R. Crampton—Fred C. Robinson (eds.), Modes of interpretation in Old English. Essays in honour to Stanley B. Greenfield. Toronto—Buffalo —London: University of Toronto Press, 197-225. 1987 "Where philology and linguistics meet: Reference, (in)definiteness and Old English sum", in: Dieter Kastovsky—Gero Bauer (eds.) in collaboration with Jacek Fisiak, Luick revisited. Papers read at the Luick-Symposium at Schloss Liechtenstein 1985. Tübingen: Gunter Narr, 295-310. forthc. "Whatever happened to the Middle English indefinite pronouns?" Paper presented at the International Conference of Middle English, Rydzyna, 1994. Rissanen, Matti—Merja Kytö—Minna Palander-Collin (eds.) 1993 Early English in the computer age. Explorations through the Helsinki Corpus. Berlin—New York: Mouton de Gruyter. Safir, Kenneth J. 1987 "What explains the definiteness effect?", in: Eric J. Reuland—Alice G. B. ter Meulen (eds.), 71-97.

Indefinite pronouns with singular human reference

85

Sahlin, Elisabeth 1979 Some and any in spoken and written English. (Acta Universitatis Upsaliensis. Studia Anglistica Upsaliensia 38.) Uppsala: Almqvist— Wiksell. Sankoff, David—Henrietta Cedergren (eds.) 1981 Variation omnibus. (Current Inquiry into Language, Linguistics, and Human Communication 40.) Carbondale—Edmonton: Research Inc. Searle, John R. 1969 Speech acts. An essay in the philosophy of language. Cambridge: Cambridge University Press. Suojanen, Μ. K. 1977 Mikael Agricolan teosten indefiniittipronominit: totalitiivit. Helsinki: Suomalaisen Kirjallisuuden Seura. Tannen, Deborah 1982 "Oral and literate strategies in spoken and written narratives", Language 58: 1-21. Tesch, Felicitas 1990 Die Indefinitpronomina some und any im autentischen englischen Sprachgebrauch und in Lehrwerken. (Tübinger Beiträge zur Lingvistik 345.) Tübingen: Gunter Narr. Tottie, Gunnel 1981 "Negation and discourse strategy in spoken and written English", in: David Sankoff—Henrietta Cedergren (eds.), 271-284. 1983 Much about not and nothing. A study of the variation between analytic and synthetic negation in contemporary American English. (Studier utgivna av Kungl. Humanistiska Vetenskapssamfundet i Lund 1983-1984: 1.) Lund: CWK Gleerup. 1991 "Lexical diffusion in syntactic change: Frequency as a determinant of linguistic conservatism in the development of negation in English", in: Dieter Kastovsky (ed.), Historical English syntax. (Topics in English Linguistics 2.) Berlin—New York: Mouton de Gruyter, 439-467. Vendler, Zeno 1967 "Each and every, any and αΙΓ, in: Linguistics in Philosophy. Ithaca, New York: Cornell University Press, 70-96. Wells, Stanley—Gary Taylor (eds.) 1989 William Shakespeare. The complete works. Electronic Edition. Oxford: Oxford University Press.

The pronominalization of one Matti

Rissanen And a man's life's no more than to say 'one' (Hamlet 5 2 74)

1.

Introduction

The meanings and functions of one are subtler and more varied than those of any other numeral, and their development presents a fascinating story of enrichment and regularization of forms. To give just one example, all one, as in they are all one, means 'united', while the meaning of alone is the opposite, although both expressions are etymologically identical. Let this be an excuse for my life-long love affair with Number One. The complete history of the grammaticalization and pronominalization of one has not yet been written. The earliest thorough survey of the English indefinites by Einenkel (1903-1904; 1906; 1912; 1914) contains a number of useful observations but is necessarily dated and based on unsystematically collected evidence. Mustanoja (1960: 292-305) gives an insightful survey of the Middle English uses and developments of one, and Rissanen (1967) discusses the Early Middle English pronominalization in some detail. Franz makes a few comments on the pronominal one in Shakespeare's English (1939: 317; 329-332; cf. also Bald 1984). Three studies published in Swiss studies of English by Fröhlich (1951), Meier (1953) and Jud-Schmid (1962 [1956]) discuss the expression of the indefinite subject from Old to Modern English and also pay attention to one in this function. There are a number of other works dealing with special aspects of the pronominal one (see the list of references); amongst them, Jespersen's discussion of the prop-word one in his Modern English grammar (II: 245-271) deserves a special mention. The rise of the pronominal uses of one offers an excellent example of grammaticalization. This development begins as early as Old English, as

88

Matti Rissanen

the range of the meanings of the numeral is extended through subjectification.1 The speaker's subjective attitude to the concept of 'oneness' results in the weakening of the indication of contrast with a higher number of items, whether persons or things; attention is focused on the individual or generic characteristics of the referent of one. Gradually the function of one, Old English an, in individualizing or generalizing/generic uses becomes mainly syntactic; we can describe this as fading or bleaching or simply in terms of the emergence of new (grammatical) meanings.2 At this stage of development, we can say that a category shift takes place and the new grammaticalized uses of one are pronominal rather than numerical (see also pp. 100-101, below). The purpose of the present study is to describe the grammaticalization and pronominalization of one in detail from Old to Modern English, with particular focus on Middle and Early Modern English.3 The emergence of the pronominal one will be linked with more general structural changes that took place mainly in Middle English, i.e. the levelling and loss of the inflexional endings and the subsequent demand for new cohesive devices in the language. The evidence is mainly taken from the Helsinki Corpus, although other corpora, concordances, dictionaries and primary texts have also been studied. In the following discussion the pronominal uses are divided into six types, some of them closely related:4 1. Personal-specific (1) One having a Son of two years and a half old, that could but even go about the house, ... observing him one day ... (Hoole, A New Discovery of the Old Art of Teaching Schoole 8) 2. Personal-non-specific (2) It is safer to leave them wholy to nature then to put them into the hands of one forward to tamper [= temper] or that thinkes children are to be curd in ordinary distempers by any thing but diet ... (Locke, Directions Concerning Education 48) 3. Personal-generic (3) he is such an odd fellow, that one may have reason to fear both his folly and knavery. (John Somers, Original Letters (Illustrative of English History) 327)

The pronominalization of one

89

4. Substitutive (4) Now my mind is wholy set upon comeing to see Thee, in order to which pray let my Horse (if I have one) be sent to the Dark House (Henry Oxinden, The Oxinden and Peyton Letters 292) 5. With indefinite pronouns (5) I saw him there, from the 18th to the 25th, every one of the days but one, and that I cannot be positive in. {State Trials, The Trial of Titus Oates IV 82.C1) 6. Prop-word (6) Ka. This will be brave indeed John, but what shall we do with our Ale. Jo. Sell it my sweet one\ (Penny Merriments 117) In the personal-specific use (1), the referent of one is a particular individual, although not necessarily defined in detail in the context. In the personal-non-specific use (2), the referent is non-specific, 'anyone'; this use is typical of non-assertive contexts, such as questions, negations and conditional clauses. The most common personal-generic use (3) of one is the generic subject but one can also be used in other functions in the sentence. Unlike (2), the personal-generic one cannot be post-modified, and it would often be possible to substitute such pronoun as you for it. In types (4), (5) and (6) one is a pro-form. The handy term 'prop-word' for (6) is taken from Jespersen (MEG II: 245ff.). It here signifies the use of one premodified by a descriptive adjective, in contrast to its uses with indefinite pronouns (5) or without a premodifying element (4).

2.

Old English background

A look at the variety and richness of the uses of Old English an (>one) reveals at once that its semantic scope differs sharply from that of all other numerals. In one, the effect of subjectification as a source of semantic change can be easily seen. 'One' is separated from all others; it is unique and singular, but 'oneness' also indicates union, togetherness and belonging. This seemingly paradoxical semantic profile derives from the

90

Matti Rissanen

fact that one not only indicates an opposition to 'two' or 'three', or any other higher number, but also, and very fundamentally, to 'other(s)'.5 The most important uses of Old English an are illustrated in Figure 1. The thick horizontal line depicts the numerical-individualizing continuum. Examples (7)—(13) illustrate this basic distinction in the meanings of an. In examples (7)-(9) the numerical contrast is obvious; in (10)—(11) the attention is called to a particular individual, and in (12)—(13) to any individual amongst a set formed by many of the same kind, rather than to numerical contrast:6 Numerical

contrast

(7)

flotmen swa strange ... J>aet oft on gefeohte an feseö tyne & hwilum laes, hwilum ma, (Wulfstan's Homilies 20 271) 'so strong pirates ... that often in a battle one puts to flight ten, and sometimes fewer, sometimes more'

(8)

He awrat to J>am Romaniscum anne, to Corinthios ii., eac to Galathas anne, to Ephesios anne, to Philipenses anne, to Thesalonicenses twegen ... (The Old Testament Prologue 57) 'He wrote to the Romans one [letter], to the Corinthians two, also to the Galatians one, to the Ephesians one, to the Philippians one, to the Thessalonians two ...'

(9)

Wiglaf maöelode, Wihstanes sunu: Oft sceall eorl monig anes willan wrsec adreogan, swa us geworden is. CBeowulf 3077) 'Wiglaf spoke, Wihstan's son: "Often must many a warrior endure distress on account of the will of one, as has happened to

Individualizing, specific reference (10)

{Da inn eode an J)aes cyninges £>egna J>e his aelmyssan bewiste, and sasde ... (ALlfric's Lives of Saints 26 130) 'then one of the king's servants who supervised his almsgiving entered and said ...'

The pronominalization of one

•H

s

c

.2

2

3

«α 60 c w TS

δ .a >> H •a £ ο „ 'C ö Λ S £ Ζ ο

3

C

•a

s .oo

£

92 (11)

Matti Rissanen beseah öa to Öam [)rim cnihtum and cwasö: Hwilc eower is forliden? öa cwasö heora an se hatte Ardalius: Ic eom forliden. (Apollonius of Tyre 32) 'then he looked at the three boys and said, "which of you is shipwrecked?" Then said one of them, whose name was A. "I am shipwrecked'"

Individualizing, non-specific reference (12)

Nu bidde we [)e Jaaet ]ju geceose {)e cenne of us Jjrym hwilcne {)u wille jje to aöume habban. (Apollonius of Tyre 30) 'Now we ask you to choose one of us three, whichever you wish, for your son-in-law'

(13)

ne faraf) ge ealle heonon ... Ac fare eower an & bringe hine hider (The Old Testament Genesis 42 15-16) 'ye shall not go hence ... One of you should go and fetch him hither'

In (10) and (11) an does not indicate contrast with 'two' or 'many' any more than the PresE sentence one of my sons is a doctor would primarily indicate contrast in most contexts with 'two of my sons'. The case is, mutatis mutandis, the same in (12) and (13), in which the referent is nonspecific. An refers to any of the members of a set, typically in nonassertive contexts, i.e. in questions, negations, exhortations or conditional clauses. The main difference between the strictly numerical and the individualizing an seems to be that in the latter use the primary purpose of an is to put the referent into the focus of the discourse.7 This easily attracts subjective attitudes from the speaker and the way is paved for semantic change through subjectification. A special type of individualizing an is the one in which the numeral is used in contrast to oper (pridda, etc.) 'the other' ('the second', 'the third', etc.), as in (14)—(15). Although the function of an is not quite the same as in (7)—(13), the numerical contrast is here of secondary importance only. (14)

twegen englas ... anne aet |)am heafdon and oöerne aet J)am fotum (West-Saxon Gospels John 20 12) 'two angels ... one at the head, and the other at the feet'

The pronominalization of one (15)

93

Ispania land is Jjryscyte ... An öaera garena lid suöwest ... & ojjer ... & se Jjridda (Alfred's Orosius 24) 'Spain is three-sided ... One of the coasts faces to the south-west ... & and the second ... and the third ...'

The individualizing an is obviously the main source of the Middle English pronominal developments of one, but to understand all the facets of these developments it is necessary to pay attention to the other main uses of an as well. Other basic distinctions in the meanings of an are represented by the lines branching from the numerical-individualizing axis towards unity on the one hand and exclusiveness or isolation on the other. The shades of unity can be defined as identity (16), similarity (17), union (18), unanimity (19), etc. It is of course the context that gives an its finer shades of meaning; the translation 'one and the same' is appropriate in most contexts. (16)

hyt byö embolismus oööe embolismaris, J>ast ys eal an {Byrhtferth's Manual 20) 'it is embolismus or embolismaris, that is all one and the same [thing]'

(17)

Jjonne rasde he Jjaes eadigan weres getingnyssa, Bedan, J)aes asöelan boceres, oööe Rabanes, J)e wel gehende an asmeadun ymbe jjisum craefte (Byrhtferth's Manual 72) Met him read the writings of the blessed man Bede, the noble scholar, or of Raban, who thought almost in the same way about this science'

(18)

meng togaedere ... oj^aet hit sie gemenged, jjaet hit sie an & hasbbe huniges {jicnesse (Lceceboc 67) 'mix ... until it is mixed so that it is uniform and has the thickness of honey'

(19)

gewearö him & Jjam folce on Lindesige anes. J)t hi hine horsian sceoldan (Chronicle MS Ε 1014) 'he and the people at Lindsay were unanimous that they should provide him with horses'

In (20) and (21), an has an exclusive meaning, 'alone', 'only'. On the whole, the concept of exclusiveness or isolation is more closely related to

94

Matti

Rissanen

individualization than the concept of unity. This is due to the implications of uniqueness, singularity, and the general relevance of an emphatic contrast to 'others'. As will be shown below, the exclusive use may have played some albeit minor role in the development of the pronominal uses. (20)

Se mann ana gasö uprihte {Alfred's Boethius 147) 'Only the man walks upright'

(21)

min sweostor me laet ane Jjegnian (The Blickling Homilies 67) 'my sister let me serve alone'

All other uses of an are related to, and can be derived from, these four basic ones. It is only to be expected, in view of the implications of uniqueness and singularity that various degrees of intensification can be easily traced in the uses of Old English an (Rissanen 1967: 189-216; for Middle English intensifying uses, see Mustanoja 1958; 1960: 295-301. For the Presentday English intensifying use of one, see Zandvoort 1940). In most of these instances the meaning of an is associated either with exclusiveness (22)-(23) or unity (24)-(25), as indicated in Figure 1. It seems, however, that in some contexts an simply introduces the referent with particular emphasis, as in the early poetic examples (26)-(29). Expressions of this type are also of relevance for the discussion of the pronominalization of one. While an is typically in determiner position in the uses exemplified by (22)-(25), its link with the following noun is looser in (26)-(29), since the noun can be either the head of the NP or in apposition to an. On the whole, the development of the intensifying uses of one is graphic proof of the effect of subjectification on the meaning of a lexical item. Emphasis on uniqueness: (22)

Sum brojjor is gyt in öisum mynstre mid me wuniende & lifigende, se is se an geornfullesta godcundra gewrita (Gregory's Dialogues 218) 'There is yet a brother in this monastery, living with me, who is the one most eager to study divine scriptures'. [This passage translates the Latin 'scriptum sacra studiosissimus\~\

The pronominalization of one (23)

95

jja wass on gange gifu HroÖgares oft geaehted; Jjaet wass an cyning, asghwass orleahtre, ojijjaet hine yldo benam maegenes wynnum, (Beowulf 1885) 'Then, as they went, Hrothgar's generosity was amply praised; he was one king, blameless in every respect, until death deprived him of the joys of power'

Emphasis on unity: (24)

lob sast Öa sarlice, eal on anre wunde, up on his mixene {/Elfric's Catholic Homilies II Thorpe 452) 'Then Job sat sorrowfully, covered by wounds all over [lit. all in one wound] on his mixen'

(25)

J)a geseah ic öasr Jjone rumestan feld & f)one faegerestan, se wass eall swetnesse anre full growendra blostmena (Bede's Ecclesiastical History 430) 'then I saw there the most spacious field and the most beautiful, which was all full of one sweetness of growing flowers'. This passage translates the Latin 'tantaque flagrantia uernantium flosculorum plenus\

Emphatic introduction of the referent (in poetical texts): (26)

Swa Öa drihtguman dreamum lifdon eadiglice, oööast an ongan fyrene fremman feond on helle. Wass se grimma gasst Grendel haten, (Beowulf 100) 'So the warriors lived joyfully until one began to commit crimes, a hellish fiend. The grim spirit was called Grendel'

(27)

he geheold tela fiftig wintra wass öa frod cyning, eald ejjelweard, oööaet an ongan deorcum nihtum draca ricsian, se öe on heaum hofe hord beweotode, (Beowulf 2210) 'He ruled rightly fifty winters — he was a wise king, old guardian of his people — until one began, on a dark night, a dragon to rule, who in a high dwelling-place guarded a hoard'

96

Matti Rissanen

(28)

Swa se öeodsceaöa J)reo hund wintra heold on hrusan hordaerna sum, eacencrasftig, oööast hyne an abealch mon on mode; mandryhtne baer faeted wasge, {Beowulf 2280) 'So the powerful people's foe occupied a treasure-house for three hundred years, until one angered him in his heart, [a] man [who] brought to his lord an ornamented flagon'

(29)

f>asr wass wop wera wide gehyred, earmlic ylda gedraeg. J)a f>asr an ongann, feasceaft haeleö, folc gadorigean {Andreas 1555) 'there was the lament of men heard widely, miserable tumult of men; then there one began, a destitute man, to collect people'

As can be seen in (22) and (23), an often occurs with superlatives, or in other expressions indicating the uniqueness or exceptional characteristics of the referent. It seems that an in (22) and (23) shares features of the numerical, exclusive and individualizing uses (cf. Figure 1). The reference is to a particular individual, unique and singular, contrasted with all others. When intensification is combined with unity instead of exclusiveness, as in (24) and (25), individualization is not obvious. Number 5 of the pronominal functions of one listed in section 1 above is the use of the word with indefinite pronouns. Even in Old English, an is common with pronouns meaning 'every', 'any', such as (ge/ceg)hwilc, (ge)hwa, celc. The use of the numeral in these contexts differs, however, from its Present-day English use with indefinites (see e.g. Kahlas-Tarkka 1987). An can either precede (30) or follow (31) the pronoun, and can be used either in agreement with the indefinite pronoun or in the genitive plural form (32). Furthermore, it can be used with pronouns which are in the determiner position (33). The meaning of an here seems to contain both individualizing and intensifying features (cf. Latin unusquisque, French chacun, German ein jeder, jeder einzelne, Swedish envar, var och en, varenda, and Finnish joka ainoa). It is obvious, however, that despite its much wider syntactic-semantic scope this Old English usage played a role in the Middle English development of the indefinite pron. + one combinations.

The pronominalization of one

97

(30)

se Mihtiga God ... seöe cenne gehwilcne {)urh his Gast geneosaö (JElfric 's Homilies, ed. Thorpe, I 628) 'Mighty God ... who visits every single person through his Spirit'

(31)

Jjaet ... sie sio soöe gesaslö fie an öissa fifa masg fullice forgifan, foröaemfje on celcum anum hi sint eall. (Alfred's Boelhius 78) 'that may be the true happiness that any one of these five may fully yield, because in each individual one they are all'

(32)

£>eah öe JDIS fyr sy egeslic ... hwaeöre hit after weorca gegearnunge anra gehwylcne demeö & basrneö (Bede's Ecclesiastical History 214) 'although this fire is terrifying ... however it deems and burns everyone according to his desert'

(33)

syx dagum asr fiissum daege gelimpej) syllice tacn asghwylce ane daege (The Blickling Homilies 91) 'six days before this day such a sign will be visible on each one day'

With other indefinite pronouns, an is used only sporadically: (34)

genim eorögeallan grenes seaw, ο{φβ hunan seaw, ο£φε wermodes seaw, swilc £>ara an swa |)u wille (Lceceboc 14) 'Take the green juice of earth-gall or the juice of horehound or wormwood, such a one of those as you wish'

(35)

Nyle he aengum anum ealle gesyllan Gaestes snyttru (Christ 683) 'He does not want to give any single individual all the wisdom of spirit'

(36)

to jjon jsaet he maege aet nehstan jja sawle mid sumre anre leasunge gescraencan & beswican (Gregory's Dialogues 339) 'in order that he might finally destroy and deceive the soul with some particular falseness'

98 (37)

Matti Rissanen hit nis nanum anum men getiohhod, ac is eallum monnum (Alfred's Boethius 112) 'it is not assigned to any individual person but to all people'

Finally, the use in which Old English an approaches the present-day indefinite article is of considerable interest for the discussion of the pronominalization of one. The development of the indefinite article from the individualizing an in the determiner position forms an almost exact parallel to that of the pronominal uses of one. Chronologically, too, the two developments are interrelated; the pronominalization takes place slightly later than the regularization of the indefinite article. In Old English, the pre-modifying an singles out an individual as the topic of the discourse, as in (38) (see Mustanoja 1960: 230; Rissanen 1967: 268-74; 1988; and cf. Givon 1981). The referent of the determiner an can also be less central to the discourse, as in (39), or non-specific (40). The generic use of the article only develops in Middle English. (38)

f)a wass pxr an maeden liegende on paralisyn lange gebrocod; ... mann ferode ^ast maeden to Jsasre foresasdan stowe. (JElfric 's Lives of Saints 26 138) 'Then there was a maiden afflicted by paralysis for a long time ... the maiden was brought to the afore-mentioned place'

(39)

ascraep jjone wyrms of his lice mid anum crocscearde (JElfric's Catholic Homilies II 452, ed. Thorpe) 'he scraped corrupt matter from his body with a crocksherd'

(40)

he riht & rast eallum gesceafitum swa swa good stiora anum scipe (Alfred's Boethius 97) 'he corrects and advises all people as a good helmsman would steer a ship'

In the same way as the pronominal one, the indefinite article is a latecomer in the grammatical system of English. The category of the article is gradually established by the end of the Middle English period. Another use in which the Old English individualizing determiner an is clearly related to the indefinite pronoun one is in expressions of time indicating 'one day', 'one night' etc. (41):

The pronominalization of one (41)

99

He com to him anes nihtes on swefne (St Neot 112) 'He came to him one night in a dream'

In Old English poetry, the position of the individualizing an in relation to the head is free, owing to the demands of rhythm and metre. It can be separated from the following noun, cf. (26}-(29) above, and it can follow the noun in apposition (42)-(44): (42)

Her is faemne, freolecu maeg, Ides Egyptisc, an on gewealde (Genesis 2229) 'Here is a maiden, a noble woman, Egyptian woman, one having power'

(43)

Ic wat hea burh her ane neah Lytle ceastre (Genesis 2519) Ί know a town here in the neighbourhood, a small town'

Also with non-specific an: (44) Ge J)onne sweltaÖ samod mid mannum, Swa ealdormann an gefealleÖ (The Metrical Psalms of the Paris Psalter 81 7) 'Then you will die together with men, as a chieftain falls'. [This passage translates the Vulgate, 'sicut unns de principibus cadetis'.] This poetical freedom in the position of one continues in Middle English and may be taken into account in the discussion of the development of the pronominal uses.

3.

Middle English enrichment

The development of the pronominal forms of one was probably initiated by a profound change in the structure of English which is characterized by the levelling and loss of the inflexional endings. This simplification of the morphological system imposed new demands on the marking of the

100

Matti

Rissanen

head of the noun phrase by lexical means; indefinite pronouns such as one offered an easy way to provide the noun phrase with a head which did not radically affect the meaning of the phrase. Furthermore, morphological losses necessitated the development of other ways of maintaining cohesion in the expression; this must have supported the rise of new forms, particularly in the indefinite and relative pronoun categories. One factor to be taken into consideration in the development of the pronominal uses of one is the disappearance of some important items from the paradigm of indefinite pronouns. One is the indefinite subject man (cf. the corresponding use of man in German and Swedish); the other is the set of wA-indefinites, i.e. hwa, hwcet, hwilc 'someone/thing', 'anyone/thing'. These losses left a gap in the pronominal system which could be filled by one (cf. Rissanen forthcoming). In the discussion of the pronominalization of one we should of course define the point in the development at which a category shift took place and one assumed syntactic-semantic characteristics which made the label 'indefinite pronoun' 8 more appropriate than 'numeral'. We could say that one is used pronominally when the individualizing one — the primary source of the pronoun — loses its connection, explicit or clearly implicit, to a set from which the referent is singled out, when expressions of the type 'She is married to one whom she will love for ever' (cf. ex. 48) begin to occur beside 'She is married to one of my brothers'. This is of course a gradual development and borderline cases can easily be found. The first steps in this direction can be seen in Old English although it seems that, until the 13th century, written language9 avoided the use of one in what could be described as pronominal contexts by the aforementioned criterion. The following Old English examples represent constructions in which the link between an and the set seems to be weak: (45)

Jpaer wearÖ wicingum wi]Derlean agyfen. Gehyrde ic J>aet Eadweard anne sloge swiöe mid his swurde ... jjset him aet fotum feoll fasge cempa {Battle of Maldon 117) 'there the Vikings were given retribution. I heard that Edward smote one [Viking] powerfully with his sword ... so that the doomed warrior fell at his feet'

The pronominalization of one

101

(46)

jja £>a he his oxan raspte. \>a scoc an his heafod (JElfric's Lives of Saints 31 786) 'when he captured his oxen, then one shook its head'

(47)

se ... wer hof upp glassene leohtfatu ... J)a feoll an of his handurn, £>aet hit wearö tobrocen {Gregory's Dialogues 49) 'the ... man lifted up glass lamps ... then one fell from his hands so that it was broken'

The first unambiguous instance of one in a context in which there is no reference to a set occurs in the early 13th century Juliane. Not unexpectedly, the referent of one is a specific person introduced as the focus of discourse (Type 1, p. 88 above): (48)

Ich chulle £>t he wite hit ful wel ... ich am to an iweddet f>t ich chulle treowliche wiöute leas luuien. f)t is unlich him & alle worltliche men. (Juliane 100) Ί wish him to know well that ... I am wedded to one that I should love faithfully without deceit, that is unlike him and all worldly men'

In Early Middle English, too, the use of one with indefinite pronouns (Type 5, above) becomes more common and gradually loses its individualizing-intensifying function, as in (49).10 (49)

Nihe wordes J>er beoö. ah hu ha beoö i-ordret & sunderliche isette. |>e an buue J)e oöre. & euchanes meoster were long to teilen. (Sawles Warde 180) 'There are nine words, but how they have been ordered and separately arranged, one above the other, and each one's function, would take a long time to tell'

Examples of the substitutive pronominal use (Type 4, above) are also early. This is not surprising as there seems to be little semantic difference between the personal-specific one and the anaphoric one referring to a person, as can be seen in (50). The earliest instances of the substitutive one with non-personal referents occur in texts dating from c. 1300 (51).

102

Matti

Rissanen

(50)

3ho wass ec ... \νί{φ weppmann wedded, \νφ|3 an Jjatt wass of hire kinn (The Ormulum 2051) 'She was also ... wedded to a man, to one that was her kinsman'

(51)

He haues a wunde in J)e side With a gleyue, ful unride; And he haues on J)oru his arum (J>er-of is ful mikel harum); And he haues on J>oru his t>e unrideste Jjat men may se. (Havelok 1982-1984) 'He has a wound in the side, made by a spear, most deep, and he has one in his arm, which causes him great harm, and he has one in his thigh, the deepest that has ever been seen'

It seems that the remaining three pronominal uses (Types 2, 3, and 6) develop from those described above. The earliest instance of the propword one, with adjectives, is recorded from the thirteenth century (52), but, as will be shown below (3.6), this use is not established until the Early Modern English period. Two subtypes of the prop-word can be distinguished, non-anaphoric (52) and anaphoric (53). The prop-word construction is of course closely related to the use of one with indefinite pronouns (Type 5); furthermore, the anaphoric prop-word is actually a subtype of the substitutive one, while the non-anaphoric type is semantically close to the personal-specific one. (52)

'Nai,' quoÖ t>e cuddeste an of ham alle (Katherine 822) '"Nay", said the most famous of all of them'

(53)

{)at blisful berde, J>e worthiest an of all in werlde (Cursor Mundi, Gött. MS 10078) 'the joyous bird, the worthiest one of all in the world'

The personal-non-specific one (Type 2) is first attested in fourteenth-century texts (54) and the generic personal one (Type 3) in the following century (55). It seems that these uses both derive from the personalspecific one and the substitutive one, whose referent can also be nonspecific.

The pronominalization of one

103

(54)

Quat! hit clatered in f>e clyff, as hit cleue schulde, As one vpon a gryndelston hade grounden a sy{)e (Gawain and Green Knight 2202) '"What", it clattered against the cliff, as if it would split, as if someone had been sharpening a scythe against a grindstone'

(55)

The feest [= feast] was ful of melodye / One myght haue luste to see suche a feeste (Caxton, The History of Reynard the Fox 51)

Before starting a detailed survey of the Middle English occurrences of the pronominal one, attention may be called to the Early Middle English use in which the function of one approaches the pronominal one but seems to have an additional exclusive implication 'the only person' (cf. Figure 1 above). There are only a few instances in extant texts, all occurring in the Katherine Group. This use was probably one of the side-tracks connected with the earliest development of the pronominal one; it disappears when the grammaticalization of the word is complete: (56)

Hire uostermoder wes an [>t frourede hire. & com to J>e cwalmhus & brohte hire to fode bred (Margarete 68) 'Her fostermother was [the only] one who comforted her and came to the torture-chamber and brought her bread for food'

(57)

f)u art rihtwises weole. & sunfules salue. Jju art an |>t al maht. & nult nawt bute riht. (Juliane 120) 'you are the good fortune of the righteous and the healing of the sinful. You are [the only] one that can do everything and you do not will anything but [what is] right'

The frequencies of occurrence of the different types of the pronominal one in the Middle English section of the Helsinki Corpus can be seen in Table 1.

104

Matti Rissanen

Table 1. Occurrence of the different types of the pronominal one in the Middle English section (1150-1500) of the Helsinki Corpus. Personal-specific Personal-non-specific Personal-generic Substitutive With indefinite pronouns Prop-word

23 3 1 7 125 1

In the following sections (3.1-3.6) the Middle English development of each pronominal type is briefly discussed in the light of the examples derived from the Helsinki Corpus and some other sources. The order of the sections follows the order in which the various pronominal types emerge in the language.

3.1. One with a personal-specific referent As was pointed out above, the gradual weakening of the reference to the set from which an individual was singled out meant a decisive step towards the development of the personal-specific pronominal use of one, cf. (45)—(47) above. It is also possible that the traditionally free placement of Old English an linked with a noun, as evidenced by Old English poetry, as in examples (26)-(29) and (42)-(44) above, contributed to the development of its pronominal uses. Examples of the 'irregular' placement of one in complement position, in a construction with a fronted element, can be found in Early Middle English verse: (58)

Buten witie |)er wes an\ he wes ihaten Ioram. (Layamon C 7745) 'But wizard there was one; he was called Ioram'

(59)

Roberd of caumpedene. £>at hosebonde was on. Vor he was a lute clerc. he ssrof hom echon. (Robert of Gloucester 746) 'Robert of Camden that husbandman was one [= that was a husbandman], because he was a minor cleric, he administered a sacrament of penance to each one of them'

The pronominalization of one

105

When the fronted noun is preceded by an indefinite article, the seemingly pronominal character of one is even more obvious. This poetic use of one was probably originally intensifying, (note its use with wonder in (61)(62), and with an impersonal referent in (61);" later, it is probably used as a handy rhyming device (63)-(64). A similar construction can also be found with the universal (everyone) and negative (none) pronouns; cf. Raumolin-Brunberg—Kahlas-Tarkka (this volume). (60)

a sori woman was she on (Ipomadon A 872; quoted in Mustanoja 1960: 296)

(61)

A wonder bataille it was on (South English Legendary 426)

(62)

A wonder maister he was on. J>at hom so coujje agaste (Robert of Gloucester 405) Ά wondrous master he was one that could so terrify them'

(63)

A seli sin-ful sco was an, And first als a comun wommen (Cursor Mundi 13972) Ά blessed sinner she was one, and first also a common woman'

(64)

1 woot [= know] it by myself fill yore [= long] agon, For in my tyme a servant was I oon (CT A 1814)

Additional instances are given in Langenfelt (1946: 108). Example (48), quoted above, is the earliest instance of the personalspecific one and the only one found in the texts dating from the early part of the 13th century.12 In texts dating from around 1300, a number of instances can be found: (65)

öor woren to öat prisun dragen, On öat Öe kinges kuppe bed, And on öe made Öe kinges bred (Genesis and Exodus 2047-2048) 'there were dragged to prison one that offered the king's cup [= the king's cup-bearer] and one that made the king's bread'

106

Matti Rissanen

(66)

As I com bi an waie, Hof on ich herde saie, Ful modi mon and proud; Wis he wes of lore, (Dame Sirith 2) 'As I was walking on a road, of one I heard tell, a brave man and proud. Wise he was in learning'

(67)

Of on ]?at is so fayr and brijt 13 velud maris Stella Brijter J)an J)e dayis 113t parens et puella, 1c crie to {je, {)ou se to me, Leuedy, preye jji sone for me, {Engl. Lyr. 13th Century 17 1) O f one that is so fair and bright as the star of the sea, brighter than daylight, parent and virgin, I cry for you, see to me, Lady, pray your son for me'

(68)

Starke laddes, stalwor^i men, And on J>e mayster of hem alle, pat was |)e name Griffin Galle. (Havelok 56) 'Strong lads, sturdy men, and one, the master of them all, whose name was G.'

The discourse function of one is particularly clear in the conventional, almost formulaic opening of the poem in (66) and (67). The construction in (66) graphically resembles that quoted in the Old English passages in (26)-(29) above. It is, in fact, possible to analyse one here as a determiner separated from the head {ful modi) mon by the subject and the verb. The order of the elements could be explained by the demands and freedom of poetic expression and the opening-formula character of (h)of on. The shift from the numerical to pronominal use of one is, in any case, unquestionable. In the other examples, too, one introduces a character who is of relevance to the course of events in the narration. As to the syntactic aspects of the pronoun, it is worth noting that one is not necessarily the grammatical or logical subject of the sentence; cf. (48), (66), (67). It is often followed by a relative clause or some other post-modifying element, not unexpectedly in view of the discoursal importance of the referent. A post-modifying element is one of the natural means of giving information concerning the individual in focus, particularly when reference is made by a semantically empty pronoun. Converse-

The pronominalization of one

107

ly, it could be assumed that one of the reasons for the emergence of the pronominal personal-specific one is to mark the antecedent of the relative clause. This change in pronominal usage can be illustrated by the readings of the different Old and New Testament versions included in the Helsinki Corpus: (69) a.

Iohannes him andswarode, ic fullige on wastere. Tomiddes eow stod pe ge ne cunnon. He is J)e aefter me toweard is se wass geworden beforan me. (West-Saxon Gospels John 1 26—27) [Vulgate: Respondit eis Iohannes dicens ego baptizo in aqua medius autem vestrum stetit quem vos no scitis]

b.

Joon answeride to hem, and seide, Y baptise in watir, but in the myddil of 30U hath stonde oon, that 3ε knowen not; he it is, that schal come aftir me, that was maad bifor me, (Wycliffe and Purvey, The New Testament John 1 26-27)

Tyndale and The King James Bible also have the antecedent one. The latter translation reads: 'lohn answered them, saying, I baptize with water, but there standeth one among you, whom ye know not'. (70) a.

Witodlice ne sece ic min wuldor, se is pe secÖ & demÖ. (West-Saxon Gospels John 8 50) [Vulgate: ego autem non quaero gloriam meam est qui quaerit et iudicat]

b.

For Y seke not my glorye; there is he, that sekith, and demeth. (Wycliffe and Purvey, The New Testament John 8 50)

c.

I seke not myne awne prayse: but ther is one that seketh and iudgeth. (Tyndale, The New Testament John 8 50)

The King James Bible: 'there is one that ...'. (71) a.

Ne wene ge f>aet ic eow wrege to Fasder. Se is f>e eow wregö, Moyses on J)one ge gehyhtaÖ. {West-Saxon Gospels John 5 45) [Vulgate: nolite putare quia ego accusaturus sim vos apud Patrem est qui accuset vos Moses in quo vos speratis]

108

Μ alt i Rissanen b.

Nyle 3ε gesse, that Y am to accuse 3011 anentis the fadir; it is Moises that accusith 30U, in whom 3ε hopen. (Wycliffe and Purvey, The New Testament John 5 45)

c.

Doo not thinke that I wyll accuse you to my father. Ther is one that accuseth you, even Moses in whom ye trust. For had ye beleved Moses, ye wold have beleved me: (Tyndale, The New Testament John 5 45)

The King James Bible reads 'there is one that ...'. (72) a.

Öa aetbasrst him sum man, & se hyt sasde Abrame, hu man hys broöor sunu on bendum aweg laedde. {The Old Testament Genesis 14 13) [Vulgate: et ecce unus qui evaserat nuntiavit Abram Hebraeo]

b.

thei token awey Loth and his catel, the sone of the brother of Abram, which Loth dwellide in Sodom. And, lo! oon that ascapide, telde to Abram Ebrew, that dwellide in the valei of Mambre of Amorrei, (Wycliffe, The Old Testament Genesis 14 12-13)

Tyndale and The King James Bible read, 'Then/there came one that ...' In three of the four passages quoted above (69)—(71), the referent of the relative pronoun is known both to the reader and to the hypothetical audience of the direct-speech utterance: God, Christ, or Moses. The Vulgate reading indicates this by using a construction with no expressed antecedent to the relative pronoun. In (69), the Old English rendering follows the Latin construction; in the other two, the antecedent se is added. In the fourth passage (72) the referent is introduced for the first time in the narration and is more indefinite than in the other passages. Vulgate uses the pronominal unus, and Old English the collocation sum man, typical of specific indefinite reference. The translations suggest that Old English does not have (or need) a special pronominal form to introduce indefinite personal-specific referents. Non-expression of the antecedent or the use of the definite antecedent pronoun se are possible reference types in many contexts,14 and the combination of indefinite determiner + noun can be used in others.

The pronominalization of one

109

The situation changes in Middle English. The non-expression of the antecedent is no longer possible in non-generalizing relative clauses, and a personal and demonstrative pronoun antecedent probably came to be considered inappropriate with a referent representing new information in the context, as in (69)—(71). The translation there is he in (70b) of the Wyclifite Bible is worth noting; in one instance the translator prefers to use a cleft sentence, it is Moises that ..., in (71b). This construction makes good sense; it is obvious, however, that the pronominal one as the logical subject of the existential sentence, there is one that ..., used by Tyndale and The King James Bible makes a more refined distinction as to the type of reference. In (69b), even the Wyclifite Bible uses the pronoun one. The personal-specific pronoun one becomes somewhat more frequent in Late Middle English, but the total number of the Helsinki Corpus instances between 1350 and 1500 remains below twenty. Once again, one is often, but not regularly, followed by a relative clause: (73)

'Sun,' he said, 'right nou was an jjat first me fedd, and sythen me kist, And me be-suak [= deceived], J>at i ne wist, Mi benisun [= blessing] now has J)i broiler.' {Cursor Mundi 220)

(74)

As he stode stylle, ... One come with an asse charged with brede [= bread]; (Mannyng, Handlyng Synne 183)

(75)

Blinde Avarice of his lignage [= lineage] For conseil and for cousinage [= kinship], To be withholde ayein largenesse [= against generosity], Hath on, whos name is seid Skarsnesse [niggardliness], The which is kepere of his hous, And is so thurghout averous [= avaricious], That he no good let out of honde; (Gower, Confessio Amantis II 74)

(76)

at Norwyche on scholde haue had large langage to yow and callyd yow traytore, and pyked [= picked] many quarellys to yow. (John Paston, Paston Letters 440)

110 Matti Rissanen (77)

Galien ... beynge sometyme in a felde, wantynge diamoron in one [= a certain patient] J>at hadde sekenesse [= sickness] in f)e {3 rote, and he fonde dyanucum. (The Cyrurgie of Guy de Chauliac 576)

A survey of individual texts and authors confirms the impression of the slow increase in frequency of the personal-specific one in written texts. In King Alisaunder, which dates from the early part of the 14th century (the MS used is dated c. 1400), there are only three instances of one which can be quoted as evidence of the development of the pronoun. In all of them, one introduces a person in the collocation 'one there was, called X': (78)

Barouns weren whilom wijs and gode Jjat £>is art wel vnderstoode, Ac on t>ere was hoten [= called] Neptenabus Wijs in £>is ars [= art] and maliciouse. (Kyng Alisaunder 74)

Also, On pere was, hith Abyron (503), and On pere was pat hete Antyfon (587). In all these instances, however, a weak reference to a set still exists; cf. Barouns in (78). In Chaucer's texts the use of the personal-specific one is varied and the word occurs in a number of syntactic constructions. There are eight instances of the personal-specific one in Canterbury Tales: (79)

'Pardee,' quod oon, 'somwhat of oure metal Yet is ther heere (CT G CY 942)

(80)

And herde oon crien 'water!' as he were wood [= mad], (CT A Ml 3817)

(81)

And to be bounden [= bound] under subjeccioun Of oon, she knoweth nat his condicioun? (CT Β ML 271)

Also CT A KN 2340; A RV 4290; Β ΤΗ 844; Β MK 2128; F SQ 212. In Malory, in the mid-fifteenth century, it seems that one has firmly established itself in this use; there are as many as 34 instances. It is obvious that by the beginning of the sixteenth century, the pronominalization of the personal-specific one was complete.

The pronominalization of one

111

3.2. One with indefinite pronouns As pointed out above, the use of an with indefinite pronouns meaning 'every', 'each', is common in Old English. With other indefinites, such as 'many', 'such', 'some', 'any', examples (34)-(37) above, it occurs sporadically. Its meaning seems to contain individualizing and intensifying features. Syntactically, too, this use of an differs from PresE one with indefinite pronouns; among other things, an is common with pronouns in determiner position and it often occurs in anteposition to the indefinite pronoun (30)-(33). In texts dating from the twelfth and thirteenth century the development from the individualizing-intensifying to pronominal use can be clearly seen. Kahlas-Tarkka's very thorough analysis (1987) of the pronouns meaning 'every', 'each', gives 104 instances of one out of a total of 239 occurrences of these pronouns in a non-determiner position, (KahlasTarkka 1987: 167-168), while one occurs only 36 times with these pronouns in determiner position, out of a total of more than a thousand instances (Kahlas-Tarkka 1987: 163-164). The Old English combination with a genitive plural (32) is not attested in Early Middle English texts. It is obvious that in these contexts one is rapidly grammaticalized. By the fourteenth century at the latest, each one is part of the pronominal system of English, even though forms without one can be found much later (cf. Raumolin-Brunberg—Kahlas-Tarkka, this volume). Another indication of the early pronominalization of one with indefinite pronouns is its growing popularity with many in the thirteenth century and, somewhat later, with such (note, however, the very early occurrence of 'such one' in example (82)). With these pronouns, one is common both in the determiner position and as the head, but the latter type clearly gains ground in the course of the fourteenth century. The earliest ME instances of one as the head occur in The Ormulum, Layamon's Brut, Ancrene Wisse and Hali Meidhad:15 (82)

f>at swillc an shollde mujhenn beon shippend of alle shaffte {The Ormulum 11595) 'that such [a] one would be the Creator of all creation'

(83)

monianne he dude scome [= shamed] (Layamon C 3766)

(84)

& te [= to you] moni an seiÖ. & Jwncheö [= thinketh] J)t... (Hali Meidhad 157)

112 Matti Rissanen As early as the South English Legendary, which dates from c. 1300, there are some fifty instances of many one; about the same number as of many one + noun. The number of the instances of such one is lower (4). In the short extract of the late 13th-century Robert of Gloucester's Chronicle in the Helsinki Corpus, there are no less than 16 instances of the pronominal manion ('many one'), while there are none of many + one preceding a noun. In the course of the fourteenth century, even such one increases in popularity, and both combinations are probably well established in the pronominal system by the end of the century. The pronominal any one and some one occur only sporadically in Middle English (85)-{86), while no instances of no one are recorded for pronouns with personal reference (cf. Raumolin-Brunberg—Kahlas-Tarkka, this volume): (85)

With rent cokrej at f>e kne & his clutte trasches, & his tabarde to-torne & his tote3 oute; 0{)er ani on of alle jjyse (.Purity 42) 'with stockings rent at the knee and his patched rags, and his upper garment torn and his teeth protruding; or any one of all these,'

(86)

we syjen sum oon for to caste out fendis (Wycliffe and Purvey, The New Testament Mark 9 38) 'we saw some one casting out devils in thy name'

3.3. The substitutive one The rise of the substitutive pronoun one was probably to some extent influenced by the personal-specific pronoun. But it is also possible that one source for this usage is the Old English individualizing numeral in contexts of the type 'one ... the other', or 'one ... the second ... the third', (14)—(15) above. This suggestion is supported by the fact that even in early instances the substitutive one can be used with non-human referents (cf. example (51) above). On the other hand, when the referent of this one is personal, as in (87), it is in many cases difficult to say whether one is primarily substitutive or personal-specific. The close relationship between these two types is underlined by the relative clause modifying one in many of the instances quoted below. The link between one and the antecedent noun seems particularly loose in (87), in which ten lines separate an from king.

The pronominalization of one (87)

113

Seoööen [= then] wes his sune king ... Seoööen com an f>e leouede [= lived] wel; he haehte [= was called] Famul-penicel (Layamon C 3511)

In (88), the reference of one is non-specific within the frame of the exhortation in indirect discourse, though the referent may have been familiar to the reader of the Life. (88)

So J)at seinte peter heom radde [= advised], ase ore louerd hadde i-bede [= ordered] For Iudas J>at apostle was: to nimen on in his stude [= place] {South English Legendary 389)

In Chaucer's texts there are a considerable number of instances, but all of them occur in verse, mostly in rhyme (89). In Malory, there are four instances of which two are in comparisons (94)-(95). (89)

... Leve [= dear] brother Osewold, Who hath no wyf, he is no cokewold [= cuckold]. But i sey nat therfore that thou art oon; (CT A 3153)

(90)

"Sir," seyde sir Gawayne, "can thou teche [= lead] us to ony ermyte? [= hermit]" 'Sir, here ys one in a litill mowntayne, (Malory, Morte Darthur 943, see also p. 71)

(91)

We women may wary all ill husbandys; I haue oone, bi Mary that lowsyd [= freed] me of my bandys [= confinement]! (The Wakefield Pageants in the Towneley Cycle 19)

The figures for the substitutive one remain low throughout the Middle English period. This is not surprising; this usage seems to have a relatively low frequency even in present-day written English. The early instances, e.g. (50), indicate, however, that the substitutive use may have become a part of the pronominal system of English roughly at the same time as the personal-specific pronoun.

114 Matti Rissanen In this connection, it is worth quoting the instances in which one, apparently substitutive, is used in comparative phrases introduced by so or as. This use of one may originally have been intensifying, but it probably became a stereotyped idiom in the course of the ME period. Note the use of one as a rhyming word in (92)-(93), and the syntactic frame closely resembling that of the personal/specific one discussed under 3.1 above, exemplified by (60)-(64): (92)

jaat als in warld i tru es nan Sua sinful man as i am an (ιCursor Mundi 28601) 'Also that in the world, I believe, is none so sinful man as I am one'

(93)

So wel arrayed hous as ther was oon Aurelius in his lyf saugh [= saw] nevere noon {CT F 1187)

(94)

Also there lyvith [= lives] nat a bygger knyght than he ys one·, and affitir this he shall do you good servyse. (Malory, Morte Darthur 51)

(95)

Alas, he myght nat endure, the whych was grete pite! So worthy a knyght as he was one, that he sholde be overmacched [= defeated], that of late tyme before he had bene a knyght of kynge Arthurs, (Malory, Morte Darthur 77)

3.4. One with a personal-non-specific (non-generic) referent One with a personal-non-specific referent typically occurs in non-assertive contexts, i.e. questions, negations, exhortations, comparisons and conditional clauses. In Middle and Early Modern English expressions of the type 'If one (= 'someone', 'anyone') comes to see me while I'm away tell him to leave a message', the referent is not a specific individual known to the speaker; cf. 'There's one (= somebody) at the door who wants to speak to you'.

The pronominalization of one

115

The relationship between the non-specific, non-generic pronoun one and the personal-specific one is the same as that between the specific and non-specific uses of the Old English 'pre-article' an mentioned above, cf. examples (38)—(40). It has been argued (e.g. Siisskand 1935; Christophersen 1939; Rissanen 1967) that the non-specific use of the indefinite article develops from the specific one. Thus it is not unlikely that the non-specific pronominal use develops from the personal-specific one, possibly with the support of the substitutive use. The non-specific pronoun is attested from the fourteenth century onwards, cf. (54) above, but the number of instances remains low until the end of the ME period. (96)

But wene ye that every wrecche woot [= knows] The parfite blisse of love? Why, nay, iwys [= indeed]! They wenen [= think] all be love, if oon be hoot. (.TC 2.892)

(97)

it happeth often so, That oon that excesse doth ful yvele [= ill] fare By good counseil kan kepe his frend therfro. {TC 1.626)

(98)

{jere one hajj smylyng semelaunt And behety]) jje to holde cunnaunt, kepe J>e J)an fro losengrye, For feyre spekyng man kan weyl lye. (Mannyng, Handlyng Synne 121) 'when somebody has a smiling appearance and promises you to keep an agreement, beware of flattery, for a fair-speaking man can easily lie'

(99)

Also hyt longef) to couetyse 3yf jx>u hyre one out of seruyse Purgh 3yft or {)urgh procurment, (Mannyng, Handlyng Synne 193) 'Also it belongs to covetuousness if you hire one who is at somebody else's service with gifts or bribery'

116 Matti Rissanen (100)

3yf {)ou receyudyst any what [= anything] Of one J)at hys J)yng forgat [= mislaid]; art falle J)an yn {>e vyce Of, couetyse, )3eft, and auaryce. (Mannyng, Handlyng Synne 193)

JDOU

Also Handlyng Synne pp. 121, 193. The difference between this use and the generic one is further evidenced by the fact that in the passages quoted above one can be postmodified, as in (100).

3.5. Generic one In the instances quoted in the previous section (3.4), such pronouns as someone or anyone could be substituted for one in Present-day English. With generic one, as in (101)—(103) below, this kind of substitution does not seem possible. It is fairly obvious that the generic one develops from the personal-non-specific use. Once again, the development of the generic use of the indefinite article from the non-specific, non-generic use offers an interesting analogy. The generic one develops slowly; there is only one unambiguous instance in the 15th-century samples of the Helsinki Corpus (55). It is mostly used as the subject but can also occur in other syntactic positions, as in (103). (101)

he herde a man say that one was surer in keping his tunge than in moche speking, for in moche langage one may lightly erre (Rivers Diets 57; OED, s.v. one 21)

(102)

every chambre was walled and closed rounde aboute, and yet myghte one goo from one to another (Eneydos 117)

(103)

doo thus fro be [= bee] to be; thus wol thai lede oon to thaire dwellyng place (Palladius on Husbondrie 146)

For further 15th-century examples, see Jud-Schmid (1956: 61, 117). The generic one is much more common in English than the corresponding use of the numeral is in the other Germanic languages, which still

The pronominalization of one

117

have the grammaticalized indefinite subject man. I have shown elsewhere (Rissanen forthcoming) that the obsolescence of the /zw-pronouns and the indefinite, strictly pronominal man (and its weakened form me) precede the establishment of one in this pronominal use.16 It is not unlikely that these developments are connected. The chronology supports this suggestion to a certain extent, the generic one appearing in writing roughly at the same time as man becomes restricted both collocationally and textually. It would be a gross oversimplification, however, to suggest that a oneto-one replacement of the indefinite /zw-forms and man/me by one would have taken place in Middle English. Although allowance must be made for the temporal lag inherent in the evidence of written texts, it seems that the personal/generic one was popularized only in Early Modern English, while the pronominal man/me became obsolete in many dialects by about 1400.17 But there is little doubt that the loss of the afore-mentioned indefinite pronouns, which were highly popular in Old English, created a systemic gap that supported the development of the non-specific and generic pronominal uses of one.

3.6. Prop-word one The theories of the rise of the prop-word, i.e. one after an adjective, are discussed e.g. in Mustanoja (1960: 301-305), Rissanen (1967: 69-70; 7377) and Fischer (1992: 223-225). It is unlikely that the prop-word goes back to the occasional Old English poetic use of an after an adjective as suggested by Luick (1906; 1913) and Einenkel (1914). In all probability, it is related to the personal-specific and substitutive pronominal uses of one. Furthermore, the use of one with indefinite pronouns offers a syntactic model for the prop-word (cf. Jespersen MEG II: 248; Langenfelt 1946: 13 Iff.). In view of the early instances, it is possible that the intensifying capacity of one supported its use with an adjective. It seems, however, that the most important factor contributing to the establishment of the prop-word was the wish to provide the noun phrase with a (pro)nominal head (Fischer 1992: 222). According to Strang (1970: 96; cf. Bald 1984: 143), the use of one with adjectives is a feature peculiar to English. This is strictly speaking not true; at least in colloquial Swedish expressions of the type Han a en toki' en 'He is a foolish one' occur. The prop-word becomes common fairly late in written language, but the Swedish parallel implies that it may have been popular at the spoken level of expression much earlier.

118

Matti Rissanen

The prop-word one can be divided into two types: the anaphoric and the non-anaphoric. In the former type, one refers to a noun occurring earlier in the context, as in (53) above; in the latter, there is no such reference: (52) above. The earliest instances of the prop-word seem to occur in contexts in which intensification is present, with superlatives and after so + adj. The referent is always personal. Only one instance can be found in EME texts dating from the thirteenth century (52); cf. Rissanen (1967: 72). In the fourteenth century, this usage becomes somewhat more common, but, judging from the figures in the Helsinki Corpus, the prop-word is established in the language only in the Early Modern period. In Canterbury Tales, there are two instances, in Malory, none.18 The referent of the prop-word one is in most instances personal; impersonal reference is restricted to the anaphoric type, as in (106), (109)(110), (113ΜΠ4): (104)

I have the mooste stedefast wyf, And eek the mekeste oon that bereth lyf {CT Ε 1552)

(105)

{je kyng of J>at contre was a clerke noble, J>e athelest ane of f>e werd [= world] (Wars of Alexander 40)

(106)

£>e secund day with a saute [= assault] a cite he takis, J>e proddest [= proudest] ane at porrus possessour was euire (Wars of Alexander 3660-3661)

With so + adj.: (107)

What is he? £>at so myjty oon, So myjty & of deef> [= death] so ferd [= afraid] (Cursor Mundi 17994)

(108)

Ne myghte for so goodly oon be lorn (TC 1.373)

(109)

a moche felde; so grete one never he behelde (Mannyng, Handlyng Synne 3268)

(110)

a chapelet so seemly oon (.Romaunt of the Rose 563)

The pronominalization of one

119

In other contexts: Non-anaphoric: (111)

Gat-tothed [= goat-toothed] I was, and that bicam me weel; I hadde the prente [= imprint] ο seinte Venus seel [= stamp]. As help me God! I was a lusty oon, And faire, and riche, and yong, and wel bigon: (CT D 605)

Anaphoric (112)

A fair knyjt a [= he] was to se, A jolif on wyjj oute lak [= blemish] (Ferumbras 251)

(113)

A mare yboned ... And though she be not swyfte, a strong one gete (Palladius on Husbondrie 136,873)

(114)

to William Sennowe oon of my short gownys, a good oon wiche as is convenient for hym (Bury Wills 41; OED, s.v. one 23)

4. Modern English establishment In the Early Modern period, in the 16th and 17th centuries, all the types of pronominal one discussed in the previous section exist side by side. The substitutive and generic uses and the prop-word steadily increase in popularity towards the end of the 17th century, and one comes to be used in more varied ways with indefinite pronouns. On the other hand, there is no increase of the specific and non-specific personal uses in the 17th century. Table 2 shows the developments quite clearly.

120

Matti Rissanen

Table 2. Pronominal uses of one in the Early Modern subperiods in the Helsinki Corpus, in a sample of six plays of Shakespeare and in the early and late texts of the Century Corpus. Figures per 100,000 words in brackets. EModEl 1500-

EModE2 1570-

EModE3 1640-1710

Shakesp. 6 plays

Century of Prose 16801731-1780

Personal-specific 20 19 (10.5) (10.0)

16 (9.4)

23 (14.3)

13 (5.4)

7 (2.7)

Personal-non-specific 4 10 (2.1) (5.3)

8 (4.7)

9 (5.6)

10 (4.2)

19 (7.3)

Personal-generic 2 2 (1.1) (1.1)

30 (17.6)

10 (6.2)

40 (16.7)

61 (23.5)

Substitutive 7 (3.7)

17 (10.0)

11 (6.9)

17 (7.1)

25 (9.6)

With indefinite pronoun 20 37 (9.5) (12.1)

49 (30.0)

17 (10.6)

22 (9.2)

27 (10.4)

Prop-word 4 (2.1)

41 (24.1)

22 (13.8)

44 (18.3)

72 (27.7)

13 (6.8)

10 (5.3)

In addition to the Helsinki Corpus, the number of occurrences in a sixplay sample based on the Oxford Shakespeare Corpus and in the Century of English Corpus (see Milic 1990a-b, and Wells et al. 1989) are given in Table 2, both in absolute figures and converted into occurrences per 100,000 words. 19 The texts of the Century of Prose Corpus are divided into two groups, representing the periods 1680-1730 and 1731-1780.

The pronominalization of one

121

The following discussion concentrates on the gradual obsolescence of the personal-specific use, the more restricted use of the personal-nonspecific, and on the rapid increase in the generic and prop-word use. The Early Modern English post-pronominal uses are not discussed in a separate section; the one-combinations reach a peak frequency in the second half of the seventeenth century, after which the increasing popularity of the -body combinations causes a decrease in frequency (cf. RaumolinBrunberg—Kahlas-Tarkka, this volume).

4.1. One with a personal-specific referent Although the personal-specific use of one exists throughout the Early Modern period, the evidence of the texts shows how its strongly referential, introductory use is gradually becoming obsolete. This development completes the pronominalization of the numeral. This is only to be expected in view of the rapidly increasing popularity of the more grammaticalized uses: the generic, substitutive, post-pronominal and prop-word one. In the first subperiod, 1500-1570 (EModEl), there are ten instances in which one clearly introduces an individual central to the discourse.20 This is often the case when one is followed by a proper name or a noun denoting an occupation, as in (115)—(118). The introductory function of one is also obvious when it appears at the beginning of a new sequence in the discourse, as in (119)—(120). (115)

In this yere and moneth of Nouembre one named the Walshe Clerke apeched [= accused] a knyght called Sir Percyuall Sowdan of Treason (Fabyan, The New Chronicles of England and France 172V.C1)

(116)

One callyd Oconer an yrish lorde toke an horseman prysoner (A Hundred Mery Talys 23)

(117)

So yl vpon ye proclaming therof, one y* was scole master of Poules of chaunce standing by, & comparing ye shortnes of ye time wl the length of ye matter, (More, The History of King Richard III 54)

122

Matti Rissanen

(118)

Than came one that was secrytorye unto the knyghte marshall, who wylled me with spede to departe owte of the sytye, (Mowntayne, Narratives of the Days of the Reformation 210)

(119)

28. The same went to see Hampton court, where thei did hunt, and the same night retourne to Durasme place. 25. One that, by way have mariage, had thought to assemble the peple, and so to make an insurrection in Kent, was taken by the gentlemene of the shier, and afterward punished. (Edward VI, Journal 273)

(120)

Bromley. Bicause you replie so sore upon the principle, I will remember, where one taking the Great Seale of Englande from one Writing, and putting it to another, was adjudged a Traytour in Henry the IVth's tyme, (State Trials, The Trial of Sir Nicholas Throckmorton I 74.C1)

The use of one with proper names or definitions of profession or occupation cannot be found in the post-1570 subperiods (EModE2 and EModE3) of the Helsinki Corpus. There are no examples in the Shakespeare Corpus or the Century Corpus, either. The use of the personal-specific one as an anaphoric pronoun, referring to a person specified earlier in the context, stands in evidence of further grammaticalization and ties this use to the substitutive one:21 (121)

Yn Awguste followinge was the sessyones; unto the which there came my lorde chyffe justyes of Ingland, one that before was recordare of London and callyd Mr. Broke; (Mowntayne, Narratives of the Days of the Reformation 205)

This anaphoric use also occurs in the phrase as one (five instances in the EModEl subperiod of the Helsinki Corpus): (122)

you were specially appointed to go away with the Earle of Devon, as one that would direct all thinges, and giue order to all Men; (State Trials, The Trial of Sir Nicholas Throckmorton I 67.C1)

The pronominalization of one (123)

123

... certaine places of scripture that somewhat seemed to serve his appetite; which when he had pervsed, and thervppon, as one that had never professed the studye of divinity, himself excused to be vnmeete many waies to medele with suche matters, (Roper, The Lyfe of Sir Thomas Moore 32)

The introductory one is not completely absent from the later 16th- and 17th-century samples in the Helsinki Corpus, but the number of the instances is low, and they mostly come from plays or fiction (Shakespeare, Vanbrugh, Armin, Deloney), or from the Bible:22 (124)

Jack comes — lie tell thee, Willy, quoth hee. As I was playing at cardes, one seeing I wonne all I playd for, would needes haue the knaue from mee, which, as very a knaue as hee seeing, would needes beare him knaue for company; (Armin, A Nest of Ninnies 8)

(125)

... on his head an old greasie cap, which had so many holes in it, that his haire started through it: who assoone as hee heard one call for a Porter, made answere straight: here Master, what is it that you would haue caryed? Mary (quoth hee) I would haue this Trünke borne to the spread Eagle at Iuiebridge. (Deloney, The Pleasaunt History of ..lack ofNewberie 75)

(126)

Ford. Master Page, as I am a man, there was one conuay'd out of my house yesterday in this basket: (Shakespeare, The Merry Wiues of Windsor 4 2 133)

(127)

Lov. You have done well to engage a Second, my Dear; for here comes one will apt to call you to an account for your Country-Principles. (Vanbrugh, The Relapse or Virtue in Danger I 36)

(128)

Nurse. What do I din your Ears for? Here's one come will din your Eares for you. (Vanbrugh, The Relapse or Virtue in Danger I 59)

124

(129)

Matti

Rissanen

He told me of another odd Presage that one had of his approaching Death in the Lady Warre, his Mother in Laws house: the Chaplain had dream't that ... (Burnet, The Life and Death of John Earl of Rochester 19)

It seems that the topic-introducing one becomes more and more restricted both textually and contextually: in the Shakespeare sample all instances occur in clauses introduced by the existential there or the local adverb here, cf. (126M128) above.23 In the EModE2 and EModE3 subsections of the Helsinki Corpus, the personal-specific one typically does not introduce the individual as the topic of the discourse. The identity and the individual characteristics of the referent are of minor importance. One often occurs in a non-subject position; as the subject it is typically linked with verbs of saying (132)(133). The substitution of a pronoun of the sowe-series for one would be possible in many of these instances: (130)

after, talked with one that Came to se me, and then [I] went to praier and examination: (Hoby, Diary of Lady Margaret Hoby 80)

(131)

I am charged to carry a pistoll: I had none when I went owte. when I came into London I sawe one havinge a pistoli. I desired it of him and had it. But it had never a stone nor cold ... (The Arraignment of the Earles of Essex and Southampton 11)

(132)

A hurly burly went through the house, and one comes and whispers the lady with the newes: she [= the lady] tels Sir William how Jack Oates had stolen the pie. (Armin, A Nest of Ninnies 14)

(133)

Whereat all the people in the streete fell into such a laughing, that the teares ran downe againe. At last one made answer, saying: goodfellow she dwels in the North brooke street, (Deloney, The Pleasaunt History of ..lack of Newberie 81)

Table 2 shows that the two sections of the Century Corpus (1680-1730) and (1730-1780) show a significant decrease in the frequency of the personal-specific one (13 and 7 instances). The use varies, both as concerns

The pronominalization of one

125

linguistic and textual environment, but one occurs in subject position only three times, all instances coming from the earlier part of the Century Corpus. (134)

We find Ben Jonson using them in his "Magnetic Lady", where one comes out from dinner, and relates the quarrels and disorders of it, (ADryden Essay[1668]016/071-P03)

(135)

... boldly breathe their pretensions in the king's ears, and that to such a degree of freedom that one who was but an ensign in England durst pull him by the sleeve because he passed without taking notice of him; (Stevens.Fil Travel [1689]0023/115-PO)

(136)

One, who either through knavery or folly, or both, always cried, yet a little sleep, a little slumber, (Atterby.Doc Polemics 4K91 [1714J0040/076-P0)

In other respects, the Century Corpus instances offer no surprises: the introductory function of one is not conspicuous in the instances in which one occurs in a non-subject position. In eight out of the 21 instances, one is used anaphorically: (137)

he was slow of speech, mightily opinionated of his own wit, one who delighted in hard words, and admired himself for his discourses, (Anonymi.Doc Fiction 3F36[1702]0001/149-P1)

(138)

he excelled in most of these to perfection. It was not difficult, or possible, but that one who was endowed with so many perfections, joined to an affable and courteous temper, should be insensible to love. (Costeker.Doc Fiction 6F37[ 1731 ]0003/028Pl)

In conclusion, the Early Modern English history of the personal-specific one shows how the grammaticalization of the pronoun goes hand-in-hand with the loss of its individualizing, topic-introducing capacity. The decrease of this pronominal use also coincides nicely with the increase of the items of the some-series with personal-specific reference described by Raumolin-Brunberg—Kahlas-Tarkka in this volume. The latest instances of the personal-specific one are recorded from the 19th century by the OED (s.v. one V 20c), but these instances (with the exception of the ones

126

Matti Rissanen

in which One refers to God), seem to be borderline cases between the specific and non-specific uses; cf. (141)-{143), below.24 (139)

The first time that I have heard of one with a beard avouch himself a coward. (Scott, Betrothed iii)

(140)

Mr Carlyle is as one who does not hear the question. (Morley, Carlyle in Crit. Misc. Ser. I (1878) 198)

4.2. One with a non-specific personal referent Judging by the figures in Table 2, this use seems to remain in current use somewhat longer than the personal-specific one. In Early Modern texts the non-specific one is typically used in contexts in which the referent, although not a particular individual, is equipped with individualizing characteristics, often with a remarkably high degree of specificity. As borderline cases we may regard the expressions in which one has a specific referent from the pragmatic point of view, but the reference is expressed obliquely, in more generalizing terms. A comparative element is often involved, as in (142) and (143). Not surprisingly, instances of this kind can be found in plays and other contexts with first-person discourse: (141)

It cannot be imagined, that I would venture the hazard of my own Life, and the Ruin both of myself and Children, to conceal one that I never knew in my Life, as I did not know Mr. Nelthorp, (State Trials, The Trial of the Lady Alice Lisle IV 123. CI)

(142)

it makes me almost readie to wrangle with mine owne honesty: lie entertaine my selfe like one that I am not acquainted withall: (Shakespeare, The Merry Wiues of Windsor 2 1 82)

(143)

I am colder towards her in my charity then I should be to one so painful as he and she have been, (Pepys, The Diary of Samuel Pepys VII 417)

The referent of one is non-specific in non-assertive expressions, that is in conditional clauses, questions, negations and exhortations. Yet the defini-

The pronominalization of one

127

tion of the characteristics of the referent clearly distinguishes this use from the generic use of one", the pronoun somebody/-one could be substituted for one in most of these instances: (144)

if a gouernour ... do gyue sentence agayne one that hath transgressed the lawes, or punissheth hym ... Beneuolence therby is nat any thing perisshed; (Elyot, The Boke Named the Gouernour 150)

(145)

I am a stranger both vnto the Citie, And to her carnall stricktnesse. 1 Prom. Good, Your will Sir? All. Pray tell me where one dwells that kils this Lent. 1 Prom. How kils? Come hither Dicke, A Bird, a Bird. (Middleton, A Chaste Maid in Cheapside 22)

(146)

Sence the worlde begänne was it not hearde that eny man opened the eyes of one that was borne blynd. (Tyndale, The New Testament John 9 20)

(147)

It is good (sayes the knight) ... and one goe call Jack Oates hether, (Armin, A Nest of Ninnies 10)

Also in contexts in which the action is iterative and one may consequently have more than one referent: (148)

at mydnyghte alwaye whan they searched the prysonars' iornys [= irons] than one shold come and knock at my dore and aske me yf I were withyn. (Mowntayne, Narratives of the Days of the Reformation 203)

As evidenced by the Century Corpus, this use of one is relatively popular in 18th century texts, not only in non-assertive expressions but also in assertions. The fairly frequent occurrence of one in the Century Corpus may well be due to the quality of the texts sampled: the selection consists of a number of philosophical and other scholarly texts and essay-type writings. But this shift from specific to non-specific use of one is also in accordance with the grammaticalization trend of the pronominal one and it shows convincingly that the non-generic pronoun one was not entirely superseded by such pronouns as some/anyone, etc., until the nineteenth

128

Matti Rissanen

century. The descriptive elements appended to one, mostly through postmodification, clearly distinguish this usage from the generic one discussed in the next section. The following examples illustrate the 18th century use of non-specific one: (149)

almost wholly within the compass of the three concords: and has no difficulty in it at all; especially to one already acquainted with the Latin tongue. (Clarke.Doc Educatio 5D19[1731]0022/ 056-P0)

(150)

Suppose one who had so vitiated his palate as to take more pleasure in the taste of opium than in that of butter or honey, to ... (ABurke Taste[1759]060/099-P06)

(151)

One who has indulged this despicable state of mind becomes a mere carrier of other people's sentiments, which he delivers ... (ABoswell Hyp21 [ 1779]065/065-P 18)

4.3. Generic one The step from personal-specific or non-specific to the generic use of one is not long. The borderline cases include usages in which the referent of one is, pragmatically, the speaker (or some other person involved in the discourse situation), but the reference is made, for some reason or other, in a more generalizing vein; cf. the distinction between the specific and non-specific uses of one in the previous section, examples (141)—(143). The following passage illustrates the point: (152)

Pray one word with you; prithee Nurse don't stand ripping up Old Stories, to make one asham'd before one's Love; do you think such a fine proper Gentleman as he, cares for a fiddlecome Tale of a draggle-tail'd Girl; if you have a mind to make him have a good Opinion of a Woman, don't tell him what one did then, tell him what one can do now. (Vanbrugh, The Relapse or Virtue in Danger I 63)

The Corpus evidence implies that although the first examples of the generic use of one go back to the 15th century, the usage only becomes common in the seventeenth. In the sixteenth- and early seventeenth-cen-

The pronominalization of one

129

tury samples of the Helsinki Corpus the number of examples is remarkably low (two unambiguous instances in each), while in the 1640-1710 subsection (EModE3), there are more than thirty. It is of interest, however, that in the Oxford Shakespeare Corpus the generic one is frequent: in the six-play sample there are ten instances, and in the entire corpus, the figure is around 50. The Century Corpus confirms the rapid growth of this usage of one from the late 17th century on; the total number of instances is close to a hundred. The fact that one is more common in Shakespeare's drama texts than in the other text types dating from the same period implies that the generic one goes back to a subjectified use which could be described as a tendency towards oblique reference either to the speaker or to another participant in the discourse situation. If this is the case, we have here an interesting instance of cyclic subjectification: first, the Old English numeral one (an) is given an individualizing meaning through subjectification, as described in section 1 above; after the Middle English pronominal ization and the Modern English 'desubjectification' of the individualizing one, a new pragmatic-subjectified use emerges in the seventeenth century; this use results in the popularization of the generic one even in non-subjectified contexts.25 The following instances quoted from Shakespeare show a varying degree of involvement and subjective attitudes: (153)

What then? What rests? Try what repentance can. What can it not? Yet what can it when one cannot repent? (.Hamlet 3 3 66)

(154)

Thy tongue, thy face, thy limbs, actions, and spirit Do give thee five-fold blazon. Not too fast. Soft, soft Unless the master were the man. How now? Even so quickly may one catch the plague? Methinks I feel this youth's perfections With an invisible and subtle stealth To creep in at mine eyes. (Twelfth Night 1 5 286)

130

Matti Rissanen

(155)

I'll devise some honest slanders To stain my cousin with. One doth not know How much an ill word may empoison liking. {Much Ado 3 1 84-86)

(156)

Swagger, swear, and discourse fustian with one's own shadow? (iOthello 2 3 274-275)

(157)

If I have horns to make one mad, let the proverb go with me: I'll be horn-mad. (Merry Wives of Windsor 3 5 139)

(158)

What valour were it when a cur doth grin For one to thrust his hand between his teeth (3Henry VI 1 4 57-58)

As early as Shakespeare, the generic one can occur in various syntactic positions, as in (156-158).26 Unfortunately, space does not allow a discussion of the interesting question of the variation between one and the personal pronoun in the non-subject positions following the generic subject one.

4.4. Substitutive one The substitutive use of one is established in Early Modern English. As pointed out in the discussion of the Middle English pronominal uses of one, with personal referents the distinction between the substitutive and the personal-specific and non-specific uses is often difficult to make, especially in cases in which one is used anaphorically but has no clearly definable NP as its antecedent. The following passages can be regarded as borderline cases: With specific referent: (159)

the master of the house his Beauer being his blue Bonnet, one that will weare no other shirts, but of the Flaxe that growes on his owne ground, and of his wiues, daughters, or seruants spinning; (John Taylor, Pennyles Pilgrimage 139.C1)

The pronominalization of one

131

With non-specific referent: (160)

He who cites the opinion, or remark, or saying of another, whether it has been written or spoken, is certainly one who quotes; (ABoswell HYP21[1779]059/032-P15)

With personal referents, attention may be called to the pattern in which the substitutive one, mostly post-modified, follows the conjunction and (10 instances in the Helsinki Corpus): (161)

Let the kyng therfore chose a wyfe whych feareth god, let hym not seke a proude, wanton and one ful of riche treasures and worldelye pompe. (Latimer, Sermons 37)

(162)

I confesse I am but a greene housewife, and one that hath but small tryall in the world, (Deloney, The Pleasaunt History of... lack of Newberie 71)

In the Helsinki Corpus and the Shakespeare Corpus, the substitutive one is somewhat more frequent with personal referents than with the impersonal ones (28 personal as against 19 impersonal). In the Century Corpus, which represents a later stage of development, non-personal referents are in a clear majority (16 personal as against 27 impersonal). This can be regarded as further evidence for the advancement of the grammatical ization process: the pronominal one is losing its connection with the topicintroducing personal reference.27

4.5. Prop-word one The increase of the prop-word follows the same pattern as that of the generic one. It seems that subjectification played a role in the popularization of this use of one, in the same way as it may have done with the generic pronoun. In the EModEl subsection (1500-1570) of the Helsinki Corpus there are only four instances of the prop-word, all of the anaphoric type. (163)

We haue nowe a prety litle shilling, in dede a verye pretye one. (Latimer, Sermons 35)

132

Matti

Rissanen

(164)

... I pray you take the Paynes, and reade further, that hereafter whatsoeuer become of me, my Words be not peruerted and abused to the hurt of som others, and especially against the great Personages, of whome I haue bin sundry times examined; for I perceiue the Net was not cast only for little Fishes, but for the great ones, {State Trials, The Trial of Sir Nicholas Throckmorton I 70.CI)

(165)

and at Ledyne-hall was nodur pagant hangyd with cloth of gold, ... at the conduyt in Cornhyll a-nodur of the sete; and at the grett condutt a-nodur goodly on, and the standard pentyd and gyldyd, and the Crosse pentyd; and at the lytyll conduyt a goodly pagant; (Machyn, The Diary of Henry Machyn 45)

(166)

7. Sir Thomas Paulmer came to th'erl Warwike, since that time duke of Northumberland, to deliver him his cheine, being a very faire one, for every lince wayed ane ounce, (Edward VI, Journal 353)

All these instances occur in contexts with first-person oriented discourse. Three of the four adjectives to which one is appended contain subjective evaluation (pretty, goodly, fair). Because of the low number of instances no great significance can be given to these features, but they support the line of development shown by the occurrences of the prop-word in the later subsections of the Helsinki Corpus. In the next subsection of the Helsinki Corpus, EModE2 (1570-1640), the number of instances is still low (10); it is noteworthy that five of these occur either in comedy (Middleton) or in Deloney's and Armin's light-weight humorous fiction. The other instances, too, occur in firstperson oriented discourse: (167)

T.I. Of some hälfe ounce, stand faire and comely, with the Sparke of a Diamond. Sir 'twere pittie to lose the least grace. Yell. Pray let's see it, indeed Sir 'tis a pure one. (Middleton, A Chaste Maid in Cheapside 7)

(168)

lie [= I'll] foe behind the arras, and study three questions, and come againe; ... I, quoth the king: they will be wise ones, no doubt. (Armin, A Nest of Ninnies 45)

The pronominalization of one

133

(169)

my gold chaine and bracelets are none of the worst sort, and I will shew them you, ... and therewithall shee stept into her chamber and fetcht them foorth. When her Gossip saw them, she sayd: Now beshrew my fingers but these are fayre ones indeede. And when doe you meane to weare them Gossip? At Whitsontide (quoth shee) if God spare mee life. (Deloney, The Pleasaunt History of...lack of Newberie 71)

(170)

I am fainte to play the Sirgion, with my Tutors finger which was a pitteful one, (William Paston II, Correspondence of Lady Katherine Paston 64)

(171)

there be imperfect fellicitie of a frayle good, It can not be doubted but that ther is a solide & parfet one. (Elizabeth I, Queen Elizabeth's Englishings of Boethius, Plutarch, &c 62)

(172)

But your little ones, which yee said should be a pray, them will I bring in, and they shall know the land which ye haue despised. {King James Bible, Numbers 14 31)

(Also Deloney 71, 79; Armin 11; Coverte 38). Once again, the prop-word seems to be favoured with adjectives containing subjective evaluation. The Shakespeare sample confirms the suggestion that, like the generic one, the prop-word is first popularized at the level of first-person oriented discourse in contexts characterized by involvement and subjective attitudes. There are no fewer than 22 instances of the prop-word; in proportion, the number is almost three times as high as in the EModE2 subsection of the Helsinki Corpus. The evidence provided by the Helsinki Corpus shows an explosive increase in the popularity of the prop-word in the latter half of the 17th century. There are forty-one instances of the prop-word in the EModE3 section of the Helsinki Corpus, in all text types, both in involved and purely factual contexts. There are numerous instances of the plural form ones, although none of the genitive one's. Instances of the anaphoric type clearly outnumber those of the non-anaphoric one. The last-mentioned type (173)—(177), contains both fixed collocations, such as the young ones, the little ones, the great ones, and affective expressions such as pretty one, sweet one:

134

Matti Rissanen

(173)

(for an Otter does so) sometimes five or six, or ten miles in a night to catch for her young ones, or to glut herself with Fish, (Walton, The Compleat Angler 211)

(174)

I have known some that... have taught little ones to pronounce all the letters, and to spell pretty well, (Hoole, A New Discovery of the Old Art of Teaching Schoole 5)

(175)

But it is a fond conceit in many great ones, and pernicious in the end, to cease from no violence till they have attain'd the utmost of thir ambitions (Milton, The History of Britain X 280)

(176)

Ka. This will be brave indeed John, but what shall we do with our Ale. Jo. Sell it my sweet one·, (Penny Merriments 117)

(177)

Ka. Must I dance too. Jo. Ay pretty one, every body will strive to dance with the Bride. (Penny Merriments 118)

Of the three versions of Boethius' De Consolatione Philosophiae included in the Early Modern samples of the Helsinki Corpus, Preston's translation (1695) contains five instances of the prop-word: (178)

... they seem to give me some imperfect Good; but the true and perfect one this can by no means afford. Bo. I agree with you. Ph. Seeing then thou knowest already which is the true Happiness, and which the false one, it remains thou shouldst be informed from what Fountain to derive that true one. (Preston, Boethius 128)

(179)

if ... there be a certain imperfect Felicity, a fading Good, there must also be, without doubt, a sollid and perfect one. (Preston, Boethius 134; cf. example (171) above)

Also Preston, Boethius 180. In Queen Elizabeth's translation (1593), only one of these passages is translated with a prop-word, example (171) above, and in the sample from Colville's rendering (1553), there is not a single example of the

The pronominalization of one

135

prop-word. Colville translates the passage quoted in (171) as som blisfiiltiesse that is sad stedefast, and parfyt. The 18th-century section of the Century Corpus shows a steady growth in the popularity of the prop-word. This is not surprising in view of the general establishment of the syntactic structure of English in the eighteenth century: the prop-word one effectively restricted the use of the adjectival head of noun phrases to the environments in which it can be used today.

5. Concluding remarks In this paper, I have tried to describe semantic developments of a closedclass item involving 'weak' grammaticalization (see the Introduction to this volume). One has offered a particularly rewarding topic for tracing gradual grammaticalization, as the developments have mostly taken place fairly late, in Middle and Early Modern English, and there is ample textual evidence of this process. It is also obvious that, throughout its history, one has been an easy target for subjectification, a phenomenon which results in changes of meaning due to personal affect or speaker-orientedness. It seems that subjectification played a role in the Old English developments of the individualizing uses of one and, as late as Modern English, in the development of generic one. The pronominalization of one can be linked with more general structural-typological developments of English, particularly with the new demands set on cohesion after the levelling and loss of the inflexional endings of the language and the consequent collapse of the Old English case system. The triggering of the pronominalization resulted in a rich system of six easily distinguishable usages. In the course of the Modern English period this variety was, however, simplified, and two of the functions of the pronominal one were largely taken over by compound forms someone, anyone. The four remaining types, i.e., generic and substitutive uses, the prop-word, and the post-pronominal use are easily disambiguated and there is little risk of syntactic homonymy. All in all, the history of the pronominal one offers us an excellent test-tube case of the enrichment and regularization of the system of grammatical lexis in English.

136

Matti Rissanen

Notes 1.

2.

3.

4.

5.

6.

7.

The role played by subjectification in the process of grammaticalization is discussed e.g. in Traugott (1989), Traugott—König (1991), Nevalainen (1991), and Hopper—Traugott (1993). Hopper—Traugott (1993: 88) point out that, at the early stages of grammaticalization, there is a redistribution or shift, rather than loss, of meaning. For this reason, they avoid the term 'bleaching' in their study. The determiner uses of one, or the development of the indefinite article, will not be discussed in this context. The process of grammaticalization resulting in the establishment of the article category is discussed, for instance, in Süsskand (1935); Christophersen (1939); Rissanen (1967). The types are given in the order of structural and semantic relatedness; first those without a premodifying pronoun or adjective (1-4), and then those with a premodifying element (5-6). Types 1-3 have a personal referent by definition. The chronological order of the appearance of these types differs from the order of this list, as will be shown below, under 3.1-3.6. The examples illustrating the six types date from the period 1640-1710 (EModE3 in the Helsinki Corpus), which shows perhaps the richest variety of the pronominal uses of one in the history of English. The examples in this chapter are drawn from the samples of the Helsinki Corpus (for the abbreviated titles, see the Appendix at the end of this volume), from the Helsinki Corpus texts outside the samples, from the Shakespeare Corpus (Wells et al. 1989), from the Century of Prose Corpus (Milte 1990) and from other primary texts (see References). In fact, it is a matter of some uncertainty which use of the Indo-European lowest cardinal should be regarded as its 'original meaning'. This question has been thoroughly discussed by Gonda (1953). For a brief summary see Rissanen (1967: 1-3). As this survey deals with head-word uses of one, the examples quoted below will mostly be given of an as the head of the noun phrase. The same meanings can be seen with the pre- or post-modifier uses of an. This is, of course, a simplification. As in most semantically-based categories there is a large number of border-line cases. Is an individualizing, or does it primarily indicate numerical contrast with reference to ealle in the following passage?

The pronominalization of one

8. 9.

10.

11.

12.

13.

14. 15.

137

Sceotend swaefon, £>a {)aet hornreced healdan scoldon, Ealle buton anum. ... ac he [Beowulf] waeccende wrajjum on andan bad bolgenmod beadwa geringes. (Beowulf 703-709) 'The warriors were asleep, those who should have guarded the gabled house, all except one. ... But he, angry and enraged, was awake and waited for the outcome of the battle'. For the use of the term 'indefinite pronoun', see Raumolin-Brunberg— Kahlas-Tarkka, this volume. It is not impossible, however, that this kind of use existed in spoken Old English. There are indefinite uses of yksi 'one' and joku 'some' in Finnish, which are common in colloquial expression although not accepted in the written standard (cf. Rissanen 1988) and Nykysuomen sanakirja ('Dictionary of Modern Finnish') s.v. yksi II 2 b-c; IV. The development of the compound pronouns of the type everyone, anyone, someone, no-one, with personal referents, is discussed in detail by Raumolin-Brunberg—Kahlas-Tarkka in this volume. This development is only briefly dealt with in the present survey. Cf. the unique intensifying use of one after wunder in Layamon's Brut C, as in pe see wes wunder ane wod 'The sea was wondrously raging' (5990). This usage is discussed in Rissanen (1967: 206-208). Cf. also RaumolinBrunberg—Kahlas-Tarkka, this volume, sections 3.5.2 and 3.5.4. According to the Middle English Dictionary, the extant manuscripts date from c. 1225 and the original perhaps from c. 1200. For other early examples, see Rissanen (1967: 62-63); OED s.v. one 20; MED s.v. on pron. 1. It must be pointed out, however, that the MED classification of the uses of one is unsystematic. The beginning of this poem is either anacoluthic or corrupt, but the pronominal character of one in the conventional opening is unquestionable. (Cf. example (66)). Cf. Mitchell (1985: 193-195). Mitchell does not give instances of the nonexpression of the antecedent with a specific referent. Note also the single pre-1300 instance of which one: wulche weis heo mihten don. & heu heo mihten taken on: & wulc an of Rissen children, heo mihten habben to kinge, (Layamon C 6484; Ο woche of) 'what way they might do, and how they might take on, and which one of these children they might make for king'.

138 16.

Matti

Rissanen

Both the indefinite who and me occur in the following 12th-century passage: 3if hwa is swa sunfiil and mid deofle biuon {)et nulle ... his scrift ihalden. Jjenne segge ic eou to soöe J>et nis hit nan Jjerf Jjet me her on J)isse Hue for his saule bidde pater noster. (Lambeth Homilies 9) 'If anyone is so sinful and possessed by the Devil that (he) does not want to do his penance, I tell you truly that there is no need to say pater noster for his soul in this life' The disappearance of the pronominal man seems to be due to its early weakening into me in positions in which it corresponds with the German and Swedish uses of the indefinite subject man. The obsolescence of the indefinite use of the wh- pronouns is probably linked with the introduction of the non-generalizing relative use of which, whom, whose in Early Middle English. Increased syntactic homonymy would be particularly dysfunctional at the time of the rapid loss of the inflexional endings. The forms man, men, continue their existence as indefinite subjects in the meaning 'human being(s)' until Early Modern English. Cf. the use of people in this meaning, from Middle English on (Mustanoja 1960: 223).

17.

Side by side with one, various other devices to indicate indefinite subject developed, such as men, people, compound pronouns with some and any, the indefinite use of the personal pronouns you, they, etc. (cf. RaumolinBrunberg—Kahlas-Tarkka, in this volume; Jud-Schmid 1956: passim). JudSchmid (1956: 95) also points out that the use of the passive is conspicuously frequent in 15th-century texts.

18.

The instances given below are quoted from Rissanen (1967: 71-72; MED s.v. on lc; OED s.v. one 23-24). 19. The Shakespeare plays sampled are Julius Caesar, Twelfth Night, Hamlet, 1 Henry IV, Macbeth, and The Tempest. The total number of words in this sample is c. 162,000. The sizes of the two parts of the Century of Prose Corpus are c. 240,000 and c. 260,000 words, respectively. 20. In addition to those quoted under (115-120), they are Tyndale Gen. 14,13; John 1,26; 5,45; 8,50. 21. This use cannot, however, be properly called substitutive as there is no one noun or pronoun earlier in the context that could be said to be replaced by one. 22. In addition to the passages quoted below, Hoole 8, and the verses from the King James Bible corresponding to the Tyndale passages referred to in note 20.

23.

It is perhaps of some significance that in The Tempest, which is regarded as the last play written by Shakespeare, there are no instances of the personal-specific one.

The pronominalization of one 24.

139

Cf., however, the use of one in 19th-century letters: She is now happily recovering, but is terribly weak and pale. I know you laugh at my enthusiasm about her; but it is something for me, too, to have one who loves me for my own sake, not as some do, for my head, and who gives me, what I have never known — a home. (Green, Letters 87: 8)

This example is taken from David Denison's Corpus of Late Modern English Prose; see Denison (1994). 25. As a curiosity, it might be pointed out that the colloquial use of the expression (I for) Number One, whatever its origin, may be an instance of still another cycle in the subjectification phenomenon of one. This usage is recorded by the OED from the 18th century on. Cf. OED s.v. number sb. 5b; s.v. one 6d. 26. The few occurrences of this one in the 1570-1640 subsections of the Helsinki Corpus are all in subject position. 27. However, the potential influence of the types of the texts selected in each corpus should not be forgotten. It is possible that the more abstract topics of the samples of the Century Corpus affect the personal/impersonal distribution of the referents.

References Texts quoted from outside the corpora JElfric's Catholic Homilies, ed. Thorpe 1844-46 The Homilies of the Anglo-Saxon Church I-Il. Ed. Benjamin Thorpe. London. Battle of Maldon 1942 The Battle of Maldon. In: The Anglo-Saxon Minor Poems. Ed. Elliott van Kirk Dobbie. (ASPR VI.) New York. CT = Canterbury Tales 1957 In: The Complete Works of Geoffrey Chaucer. Ed. F. N. Robinson. Eneydos 1890 Caxton, William. Eneydos. (Early English Text Society, Extra Series 57.) Eds. W. T. Culley—F. J. Furnivall. English Lyrics 13th Century 1932 English Lyrics of the Xlllth Century. Ed. Carleton Brown. Oxford.

140

Matti Rissanen

Ferumbras 1879

Sir Ferumbras. (Early English Text Society, Extra Series 34.) Ed. S. J. Herrtage. Gawain and Green Knight 1925 Sir Gawain and the Green Knight. Ed. J. R. R. Tolkien—Ε. V. Gordon. Oxford. Genesis and Exodus 1873 The Story of Genesis and Exodus. (Early English Text Society, Original Series 7.) Ed. Richard Morris. (Second revised edition.) Palladius on Husbondrie 1873 Palladius on Husbondrie. (Early English Text Society, Original Series 52.) Ed. B. Lodge. Purity 1864 Purity. In: Early English Alliterative Poems. (Early English Text Society, Original Series 1.) Ed. Richard Morris. Robinson, F. N. (ed.) 1957 The Complete Works of Geoffrey Chaucer. (Second edition.) Cambridge, Mass. Romaunt of the Rose 1957 The Romaunt of the Rose. In: The Complete Works of Geoffrey Chaucer. Ed. F. N. Robinson. (Second edition.) Cambridge, Mass. St Neot 1880 Ein angelsächsisches Leben des Neot. Ed. R. P. Wülcker. Anglia 3. TC = Troilus and Criseyde 1957 In: The Complete Works of Geoffrey Chaucer. Ed. F. Ν. Robinson. Wars of Alexander 1886 The Wars of Alexander. (Early English Text Society, Extra Series 47.) Ed. W. W. Skeat. Studies Bald, Wolf-Dietrich 1984 "Form and functions of ONE: Diachronic aspects", in: Wolf-Dietrich Bald—Horst Weinstock (eds.), 143-153. Bald, Wolf-Dietrich—Horst Weinstock (eds.) 1984 Medieval studies conference in Aachen 1983. Language and literature. Frankfurt am Main: Verlag Peter Lang. Björling, Augusta 1925 "The propword one in Early Modern English", Anglia Beiblatt 36: 76-78. Christophersen, Paul 1939 The articles: A study of their theory and use in English. Copenhagen: Munksgaard.

The pronominalization of one

141

Denison, David 1994 "A corpus of late Modem English prose", in: Merja Kytö—Matti Rissanen—Susan Wright (eds.), 7-16. Einenkel, Eugen 1903-4 "Das englische Indefinitum", Anglia 26, 27: 461-472; 1-204. 1906 "Nachträge zum englischen Indefinitum", Anglia 29: 542-544. 1912 "Der Ursprung der Fügung a good one", Anglia 36: 539-544. 1914 "Nochmals zur Fügung a good one", Anglia 38: 193-212. Fischer, Olga 1992 "Syntax", in: Norman Blake (ed.), The Cambridge history of the English language II. Cambridge: Cambridge University Press, 207408. Fisiak, Jacek (ed.) forthc. Studies in Middle English. Berlin—New York: Mouton de Gruyter. Franz, Wilhelm 1939 Die Sprache Shakespeares in Vers und Prosa. (Fourth edition.) Halle/Saale: Niemeyer. Fröhlich, Jürg 1951 Der indefinite Agens im Altenglischen. (Schweizer Anglistische Arbeiten 25.) Bern: Francke. Givön, Talmy 1981 "On the development of the numeral 'one' as an independent marker", Folia Linguistica Historica 2: 35-53. Gonda, J. 1953 Reflections on the numerals 'one' and 'two' in ancient Indo-European languages. Utrecht: N.V. A. Oosthoek's Uitgevers-Mij. Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Hopper, Paul J. 1991 "On some principles of grammaticization", in: Elizabeth Closs Traugott—Bernd Heine (eds.), 1: 17-35. Hopper, Paul J.—Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Jespersen, Otto 1949 A modern English grammar on historical principles II. London: George Allen and Unwin. Jud-Schmid, Elisabeth 1956 Der indefinite Agens von Chaucer bis Shakespeare: Die Wörter und Wendungen für 'man'. Meisenheim/Glan: Hain. (Also Schweizer Anglistische Arbeiten 39. Bern: Francke, 1962.)

142

Matti Rissanen

Kahlas-Tarkka, Leena 1987 The uses and shades of meaning of words for 'every' and 'each' in Old English, with an addendum on Early Middle English developments. Helsinki: Societe Neophilologique. Kastovsky, Dieter—Gero Bauer 1988 Luick revisited. Papers read at the Luick-Symposium at Schloss Liechtenstein 1985. Tübingen: Gunter Narr. Kytö, Merja—Matti Rissanen—Susan Wright (eds.) 1994 Corpora across the centuries: Proceedings of the First International Colloquium on English Diachronic Corpora. Amsterdam—Atlanta, GA.: Rodopi. Langenfeit, Gösta 1946 "The roots of the propword one ", Studier i modern spräkvetenskap utgivna av Nyfilologiska Sällskapet i Stockholm 16: 97-138. Luick, Karl 1906 "Beiträge zur englischen Grammatik IV: Der Ursprung der Fügung a good one", Anglia 29: 339-346. 1913 "Zur Vorgeschichte der Fügung a good one", Anglia 37: 543-548. Meier, Hans Heinrich 1953 Der indefinite Agens im Mittelenglischen 1050-1350: Die Wörter und Wendungen für 'man'. (Schweizer Anglistische Arbeiten 34.) Bern: Francke. Miliö, Louis T. 1990a "A new historical corpus", IC ΑΜΕ Journal 14: 26-39. 1990b The Century of Prose Corpus. Cleveland State University, Cleveland. Mustanoja, Tauno F. 1958 The English syntactical type One the best man' and its occurrence in other Germanic languages. Helsinki: Societe Neophilologique. 1960 A Middle English syntax I. Helsinki: Societe Neophilologique. Nevalainen, Terttu 1991 BUT, ONLY, JUST: Focusing adverbial change in Modern English 1500-1900. Helsinki: Societe N6ophilologique. Rissanen, Matti 1967 The uses of 'one' in Old and Early Middle English. Helsinki: Soci£t6 Nöophilologique. 1988 "Where philology and linguistics meet: Reference, (in)definiteness and Old English sum", in: Dieter Kastovsky—Gero Bauer (eds.), 295-310. forthc. "Whatever happened to the indefinite pronouns in Middle English?", in: Jacek Fisiak (ed.), Studies in Middle English. Strang, Barbara 1972 A history of English. London: Methuen.

The pronominalization of one

143

Siisskand, Peter 1935 Geschichte des unbestimmten Artikels im alt- und frühmittel-englischen. Halle/Saale: Niemeyer. Traugott, Elizabeth Closs 1989 "On the rise of epistemic meanings in English: An example of subjectification in semantic change", Language 65: 31-55. Traugott, Elizabeth Closs—Bernd Heine (eds.) 1991 Approaches to grammaticalization 1-2. Amsterdam: Benjamins. Traugott, Elizabeth Closs—Ekkehard König 1991 "The semantics-pragmatics of grammaticalization revisited", in: Elizabeth Closs Traugott—Bernd Heine (eds.), 1: 189-218. Wells, S. G. Taylor—J. Jowett—W. Montgomery 1989 William Shakespeare: The complete works. (Electronic edition for the IBM PC.) Oxford: Oxford University Press. Zandvoort, R. W. 1940 "Pregnant one", English Studies 22: 199-200.

The processes of adverb derivation in Late Middle and Early Modern English Terttu Nevalainen

1. Introduction Adverb formation is one of the basic issues covered by most grammars of the English language. What they offer us is the synchronic rule of turning adjectives into adverbs by means of suffixation, accompanied by a list of exceptions that do not differ in form from the corresponding adjectives. 1 On closer inspection, things turn out to be less obvious and clear-cut both synchronically and diachronically. If adverb formation is regular, we may first of all wonder whether it is an inflectional or a derivational process. What, if any, are the limitations on the productivity of -LY suffixation? Another intriguing issue is the relation between adjectives and suffixless adverbs. Is zero-formation still a productive process? If not, at what stage in the history of the language did it cease to be so? Taking a diachronic view of adverbialization, one may ask to what extent the present situation is simply the outcome of past processes of adverbialization. It is questions like these that have prompted my diachronic study of adverbialization in Late Middle and Early Modern English (c. 1350-1700), a period that leads up to the present-day system of adverb formation, but still reflects its Old English origins. I shall cover the primary processes of -LY suffixation and zero-formation as evidenced in the Helsinki Corpus materials for these periods. Secondary processes will also be discussed. Both kinds of process relate to different adverb classes, which therefore have a central role in accounting for the attested variation.

146

Terttu Nevalainen

In his Old English Syntax (1985: 471-472), Bruce Mitchell states with reference to Old English that As in MnE, there is a certain amount of overlap in the functions of adjectives and adverbs. [...] When the forms are similar, we cannot always be sure which we have. In Old English, ambiguity arises in cases where deadjectival adverbs and adjectival inflections both end in -E. In later stages of the language these old adverb forms regularly coincide with adjectives, which no longer take inflectional endings. New adverbs are produced either by means of zeroformation or, more commonly, by suffixation. The Old English adjective suffix -LIC followed by the adverb suffix -E had come to be reanalysed as -LYCHE, an adverb suffix later shortened to -LY.2 Below I shall argue that syntactic homonymy between adjectives and adverbs, both suffixed and suffixless, presents problems of formal and functional interpretation in Late Middle and Early Modern English. In certain syntactic environments this indeterminacy continues until the present day. We may refer to them as weakly codified areas in the grammar of English. My article falls into two parts. I shall first discuss adverbialization from the synchronic perspective, showing that adverbialization appears to be a derivational process in Present-day English. I shall then analyse the functional overlap between the two word classes, adjectives and adverbs, suggesting ways in which they could be distinguished both synchronically and diachronically (sections 2-3). Although a separation between the two may not always be desirable, a principled division is necessary if we wish to distinguish between adjectives and adverbs in the lexicon. Focusing on Late Middle English and Early Modern English, I shall devote the second half of my paper (sections 4-5) to the diachronic relations between -LY suffixation and zero-formation as productive processes of adverbialization. One of the key issues here is the one-form one-meaning principle, or the amount of redundancy between the two processes in the history of English. Studying both adverb types and adverb tokens enables me to compare lexical potentiality with actual use, and to relate adverbialization to certain more general principles of grammaticalization.

The processes of adverb derivation

147

2. Adverb formation: inflection or derivation? The title of my article suggests that I will argue that adverbs were derived rather than inflected in Late Middle and Early Modern English. This is not the traditional view, which basically tends to regard adverb formation as an inflectional process. The inflectional view is upheld by a long line of Latin-based scholarship, in which both adverb formation and comparison come under accidence. Campbell (1959: 275-280) and Bourcier (1981: 96-98) continue this line in their accounts of Old English. 3 It is not surprising to find that in Present-day English studies approaches vary depending on the criteria used. Hence a lexicologist like Marchand (1969) excludes the -LY suffix from his discussion of adverb suffixes. He considers the suffix fully productive, and thus part of the inflectional system of the language. Many grammarians, on the other hand, find justification for the derivationist approach. Huddleston (1984: 330-335) regards -LY as a derivational suffix partly because of the unstable semantic relation between the deadjectival adverb and its adjectival base. Quirk et al. (1985: 1556) similarly include -LY in their lexical inventory of adverb suffixes, but not without reservations. They remark that -LY can be very generally added to an adjective in a grammatical environment requiring an adverb [...] so that it could almost be regarded as inflectional. Zero-formed adverbs also fall outside Marchand's account of Present-day English word-formation, presumably because of their low productivity in Standard English. Quirk et al. (1985: 406) adopt a wider framework and draw attention to register variation. There is a limited number of adverbs formally identical to adjectives in standard use, but in nonstandard or very familiar English, zero-forms are common. Claudia Ross (1984) is even prepared to argue — to my mind without sufficient evidence — that Present-day North-American English is undergoing a change of dropping the -LY suffix from verb-modifying manner adverbs. I basically agree with Huddleston (1984) and Quirk et al. (1985) on adverb formation being a derivational process. In both -LY and zero-formations a regular change of lexical category is involved, adjectives becoming adverbs. The opposite is true of inflectional processes, where category labels are retained. What is more, both kinds of adverb may assume new meanings independently of the adjectives that they were originally asso-

148

Terttu Nevalainen

ciated with. These new meanings commonly coincide with functional reanalysis, for instance, the change from a verb-modifying manner adverbial, using the terms of Quirk et al. (1985), to a truth-emphasizing subjunct, as in example (1); from a manner adverbial to an item-modifying intensifier or focusing subjunct, as in (2); or from a manner adverbial/intensifier to a sentence adverbial, as in (3). (1)

(2)

(3)

a.

She likes to dress plainly.

b.

She plainly likes to dress. (Quirk et al. 1985: 584)

a.

We stated that precisely.

b.

We stated precisely that. (Nevalainen 1991: 51)

a.

The silk seemed strangely white.

b.

Strangely, the silk is white. (Swan 1988: 10)

Given that adverbialization is a form of grammaticalization, as suggested in the introduction to this volume (see also Traugott 1988: 132-134), these processes of functional specialization may be characterized as further or secondary grammaticalization within the adverb category. It is particularly evident if we look at historical material, as will appear from my Middle and Early Modern English data. If we take full productivity as the main criterion for inflectional morphology, -LY suffixation does not fulfil it, let alone zero derivation. Kjellmer (1984) convincingly demonstrates that the constraints on adverbialization are mostly semantic. So -LY suffixation does not freely apply to nondynamic adjectives denoting colour, for instance, unless they are associated with dynamic nouns or clauses. Exceptions of the latter kind may occur especially in fictional writing; see example (4). (4)

'Nothing is lost for ever,' said Slartibartfast, his face flickering redly in the light of the candle which the robot waiter was trying to take away (Douglas Adams (1982) Life, the universe and everything, p. 79)

The limited productivity of zero-derivation has already been pointed out. Here, too, semantic constraints are in evidence, as with verb-modifying

The processes of adverb derivation

149

adverbs functioning as process adjuncts zero-derivation is largely limited to what Ungerer (1988: 31) calls elementary adjectives. The majority of these adverbs found in Present-day Standard English appear to have been lexicalized very early on, some of them in fact going back to Old English -E formations. New zero-adverbs based on elementary adjectives were nevertheless introduced into Late Middle and Early Modern English, including such common ones as CHEAP and SLOW (see Nevalainen 1994a, and 4.2, below). It is, of course, a different matter altogether whether we wish to make a distinction between inflexional and derivational morphology at all. It is discarded by many recent writers on morphology (see Bauer 1988: 7 3 87). For the above reasons, and because I regard adverb derivation as lexeme formation, I find the distinction worth preserving. This approach unifies the three main kinds of adverbialization, viz. suffixation, zero-formation and compounding. Although derivation by means of suffixes other than -LY will not be discussed below, the approach would naturally accommodate formations with -WISE and other denominal suffixes.

3.

Formal and functional overlap between adverbs and adjectives

3.1. Criteria of analysis Inflectional or derivational, zero forms are notoriously problematic. The situation becomes even trickier in diachronic studies, where we have less than satisfactory distributional data for identifying zero-derived adverbs. Our criteria of inclusion may also vary. It is therefore necessary to devote some space to a discussion of the distributional overlap between adverbs and adjectives. 4

3.1.1. Adjective versus suffixed adverb Although distributional criteria are on the whole reliable indicators of adverb status, some conflicts may arise. Even suffixed adverbs may behave in unexpected ways. Mustanoja (1960: 315, 649-650) points out that in certain functional areas adverbs are often used in Middle English in-

150

Terttu

Nevalainen

stead of adjectives. An adverb of manner, or modal adverb, as they are called in older literature, may be found in expressions where one would expect an appositive or predicative adjective. Cases like (5) are not uncommon in Chaucer. (5)

As leene was his hors as is a rake, And he nas nat right fat, I undertake, But looked holwe, and therto sobrely. (Chaucer, The Canterbury Tales. The General Prologue 289)

Mustanoja regards this usage in Chaucer, for instance, largely as a literary tendency influenced by Latin models. Otto Jespersen (1961, Part 3, 367368) is more specific when he notes that predicative adverbs may occur instead of adjectives with verbs of appearance. This is common with verbs such as LOOK in Shakespeare. The usage remains variable after some copulas until the present day. Quirk et al. (1985: 4 0 7 ^ 0 8 ) point out that there is uncertainty in the use of adjective and related adverb forms in cases like (6) and (7): (6)

The flowers smell sweetHsweetly.

(7)

She felt bad/badly about it.

I had no problems in finding similar distributional variation in the Early Modern English section of the Helsinki Corpus, which means that it is not uncommon. The example in (8) comes from The Merry Wives of Windsor, and (9) and (10) from late 17th-century travelogues. (8)

I warrant you Coach after Coach, letter after letter, gift after gift, smelling so sweetly; (Shakespeare, The Merry Wiues of Windsor 2 2 66)

(9)

and there being another house pretty close to it high built with such a tower and lanthorn also, with the two churches towers and some other buildings pretty good made it appear nobly at a distance; (Fiennes, The Journeys of Celia Fiennes 151-152)

(10)

it can receive no Light but at the Doors and Windows of the Porch, whereby it looks more solemnly, (Fryer, A New Account of East India and Persia I 186)

The processes of adverb derivation

151

These cases can be approached from two different angles. Proceeding from purely formal criteria, it is not impossible to analyse these predicate complements as adjectives. Deadjectival adjectives could be derived by means of the -LY suffix in Middle and Early Modern English, although the process was much more limited than in Old English. These latish forms include such items as LOWLY, MEETLY, SICKLY in Middle English, and WEAKLY and POORLY in Early Modern English (Marchand 1969: 330, Mcintosh 1991: 300). In view of the low productivity of the process, however, this analysis does not appear plausible, especially as most deadjectival -LY adjectives have specialized senses — metaphorical and moral meanings attributed by Guimier (1985: 164) to the 'virtual character' of the adjective suffix. I shall therfore abandon this alternative. The other approach is distributional. Lexemes like NOBLY and SOLEMNLY typically occur in grammatical environments requiring an adverb, but they are not used as premodifiers of nouns, for instance. On distributional grounds they are prototypical adverbs, and should retain their word class label even when used as predicate complements of a small number of verbs. Further study into the complementation structure of this group of verbs would no doubt provide a more adequate account of what looks like a weakly codified area of grammar.

3.1.2. Adjective versus suffixless adverb I shall use the same distributional principle to distinguish between adjectives and zero-derived adverbs. In most cases, the issue is clear-cut with copulas such as LOOK, FEEL and SMELL: they require complementation that is typically realized by an adjective. But sometimes adjectives also encroach on the functional domain of adverbs. After intransitives, as in (11) and (12), the adjective phrase assumes an adverbial function in the clause structure (Quirk et al. 1985: 737). (11)

He came in drunk.

(12)

They married young.

Quirk et al. (1985) solve the conflict between adjectival form and adverbial function in two ways. They relate instances like (11) and (12) to verbless clauses that function as adverbial modifers in the sentence (giving rise to paraphrases such as 'he was drunk when he came in' and 'they

152

Terttu Nevalainen

were young when they married', respectively). Other cases are associated with the word-formation process of conversion. The instances in (13) and (14) are analysed as involving a change in secondary word class, which changes the intransitive verbs into copular (Quirk et al. 1985: 1565). (13)

We stood

motionless.

(14)

He fell flat.

Slightly different but related distinctions are made between predicative adjective phrases and subject adjuncts in Allerton (1982: 85-86), and clausal and extraclausal adjectives in Ferris (1993: 62-80, 147-166). I suggest that another way of solving the problem is to approach it within a more general dependency framework. Instead of converting verbs like STAND or FALL into copulas, and then classifying the zero-forms as adjectival subject complements, we could view all these cases as weakly codified. In examples (11) to (14), the adjectival Ά-element' could be seen as having the dual function of modifying the verb and the subject at the same time. In dependency terms, it would govern the subject while being itself governed by the verb. The Α-elements in cases like this appear to be good examples of what Matthews (1981: 185) calls "fused constructions". They also have the same dependency analysis as adjectives — or indeed adverbs in cases (5) to (10) — functioning as complements in copula sentences. But from the point of view of the present topic, the important thing is that lexically the Α-elements in (11) to (14) continue to be labelled as adjectives. A similar solution has been suggested for some Middle English data by Ohlander (1984: 442) and for Early Modern English subject modifiers by Killie (1993: 139-140). Ohlander analyses cases like (15) in terms of predicate attributes. By predicate attributes Ohlander means adjectives like joyles in (15), which qualify the subject, but also describe the circumstances that belong to the background of the verbal activity. (15)

Jonas al joyles and janglande

vp-rises. (Patience 433)

So far my search for zero-derived adverbs has been frustrated by what I have termed as imperfectly codified areas of use in cases where adverbs and adjectives overlap functionally. But there may also be collocational restrictions between the verb and the Α-element, the Α-element showing a closer connection between the verb than with the subject. This is where

The processes of adverb derivation

153

zero-derivation finally comes in. In Present-day Standard English there is only a limited number of verb-modifiers like DEAR and CLOSE, and they are usually analysed as lexicalized adverbs, as in Quirk et al. (1985: 407). As pointed out in section 2, some of these items go back to Old English -E formations (e.g. DEAR < DEORE, FAST < F^STE, LOUD < HLUDE). Some of them already had a suffixless form in Old English (e.g. FULL, the accusative form of the adjective), while others had dual forms (e.g. HARD ~ HARDLY < HEARDE ~ HEARDLICE; Campbell 1959: 275). In Late Middle and Early Modern English, these suffixless cases served as one of the productive models for adverb derivation, a model recognized by most early grammarians (see section 3.3). So new suffixless adverbs were also produced from Romance loan words (e.g. CLOSE < CLOS, JUST < JUSTE; Donner 1991: 10). My examples in (16) and (17) illustrate both the earlier and the later stratum, respectively. (16)

"Ihesu cryst, god almyghty, Of J^ys folk haue {)ou mercy, T»at JDOU boghtyst on rode so dere; (Mannyng, Handlyng Synne 64)

(17)

Look you here, Sir, do you see? (but you must stand very close) (Walton, The Compleat Angler 215)

In his study of Middle English manner adverbs recorded in the Middle English Dictionary, Donner (1991) makes the generalization that the zeroadverbs of that period usually have more concrete meanings than the corresponding -LY derivations. He illustrates his point by contrasting contexts of use: To indicate the range of contrasts by some illustrative pairings, foul may refer to how pigs root, foully to how men sin; bright to how the moon shines, brightly to how anchoresses should see and understand God's runes; heavy to how prisoners are fettered with iron, heavily to how men are burdened with God's commands; high to how a sword is raised, highly to how ladies are attired. (Donner 1991: 4) It is noteworthy, however, that the two major historical dictionaries, The Oxford English Dictionary and Middle English Dictionary, do not always agree on their word-class assignments when it comes to suffixless adverbs.

154

Terttu Nevalainen

Of the two, the Middle English Dictionary appears to be more frequently in favour of adverbialization. Unlike the OED, it interprets as adverbs the Α-elements in collocations such as liven chaste ('live in chastity'), sauf and sone ('safely and quickly'), and dwellen/lasten/stonden stable ('steadfastly', 'securely', 'firmly'). Making dichotomous decisions is of course difficult, and in view of the scalar ('squishy') nature of many linguistic categories, is not always fully justified. The Oxford English Dictionary avoids the problem by labelling some of the borderline cases as quasiadverb(ial)s or referring to adjectives being used as adverbs. They include the Α-element in cases such as talk idle ('idly') and speak meek ('meekly'), for instance. Zero-derivation also produces more abstract adverbs. In cases of prim'ary adverbialization these senses may be inherited from the adjectival base of the adverb, or may be metaphorically connected with it. They need not be present in the corresponding Old English adverb in cases where there is one. Abstract senses are associated with the items illustrated in (18) and (19a). Their dictionary paraphrases are given in parentheses. The suffixed adverb in (19b) retains the more concrete sense 'firmly'. (18)

But nowe to your Widowe whome you loue so hotte ['ardently'] (Udall, Roister Doister 1. 199)

(19) a.

and ouer J)is we schulden stonde sad ['steadfastly', 'firmly'] in byleue of God and lyuen in vertewys, as Godis lawe bydde]) vs, (Wycliffite Sermons I 376)

b.

Tak thanne this epicicle & ley it {sadly &) euene vp on the visage of thin equatorie ... {The Equatorie of the Planetis 24)

When analysing my own data, I consulted the systematic collocational evidence provided by historical dictionaries. Although I did not always adopt their word-class labels for individual cases (especially if there was disagreement), the information on the lexical potential of a given item was very useful. In borderline cases I applied the principle of excluding fully transparent single attestations from my data, such as IDLE in go idle and WORTHY in the biblical context of that ye walke worthy off the vocation wher with ye are called (Tyndale, 1526, Ephes. IV.I). On the other hand, on the strength of lexicographical evidence, I included some marginal metaphorical instances, such as (18), above.

The processes of adverb derivation

155

As illustrated above in section 2, semantic change also regularly produces intensifiers, adverbs that scale up or down from an assumed norm. It is this class of adverbs that has been particularly augmented by means of zero-derivation in the course of time. Although they are formally identical, an intensifying adverb rarely overlaps distributionally with the corresponding adjective once it has has become lexicalized. The Helsinki Corpus data show that some of the suffixless intensifiers have much higher token frequencies than other zero-adverbs in Late Middle and Early Modern English. These high-frequency items include FULL, RIGHT, and VERY — the latter only generalized at the beginning of the Early Modern period (see below, section 4.3). 5

3.2. Middle English -LICH(E)/LYCH(E) Before going into the details of -LY and zero derivation, it is useful to have a closer look at the levelling of the purely formal distinction between the Middle English suffixes -LYCH and -LYCHE in the light of the Late Middle English data in the Helsinki Corpus. This levelling process had, of course, begun much earlier, and was nearing its completion in the LME period. In the Restoration data, the only form found in the corpus is -LY, with the occasional alternative spellings -LI and -LIE. The Chaucerian subperiod in the Helsinki Corpus, ME3, yields a total of 224 deadjectival adverb types ending in -LY and its variant forms (altogether 935 tokens; see 4.3). The vast majority of the cases have the modern form -LY. There is also a fair number of instances ending in -LI, and one in -LIE, but none in -LYE. The total also includes the instances ending in the older forms -LYCH/LICH and -LYCHE/LICHE as well as their northern variants -LIK and -LIKE. These minority variants are summarized below in Table 1. The figures in Table 1 indicate the extent to which the processes of morphological simplification had advanced at the time. There were no adverb variants containing the older forms in -C (-LIC, -LICE) or yogh (-LI3, -LI3E; for the regional distribution of the variants in Late Middle English, see Mcintosh et al. 1986, Vol. 1: 4 5 5 ^ 5 7 ) . Table 1 further shows that the number of different Ε-less forms in -LIK and -LYCH/LICH used as adverbs is 24 as opposed to 41 E-ful ones.

156

Terttu

Nevalainen

Table 1. Minor variants of the deadjectival adverb suffix -LY in ME3 (13501420) in the Helsinki Corpus. Suffix form

Ν of adverb types represented by the suffix form

-LIC, LI3 -LIK -LYCH -LICH

0 5 5 14

-LICE, LI3E -LIKE -LYCHE -LICHE

0 3 4 34 Derivations from the same base (types)

-LIK/-LIKE -LICH/-LICHE

1 8

To return to the formal overlap of adverbs and adjectives, it should be added that there are a number of adjectives ending in -LYCHE/LICHE in the corpus, such as HOMELICHE, SHAMELICHE, UNSEMELICHE, and WORDLICHE. Only one of them occurs in a syntactic position that in earlier times would have required the weak adjectival ending -E. If we want to make sure that the attested variation between -LICH and -LICHE forms is not functional or semantic, we should have a closer look at items that appear with both endings. As appears from Table 1, there are nine cases like that. They are listed below. BLITHLIK/-E, FRELICH/-E, HASTELICH/-E, OPENLICH/-E, OUTERLICH/-E, PRINCIPALICH/-E, PRIUELICH/-E, SPECIALICH/-E, TREWLICH/-E A comparison of these dual cases shows that they are used in the same syntactic functions, for instance, as focusing item-modifiers and as verbmodifying manner adverbials. Consider the pairs of examples in (20) and (21).

The processes of adverb derivation

157

(20) a.

we prayen wyth mekenesse this specialich. that the Statut ordeigned & made bi parlement holden at Westmystre in the sexte year o f . . . (Petitions 197)

b.

she hadde ofte i-counsailled her housbonde to forsake J>e riches of J)e world, and specialliche in £>e laste 3eres of his lyf, but sehe spedde noujt. (Trevisa, Polychronicon Ranulphi Higden VI 217)

(21) a.

& that al strangfe] vitaillers sholden with thair vitailles frelich kome to the Cite, to seile thair vitailles as wel be retaile as in other wyse, (Usk, Appeal(s) 24)

b.

Jjere £is William dredde leste he schulde noujtfreliche passe J)e see. and took a womman clooj) above his owne cle|)inge, (Trevisa, Polychronicon Ranulphi Higden VIII 103)

Cases like this show that these adverbs cannot be assigned to different functional classes, and that we must consider them mere spelling variants. They may be further related to regional or register differences or indeed to idiosyncratic variation in the Helsinki Corpus. It it worth noting that the Middle English section of the corpus was compiled bearing dialectal variation in mind, and that almost one quarter of the material (24%) is in verse form (as opposed to some 2% in the first Early Modern English subperiod, and none in the last). Interestingly, most of the -LICH forms appear in texts labelled as documents.

3.3. The evidence of early grammarians Early Modern English grammars recognize both -LY and zero-derivation as means of adverb formation in English. Both were discussed by William Bullokar in his Pamphlet for Grammar (1586: 40), which also mentions the denominal adverb suffix -WISE (as in tablewise, heartwise). Bullokar notes with reference to adverbs denoting quality that "Qualities end in ly, wise, or are adiectiues, compounded with a, or vsed aduerbially: al generally answering to how?"

158

Terttu Nevalainen

Table 2 gives an overview of the treatment of adverb formation in some early treatises of English grammar. They all mention -LY suffixation as the chief means of forming adverbs of quality. Zero-derivation is also illustrated and referred to in most of the treatises examined, as in Jonson (1640: 68): "Here also Adjectives, as well positive, as compared stand for Adverbs", and Greenwood (1711: 160-161): "there are great Numbers of Adjectives that are used Adverbially, or as Adverbs." Cooper (1685: 162) and later Greenwood (1711: 160) — repeating Cooper's examples — relate adverbs and adjectives 'transformationally' showing the correspondence between that was nobly done and that was a noble deed, and between God's mercy is infinite and God is infinitely merciful (for this method of exposition in present-day grammars, see e.g. Quirk et al. 1985: 404-405).

Table 2.

Means of adverb formation discussed in early grammars (plus-sign indicates the processes covered).

Grammar Bullokar 1586 Greaves 1594 Jonson 1640 Poole 1646 Wallis 1653 Cooper 1685 Miege 1688 Greenwood 1711 Maittaire 1712

-LY

Zero

+ + +

+

+ + + + + +

-

Comment

illustrations of zero

+ -

+ + -

illustrations of zero

+ +

4. Deadjectival adverbs in the Late Middle and Early Modern English sections of the HC After this preparatory discussion it is time to turn to the empirical part of the study. It was initially sparked off by Morton Donner's survey (1991) of manner adverbs in the Middle English Dictionary. Donner finds that

The processes of adverb derivation

159

adverbs ending in -LY clearly dominate in the ME period, but suffixless forms occur frequently in texts characterized by native lexis. As pointed out above, he further notes that zero-forms usually have more concrete senses than the corresponding -LY forms in Middle English. What we do not learn from this comparison is the realization and relation of the two kinds of adverbs in texts. In particular, we do not learn the degree of overlap between the two processes of derivation in actual practice. How much free variation was there between the two? Although Donner must have accepted the MED adverb labels without trying to sort out the criteria used — and there are inconsistencies (see section 3.1.2 and Nevalainen 1991: 130) — his results suggest that the overlap was far less extensive than is often assumed.

4.1. Delimitation of the material In order to explore the matter further, I searched all deadjectival adverbs with dual forms in the Late Middle and Early Modern English sections in the computerized Helsinki Corpus. At this stage I have confined myself to positive forms only, but hope to return to comparatives and superlatives later. My systematic findings are based on the first Late Middle English subperiod (1350-1420), and the first (1500-1570) and last Early Modern English subperiod (1640-1710) of the Helsinki Corpus. The material thus covers the time span from about 1350 to 1710 in three phases. The obvious problem with my topic is locating the suffixless forms in a corpus that has not yet been lemmatized and provided with word-class taggings. In this situation my first task was to find all the relevant instances of -LY forms in the Late Middle English and Early Modern English data. The WordCruncher substring search was first applied to all the variants of the adverb suffix in all the LME and EModE material in the corpus. This task was relatively easy to accomplish; the number of variant spellings of the suffix was no larger than fourteen (see above, 3.2). My second step was to extract all the zero forms corresponding to the -LY adverb types found in the corpus. This was more time-consuming. I started by pruning away those items that could not possibly have zeroforms. The obvious cases were denominal -LY adverbs, which do not have corresponding adjective forms. In my data, this class included such items as NAMELY, PARTLY, PURPOSELY, SECRETLY, and YEARLY.

160

Terttu Nevalainen

As was shown in section 3.2, forms ending in -LICH/LYCH and -LICHE /LYCHE couid be used indiscriminately as adverbs and adjectives in Late Middle English. There are therefore no reliable morphological criteria for distinguishing between the adjective and adverb forms of items like ONLY and EARLY in Late Middle and Early Modern English (see Nevalainen 1991: 130). As borderline cases, I therefore omitted all dual-function instances ending in -LY, earlier -LYCH(E) and -LICH(E). On the other hand, dual-function items like LOWLY were retained for the present purposes because they could also have a zero form. After these measures, I was left with a total of 678 different suffixed adverb types representing the LME and EModE sections in the Helsinki Corpus. My final task was to locate their corresponding zero derivations by going through all their suffixless spelling variants in their respective contexts in the three subperiods under study. At this stage I had to decide whether to store all of them on disk or to work my way through them all and determine, as I went on, whether a given instance was an adverb or not. I decided on the first alternative. It yielded a large KWIC concordance consisting of homonymous adjectives and adverbs, but it also made it possible for me to to go back and reanalyse the data at a later stage, which proved useful.

4.2.

Adverb types in -LY and zero

4.2.1. Long-term trends The Helsinki Corpus materials reveal the degree of potential and actual overlap between suffixed and suffixless adverbs in Late Middle and Early Modern English. Table 3 shows the number of adverb types found in the three subperiods; their relative frequencies are presented in Figure 1. The figures for -LY adverbs stand for the suffixed forms attested in a given period. As mentioned in 4.1, zero adverbs were searched on the basis of all the -LY forms found in the five Late Middle and Early Modern English subperiods in the Helsinki Corpus. This means that those zeroes that do not have an -LY variant in the corpus are not included. A comparison with historical dictionaries suggests, however, that not many common items have been omitted. 6 The number of zero-adverb types found in the three subperiods studied here varies from 60 to 48.

The processes of adverb derivation

161

Table 3 also shows the degree of overlap between the suffixed and suffixless forms derived from the same adjectival base. It is remarkably small and stable throughout the time span studied. A larger corpus might alter the figures somewhat. The three Helsinki Corpus subperiods studied cover some 180,000 running words of Late Middle English and 360,000 of Early Modern English. As the long-term trends attested in the data proved to be statistically significant, however, the data would appear to serve as an adequate diagnostic sample of actual usage, providing a welcome supplement to the lexicographical evidence in studies like Donner's. Table 3.

Period 1350—1420

Variations between adverb types in -LY and zero in three periods in the Helsinki Corpus. -LY 224 (79%)

Zero

Total

60 (21%)

284 (100%)

Derivations from the same base (types)

1500-1570

283 (84%)

38 (13%)

284 (100%)

54 (16%)

337 (100%)

Derivations from the same base (types)

1640-1710

324 (87%)

34 (10%)

337 (100%)

48 (13%)

372 (100%)

Derivations from the same base (types) 36 (10%)

372 (100%)

162

Figure

Terttu

1.

Nevalainen

Relative frequencies of suffixed and suffixless adverbs in Late Middle and Early Modern English in the Helsinki Corpus.

The processes of adverb derivation

163

The figures in Table 3 support Donner's (1991) findings that there are much fewer suffixless than suffixed forms in Middle English. Suffixless forms constitute only 30% of the manner adverbs recorded in the Middle English Dictionary for the twelfth century.7 In my Late Middle English data they correspond to about one fifth of the adverb types considered, and their token frequency is often low (see section 4.3, below). Table 3 further suggests that zero-forms are generally losing ground in the Early Modern period. In the Restoration data, their share has dropped to 13% of the total. This change is gradual but statistically significant. For the three periods, the overall chi-square value amounts to 7.76, which makes the distributional differences statistically significant at the 5% level. Further testing shows the gradualness of the process. The difference between the late fourteenth and late seventeenth-century data is highly significant at the 1% level (chi-square 7.69). At the same time, the frequency differences between adjacent periods, the first and the second, and the second and the third, both fail to reach the required 5 per cent level of significance. As the amount of text included in each period varies to some extent, it is interesting to find out how stable the number of adverb types remains. I therefore calculated the number of adverb types per ten thousand words in each of the three periods. These figures are presented in Table 4. Table 4.

The frequencies of adverb types per 10,000 words.

Period

-LY

Zero

Total

1350-1420 1500-1570 1640-1710

12.2 14.9 18.9

3.2 2.8 2.8

15.4 17.7 21.7

The information in Table 4 confirms the impression suggested by Table 3 that the variety and use of -LY derivations increases in the course of time, while zero-derived adverbs remain fairly stable, declining only slightly in the course of the period under study. It remains a topic for further study to investigate how the well-documented general increase in vocabulary intake in the decades around 1600 is reflected here (on Early Modern vocabulary developments in general, see Nevalainen, forthcoming). Somewhat unexpectedly, the number of dual forms varies very little. Donner (1991: 3-5) finds that no more than 16% of his twelfth-century

164

Terttu Nevalainen

manner adverbs display dual morphology. As we can see from Table 3, the proportion of dual forms of all adverbs falls from 13% in my Late Middle English data to 10% at the beginning of the Early Modern period. The distributional difference is not statistically significant. This extended period of stable variation could, of course, be partly accidental because the figures are small. For what they are worth, they suggest that towards the end of our period there was a slight increase in dual forms in the total of zero-forms: in the Late Middle English and the first Early Modern English period, dual forms cover 63% of the total of suffixless adverbs, and in the last period 75% of the total. It therefore looks as though dual forms occupied a slightly larger share of the shrinking domain of zero adverbs in the latter half of the Early Modern period than earlier on. If this is the case, the question arises whether we are also witnessing a simultaneous increase in functional overlap in the lexicon.

4.2.2. Late Middle English In order to see whether there was any functional convergence and hence increased redundancy across time, I related the dual form adverbs to their adverbial functions. As the numbers of these adverb types are relatively small, the following generalizations must remain tentative. I started with the assumption that since most suffixed forms could be expected to display at least their canonical verb-modifying manner adverbial function, a good starting-point for the comparison would be the corresponding zeroes. Results showed that out of the 38 suffixless dual forms in Late Middle English one third did not have a manner adverbial function. The functions attested included, to use the classification of Quirk et al. (1985), time and place adjuncts (INWARD, LONG, NEW), intensifies (CLEAN, FULL, MARVELLOUS), focusing subjuncts (EVEN, PRECISE, RIGHT), and disjuncts (CERTAIN). They are illustrated by examples (22) to (25). (22)

And if it be so f>at J)is manere of swellynge aforseide haue y-be longe with-oute helpe {)en schalt J)ou sclyte J)e skyn in myddes j}e bolnynge. (A Late Middle English Treatise on Horses 127)

(23)

No knowlechinge of him sehe fond, Bot al was clene out of memoire, As me mai rede in his histoire. (Gower, Confessio Amantis I 319)

The processes of adverb derivation

165

(24)

And God woot that in alle these langages and in many moo han these conclusions ben sufFisantly lerned and taught, and yit be diverse reules; right as diverse pathes leden diverse folk the righte way to Rome. (Chaucer, Astrolabe 662.C2)

(25)

and {)ey seyn to doo charite and almesse, for J)ey wolde haue what for-to 3yue. Bote certayn, |)yse wel ofte fayled of [)e heje parfeccioun off charite. (Aelred of Rievaulx's De Institutione Inclusarum 39)

Some of the suffixed counterparts of these zero forms can assume the same function and, as far as one can judge, there does not appear to be any syntactic or semantic difference between the two. Compare (25) with the instance in (26) of the disjunct CERTAINLY from Trevisa's translation of Higden's Polychronicon; both texts represent the same southern regional variety. (26)

Certeynliche £>e Kyng of hevene allowed noujt service f>at is so todeled, for wolde be served with al J)e myjt of J)e soule. (Trevisa, Polychronicon Ranulphi Higden VIII 87)

The two variants may also occur in one and the same text, as LANG and LANGLY in (27) from The Prick of Conscience. The obvious explanation here is the metrical structure of the verse text (see also Donner 1991: 1). (27) a.

For sumtyme J>ai sal be pyned lang With hete, and som tyme with cald omang. (The Pricke of Conscience 89)

b.

t>e mast veniel syns sal J)ar bryn langly, Als wodde brinnes, jjat es sadde and hevy, (The Pricke of Conscience 87)

In a number of cases there appear to be syntatic or semantic reasons for preferring one or the other of the two forms. As in Modern English, there is a tendency for the zero-forms to occur in compounds, as is the case with NEW in (28a) (see also Nevalainen 1994a). That this still leaves room for variation between NEW and NEWLY can be seen from (28b) and (28c).

166

Terttu Nevalainen

(28) a.

A new born barn lay in croppe, Bondon wit a suef)elband, (Cursor Mundi 84)

b.

and {)ere he was prisonede in his owen castel J)at he had new made, fiat stode a3eins Jje Abbay of Kyng Edward. (The Brut or the Chronicles of England 221)

c.

f)ou sal sai jx>u ert esau, Fra jje forest newli comen, Venison J)ou has him nommen, (Cursor Mundi 216)

Many intensifies are syntactically context-sensitive in the sense that zeroforms tend to co-occur with adjective and adverb heads, while -LY forms are preferred with verbs and participles (see Mustanoja 1960: 319-320). However, we can again only talk about a strong tendency, not a hard-andfast rule. Compare examples (29) and (30). (29) a.

First 3ee schull wel knowe J)at the naturell bawme is full cleer & of cytryne colour & strongly smellynge. (Mandeville's Travels 33)

b.

he ... sente anon to the court of Rome men of answere ful informed, wij? lettres of bisshops witnessynge J)at J)e monkes had forsake Christes chivalry, (Trevisa, Polychronicon Ranulphi Higden VIII 95)

(30) a.

Fürst, Cristes rewle were fully sufficient to alle men, and more free and more 113t and of more auctorite. (Wycliffite Sermons 32 I 357)

b.

For and I couj^e any schorter wordes, so fully comprehendyng in hem all goodes & alle yuelles, as Jjees two wordes don, (The Cloud of Unknowing 78)

The majority of the zero-derived adverbs in the Late Middle English Corpus assume verb-modifying functions. According to my preliminary analy-

The processes of adverb derivation

167

sis, almost one third of the suffixless dual-form adverbs come under Donner's concrete cases referred to above in section 3. Donner (1991: 4) notes that in his Early Middle English data these distinctions preponderate by a ratio of approximately three to one. Most of these zero-adverbs in the Late Modern English data are derived from elementary adjectives, that is, adjectives denoting physical property, dimension, speed, and value (for further discussion, see Ungerer 1988: 31-32, Nevalainen 1994a). Consider examples (31) to (33). (31)

And whanne this was herd, Moises and Aaron felden lowe to erthe, bifor al the multitude of the sones of Israel. (Wycliffe, The Old Testament Numbers 14 5)

(32)

A MARCHANT was ther with a forked berd, In mottelee, and hye on horse he sat; (Chaucer, The Canterbury Tales. The General Prologue 270)

(33)

And sekirly soche rude streynynges ben ful hard fastnid in fleschlines of bodely felyng, and fill drie fro any wetyng of grace. (The Cloud of Unknowing 87)

Some of their more abstract counterparts in -LY are illustrated by the cases in (34) to (36). (34)

Our most dred souueraign liege lord & noblest kyng, we lowly besech J>e kyng of kynges, whos reaume shal endelesly Last, your blessed soule and noble body from either of ther enemys euermore to kepe and defend. (London Letters 73)

(35)

Of which spede and welfare, and al your ojjer kyngly lustes and plesaunces, we desire highly be the sayd berers of thes letters ... fully to be lerned and enfourmed. (London Letters 76)

(36)

3yf 3ε do J)us, y sey hardly, 3e mow here joure mynstralsy. (Mannyng, Handlyng Synne 159)

168

Terttu Nevalainen

It is worth pointing out again that although they are normally labelled as adverbs in lexicographical sources like the Middle English Dictionary, many of these verb-modifying suffixless adverbs, concrete and abstract alike, resemble adjectives not only in form but also in their predicative function; as in (37) and (38). (37)

And for he bare hym so meke and softe, Schrewes mysded hym ful ofte, And helde hym folted or wode, For he was so mylde of mode. (Mannyng, Handlyng Synne 189)

(38)

and he might come into J?e temple, f>ere shulde no man him misdo, but gone f)erin sauf and in pees, and aftre gone into what contre |)at he wolde, withouten eny harme; {The Brut or the Chronicles of England 24)

In order to grasp the diachronic development in full, one should go beyond prototypical adverbs, and study the extent to which adjectives were used in adverbial functions in cases such as (11) to (15). An investigation of this kind would provide us with a better picture of any possible increase in codification in these syntactic environments across time.

4.2.3.

Early Modern

English

A qualitative comparison of the Late Middle and Early Modern data shows that, despite their gradual decline in terms of frequency, new zeroderived adverbs are also Iexicalized in the course of time. Two-thirds of the zero-forms in the first Early Modern English period (36 out of the total of 54 types) already appear in the Late Middle English data, while only half of the third period zero-adverbs (25 out of 48) do. A number of those zero-adverbs that do not occur in my Late Middle English data from 1350-1420 go back to Romance loans. In the 1500-1570 subcorpus they include, for instance, ABSOLUTE, CHIEF, CLOSE, ESPECIAL, EXCEEDING, INCONTINENT (in the sense 'immediately'), JUST, SURE, TENDER, and VERY.

The processes of adverb derivation

169

The base forms of these Romance-based adverbs mostly came to the language prior to the Early Modern period, many of them in the fourteenth century; many were also first adverbialized in Late Middle English (see Donner 1991: 8-9). It is a sign of their increased token frequency that they occur in the Helsinki Corpus texts. The fact that a great majority of these zero-forms display dual morphology in my data, moreover, means that the two processes of adverb derivation must have been available in both Late Middle and, to a lesser extent, Early Modern English. This conclusion is also supported by my earlier findings based on lexicographical sources (Nevalainen 1994a). The proportion of zero-forms assuming a variety of adverbial functions increases in the Early Modern data: almost half of the zeroes with a dual form function as something other than verb-modifing manner adverbials. The number of intensifiers is especially noticeable. In the last subperiod (1640-1710) it includes such items as CLEAN, EXCEEDING, EXTRAORDINARY, FULL, MIGHTY, MOST, NEAR, PRETTY, SCARCE, VERY, and WONDERFUL. Although it appears that the relative share of suffixless manner adverbs is diminishing in the data, it is worth bearing in mind that quite a few of the suffixed forms corresponding to the above intensifiers, for instance, may also assume intensifying functions (e.g. EXCEEDINGLY, EXTRAORDINARILY, FULLY, MIGHTILY, MOSTLY, NEARLY, SCARCELY, and WONDERFULLY). As in Middle English, the choice may be triggered by the kind of headword the adverb modifies, adjectives and adverbs favouring the suffixless form, and verbs taking the suffixed one. Compare the instances in (39) and (40). The use of the intensifier NEAR in (41a) as opposed to NEARLY in (41b) also seems to be prompted by lexeme-specific considerations of the kind of headword associated with the adverb, here noun and participle, respectively. (39) a.

And one of the lustiest saide: Syr, we be yong ientlemen, and wisemen before vs, haue proued all facions, and yet those haue done fiill well: (Ascham, The Scholemaster 185)

b.

Ye must beware of such meats as are hard of substance ... and those that are conueyed in by heapes into the bodie, before they be fully digested, or made ripe, (Turner, A New Boke of the Natures and Properties of All Wines D2V)

170

Terttu Nevalainen

(40) a.

and that there is not a good word said of any of us but of me, and me they do all speak mightily of — which, whether true or no, I am mighty glad to hear. (Pepys, The Diary of Samuel Pepys VIII 315)

b.

That of a Tarras-walk may do well for Vines, and the Gravelwalk under the Wall will mightily encrease the heat about them. (Langford, Plain and Full Instructions to Raise All Sorts of Fruit-trees 127)

(41) a.

although it was near Evening e're we set out, and Yesterday's Showers had benignly distilled on the Fiery Drought, to cool the parched Earth. (Fryer, A New Account of East India and Persia II 187)

b. None so nearly disposed to scoffing at Religion as those who have accustomed themselves to swear upon trifling occasions. (Tillotson, The Folly of Scoffing at Religion II:ii 428) Exceptions to this tendency also occur, especially towards the end of our period, when the number of intensifiers increases; consider the instances in (42) and (43). With some exceptions such as PRETTY, RIGHT and VERY, where the suffixed form is reserved for other functions, the general trend is for the suffixed form to spread to all environments. Given that prescriptive grammars mostly date from the 18th century, it is difficult to attribute this trend to any conscious motivation to regulate usage. The examples in (42) and (43) illustrate cases of functional overlap. (42) a.

Ο thou call'st upon me to declare to thee the greatest thing which could be asked, and which indeed can scarce be answered: (Preston, Boethius 190)

b.

there is a very high hill quite round stands up on one side of it and can scarcely be ascended so steepe; (Fiennes, The Journeys of Celia Fiennes 150)

(43) a.

her husband has a thoughtfull Countenance, is wonderfull serious & silent, seemes to treate all persons alike gravely: (Evelyn, The Diary of John Evelyn 902)

The processes of adverb derivation b.

171

together with the trees and gardens thick about it the prospect was wonderfully pleasant; (Fiennes, The Journeys of Celia Fiennes 151)

Similarly with verb-modifying manner adverbs, the earlier distinction between the more concrete and objective zero-forms and the abstract and subjective -LY derivations is becoming blurred. It is not difficult to find instances where the difference still obtains, as cases like (31) to (36) freely occur in Early Modern English (as in Present-day English, for that matter). But there are also counterexamples in the Corpus. In cases like (44) and (45) the semantic relations are, if anything, reversed: the zeroform is the more abstract — a focusing adverb in (44a) and an intensifier in (45a) — while its suffixed counterpart retains the original lexical sense of the base and functions as a manner adverbial modifying the verb. (44) a.

Why, thou vile Wretch didst not thou tell me just now that thou pluck'd up the Latch? (State Trials, The Trial of the Lady Alice Lisle IV 114.C2)

b.

Whereunto I answered, it was the Plague of God come justly vpon vs; and now Almighty God dealt with vs as he did with the Israelites, (State Trials, The Trial of Sir Nicholas Throckmorton I 69.C2)

(45) a.

Thus; they have proceeded from syllables of two or three; or more letters, till a child hath been pretty nimble in the most. (Hoole, A New Discovery of the Old Art of Teaching Schoole 6)

b.

Madam, I have guilded 'em, and rang'd 'em, so prettily, before Gad, it is the most entertaining thing in the World to walk and look upon 'em. (Vanbrugh, The Relapse or Virtue in Danger I 36)

The explanation here lies in semantic change, which can be witnessed with both suffixed and suffixless adverbs. Examples of the former include the sentence adverbial developments discussed by Swan (1988); for the latter, see the discussion of intensifiers in Peters (1993) and of focusing adverbs in Nevalainen (1991). As shown above in section 2, many of these adverbs result from secondary grammaticalization.

172

Terttu Nevalainen

4.3. Adverb tokens in -LY and zero The above analysis of adverb types yields a fairly coherent overall picture of adverb derivation in Late Middle and Early Modern English. In order to explore the actual realization of these processes in texts in more detail, we should study their token frequencies. The results of this analysis are presented in Tables 5 and 6. Table 5.

The frequencies of adverb tokens in -LY and zero.

Period

-LY

1350-1420 1500-1570 1640-1710

935 1219 1236

Total:

3390

Table 6.

Zero (54%) (62%) (57%)

794 743 922 2459

Total (46%) (38%) (43%)

1729 1962 2158 5849

The occurrences of adverb tokens normalized to a frequency of 10,000 words.

Period

-LY

Zero

Total

1350-1420 1500-1570 1640-1710

50.7 64.1 72.3

43.1 39.1 53.9

93.8 103.2 126.2

Overall, the figures in Tables 5 and 6 confirm the trend suggested by type frequencies in that they show a progressive increase in adverb use in both absolute (Table 5) and relative terms (Table 6) across time. This trend is carried by -LY adverbs, whose token frequency increases linearly from 50.7 per 10,000 words in Late Middle English to 72.3 in the latter half of the Early Modern period. A comparison of Tables 3 and 5 clearly suggests that some zero-adverbs must be generally more frequent in texts than -LY adverbs. Their relative type frequency (21%-13%) is more than doubled by their relative token frequency in the data (46%-38%). These results agree with Donner's

The processes of adverb derivation

173

(1991: 7-8) findings on the high frequency usage of suffixless adverbs in Middle English. Rather unexpectedly, the figures in Table 6 imply a nonlinear course of development in the diachronic distribution of zero forms. They reach their peak in the last subperiod with a noticeable drop in the first half of the sixteenth century. The same drop in the early 16th-century data can be seen in Table 5. This discrepancy is, however, more apparent than real, and does not correspond to a sudden revival of zero-derivation in the 17th century (cf. Table 4). It can be explained in terms of some highly popular individual items. In the Restoration data the 424 occurrences of the adverb VERY alone cover more than a half of all the zero tokens. No such trendinfluenced item appears among the -LY formations.

4.3.1. High-frequency tokens The ten most frequent zero adverbs in each subperiod are shown in Table 7 with their absolute frequencies in brackets (see below). The table indicates first of all that the most even distributional pattern among the top ten can be found in the middle period, 1500-1570, where only one adverb appears with a frequency higher than one hundred instances. The relative proportion of the ten most frequent adverbs of the zero adverb total is 85% in the Late Middle English data (1350-1420), 76% in the first Early Modern English subperiod (1500-1570) and 83% in the last (1640-1710). The Middle English data largely replicate the frequency information given in Mustanoja (1960: 330; derived from Fettig 1934) for the most frequent intensifiers in the period 1350-1470. In Fettig's extensive corpus FULL is by far the most common adverb used to modify adjectives and adverbs (3518 instances), but is very rare with verbs (41 instances). The next most frequent intensifier is RIGHT (1349 instances as adjective and adverb modifier). SORE comes third with 237 instances as a verb modifier and only three as an adjective and adverb modifier. The most common Early Middle English intensifier SWITHE is only represented by 15 instances in Fettig's Late Middle English data, while VERY is already attested in 50 cases as an adjective and adverb intensifier in texts dating from 1400-1470; no earlier instances are recorded by Fettig (1934: 178— 185).

174

Terttu Nevalainen

Table 7. The ten most frequent zero adverbs in the three periods (N). Period:

1350-1420

1500-1570

1640-1710

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

FULL (253) RIGHT (202) LONG (56) ALONE (32) MOST (30) SORE (24) STILL (24) EVEN (23) LATE (17) NEAR (17)

VERY (174) EVEN (96) LONG (96) RIGHT (71) STILL (38) SORE (23) FULL (18) NEAR (18) ALONE (17) LIKE (17)

VERY (424) EVEN (69) LONG (63) STILL (52) JUST (39) PRETTY (31) ALONE (26) NEAR (20) RIGHT (20) MOST (17)

In the Helsinki Corpus materials for the period 1350-1420, VERY appears a number of times as an intensifying adjective modifying nouns, basically serving as a metalinguistic comment on the appropriateness of a designation (see Brugman 1984: 27-29). Instances like (46b) with another premodifier are by no means atypical, and show VERY in an environment where it would later be unambiguously interpreted as an adverb. Compare (46a) and (46b). The adverb is already fully established, for instance, in Palsgrave's (1530: cccc.xlv) list of Early Modern English adverbs, where VERY translates the corresponding French adverbs BIEN, MOULT and FORT ("but nat indifferently"). (46) a.

b.

Thou seist to me it is a greet meschief To wedde a povre womman, for costage; And if that she be riche, of heigh parage, Thanne seistow that it is a tormentrie To soffre hire pride and hire malencolie. And if that she be fair, thou verray knave, Thou seyst that every holour wol hire have; (Chaucer, The Wife of Bath's Prologue 248-254) Thow thynkist in thyn wit ... That he nys but a verray propre fole. (Chaucer, Legend of Good Women 259 Prol.; OED s.v. very a. etc. 1.3.a)

The processes of adverb derivation

175

FULL and RIGHT, the items preferred in the Late Middle English corpus, are both on their way out in the Early Modern data, while the most popular Early Modern English adverbs, VERY and EVEN, maintain their rankings until the end of the period. LONG, by contrast, remains the third most frequent zero item throughout the entire time span studied. It is indicative of the functional specialization of the most popular zero adverbs that none of them primarily functions as a manner adverbial. They are mostly intensifiers (e.g. FULL, PRETTY, SORE, and VERY), focusing adverbs (EVEN, JUST, RIGHT), and time or place adjuncts (LATE, LONG, NEAR, STILL). As I have shown in Nevalainen (1994a), adverbs like EVEN and JUST are also undergoing further semantic and grammatical specialization in Early Modern English, acquiring an additive and an exclusive focusing adverbial function, respectively. Although they follow the general pattern of functional differentiation to be discussed below (4.3.2), pairs such as LIKE and LIKELY, on the other hand, remind us of the fact that some high-frequency items may be collocationally restricted (see Greenbaum 1970). The adverb LIKE is one of them. With few exceptions, such as (47a), it is almost uniquely confined to a subjunct function in the context LIKE AS ('just as') in the present data; see (47b) and (47c). There are some illustrations in The Oxford English Dictionary (s.v. like B. adv. 5, like as) from the Late Modern period, where the combination is spelled as one word, thus suggesting that for some speakers it later grammaticalized as a conjunction. (47) a.

For if it be lyke the figure of a circle pressed in length, and bothe endes lyke bygge, then is it called a tunne forme, or barrel 1 forme. (Record, The First Principles of Geometrie B2R)

b. And why it is so shapen, is, that it should be plycable to the stomacke, like as a hande dothe to an apple, to comforte her digestion. (Vicary, The Anatomie of the Bodie of Man 69) c.

And for proofe thereof, like as, amonge diuers other reasons and aucthorities, he declared that this Realme, being but one member and smale parte of the Church, might not make a particuler lawe disagreable with the generali lawe of Christes vniuersall Catholike Churche, (Roper, The Lyfe of Sir Thomas Moore 93)

176

Terttu Nevalainen

4.3.2.

The degree of overlap

Limiting ourselves to the ten most frequent zero adverbs, just how much functional overlap is there between them and their suffixed dual forms in numerical terms? It would appear very little. The figures in Table 8 show the frequencies of the corresponding -LY adverbs in the present data. Table 8. Frequencies of the -LY adverbs corresponding to the ten most frequent zero-adverbs in the three periods (N). Period:

1350-1420

1500-1570

1640-1710

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

FULLY (20) RIGHTLY (1) LANGLY (1) ALONELY (-) MOSTLY (-) SORELY (-) STILLY (1) EVENLY (1) LATELY (-) NEARLY (-)

VERILY (5) EVENLY (-) LONGLY (-) RIGHTLY (4) STILLY (-) SORELY (-) FULLY (9) NEARLY (-) ALONELY (2) LIKELY (-)

VERILY (11) EVENLY (-) LONGLY (-) STILLY (-) JUSTLY (7) PRETTILY (1) ALONELY (-) NEARLY (1) RIGHTLY (4) MOSTLY (9)

Only half a dozen items in Table 8 appear with absolute frequencies higher than five, FULLY and VERILY heading the list. As shown by the illustrations in (29), (30) and (39), the choice of FULL as opposed to FULLY could be regulated by the kind of headword modified by the intensifier in both Late Middle and Early Modern English. The figures cited by Mustanoja (1960: 330) convincingly demonstrate that this was overwhelmingly the case in Late Middle English. Although there is some sporadic counterevidence to this tendency, FULL appearing as a verb-modifier and FULLY modifying a nominal head in Early Modern English, as in (48), overall their overlap remains negligible. The use of the intensifier FULL rapidly declines towards the end of the Early Modern period, where it is only represented by two instances in the corpus data. (48) a.

and sum hold opinion ... that the other course, part for pollicy and part by digging of stones for building of the town and minstre, was made a valley, and so the water-course was conveyid that way, but I approve not ful this conjecture. (Leland, The Itinerary of John Leland I 73)

The processes of adverb derivation b.

177

Galen in his booke of making of medicines, seemeth to call all Wine that is not fiilly fiue yeares olde, newe wine, (Turner, A New Boke of the Natures and Properties of All Wines B2R)

The dual forms MOST and MOSTLY display a distributional difference of precisely the opposite kind. Excluding the use of MOST in superlatives, the amplifier MOST tends to be associated with verbal headwords, and the particularizer MOSTLY with nominal. Their occurrences are illustrated in (49). The adverb MOSTLY is confined almost exclusively to one text in the Early Modern English data, The Journeys of Celia Fiennes. (49) a.

He wou'd give way to that hope, because it pleas'd him most, and flatter'd best his heart. (Behn, Oroonoko 160)

b.

in the best drawing roome was a very rich hanging gold and silver and a little scarlet, mostly tissue and of gold and silver and border of green damaske round it; (Fiennes, The Journeys of Celia Fiennes 153)

EVENLY, JUSTLY, LIKELY, RIGHTLY and VERILY all have similar relations to their suffixless forms. In the system of adverbial functions proposed by Quirk et al. (1985: 440-653) they operate as either process adjuncts or disjuncts, while EVEN, JUST, LIKE, RIGHT and VERY are used as subjuncts, "having a subordinate role in comparison with other clause elements" (Quirk et al. 1985: 566). In terms of their functional domains, rather than overlapping, these items are thus in complementary distribution. Despite these basic similarities, the adverb pairs have their lexemespecific differences. RIGHTLY is preferred in the manner adjunct function, whereas RIGHT is used both as a focusing subjunct ('just') and an intensifier ('very', 'quite'); see examples (50), and (24), above. (50) a.

neyther it was Treason for me to saye, I woulde hynder their commyng hither as much as I coulde, vnderstanding me rightly as I meant it; (State Trials, The Trial of Sir Nicholas Throckmorton I 72.C2)

b.

Yea, My lord, I know it right well; I meant not to touche your Lordship, or youre Service and Paines is euidently knowen to all men. (State Trials, The Trial of Sir Nicholas Throckmorton I 64.C1)

178

Terttu Nevalainen

VERILY, by contrast, is favoured as a disjunct. The examples in (51) illustrate the disjunctive function of VERILY as what Swan (1988: 2-3) calls a truth-intensifier at the sentence level. It is already common in this use in the Late Middle English data before the adverbialization of VERY. The high-frequency item-modifying intensifier VERY is shown in (49b), above. (51) a.

And if we diligently so do, I verily trust we shall find therin great comfort and consolacion. (The Correspondence of Sir Thomas More 509)

b.

I think verily hee is better contented that I shall be neerer home then those you and I have said are not contented, whose design is to have mee placed as far as may be. (Henry Oxinden, The Oxinden and Peyton Letters 277)

c.

Qua. Now Mary I plainly say, thy Dame is again departed. M. Yes, verily, thy Hand-maid doth see, and how easily good Woman, she is deceiv'd by the working of the Inward man; (Penny Merriments 150)

The extract from Walton's The Compleat Angler in (52a) illustrates the relative newcomer JUST both as a time subjunct ('right now') and as a particularizing subjunct focusing on now. The functions of JUSTLY as a verb-modifying manner adjunct and as a sentence-modifying content disjunct are shown in (52b) and (52c), respectively; see also (44), above. (52) a.

My friend Piscator, you have kept time with my thoughts, for the Sun is just rising, and I myself just now come to this place, and the dogs have just now put down an Otter; (Walton, The Compleat Angler 210)

b.

And there is no doubt but that all these things are done justly and in order, and for the Good of those to whom they happen. (Preston, Boethius 198)

c.

And now that those judgments have done no good upon us we may justly fear that he will appear once for all. (Tillotson, Of the Tryall of the Spirits II: ii 430)

The processes of adverb derivation

179

Syntactic and semantic differentiation also characterize most of the other dual cases, which display highly disparate quantitative relations in Tables 7 and 8. Functional overlap could perhaps be expected to occur with ALONE, LONG and STILL, as the -LY form can in principle closely duplicate at least one of the functions of the zero form. In practice overlaps turn out to be marginal. Although possible in both Late Middle and Early Modern English, LONGLY and STILLY are only found in the Late Middle English subperiod in the present data, and ALONELY in the first Early Modern subperiod.8 The contextual reasons for the choice of LONGLY are discussed above; see the examples in (27). STILLY ('silently', 'quietly') also occurs in verse data; see the case in (53a). In later English it does not appear to assume the most common function of the suffixless STILL as a time-relation subjunct (53b). In the case of ALONE and ALONELY, the two adverbs typically overlap in their focusing subjunct functions, as in the two cases in (54). A closer study of more data reveals that, besides their different frequency distributions, the two also tend to have different positional and focus preferences — a fact no doubt linked with their derivational histories (see note 8). (53) a.

Wit him allan stilli Jsai went, Vn-to jje kings aun tent, king and all his folk J>ai fand, In J>air beddes ful fast slepand. (Cursor Mundi 444)

b.

For it is well knowne, that I both like and loue, and haue alwaies, and do yet still vse, all exercises and pastimes, that be fitte for my nature and habilitie. (Ascham, The Scholemaster 216)

(54) a.

I hard that afore Rafe of Raby tyme ther was that alonly that now is the south isle (Leland, The Itinerary of John Leland I 76)

b.

This is a liuely and perfite waie of teaching of Rewles: where the common waie, vsed in common Scholes, to read the Grammar alone by it selfe, is tedious for the Master, hard for the Scholer, colde and vncumfortable for them bothe. (Ascham, The Scholemaster 184)

180

Terttu

Nevalainen

Overall, although there is a slight increase in the frequency of dual form tokens towards the end of the Early Modern period, there appears to be very little actual overlap between the most common suffixed and suffixless forms. Hence what we see is not so much convergence or increased redundancy in abverb formation as an increase in polysemy and functional homonymy. How much overlap there is in the semantic fields of the new adverbs is another matter and falls outside the scope of this study.

5.

Grammaticalization

Although adverbialization normally entails both a change of word class and a concomitant change in syntactic function, the functional division of labour between adverbs and adjectives may be blurred in practice. Zeroderivation does not always produce prototypical adverbs. This is the case with adjective-like modifiers that enter into collocational relations with the verb and the subject at the same time. In cases like these, categorical word-class assignments may not be syntactically justified. As suggested in 3.1.2, a more adequate analysis of the data could be carried out within a general dependency framework. In lexicography, where stricter categorization is expected, one might do worse than adopt the time-honoured 0£D-label of quasi-adverb(ial), which relates adverbs to the general notion of 'squishiness' or scalarity of grammatical categories. Prototypical adverbs would come at one end of the scale and the most adjective-like quasi-adverbs at the other. The scalar model could accommodate, for instance, those borderline cases between adjectives and zero-derived adverbs where the element is collocationally related with the verb, but has distributional limitations as an adverbial — or where our diachronic data do not warrant further generalizations. In this context it is interesting to find that even a well-established zero-adverb like SLOW can be restricted to a position after the verb or the object in Present-day English. Quirk et al. (1985: 405) illustrate the situation by the two examples cited in (55). (55) a. b.

He slowly/*slow drove the car into the garage, He drove the car slowly/?*slow into the garage.

The processes of adverb derivation

181

What then are the criteria for prototypical adverbs? The principal ones discussed above have been morphological (the adverbial suffix -LY as opposed to the undifferentiated zero) and distributional ("grammatical environment requiring an adverb"). The third is frequency of occurrence (see Nevalainen 1991: 94-95). It has been used as a descriptive measure to demonstrate that there is in practice very little overlap between the adverbs produced by the two processes of adverbialization discussed. If adverbs overlap in function, they rarely do so in frequency. In recent literature, frequency of occurrence is promoted as one of the key factors leading to grammatical ization. To quote Traugott and Heine (1991: 9): Given that a form A is a candidate for grammaticalization both because of its semantic context and salience, a further condition has to apply for grammaticalization to take place: The form has to be used frequently. The more grammaticalized a form, the more frequent it is. In my systematic corpus data the two processes of adverbialization differ greatly as to their type frequency. It is the relatively unambiguous process of -LY suffixation that dominates with frequencies ranging from 79% of the adverb types in the 14th century to 87% in the seventeenth. Interpreted in Halliday's (1992) terms of systemic markedness, the two processes approach the ratio of 0.9 to 0.1 in Early Modern English. This means that, within the category of adverb types with dual morphology, zero-derived adverbs are on their way to becoming a marked choice. The picture is diversified by functional specialization within the adverb category. Unlike studies such as Donner's would lead us to expect, zeroadverbs are not confined to concrete meanings or manner adverbial function in either Late Middle or Early Modern English. They can function as verb-modifying time and place adverbials, and become increasingly frequent as item-modifiers (intensifiers, emphasizers, focusing adverbials), that is, in less prototypical adverbial functions, because they operate on the phrasal rather than clausal level. It is in these functions that many of them were originally lexicalized, and have been preserved until the present day. Some of the items have undergone secondary abverbialization and become polyfunctional. Scalar focusing subjuncts, such as JUST and EVEN, are good examples of this principle of divergence in grammaticalization (see Hopper 1991: 22-25, Nevalainen 1994a). Bearing in mind the weakly codified contexts referred to above, it seems that zero-derivation is a marked process in terms of both frequency and

182

Terttu

Nevalainen

adverbial function. However, we should not forget that inflected zeroforms are expected to be more frequent with comparatives and superlatives than we have seen to be the case with positive forms (Quirk et al. 1985: 406). A systematic diachronic study of adverb comparison and participial compounding would therefore be needed to redress the balance between zero-derivation and suffixation. Overall, there appears to be a long-term diachronic trend towards increased grammaticalization in adverbialization in the English language. Deadjectival adverbs were regularly inflected for comparison in Old English (e.g. HEARDE, -OR, -OST; Campbell 1959: 277). In Present-day Standard English analytic comparison has largely replaced inflectional comparatives and superlatives with suffixed deadjectival adverbs in -LY. Given that inflections are associated with lexical rather than grammatical categories, we may thus interpret this as a tendency for the English adverb to become morphologically invariable, and hence more grammaticalized. It is noteworthy that regular morphological marking is typically associated with verb-modifying manner adverbials. It is this prototypical adverbial domain that favours the -LY suffix in cases of functional overlap between the suffixed and suffixless form from the Middle English period onwards. However, despite its predominance, -LY suffixation is not fully productive even in this functional domain. Certain large adjective classes are excluded from the process on semantic grounds, typically non-dynamic adjectives, such as colour terms. We must therefore conclude that although -LY suffixation continues to gather momentum, the process cannot be assigned to inflectional morphology on distributional grounds.

6. Summary and conclusion My systematic diachronic data show that at no point in Late Middle or Early Modern English was zero-derivation a serious contender for -LY suffixation as the prototypical process of adverbialization. Despite its rather generous allowance for suffixless adverbs, the Middle English Dictionary evidence studied by Donner (1991) suggests the same for Early Middle English. The type frequencies of adverbs with dual morphology reveal the gradual decline of zero-derivation in Early Modern English. The proportion of zero-derived adverb types in the Helsinki Corpus falls from 21% in Late Middle English to 13% at the end of the Early Modern pe-

The processes of adverb derivation

183

riod. With some notable exceptions, such as VERY, this is also true of their token frequencies. Zero-derived adverbs are more prominently represented at the token level, where they constitute about 40% of the adverb total in my material. These figures conceal, however, the heavy concentration of zero-tokens on some highly popular lexemes. In Late Middle and Early Modern English alike, the ten most frequent items cover some 80% of all the zero adverb tokens; in the last period studied, the most popular one, VERY, in fact covers more than half of them. No similar concentration occurs with -LY adverbs. The suffixed forms steadily gain ground in terms of both types and tokens in the corpus data. This happens no doubt partly as a function of the general increase in vocabulary intake in the Middle and Early Modern periods on the one hand, and the variation in the intensifier and sentence adverbial subcategories, on the other. Swan (1988) demonstrates that sentence adverbs come to be particularly associated with -LY in English. The present corpus evidence of polyfunctional adverbs largely supports her findings. The functional and semantic diversification of the adverb category in the course of time will then partly account for the attested numerical increase of -LY adverbs. The Helsinki Corpus material suggests that there is not much free variation between suffixed and zero-adverbs in running texts. Functional differentiation is often linked with differences in frequency of occurrence. So -LY forms are not found to overlap with high-frequency zero-adverbs derived from the same base, even when the two have separate functions to perform. In cases of functional parallelism, distributional differences commonly emerge, typically in the choice of headword. In diachronic terms, dual-form adverbs do not always run exactly parallel in time. We may question the whole notion of duality in cases with considerable time gaps between the two forms, although the potentiality for dual morphology of course cannot be denied so long as both processes are productive. With sufficient time depth, semantic and functional specialization will sever links between dual adverbs and their shared derivational base, and hence between the adverbs themselves. This becomes apparent from present-day dictionaries, which do not necessarily connect JUST and JUSTLY, PRETTY and PRETTILY, or VERY and VERILY (labelled as archaic or biblical). In grammars of Standard English the number of dual adverbs has dwindled to a dozen or two. They thus indirectly recognize the lexicalization of dual forms together with the demise of zero-derivation as a regular process of adverbialization in Standard English. 9

184

Terttu

Nevalainen

On the productive side, the secondary processes o f adverbialization, typically sentence adverbs formation, increase the number of adverbs whose interpretation is subjectively linked with their adjectival bases. The productive process o f -LY suffixation yields morphologically invariable deadjectival adverbs. If compared at all, they cannot be inflected but must be compared periphrastically. So although adverbs are heterogeneous in function and do not constitute a grammatical category in the functionword sense o f the term, I conclude from all this evidence that English today has a more grammaticalized adverb category than it had in the past.

Notes 1.

The number of adverbs with dual morphology introduced in grammars seems to vary between forty-two in Jespersen (Vol. 7, 1954: 48-52), some of which Jespersen in fact interpretes as adjectives (e.g. EASY in take it easy), and a dozen in Quirk et al. (1985: 405^107), including CLEAN/LY, CLOSE/LY, DEAR/LY, DEEP/LY, DIRECT/LY, FINE/LY, HIGH/LY, LATE/LY, LIGHT/LY, SHARP/LY, SLOW/LY, WIDE/LY. Interestingly, Jespersen does not include FINE, LIGHT, SHARP or SLOW in his list.

2.

The suffix -LIC itself is the result of a process of grammaticalization. It can be traced back to a Germanic noun meaning 'a body, dead or living'; for a detailed discussion, see Guimier (1985).

3.

Campbell (1959: 275) states that "the adjectival adverb is formed in OE by means of the suffix -e (< ce, Gmc. -ce), usually added to the same stem form as other inflections beginning with vowels". It is noteworthy that in Old English adverbialization can take place within the inflectional system as well, as living case forms are used adverbially and become lexicalized. In later English this is no longer possible, which strengthens the argument in favour of the derivational approach. For further discussion of the problem, see Nevalainen (1994b).

4.

A battery of tests for distinguishing between adverbs and adjectives in German has been devised by Bresson (1983); the problem here is that German adverbs are suffixless and thus formally identical with the base forms of the corresponding adjectives. See also the discussion of the problem in English by Bolinger (1979: 16-18). Elicitation tests are used by Opdahl (1990) and (1992) to study factors conditioning the choice of suffixed versus suffixless verb-modifiers in Present-day English.

5.

Item-modifying intensifiers often go two co-ordinate adjectives, which get tion. For further discussion of items (1969: 39-40), who suggests that, as

back to asyndetic constructions with reanalysed as structures of subordinalike VERY and PRETTY, see Peltola soon as the first member in the asyn-

The processes of adverb derivation

6.

7.

8.

9.

185

detic construction is felt to be semantically subordinate, it can develop into a general intensifier through 'catachrestic' extension. Similar processes would explain the rise of boosters from intensifying participles in -ING, especially common with those of the f-ing kind (my thanks to Elisabeth Eyes for drawing my attention to this). There is no suffixed counterpart in the data for the zero-derived adverbs FIRST and WRONG, for instance. FIRST occurs in all three periods with a very high frequency, but FIRSTLY is only attested in the Oxford English Dictionary for the first time in 1532. A larger database would obviously be needed to find these and similar correspondences. Donner's (1991) data consist of all the entries for manner adverbs (or "modal" adverbs, as he calls them) in the Middle English Dictionary, from A to sheden, altogether some 1,500 adverb types, illustrated by some 16,500 examples. Unfortunately, his statistics for the later Middle English periods are not very systematic. ALONELY is not directly related to ALONE derivationally, but both result from compounding. ALONE goes back to an Early Middle English collocation of the intensifier ALL and the polyfuntional ONE 'by oneself, 'alone', while ALONELY was adverbialized in the early 14th century as a combination of the intensifier ALL and the exclusive ONLY, the latter itself derived from ONE (Nevalainen 1991: 127-129, 135-136; for the adverb ONE, see Rissanen 1985). We are basically left with four kinds of relationship between variable dualform adverbs in Present-day English (see Quirk et al. 1985: 406-407, Opdahl 1990, 1992): 1) functional and semantic differentiation (HARD versus HARDLY), 2) semantic differentiation within the same adverb category (LATE versus LATELY), 3) vague, multiple conditioning of choice (CLOSE versus CLOSELY), 4) stylistic and register differentiation (SLOW versus SLOWLY). Register variation no doubt also occurs with adverbs in categories 1-3.

References Allerton, D. J. 1982 Valency and the English verb. London—New York: Academic Press. Bolinger, Dwight 1979 Meaning and form. London—New York: Longman. Bourcier, Georges 1981 An Introduction to the history of the English language. English adaptation by Cecily Clark. Cheltenham: Stanley Thornes.

186

Terttu Nevalainen

Brugman, Claudia 1984 "The very idea: A case study in polysemy and crosslexical generalizations", in: David Testen—Veena Mishra—Joseph Drogo (eds.), Papers from the Parasession on Lexical Semantics. Chicago: Chicago Linguistic Society, 21-38. Bullokar, William 1586 Pamphlet for grammar, in: The works of William Bullokar, Vol. 2, [1980] Pamphlet for grammar. (Leeds Texts and Monographs N. S. 1.) Facsimile Edition. Leeds: The University of Leeds, School for English. Campbell, Alistair 1959 Old English grammar. Oxford: Clarendon Press. Cooper, Christopher 1685 Grammatica Linguae Anglicanae. (English Linguistics 86.) Menston: [1968] The Scolar Press. Donner, Morton 1991 "Adverb form in Middle English", English Studies 72: 1-11. Ferris, Connor 1993 The Meaning of syntax. A study of the adjectives of English. London: Longman. Fettig, Adolf 1934 Die Gradadverbien im Mittelenglischen. (Anglistische Forschungen 79.) Heidelberg: Carl Winter. Finell, Anne 1992 "The repertoire of topic changers in personal, intimate letters: A diachronic study of Osborne and Woolf', in: Matti Rissanen—Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.), History of Englishes. New methods and interpretations in historical linguistics. (Topics in English Linguistics 10.) Berlin—New York: Mouton de Gruyter, 720-735. Greaves, Paul 1594 Grammatica Anglicana. London. Greenbaum, Sidney 1970 Verb-intensifier collocations in English. (Janua Linguarum, Series Minor 86.) The Hague—Paris: Mouton. Greenwood, James 1711 An essay towards a practical English grammar. (English Linguistics [1968] 128.) Menston: The Scolar Press. Guimier, Claude 1985 "On the origin of the suffix - / / ' , in: Jacek Fisiak (ed.), Historical semantics, historical word-formation. (Trends in Linguistics 29.) Berlin—New York—Amsterdam: Mouton Publishers, 155-170.

The processes of adverb derivation

187

Halliday, Μ. A. K. 1992 "Language as system and language as instance: The corpus as a theoretical construct", in: Jan Svartvik (ed.), Directions in corpus linguistics. (Trends in Linguistics 65.) Berlin—New York: Mouton de Gruyter, 61-77. HC = Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Hopper, Paul J. 1991 "On some principles of grammaticization", in: Elizabeth Closs Traugott—Bernd Heine (eds.), 17-35. Hopper, Paul J.—Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Huddleston, Rodney 1984 Introduction to the grammar of English. Cambridge: Cambridge University Press. Jespersen, Otto [1961] A modern English grammar on historical principles. Parts 3 and 7. [1954] London: Allen—Unwin. Jonson, Ben 1640 The English grammar. (English Linguistics 349.) Menston: The Sco[1972] lar Press. Killie, Kristin 1993 Early Modern English subject modifiers. (Tromso Studies in Linguistics 13.) Oslo: Novus. Kjellmer, Göran 1984 "Why great: greatly but not big: *bigly1 On the formation of English adverbs in -ly", Studia Linguistica 38/1: 1-19. Maittaire, Michael 1712 The English grammar. (English Linguistics 6.) Menston: The Scolar [1967] Press. Marchand, Hans 1969 The categories and types of Present-day English word-formation. (2nd edition.) München: C. H. Beck'sche Verlagsbuchhandlung. Matthews, Peter H. 1981 Syntax. Cambridge: Cambridge University Press. Mcintosh, Angus 1991 "Old English adjectives with derivative -lie partners: Some semantic problems", Neuphilologische Mitteilungen 92: 297-310. Mcintosh, Angus—M. L. Samuels—Michael Benskin 1986 A linguistic atlas of Late Mediaeval English. Vol. 1. Aberdeen: Aberdeen University Press.

188

Terttu Nevalainen

Miege, Guy 1688 The English grammar. (English Linguistics 152.) Menston: The Sco[1969] lar Press. Mitchell, Bruce 1985 Old English syntax. Vol. 1. Oxford: Clarendon Press. Mustanoja, Tauno F. 1960 A Middle English syntax. (Mdmoires de la Societe Ndophilologique 23.) Helsinki: Societe Niophilologique. Nevalainen, Terttu 1991 BUT. ONLY, JUST. Focusing adverbial change in Modern English 1500-1900. (Memoires de la Soci6te Neophilologique 51.) Helsinki: Societe N6ophilologique. 1994a "Aspects of adverbial change in Early Modern English", in: Dieter Kastovsky (ed.), Studies in Early Modern English. Berlin: Mouton de Gruyter, 243-259. 1994b "Diachronic issues in English adverb derivation", in: Udo Fries— Gunnel Tottie—Peter Schneider (eds.), Creating and using English language corpora. Amsterdam: Rodopi, 139-147. forthc. "Early Modern English lexis and semantics", in: Roger Lass (ed.), The Cambridge history of the English language, Vol. 3, Early Modern English 1476-1776. Cambridge: Cambridge University Press. Ohlander, Urban 1984 "Predicative complements and predicative attributes: A note on Middle English usage in coordinate structures", Neuphilologische Mitteilungen 85/4: 438-444. Opdahl, Lise 1990 "Close or closely as verb modifier? In search of explanatory parameters", in: Graham Caie—Kirsten Haastrup—Arnt Lykke Jakobsen—Jargen Erik Nielsen—Jorgen Sevaldsen—Henrik Specht—Arne Zettersten (eds.), Proceedings from the Fourth Nordic Conference for English Studies, Vol. 1. Copenhagen: Department of English, Copenhagen University, 119-131. 1992 "Sex and preference of adverbial suffix: Do women favour the prestigious form?", in: Anne-Marie Langvall Olsen—Aud Marit Simensen (eds.), Om spräk og utdanning, Essays in Honour of Eva Sivertsen. Oslo: Universitetsforlaget, 91-109. Palsgrave, John 1530 Lesclarcissement de la langue francoyse. (English Linguistics 190.) [1969] Menston: The Scolar Press. Peltola, Niilo 1969 "Contributions to the study of intensives", Neuphilologische Mitteilungen 70: 33-53.

The processes of adverb derivation

189

Peters, Hans 1993 Die englischen Gradadverbien der Kategorie booster. (Tübingen Beiträge zur Linguistik 380.) Tübingen: Narr. Poole, Joshua 1646 The English accidence. (English Linguistics 5.) Leeds: The Scolar [1967] Press. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman. Rissanen, Matti 1985 "Expression of exclusiveness in Old English and the development of the adverb only", in: Roger Eaton—Olga Fischer—Willem Koopman—Frederike van der Leek (eds.), Papers from the 4th International Conference on English Historical Linguistics. (Current Issues in Linguistic Theory 41.) Amsterdam—Philadelphia: Benjamins, 253267. Ross, Claudia N. 1984 "Adverbial change: Implications for a theory of lexical change", in: David Testen—Veena Mishra—Joseph Drogo (eds.), Papers from the Parasession on Lexical Semantics. Chicago: Chicago Linguistic Society, 243-249. Swan, Toril 1988 "The development of sentence adverbs in English", Studia Linguistica 42/1: 1-17. Traugott, Elizabeth 1988 "Is internal semantic-pragmatic reconstruction possible?", in: Caroline Duncan-Rose—Theo Vennemann (eds.), On language, rhetorica, phonologica, syntactica. A Festschrift for Robert P. Stockwell, London —New York: Routledge, 128-144. Traugott, Elizabeth Closs—Bernd Heine (eds.) 1991 Approaches to grammaticalization, Vol. 1. (Typological Studies in Language 19:1.) Amsterdam—Philadelphia: John Benjamins. Traugott, Elizabeth Closs—Bernd Heine 1991 "Introduction", in: Elizabeth Closs Traugott—Bernd Heine (eds.), 1 14. Ungerer, Friedrich 1988 Syntax der englischen Adverbialen. (Linguistische Arbeiten 215.) Tübingen: Niemeyer. Wallis, John 1653 Grammatica Linguae Anglicanae. Grammar of the English language, [1972] ed. and transl. by J. A. Kemp. London: Longman.

There compounds in the history of English Aune

1.

Österman

Introduction

There compounds are pronominal adverbs formed by compounding. The first element of the compound consists of the deictic adverb there and the second element of a prepositional adverb in, of, on, by, fore, etc. (Mitchell 1987: §1064), which defines the meaning of the first part of the compound more closely.1 This kind of adverb formation has been common to all Germanic languages; Present-day Swedish has e.g. därför 'therefore, for that reason', Danish derunder 'under that', Norwegian dermed 'with that', Icelandic par eftir 'after that', German daraus 'about that', Dutch daarbinnen 'inside that'. The above kind of adverb formation is regarded as adverbialization and this compounding process can be considered the secondary stage of grammaticalization. Most early compound adverbs were local, but in the long run they underwent semantic changes and developed more abstract meanings and were further grammaticalized. Compound adverbs represent the most grammaticalized end of the adverb category. Speaking about grammaticalization, Heine et al. (1991: 43) point out that "It is abstraction of the metaphorical type, rather than any other, that underlies grammaticalization." This seems to hold true with the grammaticalization of there compounds. They developed abstract meanings which were necessary for their being grammaticalized. Speaking about there compounds in Presentday English it seems that abstractness is the term that would characterize the compound adverbs most conveniently, e.g. therein is felt to be a more abstract spatial expression than the combination preposition + it/that/them/ those even on the most concrete local level. The pronoun it in the combination in it refers to an earlier-mentioned local expression in a more concrete way than there in therein.

192

Aune Österman

There compounds have existed in English from earliest times; the first mention in the OED dates from the year 888 (doer mid in King Alfred's Boethius), the first citation in the Helsinki Corpus dates from the earliest Old English subperiod (percet and peran 'thereon' in the 9th century documents). Due to the loss of inflexional endings and the loss of the instrumental case, synthetic forms came in very conveniently, on the one hand, to replace the earlier expressions formed by a preposition + a pronoun in different cases. On the other hand, such deictic adverbs as peer, pider and ponan coexisted along with there compounds occurring in similar functions. The present study is a continuation to my earlier study of there compounds in Old to Early Middle English (see Österman 1993) now extending up to the Early Modern English period and finally briefly tracing the Present-day English usage in spoken and written corpora. According to my previous study there compounds spread rather quickly in OE; the number of tokens per 10,000 words tripled from OE3 to OE4 and more than doubled from OE4 to ME1. The number of compound types in OE was highest in OE3 (23 types), diminished to OE4 (13 types) to rise again in ME1 (31 types) and remained at the same level in ME2. The most dramatic change was between ME1 and ME2 when a new compound type, therefore appeared in ME1 (with 6.5% of all the there compounds in its subperiod) and increased strongly in ME2 (to 22.3%). The present study tries to provide an answer to what happened after ME2, which was the last subperiod studied in my previous work. The present study shows that there compounds reached their peak in Middle English and started to decline from the ME4 subperiod onwards and have continued to decline throughout the Early Modern English period. They were grammaticalized and gradually regularized to the present level. Thereof and therefore especially, increased strongly. The circle was closed when the analytic forms began to appear in the Middle English period (for instance in it to replace therein in ME3 in the HC). The synthetic forms, especially those that did not have a long history behind them began to disappear but those that had existed in OE remained. It seems that this closed class adverbial formation is no longer productive. The latest there compounds to emerge in English according to the OED are thereward in 1922 in Joyce's Ulysses (not found in the Helsinki Corpus) and thereafterward in 1867, which, however, is found in the Helsinki Corpus as early

There compounds in the history of English

193

as ME2 in The Life of St. Edmund. The latest compounds to appear in the Helsinki Corpus are thereunto (ME4) and thereabouts (EModE2), of which the latter continued into Present-day English. Eight compounds that emerged in OE have continued their existence into EModE3 and seven of these into the present day. Three compounds that emerged in ME have continued into EME3 and two of them up to the present day. These and a few others still occur to some extent in Present-day English and are found slightly more often in written than in spoken language. The present study falls very convincingly in line with the theme of the present volume: enrichment and regularization. The whole of the ME period is the period of enrichment, regularization begins to be felt in the EModE period. The development of there compounds throughout the whole of the Helsinki Corpus will be covered. First some statistical data will be given concentrating on the entire corpus. Then some OE and ME developments will be discussed. After that the compounds that range from OE to EModE will be given a closer scrutiny concentrating on the semantics and variation within one single there compound and variation within several compounds. Finally, the remaining compounds will be discussed and a short survey of the use of there compounds in Present-day English based on the existing written and spoken corpora will be given. The starting point of the study is change through variation. Attention is also paid to grammaticalization as it is presented by e.g. Hopper— Traugott (1993), Traugott—Heine (1991), Heine—Claudi—Hiinnemeyer (1991) and Hopper (1991).

2.

There compounds from OE to EModE in the Helsinki Corpus

The data in the present study for OE1-ME2 are about the same as those in Österman (1993) (with only minor revisions),2 but they are included in the present study because they are relevant for the further analyses of there compounds from OE1 to EModE3. ME3 adds four new types of compounds, ME4 one and EModE2 another new compound. Thus the number of types of there compounds in the present study is 57. Table 1 shows the situation from OE to EModE.

194

Aune

Österman

co % ο

•ΊI

s w fN w Ό Ο

(Ν

Ο

υ1-1 »Hυ

£ o

00

η Ν

m (S

(Ν 00

σ\

ο (Ν

γο οο "" fN

ο

Γ^ fN

(Ν S ω 2

ο (Ν

Ο

CO

Ο 00 - od

ω Ο

rö

οο α\

fN w Ο

οο Γ-

^ ΟΟ νο Λ.

Ο Γe hony in a panne & scome it & do per-in J>e gres (MEX/4 The 'Liber de Diversis Medicinis' in the Thornton Manuscript 70) 'and put some honey in a pan and stir it and put the fat into it...''

Similarly, Virtanen (1992), discussing the discourse functions of local and temporal adverbials, shows that the placement of these adverbials in the clause-initial position is characteristic of certain text types, e.g. narration. The narrative text types in the present study, such as history, biography, fiction and romances, which very often deal with concrete localities, have developed their own conventions where the textually important locality is introduced first and then, in the course of narration, substituted by a there compound as for instance in: (17)

And than Seynt Edmonde made a chapell at Catysby, and after bothe his susters were buryed therein, ... (ME4 The Life of St. Edmund (in Middle English Religious Prose) 166) 'And then St Edmund made a chapel at Catsby and thereafter both his sisters were buried in it...'

The same convention is also used with abstract localities or abstract matters, often for instance in religious treatises:

There compounds in the history of English (18)

219

and he dwellis noght byfore godis lyght. for if he any tyme see oght what he sould doe, he dwellis noght tharin: for he lufes that thynge that turnys him fra god. (ME2/4 Rolle, The Psalter 19) 'and he does not live before God's light, for if he at any time sees anything he should do, he does not continue steadfastly in that matter, for he loves the thing that turns him away from God'.

Similarly, later laws, e.g. in EModE3, seem to have developed a convention where they repeatedly make use of the same construction with therein in an abstract meaning in 9 instances, e.g. (19)

... this Act or any thing therein contained (EModE3 The Statutes of the Realm VII 459)

They also use therein repeatedly in an increasingly more abstract meaning of 'in that (those) affair(s) or matter(s)' in 7 instances, e.g. (20)

... or upon Entry and Discovery of such Frauds cannot find out or discover the Person or Persons concerned therein. (EModE3 The Statutes of the Realm VII 457)

The last-mentioned meaning is highly characteristic of legal language in the EModE period. The first subperiod where English law texts occur in ME is ME4. Of all the instances of therein per subperiod, the share of legal language is about 8% in ME4, 8% in EModE 1, 26% in EModE2 and 76% in EModE3, and the share of therein compounds in the legal language per each subperiod is about 9% in ME4, 7% in EModE 1, 15% in EModE2 and 23% in EModE3. Thus it seems that the increased use of therein in law texts and its increased abstractness in them is an indication of therein having been grammaticalized rather early in formal usage. Finally, in one instance in the Helsinki Corpus therein approaches what might be interpreted as a relative adverb 'in which', 'where': (21)

Py feorJ)an daege he wass on Simones huse £>aes lic|)roweres, pterin geat f>«t wif J)a deorwyrjjan smerenesse on his heafod.12 (OE2/3 The Blickling Homilies 73) 'On the fourth day he was in the house of Simon the leper, where-in the woman poured out the precious ointment on his head.' (translation by the editor)

220

Aune Österman

Why then did thereinne, the spelling variant of therein, disappear in the ME period? When the synthetic forms began to give way to the analytic forms therein began to be replaced by 'in it', e.g. (22)

Tak a hate hauyre cake & lay it down & lay thyn ere {jer-on als hate als JJU thole it &, if {)er be schepe louse or any ojjer qwik thynge in it, it sail sone crepe owte. (MEX/4 The 'Liber de Diversis Medicinis' in the Thornton Manuscript 7) 'Take a hot oatcake, put it down and put your ear on it as hot as you can endure, if there should be a sheep louse or any other living thing in it it will soon creep out'.

The spelling variant thereinne, however, could not be replaced by 'inne it' (not a single example in the Helsinki Corpus). In other words: in in therein could be either an adverb or a preposition, inne in thereinne was originally an adverb only.13 This disadvantage may have meant that thereinne began to feel obsolete. Another reason for its disappearance could be that it was used as a spelling variant of therein14 and was felt to be unnecessarily long and heavy by the side of a shorter variant.

4.3. Thereto The first instances of thereto occur in OE2, six times (8%) in history with a directional concrete local meaning 'to it/that place, into it' and twice with an abstract meaning, once in philosophy, once in history, e.g. (23)

for f>on £>u, fasder Agustinus, hie haefst on J)inum bocum sweotole gesasd; & ic gehwam wille pcerto taecan J)e hiene his lyst ma to witanne. (OE2 Alfred's Orosius 102) 'because you, Father Augustinus, have spoken about them clearly in your books and I would like to direct everybody to them [to the books] who desires to know more about them.'

Thereto occurs in medicinal or remedy book recipes in subperiod OE3 in Lacnunga three times, in Medicina de Quadrupedibus five times, in ME1 eighteen times in Peri Didaxeon, in ME3 three times in A Late Middle English Treatise on Horses and once in Mandeville's Travels, and in

There compounds in the history of English

221

ME4 eleven times in the Thornton Manuscript and four times in chemical experiments in The Cyrurgie of Guy de Chauliac, and once in EModE2 in Clowes' Treatise for the Artificiall Cure of Struma with the meaning '[put] into it', e.g. (24)

nim ]?anne mintan & seoÖ hi on waetera & do peer to jehwaede wyn. (MEX/1 Peri Didaxeon 136) 'then take mint and seethe it in water, and add thereto a little wine.' (translation by the editor)

These are the only instances where thereto means 'into it'. Thereto can also have another, non-directional meaning, 'on it', in the recipe book texts, e.g. (25)

& a-non ley it upon Jje ded flesche. But let it not longe leye perto ... (ME3 A Late Middle English Treatise on Horses 123) 'and lay it quickly on the dead flesh. But do not let it lie long on if

Thereto can also mean directional 'to it', which in Present-day English would be interpreted as 'on/onto it', for instance in eleven of all the seventeen instances in A Late Middle English Treatise on Horses (ME3): (26)

take netus scheme & ley f>er-vpon. Or ellus take 3elkus of eyrin & herdes. & wete hem £>er-ynne & ley per-to (ME3 A Late Middle English Treatise on Horses 123) 'take dung of cattle and put it on it. Or else take yolks of eggs and hards, wet them in it and put them on if

In addition to the previous meanings thereto occurs both in a concrete local and in an abstract local meaning. In OE3 (the total number of occurrences 39), in ME1 (78 occ), and in ME2 (44 occ) the concrete meaning is more common but the abstract meaning is more prominent in OE4 (34 occ), in ME3 (71 occ) and in ME4 (63 occ). In the Late ME period abstract or figurative meanings seem to gain ground. The increased abstract usage, in its turn, can be regarded as an indication of grammaticalization. Of the following examples the first is local, the others abstract:

222

Aune Österman

(27)

'...foröan öu scalt bien 3efonded on Öa hali Hue al swo is J)e pott öe is idon on öe barnende ofne. Gif he öar inne bersteÖ and brekÖ, he is forloren and sone ut-ßeworpen; y f he belaefÖ hal and 3esund, Öe pottere hine deö öar to öe he iscapen was.' (ME1 Vices and Virtues 73) 'because you will be received in the holy life as is the pott that is placed in the burning oven. If it bursts and breaks in it, it is lost and soon thrown away, if it remains whole and uninjured the potterer puts it (to the place/there) where it was created for." 5

(28)

Some man settis )?e thoghte Ihesu, and stedfastly haldis Treatises 19) 'Some man sets the thought only and steadfastly holds it

(29)

of his herte anely in ]je name of it pare-too·, (ME2/4 Rolle, Prose of his heart in the name of Jesus therewith.'

that the processe of myn outelawerye was vnlawefully made and al that longeth ther to. (ME3 Petitions 198) 'that the process of making me an outlaw was unlawfully made and everything that belongs to it.'

In the following example the "official" translation by Garmonsway suggests that thereto is used in a local sense: (30)

On J)isum geare eac foröferde se arcebiscop Girard of Eoferwic toforan Pentecosten. & wearö syööan Thomas pcerto gesett. (OE4 Chronicle MS Ε 242) 'In this year too passed away archbishop Gerard of York before Whitsuntide, and Thomas was afterwards appointed to succeed him there.' (translation by Garmonsway)

There are nine (27%) similar examples in OE4 Chronicle MS Ε and one in ME2 The Life of St. Edmund (South-English Legendary) where thereto should be interpreted as both concrete and abstract. Besides a concrete place, all these examples include the idea of a high office to which a person succeeds/is appointed/consecrated (gehalgod), for instance an archbishop, bishop, abbot etc. Thus thereto is an excellent, precise means of suggesting both meanings in OE. In the above example Garmonsway's

There compounds in the history of English

223

translation also suggests both meanings, dividing them between the abstract succeed (to the office) and the local there. Thereto also occurs with the meaning 'in addition to that' (OE3 3 occ, OE4 4 occ, ME1 3 occ, ME2 15 occ (34%), ME3 14 occ (19%), ME4 1 occ). This meaning is especially favoured by ME2 Robert of Gloucester (12 of all the 15 occurrences), by ME3 The Cloud of Unknowing (8 out of 12) and by ME3 Chaucer The Canterbury Tales, The General Prologue (6 out of 7) (31 a-b), 'concerning that' (ME2 4 occ, ME3 3 occ, ME4 10 occ (16%)) (32), 'for that matter/purpose' (ME1 2 occ, ME2 1 occ, ME3 6 occ, ME4 6 occ) (33): (31) a.

Pe marcheis he adde sone . as sir Roger ]3e Mortimer . & sir warin of bassingbourne ... Sir roger of leibourne . & mani oJ)er per to . (ME2 Robert of Gloucester 739) 'He soon added the marquises like Sir Roger of Mortimer, and Sir Warin of Β. ... and many others in addition to these.''

b.

f»is sorow, when it is had, clensi{) Jje soule, not only of synne, bot also of peyne jjat he ha|3 deseruid for synne. & perto it makij) a soule abil to resseiue J)at ioye, (ME3 The Cloud of Unknowing 84) 'This sorrow, when it is had, purifies the soul not only of sin but also of pain that he has deserved for sin. And, in addition to that, it makes a soul able to receive that joy,'

(32)

and said these wordes: I didde hit. what wolle ye sey ther to: (ME4 Depositions 297) '... I did it. What will you say concerning thatV

(33)

Bot parto moght {jai neuer win. (ME2/3 Cursor Mundi 444) 'Remedy for that (matter) you can never obtain.'

In these abstract meanings thereto can be regarded as grammaticalized in this function. In a few scattered instances thereto occurs with adjectives or past participles which require a preposition, e.g. in ME2: nojt worpi perto 'not worthy of it' (The Life of St. Edmund (South-English Legendary) 438), η is per non perto iliche 'similar to that' (Beues of Hamtoun 101), ME2/4: contrayrie thare-to 'contrary to that' (Rolle, Prose Treatises 14), disposed

224

Aune Österman

ther-to (Rolle, Prose Treatises 26), EModEl: [I am] vsed therto (Udall, Nicholas, Roister Deister 1. 330). In some instances thereto occurs in connection with Type I prepositional verbs (see note 9), e.g. in ME2: hopiep per-to 'hope for it' (Song of the Husbandman 7), ensentede parto 'consented to it' (The Life of St. Edmund (South-English Legendary) 434), ME3: attende ther-to (Proclamations 32), ME2/3: Par-to Pers granted noght 'P. did not consent to it' (Mannyng, Handlyng Synne 190), ME4: [nor] consentyd yerto {The Life of St. Edmund (in Middle English Religious Prose) 165), EModE2: agreed there-to (Coverte, A Trve and Almost Incredible Report of an Englishman 11). The number of occurrences of thereto in EModE is rather small and they occur in much the similar instances as those in earlier subperiods. To sum up a few important developments in the use of the three previous there compounds therein, thereon and thereto, the following can be observed: excluding the first occurrence of thereon in OE1 with the meaning 'on that occasion', the earliest occurrences of thereon in OE2 (40% of all the there compounds in OE2) have the meaning 'in it', 'into it', 'on/ onto it'. Thereon undergoes a semantic change or rather a narrowing of its meaning 'in/into it', 'on/onto it' to 'on/onto it' some time after the early ME period. Medicinal or remedy book recipes show interesting diachronic variation: the meaning '[put] into it' is expressed by thereon in Lceceboc in OE2, in The 'Liber de Diversis Medicinis' in the Thornton Manuscript in ME4 by therein, and thereto expresses the same idea in Lacnunga and Medicina de Quadrupedibus in OE3, in Peri Didaxeon in ME1, in A Late Middle English Treatise on Horses and Mandeville's Travels in ME3 and in The 'Liber de Diversis Medicinis' in the Thornton Manuscript and in The Cyrurgie of Guy de Chauliac in ME4. Synchronic variation is seen in The 'Liber de Diversis Medicinis' in the Thornton Manuscript in ME4, where therein and thereto occur with exactly the same meaning 'into it'. Further grammaticalization of the above three there compounds is most often observed with thereto, developing such meanings as 'in addition to that', 'concerning that', 'for that matter' and to some degree with therein in some special formal expressions in law texts. It should also be noted that the three above mentioned compounds thereon, therein and thereto could replace or be replaced or used in the same meaning as the deictic adverb peer and its directional form pider, or by the compound adverb piderin, e.g.

There compounds in the history of English

225

(34)

Da jje munecas of Burch hit herdon saegen. J)a wasron hi swa sari swa hi nasfre asr ne wasron. forjsi f>t he waes swiöe god & softe man. & dyde mycel to gode wiöinnan & wiöutan. J)a hwile t>e he peer wunode. (OE4 Chronicle MS Ε 246) 'When the monks of Peterborough heard the news, thay had never been so overcome with grief, for he was a very good and kind man, and did much good both inside and outside the monastery whilst he was there.' (translation by Garmonsway)

(35)

pa com Swegen eorl pider, (OE3/4 Chronicle MS Ε 168)

(36)

& Eadward cyng sende pider Hereman biscop. (OE3/4 Chronicle MS Ε 171)

(37)

& sasdon jwn kinge {>t he haefde swyöe agylt wiÖ Crist f)t he asfre sceolde niman senig J?ing of Cristes cyrcean £>e his foragengceon dydon piderinn. (OE3 Documents 3 176) 'and they said to the king that he had greatly sinned against Christ in taking anything from Christchurch which his predecessors had assigned to it.' (translation by the editor)

4.4.

Therefore

Therefore is a compound adverb that has the largest number of occurrences (1253) in the Helsinki Corpus (see Table 6 in section 4). Its share of all the there compounds in the whole corpus (3605 occurrences) is 34.7%, or of all the subperiods where it occurs 38.2% (3273 occurrences). It is a high-frequency word: 22.3% in ME2, 35.5% in ME4, 41.2% in EModEl and in three subperiods, ME3, EModE2 and EModE3, therefore compounds form a half or more of all the there compounds in their respective subperiods. It emerges as late as the ME1 subperiod to replace such OE expressions as e.g. fordcem, forpy, forpi, fordon, forpan (also written in two words) in the meaning of 'for that/for that purpose/for that reason/therefore/as a consequence of that/consequently'. Therefore spread quite rapidly in the language but the old forms still existed side by side with the new compound adverb throughout the ME period (no examples with the older forms were any longer found in EModEl in the Helsinki Corpus), e.g.

226

Aune Österman

(38)

Piss sefennfald godle33C J>att Crist Uss dide J)urrh hiss are, Forrpi birr{) all Crisstene folic Goddspelless lare fol^henn. & tcerfore hafe icc turrnedd itt Inntill Ennglisshe sprasche, Forr patt I wollde blijjeliß Patt all Ennglisshe lede WiJjJ) aere shollde lisstenn itt, (ME1 The Ormulum 1. 285) 'This sevenfold benefit that Christ made us through his mercy. Therefore all Christian people should follow the preaching of his gospel. And therefore I have translated it into the English language, therefore 1 would willingly all English people listen to it with reverence,'

(39)

For therfore is suffisaunce requerid, for it is demyd to ben good; and forthy is power requirid, (ME3 Chaucer, Boethius 433) 'For therefore sufficiency is required because it is deemed to be good and for that reason power is required'

The earliest occurrences of therefore with the meanings 'for that, for that purpose, for that reason, therefore' are recorded by the OED c. 1175 in Lambeth Homilies. The meaning 'in consequence of that, consequently, as a result or inference from what has been stated' is recorded c. 1400 by the OED. All these meanings can be found with therefore in the Helsinki Corpus from the beginning of the ME period. As can be seen in the instances below, therefore can occur in its different senses either alone or in combination with 'for that reason'. In most cases, however, the meaning 'therefore/for this reason' can alone be considered by far the most common meaning of therefore. The meaning, 'for that', occurs in all subperiods, but its occurrences are very few: about 15% (4 occ) in ME1 and 16% (10 occ) in ME2, 6% (20 occ) in ME3, 10% (23 occ) in ME4, 2% (4 occ) in EModEl, less than 1% (2 occ) in EModE2 and none in EModE3.16 The following examples illustrate the use of therefore with different meanings: 'for that reason' (40a) often occurring in connection with admonition, (40b) very often occurring with and at the beginning of the formula to express reason And therefore, often also occurring with the verb

There compounds in the history of English

227

'say' introducing an authority to enhance the message of the text, e.g. in Chaucer where 13 cases out of 17 show this formula in The Tale of Melibee: (40) a.

(given as an example by the OED): and ec sette us bisne t>a3 habbe wele to ouer stohwennese on jjisse Hue ne beo J)u pereuore prud ne wilde ...(MEX/1 Lambeth Homilies 5) 'and moreover set us example, that when we have wealth in abundance in this life be ye not therefore proud, nor wild (elated),' (translation by the editor)

b.

'...for overmuchel reste norisseth and causeth manye vices.' And therfore seith Seint Jerome, 'Dooth somme goode dedes that the devel, which is oure enemy, ne fynde yow nat unocupied.' (ME3 Chaucer, The Canterbury Tales. The Tale of Melibee 233.C2) 'for excessive rest nourishes and causes many vices.' And for that reason St Jerome says: 'Do some good deeds so that the devil, who is our enemy, would not find you unoccupied.'

'as a consequence of that, consequently' (41a), 'so, thus' (41b): (41) a.

I concluded there must be neer eleven hundred of them, or somewhat more then a thousand in the length of an Inch, and therefore in a square Inch above a Million, (EModE3 Hooke, Micrographia 13 5 114)

b.

... despiseth not them, but me, saith Christ. If any man therefore doth loue the Lord Iesus ... (EModE2 Hooker, Two Sermons upon Part of S. Judes Epistle 8) 'Thus/So, if any man loves Jesus...'

'for that, as a recompense for that': (42)

the saide crafte of Surgeons have small cooning, yet they wooll take greate soomes of money and doo litle therfore. (EModEl The Statutes of the Realm III 906)

228

Aune

(43)

Österman

I pray yow b[y] for Alyson Myhell a mantell of fyn blake schankys, for I haue mony therfor, (ME4 Richard Cely, The Cely Letters 107) Ί ask you to buy A. M. a cloak of fine black fur for I have money for it.'

'for that purpose'; this usage is very often found with scientific exposition where an explanation is given for the purpose for which something is done. For instance four of the 15 instances in Record, The First Principles of Geometrie show this usage: (44)

I must make a perpendicular line to light in the pricke assigned, which is A. Therfore doo I set one foote of the compas in A, (EModEl Record, The First Principles of Geometrie C3V)

Sometimes it seems to be impossible to distinguish on the one hand between reason and consequence (of which the OED also admits (s.v. therefore II 2) that in early use they are often indistinguishable), and on the other hand between reason and purpose. The shades of meaning can be of equal strength as in (45a) or one of the shades of meaning can be stronger as in (45b) and (45c): (45) a.

& ouer sae fusen. ihiren of Cordoille. wat beon hire wille; Hire seohÖe word ich nam to grame, par-fore ich habbe nu muchele scame; (ME1 Layamon's Brut I 90) 'and travel over the sea to hear about Cordelia what her will is. I became angry with her true words. For that reason/as a consequence of that I now have great shame.'

b.

I see thou wilt answer nothing ingenuously, therefore I will trouble myself no more with thee: (EModE3 State Trials, The Trial of the Lady Alice Lisle IV 122.C1) ('therefore/for this reason)/consequently/thus/so'

c.

it must be borne with wisdome and patience as an heavy burden. Patience shall make it much more light. And therefore every one is to doe his best indeavour, (EModE2 Brinsley, Ludus Literarius, or the Grammar Schoole 13) ('therefore/for this reason/for that purpose')

There compounds in the history of English

229

Educational treatises, especially those by Bacon and Brinsley in the subperiod EModE2 with their rich and lively treatment of the subject matter, are especially interesting in the sense that their texts seem simultaneously to include different possibilities of interpretation in their use of therefore. E.g. with the meaning 'for that reason' somewhere in the background, therefore is used in the sense (to quote Collins Cobuild Dictionary) 'to introduce a conclusion that you are about to make', e.g. (46)

how to proceede in that easie and playing kinde. Therefore if you be satisfied in this, let us come unto the next point. (EModE2 Brinsley, Ludus Literarius, or the Grammar Schoole 12)

Along with therein, therefore also occurs in law texts from ME4 to EModE3 in formulaic expressions like Be it therefore ordeyned/enacted fourteen times out of a total of 40 instances. Table 7 (see p. 230) shows the distribution of the occurrences of therefore in the prototypical text categories in different subperiods. It shows that religious instruction is very much favoured by therefore in ME1 (73%) and ME2 (62%), e.g. Dan Michel, Ayenbite of Inwyt in ME2 includes more than half of all the instances of therefore, in ME3 The Cloud of Unknowing (14%) and Wycliffe and Purvey, The New Testament (John) (31%) and Wycliffe, The Old Testament (11%). In addition to these, The Book of Margery Kempe (11%) in ME4, and Smith's sermons (14%) in EModE2 show a special fondness for therefore. These figures show clearly that it is the text type and the subject matter of the text itself that is decisive in its demand and use of a certain there compound. For instance, religious texts that are concerned with hortatory advice, and admonitions how to live and behave as a good Christian, make use of therefore in explaining the reason 'why'. In the same way therefore also occurs in certain imaginative and non-imaginative narrative texts (about 27% in ME1 and ME2, about 15% in ME4 and EModE3, about 10% in EModEl and EModE3). Therefore is also used in trials in EModEl (12%), in secular instruction (handbooks and education) in EModE3 (about 23%), in expository texts (especially in educational treatises) in EModE2 (about 18%), in philosophy in EModE3 (about 13%) and in mystery plays in ME4 (about 20%).

230

Aune

Österman

-—\ ^—ν CI fN C) oo u-i fN vi Ö rn ,—t i-H fN m 1 r- VO Ο T-H 1 oo fN

m W •α ο s ω

00 rn

fN ω Ό a s ω

Ο 00

^

T-H c> ci oo 00 ov VO

N fN r- — I< 1 < fN •"Τ

1—