English in Transition: Corpus-based Studies in Linguistic Variation and Genre Styles [Reprint 2011 ed.] 9783110811148, 9783110156324

Citation preview

English in Transition

W G DE

Topics in English Linguistics 23

Editor

Herman Wekker

Mouton de Gruyter Berlin · New York

English in Transition Corpus-based Studies in Linguistic Variation and Genre Styles Edited by

Matti Rissanen Merja Kytö Kirsi Heikkonen

Mouton de Gruyter Berlin · New York

1997

M o u t o n 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 A N S I to ensure permanence and durability.

Library of Congress Cataloging-in-Publication

Data

English in transition : corpus-based studies in linguistic variation and genre styles / edited by Matti Rissanen, Merja Kytö, Kirsi Heikkonen. p. cm. — (Topics in English linguistics ; 23) Includes bibliographical references and index. ISBN 3-11-015632-6 (alk. paper) 1. English language - Early modern, 1500—1700 — Discourse analysis. 2. English language — Early modern, 1500—1700 — Variation. 3. English language - Early modern, 1500-1700 - Style, 4. Literary form. I. Rissanen, Matti. II. Kytö, Merja. III. Heikkonen, Kirsi, 1962— IV. Series. PE881.E54 1997 420M '41 —dc21 97-25407 CIP

Die Deutsche Bibliothek

— Cataloging-in-Publication

Data

English in transition : corpus based studies in linguistic variation and genre styles / ed. by Matti Rissanen ... - Berlin ; New York : M o u t o n de Gruyter, 1997 (Topics in English linguistics ; 23) ISBN 3-11-015632-6 D B N : 95.102774.3© 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 form 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, English in transition: Corpus-based studies in linguistic variation and genre styles and Grammaticalization at work: Studies of long-term developments in English. 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 some central verb constructions (with be and have), expository apposition, and genre-specific features of expressions of affect and attitude in text. The Grammaticalizaton at work volume sheds light on the development of adverbs and indefinite pronouns and on the means of reflexivization, in relation to various grammaticalization processes. 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 Pyfouton 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, Matti Kilpiö, Merja Kytö, Anneli Meurman-Solin, Saara Nevanlinna, Päivi Pahta and Irma Taavitsainen

1

Be/have + past participle: The choice of the auxiliary with intransitives from Late Middle to Modern English Merja Kytö

17

On the forms and functions of the verb be from Old to Modern English Matti Kilpiö

87

Re-phrasing in Early English: The use of expository apposition with an explicit marker from 1350 to 1710 Päivi Pahta and Saara Nevanlinna

121

Genre conventions: Personal affect in fiction and non-fiction in Early Modern English Irma Taavitsainen

185

Towards reconstructing a grammar of point of view: Textual roles of adjectives and open-class adverbs in Early Modern English Anneli Meurman-Solin

267

Bibliography Index of subjects

345 365

List of abbreviations

The following parameter codes included in the Helsinki Corpus appear in the present volume as such or in abbreviated versions (see also Bibliography at the end of the volume).

Prototypical text category: EX IR IS IS/EX NI NN STA

= = = = = = =

'expository' 'instruction religious' 'instruction secular' 'instruction secular'/'expository' 'narration imaginative' 'narration non-imaginative' 'statutory'

= = = = = = = = = = = = = = =

'biography, autobiography' 'Bible' 'biography, life of a saint' 'biography, other' 'drama, comedy' 'correspondence, non-private' 'correspondence, private' 'proceeding, deposition' 'diary' 'document' 'educational treatise' 'fiction' 'geography' 'handbook, astronomy' 'handbook, medicine'

Text type: BIA BIBLE BIL BIO COME CORO CORP DEPO DIARY DOC EDUC FICT GEO HANDA HANDM

χ

Abbreviations

HANDO HIST HOM LAW LET PRIV LET NON-PRIV MYST NEWT OLDT PHILO PREF RELT ROM RULE SCI A SCIM SCIO SERM TRAV TRI

= = = = = = = = = = = = = = = = = = = =

'handbook, other' 'history' 'homily' 'law' 'letter, private' 'letter, non-private' 'drama, mystery play' 'New Testament' O l d Testament' 'philosophy' 'preface' or 'epilogue' 'religious treatise' 'romance' 'rule' 'science, astronomy' 'science, medicine' 'science other' 'sermon' 'travelogue' 'proceeding, trial'

= = = = = =

'audience, professional' 'audience, non-professional' 'interaction', 'interactive' 'informal setting' 'formal setting' 'unspecified'

Other: PROF NON-PROF INT INF FOR Χ, XX

Introduction Matti Rissanen, Matti Kilpiö, Merja Kytö, Anneli Meurman-Solin, Saara Nevanlinna, Päivi Pahta and Irma Taavitsainen

1.

Methodological considerations: linguistics and philology in interaction

The last two decades have seen a rapid increase and methodological development in the studies of variation in language. The basic assumption in these studies is "orderly heterogeneity", i.e. variability which is not random but affected by linguistic and extralinguistic factors or constraints (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). Within this framework, it has been our aim to establish how linguistic variation is patterned not only socially, regionally and temporally, but according to genres defined by extralinguistic criteria. When the variationist approach is adapted to discourse studies, comparisons of text types defined by their linguistic properties become the key (Schiffrin 1994: 314); the last two chapters of this book extend the application to historical stylistics with the aim of charting genre styles and genre conventions. In diachronic studies, the variationist approach provides us with a good opportunity to observe the actual process of change. We can trace the birth and death of variant expressions, but perhaps more interestingly, their changing frequencies and distributions within a variant field at subsequent periods of time and in various genres, and we can analyse changes in the intricate mesh of linguistic and extralinguistic factors conditioning the occurrence of these variants. Within this approach our philo-

2

Matti Rissanen et al.

logical training has been a great asset as it emphasizes the context of expressions and the context of culture in interpreting them. The importance of extralinguistic factors in the analysis of development and change has necessitated an intensive study and discussion of genre typologies from the point of view of the historical study of language; cf. e.g. Douglas Biber's and Edward Finegan's studies of the 'dimensions' characterizing texts and offering a new basis for their grouping. In the structure of text corpora, the labelling of genres provides a general framework for the discussion of text-related changes at different periods of time. However, the varying importance of conventions or innovative pressures in the evolution of each genre or group of genres may decrease the usefulness of such classifications, as genres are also internally heterogeneous, and the pace and direction of change may be different in individual texts representing a particular genre. A comparison of the occurrences and frequencies of variant expressions in different texts allows the reconstruction of the various levels of past expression: written and speech-based, literary and non-literary, formal and informal, etc. This method is also necessary for all attempts to describe the relationship of the standard(s) to regional or social dialects. Of the five chapters included in this volume, three discuss morphological, syntactic or lexical questions with special attention paid to variation relating to text type, dialect etc., while two concentrate on the cooccurrence patterns of linguistic features in various types of texts and on questions of genre classification, genre markers and distinguishing features. The main focus is on long-span diachrony, 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 long time span and the wealth of primary data set specific demands for the grammatical models used in this volume. 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 the very wide range of morphosyntactic issues involved in classifying linguistic features in genre studies. 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 type of grammar provides an adequate basis for analysis at a relatively low level of abstraction, enabling the researcher to deal even with the more problematic borderline cases. On the other hand, it is clear that, as a grammar of Present-day English, Quirk et al. cannot be directly applied to different historical phases of

Introduction

3

English. In our case, it has provided the basis which the writers of this book have employed in different ways and to different degrees. We have made use of various approaches, from traditional grammar to semantic, pragmatic and textual theories.

2.

New openings offered by the Helsinki Corpus and other computerized material

Scholars working on variationist studies benefit from having access to computerized collections of texts. Computerized diachronic corpora, increasingly available in international distribution, make it relatively easy to collect evidence of the occurrence of variant expressions; they also encourage the researcher to approach topics that would earlier have meant an unreasonable amount of time-consuming routine work. The studies reported in this volume are based on the Helsinki Corpus of English Texts, the first large computerized corpus to cover the timespan 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 offers reliable indicators concerning the structural and lexical developments of English for over a millennium. In many cases, the results are tentative and 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. 1 Each text or group of related texts of the Helsinki Corpus is equipped with a set of parameter values containing information on the text and its author, if known. In Old and Early Middle English, this information is concentrated 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, and so forth. 2 In these periods, sociolinguistic information is given on the authors of the texts (their rank, sex, and age) and, in the case of letters, on the relationship between the writer and the receiver (intimate family members are distinguished from more distant addressees; addressees superior to authors are distinguished from those inferior to them).

4

Matti Rissanen

et al.

A large and structured corpus such as the Helsinki Corpus, equipped with textual parameter codings, has made it possible for us to combine effectively qualitative and quantitative analysis through variation. This approach has been extensively used in the analysis and interpretation of the data (cf. McEnery—Wilson 1996: 62-63). Our background knowledge of texts, also capitalized on in the compilation of the Helsinki Corpus, has been utilized to place our observations of linguistic phenomena within a larger framework. We have been able to reanalyse and re-evaluate earlier genre and text-type classifications of the older periods of English and to test the validity of the suggestions made concerning the co-occurrence of linguistic features with various types of text. The parameter coding and extensive textual basis have also allowed us to observe the role played by dialectal distribution and the differences between prose and verse texts in the survey of Old and Middle English variant forms. In the discussion of Late Middle and Early Modern evidence, emphasis has been given to the distributions shown by speech-based texts, or texts showing a relatively high degree of orality. In a few cases, it has been possible to draw conclusions from sociolinguistic variables: the genre, degree of interactiveness, level of formality, sex, age, rank or education of the author, or the relationship between the sender and receiver of letters.3 The insights derived from recent trends in sociolinguistics have added a new angle to the discussion of these factors (cf. e.g. Romaine 1982; Milroy—Milroy 1985; Milroy 1992). We believe that, among these variables, genre is the most complex one, as the varying types of genre can be claimed to derive from a combination of other variables such as topic, audience, setting, text category, etc. The corpus-based framework lends itself to various statistical applications by which it is possible to verify the reliability of the results obtained. In addition to traditional frequency surveys, a number of more advanced statistical analyses have been carried out for the purposes of some studies included in the volume. Thus Merja Kytö in her study on the be/have variation uses logistic regression analysis to assess the impact of various factors on the use of the variant forms. This analysis tests out how statistical models based on various combinations of factors account for variation in the data and assess the significance of the individual factors and their interaction. Several statistical methods are combined in Irma Taavitsainen's chapter on personal affect and genre styles. Factor analysis is applied first to identify adjoining text types and possible texttype markers. This method serves to indicate the overall patterns of genre styles and how they relate to one another. Then t-tests and f-tests are used

Introduction

5

to evaluate the significance of the features in telling fiction apart from the adjoining genres. The combination of these methods yields results which are then assessed in a larger sociohistorical context. In the chapter by Anneli Meurman-Solin on the concept of point of view in texts, the different informative value of mean frequencies and percentages is stressed and the findings are presented by focusing on one feature or factor at a time and, after a detailed analysis of this kind, by clustering them by both syntactic and semantic criteria. This adds to the reliability of data and also allows the mapping of significant correlations in a network-like pattern of dimensions that usefully describe focal features in genre styles and text types.

3.

Variation on the level of morphology, syntax and lexis: the verbs be and have

The studies by Merja Kytö and Matti Kilpiö in this volume focus on the use and development of the verbs be and have, which have played a vital role in the shaping of the English verb phrase. The paths of be and have are parallel, both having occurred as lexical and auxiliary verbs providing variant expressions in the auxiliary function (e.g. compound tenses with mutative verbs, and expression of obligation). Owing to complexities in the development of these verbs, there is no consensus about the role played by such crucial processes as grammaticalization. With the verb be, for instance, the status of the verb in progressive forms has clearly changed from the copula in Old English into an auxiliary from Middle English onwards. The uses of be and have grammaticalize at different stages in the history of English, and the two studies only concentrate on certain key periods in these developments. The study of the be/have variation with mutative intransitives examines the process by which have finally supersedes be in present and past perfect constructions; the study on be focuses on the forms and functions of the verb, with an eye on the developments in its functional load. The former study covers the period from Late Middle to Modern English, the latter from Old to Modern English. Among the extralinguistic factors included in the two studies are chronology, region and dialect, and foreign influence; moreover, with the be/have variation, such factors as text type, relationship of the text to spoken language, level of formality, orality and

6

Matti Rissanen et al.

authorial properties are taken into account. Among the linguistic factors observed in both studies are tense and certain verb constructions (finite/ non-finite forms, -ing constructions). Furthermore, with the study of the be/have variation, attention is paid to the status of the verb (stative and mutative; action and process; frequent and rarer verbs), durative, iterative and conditional contexts, negation, and object-like and other complements. With be, additional factors include developments in the morphology of be, the function of the verb (its use as an auxiliary or a lexical verb, copular or non-copular), and semantic-syntactic functions of ft-forms as against non-0-forms (for Old English). These studies have brought up new evidence pinning down trends of development in greater detail than found in previous research, thanks to the combination of quantitative and qualitative corpus linguistics and philological assessments. With the be/have variation, the use of have is shown to increase gradually from the Late Middle to the Early Modern English period, gain in momentum in the late 1700s and supersede be in the early 1800s. The more powerful extralinguistic factors influencing the process of change include chronology and text type, and a number of linguistic factors (relationship to tense, aspect, complementation etc.). In the study dealing with be, the most striking morphological developments are the rapid disappearance in the Early Middle English period of the coexistent Old English present tense paradigms (beon/wesan), the diversification of forms in Middle English and the subsequent process of simplification and regularization as Early Modern English is reached. In the survey of the main functions of be, the remarkable stability seen in the relative share of the three main functions of be over the centuries is the most important finding. Within the auxiliary category, the predominance of the use of be as a passive auxiliary is the most noticeable feature. These two studies have shown that the Helsinki Corpus, supplemented by other diachronic corpora and other standard reference works (LALME, MED, OED), offers a solid basis for the empirical approach aimed at accounting for variation and change.

Introduction

7

4. Variation in re-phrasing: expository apposition across the centuries The chapter by Päivi Pahta and Saara Nevanlinna forms a bridge between the studies concentrating on the syntactic and morphological developments described above and those discussing the characteristics of genres and text types. It examines characteristics of expository apposition — the grammatical category connected with re-phrasing. Re-phrasing as a communicative phenomenon occurs in both written and spoken media, and in both planned and unplanned discourse. It can generally be analysed as the writer's or speaker's attempt to reformulate an utterance in order to achieve successful communication. On closer inspection, the decision to re-phrase may be based on various factors, including stylistic and didactic considerations, the author's assessment of the addressee's ability to process given information, or the author's wish to add to the flow of discourse by providing more information about the topic of discussion. This chapter focuses on the development and use of expository apposition with an explicit marker in Late Middle and Early Modern English. Apposition is seen as a broad notional category containing both nominal and non-nominal phrases, clauses and sentences. There are no previous detailed studies of apposition in this period, and the theoretical framework adopted as the starting-point in this study is the recent discussion of apposition in Present-day English by Meyer (1987 and 1992), where apposition is seen as a syntactic, semantic and pragmatic relation. The main emphasis in Pahta and Nevanlinna's study is on the semantic characteristics of appositional constructions and their distribution across different text types. Attention is also paid to the devices used in linking appositional units, i.e. explicit markers of expository apposition. The study shows that the use of expository apposition links up with some of the most central lexical phenomena of the Middle English and Early Modern English periods, such as dialectal variation and the adoption and accommodation of loan-words. It also indicates a clear tendency for some text types to favour the use of appositional constructions in general, and certain semantic and syntactic types in particular. This is so throughout the period, although there is internal variation within most text

8

Matti Rissanen et al.

types. Most of the markers of expository apposition available in Presentday English were found to exist even in the late medieval period, with many others which have since gone out of use. In this respect, the results obtained point to a difference in the use of coordinative apposition, particularly binomial constructions with the markers and and or. In the course of the analysis, the use of traditional philological tools (knowledge of textual background, cultural status of texts, etc.) proved helpful. With certain limitations, the corpus-based approach offered a fruitful way of collecting data for the study of appositional constructions. Considering the open-class nature of apposition as a linguistic phenomenon, the results obtained in this study show the way for further work on the topic.

5. Expressions of personal affect and stance marking: identifying genre-specific choices The last two chapters of this volume, by Irma Taavitsainen and Anneli Meurman-Solin, have a somewhat different problem-setting, but the approach combining the quantitative and qualitative methods applies here as well. Their main topic is identifying genre-specific features in the linguistic choices related to participant roles by analysing expressions of affect and attitude in texts. Taavitsainen discusses the use of personal pronouns, exclamations, direct questions and other expressions of personal affect, while Meurman-Solin's study focuses on the frequencies and distributions of adjectives and open-class adverbs as stance markers. In recent years genres have been looked at from many different perspectives. Besides thorough comparative studies of features of individual genres (such as fiction in Fludernik 1993 and 1996) or a wide range of genres in a particular time period (such as Renaissance genres in Lewalski 1986), we find the interdisciplinary approach (for example in Sell and Verdonk 1994) and the important advances in discourse analysis (Coulthard 1994) particularly relevant. Taavitsainen and Meurman-Solin approach the problems of genre studies from a variationist's and sociolinguist's point of view. Rather than restrict the focus to the dimension of written genres as against genres reflecting usages more typical of spoken language, they aim at pointing out clusters of features which position texts on a number of other dimensions, particularly those which reflect

Introduction

9

idiolectal or genre-specific characteristics of participant roles, or focus on genre markers. Traditional genre labels such as 'history', 'private letter', 'autobiography' or 'sermon' are used in the majority of recently produced text corpora. A typology of this kind has also been adopted as a working tool in the Helsinki Corpus and its supplement, the Helsinki Corpus of Older Scots (see note 1). The labelling in these corpora is based on extralinguistic factors such as the social and communicative function of texts or their subject matter. In a number of studies based on the Helsinki Corpus and/ or its supplement of Scots (see the bibliographies in the chapters by Taavitsainen and Meurman-Solin in this volume), medieval and Renaissance prose genres have been shown to be linguistically relatively heterogeneous. This heterogeneity is partly due to the compilers' decision to polarize the samples: they have intentionally selected the representatives of a genre from different stylistic traditions when such stylistic variation has been established in earlier research (Nevalainen—Raumolin-Brunberg 1989: 99). This should always be taken into consideration when generalizing from the results. Patterns of co-occurring features, illustrated in Taavitsainen's and Meurman-Solin's studies, provide evidence for a classification of texts into text types. Each text type may thus comprise texts which represent different genres; in addition, intertextuality phenomena between genres complicate the issues. Biber's pioneering work (1988) in corpus-based stylistics with its multifeature and multidimensional statistical assessments is strictly linguistic; our innovation is the firm philological anchoring, limiting the comparisons to texts that share common features and that belong to related genres (Taavitsainen) and recategorizing multifunctional linguistic features by means of a thorough analysis of their varying syntactic and semantic properties in different time periods (Meurman-Solin). We have also profited from other studies which tackle related questions. The two studies aim at making it applicable to the analysis of early prose texts by selecting features other than those in Biber's factors, and by introducing a more detailed semantic subcategorization of a more comprehensive set of realizations of some features included in his factor analysis, and by developing the research tasks for different aims. Because of the emphasis on semantic features, the relevant examples are carefully selected by qualitative reading and analysed in the wider context of running text. Both studies thus highlight the importance of combining the quantitative approach of corpus linguistics with a detailed analysis of discourse function and meaning, central in the philological tradition.

10

Matti Rissanen et al.

To compensate for the low number of texts per genre and subperiod, the Helsinki Corpus offers relatively long samples of texts, so that statistically significant amounts of data can be recorded in them. This is also true of numerous lexical items, which often have an insignificant role in corpora where the sample size is smaller (cf. Biber 1988, Devitt 1989). Further study will show whether the texts in the corpus are prototypical representatives of the different genres, but they can perhaps be claimed to have an established position in their sociocultural context. The findings in the two chapters identify basic similarities and differences between texts and genres, which may be selected as diagnostic features in subsequent studies based on a larger corpus of texts. Taavitsainen's study tests the methods provided by corpus linguistics for a new aim, using structuralist literary criticism and a variationist approach as her theoretical basis. Her focus is on personal affect which is an optional component of participant relations. Personal affect offers a fruitful object of study since it is realized in various ways in texts, and the co-occurrence patterns of its linguistic features show a great deal of variation. Personal affect has not been defined in an adequate way in the literature, and one of the aims of the study is to define its quality in these texts in more detail. Several quantifying methods are employed in this study and they are combined with qualitative analysis. Factor analysis is applied first to reveal the co-occurrence patterns of linguistic features marking subjectivity. It arranges these features in a hierarchical order according to their importance in reflecting the underlying dimensions of variation. It also indicates textual affinities and points out genres that are close to one another. Further statistical tests are then applied to a smaller number of genres with close affinities. The aim is to test how significant the features with highest factor loadings are as markers between adjoining text types, with special reference to the corpus parameter of imaginative versus nonimaginative narration. Qualitative analysis is used to complement quantitative assessments by relating the texts to their sociohistorical background and by considering the means available in the language for expressing personal affect at that time. In the conclusions, the quality of personal affect in fiction and the adjoining text types are defined in terms of "surge" features of personal affect (i.e. expressions, such as interjections and expletives, conveying intensified personal charge between the participants of communication), interactive versus egocentric focus, indexical features of proximity, and space-building modality.

Introduction

11

In Meurman-Solin's study the hypothesis is that the frequencies and distributions of adjectives and open-class adverbs representing various semantic categories provide evidence of the relative importance of descriptive features as against the importance of stance marking in texts (cf. Biber and Finegan 1989). The study discusses the correlation patterns between these linguistic features and a number of extralinguistic factors such as genre, text category, degree of interactiveness, level of formality, the audience or the author's sex. Moreover, it illustrates how the choices made between adverb and adjective realizations, and, in the case of the latter, between integrated and fragmented structures, reflect varying degrees of informational density, and identify the relative prominence of the author's voice in the individual texts and in the different genres and text categories. Methodologically, the study aims at solving problems related to the traditional ways of grouping texts. It presents an approach in which a detailed analysis of individual features in each idiolect or text leads to the accumulation of evidence as regards co-occurrence patterns of these features; it is the cumulative effect of the carefully analysed patterns that then serves to identify text types. Ideally, these classifications can be meaningfully related to extralinguistic variables, so that ultimately the interrelatedness of the varying social functions of texts and the linguistic expression of participant roles becomes evident. Finally, as concerns the theoretical framework adopted, the studies presented in this volume clearly show the power of the two methods, linguistic and philological: accounting for language change in all its complexity becomes a possible — and a rewarding — task when based on the principles of corpus linguistics, variationist approach, and genre studies.

Notes 1.

For information on the Helsinki Corpus, see Rissanen et al. (1993) and Kytö (1996). Among the satellite corpora projects closely related to the Helsinki Corpus is the Helsinki Corpus of Older Scots, compiled by Anneli Meurman-Solin and currently in international distribution (see Meurman-Solin 1995). In addition, there are three projects in progress which will result in new diachronic corpora: the Corpus of Early American English (see Kytö 1993); the Corpus of Early English Correspondence (see Nevalainen— Raumolin-Brunberg 1996) and the Corpus of Early English Medical Writing (Taavitsainen—Pahta 1997).

12

2.

3.

Matti Rissanen et al. The Penn-Helsinki Parsed Corpus of Middle English, compiled by Anthony Kroch and Ann Taylor, is based on the prose texts drawn from the Middle English section of the Helsinki Corpus accompanied by a number of supplementary texts. The texts have been annotated for syntactic analysis. A new version of this corpus, with a more versatile linguistic coding and further texts, is in preparation. The Brooklyn-Geneva-Amsterdam-Helsinki Parsed Corpus of Old English, based on the Old English section of the Helsinki Corpus, is another morphologically and syntactically annotated corpus due to appear within the next five years. The team of compilers include Susan Pintzuk, Willem Koopman and others. Further information on historical corpora of English, either completed or in preparation, is found in Kytö et al. (1994), Kytö — Rissanen (1996), and Hickey et al. (1997). The following text types, defined on the basis on extralinguistic and situational criteria, are represented in the Helsinki Corpus: law, documents, handbooks, science, homilies, sermons, rules, religious treatises, the Bible, philosophy, prefaces, history, geography, travelogue, (auto)biography, fiction, romances, depositions, trial records, drama (mystery plays and comedies), private and official letters, educational treatises, and diaries. There are also texts which have not been given any particular text type definition (e.g. Old and Middle English verse texts). The articles in the present volume do not presuppose detailed knowledge of the contents and structure of the Helsinki Corpus. However, as always, some knowledge of the material will make it easier to appreciate the findings presented. For those interested in learning more about the corpus, Rissanen et al. (1993) and Kytö (1996) will offer a good starting-point. The typographical conventions in the examples cited from the Helsinki Corpus are 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.

References Biber, Douglas 1988 Variation across speech and writing. Cambridge—New York etc: Cambridge University Press. Biber, Douglas—Edward Finegan 1989 "Styles of stance in English: Lexical and grammatical marking of evidentiality and affect", TEXT 9/1: 93-124.

Introduction

13

Coulthard, Malcolm (ed.) 1994 Advances in written text analysis. London—New York: Routledge. Devitt, Amy J. 1989 Standardizing written English. Diffusion in the case of Scotland 1520 -1659. Cambridge—New York etc.: Cambridge University Press. Fludernik, Monika 1993 The fictions of language and the languages of fiction. London—New York: Routledge. 1996 Towards a 'natural' narratology. New York: Routledge. Halliday, Μ. A. K. 1973 Explorations in the functions of language. London: Edward Arnold. [Reprinted in 1974.] 1978 Language as social semiotic. The social interpretation of language and meaning. London: Edward Arnold. 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. Kytö, Merja 1993 "Early American English", in: Matti Rissanen—Merja Kytö—Minna Palander-Collin (eds.), 83-91. Kytö, Merja—Matti Rissanen—Susan Wright (eds.) 1994 Corpora across the centuries: Proceedings of the First International Colloquium on English Diachronic Corpora, St Catherine's College Cambridge, 25-27 March 1993. Amsterdam—Atlanta, GA: Rodopi. 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. Lewalski, Barbara Kiefer (ed.). 1986 Renaissance genres: Essays on theory, history and interpretation. Cambridge, Mass.—London, England: Harvard University Press.

14

Matti Rissanen et al.

McEnery, Tony—Andrew Wilson 1996 Corpus linguistics. (Edinburgh Textbooks in Empirical Linguistics.) Edinburgh: Edinburgh University Press. Meurman-Solin, Anneli 1995 "A new tool: The Helsinki Corpus of Older Scots (1450-1700)", /CAME Journal 19: 49-62. Meyer, Charles F. 1987 "Apposition in English", Journal of English Linguistics 20/1: 101121.

1992

Apposition in contemporary English. Cambridge: Cambridge University Press. Milroy, James 1992 Linguistic variation and change. On the historical sociolinguistics of English. (Language in Society 19.) Oxford—Cambridge, Mass.: Blackwelk Milroy James—Lesley Milroy 1985 "Linguistic change, social network and speaker innovation", Journal of Linguistics 21: 339-384. Nevalainen, Terttu—Helena Raumolin-Brunberg 1989 "A corpus of Early Modern Standard English in a socio-historical perspective", Neuphilologische Mitteilungen 90/1: 67-111. Nevalainen, Terttu—Helena Raumolin-Brunberg (eds.) 1996 Sociolinguistics and language history. Studies based on the Corpus of Early English Correspondence. Amsterdam—Atlanta, GA: Rodopi. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London—New York: Longman. 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—New York: Mouton de Gruyter. Romaine, Suzanne 1982 Socio-historical linguistics, its status and methodology. (Cambridge Studies in Linguistics 34.) Cambridge—London etc.: Cambridge University Press. Samuels, M. L. 1972 Linguistic evolution, with special reference to English. (Cambridge Studies in Linguistics 5.) Cambridge: Cambridge University Press.

Introduction

15

Schiffrin, Deborah 1994 Approaches to discourse. Oxford—Cambridge, Mass.: Blackwell. Sell, Roger D.—Peter Verdonk (eds.) 1994 Literature and the new interdisciplinary, poetics, linguistics, history. Amsterdam—Atlanta, GA: Rodopi. Taavitsainen, Irma—Päivi Pahta 1997 "Corpus of Early English Medical Writing 1375-1750", ICAME Journal 21: 71-78. 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—London: University of Texas Press, 95-195.

Be/have + past participle: The choice of the auxiliary with intransitives from Late Middle to Modern English Merja

Kytö

1. The be/have variation in the history of English In many languages the choice of the auxiliaries be and have (and their equivalents) has been subject to variation in perfective constructions with intransitives indicating "transition" or "change" (cf. they are arrived versus they have arrived). In some Germanic and Romance languages the development has resulted in the generalization of one construction (e.g. the almost total dominance of "be" in Present-day Danish and "have" in Spanish and Portuguese); in others both variant forms occur in certain grammatically, stylistically or regionally restricted contexts (as, e.g., in Present-day Swedish, German, French and Italian).1 This cross-linguistic variation reflects the differences in the systemic realizations of the distinction between state (favouring be) and action/process (favouring have). Over the successive stages of development, various linguistic and extralinguistic factors have influenced the choice of one form or the other. In the Old English period the be/have + past participle construction denoted "state" in intransitive and transitive uses (cf. hie wceron gecumene; hie hcefdon hine gebundenrte). The past participle originally functioned as an adjective and was sometimes inflected; it could also be preceded by an object complement (Mitchell 1985:1, §709). However, the grammatical concord was gradually lost and the past participle was placed immediately after the auxiliary. Have originally occurred with transitive verbs only, but early on came to be used with intransitives, too. In Early Middle English be prevailed with mutative verbs, but have started gaining ground slowly in uses with the emphasis on "action" and the notion of per-

18

Merja Kytö

fectivity (Ryden—Brorström 1987: 16-18). Signs of the rise of have are apparent from around the 1400s on (for a recent corpus-based study, see Elsness 1989: 100; 1991: 276-283). The development of be and have perfects show features attributed to the process of grammaticalization, but scholars disagree about the exact chronology and nature of the process. A recent discussion on the topic can be found in Denison 1993: 340—368 (who uses the term "grammaticisation"). With have perfect the relevant factors have been the loss of inflections in the participle, the word order, and certain VP types (according to the semantics of OE habban and the valency of the main verb). The process of grammaticalization would have reached a stage of fulfilment "when the have perfect became available for any lexical verb which did not conjugate with be", i.e. when it became an auxiliary verb, the suggested (but debatable) date for this being the late Old English period (Denison 1993: 352). The other possible stages of grammaticalization include the point when the construction became a tense equivalent (probably in late Old English); when it had developed its present-day meaning and superseded be (probably in the seventeenth century); when it became used with all non-auxiliary verbs (in late Modern English) (Denison, 1993: 352). There is, similarly, disagreement over the grammaticalization process of the be perfect (for a summary, see Denison 1993: 360—361); the factors regarded as having influenced the process include, e.g., the increasing use of be as the auxiliary of the passive and the relatively light functional load of have, the possibility of neutralizing the present tense third-person singular forms into the clitic 's, and the prescriptions of normative grammarians. In this study the emphasis is on the variational method, which means excluding from analysis the contexts in which there is no choice of one variant form or the other, e.g. the uses with transitive verbs, which have always been associated with the have perfect. However, many of the factors pointed out in previous studies have been of great help when specifying the scope of the study and the distributional factors considered.2

2. The scope of the study This study focuses on the development of the be/have paradigm in the history of English from the Late Middle English period up until modern

Be/have + past participle

19

times. The reasons for setting Late Middle English as a starting point for observing the development are two-fold. Firstly, as pointed out above, a change starts to take place in the use of the auxiliaries in the case of the mutatives in the fifteenth century, a justification for including the data from the Late Middle English period in the analysis. Secondly, the early stages of the increasing use of have coincide with the crucial transition from the Middle to the Early Modern English period. As for the timespan covered, pursuing the development up until modern times is of great interest, as it was only in the early 19th century that have finally superseded be (Ryden—Brorström 1987). Be is, moreover, still available with some verbs in present-day received usage (they are/have gone, the sun is/has set; Ryden—Brorström 1987: 211) and in some dialects (see, e.g. Kallen 1989: 18-23; Melchers 1992; Filppula 1994). Though the be/have variation has been dealt with in a good number of studies, many of the previous works have largely been based on idiolectal corpora of single authors or restricted to cover only certain periods or a few text types. The present study, based on material that covers a period of some 650 years and representative of a variety of authors and text types, should provide new evidence on the role of the linguistic and extralinguistic factors in the process of change. The study will be carried out within the framework of socio-historical variation analysis (see the Introduction in this volume). The distribution patterns of the variant forms will be followed across the successive periods of time distinguished for the corpora studied (see below). Conclusions will be based on simple frequency tables and results yielded by the logistic regression analysis used to assess the combined effects of the factors and their interaction (see section 9 below).

3. The sources for data In variational terms, the be/have paradigm presents a two-term notional case, involving various aspects of syntactic, semantic and lexical change. For obvious reasons, this paradigm is a particularly rewarding topic within the framework of computer-assisted corpus analysis. Relevant examples with the construction are, if not overwhelming in number, still frequent enough for a thorough-going analysis; they are also relatively easy to retrieve from machine-readable material on the basis of lexical forms. As

20

Merja Kytö

there is no single corpus available at the moment that would cover the time-span included in this study, several diachronic corpora will need to be consulted. The differences in the size and structure of the corpora will be taken into consideration when weighing the evidence. The corpora that were used as sources for data are introduced below. The Helsinki Corpus of English Texts (HC; Late ME, EModE) The Late Middle English data and a good deal of the Early Modern English data have been drawn from the Helsinki Corpus (see Rissanen— Kytö—Palander-Collin 1993; Kytö 1996); the subperiods included in the Helsinki Corpus material are ME3 (1350-1420) and ME4 (1420-1500), and EModE 1 (1500-1570), EModE2 (1570-1640) and EModE3 (16401710). All types of texts represented in the corpus will be considered, but priority will be given to those that have counterparts in the other corpora studied. The Century of Prose Corpus (COPC) As regards the rise of have, the Century of Prose Corpus covers a crucial period of development, extending from 1680 to 1780. Part A of the corpus includes extracts from works of 20 major prose authors of the period (three selections of 5000 words representing each author), while Part Β consists of writings (sample length 2000 words) drawn from the pens of one hundred authors writing as journalists, scholars, men of letters and so forth (Milic 1990: 27-29; 1994: 70). Text type definitions are given by the compiler for texts included in Part Β only. The first three decades of writing in the Century of Prose Corpus (1680-1710) coincide with the last three decades of the Early Modern English section of the Helsinki Corpus. The whole century covered also coincides with the early decades represented by the ARCHER Corpus (1650-1990, see below). Contrary to the conventions adopted for the other corpora included in this study, the spelling of the Century of Prose Corpus has been modernized. The ARCHER Corpus This study owes a great deal to the compilers of the ARCHER Corpus, or A Representative Corpus of Historical English Registers, for access to relevant data drawn from parts of the pilot version of the corpus. When complete, this corpus will include c. 1,000 texts and c. 1.7 million words

Be/have + past participle

21

(sample size at least 2,000 words). The aim of the corpus is "to enable the analysis of historical change in the range of written and speech-based registers of English from 1650 to the present" (Biber et al. 1994a: 3; see also Biber et al. 1994b). Out of the text types found in the corpus, six have been included in the present study: fiction prose, drama, journals, letters, science and sermons. The six text types are of special interest for this study, as they all bear affinities to those found in the Helsinki Corpus and parts of the Century of Prose Corpus. Texts by British and American authors have been considered; while the British authors are represented more or less evenly over the subperiods distinguished, the texts by American authors are all from three subperiods, 1750-1799, 1850-1899 and 1950-1990 (all scientific writings are by British authors). The corpus is being tagged for grammatical and functional categories; for the purposes of the present study, the data was drawn from the untagged version on the basis of lexical forms. 3 Summary of the corpora

studied

The main characteristics of the size and the structure of the corpora included in the study are given in Table 1. Table 1. The structure and size of the (sub)corpora studied.

Helsinki Corpus

Text type

Words

ME3 ( 1 3 5 0 - 1 4 2 0 )

fiction diary private letter official letter drama = mystery play science sermon document handbook philosophy homily other

14,300

Total

5,000 3,600 18,900 13,900 19,200 12,600 7,300 94,900 189,700

Total

22

Merja Kytö

Table 1. Continued.

Helsinki Corpus

Text type

ME4 (1420-1500)

fiction diary private letter official letter drama = mystery play science sermon document handbook philosophy homily other

Total

Words 8,800 -

19,500 3,200 20,100 6,400 25,000 10,500 20,100 -

100,300 213,900

Total Late ME EModEl (1500-1570)

Total

Total

403,600 fiction diary private letter official letter comedy science sermon document handbook philosophy other

11,600 13,100 10,600 6,300 10,600 12,900 9,500 10,000 9,900 95,600 190,100

Be/have + past participle Table 1. Continued.

Helsinki Corpus

Text type

Words

EModE2 (1570-1640)

fiction diary private letter official letter comedy science sermon document handbook philosophy other

12,500 12,500 11,600 5,700 11,800 13,000 10,300

Total EModE3 (1640-1710)

Total Total EModE

Total

-

12,300 6,900 93,200 189,800

fiction diary private letter official letter comedy science sermon document handbook philosophy other

12,000 11,200 13,100 5,900 12,700 11,300 12,500 -

11,400 8,800 72,100 171,000 550,900

23

24

Merja Kytö

Table 1. Continued.

COPC

Text type

Part A (1680-1780)

expository prose by 20 authors (15,000 words each)

Total Part Β (1680-1780)

Words 300,000

300,000 fiction letter science biography periodical education essay history polemics travel

Total

20,000 20,000 20,000 20,000 20,000 20,000 20,000 20,000 20,000 20,000 200,000

Total COPC

500,000

ARCHER Corpus

Text type

Words

ARCHER 1 (1650-1700)

fiction journal letter drama science sermon

29,500 21,900 14,900 32,600 20,900 11,400

Total Total ARCHER 1

Total

Total

131,200 131,200

Be/have + past participle Table 1. Continued.

ARCHER Corpus

Text type

Words

ARCHER2a (1700-1750)

fiction journal letter drama science sermon

44,000 22,000 20,700 24,700 21,400 6,500

Total ARCHER2b (1750-1800)

139,300 fiction journal letter drama science sermon

86,200 45,300 32,200 44,400 20,900 26,900 255,900

Total

395,200

Total ARCHER2 ARCHER3a (1800-1850)

Total

Total

fiction journal letter drama science sermon

53,900 22,800 17,200 34,400 21,400 4,600 154,300

25

26

Merja Kytö

Table 1. Continued.

ARCHER Corpus

Text type

Words

ARCHER3b (1850-1900)

fiction journal letter drama science sermon

79,700 46,600 28,300 77,200 22,400 29,100

Total

283,300

Total ARCHER3 ARCHER4a (1900-1950)

437,600 fiction journal letter drama science sermon

12,400 22,700 15,200 35,700 22,500 4,300 112,800

Total ARCHER4b (1950-1990)

Total

fiction journal letter drama science sermon

Total Total ARCHER4

68,100 45,100 32,100 68,400 22,900 28,000 264,600 377,400

Total ARCHER

1 341,400

TOTAL HC+COPC+ARCHER

2 795,900

Be/have + past participle

27

4. On the criteria of inclusion Selecting examples for a study of the be/have variation with intransitives poses some problems. There are two main questions related to (1) the type and history of the main verb and (2) the rich functional range of the auxiliary be. Further selectional criteria to be considered include the role of object-like complements (e.g. nouns of extent or measure), and some relatively infrequent constructions in which the variational role of the auxiliaries is blurred. The type of the verb The main verbs are an open-ended class, characterized by a number of semantic features not always too precise to determine. The studies on the history of the be/have variation traditionally deal with the use of the more frequent verbs indicating motion (e.g. arrive, come, go, pass, ride), process or change (e.g. alter, change, improve, turn), happening (e.g. befall, chance, hap(pen)), appearing or originating (e.g. appear, arise, become, begin) and finishing or disappearing (e.g. cease, decay, decline, expire, die).4 The inclusion of the less frequent verbs (e.g. pirouette, penetrate) has been the researcher's decision. Most intransitive verbs considered in previous studies have here been included for preliminary analysis (for some individual verbs excluded, see below); along the lines of variational method, some categories dominated by one or other variant form have been excluded from further analysis on the basis of the distribution patterns obtained. Similarly, as seems sensible within the framework of variation analysis, some verbs (e.g. follow, misgo), which on the basis of the previous studies have been used with the form have only, have been excluded altogether (for a provisional list of verbs and their be/have history, see Visser 1952: 661-663). Some verbs are frequently indeterminate in nature as regards the transitive versus intransitive dimension. For this reason, such verbs as agree, assemble, end, finish, gather, meet, marry, wed and die {be dead) have been excluded from the analysis. In some studies attention has also been paid to non-mutative or stative verbs (e.g. cling, lie, rest, stay; see Visser 1963-1973: 2044-2084). This category has also been considered at the first stages of the present study.

28

Merja Kytö

The functional range of the verb be The use of be as a copula as well as a perfective and a passive marker renders the functional load of the verb heavy (Ryden—Brorström 1987: 24; Elsness 1991: 266-268). With some verbs it is not always possible to distinguish between the copula and the perfective uses (cf. he is changed 'he is different' versus he is changed 'he has become different'); see example (l). 5 (1)

The cannon, however, did not stand proof, and the Indians, who made a close attack, were beaten off and the garrison relieved. The fort is now totally decayed, and Captain Zane, the only inhabitant at or near the place, makes use of it for firewood. (ARCHER2b/Journal/Peter Muhlenberg)6

Nor can one always distinguish between active (perfective) and passive (present/past tense) uses with verbs which can appear in both transitive and intransitive contexts {he is changed 'he has become different' versus he is changed 'he has been made different'); for a discussion, see Ryden—Brorström (1987: 24), who for this reason exclude an example such as 'My prospect of getting to London this spring is rather darkened* (Wordsworth II: 772, 1836) from their data; for further points, see also Kakietek 1987 [1976]: 310-311); for a corpus example, see (2). (2)

.. and such is their methode, that rests not so much vppon euidence of truth prooued by arguments, authorities, similitudes, examples; as vpon particular confutations and solutions of euerie scruple, cauillation & objection: ... so that the Fable and fiction of Scylla seemeth to be a liuely Image of this kinde of Philosophie or knowledge, which was transformed into a comely Virgine for the vpper parts; (Bacon, Advancement of Learning 20V)

Ambiguous instances such as (1) and (2) have been excluded from this study; the basic criterion of inclusion has been that in its context the construction conveys the notion of perfectivity.7 All instances of dynamic intransitives (e.g. go, come, rise, arrive) have been included, even though the construction in some examples may come closer to a stative than a perfective meaning (cf. they are/have gone, the sun is/has set).8

Be/have + past participle Object-like

29

complements

Though this study is primarily concerned with intransitive uses as in (3a), there are verbs that can take complements in the form of nouns of extent or measure as in (3 b).9 (3)

a.

My godfathyr has be syke byt he ys whell mendyd, thankyd be God. [Thys same] day my Loord ys comyn to London to aske the Kyng leue to go to the Rodys for he ys sent for. (Richard Cely, The Cely Letters 107)

b.

... and when mattens vhos done thay whente to a kynnyswhoman off the 3ewnge genttyllwhomane; and I sent to them a pottell of whyte romnay, and thay toke hyt thankefully, for thay had cwm a myle a fote that mornyng; (Richard Cely, The Cely Letters 152)

With verbs there is variation in whether prepositions or prepositional phrases are used to link the destination or other complement with the verb; cf. (4a) with enter + into + object pronoun and (4b) with enter + object pronoun.10 (4)

a.

They have not perhaps received precisely what they expected when their Christian life began, for the kingdom of heaven cannot be really known until a man has entered into it; (ARCHER3b/Sermon/Robert Dale)

b.

Feb. 26th The works of Ciudad Rodrigo having been completely put in order, and a garrison of Spaniards having entered it, the army was ordered to proceed towards Estremadura. (ARCHER3a/Journal/George Simmons)

In the interest of the emerging individual verb profiles, instances such as (3b) will be included in the analysis. Out of the examples illustrated in (4b), two frequent verbs, enter and pass, have been sampled systematically for full coverage; as for the other verbs, some of them presenting ambiguous borderline cases between transitive and intransitive uses, the instances with object-like complements have been omitted from the analysis.

30

Merja Kytö

Special

constructions

The construction have been + past participle was first attested in the 14th century and disappeared from usage after the 1850s (for discussion and references, see Ryden—Brorström 1987: 24-26; Denison 1993: 361, 363, with reference to the "double perfect" construction). This category would, strictly speaking, have presented a third variant form to compete with be and have, but as no more than a score of examples were found (an instance given in (5)), the category has been excluded from the statistics. (5)

Ay, at home, you're as sure of finding an old Mistress, as a Creditor that expects you to pay him an old Debt, in good Humor too, I warrant; I was afraid, Madam, you had not been come home yet. (ARCHER2a/Drama/Catharine Trotter)

Another marginal category is presented by the instances of coordinated past participles (see example (6)); as the choice of the verb may have depended on verbs other than those included in the study, these instances have been omitted from the analysis. (6)

For Sr. Allmighty God I take to my record, I have not meant, intended, or gone about, ne also have willed mine Officers, to do any thing concerning the said Suppressions, but under such forme & manner, as is & hath largely been to the full satisfaction, recompence, & joyous contentation of any person which hath had, or could pretend to have right or interest in the same ... (Wolsey, Original Letters (Illustrative of English History) (1846) II 20)

Similarly, the examples with the reduced form 's, which could stand either for is or has, are excluded from the analysis; see (7); the same holds for the examples with the form 'd in contexts where it could stand either for had or should/would. (7)

Every body, no body, I can't tell who; such a Mixture of the pert and formal! but the most conspicuous Fop was Beau, what d'you call him, with the fine lac'd Liveries? he's so lately come into the Town, that I don't know his Name yet, ... (ARCHER2 a/Dram a/John Mottley)

Be/have + past participle

5.

31

Distributional factors

5.1. The main factors considered Over the centuries, the distinction between state/result (indicated by be) and action (indicated by have) seems to have been one of the main distributional factors influencing the choice of the auxiliary.11 The other factors, extralinguistic and linguistic, taken into consideration in the present study are the following: Extralinguistic factors·. time region (British versus American English in 1750-1990) text type the relationship of the text to spoken language (speech-based/ written/script) the level of formality (formal/informal) oral versus non-oral genres author properties (men/women writers) for foreign influence, see the section on main verbs and loan words, below Linguistic factors: stative versus mutative verbs (or stand, live, etc. versus come, arrive, run, grow, become, etc.) action versus process verbs (or come, arrive, etc. versus grow, become, etc.) frequent verbs (e.g. come and go) versus other less frequent verbs; individual verb profiles past perfect perfect infinitive -ing constructions durative and iterative contexts optative / conditional contexts negation object-like and other complements loan verbs

32

Merja Kytö

After introducing the overall distributions of the data, I will discuss the role of the two fundamental factors, Stative versus mutative and action versus process uses. Next, the influence of a number of major extra-linguistic factors will be dealt with, followed by a discussion of the role of further linguistic factors. As many of the linguistic factors considered may be effective simultaneously, attention will also be paid to possible interaction and combined effects of factors. This will be done by discussing the distribution patterns obtained and, more systematically, by applying the principles of logistic regression analysis to the data.

5.2. Overall distributions of variant forms The data included in the analysis total 2868 examples. Out of the 194 different verbs, come and go are the most frequent (with 512 and 445 examples, respectively; see Appendix 1). Ten verbs are represented by more than 30 examples, i.e. pass (178), become (165), fall (131), grow (90), arrive (62), get (63), return (56), enter (43), run (39), depart (32). Of the 194 different verbs, 71 are represented by one occurrence only. The overall figures point, as might be expected, to a gradual rise in the use of have over the centuries.12 While be still dominates in the Late Middle English and Early Modern English periods, during the eighteenth century have is already used in more than 50% of the instances; during the nineteenth century have supersedes be (see Table 2; in the interest of chronology, the list is organized so that the results obtained for the Century of Prose Corpus (1680-1780) follow the first subperiod of the ARCHER Corpus (1650-1700)). The percentages are calculated from the total of examples found with be or have within a subperiod. With twoway tables the chi-square test is used to test the significance of the effect of the time factor (or period division) on the distributions obtained; the chi-square figures are nearly significant when ρ < 0.05, significant when ρ < 0.01 and highly significant when ρ < 0.001, e.g. in Table 2, the influence of the time factor on the distributions of the variant forms is highly significant.13 The breakdown figures point to the post-1750s as the final turn-over period in the history of the paradigm (see Table 3). The results are, by and large, convergent with the trend of development presented by Ryden—Brorström (1987) in their study of be/have variation in eighteenth-century letters and drama. However, the figures obtained for have are somewhat higher than expected on the basis of previous studies:

Be/have + past participle

33

Table 2. Be/have variation with intransitives in 1350-1990 (the percentages are calculated from the total of examples found with be or have). be

have

254 363 95 135 193 79 27

106 188 62 273 250 447 396

(29%) (34%) (39%) (67%) (56%) (85%) (94%)

360 551 157 408 443 526 423

1146

1722

(60%)

2868

Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Total

Total

Chi-square = 668.0, df = 6, Ρ < 0.001

Table 3. Be/have variation with intransitives (per subperiods).

ME3 (1350-1420) ME4 (1420-1500) EModE 1 (1500-1570) EModE2 (1570-1640) EModE3 (1640-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2a (1700-1750) ARCHER2b (1750-1800) ARCHER3a (1800-1850) ARCHER3b (1850-1900) ARCHER4a (1900-1950) ARCHER4b (1950-1990) Total

be

have

134 120 105 152 106 95 135 80 113 40 39 10 17

50 56 49 74 65 61 273 71 180 149 298 95 301

(27%) (32%) (32%) (33%) (38%) (39%) (67%) (47%) (61%) (79%) (88%) (90%) (95%)

184 176 154 226 171 156 408 151 293 189 337 105 318

1146

1722

(60%)

2868

Chi-square = 685.0, df = 12, Ρ < 0.001

Total

34

Merja

Kytö

the overall ratio of have has been estimated at some 10% around 1600, and at 20% around 1700 (Ryden—Brorström 1987: 200; Ryden 1991: 346-347). The varied textual background of the material analysed (see below) could partly explain the difference. There may also have been differences in the way of applying the criteria of exclusion when selecting examples. As pointed out above, the examples with be are particularly problematic; the instances with have are much easier to deal with. As my principles for excluding the ambiguous examples with be were rather strict, this may be reflected in the figures obtained for have.

6.

Two main axes: stative versus mutative and action versus process

6.1. Stative versus mutative verbs Stative verbs (lie, rest, stand, stay, etc.) are known to have favoured the use of have. This holds for the data included in this study as well (see Table 4; percentages are given for the better represented categories of have only):

Table 4. Stative verbs. be

have

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

2 2 0 0 0 0 0

22 35

Total

4

102

Total (92%) (95%)

6

24 37

6 11

11 9 8 11

9 8

11 (96%)

106

Be/have + past participle

35

As the 106 examples attested with stative verbs represent only some four percent of the data and are distributed fairly evenly over the subperiods, the figures obtained for the mutative verbs (come, arrive, run, etc.) come close to the overall figures given in Tables 2 and 3 (see Table 5).

Table 5. Mutative verbs.

ME3 (1350-1420) ME4 (1420-1500) EModEl (1500-1570) EModE2 (1570-1640) EModE3 (1640-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2a (1700-1750) ARCHER2b (1750-1800) ARCHER3a (1800-1850) ARCHER3b (1850-1900) ARCHER4a (1900-1950) ARCHER4b (1950-1990) Total

be

have

132 120 104 151 106 95 135 80 113 40 39 10 17

38 46 42 54 57 55 262 65 177 147 292 94 291

(22%) (28%) (29%) (26%) (35%) (37%) (66%) (45%) (61%) (79%) (88%) (90%) (94%)

170 166 146 205 163 150 397 145 290 187 331 104 308

1142

1620

(59%)

2762

Total

Chi-square = 744.7, df = 12, Ρ < 0.001

Within the group of mutative verbs, a number of verbs came to be used with have at an early stage. These can perhaps be best described by the absence of the semantic feature that would assign them to verbs indicating motion or process. In Visser's classification of intransitive verbs these verbs fall in the groups of verbs indicating appearing, originating (e.g. appear, begin), or finishing, ceasing, disappearing (e.g. cease, expire, do, perish), or happening, befalling and like concepts (e.g. befall, hap(pen) chance·, Visser 1963-1973: 2044-2084). The figures obtained for the use of be and have with these verbs are given in Table 6; the verbs included in the figures are appear, befall, begin, cease, chance, disappear, do, expire, fail, hap, happen, occur, perish, start, stop, vanish.

36

Merja Kytö

Table 6. Be versus have with appear, befall, begin, cease, chance, disappear, do, expire, fail, hap, happen, occur, perish, start, stop, and vanish. be Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Total

5 8 1 1 1 0 2 18

have

Total

6 (55%) 22 (73%) 7 (88%) 70 (99%) 46 (98%) 63 (100%) 64 (97%) 278

(94%)

11 30 8 71 47 63 66 296

As the use of have with stative verbs and verbs indicating appearing, finishing, happening and like notions comes close to a knock-out feature, they will be excluded from further analysis. The remaining sections of the study will thus focus on the choice of be and have in the 2466 examples with mutative verbs expressing action or process.

6.2. Action versus process uses The fundamental differences between the use of be and have with verbs indicating action (typically motion as in, e.g. arrive, return, enter) or process (typically change of state as in, e.g. grow, become, wax) have been pointed out in previous studies. 14 The main difference is that, by and large, the process verbs are slower to adopt have than the motion verbs. There are, of course, some verbs, e.g. turn and get, which appear to have a double function, indicating both motion and process; there are also some phrase-like uses such as fall ill/sick, go mad/crazy. However, one or the other function seems to dominate the meaning profile of an individual verb, and instances of indeterminate meaning are relatively rare (for a discussion, see the individual verb profiles below). The distributions of the variant forms with action and process uses are given in Table 7; when coding the examples, the "use" rather than the "main meaning" of the verb has been followed, i.e. (8a) has been coded as an instance of action use and (8b) as that of a process use.

Be/have + past participle (8)

37

a.

He was therefore no sooner got to his apartment, but he sent the royal veil to Imoinda; that is, the ceremony of invitation: (Behn, Oroonoko 158)

b.

The church might have got tired of him: ... he forestalled any disagreeable circumstances which might have attended that event, by getting tired of the church: (ARCHER3b/Fiction/John Cooke)

In discussion and tables, the terms action/process verbs/uses will be used indiscriminately to cover the coding principle adopted. Table 7. Action versus process verbs. Action be 102 ME3 (1350-1420) 101 ME4 (1420-1500) 91 EModEl (1500-1570) EModE2 (1570-1640) 127 80 EModE3 (1640-1710) ARCHER 1 (1650-1700) 69 80 COPC (1680-1780) ARCHER2a (1700-1750) 52 ARCHER2b (1750-1800) 70 ARCHER3a (1800-1850) 33 ARCHER3b (1850-1900) 34 ARCHER4a (1900-1950) 9 ARCHER4b (1950-1990) 11 Total

Process

have 35 41 34 43 40 45 166 46 124 107 205 66 176

Total (26%) (29%) (27%) (25%) (33%) (39%) (67%) (47%) (64%) (76%) (86%) (88%) (94%)

137 142 125 170 120 114 246 98 194 140 239 75 187

859 1128 (57%) 1987 Chi-square = 512.8 df = 12, Ρ < 0.001

be have 30 14 12 20 23 25 28 54 42 7 5 1 4

1 1 4 4 6 3 3 26 23 18 46 16 63

Total (3%) (7%) (25%) (17%) (21%) (11%) (10%) (33%) (35%) (72%) (90%) (94%) (94%)

31 15 16 24 29 28 31 80 65 25 51 17 67

265 214 (45%) 479 Chi-square = 215.8 df = 12, Ρ < 0.001

The relative figures obtained for the action uses follow closely the figures obtained for the total of the examples with mutative verbs: have occurs in over 60% of the instances from the late 1700s on (the figures for the Century of Prose Corpus even point to an earlier breakthrough of the

38

Merja Kytö

form; see Table 5 above). However, the process uses retain the verb be in nearly or over 70% of instances up until the early nineteenth century. A levelling process then takes place and over the first half of the nineteenth century have supersedes be, coming closer to the figures obtained for action uses.

7.

Extralinguistic factors

7.1. Regional variation (1750-1990) The material included in the ARCHER Corpus contains some 328,000 words sampled from the latter halves of the eighteenth, nineteenth and twentieth centuries: the data offers the possibility to compare the rate and directions of change in British and American English (the overwhelming use of the form be in the British English data from the pre-1700 period makes comparisons of the earlier regional variation less interesting). Table 8 presents the figures obtained for the distributions of forms in the two varieties. Table 8. Be versus have in British and American English (action and process uses). British English Action/process

be

have

ARCHER 1 (1650-1700) ARCHER2b (1750-1800) ARCHER3b (1850-1900) ARCHER4b (1950-1990)

94 66 21 7

48 68 136 162

Total

188

American English Total

be have

Total

142 134 157 169

_ 46 79 (63%) 18 115 (86%) 8 77 (91%)

_ 125 133 85

414 (69%) 602

72 271 (79%)

343

(34%) (51%) (87%) (96%)

Chi-square = 182.2 df = 3, Ρ < 0.001

—

Chi-square =30.17 df = 3, Ρ < 0.001

Be/have + past participle

39

According to the results obtained, there are signs of an innovative tendency in late eighteenth-century American English. The breakdown figures for action and process verbs point to the process verbs, in particular, as the catalyst for the faster rate of change (see Table 9). Table 9. Action versus process verbs in British and American English.

British English be

have

69 41 16 3

45 62 113 124

American Engish Total

be

have

114 103 129 127

_ 29 18 8

_ _ 62 (68%) 91 92 (84%) 110 52 (87%) 60

Total

Action verbs ARCHER 1 ARCHER2 ARCHER3 ARCHER4

(1650--1700) (1750--1800) (1850--1900) (1950-1990)

129

Total

(39%) (60%) (88%) (98%)

344 (73%) 473

Chi-square = 125.8 df = 3, Ρ < 0.001

55 206 (79%) 261 Chi -square = 10.00 df : = 3, Ρ < 0.01

Process verbs ARCHER 1 ARCHER2 ARCHER3 ARCHER4 Total

(1650--1700) (1750--1800) (1850--1900) (1950--1990)

25 25 5 4

3 6 23 38

28 31 28 42

17 0 0

_ 17 (50%) 23 (100%) 25 (100%)

_ 34 23 25

59

70 (54%) 129

17

65 (79%)

82

(11%) (19%) (82%) (90%)

Chi-square = 67.58 df = 3, Ρ < 0.001

—

Chi -square = 30.28 d f := 3, Ρ < 0.001

40

7.2.

Merja Kytö

Text type variation

7.2.1. Organizing data As pointed out above (see Table 1), the corpora studied differ in structure and size. Moreover, there are differences in the way that texts have been selected, defined and grouped to represent one type or another (e.g. drama, (private) letters). The subsections drawn from the Helsinki Corpus and the Century of Prose Corpus present a more varied selection of text types than the material drawn from the ARCHER Corpus. 15 To facilitate comparisons, the textual distributions of the variant forms will be organized as follows: - the tables will be organized to preserve the original text type definitions assigned by the compilers of the corpora, and the affinities within related text types will be taken into account when interpreting the figures. - with the (sub)types that could be thought to display some degree of diachronic continuity, the results obtained will be organized as follows: - the types of texts followed across the corpora include fiction, science, sermons, drama, diaries/journals, and (private) letters. - the results obtained for the diaries in the Helsinki Corpus will be compared to those obtained for the journals included in the ARCHER Corpus. 16 - in the Helsinki Corpus a distinction is made between private letters (or correspondence between close family members) and official letters; this distinction is preserved in the present study. In the other corpora studied no such rigid distinction is made. 17 - the extracts of representative writers in Part A of the Century of Prose Corpus will be grouped together under the heading "COPC A" (no attempt has been made to distinguish types of text in further detail).

Be/have + past participle

41

- the term "other" groups together the results obtained for texts which have no direct counterparts in the other corpora or which contribute only a relatively small number of examples to the study (e.g., law, documents, official letters, homilies, prefaces and rules from the Helsinki Corpus, and essays, periodicals and polemics from Part Β of the Century of Prose Corpus).

7.2.2. The text types followed across the corpora The textual distributions obtained for the variant forms will be presented in four tables (Tables 10 to 12, and 15). Table 10 gives the overall figures for the (sub)types of texts represented across the corpora studied (neither drama nor sermons are included in the Century of Prose Corpus): have occurs in over 50% of the instances recorded for the texts assigned the labels fiction, diary/journal, (private) letter, drama,18 science and sermon. Interestingly, the relative figures are highest for diary/journal and fiction, both types of texts approaching the more informal end of the stylistic continuum. Table 10. Be versus have in various text types (I).

Fiction Diary/journal (Private) letter Drama Science Sermon

be

have

195 96 110 145 59 65

402 217 111 162 73 93

Total (67%) (69%) (50%) (53%) (55%) (59%)

597 313 221 307 132 158

Chi-square = 149.3, df = 8, Ρ < 0.001

To follow the process of change within each text type, the breakdown figures for the different corpora are given in Table 11 (the figures obtained for the ARCHER Corpus are given by century).

42

Merja Kytö

Table II. Be versus have in various text types (II).

be

have

Total

Fiction

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

24 38 18 15 71 23 6

16 16 17 14 72 153 114

(40%) (30%) (49%) (48%) (50%) (87%) (95%)

40 54 35 29 143 176 120

Diary

EModE (1500-1710)

29

17

(37%)

46

Journal

ARCHER 1 ARCHER2 ARCHER3 ARCHER4

21 24 15 7

6 53 67 74

(22%) (69%) (82%) (91%)

27 77 82 81

Private letter Late ME (1350-1500) EModE (1500-1710)

18 40

2 7

(10%) (15%)

20 47

Letter

ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

15 7 17 11 2

6 14 24 27 31

(29%) (67%) (59%) (71%) (94%)

21 21 41 38 33

Drama

Late ME (1350-1500)

28

2

(7%)

30

EModE (1500-1710) ARCHER 1 (1650-1700) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

25 19 44 23 6

5 4 17 78 56

(17%) (17%) (28%) (77%) (90%)

30 23 61 101 62

(1650-1700) (1700-1800) (1800-1900) (1900-1990)

Be/have + past participle

43

Table 11. Continued.

Science

Sermon

be

have

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

2 6 16 8 23 2 2

0 3 12 7 16 20 15

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

28 12 5 13 5 2

9 5 4 13 31 31

Total

(43%) (47%) (41%) (91%) (88%) (24%) (29%) (50%) (86%) (94%)

2 9 28 15 39 22 17 37 17 9 26 36 33

Chi-square = 484.4, df = 6, Ρ < 0.001

While the nineteenth century is, as expected, the period when the final rise of have takes place in most text types shown above (with relative figures all 70% or above recorded for ARCHER3), journals show relatively high figures as early as the eighteenth century (ARCHER2). In journals the rise of have is most marked from ARCHER 1 to ARCHER2 (47%), while in fiction, drama and science a corresponding j u m p takes place only between ARCHER2 to ARCHER3. Further, journals and letters are the types of text in which have is used in nearly or over 60% of the instances. All in all, the evidence is scantier for letters and sermons. To highlight the crucial stages of development, further breakdown figures will be given to the subperiods distinguished for the eighteenth and nineteenth centuries in the ARCHER Corpus (see Table 12; percentages are given for the better represented categories only).

44

Merja Kytö

Table 12. Be versus have in various text types (III).

be

have

Total

Fiction

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

33 38 8 15

20 52 65 88

(38%) (58%) (89%) (85%)

53 90 73 103

Journal

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

13 11 9 6

9 44 22 45

(41%) (80%) (71%) (88%)

22 55 31 51

Letter

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

6 11 8 3

5 19 5 22

(45%) (63%) (38%) (88%)

11 30 13 25

Drama

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

18 26 13 10

4 13 22 56

(18%) (33%) (63%) (85%)

22 39 35 66

Science

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

9 14 1 1

10 7 10 10

(53%) (33%) (91%) (91%)

19 21 11 11

Sermon

ARCHER2a ARCHER2b ARCHER3a ARCHER3b

1700-1750) 1750-1800) 1800-1850) 1850-1900)

1 12 1 4

1 12 1 30

(50%) (88%)

2 24 2 34

Chi-square = 359.1, df = 12, Ρ < 0.001

The figures show that the rise in the use of have in journals takes place in the latter half of the eighteenth century (for reasons, see below). The results point to meaningful differences in the process of change along the

Be/have + past participle

45

textual continuum over the centuries (the influence of the action versus process uses does not add much to the picture shown in Table 7; for exact figures, see the table given in Appendix 2).

7.2.3. The influence of come and go on textual distributions The occurrence of come and go, two of the most frequent verbs in my data, influences the figures obtained for the textual distributions of be and have. By way of introduction, Table 13 below presents the figures for come and go contrasted with the other verbs in the data considered. Table 13. Come and go versus other (action/process) verbs.

Come and go

Other

be

have

Total

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

131 198 39 41 76 57 16

32 48 12 33 37 126 111

Total

558

399 (42%) 957

(20%) (20%) (24%) (45%) (33%) (69%) (87%)

Chi-•square = 258.0 df = 6, Ρ < 0.001

163 246 51 74 113 183 127

be have 116 155 55 93 116 22 9

46 83 37 159 158 250 210

Total (28%) (35%) (40%) (63%) (58%) (92%) (96%)

162 238 92 252 274 272 219

566 943 (62%) 1509 Chi-square = 384.7 df = 6, Ρ < 0.001

The rise of have is clearly less rapid with come and go than with other verbs: the tendency is evident from the Late Middle English period on, and the two verbs are equally numerous as the other verbs as late as the twentienth century. One reason which accounts for the popularity of be with come and go lies in the Stative type of uses mentioned above (cf. note 8 and further discussion in section 8.5.2 below). The overall figures are influenced by how characteristic the occurrence of these verbs is in a text type (see Table 14 below). As shown above, the rise of have with verbs other than come and go can be seen to take place earlier in fiction

46

Merja Kytö

and journal writing than in the other text types studied across the corpora; in these text types have is, similarly, used to a greater extent than be with verbs other than come and go. In letters and drama, be is quite persistent with come and go, even during the nineteenth century, when have has already superseded be with other verbs. Table 14. Come and go versus other verbs in fiction, diary/journal, (private) letter, drama, science and sermon.

Come and go be have

Other Total

Fiction Late ME (1350-1500) EModE (1500-1710) ARCHERl (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

14 6 (30%) 18 8 (31%) 9 5 (36%) 9 4 (31%) 34 13 (28%) 17 57 (77%) 4 33 (89%)

20 26 14 13 47 74 37

Diary

23

7 (23%)

30

9 0 8 10 (56%) 10 30 (75%) 4 23 (85%)

EModE (1500-1710)

Journal ARCHERl ARCHER2 ARCHER3 ARCHER4

(1650-1700) (1700-1800) (1800-1900) (1900-1990)

Private Late ME (1350-1500) EModE (1500-1710) letter Letter

ARCHERl (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

be have

Total

10 10 (50%) 20 20 8 (29%) 28 9 12 (57%) 21 6 10 (63%) 16 37 59 (61%) 96 6 96 (94%) 102 2 81 (98%) 83 10 (63%)

16

9 18 40 27

12 6 (33%) 16 43 (73%) 5 37 (88%) 3 51 (94%)

18 59 42 54

2 (14%) 3 (14%)

14 22

6 21

0 4 (16%)

6 25

4 3 3 4 6 4 8 9 (53%) 2 15 (88%)

7 7 10 17 17

11 3 (21%) 4 10 (71%) 11 20 (65%) 3 18 (86%) 0 16 (100%)

14 14 31 21 16

12 19

6

Be/have + past participle

47

Table 14, Continued. Other

Come and go be have

Drama Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

24

Total

0

18 2 (10%) 13 3 (19%) 20 6 (23%) 19 22 (54%) 6 30 (83%) 0 0 1 2 0 0 0

0 0 0 2 2 3 0

12 3 (20%) 5 4 3 1 8 2 3 5 0 10

24

be have

4

2

Total 6

20 16 26 41 36

7 3 10 6 1 7 24 11 (31%) 35 4 56 (93%) 60 0 26 (100%) 26

0 0 1 4 2 3 0

2 0 2 6 3 9 15 12 (44%) 27 6 5 (45%) 11 23 14 (38%) 37 2 17 (89%) 19 2 15 (88%) 17

15 9 4 10 8 10

16 6 (27%) 22 8 7 1 2 3 5 5 11 (69%) 16 2 26 (93%) 28 2 21 (91%) 23

7.2.4. The corpus-specific text types Table 15 presents the results obtained for the genres that are less consistently represented in the corpora studied. The scope of these results is rather limited in view of the full span of diachrony studied, but the figures offer interesting glimpses at the development of text types other than those illustrated above (Tables 10, 11 and 12). In the text types covering more than one period, the overall lines of development point, as expected, to the rise of have.

48

Merja Kytö

Table 15. Be versus have in various text types (IV).

be

have

Total

Bible

Late ME (1350-1500) EModE (1500-1710)

23 86

11 9

(32%) (9%)

34 95

Biogr. etc.

Late ME (1350-1500) EModE (1500-1710) COPC (1680-1780)

2 32 6

0 12 7

(27%) (54%)

2 44 13

Education

EModE (1500-1710) COPC (1680-1780)

8 3

10 1

(56%)

18 4

Handbooks

Late ME (1350-1500) EModE (1500-1710)

27 15

1 4

(4%) (21%)

28 19

History

Late ME (1350-1500) EModE (1500-1710) COPC (1680-1780)

29 20 4

6 16 2

(17%) (44%)

35 36 6

Other

Late ME (1350-1500) EModE (1500-1710) COPC (1680-1780)

22 13 14

15 10 23

(41%) (43%) (62%)

37 23 37

COPCA

COPC (1680-1780)

67

118

(64%)

185

Relig. treat.

Late ME (1350-1500)

30

13

(30%)

43

Romance

Late ME (1350-1500)

14

3

(18%)

17

Travel

EModE (1500-1710) COPC (1680-1780)

19 10

9 6

(32%) (38%)

28 16

Trial

EModE (1500-1710)

10

8

(44%)

18

Chi-square = 80.48, df = 6, Ρ < 0.001

Be/have + past participle

49

7.3. The level of formality As seen above, some types of text reflecting the more informal level of usage tend to promote the use of the rising variant form. For an overall picture, the use of be and have was studied in texts labelled as "informal", i.e. comedy, diary, journal, private letter, and letter, as against those labelled as "formal", i.e. law, documents, official letter, and sermon (for the situational criteria applied when defining the level of formality, see Rissanen—Kytö—Palander-Collin 1993: 9-10). However, when pooled together in this way, the figures obtained for the "informal" texts do not offer systematic evidence for considering this factor to be of significance in the development of the be/have paradigm.

7.4. The relationship of the text to spoken language The closer a text comes to a record of spoken language, the better one could expect it to reflect ongoing linguistic change. In the classification adopted for the present study, "speech-based" texts have supposedly been taken down in real speech situations, while "scripts" include texts intended to be spoken. The only speech-based texts in the corpora studied are the trials included in the Early Modern English section of the Helsinki Corpus, while the scripts included in this study consist of drama and sermons. Given the relatively limited textual basis, it is hardly surprising that practically no evidence was found to support the hypothesis that a closer relationship of the text to spoken language would have promoted the use of have. Of the 18 examples recorded in the Early Modern English section of the Helsinki Corpus, ten occur with be and eight with have. Similarly, the figures obtained for the use of have in "script" texts are lower than those obtained for "written" texts over the centuries included in this survey.

7.5. Oral versus other genres Such factors as text type, level of formality and the influence of the relationship of a text to spoken language could also be combined to describe the position a text holds in the continuum, extending from the more

50

Μ er ja Kytö

oral to the more literate genres (see the Introduction to this volume). The texts defined as "oral" in this study include drama, fiction and (private) letters, diaries and trials. No evidence was found in this study which pointed to the role of the orality factor in the development of the be/have paradigm.

7.6. Men and women writers The data available for assessing the impact of the author's sex on the choice of the forms is scarce from the earlier periods. Women writers are represented to some extent in the Early Modern English section of the Helsinki Corpus and more systematically so in the ARCHER Corpus. A number of women writers can also be found among the letter writers in Part Β of the Century of Prose Corpus, but as there are only a handful of examples, this corpus will be omitted from this section of the presentation. The same holds for some less represented text types in the Helsinki Corpus (e.g. official letters). For meaningful results, only those types of text that can be found as produced by men and women authors in the Helsinki Corpus or the ARCHER Corpus will be considered, that is, fiction, journal, letter, drama, science, and sermon (the texts by anonymous authors will be excluded from the analysis). In the Late Middle English data women writers tend to use the rising form have to a greater extent than men writers. In the Early Modern English period the figures obtained for both men and women authors remain more or less even (see Table 16). Table 16. The use of be/have by men and women writers in the Helsinki Corpus: fiction, diary, private letter, religious treatise, travel.

be ME (1350-1500)

EModE (1500-1710)

have

Total

Women Men

17 50

11 (39%) 20 (29%)

28

Women Men

26 80

9 (26%) 33 (29%)

35 113

70

Be/have + past participle

51

This is the case with the results obtained for the first subperiod in the ARCHER Corpus (1650-1700) as well. However, from the 1700s on, up until the 1900s, women writers remain more conservative and systematically prefer the form be (see Table 17). Table 17. The use of be/have by men and women writers in the ARCHER Corpus: fiction, journal, letter, drama, science, sermon.

ARCHER la (1650--1700) ARCHER2a (1700--1750) ARCHER2b (1750--1800) ARCHER3a (1800--1850) ARCHER3b (1850--1900) ARCHER4a (1900--1950) ARCHER4b (1950--1990)

Women Men Women Men Women Men Women Men Women Men Women Men Women Men

be

have

14 75 24 50 29 83 14 26 5 34 1 9 4 11

8 34 9 34 28 118 31 94 15 234 12 70 69 170

Weighing the influence of the extralinguistic

Total (36%) (31%) (27%) (40%) (49%) (59%) (69%) (78%) (75%) (87%) (92%) (89%) (95%) (94%)

22 109 33 84 57 201 45 120 20 268 13 79 73 181

factors

By way of summary, of the main extralinguistic factors thought to have influenced the development of the be/have paradigm, text type and the author's sex seem to have been of most potent influence across the centuries (the role of foreign influence will be discussed under the treatment of linguistic factors in section 8.5.1, concentrating on the loan words used as main verbs). I shall now turn to the role of a number of essential linguistic factors; for the fundamental axes (stative versus mutative, and action versus process uses), see section 6 above.

52

Merja Kytö

8.

Linguistic factors

8.1. Three verb constructions As a variant form necessarily occurs in one of the four verbal constructions considered in the present study — present perfect, past perfect, perfect infinitive and the -ing construction — the influence of this structural context is intrinsic. This means that the other linguistic factors to be dealt with are, in some sense, optional contextual co-occurrence features rather than independent factors. They are of interest not only for their own value but also vis-ä-vis the verb construction found in the context. The possible combined effect of these contextual factors will be taken into consideration in the discussion which follows.

8. J. 1. Perfect versus past perfect The past perfect, which highlights the perfectivity of action, paved the way for the rise of have; for corpus examples, see (9a) and (9b). (9)

a.

Nothing material happened during our journey to Mosco, where we arrived on the thirtieth of July, and joined the ambassador, who had arrived there two days before, having passed us on the road. (COPC_B/Travel/John Bell/1719)

b.

By this I would be understood to mean, the additional articles of news collected out of the daily historians, from all which, I never could conceive, what kind of advantage can redound to a rational creature, who can neither receive instruction or entertainment, by reading that Mr Suchaone died at his country house, when perhaps the gentleman is in perfect good health; and if Squire Rentroll is arrived in town with a grand retinue, I apprehend it in no way interesting to any man breathing, except his tailor. (COPCB/Periodical/Gray 's Inn Journal! 1753)

From early on, the use of have is more common in past perfect than in present perfect constructions (see Table 18); the present perfect remains one of the last strongholds of be, up until this century.

Be/have + past participle

53

Table 18. Present perfect versus past perfect constructions. Present perfect be

have

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

154 166 43 73 110 48 11

26 37 14 72 63 166 158

Total

605

Past perfect Total

have

Total

180 203 57 145 173 214 169

93 26 (22%) 158 52 (25%) 36 21 (37%) 45 77 (63%) 62 95 (61%) 27 142 (84%) 11 124 (92%)

119 210 57 122 157 169 135

536 (47%) 1141

432 537 (55%)

969

(14%) (18%) (25%) (50%) (36%) (78%) (93%)

Chi-:square = 390.7 df = 6, Ρ < 0.001

be

Chi-:square = 275.2 df == 6, Ρ < 0.001

Again, the rise of have is more noticeable with action verbs than with process verbs (for evidence, see Appendix 3).

8.1.2. Perfect

infinitive

The perfect infinitive, like the past perfect, emphasizes perfectivity and thus promotes the use of have; for corpus examples, see (10a) and (10b). (10) a.

Jay called upon me at half-past three in his gig, to take me to dinner at his father's, Hartham Park. Company at dinner: the Hawkins; Lady Frances W. was to have come with them, but, to my somewhat disappointment, she had been called away to London the day before; (ARCHER3a/Journal/Thomas Moore)

b.

The parliament writs were delivered yesterday. Mr Skelton is still in the Tower, and is in more danger than was at first apprehended. We hear no further of the Dutch fleet. Captain Ouseley is said to be come to town to give his reasons for tossing the Mayor of Scarborough in a blanket. (COPC_B/Letters/1688)

54

Merja Kytö

The use of be was rare as early as the Late Middle English period (see Table 19). Table 19. Perfect infinitive versus other constructions.

Perfect infinitive

Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Total

Other

be

have

Total

0 9 6 4 6 4 2

26 40 6 31 17 55 30

26 49 12 35 23 183 32

247 52 344 91 88 42 130 161 186 179 75 321 23 291

31

205

(87%) 236

1093 1137

(100%) (82%) (50%) (89%) (74%) (93%) (94%)

Chi-•square = 26.29 df =: 6, Ρ < 0.001

be

have

Total (17%) (21%) (32%) (55%) (49%) (81%) (93%)

299 435 119 291 365 396 314

(51%) 2230

Chi·-square = 674.2 df == 6, Ρ < 0.001

Among the various types of perfect infinitive constructions, the presence of a modal auxiliary has been a further Äave-triggering factor (see examples (11a) and ( l i b ) , and Table 20). (11) a.

And trewely, most dred and souueraign lord, gladder ne moor confortable tithinges might neuer haue come, nor in better tyme, for to satisfie and refresshe J>e feruent desir of your poure lieges, J)at haue loong thrusted aftur knowlech of your prosperite, than were your sayd gracious lettres, ... (London Letters (LLETT) 79)

b.

... for my Affairs are a little dubious at present, perhaps I may be gone in half an Hour, perhaps I may be your Guest till the best part of that be spent; (Farquhar, The Beaux Stratagem 8)

Be/have + past participle

55

Table 20. Types of perfect infinitive constructions (the breakdown figures are given for the instances with modal auxiliaries only).

Perfect infinitive + modal auxiliary

Late ME (1350-1500) EModE (1500-1710) ARCHER 1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990) Total

be

have

0 3 5 0 4 4 2

22 30 6 18 13 47 24

(100%) (91%) (55%) (100%) (76%) (92%) (92%)

22 33 11 18 17 51 26

18

160

(90%)

178

45

(78%)

58

Total

Chi-square = 23.48, df = 6, Ρ < 0.001 Perfect infinitive + other than modal auxiliary 13

Total

8.1.3.

-Ing constructions

As -ing constructions express the prolongation and durativity of action, one could expect them to favour have with mutative intransitives (Ryden—Brorström 1987: 193-194; see examples (12a) and (12b); see, also, the discussion on durative contexts, below). In my data the Aave-promoting influence of the -ing constructions is less evident (see Table 21). (12) a.

Saturday, September 20th Having fallen into a sound sleep while drifting with the tide into New York, I was awoke about the first dawn of the morning, ... (ARCHER2b/Journal/William Strickland)

56

Merja Kytö b.

... not only deprived of the favor of God, but also of His image; of all virtue, righteousness, and true holiness, and sunk partly into the image of the devil, in pride, malice, and all other diabolical tempers; partly into the image of the brute, being fallen under the dominion of brutal passions and groveling appetites. (ARCHER2b/Sermon/John Wesley)

Table 21. -Ing constructions. -ing

Other

be

have

Total

be

have

Late ME (1350-1500) EModE (1500-1710) ARCHER1 (1650-1700) COPC (1680-1780) ARCHER2 (1700-1800) ARCHER3 (1800-1900) ARCHER4 (1900-1990)

0 20 9 12 14 0 1

0 2 (9%) 8 (47%) 12 (50%) 20 (59%) 13 (100%) 9

0 22 17 24 34 13 10

247 333 85 122 178 79 24

78 129 41 180 175 363 312

Total

56

64 (53%) 120

Chi·•square = 34.87 df = 6, Ρ < 0.001

Total (24%) (28%) (33%) (60%) (50%) (82%) (93%)

325 462 126 302 353 442 336

1068 1278 (54%) 2346 Chi--square = 620 .0 d f := 6, Ρ < 0.001

Given that the past perfect, perfect infinitive and the -ing constructions all tend to promote the use of have, the influence of the co-occurrence features might be of most interest in instances of the present perfect, the last stronghold of the receding form be. This co-occurrence effect of contextual factors will be taken into consideration in the discussion which follows.

8.2. Durative and iterative contexts; conditional / optative uses As shown above, the semantic distinction between action and process uses has had a profound influence on the long-term development of the

Be/have + past participle

57

be/have paradigm. Other contexts found to have been of importance are durative (see example (13a)), iterative (13b), conditional ((11a) above) and optative ((13c)) contexts, all prone to promote the use of have (the examples cited illustrate the early uses, in particular); for extensive discussion, see Ryden—Brorström (1987). (13) a.

There were in hominum memoria 3. clothiers at one tyme, thus nam id, Style, Kent and Chapman, by whom the toun of Bath then florishid. Syns the death of them it hath sumwhat decayed. (Leland, The Itinerary of John Leland I 143)

b.

On whom [= six grete dogges] Reynart the theef had grete enuye by cause they were so sure that he cowde none gete of them / how wel oftymes hath this fel theef goon rounde aboute this wal / and hath leyde for vs in suche wyse that the dogges haue be sette on hym and haue hunted hym away (Caxton, The History of Reynard the Fox 11)

c.

God wolde ye had nat come here; but ye com never in felyship of worshipfful folke for to do good, but allwayes grete harme. (Malory, Morte Darthur 50)

Though the relevant examples found for the above categories are in the minority if compared with the total of examples studied, these contexts form a strong co-occurrence pattern with the use of have (see Table 22; as different semantic distinctions may overlap in the same context, durative contexts are contrasted against non-durative, iterative against noniterative and conditional / optative against non-conditional / non-optative contexts). Restricting the set of examples included at this stage of the analysis to constructions with the present perfect does not change the relative figures obtained for durative and iterative contexts to any great extent. More evident is the co-occurrence of the verb construction as a have-promoting factor in the conditional / optative category: apart from two examples found in IF clauses, the six examples all occur in instances of the past perfect or perfect infinitive (with a modal auxiliary).

58

Merja Kytö

Table 22. Durative, iterative and conditional / optative contexts.

Durative Other

be

have

24 1100

66 1276

(73%) (54%)

90 2376

46 1296

(78%) (54%)

59 2407

171 1171

(96%) (51%)

179 2287

Total

Chi-square = 13.47, df = 1, Ρ -forms in the pres. ind. sg. (beo, bist, bip) and by forms with initial s(ind. pi. sind(en), subj. sg./pl. sy/syn). As a result, the most typical present indicative paradigm of to be in ME1 is am, art, is in the sg. and bep (or less commonly, be(n)) in the pi. The ind. pi. form aren, a development of OE earon, is represented by four instances only. In the present subjunctive the typical paradigm is sg. be (beo, bie, bo), pi. ben (beort, bien, bon, byri). The losses suffered both by ό-forms in the pres. ind. sg. and by s-forms in the ind. pi. and subj. sg./pl. are something that could be expected from OE developments. In Kilpiö (1993) I point out that there is, generally speaking, a decline in the relative share of ό-forms in the present indicative singular (43.5% in OE2, 30.1% in OE3 and 38.1% in OE4; for a discussion of the unexpected rise in OE4 and its probable connection with the composition of the texts in OE4, see Kilpiö 1993: 100). In subperiod ME1, the corresponding percentage of ό-forms is 28.7% in texts representing MEX/1 and 2.8% in texts representing ME1. As for the losses suffered by J-forms, it has already been seen that from OE1 to OE4 the share of sforms in the present subjunctive declines steadily and the development is particularly marked in the plural. Assuming that the gradual loss of sforms began with present subjunctive forms, both singular and plural, it could be hypothetized that, having gradually lost the support of the subjunctive s-forms, the present indicative plural sind(on), the last stronghold of the Λ-forms, was under pressure of being replaced by other forms. On the analogy of the present subjunctive, the most likely candidate was a bform, and, in fact, by far the most common present indicative plural form in ME1 is bep.

2.3. Middle English, subperiod ME3, with special reference to LALME The texts in Table 3a and 3b have been arranged in two groups.5 The first, smaller group (21,450 words), marked ME2/3, consists of texts with originals dating to period ME2 (1250-1350), but with manuscripts written in ME3. The second, larger group (162,780 words) consists of texts composed in ME3 and preserved in manuscripts written in the same period.

96

Matti Kilpiö

I I (Ν

CM

I I I

I I I I I I

δ •ο -5

I —I

I I οο

I I I I I I I I I I I I

oo

I I I I I oo

I I I I

υ ft

m vi m Ό οο

λ. Ι-ι υ Ο. Ό « _«t es

1> α. δ β

h ^·-'n N H h fΝi iΝn Ή OΝ iOi»»

I I —

I I I I I I I I I I I I

I I "Τ

I I I I I I I I I I I I

(Ν ΟΟ rf

I< r1 NI

η η

I Im

I I I

I I I I

I I I

I I I I I

I 1-Η — ι » -) t/3 z—s co '— w © «Ο -η Γu...wite] 151.24 beon

118

Matti Kilpiö

Notes 1.

The (evidently) rare instances of his representing is have not been disambiguated for OE, ME3 and EModEl. 2. With the exception of sie forms from OE3 and OE4; of these, every 50th instance was coded. 3. For the abbreviated titles of texts used in Tables 2 and 3, see Kytö (1996: 167—230). The final letter in the Helsinki Corpus dialect codings is L if LALME is the source of the definition; otherwise the coding ends in O. Thus EML, EMO = 'East Midland', WML, WMO = 'West Midland', NL, NO = 'Northern', SL, SO = 'Southern'. 4. The numbers of instances given in Table 2 represent absolute frequencies. As the total numbers of words of texts representing MEX/1 is 39,050 and that of ME1 texts 72,010, the numbers of instances for the two groups of texts are not directly comparable. 5. For the abbreviated titles for texts and the dialect codings, see note 3 above. 6. Historically speaking, this construction is an amalgamation of two constructions, the OE impersonal type expressing obligation or necessity (see example (19), a type which changes into a 'personal' variant in ME, and the OE type he is to cumenne 'he is to come' with personal subject; for a discussion, see CHEL 2: 336-337). 7. Imperative forms have been left out of account here, as they are basically only met in copular, not auxiliary or non-copular main verb uses. 8. The progressive remains rare through the Early Modern English period, see e.g. Görlach (1978: 108-109). 9. For a discussion of Early Modern English developments in the verbal group, see Blake (1983: 81-103). 10. This impression is confirmed if we consider all the appropriate -ing forms, both those coded and those left outside the analysis: ME3 contains no more than 11 instances, EModEl as many as 183. Forms in -ing which have been excluded from the discussion include (1) being as a non-verbal noun not functioning in any of the main uses of be; (2) being in the set phrase 'for the time being'.

The verb be

119

References Blake, N. F. 1983 Shakespeare's language. An introduction. London—Basingstoke: Macmillan. CHEL 2 = The Cambridge history of the English language. Volume II: 1993 1066-1476. Cambridge: Cambridge University Press. Forsström, Gösta 1948 The verb 'to be' in Middle English: A survey of the forms. Lund: Gleerup. Görlach, Manfred 1978 Einführung ins Frühneuenglische. (Uni-Taschenbücher 820.) Heidelberg: Quelle and Meyer. HC = Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Hecht, Hans (ed.) 1900 Bischofs Wcerferth von Worcester Ubersetzung der Dialoge Gregors des Grossen. (Bibliothek der angelsächsischen Prosa 5.) Leipzig: Wigand. Kilpiö, Matti 1989 Passive constructions in Old English translations from Latin with special reference to the OE Bede and the Pastoral Care. (Memoires de la Soci6te N6ophilologique de Helsinki XLIX.) Helsinki: Soci6te N6ophilologique. 1992 Dictionary of Old English: BEON. With attested spellings by Robert Millar using material assembled by Haruko Momma. Toronto: Pontifical Institute of Mediaeval Studies. 1993 "Syntactic and semantic properties of the present indicative forms of the verb 'to be' in Old English", in: Matti Rissanen et al. (eds.), 97-116. 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. LALME = Mcintosh, Angus—M. L. Samuels—Michael Benskin—Margaret Laing—Keith Williamson 1986 A linguistic atlas of late mediaeval English. 4 vols. Aberdeen: Aberdeen University Press. Mitchell, Bruce 1985 Old English syntax. I—II. Oxford: Clarendon Press.

120

Matti Kilpiö

Mustanoja, Tauno F. 1960 A Middle English syntax. Part I: Parts of speech. (M6moires de la Societe N6ophilologique de Helsinki 22.) Helsinki: Societ6 Νέοphilologique. Nevalainen, Terttu—Helena Raumolin-Brunberg 1993 "Early Modern British English", in: Matti Rissanen et al. (eds.), 5373. Nevanlinna, Saara—Päivi Pahta—Kirsti Peitsara—Irma Taavitsainen 1993 "Middle English", in: Matti Rissanen et al. (eds.), 33-51. Nickel, Gerhard 1966 Die expanded Form im Altenglischen. Vorkommen, Funktion und Herkunft der Umschreibung 'beon/wesan'+Partizip Präsens. Neumünster: Wachholz. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London—New York: Longman. 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. Samuels, M. L. 1985 "The Great Scandinavian Belt", in: Roger Eaton—Olga Fischer— Willem Koopman—Frederike van der Leek (eds.), Papers from the 4th International Conference on English Historical Linguistics, Amsterdam, 10-13 April, 1985 (Amsterdam Studies in the Theory and History of Linguistic Science: Series IV — Current Issues in Linguistic Theory 41.) Amsterdam—Philadelphia: Benjamins, 269281.

Re-phrasing in Early English: The use of expository apposition with an explicit marker from 1350 to 1710 Päivi Pahta and Saara Nevanlinna

1.

Introduction

1.1. The aim of the study The aim of our study is to examine the structure and use of nonrestrictive expository apposition in Late Middle English and Early Modern English texts in the Helsinki Corpus (HC). Originally we were interested in the use of appositional constructions in different types of texts and in the factors motivating re-phrasing in the form of apposition, but since apposition as such has not been researched in any detail in this period, we felt it necessary to discuss the structure of apposition as well. We approach the problem by reflecting the present-day situation onto Late Middle English and Early Modern English in order to see to what extent the outline of modern usage was already established in those days. The focus of our interest has been to trace the development of appositional constructions containing an explicit marker during the centuries framing the transition period between Middle English and Early Modern English (1350-1710). We have also mapped the general situation in Old English and Early Middle English, which still very much reflects the Old English usage (see Nevanlinna—Pahta forthcoming).

122

Päivi Pahta and Saara Nevanlinna

1.2. Earlier research on apposition Although apposition is discussed in most works on contemporary English grammar and linguistics, it has generally been done perfunctorily. 1 The semantics of apposition is touched upon, e.g., in Halliday—Hasan (1976), Lyons (1979), Bolinger (1979) and Jackendoff (1985). The syntactic characteristics of apposition are discussed, e.g., in Matthews (1981), while Huddleston (1971) and Quirk et al. (1972 and 1985) discuss fairly thoroughly both semantic and syntactic features of appositional constructions. There are also a number of special studies, which include those by Hockett (1955), Sopher (1971), Burton-Roberts (1975) and Meyer (1987, 1989, 1991). Appositional reformulations as a stylistic device have been discussed by Blakemore (1993). The most detailed discussions of the topic so far are by Acuna (1996) and the corpus-based study of apposition in contemporary English by Meyer (1992).2 Language historians have generally shown less interest in apposition (cf. Strang 1970). Denison (1993) refers several times to interesting structures involving apposition (see General Index p. 526), but since he concentrates on verbal constructions, apposition remains only marginal to his theme. Some aspects of apposition in Old English, either verse or prose, have been discussed in a number of special studies, for instance, in Peltola (1960), Robinson (1985) and Blockley (1989), as well as in the comprehensive Old English Syntax by Mitchell (1985). The use of repetitive word pairs in Old and Middle English is discussed in Koskenniemi (1968) and (1983), and the use of the parenthetic as who say and its variants in Old and Middle English in Nevanlinna (1974) and (1993). Some appositional constructions used in the presentation of terminology in Middle English medical prose are discussed in Wallner (1987), Jones (1989) and Norri (1992).

1.3. Apposition as a grammatical relation Apposition has been defined in a number of ways. The treatment of apposition is summarized in A Dictionary of Linguistics and Phonetics by David Crystal (1985: 20), who states that "apposition(al)" is "a traditional term retained in some models of GRAMMATICAL description for a sequence of units which are CONSTITUENTS at the same grammatical level, and which have an identity or similarity of REFERENCE". He goes on to add that there are "many theoretical and methodological problems in de-

Expository apposition

123

fining the notion of apposition, because of the existence of several constructions which satisfy only some of these criteria, and where other SEMANTIC or syntactic issues are involved, as in titles and other designations".3 The narrowest type of definition is given by Fries (1952: 187), who argues that the appositives must be coreferential, juxtaposed noun phrases. Curme (1931) admits as appositions some other constructions, as well, such as appositive genitives (p. 84, the vice of intemperance), appositive adjectives (p. 93, the room above), and predicate appositives (p. 30, he came home sick). In most instances these are juxtaposed. In another work (1970) he recognizes an apposition in the form of an explanatory remark comprising a whole dependent or independent clause. He enumerates coordinative conjunctions connecting words, phrases or sentences and introducing an explanation, such as namely, to wit, that is (to say), for example, etc., but he does not yet recognize them as apposition markers. Jespersen (1961) has a fairly broad definition of apposition including certain uses of reflexive pronouns (vol. VII, p. 172, you yourself must set some tasks) and participles (vol. V, p. 406, he sat smoking). He also includes among the appositional constructions some clauses which are in apposition with noun phrases (vol. Ill, p. 27, their idea that priests are infallible). Sopher (1971: 412) points out that the head group (the first appositive) and the appositional group (the second appositive) may belong to different grammatical categories, and that the morphological or structural features are not relevant, since apposition is a syntactic category. He further postulates that apposition is a grammatical category distinct from both coordination and subordination, and that the appositional groups cannot be linked by and or any other coordinating conjunction. Sopher presents both notional and formal criteria. Both appositional groups have a single referent, and are notionally interchangeable, so that the second appositional group may replace the first without altering the syntactic structure of the sentence. Where the introduction of an apposition marker is not possible, notional equivalence and the single referent may be tested by the alternative criteria, namely the test of "interchangeability" or "replaceability", and the "and" test. Huddleston (1971: 251-5) argues that in transformational grammar there is a tendency to use the term "appositive" in the sense of "nonrestrictive", though traditional grammar uses "apposition" for both restrictive and nonrestrictive constructions. He distinguishes three main types of apposition: (i) Characterizing apposition derived from an underlying rela-

124

Päivi Pahta and Saara Nevanlinna

tive. No apposition marker is used in this type, (ii) Identifying or equative apposition, in which the second (subordinate) member may be a dependent or independent clause. The optional markers that is, viz., namely, etc., can be used, (iii) Narrowing apposition, which includes exemplification and particularization. (See also Burton-Roberts 1975 and Matthews 1981.) Quirk et al. (1985: §17.65 ff.) treat apposition primarily as a relation between noun phrases, "constituents of the same level", although they also refer to appositive postmodification by /Aa/-clauses (§17.26, the news that his team had won), appositive nonfinite clauses (§17.35, the appeal to give blood), and appositive prepositional phrases (§ 17.47, the news of the team's victory). According to Quirk et al. (1985: §17.65) the appositives are either identical in reference or the reference of one is included in the reference of the other. Apposition may be full or partial, strict or weak, and restrictive or nonrestrictive, and appositional constructions are combinations of these various categories. In full apposition the second appositive is omissible without the sentence becoming unintelligible, whereas this is not so in partial apposition. In strict apposition the two appositives belong to the same syntactic class, whereas in weak apposition they belong to different syntactic classes. In restrictive apposition the second appositive is necessary for the identification of the first. In nonrestrictive apposition, the appositives are in separate information units, providing relatively independent information. The first appositive normally acts as the defined expression and the second has a defining role. In Present-day English, the defining role of the second appositive is reflected by the fact that it is marked as parenthetic by punctuation or intonation. Quirk et al. (1985: §§17.74-17.87) also classify appositions into various semantic classes ranging from different degrees of equivalence (appellation, identification, designation and reformulation) to attribution, and inclusion (exemplification and particularization). In recent discussions of apposition, both Acuna (1996) and Meyer (1987: 101 and 1992: 3) argue that all earlier definitions of apposition, taken individually, are inadequate or incomplete. Most definitions are too narrow, whereas some admit as appositions constructions which should not be included (e.g., Curme's category of appositive adjectives, see p. 123 above). Meyer points out that the difficulties which linguists and grammarians have had in defining apposition are caused by the fact that they regard it as primarily a syntactic relation. Meyer argues that apposition is not merely a syntactic relation; it must also be viewed as a semantic and pragmatic relation, and should be defined in terms of those characteristics which distinguish it from other grammatical relations — com-

Expository apposition

125

plementation, modification, parataxis, peripheral elements, and coordination. In Apposition in contemporary English, M e y e r (1992) provides a comprehensive analysis o f the syntactic, semantic and pragmatic characteristics of apposition based on his study of appositions in three computer corpora of English. 4 H e views apposition as a grammatical relation having various realizations, which are described in Table 1, adopted from Meyer (1992: 6). Table 1. The linguistic characteristics of units in apposition (Meyer 1992: 6).

Syntactic characteristics

Semantic characteristics

Pragmatic characteristics

Syntactic form: nominal apposition, NP/clause apposition, appositions with obligatory markers, non-nominal apposition

Semantic relations: coreference, cataphoric reference, part/whole relations, synonymy, hyponymy, attribution

Information structure: new or partially new information in second unit of apposition

Syntactic function: subject, object, complement, adverbial Linear structure: single/double/ triple apposition, juxtaposed/ unjuxtaposed apposition Hierarchical structure: binary/non-binary apposition

Semantic classes: identification, appellation, exemplification, particularization, characterization, paraphrase, reorientation, selfcorrection Semantic integration: restrictive/ nonrestrictive

Functional potential: tendency of some appositions to occur more commonly in some contexts than others

126

Päivi Pahta and Saara Nevanlinna

In this study we have adopted the outline of Meyer's definition of apposition as a syntactic, semantic and pragmatic relation with some modifications. We decided to adopt this rather broad definition of apposition for two reasons. Working on historical data, we did not want our theoretical framework to impose preconceived restrictions on our findings. Having worked through the data, we felt that from a functional point of view we could not exclude non-nominal phrases, clauses or sentences from our discussion. Syntactic variation in the form of the units in apposition did not seem to affect the linguistic function of expository apposition, i.e., that of providing additional information. Some of the syntactic and semantic relations that qualify as appositional according to Meyer's definition remain outside the context of this study. We only deal with nonrestrictive apposition. In restrictive apposition the second appositive restricts the reference of the first, while in nonrestrictive apposition the second appositive provides relatively independent information about the first, explaining, illustrating or naming an object, person or idea expressed in the first appositive to which it is notionally equivalent and to which it is or may be linked by an explicit apposition marker.5 We are also exclusively concerned with those nonrestrictive appositional constructions where the appositives are linked by such markers. We have further restricted this study to expository appositions which display what Quirk et al. call "a relation of equivalence" (see p. 124 above). Expository apposition includes instances of coordinative apposition, where the units of apposition are joined by and or or. Apposition resembles coordination in that the two units are typically constituents of the same level. The units of coordinative apposition are synonymous, coreferential or related by cataphoric reference. In coordinative apposition no reduction is implied, whereas, for instance, non-appositional (segregatory) coordination can be regarded as implying the reduction of two or more coordinated clauses (Quirk—Greenbaum 1973: §7.21, §9.41). Table 2 shows the linguistic characteristics of the type of apposition to be discussed in this study.

Expository apposition

127

Table 2. The linguistic characteristics of nonrestrictive expository apposition with an explicit marker discussed in this study. Syntactic characteristics Syntactic form: nominal apposition, NP/clause apposition, non-nominal apposition Syntactic function: subject, object, complement, adverbial, verb Semantic characteristics Semantic relations: coreference, cataphoric reference, synonymy Semantic classes: identification, appellation, characterization, paraphrase, revision Pragmatic characteristics Generic distribution: tendency of some structures to occur more commonly in some contexts than in others

1.4. Apposition markers In Present-day English numerous expressions are available for marking nonrestrictive apposition (see Quirk et al. 1985: §17.73). Markers of apposition are used to explicitly express the semantic relationship existing between the appositives. They indicate whether the information supplied by the second appositive is expository or exemplifying. 6 In an expository relation, where the appositives are referentially or semantically equivalent,

128

Päivi Pahta and Saara Nevanlinna

i.e., they present the same truth value (Leech 1988; Allwood et al. 1977), markers such as that is (to say), namely or in other words can be used in Present-day English to introduce the second appositive. In an exemplifying relation, where the units are referentially and semantically only partially equivalent, i.e., the reference of one is included in the reference of the other, markers such as for example, e.g. and say can be used in Present-day English (Meyer 1992: 97). Markers of apposition can be either obligatory or optional (see Meyer 1992: 25-26). The use of obligatory markers is governed by semantic and syntactic reasons and their omission leads to an ungrammatical construction. An obligatory marker of apposition is required, e.g., in exemplifying apposition, to unambiguously mark the part/whole relation which exists between the appositives. In the use of optional markers of apposition, pragmatic considerations are essential. The only marker of expository apposition which Meyer has classified as obligatory is or. The other expository markers included in his study (that is, that is to say, i.e., viz, namely, and in other words) are classified as optional. The use of and as an apposition marker is not discussed by Meyer at all, although and is included in the list of apposition markers in Quirk et al. (1985: §17.73). In Meyer's study of Present-day English usage, optional markers of expository apposition were fairly rare, occurring in less than three percent of the appositional constructions in the corpora. The majority of the optional markers in his study appeared in the learned genres of the corpora, which suggests that they are indicators of formal style and would accordingly be inappropriate in less formal Present-day English styles such as fiction, press reportage or spontaneous conversation (Meyer 1992: 96-98). One of the aims of this study is to see whether this could be true for earlier periods of English as well. The device of using a special marker to indicate apposition is found in Old English, though appositional constructions without a marker are more common.7 Apposition was not uncommon in Old English as a stylistic device, modelled on Latin rhetoric (cf. Bloomfield 1965: 6). Robinson (1985: 60) points out that in Beowulf apposition relating to a particular subject serves as a means of emphasis. It is also a "retarding device" that enables the reader "to consider an object as an action from more than one perspective", as well as making the poet "move on swiftly and easily".8 The markers of nonrestrictive expository apposition that occurred in our corpus are listed in Table 3. The list is somewhat different from the corresponding list of markers in contemporary English (cf. Quirk et al. 1985: §17.73 and Meyer 1992: 26, 97). One of the commonest present-day

Expository apposition

129

markers, in other words, did not occur in our data. The earliest example of the phrase in the Century of Prose Corpus (1680-1780) is from 1690,9 which indicates that it appeared towards the end of the period we have studied. On the other hand, the list contains some markers which are no longer used, such as yea, variants of as who say, i. (an earlier variant of i.e.), and a number of unique markers. Some of the unique markers were detected quite accidentally, which suggests that other similar phrases could perhaps be found in contemporary texts. Furthermore, particularly in the Middle English sections of our corpus, the markers were manifested in a variety of different spellings. In Table 3, the markers occurring several times are given in their Present-day English form, whereas the unique markers are given in the form in which they were found in the data. Table 3.

The markers of expository apposition in the Late Middle English and Early Modern English subperiods of the Helsinki Corpus.

Markers occurring several times

Unique markers

and, & or (else/rather) (and) that is that/this/which/it is to say (that is) to wit yea (and) as who say/saith (and/as) namely (that is) to understand nay as scilicet, sc. id est, i.e., i. viz.

pat is to menynge that is to meane this to seyn as who seith and it is this pat ben pese as moche to say as that is as myche to say as which is to say by interpretation which is to say being interpreted vel

In Present-day English, some indicators of apposition can either precede or follow the second appositive, though structures of the latter type are very rare.10 In our data we found no examples of apposition where the marker was placed after the second appositive. Words used as apposition markers are not necessarily restricted to appositional use only. The apposition markers and and or are also commonly used as explicit coordination markers. Gradual semantic change may

130

Päivi Pahta and Saara Nevanlinna

also cause changes in the use of an apposition marker, as with namely, originally used as a marker of particularization. Its use in the modern function dates from the period surveyed in this study.

2.

Classifying appositional constructions

Appositional constructions can be classified in a number of ways which focus on their various syntactic, semantic or pragmatic characteristics, some of which we concentrate on here. Attention is primarily paid to the use of the explicit markers and the semantic classes of expository apposition and their distribution across different types of text. First the syntactic form and function of units in expository apposition will be briefly discussed.

2.1. Syntactic form and function of apposition Defining apposition in the way we have done, a wide range of syntactic constructions can be admitted as appositions, including phrases, clauses, or sentences. Phrasal appositives may be, to list the central constructions, noun phrases, verb phrases, prepositional phrases, predicative phrases, adverb phrases or adjective phrases. Appositional clauses may be predicates, dependent or independent clauses. The appositions in the corpora consisted of two or, rarely, more appositives which had the following syntactic forms: nominal apposition (noun phrases in apposition), as in example (1), a noun phrase in apposition with a clause or sentence (2), or non-nominal apposition, in which the appositional units are clauses, sentences or non-nominal phrases (3)—(4). (1)

And J requyre and byseche alle suche that iynde faute or errour that of theyr charyte they correcte and amende hit. (Caxton, The Prologues and Epilogues 78) 'And I require and beseech all those who find fault or error that they should correct and amend it.'

Expository apposition

131

(2)

The office of a tutor is firste to knowe the nature of his pupil, that is to say, wherto he is mooste inclined or disposed, and in what thyng he setteth his most delectation or appetite. (Elyot, The Boke Named the Gouernour 24)

(3)

It is harde, ye, it is impossyble, that a man may have alle joye in this worlde (In Die Innocencium 12) 'It is hard to believe, or rather, it is impossible, that a man may enjoy everything in this world'

(4)

And be it further enacted, That all Justices of the Peace ... findinge any such Oxen ... shall take and seize the same as forfeited, and shall give and distribute the same to Prisoners and other poore Folkes by theire discretion. (The Statutes of the Realm IV 1058)

Nonrestrictive appositional structures can also be incomplete in that there is ellipsis in the second unit (cf. Quirk et al. 1985; Halliday—Hasan 1976; Matthews 1981). The term ellipsis is used here in the sense of grammatical reduction (Quirk et al. 1985) or contraction (Matthews 1981). The parts implied can be inferred from the context; in other words, they can be recovered from the first unit, where they are expressed in the same form. The defective second unit may thus contain only what is added to the first appositive, as in examples (5) and (6). In the examples the parts omitted are indicated in square brackets by way of illustration. (5)

Laborande in a febre efimera, sc. [a febre efimera] of hotnesse, if vertue and age suffre, mynusche he of cephalica of J>e rijt arme in somere, in wynter of J)e lifte. (A Latin Technical Phlebotomy 41) 'For treating ephemeral fever, i.e., (ephemeral fever) caused by heat, if the condition and age of the patient allow, let blood of the cephalic artery of his right arm in the summer and of his left arm in the winter.'

(6)

wine (that is to say [wine] of midle age) is hote in the seconde degree (Turner, A New Boke of the Natures and Properties of All Wines B3R)

132

Päivi Pahta and Saara Nevanlinna

In example (7) there are two appositions which both have ellipsis in the second appositive. The second apposition is embedded in the second unit of the first apposition. (7)

'Honours thy fadyre and pi modyre.' That es, [honoure thy fadyre and pi modyre] in twa thynges, fiat es, fin] bodyly and gastely [thynges], (Rolle, Prose Treatises 11) 'Honour your father and your mother, that is, in two things, that is, in bodily and spiritual.'

Units of apposition can appear in a variety of syntactic functions within phrases, clauses, or sentences. In Present-day English, appositions are most common in functions connected with the noun phrase (Meyer 1992: 34-35, Quirk et al. 1985: §17.65), and this seems to be the tendency in our historical data as well. Appositional constructions are very common in the function of subject (8) and object (9). (8)

Forsothe euer eithir was nakid, that is, Adam and his wif. (Wycliffe, The Old Testament Genesis 2 25)

(9)

Tak pentafoyloyn, id est quintfoyle, & welle it wele in water (The 'Liber de Diversis Medicinis' in the Thornton Manuscript 13) 'Take pentafilon, that is, cinquefoil and boil it well in water'

In our data appositions were quite frequent in the function of the verb, as in examples (10) and (11), whereas in Present-day English these are very rare (Meyer 1992: 36). (10)

For the reformynge of the great Mischiefes and Inconveniences that daylie growe and increase by reason of the pestering of Houses with div~se Famylies... (The Statutes of the Realm IV 852)

(11)

l>e splene ... is heled or leched aboute J)e lefte side, as Galien saij). (The Cyrurgie of Guy de Chauliac 64) 'The spleen is healed or cured along the left side, as Galen says.'

Expository apposition

133

Appositions were also found as subject complement (12), object complement (13), and adverbial (14). (12)

t>e sexte dedly synn es 'slewthe or slawenes' (Gaytryge, Sermon 13) 'The sixth deadly sin is sloth or slowness'

(13)

My deere harte, I cannot forgett the obligation that I am bound in to the, that is to write to thee (Thomas Knyvett, The Knyvett Letters 59)

(14)

Whereupon sone after that is to wit, on the friday the thirtene day of Iune many Lordes assembled in the tower (More, The History of King Richard 111 46)

2.2. The semantic classes of apposition Expository appositional constructions can be divided into various semantic classes. This section shows the semantic classification used in our study, for which we are obliged to Meyer (1987: 103 and 1992: 73-82), and to Quirk et al. (1985: §§17.74-17.81). Quirk et al. (1985) distinguish between four relations of equivalence in apposition (appellation, identification, designation and reformulation). Their classification includes only noun phrases in apposition, but it seems to be applicable to other appositional constructions as well. The same semantic classes are also found in Meyer (1987: 113). In his more extensive corpus-based study Meyer (1992) made some adjustments to his earlier classification, redefined some classes and added some classes to account for semantic relationships not included in Quirk et al. (1985). He divides appositions into three basic types according to the various ways in which the second unit of apposition conveys new information about the first unit. The new information provided by the second unit may be more specific than the information provided by the first unit (identification and appellation); it may be less specific than the information given in the first unit (characterization); or the two units may be equally specific (paraphrase, self-correction, and reorientation). We adopt Meyer's scheme for the semantic classification of appositions according to the specificity of the units. The individual classes that we have distinguished within the last of the major semantic categories differ to some extent from those that Meyer derives

134

Päivi Pahta and Saara Nevanlinna

from his modern corpus material (see pp. 140-142 below). Assigning appositional constructions to these classes is not unproblematic and in some cases the interpretation is left to the discretion of the scholar. Table 4 shows the semantic classes into which we have classified appositions and the frequency of these classes in our data. In the following sections we discuss the semantic classes in more detail and provide examples of different types of constructions belonging to these classes. Table 4. The semantic classes of apposition. Ν

%

Identification Appellation Characterization Paraphrase Revision

228 20 8 3478 135

5.9 0.5 0.2 89.9 3.5

Total

3869

100

SEMANTIC CLASS

2.2.1.

Appositions in which the second appositive is more specific than the first appositive

Identification (Meyer 1992: 74; Quirk et al. 1985: §17.77; identifying or equative apposition in Huddleston 1971). The second appositive, which is more specific than the first one, identifies what is referred to in the first appositive. Two subtypes can be distinguished. In the first subtype the appositives are coreferential, i.e., they refer to the same "piece of reality" (Leech 1981: 156), as in examples (15) and (16). (15)

who J)at is absent paie pe peyne, jDat is to witene, a pound of wex (Returns 55) 'he who is absent shall pay the fine, that is, a pound of wax'

(16)

It remaineth now, that wee consider the thing prescribed, namely wherein we must bee built. (Hooker, Two Sermons upon Part of S. Judes Epistle 41)

Expository apposition

135

The first appositive can also be a pro-form (a plural personal pronoun, an indefinite pronoun, a numeral, or the like) coreferring to the second appositive (17). (17)

It is to wete |)at in flebotomie 4 pyngis are principalli attendid: sc., custome, tyme, age, & vertue (A Latin Technical Phlebotomy 39) 'It is to be known that in phlebotomy four things are mainly to be attended to, that is to say, custom, time, age and virtue'

The second subtype of idenfication contains appositions in which one of the units has no referential value, i.e., it is a clause or a sentence (Meyer 1992: 74-75). In this type of apposition the units are related by cataphoric reference; the first unit, usually a noun phrase, refers cataphorically to the second. Examples (18) and (19) illustrate this kind of identification. (18)

But saynte Austyne saythe a comfortable worde again, to them that gyue theyr tythes truely, that is to saye: Decimae sunt tributa egentium animarum: Tithes are tributes or rewardes to nedye soules. (Fitzherbert, The Book of Husbandry 37)

(19)

And if J>at hors schal be y-reden while his back is sore £)en jx>u most use pe cure aforsaide j}at is to say. at euene take of pe sadel. & wasche it with hot wyn or vryne. (A Late Middle English Treatise on Horses 127) 'And if the horse is ridden while his back is sore then you must use the aforesaid cure, that is to say, in the evening take off the saddle and wash it with hot wine or urine.'

The commonest present-day markers in the relation of identification are namely, that is, that is to say (Meyer 1992: 74). In our data the markers used are (and) that is, that is to say, (that is) to wit, id est, scilicet, as, (as) namely, and viz., some of which have several spelling variants. Appellation (Meyer 1992: 76; Quirk et al. 1985: §17.76). Appellation resembles identification, but in addition to identifying the referent of the first unit, the second unit also names it. Examples (20) and (21) illustrate this kind of usage.

136

Päivi Pahta and Saara Nevanlinna

(20)

Aftir these thingis Jhesus wente ouere the see of Galilee, that is Tiberias. (Wycliffe and Purvey, The New Testament John 6 1) 'After these things Jesus went over the sea of Galilee, that is, Tiberias.'

(21)

And withyn shorte tyme after, I, standynge yn Cheapesyed, sawe these iiij ryed ['ride'] throwe Chepe, (that is to saye,) kynge Phyllype, quene Marye, cardynall Poole, and Steven Gardynar (Mowntayne, Narratives of the Days of the Reformation 209)

The commonest apposition markers in Present-day English are that is, that is to say, in other words (Meyer 1992: 76). In our samples only that is and that is to say are found in Middle English and that is to say/meane, (and) namely, and viz. in Early Modern English.

2.2.2.

Appositions in which the second appositive is less specific than the first appositive

Characterization (Meyer 1992: 78; designation in Quirk et al. 1985: §17.79). This is the converse of identification and appellation; the second appositive provides some general characteristic of the first. The appositives are coreferential noun phrases (or attributive, cf. Jackendoff 1985: 73). Characterization is illustrated in example (22). (22)

he Jjat lokuj) in Godus lawe, f>at is lawe of parfijt fredom, and dwelluj) parfijtly in J?is lawe by al his lif (Wycliffite Sermons 27 1 590) 'He that observes God's law, that is, the law of perfect freedom, and abides completely to this law all his life'

When that is is used as a marker, it is often difficult to distinguish the demonstrative that from the relative pronoun that, as in example (23). (23)

ye shullen do no thyng which may in any manere displese God, that is youre creatour and makere. (Chaucer, The Tale of Melibee 234.CI) 'You shall do nothing that can in any manner displease God, that is, your creator and maker.'

Expository apposition

137

In our corpus characterization is not a frequent class of apposition." The markers that is and that is to say occurring in our data may also be used in Present-day English (Meyer 1992: 78).

2.2.3.

Appositions in which the second appositive is as specific as the first appositive12

Paraphrase (Meyer 1992: 79; reformulation in Quirk et al. 1985: §17.80). The units of this type of apposition are related by synonymy and can represent lexical synonymy or be synonymous phrases, clauses or sentences (cf. Burnley 1992: 462-463). Lexical appositive units display degrees of synonymy varying from absolute synonymy at one end of the scale to speaker synonymy at the other (Cruse 1986: 265-268); this can also be used for describing the synonymous relation in phrasal apposition. In absolute synonymy the units have identical meanings in any contexts, their contextual relations are identical, as in example (24). (24)

J)ei sehe wer loth & not wylly to do swech thyngys (Kempe, The Book of Margery Kempe 1 55) 'though she were reluctant and not willing to do such things'

The second unit is often added in order to provide a more familiar variant, as in the following examples, where the second unit provides a native translation variant of a Latin term (25) or name (26). (25)

Jjre substaunces ben made of f)e chyle by decoccioun (i. sepinge) in the lyuer (The Cyrurgie of Guy de Chauliac 62) 'three substances are made of the fluid of the intestines by a process of decoction, i.e., boiling in the liver'

(26)

archebisshop of Dorobernya, £at is Canturbury (Trevisa, Polychronicon Ranulphi Higden VI 217) 'Archbishop of Dorobernia, that is, Canterbury'

Sometimes the more technical or foreign variant is contained in the second appositive.

138

(27)

Päivi Pahta and Saara

Nevanlinna

Provided alwey that this acte extend not to eny Shipp or Shippes, havyng eny of the seid Wares or marchaundisez, constreyned by tempast of Weder or enemyes to arrive in eny porte or place within this youre Reame. {The Statutes of the Realm II 535)

One of the appositives may also be a regional variant. (28)

Frende, how comeste ]30u in here — oJ)ur hidur — and haste no leueree of my weddynggus? {Middle English Sermons (MS Royal) 17) 'Friend, how do you come here, or hither, and are not wearing a gown for my wedding?'

In speaker synonymy the second unit does not serve as an exact paraphrase of the first unit but rather as an explanation or clarification of the meaning the speaker intended the first unit to have. Semantically, the two units share sufficient salient characteristics to be regarded as roughly synonymous by the speaker, though not absolutely so. Examples (29) and (30) illustrate speaker synonymy in paraphrase. (29)

for God may more do J)an any man may penke or vnderstonde. {The Book of Vices and Virtues 109) 'for God can do more than anyone can think or understand.'

(30)

And this loue or appetyte that euerye thynge hathe to it selfe, procedeth ... by naturall intencion. (Colville, Boethius 80)

As some of the preceding examples show, the class of paraphrase also includes instances of coordinative apposition. The units are either synonyms or near-synonyms and they are joined by and or or. In this capacity the 'inclusive' or and and are variants of the coordinative apposition marker. Wrenn points out (1973: 265) that the Old English oppe meant both 'or' and 'and'. The addition of the second appositive enlarges the conceptual field of the first appositive and thus strengthens the effect of the single phrase, as in (31).13

Expository apposition (31)

139

thilke cerkle that is innerest or most withinne ... it is like a mervayle or miracle. (Chaucer, Boethius 452.C1-2) 'That circle that is innermost or most central ... it is like a marvel or a miracle.'

Coordinative apposition as paraphrase was very common in Old and Middle English as a stylistic device, along with other coordinative word pairs (cf. Mitchell 1985:§ 1809). The two units often alliterated (Oakden 1930; Koskenniemi 1968). In Present-day English, a great many coordinative word pairs, especially nouns, have become idiomatic in colloquial language, with a fixed collocation (see e.g. McMordie 1976: 84). There are several examples where the same two appositives are joined both by and and by or in the same text (cf. Norri 1992: 177): (32)

Lamentation aryseth of foure affections, eyther of a great feare, or dreade, or of a great shame, or of some sorrowe, or els of some hatred ... the people were in a greate feare and dreade to be oppressed (Fisher, The English Works of John Fisher 1 397)

(33)

all thynges that be vnder or subiect to destinie be also subiect and vnder gods prouidence and ordynaunce. To the whyche prouidence or ordynaunce destinye is subiecte it selfe (Elizabeth I, Queen Elizabeth's Englishings of Boethius, Plutarch, &c 107)

In clausal or sentential apposition the principles of truth-functional semantics can be applied to describe the synonymy between the appositives since both units are synonymous if they express the same truth value (Allwood et al. 1977: 61; Leech 1981: 74). This type of usage, exemplified in (34) and (35), is frequently resorted to when there is a need to expound a difficult passage, especially in religious and philosophical writings. The apposition marker acts here as the linguistic correlate of the identity-operator in mathematics and logic (cf. Lyons 1979, 2: 472). The first appositive (i.e. the first sentence) is true for the interpretation that the speaker has given to it in the second clause or sentence, which acts as the second appositive. In apposition expressing truth-functional synonymy there are several markers that had previously undergone grammaticalization, e.g., yea and as who say (see Pahta—Nevanlinna forthcoming).

140

Päivi Pahta and Saara Nevanlinna

(34)

The riches of the sinner is laid vp for the righteous·, that is, the righteous shall inioy that which the wicked gathereth. (Smith, Sermons D8V)

(35)

"drede Jje not, maide Mari for pou hast fownden grace aneyns almyjthy God." As hoo seif). "Pou arte gracious in pi-selfe and also shall be to all Cristen pepull." (Middle English Sermons (MS Royal) 259) 'Don't be afraid, virgin Mary, for you have found grace with almighty God, in other words, you are virtuous in yourself and will be so to all Christians.'

Clauses or sentences in apposition also contain instances where the second appositive provides a translation of the clause or sentence appearing as the first appositive, as in (36). (36)

Sanat, sanctificat, et ditat surgere mane. That is to say, Erly rysyng maketh a man hole in body, holer in soule, and rycher in goodes. (Fitzherbert, The Book of Husbandry 101)

Coordinative apposition can also express truth-functional synonymy; for example, in clausal paraphrase, as in (37). (37)

to dwellyn wyth me & neuyr to departyn fro me (Kempe, The Book of Margery Kempe 1 53) 'to stay with me and never to leave me'

According to Meyer (1992: 79), the markers used in present-day appositions in the class of paraphrase are or, and that is, that is to say, and in other words. In the diachronic data several markers are found. That is, and, or, or else/rather, that/this is (for) to say, as who say/saith, yea (and), id est, i.e., /., viz. and scilicet are all frequent, but there are also a number of rare variants including as, (the which is) as moche to saye as, pat is to vnderstande, pat is to menynge, and this to sayn as who saith.14 Revision (Quirk et al. 1985: §17.80;15 reorientation and self-correction in Meyer 1992: 80-81). In revision the units of apposition are coreferential, but not synonymous as in paraphrase. Two subtypes can be distinguished according to the nature of coreference between the units. In the

Expository apposition

141

first type, which corresponds to what Meyer (1992) calls reorientation, the two appositives are strictly coreferential and there is a close connection between the meaning of the units and their referents in the external world. The second appositive refocuses the reference of the first, that is to say, it provides a different way of viewing it. Examples (38) and (39) illustrate this kind of revision. (38)

Who is elles kepere of good or dryvere awey of yvel but God...? (Chaucer, Boethius 453.C1) 'who else is a keeper of good or a dispeller of evil but God?'

(39)

The tend commandement an pe laste es, Jjat we 3erne noghte J>e wyefe of oure neghteboure. (Gaytryge, Sermon 6) 'The tenth and the last commandment is that we should not desire the wife of our neighbour.'

The writer can use reorientation for reflecting what semantic features of a notion are uppermost in his mind. A good example is Chaucer's illustration of the various sides of the profession of an advocate: (40)

thise oratours or advocattes\ the accusours or advocattes; the deffendours or advocatz (Chaucer, Boethius 449.C1) 'these orators or advocates; accusers or advocates; defenders or advocates'

In the second type of revision, which is equivalent to what Meyer (1992) calls self-correction, the units are related by speaker coreference, i.e., the speaker intends the units to have the same extralinguistic referent, although they have distinct meanings. The second appositive corrects a mistake in the first appositive or provides more accurate information and is usually more emphatic than the first. In Present-day English this kind of reformulation is generally analyzed as a performance error in impromptu speech, and can be appropriately called self-correction. In our written historical data this type of revision occurs most often in constructions where the second appositive provides an expression which is more appropriate or precise than the first unit. Examples of this kind of usage are provided in (41) and (42).

142

Päivi Pahta and Saara Nevanlinna

(41)

Some men here present, ye & not a fewe. woll perauenture muse why & to what entent I brynge in thys (Fitzjames, Sermo Die Lüne ΒIV) 'Some here present, and not just a few, will perhaps wonder why and for what purpose I bring this in'

(42)

in the forepart of his head, between the two fore-leggs, [a flea] has two small long jointed feelers, or rather smellers, MM, which have four joints, and are hairy, like those of several other creatures (Hooke, Micrographia 13.5 210)

In some Early Modern English examples the second unit reveals the ironic attitude of the speaker or writer, as in examples (43) and (44). (43)

For, the scholer, is commonlie beat for the making, whe~ the master were more worthie to be beat for the mending, or rather, marring of the same (Ascham, The Scholemaster 182)

(44)

When he saw the king's temper, and his way of managing, or rather of spoiling, business, he grew very uneasy (Burnet, Burnet 's History of My Own Time 1 I 170)

In Present-day English, the markers of apposition in revision are or, sometimes followed optionally by the conjuncts rather or better, and that is (to say) (Meyer 1992: 81). In our diachronic data and, or, or else/ rather, that is (to say), and yea (and) are found in this relation.

3.

Generic distribution of appositional constructions

Although the use of apposition is in most cases optional, in some contexts appositions are more necessary than in others. The necessity of the second appositive can depend on the linguistic context or, more often, on the situational context, which means that the speaker or writer has to determine whether the audience needs to be supplied with new information about the first appositive. He may choose to add the second appositive to ensure that his audience understands exactly what he means by a certain phrase or clause, or he may just wish to provide some further information about the subject of discussion (cf. Meyer 1987: 114-115). This makes

Expository apposition

143

apposition an additive relation, the second appositive adding information to the discourse (Halliday—Hasan 1976: 248-250). Appositions are most needed and, according to Meyer's study of apposition in Present-day English, most frequent in genres characterized by a low degree of shared knowledge (Meyer 1991: 179).16 Certain types of appositions are also found more often in some text types than in others, their communicative functions being better suited to some text types than to others (cf. Meyer 1992: 92). Furthermore, there may be variation in the frequency of apposition in texts within individual text types, as text types can be very heterogeneous in their use of linguistic features and there may be variation even within narrowly defined groups of texts (cf. Biber 1988, Meurman-Solin 1993a and Taavitsainen 1993).

3.1. Expository appositions across text types Although expository appositions with a specific marker were attested in texts belonging to all text types represented in the HC, they were not evenly distributed across the data. Tables 5 - 9 (see pp. 144-147) show the frequencies of appositions in the five subperiods (ME3-EModE3) of the HC included in our study. Some general remarks can be made about the results presented in Tables 5-9. Although there is temporal variation within all text types, there seem to be some types of writing where expository appositions containing an explicit marker are consistently rare throughout the period of our study. In all texts classified as narrative the incidence of appositions is low. This applies to both imaginative and non-imaginative types of narration, i.e., fiction, romances, diaries, biographies, travelogues and historical writings. In all texts belonging to these text types the incidence of appositions is below 4 per 1000 words. Appositions with an explicit marker are also infrequent in most texts which could be characterized as speech-based, such as drama texts and proceedings of state trials. They are also relatively rare in all correspondence, but in official letters they are more frequent than in private letters in all the subperiods containing both types of correspondence. It is also worth noting that the English Bible texts included in our data contain very few appositional constructions with an explicit marker.

144

Päivi Pahta and Saara Nevanlinna

Table 5.

Frequency of appositions in text types in ME3 (1350-1420). Ν = absolute numbers (for the abbreviations used, see pp. ix-x in this volume).

Text category

Text type

IS STA XX IR XX XX IS IR NI NN IR IS NN XX IR

Science, medicine Document Philosophy Religious treatise Official letter Document Handbook, astronomy Homily Fiction History Rule Handbook, medicine Travelogue Bible Sermon

Total

Word count

Ν

Ν/1000

3 580 1 520 12 630 37 080 5 010 13 870 13 040 7 280 14 320 24 500 5 390 6 110 5 530 20 910 13 460

61 13 105 133 15 41 38 12 21 32 7 7 6 17 7

17.04 8.55 8.31 3.59 2.99 2.96 2.91 1.65 1.47 1.31 1.30 1.15 1.08 0.81 0.52

184 230

515

Table 6. Frequency of appositions in text types in ME4 (1420-1500). Text category

Text type

IS STA IR IR IR XX XX IS XX IS NI

Science, medicine Law Religious treatise Sermon Rule Preface Official letter Handbook, other Document Handbook, medicine Fiction

Word count 6 430 11 240 41 090 25 010 1 810 6 200 3 140 11 560 10 480 5 670 8 760

Ν

Ν/1000

64 110 348 170 12 35 17 53 33 16 18

9.95 9.79 8.47 6.80 6.63 5.65 5.41 4.58 3.15 2.82 2.05

Expository apposition Table 6.

145

Continued.

Text category

Text type

XX XX NN NN XX XX NI IS

Deposition Bible History Biography, saints lives Mystery play Private letter Romance Handbook, astronomy

Total

Word count

Ν

Ν/1000

970 530 600 820 100 490 070 880

4 5 19 5 19 17 12

2.03 1.98 1.51 1.31 0.95 0.87 0.63

213 850

957

1 2 12 3 20 19 19 2

-

-

Table 7. Frequency of appositions in text types in EModEl (1500-1570). Text category

Text type

XX STA EX IS IR IS/EX EX XX NN NN NI XX NN XX NN XX NN XX

Philosophy Law Science, medicine Handbook, other Sermon Educational treatise Science, other Trial History Biography Fiction Official letter Travelogue Comedy Autobiography Private letter Diary Bible

Total

Word count

Ν

Ν/1000

890 790 180 000 470 440 700 970 090 440 550 300 100 570 740 640 060 230

262 303 107 116 75 80 41 59 22 9 19 10 22 11 5 9 8 13

26.49 25.70 17.31 11.60 7.92 7.66 6.12 3.69 1.98 1.65 1.65 1.59 1.56 1.04 0.87 0.85 0.61 0.61

190 160

1 171

9 11 6 10 9 10 6 15 11 5 11 6 14 10 5 10 13 21

146

Päivi Pahta and Saara

Table 8.

Nevanlinna

Frequency of appositions in text types in EModE2 (1570-1640).

Text category

Text type

STA EX EX EX NN IR IS NN NN NN XX XX XX XX XX NI XX NN

Law Science, medicine Science, other Educational treatise Travelogue Sermon Handbook, other Autobiography History Biography Official letter Trial Comedy Philosophy Bible Fiction Private letter Diary

Total

Table 9.

Word count

Ν

Ν/1000

780 620 420 220 780 300 290 120 090 810 660 230 810 880 190 490 590 520

221 75 42 65 58 38 44 13 26 7 7 17 13 6 15 7 5 2

18.76 11.33 6.54 5.79 3.92 3.69 3.58 3.16 2.58 1.46 1.24 1.19 1.10 0.87 0.68 0.56 0.43 0.16

189 800

661

11 6 6 11 14 10 12 4 10 4 5 14 11 6 22 12 11 12

Frequency of appositions in text types in EModE3 (1640-1710).

Text category

Text type

STA XX EX IR IS NN IS NI XX NN

Law Philosophy Science, other Sermon Handbook, other Travelogue Educational treatise Fiction Official letter History

Word count 13 8 11 12 11 10 11 12 5 11

180 820 280 470 370 470 320 040 870 640

Ν

Ν/1000

218 47 60 52 37 31 26 25 11 19

16.54 5.33 5.32 4.17 3.25 2.96 2.30 2.08 1.87 1.63

Expository apposition

147

Table 9. Continued. Text category

Text type

NN XX NN XX XX NN

Biography Trial Diary Private letter Comedy Autobiography

Total

Word count

Ν

Ν/1000

170 760 210 140 740 560

9 9 7 7 5 2

1.46 0.65 0.62 0.53 0.39 0.36

171 040

565

6 13 11 13 12 5

Throughout the period of our study the use of expository apposition with a specific marker is frequent in all statutory writing included in the data. It is also frequent in most writings that can be characterized as expository or instructional, although there is more variation between individual texts within these categories. In the field of secular instruction, appositions containing a marker are frequent in scientific writings, particularly in the medical treatises, whereas in handbooks they are generally less frequent. They are also fairly common in educational treatises, and in some texts belonging to religious instruction. Finally, they are frequently used in most philosophical treatises included in our data. In the following sections we attempt to look more deeply into those text types and individual texts in which expository appositions with an explicit marker are relatively most frequent. The use of appositional markers in the data will be briefly discussed in section 4 below, but it is useful at this point to distinguish between two main types: appositions containing the native speech-based coordinators and and or, and those containing other markers. And and or occur as apposition markers in the majority of appositions registered (see p. 168). In most instances they join two synonyms or near-synonyms in the class of paraphrase.

3.1.1.

Statutory

writings

Samples of statutory writing are contained in all five subperiods included in this study. English was reintroduced as the language of law towards the end of the fourteenth century, so that only some statutory documents are

148

Päivi Pahta and Saara Nevanlinna

available from subperiod ME3, but the other four subperiods contain well over 10,000 words of legislative writing. The language of law contains a number of distinctive features which have been commented on by several writers, including the use of archaic, foreign and uncommon words; repetition of lexical items instead of reference devices; long, thinly-punctuated complex sentences with complicated patterns of coordination and subordination; and the use of passive voice.17 All these features contribute to the distinctive style of legal writing, which has been described as "a complex intricate, even bizarre style of language" completely lacking "colour and humanity" (Maley 1987: 25). Legal language is also frequently described as formulaic, and each species of legal transaction has its own collection of linguistic formulae, which have in the course of time been thoroughly tested before the courts and found to be adequate. The use of the established formulae contributes to the extreme linguistic conservatism of present-day legal language (Crystal—Davy 1974: 194; Hiltunen 1990). According to Maley (1987: 27), the key to understanding legislative language is the fact that it aims at controlling actions by words. The binding force of the statute is attached to its verbal formulation, which has to be extremely explicit and precise, because in a court of law "attention will be paid only to what [a document] appears to declare: any intentions of the composer which fail to emerge clearly are not usually considered in arriving at what the document means" (Crystal—Davy 1974: 193). Consequently, all legal documents must be written in a manner which leaves no place for misinterpretation.18 The pursuit of certainty, explicitness and precision seems to be the explanation for the frequency of appositions in statutory writing. In our corpus, appositions in this category in each subperiod belong almost exclusively to the semantic class of paraphrase (c. 98 per cent of the total), vast majority of them consisting of synonymous or near-synonymous words linked by and or or, as in example (45) (45)

Provided that yf ther be eny lyke Forfeyture co~mytted or done by eny of the Qwenys sarvand~ beyng in her Cheker roule (The Statutes of the Realm III 8)

It has been suggested that the source of the habit of using these wordpairs lies in the multilingual origin of English law, particularly in the period when English was replacing French as the language of law. It is true that both Latin and French have influenced English legal language

Expository apposition

149

heavily, and a large part of the synonymous or near-synonymous wordpairs occurring in our data are of the type where a word of foreign origin occurs in apposition to a native word for the same or a similar concept, as in example (45) above. Further examples are provided by such pairs as act and deed, goods and chattels, lie and extend, lose and forfeit, etc. In the multilingual situation of medieval England there could be uncertainty as to whether such synonyms meant exactly the same thing and the inclusion of both alternatives contributed to the precision of the text and reduced opportunities for misinterpretation (Crystal—Davy 1974: 208). It is not, however, in the fourteenth century that the custom of using synonymous word-pairs appears for the first time in legal language. The same habit is already found in Old English legal documents (cf. Hiltunen 1990). The native tradition is still continued in the Middle English and Early Modern English legal texts (and even in Present-day legal English), and our data includes many appositions consisting of native words, such as by and with, have and hold, keep and hold, have and keep, or all and every, as in (46). (46)

honeste men of the same Parishe being Housholders, shall have full power and aucthoritie by vertue of this Acte to enter into all and everie House (The Statutes of the Realm IV 853)

The habit of using synonymous words in apposition seems to have reached its peak during the early sixteenth century (25.53 paraphrases per 1000 words in EModEl). After that there is a slight decline in their frequency (17.83 per 1000 words in subperiod EModE2 and 16.09 per 1000 words in EModE3), but they remain proportionally more frequent in statutory writings than in any other type of writing included in our data.

3.1.2.

Scientific treatises

The scientific texts in the HC are divided into medical treatises and other scientific treatises which belong to the field of physical sciences. Medical treatises are included in the Late Middle English subperiods and the first two Early Modern English subperiods, whereas other scientific treatises are only included in the three Early Modern English subperiods. With the exception of EModE3, both text types are represented by one text in each subperiod; EModE3 contains two texts labelled under other scientific treatises. Although the use of expository apposition with an explicit marker

150

Päivi Pahta and Saara Nevanlinna

is relatively frequent in all scientific treatises included in our data, it is particularly so in medical writing. In technical texts the use of appositional reformulation contributes to accurate understanding of particular concepts and accordingly helps the reader to understand the surrounding text. Authors who either argue in favour of a particular theoretical position or wish to instruct their readers about a phenomenon or theory are especially interested in presenting their utterances so that their audiences understand them (Blakemore 1993: 101). In medical treatises, which aim at maintaining or restoring health, the need to communicate successfully is vital. Although the medical treatises included in our corpus were aimed at medical professionals, i.e., an audience relatively familiar with the basic concepts of contemporary medicine, and the degree of shared knowledge between the author and the addressee is supposedly high, the complexity of the material (both terminology and procedures), seems to have required attention by way of appositional explanations, as in (47). (47)

a veyne goof) out, icleped vena concaua (i. the holowe veyne) {The Cyrurgie of Guy de Chauliac 62) 'a vein called vena concava, i.e., the hollow vein, goes out'

Appositions of this type seem to be frequent in other similar texts of the period. In a study of the names of sicknesses in Middle English (Norri 1992), appositional constructions serving as an explicatory technique in presenting terminology were frequently found in both academic medical treatises and surgical manuals (both classified as scientific treatises in the HC) in contrast to handbooks or remedybooks, which contained far fewer instances of this technique. The distinction between scientific treatises and handbooks is also visible in the present material: texts classified as handbooks generally contain fewer expository appositions with an explicit marker than those classified as scientific. The Middle English medical treatises included in the corpus are translations from Latin. In the vernacularization of scientific Latin texts at least two factors can be seen as a cause of frequent use of apposition. Firstly, the model provided by the corresponding constructions in the original Latin text (cf. Blake 1992b), and secondly, the need to provide a vernacular translation for technical terms. Both trends, exemplified in (48) and (49) respectively, are visible in the Middle English translation of a Latin technical Phlebotomy included in subperiod ME3, in which the

Expository apposition

151

incidence of apposition is 17.04 per 1000 words, 98 per cent being paraphrases of technical terms mainly joined by the markers i.e., or, and sc. (48)

Laboryng ... of agriacape gasiliontes, i.e., grene jawnes, mynusche of basilica of Jie ryjt arme. (A Latin Technical Phlebotomy 51) 'In treating green jaundice, let blood from the basilica of the right arm.' cf. Latin: Laborans agracabe gaselcontis, id est yctericia viridi (Voigts—McVaugh 1984: 50)

(49)

it is to wyt that some minision is made or done by metacentesyn, i.e., even drawyng (A Latin Technical Phlebotomy 37) 'It is to be known that sometimes blood is let by metacentesis, i.e., bleeding from the side on which the sickness is found' cf. Latin: Notandum igitur quod minutionum alia per antispasim fit, alia per metacentesim. (Voigts—McVaugh 1984: 36)

Not all paraphrases, however, provide a native variant for a foreign term; in some instances the second appositive contains a synonymous technical term of foreign origin. There are cases where an apposition in the Latin text has been translated, but in many of these cases the English translator has provided another appositional explanation, which is sometimes attached to the preceding unit with an explicit marker and sometimes without it, as in example (50). (50)

Fyrst is cephalica, i.e., capitalis or hedd veyne; 2, mediana or cardiaca, i.e. veyne of the herte; 3, epatica or bacilica, lyuer veyne (A Latin Technical Phlebotomy 39) 'The first is the cephalic vein, i.e., the capital or head-vein; the second is the median or cardiac vein, i.e., the vein of the heart; the third is the hepatic or basilic vein, that is, the liver vein' cf. Latin: Prima est cephalica, id est capitalis·, secunda est mediana, id est cardiaca vel vena cordis; tertia est epatica vel basilica (Voigts—McVaugh 1984: 38)

In the translation of Guy de Chauliac's Chirurgia Magna included in ME4 the overall frequency of appositions is 9.95 per 1000 words, which is clearly lower than in the Phlebotomy. Being a surgical manual, its aim is to provide more straightforward, practical advice on surgical procedures

152

Päivi Pahta and Saara Nevanlinna

and perhaps pay less attention to theoretical terminology, although there are some expository passages. The lower frequency of appositions may also be connected with the general nature of the particular translation included in the corpus. Of the three independent Middle English translations of this work, that included in the HC has been characterized by its editor as an attempt to render the text in a more thoroughly English idiom (Ogden 1971: vi). Providing variant terminology is, however, the main source of expository apposition in this text, too; 88 per cent of the appositions were paraphrases of technical terms mainly joined by the markers or or /., the early variant of i.e. (51). (51)

And j)an f)ou schalt byholde J)e nombre and quantite of f>e spondiles (i. pe lynkes ioynt or bones of pe bak). (The Cyrurgie of Guy de Chauliac 65) 'And then you will see the number and quantity of the vertebrae, i.e., the links or bones of the back.'

The medical treatises included in the first two Early Modern English subperiods follow the same model as far as appositions are concerned. Thomas Vicary's The Anatomie in EModEl dates from the mid-sixteenth century, although it is mainly based on the work of a thirteenth-century surgeon (Bennett 1952: 108). The incidence of apposition is 17.31 per 1000 words, and the proportion of paraphrases is 96 per cent. The majority of the paraphrases consist of synonymous or near-synonymous wordpairs joined by and or or, as in examples (52) and (53). (52)

Why it is called Piamater, is, for because it is so softe and tender ouer the brayne (Vicary, The Anatomie of the Bodie of Man 29)

(53)

The fyrst is, that they shoulde keepe and defende the Eye from duste (Vicary, The Anatomie of the Bodie of Man 35)

Subperiod EModE2 contains William Clowes's surgical text, Treatise for the Artificiall Cure of Struma from 1602. The incidence of appositions is 11.33 per 1000 words, paraphrases comprising 87 per cent of all instances. The use of and as a paraphrase marker is higher in Clowes's text than in any other scientific treatise in our data: 59 per cent of all the appositional constructions are synonymous or near-synonymous words or phrases joined by and. Although providing a translation for a foreign

Expository apposition

153

word is also an important function of appositional structures in this text, stylistic reasons seem to be important, as in (54). (54)

diuers learned mens writings, which heer to repeate were needles, and to no great vse (Clowes, Treatise for the Artificiall Cure of Struma 22)

Other pairs occurring frequently in apposition are curing and healing, diuers and sundry, and end and purpose. The treatises representing physical sciences are somewhat different from the medical texts. The incidence of expository apposition containing a marker is lower and the proportion of paraphrase is generally smaller, particularly in the sixteenth-century samples. In Robert Record's The Path-way to Knowledg, Containing the First Principles of Geometrie of 1551 (EModEl), claimed to be the first book treating questions on geometry in English (Bennett 1952: 115), incidence of apposition is 6.12 per 1000 words and the proportion of paraphrase is 56 per cent. The remaining appositions belong to the semantic class of identification, reflecting the fact that features such as listing or enumerating items are frequent in this treatise (55). (55)

Nowe adde togyther all the partes of the two lesser squares, that is to say, sixtene and nyne, and you perceyue that they make twenty and fiue (Record, The First Principles of Geometrie E4R)

A similar tendency is seen in Blundevile's Α Briefe Description of the Tables of the Three Speciall Right Lines Belonging to a Circle written on a related topic in 1597 (EModE2). (56)

Of which Zones the auncient men were wont to say that three were unhabitable, that is, the two colde, and the extreame hoat (Blundevile, The Tables of the Three Speciall Right Lines Belonging to a Circle 157R)

The incidence of apposition in Blundevile's text is 6.54 per 1000 words, paraphrases comprising 67 per cent of all instances. Robert Hooke's Micrographia of 1665 and Robert Boyle's Electricity and Magnetism of 1675-1676 (EModE3) differ greatly from each other, which shows that there can be a lot of variation within individual text

154

Päivi Pahta and Saara

Nevanlinna

types. Hooke's text contains 8.25 appositions per 1000 words, whereas in Boyle's treatise the corresponding figure is only 1.72. The proportion of paraphrase in both treatises is higher than in the earlier treatises, c. 80 per cent in each. In the majority of the paraphrases one of the units provides a more familiar variant for an originally French or Latin word (Cells or Boxes, a dore or passage, true and genuine), but there are also instances where the apposition serves merely a stylistic function (shining and bright).

3.1.3.

Religious

instruction19

One of the prototypical text categories rich in apposition is religious instruction. Religious language is not homogeneous, which is also reflected in the varying use of apposition in the texts belonging to this category. The purpose of the writer and the intended audience necessitate constraints. In the HC the following distinct text types have been distinguished: homily, sermon,20 rule, religious treatise. In the English speech-community the language of religion has had a special position from the outset. That religious language depends on tradition is recognized both by believers and non-believers. It is largely based on foreign models. In the older stages much of the vocabulary was Latin or anglicized Latin, and partly French. The language of religion is characterized by linguistic archaisms which, owing to the sacredness of the subject matter, tend to keep it apart from ordinary speech and writing. In the course of time the linguistic effect of religious terminology and phraseology has been evident in all kinds of literature. During the period of our study the various stages in the general development of English are apparent in the language of religion. Theological or learned religious texts typically contain obscure, difficult or metaphorical words, phrases, clauses and sentences that need to be clarified by some means. Terms belonging to the common core terminology such as God, Christ, Virgin Mary, Heaven and the Church may be referred to by lexically common but semantically special terms. Since there are other words or phrases that are both lexically and semantically special, religious texts typically contain expositions of difficult passages such as paraphrases of theological terms or metaphors, or translations of foreign words and phrases. There is usually some shared knowledge between the author of religious texts and his customary audience, and a

Expository apposition

155

writer aims at keeping the balance between the ordinary and the difficultto-grasp. The peak use of expository apposition with an explicit marker falls in the transition periods ME4 and EModEl, the mean incidence being 4.48 and 6.18 per 1000 words respectively. There is, however, a lot of variation both between different text types and individual texts, which will be discussed in more detail in the following two sections.

3.1.3.1.

Late Middle English religious instruction

All the four text types pertaining to the category of religious instruction in the HC are represented in the Late Middle English subsections. The text type labelled homily is only found in ME3. The only sample of homilies in the HC, an extract from the Northern Homily Cycle, contains very few appositions. It is written in verse, which may be one of the main reasons for the scarcity of appositions. Individual sermons included in the Late Middle English subsections differ greatly from each other in their use of apposition. In ME3 this text type is represented by Wycliffite sermons. Wycliff and his followers based their sermons solely on the text of the Bible. They saw English as the language of instruction and paid less attention to style, not to mention any technical theological explanations. Their language was a continuation of Alfredian speech-based prose with hardly any appositions. Early sermons such as those included in John Mirk's Festial (ME3/4) were of the type sermo ad populum, an edifying discourse without any specific spiritual theme. They were addressed to 'good men and women', and contained few expository appositions (only 3.8 per 1000 words). In the fifteenth-century sermons included in the HC the use of apposition is more frequent. Gaytryge's Sermon (ME3/4) introduces the articles of faith to laymen as a response to an episcopal command to educate them in the faith and Christian virtues (57). (57)

The fyfte comandement byddes vs pat we sla na man, fiat es to say, bodyly ne gastely noper (Gaytryge, Sermon 6) 'The fifth commandment bids us not to kill any man, that is to say, either in body or in spirit.'

The incidence of expository apposition with a marker in Gaytryge's Sermon is 12.43 per 1000 words, and and/or comprises 69 per cent of the

156

Päivi Pahta and Saara Nevanlinna

markers. In the Sermons of Fitzjames (ME4), who was a middle-aged professional preacher of high standard, the incidence of markers is 9.84 per 1000 words, and appositions with and/or comprise 82 per cent of the markers. Fitzjames' tone of voice in many of the appositions is similar to that of Gaytryge's (58). (58)

Declyne thou neuer fro them [i.e., the commandments] ne in to Ϋ ryght honde, ne yet in to the lyfte honde, that is to saye, ne in prosperyte, ne yet in aduersyte. (Fitzjames, Sermo Die Lüne A4R) 'Don't ever give up them, neither on the one hand, nor on the other, that is to say, neither in prosperity, nor in adversity.'

The religious treatises included in the Late Middle English subsections of the HC show a great deal of variation in their use of apposition. Most of the religious treatises of the period are translations or adaptations from Latin or French models. The former are generally regarded as the better language. The translations from French were influenced by the new, incoherent trailing-sentence structure. The sample from Rolle's translation of the Latin Psalter comprises 157 verses. Rolle translates 'sense for sense', and his text does not seem to contain any additions in the form of apposition. In his commentary (from ME2/4) attached to those verses there are, however, altogether 190 instances of expository apposition with a marker, the incidence being 18.30 per 1000 words in the text, which is the highest among the Late Middle English religious texts. In the majority of cases (162) the marker that is is used. Only six appositions with that is belong to the class of identification. The rest, as well as all the other appositions, represent paraphrase. The second appositive is usually a lexical synonym or an exposition of a difficult or metaphorical phrase or clause. Rolle's commentary is closely modelled on the Latin commentaries of the Church Fathers, such as Peter Lombard. In The Book of Vices and Virtues (ME3/4), which was translated from French, the incidence of expository apposition is lower, 11.00 per 1000 words. The same incidence is seen in Purvey's Prologue to the Bible (ME3), which is a vernacular text, addressed to the lay reader. Appositions belonging to the semantic class of identification are fairly frequent in it (59).

Expository apposition (59)

157

and eft Dauith seith, "the Lord schal telle in the scripturis of puplis and of these princis that weren in it, that is, in holi chirche". (Purvey, The Prologue to the Bible 56) 'And then David says, "the Lord will tell in the scripture about people and about the rulers that were in it, that is, in holy church".'

In The Cloud of Unknowing (ME3) composed in the vernacular by an unidentified mystic writer for an apparently select readership with a great deal of shared knowledge, the incidence of markers is only 2.74 per 1000 words. The native coordinative conjunctions and/or are much less frequent as apposition markers in translations of religious treatises than in original English texts. In Rolle's Psalter commentary, and/or occurs only in 7 per cent of the appositions. The corresponding percentage in The Book of Vices and Virtues is 37, whereas it is 70 in The Cloud, and as high as 74 in Purvey's Prologue, both texts being directly addressed to their audience. The two independent translations of Aelred of Rievaulx 's De Institutione Inclusarum classified as religious rules are also different from each other in their use of apposition. In ME3 samples, which derive from Latin, the incidence of apposition is very low (1.30 per 1000 words), while in ME4 samples, which contain a translation of the whole work from French, it is considerably higher (6.63 per 1000 words).

3.1.3.2. Early Modern English sermons By the end of the fifteenth century the influence of the Renaissance had reached England, and Latin received a fresh boost. The vocabulary of English was greatly enriched from classical Latin sources. The newly introduced terms necessitated a great deal of explanation and interpretation. Throughout the Middle Ages rhetoric had remained central to the culture. Since rhetoric was specially designed to give emotional or imaginative weight to the authoritative truth advocated by Church and State, rhetorical conventions could be found in all kinds of texts. When classical texts were rediscovered during the Renaissance, rhetoric was revised as a complete system. Aristotle's main means of persuasion (ethos — persuasion through personality and stance, pathos — persuasion through

158

Päivi Pahta and Saara Nevanlinna

emotion, and logos — persuasion through logic and reasoning) were important principles followed by public preachers and orators (Cockcroft— Cockcroft 1994 passim). The only text type of religious instruction in the Early Modern English subsections in the HC is sermon. Arousing emotion was especially important for a sermon to be effective, but it was also important that the audience should comprehend the message. Educated, largely bilingual humanist preachers addressed their lay audiences in a new type of language combining native simplicity and humanist latinity. The Early Modern English sermon-writers were well-versed in rhetoric and highly experienced in professional preaching. In the compilation of the HC, the sermons in each Early Modern English subsection were selected to represent two different stylistic strata, popular and more formal (see Nevalainen— Raumolin-Brunberg 1993: 63). The first Early Modern English subperiod is the peak period in the use of expository apposition. Although the preachers whose sermons are included in EModEl, Latimer and Fisher, are usually presented as being very different from each other, the incidence of expository apposition in the samples of their sermons is almost the same (nearly 8 per 1000 words). Among seven variant apposition markers that were found in Bishop Latimer's sermons, and/or comprised 78 per cent of the total. Bishop Fisher's sermons contained four variant markers, out of which and/or comprised only 40 per cent. Catholic lay audiences accustomed to Latin readings from the altar probably expected less explanation by way of paraphrase. Examples (60) and (61) show Latimer's and Fisher's typical ways of using appositions. (60)

The one is an inclosinge to let or hinder ye bodily ploughynge, and the other to lette or hynder the holiday ploughyng, the church ploughinge ... no man wyll herken it but to hinder and let it (Latimer, Sermons 25)

(61)

These fyue great woundes were ingraued wyth sharpe & vyolent pennes, that is to say, the sharpe nayles, and the speare. (Fisher, The English Works of John Fisher 1 396)

In EModE2, the incidence of markers was very low in the sermons of Henry Smith, only 1.73 per 1000 words and there were no and/or appositions. Richard Hooker had very strong feelings for the Church of England and he seems to have been a true rhetorician. Paraphrase with and/or

Expository apposition

159

comprised 86 per cent of the markers, the incidence of markers being 5.48 per 1000 words. In EModE3, the incidence of markers in the sermons of the Catholic preacher Jeremy Taylor was only 3.24 per 1000 words, the markers and/or comprising 68 per cent of the total. In Archbishop Tillotson's sermons the incidence of expositor apposition was 5.00 per 1000 words. His sermons had their emotional appeal enhanced by appositional paraphrases with and/or, which constituted as much as 94 per cent of the total. The above shows that the great protestant sermon-writers were styleconscious. It has been argued that their sermons resemble the persuasive homiletic style of ^ l f r i c and Wulfstan, who often coupled synonyms and near-synonyms with the native markers and and or.

3.1.4.

Philosophical

treatises

There is some interesting variation between the philosophical treatises included in the HC. In Chaucer's translations of Boethius' De consolatione philosophiae (ME3) after Latin and French models, and the Tale of Melibee from French, the incidence of markers is 6.78 and 7.90 per 1000 words respectively (cf. Blake 1992a: 504). Being a highly professional writer and very versatile, Chaucer makes ample use of coordinative apposition with and/or as a stylistic device, as in (62). (62)

"A," quod dame Prudence, "ye seyn youre wyl and as yow lyketh, but in no caas of the world a man sholde nat doon outrage ne excesse for to vengen hym. (Chaucer, The Tale of Melibee 232.C1) '"Ah," said Dame Prudence, "you speak your will and as you like it, but in no case should a man take excessive revenge.'

And/or occurs in 81 per cent of the instances in Boethius, and in 78 per cent in Melibee. Chaucer's main reference points in his Boethius were the Vulgate, Jean de Meun's French translation and Nicholas Trevet's Commentary. Chaucer followed in the footsteps of Trevet by providing glosses to his vernacular text (Minnis 1993: viii, 173). He wanted to be faithful to the sense (sententia), but also followed Jerome's other principle in his translation of the Bible: word for word. Expository comments and glosses, like in example (63), were considered to be part of the whole.

160 (63)

Päivi Pahta and Saara Nevanlinna telle me, yif thow accordest to leten no torment to the soules aftir that the body is ended by the deethl (This to seyn, "Undirstondestow aught that soules han any torment aftir the deeth of the bodyV (Chaucer, Boethius 448.C1) 'tell me whether you consent to assigning no torture to the souls after the body ends in death? (This is to say, "Do you believe at all that souls are tormented after the death of the body?")'

The HC contains several vernacular versions of Boethius; besides Chaucer's Boethius the samples include three Early Modern English translations. In George Colville's translation of Boethius (EModEl) the influence of the Renaissance is clearly to be seen, as in example (64). The incidence of expository apposition is as high as 26.49 per 1000 words. The markers and/or are contained in 88 per cent of all the appositions. (64)

Therfore these worldlye thynges, that is to saye, worldelye suffisaunce, power reuere~ce nobles and pleasures, semethe to geue vnto men the symylitudes or lykenes, of true good, or ells, to geue certaine vnperfit and fained goodes (Colville, Boethius 71)

Queen Elizabeth's translation from the original Latin of parts of Boethius in EModE2 provides the direct opposite to the earlier prose translations as far as apposition is concerned. Though her translation was done only some thirty years after Colville's, the incidence of expository apposition is very low, only 0.87 per 1000 words. She no longer had the medieval attitude towards comments and glosses; her text is a slavish translation from Latin. The only apposition markers occurring in the Queen's translation are and and or within the class of paraphrase. The Right Honourable Richard Preston translated Boethius from Latin towards the end of subperiod EModE3 (1695). The incidence of appositions, 5.33 per 1000 words, is lower than in Colville, but markedly higher than in Queen Elizabeth's translation. The markers and and or, constituting 96 per cent of all the markers, occur almost exclusively in lexical paraphrase. These results show how idiosyncratic preferences can be reflected in the translation of one and the same work.

Expository apposition 3.1.5. Educational

161

treatises

Six educational treatises are included in the Early Modern English section of the HC, two in each subperiod. Their external aims divide them into two groups. Each subperiod is represented by one educational treatise with more theoretical aims and one with purely practical goals (Nevalainen—Raumolin-Brunberg 1993: 64). The first group consists of Thomas Elyot's The Boke Named the Gouernour (EModEl), Francis Bacon's Advancement of Learning (EModE2) and John Locke's Directions Concerning Education (EModE3), and the second group contains Roger Ascham's The Scholemaster (EModEl), John Brinsley's Ludus Literarius, or the Grammar Schoole (EModE2) and Charles Hoole's A New Discovery of the Old Art of Teaching Schoole (EModE3). The mean incidence of appositions is highest in the earliest texts and declines steadily towards the end of the period (7.66 appositions per 1000 words in EModEl, 5.79 in EModE2, and only 2.30 in EModE3). There is, however, also synchronic variation within the text type. In the first two subperiods the use of expository apposition with a marker is clearly different between the more theoretical discussions and the more practically oriented treatises. In EModEl, Elyot's text contains 12.28 appositions per 1000 words, whereas in the same amount of text Ascham only uses 2.49 appositional constructions. Appositional constructions seem mainly to be used for stylistic purposes by both writers, as in example (65) from Elyot. (65)

Ο what misery was it to be subiecte to suche a minstrell, in whose musike was no melodye, but anguisshe and dolour? (Elyot, The Boke Named the Gouernour 27)

Similar trends continue in EModE2: there are 8.27 appositions per 1000 words in Bacon's more theoretical text (66) and the corresponding figure in Brinsley's dialogue-form treatise is only 3.25 (67). (66)

So Readers in Scyences are indeede the Gardyans of the stores and prouisions of Scyences, whence men in actiue courses are furnished (Bacon, Advancement of Learning 4R)

162

(67)

Päivi Pahta and Saara

Nevanlinna

Prove and confirme what tongue soever your Scholler learnes, even from the first reading of English (Brinsley, Ludus Literarius, or the Grammar Schoole 44)

The situation is different in EModE3, where both educational treatises contain only few appositions: 1.92 per 1000 words in Locke's more theoretical treatise and 2.61 in Hoole's text. But like in the earlier texts, the majority of appositions in Locke's and Hoole's writing are paraphrases consisting of repetitive word-pairs joined by and or or used either for rhetorical or instructional purposes, as in example (68). (68)

Now there are five organs or instruments of speech, in the right hitting of which, as the breath moveth from within, through the mouth, a true pronunciation of every letter is made (Hoole, A New Discovery of the Old Art of Teaching Schoole 2-3)

Further examples of appositions containing and/or in the educational treatises are provided by pairs like accepting or admitting, delites and pleasures, deceit or vntruth, properlie and fitlie and trayne and exercise. Other markers are rare and they are mainly used in the theoretical treatises, as in example (69). (69)

If what I have said in the beginning of this discourse be true as I doe not doubt but it is, viz. That the differences to be found in the manners and abilitys of men is oweing more to their Education then anything else, we have no reason to conclude that great care is to be had of formeing childrens mindes (Locke, Directions Concerning Education 49)

3.2. The frequency of semantic classes across text types The different semantic classes of expository apposition were unevenly distributed in the data. The majority of the appositional constructions with an explicit marker belonged to the semantic type in which the two units of apposition are equally specific, i.e., the classes of paraphrase and revision; c. 93 per cent of the appositions in our study belonged to this type. The overall distribution of the five semantic classes in our data is shown in Table 4 on p. 134 above, and Tables 10-14 (see pp. 163-166) show the distribution of the different classes in each subperiod.

Expository apposition

163

In the following sections we will briefly comment on the occurrence of the five different semantic classes in our data. Table 10. Semantic classes in ME3 (1350-1420; N/1000 words). IDE = identification, APP = appellation, CHAR = characterization, PAR = paraphrase, REV = revision. Text Text categ. type IS STA XX IR XX XX IS IR NI NN IR IS NN XX IR

Science, medicine Document Philosophy Religious treatise Official letter Document Handbook, astronomy Homily Fiction History Rule Handbook, medicine Travelogue Bible Sermon

IDE /1000

APP /1000

CHAR /1000

0.04 0.40 0.16 0.58 0.70

0.08 0.03

0.08

0.37 0.67 0.19

0.05 0.07

PAR /1000 16.20 8.55 6.97 2.91 2.99 2.38 1.84 1.24 1.47 1.27 0.74 0.49 1.09 0.48 0.37

REV /1000

0.87 0.49

0.31 0.41 0.04 0.19

0.10 0.07

Table 11. Semantic classes in ME4 (1420-1500; N/1000 words). Text Text categ. type IS STA IR IR IR XX XX IS XX

Science, medicine Law Religious treatise Sermon Rule Preface Official letter Handbook, other Document

IDE /1000 1.24 0.09 0.44 0.40 1.66 0.16 0.32 1.56 0.19

APP /1000

0.08

CHAR /1000

0.02 0.04

PAR /1000 8.71 9.70 8.01 6.04 4.97 5.00 5.10 2.94 2.96

REV /1000

0.24 0.48 0.09

164

Päivi Pahta and Saara Nevanlinna

Table 11. Continued. Text Text categ. type IS NI XX XX NN NN XX XX NI IS

Handbook, medicine Fiction Deposition Bible History Saint's life Mystery play Private letter Romance Handbook, astronomy

IDE /1000

APP /1000

CHAR /1000

0.35 0.11

PAR /1000 2.47 1.94 2.03 1.58 1.35 1.05 0.95 0.87 0.52

0.16

0.11

REV /1000

0.40 0.26

Table 12. Semantic classes in EModEl (1500-1570; N/1000). Text Text categ. type XX STA EX IS IR IS/EX EX XX NN NN NI XX NN XX NN XX NN XX

Philosophy Law Science, medicine Handbook, other Sermon Education Science, other Trial History Biography Fiction Official letter Travelogue Comedy Autobiography Private letter Diary Bible

IDE /1000

APP /1000

CHAR /1000

1.31 0.81 0.90 0.21 0.19 2.69 0.44 0.09

0.10

0.13 0.09

0.17

0.09

PAR /1000 25.18 25.53 16.51 10.60 7.18 7.38 3.43 2.94 1.80 1.65 1.56 1.59 1.49 0.85 0.70 0.75 0.61 0.52

REV /1000

0.17

0.53 0.10 0.19

0.09 0.07 0.19 0.09

Expository apposition

165

Table 13. Semantic classes in EModE2 (1570-1640; N/1000). Text Text categ. type STA EX EX EX NN IR IS NN NN NN XX XX XX XX XX NI XX NN

Law Science, medicine Science, other Education Travelogue Sermon Handbook, other Autobiography History Biography Official letter Trial Comedy Philosophy Bible Fiction Private letter Diary

IDE /1000 0.34 1.21 1.56 0.18 0.07 0.29 0.24 0.24 0.10

APP /1000

CHAR /1000

0.15 0.31 0.09 0.34

0.24 0.21

0.07

0.05 0.09

PAR /1000

REV /1000

17.83 9.82 4.36 5.17 3.45 3.20 3.01 2.43 2.48 0.83 1.06 1.05 0.76 0.87 0.63 0.56 0.35 0.16

0.59 0.15 0.31 0.36 0.07 0.19 0.33 0.24

PAR /1000

REV /1000

16.09 5.10 4.34 3.93 2.55 2.77 1.86 1.50 1.70 1.20

0.44 0.16 0.35 0.19 0.35 0.58

0.42 0.18 0.07 0.34

Table 14. Semantic classes in EModE3 (1640-1710; N/1000). Text Text categ. type STA XX EX IR IS NN IS NI XX NN

Law Philosophy Science, other Sermon Handbook, other Travelogue Education Fiction Official letter History

IDE /1000 0.46 0.28 0.44 0.08 0.26 0.09 0.17

APP /1000

CHAR /1000

0.09 0.09

0.43

166

Päivi Pahta and Saara Nevanlinna

Table 14. Continued. Text Text categ. type NN XX NN XX XX NN

3.2.1.

Biography Trial Diary Private letter Comedy Autobiography

IDE /1000

0.09 0.15 0.18

APP /1000

0.08

CHAR /1000

PAR /1000

REV /1000

1.46 0.58 0.45 0.23 0.24 0.18

0.07 0.09 0.08 0.16

Paraphrase

The most frequently occurring semantic class in our data is paraphrase (see p. 137 above), instances of which are found in all text types and practically all texts. It is particularly frequent in texts belonging to the fields of legal, medical or philosophical writing. The linguistic situation in medieval and early modern England created a need for this type of usage. It was often necessary to provide an explanation for words of foreign origin or dialectal variants (see examples (25) and (28) above). Coordinative apposition consisting of synonymous or near-synonymous word-pairs linked by and or or was also a common stylistic or rhetorical device, and is accordingly found in texts which are not otherwise rich in appositions (example (24) above). It is precisely this construction which makes paraphrase so much more frequent than the other semantic classes of apposition. Although the majority of paraphrases consist of lexical apposition, there are also a considerable number of longer paraphrases used for interpreting complex propositions to make the text more accessible to the readers or listeners (see examples (34) and (35) above). Both forms of paraphrase contribute to success in communication (cf. Blakemore 1993: 109).

3.2.2.

Revision

In revision (see p. 140), two subtypes serving different communicative functions can be distinguished. The first subtype, which can be labelled reorientation, is a fairly rare type of apposition in our data. It is mainly found in religious or philosophical texts in a context where the speaker or

Expository apposition

167

writer reviews a phenomenon from a different standpoint (examples (38) and (39) above). The second subtype can be labelled self-correction. It is more typical of spoken language and therefore mainly found in texts which are speech-based, such as homilies or sermons and drama. It also occasionally occurs in legal or scientific texts in order to provide a more appropriate expression for a lexical item or proposition. In educational treatises and narrative texts it is used style-consciously either to provide emphasis or to communicate the author's point of view by way of a metalinguistic comment, or to supply the reader with a range of implications (examples (42) and (43) above; cf. also Blakemore 1993: 111).

3.2.3.

Identification

The semantic class of identification (see p. 134) is found in several types of texts, but there is a tendency for appositions of this kind to occur more frequently in texts that can be labelled instructional or expository, such as handbooks, scientific treatises, or religious rules. As in Present-day English, appositions in the class of identification are mostly connected with enumerating or listing items or procedures (example (17) above; cf. Meyer 1992: 121).

3.2.4.

Appellation

The appositions with an explicit marker belonging to appellation (see p. 135) were rare in our data. In Present-day English this type of apposition is very common in journalistic writing (Meyer 1992: 115). In our data the scattered examples are found in a variety of text types including narrative texts, such as biography, travelogue and history, i.e., in texts in which naming individual people or places is an essential part of communication (example (21) above), but also in scientific treatises and handbooks.

3.2.5.

Characterization

Providing characteristics of individuals or incidents by way of apposition with a marker is rare in our data. Characterizing apposition, with the second unit providing less specific information, is perhaps a type of construction where the use of a specific marker is more awkward than in the other semantic classes (cf. p. 136 above). The rare instances of charac-

168

Päivi Pahta and Saara Nevanlinna

terization are found in instructional or expository texts (example (22) above). In Present-day English characterization is, like appellation, common in the press genre, where it is important to both name and attribute characteristics to individuals about whom the audience has little knowledge (Meyer 1992: 116).

4.

On the use of the markers of expository apposition

This section summarizes briefly some of the general tendencies seen in the use of explicit markers of expository apposition in our data. A more detailed discussion of the topic is contained in Pahta—Nevanlinna forthcoming. The dominance of expository appositional constructions with the markers and and or has already been pointed out and discussed in connection with some text types (pp. 147-162 above). These two markers occurred in practically all text types and individual texts included in our study, although there is a lot of variation in their frequency across different text types and individual texts as well as across time. They occur in well over half of all appositional constructions in the Late Middle English subperiods (75 per cent in ME3 and 57 per cent in ME4), making up almost 90 per cent of the total number of the markers in each of the Early Modern English subsections of the corpus (EModEl 87; EModE2 89; EModE3 89). The texts in which they are most frequently used in our data belong to the fields of law, science, education, religion and philosophy (for examples, see section 3.1). The other markers of expository apposition were found to be far more restricted in their scope, and they can be divided roughly into two types, although there is considerable overlap. Markers like that is (to say), (that is) to wit, namely, and particularly the Latin markers seem to be most common in fairly formal or technical instructive or expository texts (examples (70) and (71)), whereas markers like nay and yea, which border on pragmatic particles, are mainly found in narrative or speech-based texts (examples (72) and (73)). (70)

they tende to one conclusion or purpose, that is to saye, with a free and glad wyll to gyue to a nother that thinge which he before lacked. (Elyot, The Boke Named the Gouernour 148-149)

Expository apposition

169

(71)

if thou list to dissolue & cast out Phlegme, these Pils following must be taken, viz. Pillulae de Sagap: de Opopan, de Elleboro, de Euphorbio. (Clowes, Treatise for the Artificiall Cure of Struma 10)

(72)

I say unto thee again, make haste, nay, run, run unto the Brethren. (Penny Merriments 150)

(73)

that he sholde after his labour haue full reste. ye & for shorte labour haue eternal reste (Fitzjames, Sermo Die Lüne A2R)

Our study also showed that most of the markers which are available in contemporary English for explicitly indicating expository apposition existed in Late Middle English. The only marker current today which was not found in our data by 1710 is in other words (see p. 128 and note 9). Another marker, namely, which is also common in Present-day English as an expository marker, was undergoing a semantic change during the period studied. In the earliest samples of our data it is used as a marker of exemplifying apposition (74) and the first instances of namely in expository apposition date from the fifteenth century (75). (74)

but loke that he examyne truli his Latyn bible, for no doute he shal fynde ful manye biblis in Latyn ful false, if he loke manie, namely newe. (Purvey, The Prologue to the Bible 57-58) 'But he should take care to examine closely his Latin Bible, for undoubtedly he will find many Latin Bibles quite in error, if he examines several, especially new ones.'

(75)

Pis word 'oure' teche}) vs to hate pre pinges, namely: pride, hate, couetise. (The Book of Vices and Virtues 101) 'This word "our" teaches us to hate three things, namely: pride, hate, avarice.'

In addition to the markers used in contemporary English, our study also revealed a number of phrases which have disappeared or become rare, such as nay, yea or as who say, as in example (76) (for others see Table 3 on p. 129).

170 (76)

Päivi Pahta and Saara Nevanlinna no thing nis leveful to folye in the reaume of the devyne purveaunce (as who seith, no thing nis withouten ordenaunce in the reame of the devyne purveaunce) (Chaucer, Boethius 454.C1) 'Nothing is allowable to folly in the domain of divine providence, in other words, nothing is without divine arrangement in the domain of divine providence.'

The Latin markers scilicet, id est, and viz. and their variants occurred in both Late Middle English and Early Modern English texts. They were most frequent in appositions where one of the appositional units was Latin and the other, vernacular (77). Some of them occurred in fully Latin passages, but also in fully English passages, as in (78).

(77)

in Jje holownesse of J>e lyuere goo{> oute a veyne icleped porta (i. pe gate) (The Cyrurgie of Guy de Chauliac 62) 'in the hollow cave of the liver a vein goes out which is called porta, i.e., gate'

(78)

His daughter, viz his 2J daughter is upon recovery so as to come (Henry Oxinden, The Oxinden and Peyton Letters 277)

5.

Conclusion

In this study our aim has been to shed some light on the structure and use of nonrestrictive expository apposition containing an explicit marker during the period 1350-1710 and to relate our findings to the situation in Present-day English. As far as the structure of expository apposition is concerned, the results of our study show that the basic syntactic types current in Present-day English, i.e., lexical, phrasal, clausal and sentential apposition, can also be found in the period of our study. As today, apposition is typically a relation between two noun phrases. Similarly, the various semantic classes distinguishable in contemporary English (identification, appellation, characterization, paraphrase and revision) can also be detected in Late Middle and Early Modern English texts. The dominant semantic class in our data was paraphrase. Several factors seem to motivate re-phrasing in the form of apposition and particularly the use of paraphrase in our material. During the period of our

Expository apposition

171

concern, a large amount of new lexical items were introduced to the language from Latin and French, and there was a need for writers to provide their audience with a native synonym or near-synonym of loanwords in order to ensure successful communication. The borrowed lexicon which needed to be explained included words from all spheres of life, but in the data of this study, providing translations for technical terminology is one of the functions in which apposition containing an explicit marker is particularly frequent. In Middle English texts dialectal variation also provided a source of frequent use of apposition. Although the rise of standardisation began in late fourteenth-century London, in Late Middle English there was still a lot of regional and orthographic variation, especially by scribes who worked far away from the London region and among writers whose families had migrated to London from other parts of the country. Like foreign words, dialectal forms could be glossed by adding an explanation in the form of apposition. The need for an explanation is usually the source of apposition in clausal and sentential paraphrase, too. Religious and philosophical texts particularly often contain passages where complicated propositions are made more accessible to the audience by adding an appositional explanation. A different source of apposition is provided by stylistic considerations. The use of coordinative apposition consisting of synonymous or nearsynonymous word-pairs joined by and or or is a well-established stylistic device going back to the persuasive rhetorical style of Anglo-Saxon writers like yElfric and Wulfstan. Apposition with and/or dominates throughout the period of our study. It becomes prominent in writings of the Renaissance and remains frequent during the whole Early Modern English period. In some types of writing coordinative apposition has become formulaic, e.g., contemporary legislative writing is characterized by formulaic use of archaic appositional constructions, some of which can be dated back to Old English. The study also shows that certain types of writing clearly favour the use of appositional constructions or the use of particular semantic or syntactic categories of apposition. Expository apposition with an explicit marker was found to be most frequent in texts belonging to learned or formal genres, such as scientific or statutory writings. The dominant category in both types of writing is lexical paraphrase, and the use of apposition is primarily motivated either by the need to explain technical terminology of foreign origin or the need to achieve explicitness and precision in communication. Expository apposition is also frequent in texts belonging to the spheres of religious and secular instruction. Lexical paraphrase is also

172

Päivi Pahta and Saara Nevanlinna

common in these texts. Instructive texts, however, show a greater tendency to contain longer appositional paraphrases, which help the reader to process information and profit from instruction. In a diachronic perspective one the main lines of variation is connected with the selection of different markers used for explicitly indicating expository apposition. Various types of unique phrases (see p. 129 above) are found in this function in the first four subsections of the Helsinki Corpus included in this survey (ME3-EModE2), being most common in subperiod EModEl. By mid-seventeenth century (EModE3), the selection seems to have become stable, and with the exception of in other words (see p. 128 above), all the markers used for indicating expository apposition in Present-day English, are found in our data.

Notes 1. 2.

3.

4.

5.

6. 7.

Hudson (1990: 314) notes that apposition tends to receive less attention from theoretical linguists than it deserves. We found the model presented by Charles F. Meyer (1992) very useful for our study, and we would like to thank him for his valuable comments on this article. For reviews of his approach see Oostdijk (1994); Pahta (1994); Acuna (1996: 119-127). In Investigating English style, Crystal—Davy (1974: 59, n. 13) point out explicitly that they apply the term "apposition" only to identically functioning nominal groups and exclude all other structures which may be appositional in "some notional sense". The London-Lund Corpus of Spoken British English (see Svartvik—Quirk 1980), the Survey of English Usage Corpus of Written British English (see Greenbaum 1985), and the Brown University Standard Corpus of Presentday American English (Kucera—Francis 1967). Approximately 120,000 words from each corpus were included in the study (for details see Meyer 1992: 7-8). In Meyer's study of apposition in Present-day English, the majority of appositions (71 per cent) were either nonrestrictive nominal appositions or non-nominal appositions for which the notions restrictive and nonrestrictive were irrelevant (Meyer 1992: 82). These are the two main additive functions of appositional constructions according to Halliday—Hasan (1976: 248-250). Mitchell (1985, 1: 596-612) discusses only paratactically joined appositions. See also Nevanlinna—Pahta (forthcoming).

Expository apposition 8.

9.

173

The Old English interlinear glosses in several Latin Psalter manuscripts (e.g. the Vitellius and Canterbury Psalters) contain numerous instances of phrasal appositive structures in the form of double, triple, or, rarely, quadruple glosses expressing synonymy. The marker of apposition is an I with a slanting cross-bar short for Latin vel (see Nevanlinna 1984). The earliest example is for to say, that having in any quantity measured so much, or gone so far, you are not yet at the end, is only to say, that that quantity is greter. So that the negation of an end in any quantity is, in other words, only to say, that it is bigger. (Locke 068/067-P16)

10.

11. 12.

13. 14. 15.

16.

17.

For information on the Century of Prose Corpus see Milic (1990). Of the markers of apposition included in this study, that is and that is to say can either precede or follow the second appositive in Present-day English, whereas namely, or (rather/better) and the abbreviated forms i.e. and viz. can only precede the second appositive (Quirk et al. 1985: §17.73). Cf. Huddleston's (1971) view that in Present-day English no apposition marker is used in characterizing apposition, pp. 123-124 above. In this category our classification differs both from Quirk et al. (1985) and Meyer (1992). We divide appositions of this type into two semantic classes, Meyer (1992) distinguishes three classes, and Quirk et al. (1985) have one class with four subtypes. As regards abstract nouns in apposition it is not always possible to decide whether they represent one quality or two. In this to sayn as who saith, is is deleted through haplology (see Pahta —Nevanlinna forthcoming). Quirk et al. (1985) classify revision as a subcategory of reformulation, whereas we decided to discuss it as a separate class with two subclasses of its own. In Meyer's study, spoken genres tended to contain fewer appositions than written genres (Meyer 1992: 126). This reflects Biber's (1988: 46) characterization of spontaneous speech, together with personal letters, as communication between speakers with a high degree of shared knowledge, whereas most written texts are produced by discourse participants with little shared knowledge. Different aspects of legal language are discussed e.g. in Charrow (1982), Crystal—Davy (1974), Danet (1980), Gustafsson (1975), Hiltunen (1990), Maley (1987), and Mellinkoff (1963) and (1982).

Päivi Pahta and Saara Nevanlinna

174 18.

19.

20.

This role of legislation as canonical text has been suggested as originating in fourteenth-century England, when the function of the judiciary became separate from that of the King's Council. The judges began to interpret statutes strictly — not only as "suggestions of policy within whose broad limits the court can exercise a broad discretion" (Plucknett 1940; cited after Maley 1987: 28). For background information in this section we are indebted to Blake (1992a), Crystal—Davy (1974), Drabble (1986), Gordon (1972), Edwards (1984), Jacob (1993), Leith (1987), Swales (1990), and Webster (1988). See also Nevanlinna et al. (1993) and Nevalainen—Raumolin-Brunberg (1993). In Middle English texts, the distinction between the text types 'sermon' and 'homily' was observed until the last third of the fourteenth century (Heffernan 1984: 179). Later, the two terms were at times indiscriminately applied to either of the text types. To the medieval mind, homilia was "a name applied to an edifying discourse based on a text of sacred scripture that is itself traditionally tied to a particular liturgical day or event" (Heffernan 1984: 179). In the Middle English period, homilies were assigned to the Sundays or other festival days, even to some weekdays, mostly according to the use of York or Salisbury. In the Early Modern English period, special Books of Homilies were published by the Church of England in 1547 and 1563 for the use of parish churches (OED). The purpose of the medieval sermon (sermo ad populum) was to teach the believers the truths of scripture and the faith. There was a considerable variety in form (prose - verse) and in theme (religious - popular). Connection to a liturgical event was not obligatory. Rhetorical techniques were often borrowed from the artes predicandi. References to topical events were rare. Exempla derived from various sources were attached to sermons and homilies, which was criticized severely by the Wycliffites (Heffernan 1984).

References Acufia, Juan Carlos

1996

The puzzle of apposition. On so-called appositive structures in Eng-

lish. (Monografias de Universidade de Santiago de Compostela 190.) Santiago de Compostela: Universidade de Santiago de Compostela, Servicio de Publicacions e Intercambio Cientifico. Aijmer, Karin—Bengt Altenberg (eds.)

1991

English corpus linguistics. Studies in hounour of Jan Svartvik. London—New York: Longman.

Expository apposition

175

Allwood, Jens—Lars-Gunnar Andersson—Osten Dahl 1977 Logic in linguistics. London—New York—Melbourne: Cambridge University Press. An Anglo-Saxon dictionary 1955 (Based on the manuscript collection of Joseph Bosworth. Edited and enlarged by T. Northcote Toller.) London. AV = The Authorized Version of the English Bible 1953 London. Bennett, H. S. 1952 English books & readers 1475 to 1557. Being a study in the history of the book trade from Caxton to the incorporation of the Stationers' Company. Cambridge: Cambridge University Press. Biber, Douglas 1988 Variation across speech and writing. Cambridge: Cambridge University Press. Blake, Norman F. 1992a "The literary language", in: Norman Blake (ed.), 500-541. 1992b "Translation and the history of English", in: Matti Rissanen et al. (eds.) 1992: 3-24. Blake, Norman F. (ed.) 1992 The Cambridge history of the English language. Vol. 2. Cambridge: Cambridge University Press. Blakemore, Diane 1993 "The relevance of reformulations", Language and Literature 2: 101— 103. Blockley, Mary 1989 "Old English coordination, apposition, and the syntax of English poetry", Modern Philology 87/2: 115-131. Bloomfield, Leonard 1965 Language. (Reprinted.) London: Allen and Unwin. Bolinger, Dwight 1979 Meaning and form. (English Language Series 11.) London—New York: Longman. The Book of Common Prayer. 1945 (Reprinted.) New York: The Church Pension Fund. Burnley, David 1992 "Lexis and semantics", in: Norman Blake (ed.), 409^493. Burton-Roberts, N. 1975 "Nominal apposition", Foundations of Language 13: 391-419. Chambers, R. W.—Marjorie Daunt (eds.) 1931 A book of London English. Oxford: Clarendon Press.

176

Päivi Pahta and Saara Nevanlinna

Charrow, Veda R. 1982 "Linguistic theory and the study of legal and bureaucratic language", in: Loraine Obler—Lise Menn (eds.), 81-101. Cockcroft, Robert—Susan M. Cockcroft 1994 Persuading people: An introduction to rhetoric. (Reprinted.) London: Macmillan. Collins Cobuild English language dictionary. 1987 Ed. in chief John Sinclair. London: Collins. Cruse, D.A. 1986 Lexical semantics. Cambridge: Cambridge University Press. Crystal, David 1985 A dictionary of linguistics and phonetics. (Second edition updated and enlarged.) Oxford: Blackwell. Crystal, David—Derek Davy 1974 Investigating English style. (5th edition.) London: Longman. Curme, George O. 1931 Syntax. (A grammar of the English language, vol. 3.) Boston: D. C. Heath and Co. 1970 English grammar. (Reprinted.) (College Outline Series 61.) New York: Barnes and Noble. Danet, B. 1980 "Language in the legal process", Law and Society Review 14/3: 445563. Denison, David 1993 English historical syntax: Verbal constructions. London: Longman. Drabble, Margaret (ed.) 1986 The Oxford companion to English literature. (5th edition reprinted with corrections.) Oxford—New York—Tokyo—Melbourne: Oxford University Press. Edwards, Anthony S. G. (ed.) 1984 Middle English prose. A critical guide to major authors and genres. New Brunswick, New Jersey: Rutgers University Press. Forshall, John—Sir Frederik Madden (eds.) 1850 The Holy Bible containing the Old and New Testament with the Apocryphal Books in the earlier English versions made from the Latin Vulgate by John Wycliffe and his followers. Oxford: Oxford University Press. Fries, Charles C. 1952 The structure of English. New York: Harcourt, Brace and Co.

Expository apposition

177

Ghadessy, Mohsen (ed.) 1988 Registers of written English: Situational factors and linguistic features. (Open Linguistics Series.) London—New York: Pinter Publishers. Gordon, Ian A. 1972 The movement of English prose. (English Language Series.) (Third impression.) London-—Beccles—Colchester: Longman. Greenbaum, Sidney 1985 "25th Anniversary of the Survey of English Usage", World Englishes 4: 261-265. Gustafsson, Marita 1975 Some syntactic properties of English law language. Turku: University of Turku. Halliday, Michael A. K.—Ruqaiya Hasan 1976 Cohesion in English. London: Longman. HC = Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Heffernan, Thomas J. 1984 "Sermon literature", in: Anthony S. G. Edwards (ed.), 177-207. Hiltunen, Risto 1990 Chapters on legal English: Aspects past and present of the language of the law. (Annales Academiae Scientiarum Fennicae, Ser. B, Tom. 251.) Helsinki: Academia Scientiarum Fennicae. Hiltunen, Risto—Marita Gustafsson—Keith Battarbee—Liisa Dahl (eds.) 1993 English far and wide: A festschrift for Inna Koskenniemi. (Annales Universitatis Turkuensis, Ser. B, Tom. 197.) Turku: University of Turku. Hockett, C. 1955 "Attribution and apposition", American Speech 30: 99-102. Hopper, Paul J.—Elizabeth Closs Traugott 1993 Grammaticalization. (Cambridge textbooks in linguistics.) Cambridge: Cambridge University Press. Huddleston, Rodney D. 1971 The sentence in written English. Cambridge: Cambridge University Press. Hudson, R. 1990 English word grammar. Oxford: Blackwell. Jackendoff, Ray 1985 Semantics and cognition. (Current Studies in Linguistics Series.) (Second printing.) Cambridge, Massachusetts—London: MIT Press.

178

Päivi Pahta and Saara Nevanlinna

Jacob, E. F. 1993 The fifteenth century 1399-1485. Oxford—New York: Oxford University Press. Jespersen, Otto 1961 Α Modern English grammar on historical principles. Vol. 4. London: Allen and Unwin. Jones, Peter Murray 1989 "Four Middle English translations of John Ardeme", in: A. J. Minnis (ed.), 61-89. Kajanto, Iiro—Inna Koskenniemi—Esko Pennanen (eds.) 1983 Studies in classical and modern philology presented to Y. M. Biese on the occasion of his eightieth birthday 4.1.1983. (Annales Academiae Scientiarum Fennicae, Ser. B, Tom. 223.) Helsinki: Academia Scientiarum Fennicae. Kilpiö, Matti 1992 "Beon", in: Dictionary of Old English: Fascicle B. Published for the Dictionary of Old English Project, Centre of Medieval Studies, University of Toronto. Toronto: Pontifical Institute of Mediaeval Studies. Koskenniemi, Inna 1968 Repetitive word pairs in Old and Early Middle English prose. Expressions of the type whole and sound and answered and said, and other parallel constructions. (Annales Universitatis Turkuensis, Ser. B, Tom. 107.) Turku: University of Turku. 1983 "Semantic assimilation in Middle English binomials", in: Iiro Kajanto—Inna Koskenniemi—Esko Pennanen (eds.), 77-84. Kristensson, Gillis 1974 John Mirk's Instructions for parish priests. (Lund Studies in English 49.) Lund: CWK Gleerup. Kuöera, Η.—W. Ν. Francis. 1967 Computational analysis of Present-day American English. Providence: Dept. of Linguistics, Brown University. LALME = A linguistic atlas of late mediaeval English. 4 vols. 1986 Ed. Angus Mcintosh—M. L. Samuels—Michael Benskin. Aberdeen: Aberdeen University Press. Leech, Geoffrey 1981 Semantics. (2nd edition.) New York: Penguin. Leith, Dick 1987 A social history of English. (Language and Society.) (Reprinted.) London—New York: Routledge and Kegan Paul.

Expository apposition

179

Lewis, Charlton T.—Charles Short (eds.) 1975 A Latin dictionary founded on Andrews' edition of Freund's Latin dictionary. (Revised, enlarged and in great part rewritten.) Oxford: Clarendon Press. Lyons, John 1979 Semantics. Vol. 2. (First published in 1977.) Cambridge—London—New York—Melbourne: Cambridge University Press. Maley, Yon 1987 "The language of legislation", Language in Society 16/1: 25-48. Matthews, Peter H. 1981 Syntax. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. McMordie, W. 1976 English idioms and how to use them. (Revised by R. C. Goffin. Fourteenth impression.) Hong Kong: Oxford University Press. MED = The Middle English dictionary. 1952Ed. Hans Kurath—Sherman M. Kuhn, et al. Ann Arbor, Michigan: Michigan University Press. Mellinkoff, D. 1963 The language of the law. Boston: Little, Brown and Company. 1982 Legal writing: Sense and nonsense. New York: Charles Scribner's and Sons. Meyer, Charles F. 1987 "Apposition in English", Journal of English Linguistics 20/1: 101121.

1989

"Restrictive apposition: An indeterminate category", English Studies 70: 147-166. 1991 "A corpus-based study of apposition in English", in: Karin Aijmer —Bengt Altenberg (eds.), 166-181. 1992 Apposition in contemporary English. Cambridge: Cambridge University Press. Miliö, L. T. 1990 "A new historical corpus", ICAME Journal 14: 26-39. Minnis, A. J. (ed.) 1989 Latin and vernacular. Studies in late-medieval texts and manuscripts. (York Manuscript Conferences: Proceedings Series.) University of York, Center for Medieval Studies: D. S. Brewer. 1993 Chaucer's Boece and the medieval tradition of Boethius. (Chaucer Studies 18.) Cambridge: D. S. Brewer.

180

Päivi Pahta and Saara

Nevanlinna

Mitchell, Bruce 1985 Old English syntax. 2 vols. Oxford: Clarendon. Mustanoja, Tauno F. 1960 A Middle English syntax. Part 1. (Mimoires de la Societe N6ophilologique 23.) Helsinki: Societe Neophilologique. Nevalainen, Terttu—Helena Raumolin-Brunberg 1993 "Early Modern British English", in: Matti Rissanen et al. (eds.) 1993: 53-73. Nevanlinna, Saara 1974 "Background and history of the parenthetic as who say / saith in Old and Middle English literature", Neuphilologische Mitteilungen 75/4: 568-601. 1984 "Continuity and variation in Old English psalm glosses", in: Häkan Ringbom—Matti Rissanen (eds.), 23-36. 1990 "The occurrence of glosses in three Late Middle English texts: Lexical variation", in: Jacek Fisiak (ed.), Historical linguistics and philology. (Trends in Linguistics. Studies and Monographs 46.) Berlin: Mouton de Gruyter, 273-289. 1993 "AS WHO SAY/SAITH revisited: Form and function", Florilegium 11 (Special I.A.U.P.E. Issue. Carleton University Papers on Late Antiquity and the Middle Ages, ed. D. Wurtele.) Ottawa: Carleton University, 61-81. Nevanlinna, Saara—Päivi Pahta forthc. "Middle English nonrestrictive expository apposition with an explicit marker", in: Jacek Fisiak (ed.), Studies in Middle English. (Trends in Linguistics. Studies and Monographs.) Berlin—New York: Mouton de Gruyter. Nevanlinna, Saara—Päivi Pahta—Kirsti Peitsara—Irma Taavitsainen 1993 "Middle English", in: Matti Rissanen et al. (eds.) 1993: 33-51. The New English Bible 1961 London. Norri, Juhani 1992 Names of sicknesses in English, 1400-1550: An exploration of the lexical field. (Annales Academiae Scientiarum Fennicae. Dissertationes Humanarum Litterarum 63.) Helsinki: Academia Scientiarum Fennicae. Oakden, J. P. 1930-35 Alliterative poetry in Middle English, 1: The dialectal survey; 2 (with E. R. Innes): Survey of the traditions. Manchester: Manchester University Press.

Expository apposition

181

Obler, Loraine—Lise Menn (eds.) 1982 Exceptional language and linguistics. (Perspectives is Neurolinguistics, Neuropsychology, and Psycholinguistics: A Series of Monographs and Treatises.) New York—London: Academic Press. ODEE = The Oxford dictionary of English etymology. 1976 Ed. C. T. Onions (with the assistance of G. W. S. Friedrichsen and R. W. Burchfield). (Reprinted with corrections.) Oxford: Clarendon Press. OED = The Oxford English dictionary. Vols. 1-12 with supplement and bibliography. 1961 Ed. James A. H. Murray—H. Bradley—W. A. Craigie—C. T. Onions. Oxford: Clarendon Press. 1989 Second edition. Ogden, Margaret S. 1971 The Cyrurgie of Guy de Chauliac. (EETS, O.S. 207.) Onions, C. T. 1971 Modem English syntax. A new edition of An advanced English syntax completely revised by B. D. H. Miller. London: Routledge and Kegan Paul. 1986 A Shakespeare glossary. Enlarged and revised by Robert D. Eagleson. Oxford: Clarendon. Oostdijk, Nelleke 1994 Review of Meyer 1992, ICAME Journal 18: 83-93. Pahta, Päivi 1994 Review of Meyer 1992, Neuphilologische Mitteilungen 95: 385-387. Pahta, Päivi—Saara Nevanlinna forthc. "Markers of expository apposition in a diachronic perspective". Peltola, Niilo 1960 "Appositional constructions in Old English prose", Neuphilologische Mitteilungen 61: 159-203. Plucknett, Τ. F. Τ. 1940 A concise history of the common law. London: Butterworths. Prins, A. A. 1952 French influence on English phrasing. Leiden: Universitaire Pers Leiden. Quirk, Randolph—Sidney Greenbaum 1973 A university grammar of English. London: Longman. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1972 A grammar of contemporary English. London: Longman. 1985 A comprehensive grammar of the English language. London: Longman.

182

Päivi Pahta and Saara

Nevanlinna

Ringbom, Häkan—Matti Rissanen (eds.) 1984 Proceedings from the Second Nordic Conference for English Studies. (Publications of the Research Institute of the Abo Akademi Foundation.) Abo: Abo Akademi. Rissanen, Matti 1993 "The loss of wit 'know': Evidence from the Helsinki Corpus", in: Risto Hiltunen et al. (eds.), 195-206. Rissanen, Matti—Ossi Ihalainen—Terttu Nevalainen—Irma Taavitsainen (eds.) 1992 History of Englishes. Berlin: Mouton de Gruyter. 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—New York: Mouton de Gruyter. Robinson, Fred C. 1985 Beowulf and the appositive style. Knoxville, Tennessee: University of Tennessee Press. Ryd6n, Mats 1966 Relative constructions in early sixteenth century English with special reference to Sir Thomas Elyot. (Universitatis Upsaliensis Studia Anglistica Upsaliensia 3.) Uppsala: Almqvist and Wiksell. Samuels, M. L. 1970 Linguistic evolution. With special reference to English. (Cambridge studies in linguistics 5.) Cambridge: Cambridge University Press. Sopher, H. 1971 "Apposition", English Studies 52: 401-412. Strang, Barbara Μ. H. 1970 A history of English. London—Colchester: Methuen and Co. Svartvik, Jan—Randolph Quirk (eds.) 1980 A corpus of English conversation. (Lund Studies in English 82.) Lund: Lund University Press. Swales, John M. 1990 Genre analysis. English in academic and research settings. (Cambridge applied linguistics series.) Cambridge: Cambridge University Press. The Book of Common Prayer. See above. Thomas Thomas's Dictionarium linguae latinae et anglicanae 1587. 1945 Ed. R. C. Alston. (English Linguistics 1500-1800.) Menston: Scolar Press. Visser, F. Th. 1972 An historical syntax of the English language. Part 2. (Second impression.) Leiden: Brill.

Expository apposition

183

Voigts, Linda Ε.—Michael R. McVaugh (eds.) 1984 A Latin technical phlebotomy and its Middle English translation. (Transactions of the American Philosophical Society 74, Part 2.) Philadelphia. Vulgata = Biblia sacra iuxta Vulgatam Clementinam 1950 Ed. R. P. Alberto Colunga—Laurentio Turrado. Wallner, Björn 1987 "On the ./. periphrasis in the N.Y. Chauliac", Neuphilologische Mitteilungen 88/3: 286-294. Webster, Jonathan 1988 "The language of religion: A socio-linguistic perspective", in: M. Ghadessy (ed.), 85-107. Wrenn, C. L. (ed.) 1973 Beowulf. With the Finnesburg fragment. (Third edition. Fully revised by W. F. Bolton.) London: Harrap.

Genre conventions: Personal affect in fiction and non-fiction in Early Modern English Irma Taavitsainen

1. Aim and plan of the study This study discusses Early Modern English genres and text types and proposes a new way of looking at them. The focus is on differences between fiction and the adjoining text types. The starting point of the assessment is the level of individual texts; genres defined by external factors form the first level of abstraction and text types the second. Variation in features marking subjectivity provides a possible key to a description of text types. In this study a broad range of such features is used to assess whether there are any distinctive traits in early prose fiction that tell this mode apart from other narrative but non-literary text types; and whether and how statistical analysis of linguistic features can cast new light on our understanding of early fiction and its evolution. The aim of this study is to chart the difference between adjoining text types in terms of conventionalised discursive properties, i.e. to pinpoint text-type markers. Personal affect may provide a tool for a more detailed definition of these properties because it shows a great deal of variation in both quality and volume. The co-occurrence patterns of features marking subjectivity may also make a more detailed description of the qualities of personal affect possible. The plan of my study comprises several phases. First, I give the theoretical frame of the study. This theory is based on several sources and my previous empirical studies (Taavitsainen 1993, 1994, 1995a, 1995b, and forthcoming a and b). The inspiration for this study came from structuralist literary criticism, and I developed my research task so that it is suitable for corpus-based statistical analysis.

186

Irma Taavitsainen

The hypothesis is that the co-occurrence patterns of features marking subjectivity provide a means to a more detailed description of genres and their affinities. I have used factor analysis to identify adjoining text types and matrices of linguistic features that may be of importance in defining the quality of personal affect. The genres under scrutiny in this phase are plays, fiction, biography, autobiography, diary, travelogue, history, handbooks, science, education, and philosophy in the Early Modern English section of the Helsinki Corpus (HC hereafter). The purpose is to find out relevant groupings of adjoining and perhaps overlapping genres, and linguistic features that need to be assessed in more detail, for the ultimate aim of defining genre conventions and their changes. This is hardly possible on the basis of factor analysis only, and therefore this method has been complemented by other quantitative and qualitative methods. Factor analysis arranges the measured features into hierarchical scales according to their importance in reflecting the underlying dimensions of variation, and thus it helps to select features for further assessment. The third step is to select the genres with closest affinities to fiction for closer scrutiny and make a microanalysis of linguistic features that are potential text-type markers. This study concentrates on fiction and its nearest genres. The material of this phase consists of 6 texts of early fiction, and 24 other texts: travelogues, diaries, biographies, autobiographies, and history writing. This grouping is suggested by factor analysis. In this assessment, the factor analysis is complemented with other statistical methods for the purpose of pinpointing some conventionalised discursive properties. T-tests and variance analysis are employed to verify whether the linguistic features, indicated by factor analysis as important in describing the underlying dimensions of personal affect, are significant in telling the adjoining text types apart. The additional methods are necessary for testing whether the features with highest factor loadings really belong to the class of conventionalised discursive properties that I set out to establish in this study. The final phase of the study consists of qualitative analysis of passages that show high frequencies of features studied by statistical methods. Four different aspects prove important: "surge" features of personal affect, especially interjections; interactive focus; immediate context; and spacebuilding modality. At the end of the study, conclusions are drawn and suggestions made for future studies.

Genre conventions: Personal affect

2.

187

Genre theory

2.1. Genres and text types The HC uses genres as structural elements. When assigning texts to genres, we charted the fields of writing in each period, and classified texts according to external evidence of provenance and function. Labels like "travelogue", "religious treatises", "sermons", and "biography" are based on this external evidence. They provide a structural element in the corpus, and we chose to call them text types (1986; for the principles of text selection, see Rissanen et al. 1993). In this paper, however, I follow the usage of Biber (1988) according to which the term genre applies to classification according to external evidence, and text types are groupings made according to internal linguistic features of texts. Thus genres and text types are two kinds of abstraction used in text taxonomies, but based on different criteria. External evidence, such as textual origin, purpose and audience, provide one set of criteria, and the term genre is used of such groupings; internal linguistic features provide another way of classification, leading to text types.1 For example, moon prognostications can be classified according to their external starting point, i.e. the moon by itself or in combination with other celestial bodies, their purpose and their provenance in handbooks for doctors, almanacs and household literature (Taavitsainen 1988). When the internal evidence is examined, they can be classified as instructional texts according to their use of imperative forms, second person pronouns and prescriptive features of language. In contrast, parts of them, for example more comprehensive biblical allusions, are classed as narratives as they contain past tense verbs, a chronological sequence of events and third person pronouns. The distinction beween the two types of abstraction is useful and necessary if we wish to chart the linguistic features that serve as a matrix in distinguishing various text types and pinpoint differences between adjoining groupings (see below). Both genres and text types are abstractions made on the basis of individual texts. Thus the level of the text itself is the only reliable starting point. Genre criticism of Renaissance texts has focused on literary writings. The present study takes a somewhat different approach. Its incentive came from Tzvetan Todorov's claim that every literary genre has non-literary counterparts that stand near it (1990: 11), and the argument that distinctions between literary and non-literary styles are at best arbitrary (Carter—

188

Irma Taavitsainen

Nash 1990: 3). If this is true, it must be possible to verify it in more detail. The idea of testing it fitted into my current thinking about genres and text types, and textual conventions within them (Taavitsainen forthcoming c). It seems that text types may be very different from one another in their stylistic markedness, so that some linguistic features may be more salient than others at any one time. These features may change in the course of time though the same label may persist, for example, politeness formulae are more prominent in Renaissance texts than earlier, instructions in handbooks may be given directly in imperative forms or by descriptive passages, or in dialogue form. The recognition of such salient features could be helpful in defining genre styles of past periods. It is evident that no feature alone is sufficient for describing genre styles, but a matrix is needed. Yet the features within the matrix can be ordered, and a hierarchical scale may be helpful. Furthermore, the dynamics of textual features along a time scale may be revealed through the changing order of importance of matrix features occurring in a diachronic sequence of sychronic descriptions.

2.2. Genre conventions The social practice of texts and genres has received increasing attention. In a given society, the recurrence of some linguistic features may be institutionalised, and individual texts in that genre are produced and perceived in relation to that norm. Late medieval and renaissance dedications provide a good example, and letters with their well-defined politeness strategies are another case in point. Instructions for epistolary compositions were in great demand on the continent and in England, and guidebooks on how to write letters in the correct way were reproduced in several editions (Guillen 1986: 72-74). 2 Besides such regulation from "above", the influence could come more unconsciously from "below", as the conventions may have been created under the influence of social prestige or in imitation of older masters. Chaucer must have served as a model for later writers in the art of storytelling. The broadening of the audience, on the one hand, and of the topics, on the other hand, may have had an influence, so that features of colloquial speech may have found their way more readily into the language of comedies and fiction (see Taavitsainen forthcoming b). Furthermore, the formal difference, whether verse or prose, must be related to the changing functions of the media (see 2.4).

Genre conventions: Personal affect

189

Both literary and non-literary text types are choices among the discursive possibilities available in the language system. Both genres and text types guide text production and regulate the reception of texts; the social situation determines the choices, within the frame created by already existing writings pertaining to the same situation. Both authors and readers are thus influenced by the existing system, and changes are brought about by bringing in additional linguistic features and leaving out old ones, or by broadening the target groups, functions, etc. (cf. Fowler 1982: 11). A text type, whether literary or not, is a codification of discursive properties, i.e. linguistic features, but its individual members contain these features in various degrees and in various combinations. Text types may be very heterogeneous in respect to the occurrence of linguistic features; there is variability even within narrowly defined groups of texts (Meurman-Solin 1993a, Taavitsainen 1993), though some text types may be very limited and have strict rules. 3 It is because text types exist as conventions that they function as guidelines for both readers and writers. Conventionalised opening phrases like "once upon a time" serve as signals to the audience and tune their expectations. At the same time the same conventionalised phrases act as models for authors, as they provide the norms. Individual linguistic features, and their combinations, may work in the same way and provide models for imitation. This idea is especially relevant when dealing with periods in which imitation was a virtue. In the Late Middle Ages and the Renaissance the goal of a poet was to bring the accepted form of versification, the sonnet for instance, to perfection; the aim was to recreate the old, not to renew it. Writing belonged to the public world, and it was far removed from the private concerns of modern authors. For example, literature intended for the court circles strove to achieve the accepted norms. This principle must have worked on a broader basis as well, and influenced other writers of the period. The time span of the HC covers the transition from medieval conceptions of literary production as imitation to the modern outlook based on individual creativity. This change should also show in the selection of linguistic features. Any verbal property, optional at the level of language, may be obligatory in discourse; the choice a society makes among all the possible codifications determines its way of writing both in relation to genres and text types. Personal letters are different from official correspondence, scientific discourse has its own rules, and so on. These conventions may vary greatly in a historical perspective.

190

Irma

Taavitsainen

The institutionalised discursive properties of various text types may stem from any level of language, from semantic, syntactic or lexical sources, and the difference between one text type and another may be situated on any of these levels (Todorov 1990: 18). Other names for such discursive properties are distinctive features or text-type markers. The above view may be criticised for being reductive, but it gives an explicit formulation to the theory and indicates a case for empirical studies. The problem can be approached, for example, with a statistical study of linguistic features in various text types. It is evident that no single feature is sufficient to describe a text type, or even the style of a single writer, but a matrix is needed. The co-occurrence of linguistic features working in the same direction is important. The co-occurrence patterns of a great number of linguistic features make up several linguistic matrices that form a multidimensional scale of linguistic variation. Yet if the aim is different the approach can be modified. If the aim is to pinpoint a difference between adjoining genres, as in this study, the notion of a distinctive feature becomes important, as according to the present assumption the difference between one text type and another may be found in such features, though it is their co-occurrence that determines the style. An overall assessment with a matrix of features may suggest stylistic markers of particular period and genre styles. Linguistic features to be studied must be selected so that they form a coherent whole but still show enough variation to form different co-occurrence patterns.

2.3. Fiction and non-fiction The difference between fiction and non-fiction is one of the key questions of aesthetics and literary theory, starting from Aristotle. The difference has been defined in various ways, e.g. in terms of referentiality, truth values, modes of knowledge, speech act theory, audience response, and purpose of the text. For example, the question of mimesis can be broadened to include non-fictional writing such as prose dialogue in which characters are created for instructional purposes in works like Walton's The Compleat Angler (Imbrie 1986: 66). Other prose genres with close affinities to fiction include biography, autobiography and travelogue, in which fictive and historical elements can blend and pose the mimetic questions of "design" and "truth" (Hart 1974: 221).

Genre conventions: Personal affect

191

Early fiction and especially the early novels are well-known for their emotive qualities, with frequent outbursts of personal feelings. Literary style became highly conventionalised in this respect; an example is Euphuistic style, which was "an artistic embodiment of a world view" and had a courtly appeal with its elaborate patterns and analogies (Margolies 1985: 46). John Lyly wrote for the high social circles and was widely imitated. Yet there were other kinds of fictional writing, such as pastorals, romances, and literature with a broader social basis, and other patterns must have influenced fictional writing as well. In particular, short fiction for a wider and more general audience with different tastes must have had other models. It has been said that the purpose of early fiction was to evoke awe and wonder, the defining characteristic of romance being the immediate, powerful impact on the reader of the separate episodes, each of which calls for a full response and are both primitive and sophisticated at the same time. Characters are said to manifest states of uncontrolled passion that burst into either extravagant lamentation or violent action. It has also been said that emotion in early fiction may be mentioned but it is never particularised, and therefore remains, like "noble chivalry", a value so generalised that it means practically nothing and does not provide an experience for the reader (Margolies 1985: 36).4 The subtle and refined intricacy of early works like Arcadia or Euphues was created for a culturally learned audience that appreciated literary conventions beyond story-telling. In the course of time the readersip broadened, and there was a circular line of development. When the literature originally intended for a courtly audience spread to the bourgeoisie, the aristocracy rejected what the middle classes imitated; when Euphues gained popularity with a common audience, it became an object of mockery among the aristocracy. The same happened to romance (Margolies 1985: 14-16). This circular movement can be found even earlier in literary history. Chaucer's original audience consisted of the court circles and the upper middle classes that had dealings with the court, such as rich merchants. By the peak popularity of Euphues and Arcadia, the same stories circulated among the lowest literate classes in jest books like Me/77 Tales.

192

Irma

Taavitsainen

2.4. Prose and verse The period in focus poses special problems as the relationship between prose and verse was in a state of flux. In the 15th and 16th centuries, fiction was still largely written in verse, Sir Philip Sidney's Arcadia being a good example, but prose started to be used for this purpose increasingly. Prose fiction was firmly associated with an instructive function and moralising purposes were prominent. The conventions of prose fiction were in the process of formulation, and verse fiction of the previous period should not be ignored when charting the origin of these conventions. Some features of medieval verse fiction may have been adopted into the prose fiction of the 16th century, and its generic markers may have persisted in early prose fiction.

2.5. From manuscript to print Another important change that was simultaneously going on had to do with the transmission of literature through print. It is well known that early printers imitated manuscript lay-out and illumination. Reading was still a social activity rather than a personal exercise in the Early Modern period. Manuscripts were produced for a certain, well-defined readership, for local circulation, and they provided the model even when printing had become more common. The conventions of reader involvement were carried over so that early works often had the personal, immediate quality of direct address. Authors wrote for an assumed readership long after the beginning of printing. This is especially noticeable in translations, which often had an extensive commentary for the readers' benefit. In fiction the authors and readers were thought to share a literary experience, and the change to a more diffused presence of the author or commentator was slow. The change from manuscript circulation to print involved an objectification of the text, which happened only gradually. The first works clearly perceived as printed, in contrast to the more private circulation by manuscripts, are from the very end of the 16th century (Margolies 1985: 22-28).

Genre conventions: Personal affect

193

3. Personal affect

Personal affect means the expression of subjective emotions, feelings, moods and attitudes. It has been defined in various ways in the literature, but no generally accepted set of linguistic features expressing personal affect in Present-day English has been defined. Cultural environment has a powerful effect upon emotional expression, as different display rules are obeyed (Frijda 1986: 62). Cultural conventions are crucial to both the amount and kind of personal affect; this applies to gestures as well as linguistic behaviour, and both to spoken and written language. According to my earlier studies, the degree of involvement in the form of personal affect varies greatly in intensity and means of realisation from text to text. It provides a fruitful object of study, and a way of describing genre conventions. For example, the volume of personal affect varies greatly in Middle English religious treatises so that the differences are great even among mystical authors (Taavitsainen 1993 and 1994b). It is an essential component of participant relations and finds outlets in various forms. It may be present explicitly as direct guidance and comments to the reader, as for example in religious texts. More indirect ways of involvement include the creation of a mock reality in which the readers are supposed to live with the depicted characters, feel their emotions and imagine themselves in their situations. For example in fiction, the pretence of recorded speech creates a sense of immediate presence and widens the scope of interaction with the audience; in religious texts readers are supposed to join in the prayers. Personal attitudes are also expressed by the point of view in texts, subjectivity being the prevailing mood in texts with high "volume" of personal affect. 5 In general, emotive focalisation implies a noticeably idiosyncratic way of representation (cf. Toolan 1988: 73).

3.1. Towards a definition of the quality of personal affect In earlier studies, affect has been classified into positive and negative features. Yet it is evident that other, more subtle and perhaps more important shades can be distinguished. Participant relation is important in its various forms: feelings about oneself often pair off with attitudes to someone else. Emotions can be reflected on someone else, and a personal point of view can open one's mind and way of seeing things to others. Besides negative and positive expressions of personal affect, two other dimensions

194

Irma

Taavitsainen

have been proposed for further study. They are "surge/predisposition" and "self/other" (Martin 1992: 533). In the present study I understand "surge" as an outburst of emotion, an expression that conveys intensified personal charge between the participants of communication, whether text-internal or exophoric in reference. The dimension "self-other" includes various aspects of involvement and interactivity versus egocentric focus. Negative and positive expressions of personal affect were assessed in Biber's and Finegan's study (1989b) with adjectives and adverbs, but their results were not particularly good: in a cluster analysis only 15 out of more than 400 texts showed relatively high frequencies of personal affect features. It may well be that the features in their study did not grasp the essence, and better results may be achieved by different means, along other dimensions. My earlier studies showed that different genres have their own codes, and that these can only be grasped by close reading and detailed textual analysis (Taavitsainen 1993). The present study concentrates on fiction and non-fiction. The measurement of features reflecting involvement and emotional charge could perhaps illuminate the problem of the relationship between early fiction and other narrative text types of the same period. In addition, a diachronic approach seems to be helpful in approaching the question of conventions within a changing system of genres, and a longer time span may also reveal distinctive features and give evidence for describing the evolution of text types. According to my theory, personal affect may be realised in different ways in the participant relation, and such features may well serve as formatting factors in text production and reception.

3.2. Point of view In my definition the subjective point of view provides a key to the assessment of personal affect. Our conceptualisation of reality begins with the individual, the mental self, and "I" serves as the reference point (Benveniste 1966 [1971]; Mühlhäusler—Harre, 1990; Heine et al. 1991: 12; Fludernik 1993: 46, 431^132; Adamson 1994: 195). It has been claimed that every interpretation of history is based on the present moment, just as space starts with "here", and the notion of other people with "I". Four different spaces have been discerned in this egocentric illusion of "I-herenow-in this case"; they form a scale which orders social, spatial, temporal, and logical spaces in an inward/outward progression. Social space with "I"

Genre conventions: Personal affect

195

stands closest to the core, and the first person singular is the basic unit in this reference system. Deixis may therefore become important in this dimension. Five major types of deixis have been identified: social and discoursal, as well as those of person, place and time. Egocentricity is crucial as the speaker in any language act situates referents both temporally and spatially to him- or herself, speaking "here" and "now" (Toolan 1988: 127). Reader involvement is another important issue connected with point of view. Involvement may be created in various ways, and one challenge is to find out both the evident and subtle mechanisms used for this purpose and the various genre conventions. For example, advice on how to interpret the text may be given, or the author may assume that the reader shares his views and feelings, which is a more indirect way of influencing the audience.

4. Material The focus of this study is on early fiction. The HC includes passages of fictional texts targeted at various layers of society. The following groups can be discerned: 1) 2) 3) 4)

popular fiction; (a) and (c) sociological literature; (d) middle-class fiction; (b) and (e) romantic fiction; (f).

The popular trend is well represented: (a) Merry Tales contains comic turns and moralising authorial comments. It is the earliest English jest book, (b) Thomas Deloney represents middle class fiction. The last of the Elizabethan realists, he started with ballads but turned later to write narrative prose. His work continues and develops the tradition of jest books and anecdotes; about one third consists of separate jests, but he uses them for social and moral commentary, thus developing them into a whole. His book is a forerunner of later novels in this sense (Davis 1969: 239-244). (c) Penny Merryments belongs to the layer of popular literature written for the broadest possible audience. It contains stories like "Honest John" and "Loving Kate" describing the marriage plans of two in-servants for an

196

Irma

Taavitsainen

audience of the same level of society (Spufford 1981: 60). (d) Thomas Harman's book is meant as an "alarum" to forewarn honest citizens, but in fact it belongs to sociological literature, (e) Robert Armin was a chief actor of comic parts in the early 17th century. He wrote one play with genuine dramatic power and some prose works, (f) The final text, very different from the others, is by the first professional woman writer, Aphra Behn, a dramatist and novelist. Her most famous novel Oroonoko owes much to her childhood memories of plantation life in Surinam (Sampson 1970: 929).

4.1. Parameters of the HC: interactive textual form The parameter value "Interactive" has been assigned to plays and texts in the dialogue form in the HC, but texts with the value "X" also contain interactive passages. Thus a large part of the present material consists of dialogue without time constraints or paralinguistic gestures and intonation to aid the interpretation, but imitating and perhaps condensing some features of spoken discourse. Thus the different levels of spoken language have to be taken into account, as well as the proportion of direct and indirect narration.6 The word counts of direct speech versus narration in fiction show a great deal of variation: Merry Tales:

narration 5093 words; total 6426 words narration 3107; direct Harman: narration 3914; direct Armin: narration 3002; direct Deloney: Penny Merriments: narration 1690; direct narration 5120; direct Behn:

direct speech 1333 words; speech speech speech speech speech

2012; total 5119 1257; total 5171 4321; total 7323 4882; total 6572 357; total 5477.

The time constraints of spoken language may cause reduced features of syntax such as THAT-deletion, and the more involved quality in the use of first and second person pronouns. Immediate context and indexical use of language may be a characteristic of spoken language as the context is usually clear; in written language the context has to be created by different means. In general, spoken language is less edited and integrated, but rapid adjustments to subtle nuances of interpersonal meaning are possible, and spoken language has its own kind of complexity (Halliday 1989: 7884). Thus personal affect features are likely to be found in passages that

Genre conventions: Personal affect

197

imitate speech, and its linguistic features are likely to be among those of spoken language. Interaction proved extremely important for the present assessment, but the parameter itself was not very useful (see note 11).

4.2. Parameters of the HC: informal language Variation in language may be described in several ways, and there are subtle shades of difference between various labels. Formal versus informal is a pair of such labels connected with situation, and employed as a parameter in the HC. Non-standard language may be found at the informal end and such language may be socially or regionally marked. Imitations of vulgar or provincial speech are well known from earlier periods. Nonstandard speech may be emotionally loaded, swearing being a good example. In fiction, dialogues are important and they are likely to contain deliberate informal features. Early short fiction in the present material consists of jest books and meriy tales for the broadest possible audience. If the classical rules of adjusting the style of writing according to audience education are at all valid in this connection, the distinctive features should be expected at the more informal, spoken end of the scale. On the other hand, the moralising and edifying function may also be reflected at the level of linguistic features, and indeed the edifying purpose may be dominant. Conventionalised politeness strategies are directly relevant to the present attempt to trace personal affect features in various text types. For the influence of this parameter, see note 11.

4.3. Parameters of the HC: imaginative versus non-imaginative In the Early Modern section of the HC, fiction is assigned the parameter value "NI" ("Narration Imaginative"), while other texts have the value "NN" ("Narration Non-imaginative"). As the aim of this study is to search out possible distictive features of early fiction, this parameter was selected for further scrutiny (see section 7).

198

Irma Taavitsainen

5. Linguistic features Linguistic stylistics uses a broad range of linguistic features with various functional properties (see e.g. Leech—Short 1981). For my aim, Biber's list of involvement features (1988: 89) is helpful, but not applicable as such to the present purpose or to a historical corpus because of, for example, linguistic and semantic changes. As pointed out before, no set of linguistic features describing the quality or qualities of personal affect has been defined in Present-day English, and the problem is even greater with the past stages of language and in past sociocultural environments. The choice of linguistic features was based on my previous studies and my reading of these texts, and background knowledge of early fiction. Some of the features, like personal pronouns, were an obvious choice, but I tried to be more inventive and include features that might prove important and illuminate the problem of defining personal affect (see above 3.1). In order to achieve a more detailed description of this aspect of participant relations in texts, I selected a fairly broad range of linguistic features that might be thought to be relevant to personal affect. Personal focus and deixis are potentially important. Yet the range of features should be broader as co-occurrence patterns can be assumed to provide a key to defining the quality of personal affect in various text types. Therefore I selected a number of other features which had been established as relevant along the dimensions spoken-written / informal-formal / spontaneous-more processed in earlier studies. Their affinity to personal affect could be more indirect and grasped via matrices (see 2.1). A brief list of the selected features is given, with the working assumptions based on earlier studies. The features and the aspects that become important in the assessment will be discussed in detail later.

5.1. Personal pronouns (subject forms): I, WE; YE/YOU; HE, SHE, IT, THEY All subject forms of personal pronouns were assessed; other forms were not included in this phase of the study, but they are included in the discussion later (see section 9). In earlier studies, personal pronouns proved to be powerful in determining the kind of involvement in texts, and the interpretation is straightforward: first person pronouns express ego-in-

Genre conventions: Personal affect

199

volvement, second person pronouns involvement with the addressee, and third person pronouns involvement with the subject-matter; the alternative is to use nouns instead (Chafe 1985). In general terms, first and second person pronouns mark the presence of a narrator and an addressee, while third person pronouns mark relatively inexact reference to persons outside the immediate context of interaction. According to Biber's study (1986), third person pronouns co-occur frequently with past tense and perfect aspect forms, as a marker of narrative, or reported (versus immediate) styles. A frequent use of the pronoun it has been interpreted as a sign of a relatively inexplicit lexical content due to strict time constraints and noninformational focus, and thus it has been labelled as a feature of spoken language.

5.2. Direct questions: WH-, HOW Questions, like second person pronouns, are typically interactive and indicate a concern with interpersonal functions and involvement with the addressee (Biber 1986). Yes/no questions were excluded from the present analysis because they could not be accurately identified by automatic analysis.

5.3. Proximal versus distal context Demonstrative pronouns THIS/THAT, THESE/THOSE can refer to an entity outside the text, an exophoric referent, or to a previous referent in the text itself, often to an abstract concept (e.g "this shows ..."). Their use without a nominal referent has been interpreted as a feature of spoken language due to faster production and lack of editing (Chafe 1985). Adverbs of proximity NOW and HERE are also relevant to the immediate context and situationally bound discourse, and thus to egocentric focus.

5.4. Reductive features Reductive features have been considered typical of the spoken language. Subordinator THAT-deletion is a form of syntactic reduction relevant to the present concerns; in Biber's study it had the second highest factor loading in the involvement factor (0.91); contractions were also important

200

Irma

Taavitsainen

in his study (factor loading 0.90), but in the present study they are not as relevant as the only contractions are verbal forms like V, '// of modals and 'c/ of past tense forms. They become frequent only in the last period of the HC.

5.5. Exclamations Exclamations have been described as belonging to the purely emotive level of language (Quirk et al. 1985), but in addition to expressions of emotions, other functions can be distinguished. According to my previous study (1993) they were most important in expressing personal affect in Middle English religious texts. The following lexical forms were found in the Early Modern part of the HC: Ο! OH! A! AH! LO! LOO! LOE! LA! ALAS! ALLAS! FY! FIE! FYE! EH! BENEDICITE! GRAMERCY! AMEN! WHY! WHAT! HOW! TUT! TUSH! PSHA! They occur in a variety of functions that range from creating an illusion of spoken language to genrespecific conventions of story telling in the written form (Taavitsainen 1995a, 1995b, and forthcoming d; see also Blake 1992).

5.6. Swearing words Swearing words overlap with exclamations to some extent, but according to my pilot study they were even more restricted and occurred in few text types (Taavitsainen forthcoming b). Therefore it seemed fruitful to assess them separately. The following swear words and phrases were found: MARY, BY SAINT ..., FOR SHAME, FOR ... SAKE, BY MY ..., DEVIL, GOGS..., COCKES BONES!

5.7. Pragmatic particles Pragmatic particles also overlap with exclamations to some extent, but mostly their functions are different (see Jucker forthcoming). The lexical items included in the present assessment are: WELL, (I) PRAY (THEE/ YOU), PRITHEE, I SAY, YEA, NAY.

Genre conventions: Personal affect

201

5.8. Private verbs Private verbs had the highest factor loading in Biber's study (0.96). In his study they are treated as one class, without further subcategories. I decided to apply a more detailed semantic classification that could be helpful in defining the various kinds of personal affect, or personal point of view in texts. Certain restricted classes of verbs can be identified as having specific functions, such as verbs of cognition that refer to mental activities and sensory verbs such as SEE, HEAR, and FEEL. All occurrences were counted at first, but in the qualitative assessment the difference between the first, second and third person are discussed. In my study I applied the following semantic subcategories, based on my own judgement: 1. 2. 3. 4.

5. 6.

Emotion: LOVE, HATE, HOPE, DESIRE, COVET, LONG, ENVY, DREAD, FEAR Sensory perception: SEE, HEAR, TASTE, SMELL, FEEL Observation: OBSERVE, PERCEIVE, DISCERNE, NOTE, DISCOVER Mental activity: ESTEEM, DEEM, ESTIMATE, JUDGE, FIND, THINK, CONSIDER, GUESS, DOUBT, REASON, CONCLUDE, SUSPECT, SUPPOSE, PRESUME, ASSUME, ACCEPT, EXPECT, FORGET, REMEMBER, RECALL, BELIEVE Mental state: KNOW (WIT), UNDERSTAND, HOLD (AN OPINION) Imagination: IMAGINE, FANCY, DREAM.

5.9. Modals Modality in language can be understood broadly as encoding speaker attitudes and point of view. It is expressed in a boad range of features, among which modal auxiliaries are important and easily detectable in a computerised corpus. No definitive set of linguistic features encoding modality in language has been defined, but it can be assumed that such features are closely connected with personal affect, and thus they may occur in the matrices sought out in this study. It is possible to distinguish three functional classes of modals: 1) those marking permission, possibility, or ability; 2) those marking obligation or necessity; and 3) those marking volition or prediction (Quirk et al. 1985). Chafe (1985) includes possibility modals among the evidentials that mark

202

Irma

Taavitsainen

reliability, and necessity modals among those evidentials that mark some aspect of the reasoning process. Politeness strategies may also be relevant in the assessment of the use of modals. The following groupings were taken as the basis of assessment; the concept of boulomaic modality was introduced later, as it proved significant. Possibility: CAN, COULD; MAY, MIGHT Necessity: SHALL, SHOULD; MUST; OUGHT Volition: WILL, WOULD ('L(L), 'DE)

6.

Factor analysis

6.1. Method Factor analysis has proved to be a fruitful way of assessing textual affinities, and the method has been applied for various purposes (Biber 1988; Meurman-Solin 1993a; Taavitsainen 1993; Kytö forthcoming. For an alternative method, cluster analysis, see Biber 1989). As the aims of the present study are different from those of earlier studies using factor analysis, it may be helpful to specify the difference in more detail. In Biber's study the purpose is to give a comprehensive overview of variation across speech and writing, and of the complexity of text types (Biber 1988 and Biber—Finegan 1986). Meurman-Solin's study tests Biber's method with a historical corpus of early Scots, and she concentrates on Biber's Factor 1, interpreted as the involvement factor. In another study she elaborated her results by relating them to the audience parameter of her corpus of early Scots (1993b). Kytö's study is another application of Biber's involvement features, but this time to the subcorpus of Early American English. My own pilot study (1993) was designed to test whether various types of religious texts could be distinguished from one another on the basis of personal affect features. The linguistic features were selected on the basis of qualitative studies of Rolle's affective style, and my own reading of Middle English mystics. The features had to be detectable in an untagged corpus with a great deal of dialectal variation, and therefore lexical searches were the most appropriate. Exclamations which had not been as-

Genre conventions: Personal affect

203

sessed before proved to be most powerful in distinguishing genre/subgenre styles. The pilot study thus indicated a clear path to follow in a further study. It is obvious that stylistic features like parallelism, alliteration and repetition are not readily captured by statistical methods, but statistical approaches have to be complemented by qualitative readings. The present study is a further development along the same lines of thought. I concentrate on the volume and quality of personal affect, and whether features of personal affect are conventionalised to the degree that they can be used in distinguishing genre/subgenre styles. The aim is to proceed further with the assessment of co-occurring linguistic features in order to test with other methods whether any of them serve as distinctive discursive properties that tell the adjoining genres apart. The factor analysis is complemented by other statistical methods for this aim. Another aim is to define the quality of personal affect, and qualitative assessment is used for this purpose. Factor analysis seems to be very sensitive to the number of linguistic features under scrutiny: the broader their scope, the more general the description of a single factor. In Biber's study the number of features was 67 and factor analysis grouped them into 7 underlying dimensions. 7 Factor 1 was the most powerful in explaining underlying variation, and it was labelled "involved versus informational production". In Meurman-Solin's study the number of features was 19, all from Biber's factor one. Her results yielded a more detailed description of involvement. She suggested a division into author-centred texts, addressee-oriented texts, and a third kind of involvement focused on evidentiality. In my pilot study the number of features was five and they were deliberately selected so that they all contributed to the same direction and fell on one factor. This was desirable as the emphasis of the study was on factor scores of texts and the internal homogeneity versus heterogeneity of genres in respect to personal affect. It proved that the volume of personal affect within texts grouped according to external evidence varied a great deal, and confirmed my assumption that the quality of personal affect provides a fruitful area for more detailed research. In the present study the number of features has been greatly increased as the aim is to assess the quality of personal affect as well as the amount of variation along the various dimensions within the larger frame. Many of these features have not been included in earlier studies and thus there was no previous knowledge of how they would correlate with other features. Personal pronouns are central to the present approach, and the as-

204

Irma Taavitsainen

sumption is that the three kinds of involvement suggested by Chafe and confirmed by Meurman-Solin's factor analysis would be found in some form, and the additional features would help to define their nature in more detail. According to my pilot study (1993) exclamations were the most powerful feature in distinguishing Middle English genre styles, and it is evident that the use of interjections in a historical perspective is worth studying in more detail. The results confirm this assumption. Yet the quality of personal affect proved to be extremely complicated, and results were achieved thanks to the combination of the various methods; factor analysis alone was not sufficient for this purpose.

6.2.

Factorial structure

The linguistic features were gathered by WordCruncher and an exploratory factor analysis was performed. To start with, all features are treated equally. Factors group the linguistic features that co-occur with high frequency. The features loading on the same factor are interpreted to share a communicative purpose. The method also arranges the linguistic features into a hierarchical order according to their weights on the factor, i.e. their correlation with the factor denotes their importance in representing an underlying dimension. Consequently, factors represent co-occurrence patterns of linguistic features which are arranged in a hierarchic order according to their weight on the factor. Co-occurring features may be interpreted to be functionally related, constituting an underlying dimension of variation. The best number of factors has to be established by experimentation in order to find the solution with the most reasonable linguistic explanation. If too few factors are selected, the result is a loss of information, while too many factors lead astray in the interpretation of underlying dimensions. The features may be combined (e.g. singular and plural, / and we), some may be excluded, and others added. 8 It has also been emphasised that interpretations made on the basis of factor analysis should be verified by additional considerations (see section 9).

Genre conventions: Personal affect

Eigenvalues

1.5 -

10.5 -

0

1

1

1

Fl

F2

F3

Figure 1. Eigenvalues: scree plot.

205

206

Irma Taavitsainen

In the present study, a three-factor solution was considered best in describing the qualities of personal affect and in indicating relevant groupings of texts. It gives enough information, and the communalities of features reveal that no potential factors remain hidden. The correlation patterns reveal that exclamations, with the highest factor loading, correlate very strongly with swearing, pragmatic particles, verbs of sensory perception, verbs of emotion, expressions of proximity and first person pronouns, modals of necessity and volition; but also with second person pronouns and direct questions. Of the features of Factor 2, the pronoun it correlates strongly with verbs of observation and modals of possibility and necessity; and likewise the most salient features of Factor 3, the third person pronouns s/he, THAT-deletion and verbs of emotion and imagination, correlate with one another very strongly. Table 1. A three-factor solution (salient loadings having weights larger than 0.5). Factor 1 EXCLAMATIONS 0.888 0.582 SWEARING IT -0.145 THAT-deletion 0.275 SENSORY PERCEPTION 0.517 OBSERVATION -0.157 ACTIVITY (MENTAL) 0.127 STATE (MENTAL) 0.430 NOW 0.652 HERE 0.653 DIRECT QUESTIONS 0.838 PRAGMATIC PARTICLES 0.843 PERSON1 0.623 0.828 PERSON2 -0.176 PERSON3 POSSIBILITY 0.155 NECESSITY 0.416 VOLITION 0.693 EMOTION/IMAGINATION 0.272 Sumsqr 5.836

Factor 2 0.045 -0.046 0.682 0.026 -0.061 0.444 0.600 0.335 0.102 -0.091 0.178 0.108 -0.064 0.327 -0.129 0.720 0.696 0.312 0.369 2.570

Factor 3 0.201 0.188 -0.189 0.580 0.378 -0.179 0.294 0.487 0.104 0.004 0.215 0.185 -0.099 0.110 0.973 0.220 0.154 0.351 0.545 2.499

Sumsqr 0.831 0.376 0.522 0.413 0.413 0.254 0.462 0.534 0.446 0.435 0.780 0.756 0.402 0.804 0.995 0.591 0.682 0.700 0.507 10.904

Genre conventions: Personal affect

Figure 2a: 3 factor solution combined variables Fl

F2

207

208

Irma

Taavitsainen

Figure 2b: 3 factor solution combined variables Fl

F3

Genre conventions: Personal affect

Figure 2c: 3 factor solution combined variables F2

F3

209

210

Irma Taavitsainen

Factor 1 (eigenvalue 5.836): The strongest features contributing to this factor are exclamations (factor loading 0.888), pragmatic particles (0.843), direct questions (0.838), second person pronouns (0. 828), modals expressing volition (0.693), first person pronouns (0.623), expressions of proximity (here 0.653 and now 0.652), swearing (0.582) and sensory perception verbs (0.517). This factor can be interpreted as "Interaction" (see below) . Factor 2 (eigenvalue 2.570): The most important features contributing to this factor are modal auxiliaries of possibility (0.720) and necessity (0.696), the pronoun it (0.682), private verbs expressing mental activity (0.600) and observation (0.444). This factor is interpreted as "Reasoning" (see below). Factor 3 (eigenvalue 2.499): The third person pronouns he and she contribute to this factor very strongly (0.973); THAT-deletion is next in importance (0.580), then come private verbs expressing emotions and imagination (0.545) and mental state (0.487). This factor is named "Narration" as it seems to reflect narrative concerns. For illustration, see Figures 2a, b, and c (pp. 207-209).

6.3. Interpretation of factors Factor scores of individual texts tell how important the co-occurrence of the features constituting the factor are for each text. Thus the basic unit of measurement is an individual text, and texts of a given genre may be very different from one another in respect to various factors. Each text has its own place in this three-dimensional scale (see Table 2, pp. 211-212). Examples of texts with high factor scores on at least one factor are given below as "typical texts".

Genre conventions: Personal affect Table 2.

Factor scores of texts (for abbreviations, see pp. ix-x and 345-363 in this volume).

CAT PERIOD TEXT FICT 3 PLAY 1 PLAY 1 PLAY 3 PLAY 2 PLAY 2 PLAY 3 FICT 2 AUTO 1 HAND 3 1 FICT 2 FICT HAND 2 DIAR 2 SCIE 2 1 SCIE AUTO 3 HAND 2 EDUC 1 TRAV 1 TRAV 2 PHILO 2 DIAR 3 2 SCIE FICT 1 HAND 3 DIAR 2 TRAV 3 DIAR 3 PHILO 1 TRAV 1 TRAV 3 EDUC 3 PHILO 3

211

WORDC PARI

PENNY UDALL STEVENSO VANBR SHAKE MIDDLET FARQUHAR DELONEY MOWNT WALTON HARMAN ARMIN GIFFORD HOBY BLUNDEV RECORD FOX MARKHAM ASCHAM LELAND JOTAYLOR ELIZ EVELYN CLOWES MERRY LANGF MADOX FIENNES PEPYS COLVILLE TORKINGT FRYER LOCKE PRESTON

6560 4620 5950 7190 6170 5640 5550 7320 5740 5370 5120 5170 6190 6050 6420 6700 5560 6100 4900 6860 8850 6880 6070 6620 6430 6000 6470 5140 5140 9890 7240 5330 5200 8820

NI XX XX XX XX XX XX NI NN IS NI NI IS NN EX EX NN IS IS/EX NN NN XX NN EX NI IS NN NN NN XX NN NN IS XX

INF

INT

INF INF INF INF INF INF INF INF X X X INF X INF X X X X X X X X INF X INF X INF X INF X X X X X

INT INT INT INT INT INT INT INT X INT X INT INT X X X X X X X X INT X X X X X X X INT X X X INT

Fl 3.092 2.446 2.278 2.180 2.137 1.977 1.976 1.165 0.861 0.697 0.523 0.446 0.085 0.010 -0.054 -0.079 -0.145 -0.153 -0.215 -0.216 -0.231 -0.233 -0.236 -0.236 -0.244 -0.249 -0.268 -0.272 -0.274 -0.280 -0.348 -0.359 -0.370 -0.375

F2 0.516 0.364 -0.068 0.460 -0.082 -0.143 -0.118 -0.454 -0.383 0.822 -0.656 -0.731 0.352 -1.550 0.626 0.138 -0.806 1.083 -0.235 -1.136 -0.367 2.281 -0.937 0.339 -0.385 1.005 -0.412 -1.126 -0.474 1.947 -1.177 -0.563 1.054 2.367

F3 0.504 0.052 0246 -0.054 0.442 -0245 -0232 1.166 0303 -0.391 1.771 0.869 1.715 -1.642 -1.113 -1284 0.534 -0.470 -0.756 -1.487 -0.784 0.614 -0.980 -1.056 0.803 -0.946 -0.326 -0.893 -0.157 -0.062 -0.968 -0.980 0377 0.102

212

Irma

Taavitsainen

Table 2. Continued. CAT PERIOD TEXT BIO HAND DIAR DIAR EDUC HIST HIST HAND TRAV BIO HIST EDUC EDUC SCIE SCIE HIST EDUC HIST AUTO FICT SCIE BIO HIST

1 1 1 1 2 3 1 1 2 2 2 1 2 3 3 2 3 1 2 3 1 3 3

ROPER FITZH MACHYN EDWARD BRINSLEY MILTON FABYAN TURNER COVERTE PERROT STOW ELYOT BACON BOYLE HOOKE HAYWARD HOOLE MORERIC FORMAN BEHN VICARY BURNETR BURNETC

WORDC PARI 5440 5150 6280 6780 5540 5820 5420 4850 5930 4810 4810 5540 5680 5220 6060 5280 6120 5670 4120 5480 6180 6170 5820

NN IS NN NN EX NN NN IS NN NN NN IS/EX EX EX EX NN IS NN NN NI EX NN NN

INF

INT

X X INF INF X X X X X X X X X X X X X X X X X X X

X X X X X X X X X X X X X X X X X X X X X X X

Fl -0.433 -0.442 -0.443 -0.471 -0.480 -0.493 -0.525 -0.535 -0.567 -0.576 -0.591 -0.609 -0.612 -0.616 -0.652 -0.659 -0.669 -0.697 -0.853 -0.876 -0.883 -1.121 -1.231

F2 0.026 1.282 -1.453 -1.047 1.561 -0.717 -1.117 0.047 -0.898 -0.444 -0.943 0.344 0.432 1.303 0.858 -0.497 1.194 0.117 -0.740 -0.219 0.617 -0.439 -0.819

F3 0248 0.039 -1.127 -0.549 0221 -0.045 -0.174 -0.709 0.144 1249 -0.331 0.489 -0.949 -1.043 -1241 -0.105 -0.239 1.123 1.791 2.163 -0.6*3 2.187 2.828

A personal point of view, egocentric focus, or interactive communication are central to this study; these are all aspects of personal affect that might be relevant to defining its quality particularly in fiction. Thus their cooccurrence patterns should reflect more specialised qualities of personal affect. Unlike earlier studies of involvement, first and second person pronouns contribute to the same factor. Thus the well-established division into ego-involvement, involvement with the second person, and involvement with the subject-matter or evidentiality is not dominant. The interpretation of Factor 1 is not straightforward as both parts of the dichotomy between "self' and "other", i.e. the first and the second person pronouns, are important in this factor.

Genre conventions: Personal affect 6.3.1. Factor 1 (eigenvalue 5.836),

213

"Interaction"

This factor can be called "Interaction", and it is most powerful in explaining underlying variation in this study. It is evident that the interactive quality has to be assessed in more depth. Texts that score highest belong to the genre of fiction. Penny Merriments from the last period (EModE3) scored highest on Factor 1, and had positive values on both others; examples of texts help to define the qualities of personal affect in them. Comments of stylistic features not captured by the statistical analysis are also included in the discussion (see 6.1): FICTION: PENNY, factor scores: Factor 1: 3.092 (Factor 2: 0.516, Factor 3: 0.504) (Ka) Who shall be my Father to give me. (Jo) Thou mayest ask Jack Wheeler, but I know he had rather had thee himself. (Ka) Oh fie no, I will not ask him, he will take it for an affront, I will rather ask old father Bandol for he us'd to call me Daughter, and he will take it kindly. (Jo) Do then. (Ka) Does it not make you ashamed to talk of these things. (Jo) No I promise thee, I am proud of it, and so art thou I believe, but that thou wilt not confess it. (Ka) I would it were once over. (Jo) So would I, i'd as live as a groat. {Penny Merriments 119) This text is written in dialogue form with first and second person pronouns, direct questions, interjections, contracted forms, and some popular phrases. It is obviously written in imitation of natural language, with a high emotional loading between the participants. FICTION: DELONEY, Factor 1: 1.165 ( Factor 2: -0.454, Factor 3: 1.166) Who was it that checkt thee, I pray thee tell mee? was it not your old gossip, dame dayntie, mistresse trip and go? I beleeue it was. Why man if it were she, you know shee hath beene an old housekeeper, and one that hath known the world; and that shee told mee was for good will. Wife (quoth hee), I would not haue thee to meddle with such light braind huswiues, and so I haue told thee a good many times, and yet I cannot get you to leaue her company. Leaue her company? why

214

Irma

Taavitsainen

husband so long as she is an honest woman, why should I leaue her company? Shee neuer gaue mee hurtfull counsell in all her life, but hath alwayes been ready to tell me things for my profit, though you take it not so. Leaue her company? I am no gyrle I would you should well know, to bee taught what company I should keepe: I keepe none but honest company I warrant you. Leaue her company ketha? Alas poore soule, this reward she hath for her good will. I wis I wis, she is more your friend, then you are your owne. Well let her be what she will sayd her husband ... {The Pleasaunt History of... lack of Newberie 73) In the above text, Factors 1 and 3 are equally strong. The text is mostly in dialogue form, though a narrator is present explicating what is going on in the story. Direct questions, pronouns of the third, first and second person, interjections, assertations and repetitions are frequent in the above passage and elsewhere in the work. FICTION: HARMAN, Factor 1: 0.523 ( Factor 2: -0.656, Factor 3: 1.771) "What! your neuewes?" quoth she. "My neuewes?" quoth this parson; "I trowe thou art mad." "Nay, by god!" quoth this good wife, "as sober as you; for they tolde me faithfully that you were their vncle: but, in fayth, are you not so in dede? for, by my trouth, they are strau[n]gers to me. I neuer saw them before." "O, out vpon them!" quoth the parson; "thye be false theues, and this night thei compelled me to geue them al the money in my house." "Benedicite!" quoth this good wife, "and haue they so in dede? as I shall aunswere before god, one of them told me besides that you were godfather to him, and that he trusted to haue blessinge before he departed." "What! did he?" quoth this parson; "a halter blesse him for me!" "Me thinketh, by the masse, by your countenance you loked so wildly when you came in," quoth this good wife, "that somthing was amis." "I vse not to gest,"... (A Caveat or Warening for Commen Cursetors 40) Factor 3 is strongest in this text, but there are passages in which features of Factor 1 dominate. The above scene is in dialogue form and it provides a highlight of a story, used as an exemplum, with a strong emotional loading with exclamations, interactive questions and mild swearing.

Genre conventions: Personal affect

215

FICTION: ARMIN Factor 1: 0.446 (Factor 2: -0.731, Factor 3: 0.869) But they come; and up they came, and to the king they goe, who, being with the lord treasurer alone, merry, seeing them two, how Will had got another foole, knew there was sport at hand. How now! sayes the king, What news with you? O, Harry! sayes he, this is my owne uncle; bid him welcome. Wei, said the king, he is welcome. Harry, sayes hee, heare me tell thee a tale, and I will make thee rich, and my uncle shall be made rich by thee. Will tels the king how Terrils Frith was inclosed. Tirrels Frith! sayes the king; what is that? Why, the heath where I was borne, called by the name of Tirrels Frith: now a gentleman of that name takes it all in, and makes people beleeue it is all his, for it took the name from him; so that, Harry, the poore pine, and their cattle are all undone without thy help. And what should I doe? sayes the king. Marry, sayes Will, send to the Bishop of Hereford; hee is a great man with Terril: commaund him to set the Frith at liberty againe, who is now imprisoned by his means. And how shall I be rich by that? sayes the king. The poor will pray for thee, sayes Will, and thou shalt bee rich in heauen, for on earth thou art rich already. (A Nest of Ninnies 44) The basis of Armin's work is third person narration reflected by the high factor score of Factor 3, but direct quotations with questions, exclamations, and pragmatic particles are incorporated into the text, and first and second person pronouns alternate with the third person. These are features that fall on Factor 1. The result is a combination of different view points (see below, discussion of Factor 3). Some of the non-literary texts score high on this factor, Mowntayne's autobiography being the extreme example: AUTOBIOGRAPHY: MOWNTAYNE, Factor 1: 0.861 (Factor 2: -0.383, Factor 3: 0.303) "No, (sayed I,) as I ham borne to dye, contentyd I ham so to doo whan God wyll; but to be made awaye after sowche slyghte, I wolde be verye lothe; and therfor, yfe that yow have nothynge to showe for your dyscharge, acordynge as I have requyryd of yow, I tel yow trwe that I wyll not dye. Take yow good heed therfor to your selve, and loke that I myscary not, for yfe that awghte come unto me but good, yow and yours are lyke to knowe the pryse of yt, be yow well assur-

216

Irma Taavitsainert yd therof. Whan dyd yow ever see anye man put to deathe, before he was condemnyd to dye?" "That ys trwe, (sayd he;) and are yow not condemnyd?" "No, (sayd I,) that I ham not, nether was yet ever araynyd at anye sesyones." "Than, (sayed he,) I have been greatly myseynformyd. I crye yow marsy; for I hade thowghte that yow had been bothe araynyd, and also condemnyd to dye, beynge sent hether for to suffer yn thys plase, bycawse that yow were here agaynyste the quene with the ducke of Northethomeberland." "Well, (sayed I,) thoos materes hathe bene alredye suffysyently answeryd before your betteres; but I praye yow, syr, and a man myghte aske yow, whoos man are yow, or to whome doo yow belonge?" "Maiye! (sayd he,) I ham not ashamyd of my maister, I wolde thow showldest knowe yt, as thow arte". (Narratives of the Days of the Reformation 200-201)

The story is half in dialogue form, half in indirect narration. Some of the speech quotations have a very strong emotional loading, with first and second person pronouns, interjections, pragmatic particles and swearing. HANDBOOK: WALTON, Factor 1: 0.697; (Factor 2: 0.822, Factor 3: -0.391) (Ven.) Gentleman Huntsman, where found you this (Otter)? (Hunt.) Marry (Sir) we found her a mile from this place a fishing; she has this morning eaten the greatest part of this (Trout); she has only left thus much of it as you see, and was fishing for more; when we came we found her just at it: but we were here very early, we were here an hour before Sun-rise, and have given her no rest since we came; sure she will hardly escape all these dogs and men. I am to have the skin if we kill her. (Ven.) Why, Sir, what's the skin worth? (Hunt.) 'Tis worth ten shillings to make gloves; the gloves of an (Otter) are best fortification for your hands that can be thought on against wet weather. (Pise.) I pray, honest Huntsman, let me ask you a pleasant question, do you hunt a beast or a fish? (Hunt.) Sir, It is not in my power to resolve you, I leave it to be resolved by the Colledge of Carthusians... (The Compleat Angler 210-211)

Genre conventions: Personal affect

217

The other non-literary text that has a fairly high score on Factor 1 is Walton's handbook, written in dialogue form according to the conventions of instructive texts. It is worth noting that Factor 2 has an even higher score, in accordance with the instructive purpose of this text. DIARY: HOBY Factor 1: 0.010, (Factor 2: -1.550, Factor 3: -1.642) The positive score on Factor 1 obviously depends on the high frequency of the first person singular pronoun. See the example on p. 248.

6.3.2.

Factor 2 (eigenvalue 2.570)

"Reasoning"

Factor 2 is perhaps the easiest to interpret as its salient features, e.g. modal auxiliaries of possibility and necessity, and private verbs of mental activity and observation, mark the subjective reasoning process. This is in accordance with my earlier studies of the personal viewpoint in scientific texts in the 17th centuiy. The highest factor scores are found in philosophical and educational writing. 9 PHILOSOPHY: PRESTON Factor 2: 2.367, (Factor 1: -0.375, Factor 3: 0.102) (Ph) ... Let me ask thee, can that, dost thou think, which needeth nothing want Power? (Bo) No, I am not of that Opinion. (Ph) Thou thinkest right indeed; for if there be any thing which, upon any occasion of Performance, doth shew a Weakness or want of Power, it must, as to that, necessarily need foreign Aid. (Bo) So it is. (Ph) And therefore Sufficiency and Power are of one Nature. (Bo) So it truly seems. (Ph) And thinkest thou that things of this kind are to be undervalued and contemn'd, or rather to be reverenced of all? (Preston, Boethius 124) Boethius' De Consolatione philosophie is written in dialogue form according to the scholastic tradition. The language is different from Walton's especially in respect to Factor 1. The contrast to imitation of natural speech is clear.

218

Irma Taavitsainen

EDUCATION: BRINSLEY, Factor 2: 1.561; (Factor 1: -0.480, Factor 3: 0.221) Very reason must needs perswade every one of this. For, if they bee apt much before five yeeres of age, to learne shrewdnesse, and those things which are hurtfull, which they must bee taught to unlearne againe; why are they not as well fit to learne those things which are good and profitable for them, if they be entred and drawne on in such manner, as they may take a delight and finde a kinde of sport and play in the same? This delight may and ought to be in all their progresse, and most of all in their first entrance, to make them the better to love the schoole, and learning, as we shall see after. (Ludus Literarius, or the Grammar Schoole 9) This educational text scores high on Factor 2 only. It has high frequencies of modal auxiliaries expressing possibility and necessity, the pronoun it, and private verbs of mental activity and reasoning. Scientific writings and handbooks also score high on this factor, although the quality of involvement is different in them (see note 17). SCIENCE: BOYLE Factor 2: 1.303 (Factor 1: -0.616; Factor 3: -1.043) And this brings to my mind, that it has been observed, that Diamonds draw better whilst rough, than they do after they are cut and polish'd; which seeming to contradict what has been observed by others and by us also, that Amber, for instance, attracts more vigorously if the surface be made very smooth than otherwise, it induces me to conjecture, that, if this Observation about Diamonds be true, as some of my trials have now and then inclined me to think it, and if it do not in some cases considerably depend upon the loss of the (Electrical) Substance of the Stone, by its being cut and ground, the Reason may possibly be, that the great rapidness with which the Wheels that serve to cut and polish Diamonds must be mov'd, does excite a great degree of heat, (which the senses may easily discover) in the Stone, and by that and the strong concussion it makes of its parts, may force it to spend its effluviable matter, if I may so call it, so plentifully, that the Stone may be impoverish'd and perhaps also, on the account of some little change in its Texture, be rendred lesse disposed to emit those effluvia that are Instruments of Electrical Attraction. (Electricity & Magnetism 37-38)

Genre conventions: Personal affect

219

HANDBOOK: FITZHERBERT Factor 2: 1.282 (Factor 1: -0.442, Factor 3: 0.039) I coulde peraduenture shewe the housbandes dyuerse poyntes that the wyues deceyue them in: and in lyke maner, howe husbandes deceyue theyr wyues: but if I shulde do so, I shulde shewe mo subtyll poyntes of deceypt, than eyther of them knewe of before. And therfore me semeth beste to holde my peace, least I shoulde do as the knyght of the Tour dyd, the whiche had many fayre doughters, and of fatherly loue that he oughte to them, he made a boke, to a good entente, that they myghte eschewe and flee from vyces, and folowe vertues. In the whiche boke he shewed, that if they were wowed, moued, or styred by any man, after suche a maner as he there shewed, that they shulde withstande it. (The Book of Husbandry 98)

6.3.3.

Factor 3 (eigenvalue 2.499),

"Narration"

Factor 3 has the third person pronouns he/she as its most salient feature. THAT-deletion and emotive and imaginative verbs are also important. The third person is connected with narration, but in contrast to earlier studies, past and present tense were not important (see note 8). This, again, may be connected with the frequent use of first person narration with past tense verbs in scientific texts that have high factor scores on Factor 2. The highest scores on this factor are found in non-literary texts, though several fictional texts have high scores as well. FICTION: BEHN, Factor 3: 2.163; (Factor 1: -0.876, Factor 2: -0.219) 'Twill be imagin'd Oroonoko stay'd not long before he made his second visit; nor, considering his quality, not much longer before he told her, he ador'd her. I have often heard him say, that he admir'd by what strange inspiration he came to talk things so soft, and so passionate, who never knew love, nor was us'd to the conversation of women; but (to use his own words) he said, most happily, some new, and, till then, unknown power instructed his heart and tongue in the language of love, and at the same time, in favour of him, inspir'd Imoinda with a sense of his passion. She was touch'd with what he said, and return'd it all in such answers as went to his very heart, with a pleasure unknown before. Nor did he use those obligations ill, that love had done him, but turn'd all his happy moments to the best advantage; and as he knew no vice, his flame aim'd at nothing but

220

Irma Taavitsainen honour, if such a distinction may be made in love; and especially in that country, where men take to themselves as many as they can maintain. (Oroonoko 156)

Aphra Behn's text has a very high score on Factor 3. In contrast to other writings of the same genre, it represents long fiction, and a much later period. The story is told in the third person, with long monologues inserted into the narration. A narrator is present, intruding with first person comments as in the passage above, but such remarks are not very frequent as the negative score on Factor 1 indicates. FICTION: HARMAN, Factor 3: 1.791, (Factor 1: 0.523, Factor 2: -0.656) Α Roge is neither so stoute or hardy as the vpright man. Many of them will go fayntly and looke piteously when they see, either meete any person, hauing a kercher, as white as my shooes, tyed about their head, with a short staffe in their hand, haltinge, although they nede not, requiring almes of such as they meete, or to what house they shal com. But you may easely perceiue by their colour that thei cary both health and hipocrisie about them, wherby they get gaine, when others want that cannot fayne and dissemble. Others therebee that walke sturdely about the countrey, and faineth to seke a brother or kinsman of his, dwelling within som part of the shire; either that he hath a letter to deliuer to some honest housholder, dwelling out of an other Shyre, and will shewe you the same fayre sealed, with the superscription to the partye he speaketh of, because you shall not thinke him to runne idelly about the countrey; (A Caveat or Wareningfor Commen Cursetors 36-37) The basis of Harman's work is third person narration as in the above passage; see also p. 214 for another type of narration inserted in the work. FICTION: DELONEY Factor 3: 1.166; see the example on pp. 213-214. FICTION: ARMIN, Factor 3: 0.869, (Factor 1: 0.446, Factor 2: -0.731) Will Somers, in no little credit in the king's court, walking in the parke at Greenwich, fell asleepe on the stile that leads into the walk, and many that would haue gone that way so much loued him, that they were loth to disease him, but went another way; I, the better sort, for now adaies beggars are gallants, while gentiles of right blood seeme tame ruffians: but note the loue Will Sommers got. A poore

Genre conventions: Personal affect

221

woman, seeing him sleepe so dangerously, eyther to fal backward, or to hurt his head leaning so against a post, fetcht him a cushion and a rope; the one for his head, and the other to bind him to the post, from falling backward: and thus hee slept, and the woman stood by, attending as the groom of his chamber. (A Nest of Ninnies 42) Armin's text has positive scores on Factors 3 and 1. Third person narration forms the basis, but in places it alternates with direct speech and interactive patterns, with first and second person pronouns, direct questions, and mild swearing. The above example has an imperative form directed to the reader; controversal statements reflect the theme of carnivalism, i.e. reversal of ideas. FICTION: MERRY TALES, Factor 3: 0.803, (Factor 1: -0.244, Factor 2: -0.385) A woman ther was whych had had .iiii. husba~des. It fortunyd also that this fourth husband died & was brought to chirch vppon ye bere/ who- this woma- folowyd & made gret mone & wext very sory. In so mych that her neybours thought she wold sowne & dy for sorow/ wherfor one of her gossyps cam to her & spake to her in her ere & bad her for goddes sake to comfort her self & refrayne that lamentac~on or ellys it wold hurt her gretly & p~auenture put her in ieoperdy of her lyfe. To who- this woman answeryd & sayd I wys good gosyp I haue gret cause to morne if ye knew all/ for I haue byryed .iii. husbandys besyde thys man/ but I was neuer i~ the case y' I am now/ for there was for there was not one of the- but whe~ that I folowid the corse to chyrch yet I was sure alway of an other husba~d before that ye corse cam out of my house/ & now I am sure of no nother husband & therfore ye may be sure I haue gret cause to be sad and heuy. (A Hundred Mery Talys 19-20) Merry Tales contains third person narration with THAT-deletion and imaginative and emotive verbs. Direct quotations are given at the turning points of the plot and other highlights of the story; yet the score on Factor 1 is negative, which marks the absence or rarity of features like exclamations, swearing, pragmatic particles, and first and second person pronouns. HISTORY: BURNET, Factor 3: 2.828; (Factor 1: -1.231, Factor 2: -8.19) He thought that nobody served him out of love: and so he was quits with all the world, and loved others as little as he thought they loved

222

Irma Taavitsainen him. He hated business, and could not be easily brought to mind any: but when it was necessary, and he was set to it, he would stay as long as his ministers had work for him. The ruin of his reign, and of all his affairs, was occasioned chiefly by his delivering himself up at his first coming over to a mad range of pleasure. (Burnet's History of My Own Time 168)

BIOGRAPHY: BURNET, Factor 3: 2.187; (Factor 1: -1.121, Factor 2: -0.439) He loved to talk and write of Speculative Matters, and did it with so fine a thread, that even those who hated the Subjects that his Fancy ran upon, yet could not but be charmed with his way of treating them. Boileau among the French, and Cowley among the English Wits, were those he admired most. Sometimes other mens thoughts mixed with his Composures, but that flowed rather from the Impressions they made on him when he read them, by which they came to return upon him as his own thoughts; than that he servilely copied from any. For few men ever had a bolder flight of fancy, more steddily governed by Judgment than he had. No wonder a young man so made, and so improved was very acceptable in a Court. (The Life and Death of John Earl of Rochester 8) Two non-literary texts by the same author have the highest scores on Factor 3. They represent the extreme ends on the positive side on this factor and on the negative side on Factor 1 reflecting the absence of interactive features. Private verbs of emotion and imagination are prominent, as are third person pronouns. AUTOBIOGRAPHY: FORMAN, Factor 3: 1.791 (Factor 1 -0.853; Factor 2 -0.740) Ther was a man of good reputation and wealth, that dwelte not far from Simon's master, that had a proper fine maiden to his only daughter, the which being but yonge of yeares and younger then Simon, that loved Simon wonderfull welle, and wold suerly see him once a daie, or ells she wold be sicke. And often she wold com to Simon's master, and entreat him very kindly on holly daies that she might see him or speake with him, and somtymes to goe to pastymes with her; and she loved him soe well that yf forty youth were at play before the dore, in a spacious place which ther was, yf Simon were not amonge them, she would not be ther; but yf he were there, none

Genre conventions: Personal affect

223

could kepe her from thence. Yf Simon stode by his master or mistress at the dore she would com and stand by him, and wold not goe from him tille necessity did comple. And Simon's master well perceyvinge the grete affection of the gentlewoman towardes Simon, wold often say unto her, mistress An Yene love my boy welle methinkes; and she would aunswer, yea forsothe, yf yt will please youe to give him leave to go rone with us, wee shall give youe thankes, sir. Wherupon oftentymes he wold give him leave; and as for Simon, he loved her not but in kindnes, but because she was soe kind to Simon, he wold doe anythinge he could doe for her. And his love on her syde lasted longe, as herafter shalbe showed. (The Autobiography and Personal Diary of Dr. Simon Forman 9) Forman's autobiography and Perrott's biography score high on this factor; both have negative scores on the two other factors in spite of occasional exclamations, first person passages and direct questions. Negative scores, however, tell that they are very rare (see 6.2). BIOGRAPHY: PERROTT Factor 3: 1.249 (Factor 1: -0.576, Factor 2: -0.444) Shortly, after Sir John Perrott returned from France, and came to the Court of England, where he lived at great Charge, and at soe high a Rate, as he grew into great Dept, and ranne so farre into Arrearages, that he began to mortgage some of his Lands, and yet did owe some seven or eight Thowsand Pounds, being like to Allexander the Great in this, who agaynst his Expedition into Persia did put most Part of his Possessions belonging to the Crowne of Macedon in Pawne. And being asked by Perdica, his chiefe Commander, what he left behynd hym, answered Hope. So this Knight spending his Patrimony (as many of his Yeres and Calling do now-a-dayes, wastfully, and above their Habiliitie) had lefte but a bare hope to recover his Estate.Yet he at Length did begin to bethink hymself, and to look back into his decaying Fortune, and soe grew much agrieved at his owne Prodigality. Insomuch that on a Time he walked out of the Court, into a Place where commonly the Kinge did use to come about that Howre; and there he began (either as knowing that the Kinge would come that Way, or els by Chaunce,) to complayne as it were agaynst hymself, unto hymself: How unfortunate and unwise he was, soe to consume his Livinge, having wasted a great Part of that in few Yeares, which his Auncestors had gotten and enjoyed for many Yeres? And must I

224

Irma

Taavitsainen

(quoth he) be the man that shall overthrow my Howse, which hath continued soe longe? It had byn better I had never byn born. (The History of That Most Eminent Statesman, Sir John Perrot 31-32) The above passages of texts with analyses of their linguistic features aim at highlighting the complicated texture of these early works. No simple answer to the question posed at the beginning, whether there are any salient features that tell various text types apart and reveal relevant genre conventions, emerges. The quality of personal affect in these texts also needs further investigation.

6.4. Textual affinities Table 3. Factor scores of genres, with means (M), standard deviations (D) and frequencies (F), i.e. numbers of texts. Category Autobiography

Μ D

Factor 1

Factor 2

Factor 3

0.87573 3

-0.04595 0.86119 3

-0.64288 0.22790 3

0.80068

Biography

Μ D F

1.22783 0.96999 3

-0.70997 0.36263 3

-0.28559 0.26958 3

Philosophy

Μ D F

0.21783 0.35229 3

-0.29573 0.07251 3

2.19818 0.22169 3

Diary

Μ D F

-0.79672 0.55666

-0.28022

-0.97870 0.47631 6

Μ D F

-0.14267

Education

6

0.60611 6

0.17279 6

-0.49266 0.17382 6

0.72499 0.66015 6

Genre conventions: Personal affect

225

Table 3. Continued. Category

Factor 1

Factor 2

Factor 3

Fiction

Μ D F

1.21268 0.63329 6

0.68433 1.37172 6

-0.32160 0.45020 6

Handbook

Μ D F

-0.12693 0.96102 6

-0.09960 0.44740 6

0.76536 0.47233 6

History

Μ D F

0.54932 1.23272 6

-0.69949 0.27179 6

-0.66273 0.43532 6

Play

Μ D F

0.03489 0.27114 6

2.16561 0.18096 6

0.06882 0.26879 6

Science

Μ D F

-1.06343 0.22794 6

-0.41997 0.34377 6

0.64697 0.40757 6

Travel

Μ D F

-0.82784 0.53415 6

-0.33217 0.12920 6

-0.87796 0.34024 6

Total

Μ D F

0.00000 0.99647 57

-0.00000 0.97327 57

-0.00000 0.91928 57

The next step is to assess the means and standard deviations of the genres. In the structure of the corpus, the genres are represented by two texts in each subperiod, which means a great deal of generalisation at this level. The samples were often deliberately selected to stand far apart from one another, and thus the standard deviations are important as indicators of the possible heterogeneity of texts grouped together on external evidence. On the other hand, the representatives of some genres seem to be very close to one another in respect to their personal affect features. 10

226

Irma Taavitsainen

For the purpose of this study, textual affinities revealed by the three factors are of great importance (see Table 3 above). If the ultimate aim of genre studies is to probe into the essential differences between adjoining, overlapping genres, the above description of features that correlate and texts in which they occur serves to indicate the groupings within which the differences should be found. The overlappings are clear. Factors 1 and 3 are closely interwoven in some texts: while Mowntayne's autobiography scores high on Factor 1, some fictional works are best described by Factor 3. High scores on both factors are not uncommon, as the above examples show. Factor 2 is different as it describes philosophical, educational and scientific texts, as well as handbooks; only Walton's handbook reached a high factor score on Factor 1 because of the more colloquial features of the dialogue. In the following assessment I shall concentrate on fiction and the adjoining genres indicated by Factors 1 and 3. It is obvious that the parameter values ascribed to plays and some other texts (INTERACTIVE, INFORMAL, see 4.1 and 4.2) show up well in the above assessment." The generic conventions of plays are very different from those of the other genres, and therefore they are excluded from the second phase of my search for distinctive discursive properties in fiction and adjoining genres. They are completely in dialogue form, in imitation of spoken language, and the conventions must necessarily be different. The narration of events takes place through direct speech quotations, recreated on stage each time the play is performed. The situations that form the context of linguistic behavior are acted, and thus the reference system must be different from that of purely written texts. (See Figures 3 a, b, c. They indicate how far apart plays stand from the other genres in focus.)

7. Combining quantitative and qualitative approaches In this phase of the study the material was limited to comprise fiction and the genres that stand nearest to it: history, biography, autobiography, diary and travelogue. Factor analysis proved the close affinity of these genres, defined by external evidence, in respect to the linguistic features that reflect personal affect. The pattern revealed by factor analysis is, however, intricate and requires further consideration (see 6.3).

Genre conventions: Personal affect

Figure 3a: 3 factor solution combined variables F2 BOETH

]

HA ND

«

I BIO

FICT A JTO

HIST

TRAV DIAR

-

1

0

1

2

227

228

Irma

Taavitsainen

Figure 3b: 3 factor solution combined variables FL

PLAY

HCT

AU' ι υ

HA *JD BOETH SC :IE HIST

-1.5

-0.5

BIO

0.5

1.5 F3

Genre conventions: Personal affect

Figure 3c: 3 factor solution combined variables F2

BOETK

m

SC :IE

ίΡ

PLAY

mr HIST

AU' ro

TRAV DIAR

-1.5

-1

-0.5

0

0.5

1

1.5

229

230

Irma

Taavitsainen

The genres selected for scrutiny are all relevant to the evolution of early fiction. The difference between them and fiction may be difficult to pinpoint, but it may be attempted via the features indicated by Factors 1 and 3, though the distinction does not seem to be striking on the basis of factor scores only. Fiction may have its own devices in story telling. Questions of deixis and interactive focus may grow in importance in adjoining text types that by definition employ first person pronouns, and there may be subtle differences not captured by factor analysis. The categories may be fuzzy, and it may prove difficult or impossible to tell these non-fiction genres apart from fictional accounts with similar contents. The questions to be asked are numerous. Where are the differences, then? Are there any that separate the adjoining genres? How do non-literary texts of adjoining genres stand in relation to fiction? Are there any differences in respect to the interactive focus or narrative concerns? Can more subtle shades of personal affect be discerned? Is there any difference in the intensity, emphasis, or quality of affect? Does the moralising function of early fiction show in this analysis? And finally, can this analysis reveal mechanisms of reader involvement in these texts? According to earlier considerations, the ultimate aim of genre theory is to define the institutionalised discursive properties of various genres, which can be interpreted as distinctive features that form a matrix and contribute to genre conventions in a longer time span. If a significant difference can be verified in the occurrence of a feature in fiction and text types that stand near it, this feature must be worth studying in more detail. It may illuminate the special character and evolution of these text types.

7.1. Imaginative versus non-imaginative narration For a sharper focus, the material was divided into two groups along the corpus parameter of imaginative versus non-imaginative narration (NI versus NN) which divides fiction and the adjoining non-fictional genres into groups of their own. 12 This parameter is built into the corpus, and this is the first time to my knowledge that this parameter in the HC has been systematically applied in research. The linguistic features of personal affect showed a great deal of variation and possible matrices were indicated by factor analysis, and in the next phase of the study the significance of the individual features included in these matrices was tested along this parameter. The aim was to find possible text-type markers. T-tests were performed in order to find out whether the difference between fiction and

Genre conventions: Personal affect

231

non-fiction was significant in respect to the same variables that were applied in the factor analysis. The results of the t-tests of features discussed below are as follows: Table 4. Results of t-tests. Ν of observations: NN 24, NI 6

Probability

Means NN

NI

St. d. NN

NI

Exclamations Swearing THAT-deletion Second person Modals, volition possibility Deictic features NOW

0.046 0.028 0.019 0.041 0.040 0.030

0.156 0.046 1.24 1.73 3.71 3.52

3.98 1.24 2.84 8.88 9.91 5.33

0.30 0.15 1.55 3.37 3.14 2.13

3.54 0.95 1.16 6.15 5.27 1.40

0.031

1.43

2.69

0.98

1.03

The above table gives the results of t-tests of features with the greatest factor loadings on Factor 1 and Factor 3; only features that proved significant are given. 13 The texts of early fiction with the parameter value NI ("Narration Imaginative") formed one group, and texts in the adjoining groups with the parameter value NN ("Narration Non-imaginative") the other. In addition, an analysis of variance was done with the influence of the period and the parameter NI versus NN as dependent variables. In general the results confirmed those of the t-tests; in some cases the analysis of variance added to the earlier information, as the difference between fiction and non-fiction, given as insignificant by t-test, proved significant according to the analysis of variance. For example, of the tested features pragmatic particles (probability < 0.0028), direct questions (p < 0.0005) and this (p < 0.0043), and the second person singular (p < 0.0017) proved significant as indicators of the difference between imaginative and nonimaginative narration when the influence of period was eliminated. According to the analysis of variance, exclamations and swearing were the most outstanding text-type marker of fiction (p < 0.0001). These statistical tests were used to complement the findings of factor analysis. They indicate that there are significant differences between the occurrences of some linguistic features in imaginative and non-imaginative narration, fiction and the adjoining genres. Some of the features belong to

232

Irma

Taavitsainen

the spoken versus written dimension of language, some reflect interaction, and some are connected with the reference system. As one of the aims of this paper is to define the quality of personal affect in these texts, I have grouped these features according to my interpretation, in view of their relevance to making a difference between imaginative narration and nonimaginative narration, fiction and related non-literary text types. The aim is to chart what this difference really consists of.

8. "Surge" features of personal affect

According to my pilot studies personal affect is an umbrella term for the features among which the markers of the fictional mode can be sought. One of the main dimensions is the surge of personal feelings and attitudes. Reflections of the speaker's or writer's mental state have linguistic outlets, encoded in the texts, and they seem to be of special relevance to the present purpose.

8.1. Exclamations, swearing, and pragmatic particles Table 5. Results of t-tests and f-tests. Ν of observations: NN 24, NI 6 t-tests Exclamations Swearing f-tests Exclamations Swearing Pragmatic particles

Probability

0.046 0.028

Means NN

NI

St. d. NN

0.156 0.046

3.98 1.24

0.30 0.150.9

NI

3.54

0.0001 0.0001 0.028

In modern linguistics interjections are described as linguistic gestures which express a speaker's mental state, action or stance, or reaction to a situation. Thus they encode speaker attitudes and communicative intentions and are context-bound (see Ameka 1992). In spoken language their

Genre conventions: Personal affect

233

interpretation often depends on paralinguistic features like intonation and accompanying gestures. In written genres, the linguistic context is the only clue to their interpretation. Interjections range from onomatopoetic ad hoc formations to single words, utterances and set collocations; such secondary interjections merge with swear words and phrases, oaths, and also with greeting formulas, expressions of politeness, and other routines. Sometimes interjections have been included in the class of particles, because they are uninflected. The borderlines between various categories are fuzzy and difficult to draw. Because of the close affinity of secondary interjections and pragmatic particles, these groups are discussed together with interjections and exclamations.

8.1.1. Occurrences in Early Modern English In my statistical analysis, the difference between the occurrence of interjections in fiction and in the adjoining genres proved highly significant. The use of exclamations and swearing seems to be the most powerful texttype marker of fiction. In quantitative calculations the frequencies are given per one thousand words of text, but at this point it may be useful to return to the absolute occurrences: Table 6. (Absolute) occurrences of exclamations. Exclamations

Total

EModEl

EModE2

Fiction Autobiography History Biography Diary Travelogue

94 7 2 1 1 0

19 6 2 0 0 0

27 0 0 0 1 0

EModE3 48 1 0 1 0 0

The distribution is very uneven across the genres; of the excluded genres plays had even higher concentrations of exclamations (see Taavitsainen 1995a).

234 8.1.2.

Irma Taavitsainen Conventions English

in a longer diachronic perspective:

Late Middle

Inspired by the results of this study, I extended my assessment of interjections and swear words to the preceding period (1350-1500) with the help of the Late Middle English part of the HC, and the Chaucer Concordance (for details, see Taavitsainen forthcoming a and b). The aim was to see how far back in time the convention could be traced, and whether the distribution of these linguistic features showed a similar pattern earlier. The difference between fiction and non-fiction proved highly significant in the Late Middle English material as well.14 Interjections were particularly frequent in Chaucer's works, and the example he set must have inspired authors after him. The stylistic quality of interjections ranges from solemn invocations that only occur in some genres to burlesque exaggerations of brisk action with collocations of interjections, equally restricted in scope. In some cases Chaucer modifies and colours the meanings of these words with shades of irony, which use is not found in any other writer's work in this period. An emotive loading is often present in speech quotations and passages that employ interjections. The audience is supposed to live with the characters and their emotions. Interjections may also have a conative function, directing text internally to another character, or the reader may be requested to pay special attention; this is specially frequent with the originally emphatic particle lo. In addition to the element of personal affect, interjections have textual functions. In several genres they are used to organise discourse and mark turn-taking in the perfomance of the text, aiding the hearer to distinguish various speakers' turns when read aloud. This is especially clear in drama and romances, in which the use of interjections at the beginning of each turn is almost regular. In Chaucer's works interjections have acquired other textual functions as well. Collocations of exclamations are used to signal important turning-points in the plot; the patterns are genre-specific (see Taavitsainen 1995b). An extreme development can be seen in Chaucer's use of interjections with a special future meaning, as a foregrounding device. In these passages interjections are embedded in the narration to create apprehension in the audience by ominous exclamations like alas and lo.ls

Genre conventions: Personal affect 8.1.3.

235

The art of story-telling

Nearly all interjections in the Late Middle English and the Early Modern English period are found in direct speech quotations, frequently in sentence-initial position. When compared with the Late Middle English functions of interjections, the marking of turn-taking is not evident in the Early Modern material. This issue is connected with several important points concerning the consumption of these texts. With the spread of literacy reading habits changed, and with cheaper production methods and materials, books and booklets became more readily available for a wider audience. The change was gradual, and reflected the transition from medieval to modern (cf. 2.5). The instances of interjections in the present material are especially numerous in Merry Tales (EModEl) and Penny Merriments (EModE3) which are both collections of short funny stories. They continue the tradition of medieval fabliaux, which can also be seen in the use of interjections and swear words. Medieval narration in verse has been transformed to fit the prose diction in these stories (see below). According to the fabliau tradition, the characters represent the lower or middle layers of society and the speech quotations have the pretence of natural talk with colloquial idioms, oaths and swear words. Some of the phrases obviously belonged to the stock-in-trade of the genre. Alas occurs frequently in emotive vocatives with adjectives of sympathy like poor and good, and in passages of regret and lamentation as the very stereotype of such feelings. Chaucer's art of story-telling is reflected in the use of interjections in marking turning-points in the plot. This tradition is carried over to Early Modern short fiction. Alas is used in this way to highlight important passages and lend them vividness and emotional colouring: sayd to ye maltman alas I haue let my boget fal in to ye water & there is .xl. li. of money therin yf thou wylt wade in to ye water & go seke it & get it me agayne I shall gyue ye .xii. pence for thy labour ... This maltman within a whyle after with grete payne & depe wadynge founde ye boget & came out of the water & sawe not his felowe there & sawe that his clothys & money were not there as he left them suspectyd ye mater and openyd the boget and than founde nothynge therin but stonys cryed out lyke a mad man and ran all nak-

236

Irma Taavitsainen yd to london agayne and sayde alas alas helpe or I shall be stolen. For my capons be stolen. My hors is stolen. My money and clothys be stolen and I shall be stolen myself. And so ran aboute the stretys in london nakyd and mad cryenge alway I shall be stole. I shall be stolen ... (A Hundred Mery Talys 149)

The collocation fye, for shame, fye provides the climax of one of the short comic anecdotes in Harman. It is a watch word and when uttered at the wrong moment it provides the jest of the story around which the narration is built. The choice of a phrase like this to mark the culmination of a comic tale is in accordance with the conventions derived from medieval fabliaux and Chaucer's use of cumulative lists of interjections at the turning-points of such stories (see Taavitsainen 1995b). geue me a watche worde a loud when hee goeth aboute to haue his pleasure of the, and that shall be "fye, for shame, fye" and I wyll bee harde by you wyth helpe ... "And are you not ashamed? neuer a whyte," sayth he, "lye downe quickely." "Now, fye, for shame, fye," sayth shee a loude, whyche was the watche word. At the which word, these fyue furious, sturdy, muffeled gossypes flynges oute, and takes sure holde of this betrayed parson, sone pluckinge his hosen downe lower, ... (Harman, A Caveat or Warening for Commen Cursetors 72) In Armin's work the use of interjections is somewhat more elaborate than in the above mentioned texts. The passage with the most marked use takes an outsider's angle with third person pronouns, and yet it uses Ol several times as direct speech quotations to shift the angle; ah is also used in this way. This conflict of viewpoints is effective and it does not occur in other texts of the present material. Exclamations, together with short questions and quick changes of turns in the fictional conversation create a vivid narration with an intensive emotional tone, perhaps in imitation of pangs of pain in the first passage, and anxiety in the second. Thus there is a link to the use of interjections in passages that provide the climax of the story: The pyper and the minstrel, being in bed together, one cryed O! his backe and face; the other, Ol his face and eye: the one cryed Ol his pype! the other, Ol his fiddle! Good mussicke or broken consorts, they agree well together; (Armin, A Nest of Ninnies 11-12)

Genre conventions: Personal affect

237

Jack fed, and feeding greedily, (more to anger the cooke, than disappoint Sir William) ever as he burnt his mouth with hast, dipt the pie in the water to coole it. Ol says the cooke, it is Sir William's owne pie, sirra. Ol sayes Jack hang thee and Sir Willy too. I care not, it is mine now. Save Sir William some, sayes one; save my lady some, sayes another. By James, not a bit, sayes Jack. (Armin, A Nest of Ninnies 14) The function of foregrounding, found in Chaucer's works,16 is not obvious in the present material. The nearest usage can be found in Aphra Behn's text, which is different from the other fictional works here. It is a long narration that can be considered a forerunner of the romantic mode of story-telling. The use of interjections seems to be consciously targeted at manipulating readers' emotions, as these exclamations are embedded in long monologues or in the middle of narration. The interjection provides a side remark to provoke the readers' emotions and contributes to the overall affective tone of the text (see Taavitsainen forthcoming d).17 At this character, his old heart, like an extinguish'd brand, most apt to take fire, felt new sparks of love, and began to kindle; and now grown to his second childhood, long'd with impatience to behold this gay thing, with whom, alas! he could but innocently play. But how he shou'd be confirm'd she was this wonder, before he us'd his power to call her to court, (where maidens never came, unless the king's private use) he was next to consider ... (Behn, Oroonoko 157) He cou'd not be convinc'd he had no cause to sigh and mourn for the loss of a mistress, he cou'd not with all his strength and courage retrieve. And he would often cry, Oh, my friends! were she in wall'd cities, or confin'd from me in fortifications of the greatest strength; did inchantments or monsters detain her from me; I wou'd venture through any hazard to free her: But here, in the arms of a feeble old man, my youth, my violent love, my trade in arms, and all my vast desire of glory, avail me of nothing. Imoinda is as irrecoverably lost to me, as if she were snatch'd by the cold arms of death: Oh! she is never to be retrieved. If I wou'd wait tedious years, till fate shou'd bow the old king to is grave, even that wou'd not leave me Imoinda free (Behn, Oroonoko 160)

238

Irma Taavitsainen

8.1.4.

Non-literary text types

Although there is a significant difference between the frequencies of interjections, swear words and pragmatic particles in imaginative and nonimaginative narration, examples are found in other genres as well. Mowntayne's autobiography has several occurrences. They are all found in direct speech quotations. "Ye", sayed he. "And ys thys the beeste servys that yow can doo my lorde your master? Fye, for shame, fyel wyl you folowe now the bludye stepes of that wyckyd man your master! who ys unworthye, before God I speake yt, bothe of the name and place that he hathe and ys calyd unto. What sholde moufe yow for to handyll me after thys sharpe sorte as yow have done, so spytefiillye, beynge here not yet iij dayes under your kepyng? Wyl yow become a tormentor of Godys people and prophetes? wyl yow now seas from kyllynge of bolokes, calvys, and shepe, which ys your ockapasyon (being a bucher), and to gyve over your seife moste crwellye to sarve your mastares tourne in sheddynge of Ynnosente blode? Ο man, with what an avaye (heavy) harte maye yow laye your selve down to slepe at nyghte, yf that God of hys great marsy doo suffer yow to lyve so long yn thys your so wycked atempte and enterpryse! (Mowntayne, Narratives of the Days of the Reformation 201) After execucyon of which dedely dede ye sayd syr Piers toke great repentaunce Inso moche that lamentably he sayd alas what haue we done we haue now put to deth hym that hath ben our Soueraygne and drad lorde by the space of ,xxii. yeres by reason whereof I shall be reprochyd of all honoure (Fabyan, The New Chronicles of England and France 170R.C2) Now truly, quoth an old gentilman to a young feloe, ye ar far to blame to mislyke your aunt for she may do you pleasure and I wold God I had such an aunt. Fy, quoth he, wold I had your land on condicion you had xxty such aunts. (Madox, The Diary of Richard Madox 88) Even the texts with the lowest negative scores may contain occasional examples of interjections:

Genre conventions: Personal affect

239

Soon after I told him, I was glad to find his Style so reformed, and that he had so entirely overcome that ill habit of Swearing; Only that word of calling any "damned", which had returned upon him, was not decent. His Answer was: (Oh that language of Fiends, which was so familiar to me, hangs yet about me: Sure none has deserved more to be damned than I have done.) (Burnet, The Life and Death of John Earl of Rochester 154) In these examples the interjections are reflexions of spoken language, spontaneous reactions to the immediate situation.

8.2. THOU and YE; sociolinguistic considerations According to the statistical assessment, the difference between the use of the second person singular in fiction and the adjoining genres is significant. The absolute occurrences of the second person singular are given below: Table 7. (Absolute) occurrences of second person singular. Total Fiction Autobiography History Biography Diary Travelogue

102 23 2 1 3 0

EModEl 20 9 0 0 0 0

EModE2 29 4 0 1 0 0

EModE3 53 10 2 0 3 0

The historically plural pronoun (YE; oblique form YOU) for polite address to a single person had been in use from the Early Middle English period. By c. 1600, the plural form (with YE/YOU as nominative form) had become the unmarked form of address among the upper classes, and the second person singular was restricted to expressions of "power and solidarity" as well as heightened emotionality (Mühlhäusler—Harre' 1991: 152; for a summary of uses, see Calvo 1992). A classical example of the use

240

Irma Taavitsainen

of this pronoun in trials is the well-known adjuration to Raleigh: "I 'thou' thee thou traitor". Vacillation of the pronouns is common in Shakespeare's plays, and various functions have been attributed to these shifts. In the seventeenth century the use of thou gradually became restricted to religious language (Mühlhäusler—Harre' 1991: 153). A detailed study of the legend of St Katherine in Southwell Minster MS 7, written about 1500, gives new evidence of the state of the process at the very beginning of the Early Modern period (Nevanlinna—Taavitsainen 1993). It shows that the use of the second person pronouns was in a state of flux, as expected, but the emotional use was so well established that a change of pronoun could be employed as a deliberate device in marking an emotionally heightened passage and a change in attitude, from respect to contempt in a saint's legend; similar examples are well known from Shakespeare's works. In some other passages the use of the polite plural pronoun implies respect and emotional distance, but with growing familiarity it changes to thou. The use seems to be well under control in most examples, at least where the main characters are concerned. The second person singular was one of the surge features of personal affect available in the language in this period, and thus worth a detailed assessment in the present connection.18

8.2.1.

Unmarked and special uses in fiction

The second person singular thou occurs frequently in fiction. The explanation of the high frequency can be found in the social classes that these texts depict, as the second person singular was the normal pronoun of address among country folk and lower classes (Wales 1983: 116). It is evident that most instances are unmarked as they occur in quotations of speech among social equals of these classes. Thus it would be exaggerated to attribute affective meanings to the pronouns in most cases, for example in the first example below, but the second passage certainly contains an emotional colouring which is made explicit by other features contributing in the same direction: This poller then sayd to hym go thy way streyght to thend of yt long entre & there thou shalt se whether it be he or no & I wyl holde thy horse here tyll thou come agayn. This maltman thynkyng to fynde ... (A Hundred Mery Talys 147)

Genre conventions: Personal affect

241

(Tweedle) I drinke to thee with all my hart: why thou whoreson when wilt thou be maried? Ο that I were a young wench for thy sake: but tis no mater though I be but a poore woman... (Deloney, The Pleasaunt History of... lack of Newberie 79) Alongside the normal, unmarked use of thou, special meanings can be attributed to the use of this pronoun in fiction. In Armin's work, the fool addresses Sir William with thou·, the king addresses the fool with ye in return. This is well in accordance with the reversal of values in his work. Sir William demaunded why hee eate the pye? Because I had a stomacke, sayes Jack. Would nought else serve, sayes the knight, but my pye? No, Willy, sayes he, thou would not be angry then, and the cooke had not been turned away: but all is well — thou art rich enough to buy more. (Armin, A Nest of Ninnies 15) This lusty jester, forgetting himself, in fury draws his dagger, and begings to protest, nay; saies the king, are ye so hote? (Armin, A Nest of Ninnies 48) In EModE3 all examples in fiction are found in Penny Merriments (1685); Aphra Behn's text contains third person narration; you occurs only six times in direct quotations of fictional conversation. The high frequency of the singular form in EModE3 in Penny Merryments occurs in a dialogue between a Quaker and his maid (34 out of 53 instances of thou)·, its use in marking Quaker affiliations is well-known (Mühlhäusler—Harre 1991: 153). The text in Penny Merriments is a parody with exaggerated emotionality and comical imitation of religious language. The use of thou is in accordance with the special usage among Quakers, but there are other surge features of personal affect, and the mocking tone is obvious: (Quaker) I What a War is there even now, betwixt the Inward and the Outward Man! Satan, Satan, I say unto thee, avoid, by Yea and by Nay, I charge thee tempt me not: Oh! how the Outward Man prevails! and I can hold no longer; nay, the Light within does say unto me, That Mary is a Sister, and that Gods Lambs may play, so that they can but keep it secret from the Wicked; therefore Satan, though I defie thee and all thy Works, yet will I go unto Mary as I have said: Mary, why Mary, I say unto thee Mary.

242

Irma Taavitsainen (Qua) Then Mary, I plainly say unto thee, sit thee down, by yea and nay I must Touze thee, ingeniously I must. (Ma) I fie, Master, fie; what is't ye do? the Saints ought not to defile each other, we shall lose our Credit among the Prophaned; nay, Master, why Master, Ο fie! wherefore is it you Kiss me so? Ο if my Dame should know on't! (Qua) I say unto thee, fear not, fear not I say, thou art a Sanctified Sister, and one of the Infallible Congregation; and as for thy Dame, I say she is departed; therefore Mary, again I say unto thee, that the Spirit within does move me to refresh thee; I burn, I fry, and can forbear no longer. (Mary) Oh! Master, Master, I adjure thee, that thou forbear, nay, Master, Master, Ο Master! Ο fie! take away thy hand, what is't thee dost? I say unto thee, nay, nay, I say unto thee nay; Ο let me alone, why dost thee tempt me to go astray like one of the Wicked? (Penny Merriments 147-148)

The other passage in Penny Merriments, "Loving Kate", is equally affective, with a dialogue of high emotional charge between the participants (see the example above). There are several features that contribute to the overall tone, exclamations for instance. Thou is regularly used by John when addressing Kate, but the pronoun vacillates in Kate's answers. Intimacy and affection are strong qualities in this text, but it is difficult to know whether the vacillation reflects the change of the unmarked use among the social class in question, or heightened emotionality. (Ka.) You need not doubt that, for he cannot keep it from me, and five pound I have gathered since I came to service, besides my Mistress owes me above half a years wages. (Jo.) Ο what a happy man shall I be, what a good housewife thou hast been, thou hast good cloathes too ... (Penny Merriments 116) (Ka.) Must I dance too. (Jo.) Ay pretty one, every body will strive to dance with the Bride. (Ka.) Ide rather dance with thee John than with them all. (Jo.) So thou shalt my dear. (Penny Merriments 118)

Genre conventions: Personal affect

243

8.2.2. Surge and social hierarchy in non-fiction In non-fictional genres the characters belong to higher social classes. Autobiography employs thou frequently. Mowntayne's situation in his autobiography was much like Raleigh's, and the use of thou is in accordance with it. The first part of the passage below contains polite, civilised discussion, but personal affect is certainly present in Mr. Gryffyn's speech: "Syr, what make yow here? are you not a Londynar?" "Yes, and yt lyke your lordshyp." "Howe longe have yow be here prysonar?" "Halve a yeare, my lorde." "Who sent yow hether?" "Forsothe, my lorde, that dyd the counsel." Than sayd the hye shyryffe, "My lorde. thys ys the man that I tolde your lordshyp of; I beseeche yow be good lord unto hyme, for he hathe bene as quyete a prysonar as ever came within thys gayell, and hathe usyd hymselve as honestly toward hys keapar." "Yow speake wel for hym," sayd my lorde; "stand asyed a whyell tyl yow be called." Yn the meane tyme mr. Gryffyn had a caste at me, sayenge thus: "Thou arte bothe a traytor and a herytyke" "No, and yt lyke your worshup, I ham nother of bothe." "Ys not thy name Mowntayne?" "Yes, forsothe, I wyll never deny yt." "And art not thow he that my lorde chansler sent hether with a wiyte?" "I am the same man." "Wel! (sayed he,) and thow be not hangyd I have marvell. Thow wylte scape narrowly, I beleve." (Mowntayne, Narratives of the Days of the Reformation 206-207) In other autobiographies it also occurs in emotionally coloured quotations. It is particularly frequent in the journal of George Fox from the latter half of the 17th century. This deals with the same topic, imprisonment and justice, which are highly affective themes: Thats a snare said I: which all people may take notice of: for I ought to bee sett free from ye goaler: & this Courte: if I am a free man: as thou says I ought to bee: & yett thou tendrest ye oath before I am at liberty but thou ought to lett mee bee at liberty & then thou mightst have donne thy will. But hee cryed give him ye booke (Fox, The Journal of George Fox 82)

244

Irma Taavitsainen

In Forman's autobiography and Perrott's biography from EModE2 the pronoun of address is in accordance with the social hierarchy and power relations in addressing social inferiors. Both examples contain pious wishes and reference to suffering, and are thus connected with the very personal, even emotive level: And his master wold say to him, Simon, thou moste suffer as well as I myselfe; thou seeste we cannote remadie yt as yet, but God will send a remadie on daie. No we because Simon's master had beaten him for his mistris sake, herself being in faulte, ... (Forman, The Autobiography and Personal Diary of Dr. Simon Forman 10) To whom he sayd, Well Boy, Blessing. I wold to God thou saffe; then I should care the History of That Most Eminent

God blesse thee, and I give thee my wert a shore, and the Queenes Ship lesse for my seife (Perrott (?), The Statesman, Sir John Perrot 114)

The other examples from EModE3 are found in Evelyn's diary and Milton's history writing. All examples in Evelyn's text are from a religious context, a prayer, which use conforms to the regular pattern at the time. Milton's history writing provides a good example of the pathetic use of this pronoun in an emotionally toned passage: He caus'd his Royal Seat to be set on the shoar, while the Tide was coming in; and with all the state that Royalty could put into his countnance said thus to the sea: Thou Sea belongst to me, and the Land wheron I sit is mine; nor hath any one unpunish't resised my commands: I charge thee come no furder upon my Land, neither presume to wet the Feet of thy Sovran Lord. (Milton, The History of Britain 281)

9. "Self'-"other" dimension Another possible dimension for the definition of the quality of personal affect is "self' versus "other". The dichotomy implies an interpersonal focus, and a key to it is provided by personal pronouns.

Genre conventions: Personal affect

245

9.1. THOU and YOU — "the other" The primaiy function of second person pronouns is to express orientation towards other participants, either the reader or other characters within the textual world. Second person pronouns had a high factor loading on Factor 1 (0.821) and further statistical tests indicated the significance of this feature (t-test: ρ < 0.041). It is important to find out who is addressed by whom. Keeping the moralising function of early fiction in mind, we have to ask who this second person is in these texts. Are the readers addressed directly? And if not, how are they taken into account in these genres? I went through the examples to see if they included direct authorial comments to the readers using the second person pronouns. It was rather surprising that only few could be found: By thys tale ye may lerne a good generali rule of phesyk. By thys tale ye may se that the olde p~uerbe ys trew that yt is as gret pyte to se a woman wepe as a gose to go barefote. By thys tale ye may perceyue that women ofte tymes be wyfe and lothe to lose any tyme. (A Hundred Mery Talys 20) Α Roge is neither so stoute or hardy as the vpright man ... requiring almes of such as they meete, or to what house they shal com. But you may easely perceiue by their colour that thei cary both health and hipocrisie about them, wherby they get gain ... you shall not thinke ... (Harman, A Caveat or Warening for Commen Cursetors 36-37) here you haue heard the difference twixt a flat foole naturall, and a flat foole artificiall; one that did his kinde and the other who foolishly followed his owne minde: on which two is written this Rime ... (Armin, A Nest of Ninnies 12) The scarcity of examples of this use in HC is even more surprising when we consider the transition from manuscript transmission to print (see above), which emphasises a more personal attitude. It is necessary to go beyond the extracts included in the corpus and look at the books as a whole for this purpose. A good example is the conclusion in Penny Merriments. It is a poem, directed to the reader, which makes the parody of Quakers explicit:

246

Irma Taavitsainen Now Reader, see what here is plainly shown The Character of such who weep and groan Who hate all sins but such as are their own. Divide us from the Wicked, still they cry, And whilst they seem the Flesh to mortifie, They hatch vile frauds, and close Adultery. ... {Penny Merriments 152)

But if the moral commentary and personal involvement are not usually direct, by addressing the reader overtly, how are they incorporated in these texts? Sometimes this is done more covertly: e.g. Deloney repeats the same moral theme in his stories. Sometimes a more direct commentary is found in prefaces or titles. The name of Harman's book, A Caveat or Warening Signals for Commen Cursetors Vulgarely Called Vagabondes (1567), signals the moral content, and the pseudoeloquent term "cursetors", invented by the author, is revealing. Harman's books had a sociological bias and claimed to be doing a public service by explaining to the honest citizenry the ways of the criminal classes (Hughes 1991: 156). This is also found in the direct comment to the reader, quoted above. Although fiction could have didactic aims, there is a difference between instructional texts and fiction. Guiding the reader directly is not very prominent when compared with handbooks and 17th-century scientific writings. We cannot speak of second-person oriented texts in the same way, though the frequency of the pronouns is statistically significant. The overlap with instruction is not conspicuous in this respect, though single examples like the above can be found.19 Instead of direct reader orientation, the fictional characters of conversation provide the poles of interaction, and a two-part system with / and you, the self and the other, must be introduced. Even if the you referred to in fiction turns out to be a fictional addressee, the use of the pronoun always involves the actual reader, either by means of direct address or by the invocation of a more generalised you, with a pretense of being applicable to the current reader by virtue of its gnomic truth value (Fludernik 1995).

Genre conventions: Personal affect

247

9.2. First person — "the self' This brings us to "the self' — the first person singular, which had a fairly high factor loading on Factor 1 "Interaction" (0.623) but not significant in the present assessment (t-test: ρ < 0.22). By definition autobiography, diary, and travelogue report on personal experience in the first person singular. So it is not surprising that no great differences are found. Yet the matter can be taken further by defining the role and quality of the / in these texts. The interactive role of the second person presupposes an interactive first person, so I assessed the first person singular pronouns to find out whether the / of these texts was interactive, defined in relation to the other, or an I with an egocentric, solipsistic focus. This assessment reveals a difference. It can be stated that in fiction the self exists only in relation to other; / or we is always paired off with someone else in the second person, mostly a fictional character, but sometimes the author is paired with the reader (see the examples above). FICTION: "that," sayd she, "is your vncles house." "Nay," saith one of them, "he is not onely my vncle, but also my godfather." "It may well be," quoth she, "nature wyll bind him to be the better vnto you." "Well," quoth they, "we be weary, and meane not to trouble our vncle tonight; but to-morowe, god willinge, we wyll see him and do our duty: but, I pray you, doth our vncle occupy husbandry? what company hath he in his house." "Alas!" saith she, "but one old woman and a boy, he hath no occupying at al: tushe," quoth this good wife, "you be mad men; go to him this night, ... "Now, by my troth," quoth one of them.... "Where dwellest thou!" quoth this parson. "Alas! sir," saithe this roge, " / haue smal dwelling, and haue com out of my way; and / should now," saith he, "go to any towne nowe at this time of night, they woulde set me in the stockes and punishe me." "Well," quoth this pitifull parson, "away from my house, either lye in some of my out houses vntyll the morning, and holde, here is a couple of pence for thee." (Harman, A Caveat or Warening for Commen Cursetors 38)

248

Irma Taavitsainen

In contrast, diaries record the first person only; the extreme example is found in Hoby's text in which the viewpoint is purely egocentric and solipsistic. In the same way travelogues record personal experience, sometimes expressed with the more collective we. The texts also include third person narration and more neutral description of external reality; occasional instances of other pronouns may occur in such contexts. DIARY: Friday 7 After priuat praiers I wrett my notes in my testement, which I geathered out of the Lector the night before: then / did eate my breakfast, then / walked abroad and talked of good thinges, so that I found much Comfort: after / Cam hom I wrett my sermon that was preached the saboth day before, then / went to priuat praier, and so to dinner: after which / taked a litle with som of my frendes, and exercised my body at bowles a whill, of which / found good: then / cam home and wrought tell 4, then / praied with Mr Rhodes, and after walked abroad: and when / Came hom / praied priuatly, and sonne after went to supper: after which I went to the Lector, and then to bed: (Hoby, Diary of Lady Margaret Hoby 70) TRAVELOGUE: Wedynsday, the xv Day of Julii, the ffather Warden of Bedelem cam to vs with lordis of Jherusalem — And Rama thane beyng turkys — The great Türke havyng in Dominyon All the holl londe, And in shorte tyme they concludyd what sume ower patrone should pay for or tribute. And thanne we war suffered to com on londe. The same Day at iij of Cloke at aftir noon, we com on londe, And as we came owt of the boott we war receyvyd by the Turkys and Sarrasyns, and put in to an old Cave by name and tale, ther Screvener ever wrytyng ower namys man by man As we entyred in the presens of the seyd lordis, And ther we lay in the same Grotte or cave all nyght upon the stynking Stable grounde, as well nyght as Day, ryght evyll intretyd by the seyd Turkes Mames. (Torkington, V Oldest Diarie of Englysshe Travel! 23)

Genre conventions: Personal affect

249

TRAVELOGUE:

From London to Amwell bery /Amwell Bury/ which is in Hartfordshire 19 mile, where I staid a day or two, thence to Bishopstafford /Bishops Stortford/ in Essex 13 mile, thence to Dunmew /Dunmow/ 8 long miles thro' severall little villages, its very deep way especially after raines; this is a little Market town, they are altogether taken up about the spinning and prepareing for the Bayes /baize/: all along between that and Colchester you pass but hälfe a mile ere one comes to two or 3 houses all along the road, its from Dunmow to Collchester 22 miles, and mostly clay deep way. (Fiennes, The Journeys of Celia Fiennes 142) Some passages of dialogue are inserted in these texts. Mostly such insertions are occasional, but half of Mowntayne's autobiography is written in dialogue form. An assessment of the interactive passages embedded in these texts revealed an interesting feature. As the focus is on the I, the narrator of the travelogue, for instance, the real referent of the pronoun you may be the / of the text. Thus the movement is not reciprocal but self-centered: TRAVELOGUE:

Whereto an English Gentleman replyed, that hee was the next good voyage after at the Hands: / answered him that / was there also. He demanded in what ship I was? I tolde him in the Rainebowe of the Queenes: why (quoth he) doe you not know me? / was in the same ship, and my name is (Witherington). Sir, said I, I do remember the name well, but by reason that it is neere two and twenty yeers since / saw you, I may well forget the knowledge of you. Well said he, if you were in that ship, / pray you tell me some remarkable token that happened in the voyage, whereupon / told him two or three tokens; which he did know to be true. Nay then, said I, I will tell ... (John Taylor, Pennyles Pilgrimage 131.C2) / and you mix in the same way in Mowntayne so that the prevailing direction of address is towards the self, though some reciprocity is also found (see also the examples above):

250

Irma Taavitsainen

AUTOBIOGRAPHY:

Than sayed my lord cheffe justyes unto me, "Syr, whate make yow here? are you not a Londynar?" "Yes, and yt lyke your lordshyp." "Howe longe have yow be here prysonar?" "Halve a yeare, my lorde." "Who sent yow hether?" "Forsothe, my lorde, that dyd the counsel." (Mowntayne, Narratives of the Days of the Reformation 206)

"you"

Figure 5. Egocentric "you".

Genre conventions: Personal affect

251

In contrast to fiction, the other text types use the first person singular frequently, independently of the dimension "other", with a solipsistic illusion. In several cases the second person you is mostly a projection to I, someone addressing the self in an egocentric focus. The contrast to the interactive self in fiction is striking. The first person singular in Aphra Behn's text is different. There is a narrative frame and the narrator gives personal comments and guidance to the reader. The dual role of the narrator and character are dedicated to authenticating the story, claiming historicity for the events. Behn's work has therefore been characterised as an imaginary "true history" of travel narrative (McKeon 1987: 111-113). Direct quotations are not very common. This work seems to belong to a totally different kind of fiction, foreshadowing romantic prose. The present analysis does not catch the essence of its style.

9.3. Direct questions and THAT-deletion Direct questions had a very high factor loading on Factor 1 (0.838), and they are also significant for the difference between imaginative and nonimaginative text types. They represent the prototypical form of interaction and are closely linked with the occurrence of the second person. The quotations in fiction imitate natural speech, and contain features typical of the spoken level of language such as exclamations, swearing and pragmatic particles, but reductive features of the surface form are also present as THAT-deletion proved significant. Besides direct quotations, this feature occurs in indirect narration in Armin's text. In addition, contracted forms are found in fiction. The foole, being wakened, lookes about him; when he had thanked the woman, asked what newes? sayes the man, Sir, here is your uncle come out of the country to see you. God a mercy cousin! sayes Will Sommers; I thank thee for thy labour, you cannot uncle me so. Yes, truly, sir, I am your own deare uncle, M. William, and with that wept. Are you my uncle? sayes Will. I, sir, sayes hee. Are you my uncle? sayes hee againe. I, sure, and verely too. But are you my uncle, indeed? By my vusse I am, sayes the old man. Then, uncle, by

252

Irma Taavitsainen my vusse, welcome to court, sayes Will Sommers. But what make you heere, uncle? He ups and tels his comming to him. Will takes him by the hand: Come, saies hee, thou shalt see Harry, onckle — the onely Harry in England; (Armin, A Nest of Ninnies 43—44)

10. Indexical features of proximity Besides surge and interaction, deictic features proved important. In the factor analysis both here and now loaded on Factor 1 (factor loadings 0.653 and 0.652 respectively), but only now proved significant in the second phase (t-test: ρ < 0.031). Demonstrative pronouns did not load on any factor, but the difference between the use of this in fiction and nonfiction proved significant in the analysis of variance. Orientational features of language which locate utterances in relation to space and time form the deictic patterning in texts (Simpson 1993: 13). The main types of indexical features are the pairs of adverbs and demonstrative pronouns here/there, now/then, this/that; personal pronouns are also indexical, I and you being the basic dichotomy. Other indexical features are deictic verbs that express the direction of motion in respect to the speaker like come/go, bring/take, tense present/past, and even interjections can be included as they derive their meanings purely from the immediate context, at least in spoken language. In the present assessment, the indexical features of proximity proved significant. The deictic adverb of present time now and the deictic adverb or adjective this, indicating physical proximity to the speaker, were important. Deictics strongly invoke the point of view in texts (Ehrlich 1990: 9697; Toolan 1990: 127); they create orientation, and indicate the way in which egocentricity and the dimension "self-other" works. In addition, it has been pointed out that words like this and that include, besides definiteness, an indication of why the speaker expects the addressee to be able to identify the referent (Chafe 1975: 39). For example, the passage below contains a condensed network of this and now occurrences which are not restricted to direct speech quotations only. Thys horsman heryng ye frere thys intrete for hym sayd to oconer thys Oconer thou seeyst well by thys mannys reporte γ yf I dye now I am out of charyte & not redy to go to heuen & so it ys y' I am now

Genre conventions: Personal affect

253

out of charyte in dede but hou seest well y' this frere ys a good man he is now well dysposyd & in charyte and he is redy to go to heuen & so am not I therfore I pray the hang vp thys freere whyle that he hys redy to go to heuyn. This Oconer heryng this mad answere of hym sparyd the man ... (A Hundred Mery Talys 24) This device helps to bring the events maximally close to the time of narration, and create a "proximal" style (Simpson 1993: 18). Together these features have the function of calling for a more immediate response in the audience. In the same way as the readers are expected to feel the emotions of excitement, sorrow, hesitation, surprise or horror expressed by surge features like interjections, swearing and pragmatic particles, they are guided by indexical features of proximity. Narration is made more vivid, and the immediacy of experience is emphasised. Armin's text uses deictic elements in the same way as interjections, to shift the viewpoint (see above); come and go in the present and past tenses alternate in succession. The quick change from proximal to distal contributes to the special effect of reversal: But they come; and up they came, and to the king they goe, who, being with the lord treasurer alone, merry, seeing them two, how ... (Armin, A Nest of Ninnies 44) In non-literary texts deictic expressions are fewer and serve to make the directions explicit in a matter-of-fact way. For example, the text strategies found in the travelogues of the HC are explicitly iconic (Virtanen 1995). In the description of the scenery below with neutral expressions of the points of compass and existential sentences: TRAVELOGUE: A mile a this side Akeland Castelle I cam over a bridg of one great arch on Gaundelesse a praty ryver rising a vj. miles of by west: and renning by the south side of Akeland Castelle goith a litle beneth it to the great streame of Were. Gaundeles rising by west cummith to Westakeland, by S. Helenes Akeland, by S. Andreas Akeland, and by Bisshop Akelande. The towne self of Akeland is of no estimation, yet is ther a praty market of corne. It standith on a praty hille bytween 2. ryvers, wherof Were lyith on the north side, and Gaundelesse on the south, and an arow shot or

254

Irma

Taavitsainen

more benethe they meete and make one streame, and ren to the este. And ech of these rivers hath an hille by it, so that Bisshop Castelle Akeland standith on a litle hille bytwixt 2. great [hills]. (Leland, The Itinerary of John Leland 70)

11. Modal auxiliaries Modal auxiliaries and adverbs, evaluative adverbs, generic sentences, verbs of knowledge, prediction and evaluation, together with deictic elements, contribute to the expression of point of view and degrees of commitment in texts (Simpson 1993: 47). Modal auxiliaries were included in the present assessment; for adverbs and adjectives, see Meurman-Solin in this volume. Verbs expressing volition and possibility proved significant in t-tests (t-tests: modals of volition ρ < 0.040; modals of possibility ρ < 0.030). In factor analysis, modals expressing volition correlated with Factor 1 (factor loading 0.693), whereas modals of possibility and necessity correlated with Factor 2 (factor loadings 0.720 and 0.692). These auxiliaries mark the reasoning process together with private verbs of mental activity, and this factor describes philosophical and educational writings. Boulomaic modality, expressing "desire" and futurity, seems to be a component of the fictional world in the present selection of texts. In Penny Merriments the use of modals of volition is especially prominent, and they act as fictional space-builders of the make-believe world. The contrast to the factual, matter-of-fact narration is striking as the examples below demonstrate. The examples given below represent the far ends of the scale. More studies are needed to define the quality and shades of modality in these texts.

FICTION:

(Jo) If my master will let me have the house we will brue good Ale, and we will have mault of our own, for we'l keep a stock going in my Master's Mault-house, with his leave, and there is pasture enough to keep two beasts, and conveniences for hogs and poultery, so thou mayest have all things about thee, and keep a maid and live like a Lady.

Genre conventions: Personal affect

255

(Ka) This will be brave indeed John, but what shall we do with our Ale. (Jo) Sell it my sweet one; Let me see, there are eight rooms in the house besides the Cellar, and with a little painting and a few benches it will be very fine, & a handsome sign to draw in company. (Ka) What shall that be. (Jo) The three fair maids, I think. (Ka) Not for a hundred pound I would not have such a sign. (Jo) Why prithee. (Ka) Why man they'd think surely we kept a bawdy house ... (Penny Merriments 117)

DIARY: The xxvij day of July the duke of Suffoke, maister [Cheke] the kynges scolmaster, maister Coke, [and] ser John Yorke, to the Towre. The xxxj day of July was delevered owt of the Towre the duke of Suffoke; and the sam day rod thrugh London my lade Elssabeth to Algatt, and so to the qwens grace her sester, with a Ml hors with a C. velvett cotes. The sam tyme cam to the Flett the yerle of Ruttland and my lord Russell, in hold. The qwen['s] grace mad [sir Thomas] Jarnyngham vyce-chamburlayn and captayne of the garde, and ser Edward Hastyngs her grace mad ym the maister of horsse the same tym (Machyn, The Diary of Henry Machyn 38)

12. Conclusions The aim of this study was to test whether any differences could be found between literary and non-literary text types. Another aim was to achieve a more detailed description of the qualities of personal affect. For these purposes various methods were combined. Factor analysis was used first to establish the affinities of texts and to detect matrices of subjective uses of language, reflecting the underlying dimensions of variation. Further statistical methods were then employed to verify possible text-type markers among the linguistic features that could prove important for various qualities of personal affect in fiction.

256

Irma Taavitsainen

Four different qualities of personal affect were revealed by this study. They are all characteristic of early fiction: surge features in the form of emotional outbursts, an interactive dichotomy between self and other, indexical features of proximity, and boulomaic modality. As the ultimate aim of the study was to chart genre styles and genre conventions, some of these features were assessed in a longer diachronic perspective. The quality of affect in the fiction of the Early Modern section of the HC is veiy different from the kinds of affect I defined in my earlier studies on religious and scientific prose. The repertoires of affective/emotive features in text types vary greatly and there are differences from text to text. Some general trends of evolution can be discerned, but further studies are needed for more specific results.

12.1. "Surge" of personal affect Of the surge features, the most important are exclamations. Their use seems to be a deliberate device of story-telling. Emotional involvement is often prominent especially in fiction, and echoes of earlier writings in the Middle English fabliau tradition can be discerned; genre conventions are an important factor in the frequent occurrence of interjections in written texts. Chaucer uses them in marking turning-points of the plot and foregrounding to create awe and apprehension in the audience. The Late Middle English conventions are continued in a somewhat modified form in Early Modern short fiction, and even later. Swearing, oaths, and pragmatic particles are also important as surge features of affect, and the line of continuation from Middle English is evident. The last sample of fiction from 1706 belongs to another era and another tradition; interjections are found in it as well, but they are used in a different way. Thus the convention continues, although it finds new forms and applications. A possible surge feature available in the language at the time in focus was the use of thou. Statistically it proved significant, but a closer consideration revealed that the frequent use of thou in fiction is connected with the social classes depicted in these texts: country folk and lower and middle class people among which it was the unmarked pronoun of address. Although several texts clearly contain an emotional loading, it cannot be attributed to this linguistic feature alone; together with others, the feature may contribute to the affective tone. In Armin's text and in Penny Merriments the use of thou is marked and serves special purposes: reversal of values in carnivalism, and parody of

Genre conventions: Personal affect

257

religious language and Quaker usage.These conclusions were made on the basis of qualitative assessment and backgound knowledge of the sociohistorical context of these texts. In non-literary texts such as autobiography the use of thou to express heightened emotionality is evident. Some instances are found in outbursts of accusation and contempt, and in one example it is used in a solemn and affective speech.

12.2. Interactive focus Another salient feature contributing to the difference between fiction and the adjoining text types is the reference system of the pronouns. The dichotomy between self and other proved important in fiction, whereas the non-imaginative narrative texts have a more self-centered focus. An interactive, reciprocal system functions in these texts with questions and answers, expressions of emotions and responses to the other partner's words (cf. Factor 1). In contrast, the other text types show a system in which you is often coreferential with I, and the egocentric, solipsistic focus is dominant.

12.3. Immediate context and space-building modality Indexical features of proximity are characteristic of fiction. They indicate that the action is simultaneous with the narration and that the things referred to are close at hand. Special elements like modal auxiliaries of a certain kind help to build up the fictional world in these texts. The above-mentioned linguistic features are mostly found in direct speech quotations that imitate natural speech, and perhaps condense some of its features. It has also been suggested that interjections should be regarded as deictic elements because their interpretation is totally contextbound. This is not entirely true of written texts, as textual functions have also developed. Linguistic features reflecting personal affect may act as text-type markers in some genre and period styles; exclamations are especially distinctive in this respect. These features are essential in reader involvement. They function in various ways. Readers may be asked to pay attention to certain points by conative interjections or collocations. Proximal deictic expressions help to bring the events and characters to the immediate experience

258

Irma Taavitsainen

of the reader. Indirectly they too contribute to reader involvement. Personal affect in the form of charge between participants may be found both text-internally between characters in the text or involving the readership.

13. Evaluation of the methods There is a great deal of variation both in volume and form of personal affect. Factor analysis revealed textual affinities and helped to focus the assessment to texts that stand closest to fiction. Factor analysis also indicated potential text-type markers, and its matrix of interactive features a feasible way of interpreting the final results. Additional statistical methods were used to verify the implications of the hierarchical ordering of linguistic features marking subjectivity. T-tests and f-tests proved the significance of these features in telling fictional works apart from the adjoining non-imaginative texts. Thus it was possible to pinpoint text-type markers and achieve empirical evidence of genre conventions. This study showed that it is necessary to complement quantitative methods by qualitative analysis in order to achieve reliable results. Immediate context is not sufficient, but the books have to be considered in all and their sociohistorical connections must be taken into account; irony, mocking, and parody, or reversal of values in carnivalism, can only be detected by careful reading of the texts; only qualitative stylistic analysis can reveal the nature of involvement and the more subtle shades of meaning in these texts.

Notes I am grateful to Professor Edward Finegan for comments and suggestions for improvement on an earlier version of this study. 1.

Cf. Werlich's classification (1983: 38-41) into the basic text types of description, narration, exposition, argumentation, and instruction. They are defined according to internal linguistic features.

Genre conventions: Personal affect 2. 3.

4.

5.

6.

7.

8.

259

Guillen also points out the importance of the Renaissance letter to the development of the early novel in terms of shared functions (1986: 99-100). Görlach emphasises the role of convention in the recognition of certain well-defined and standardised uses of language as text types in a specific culture (1991: 200-201). Some recent studies concentrate on very narrow text types like wedding reports in newspapers, which are then discussed in great detail, e.g. Suter (1993). "Noble chivalry" in the same way that kitchen furniture may be marketed as "sophisticated elegance" (Margolies 1985: 36). For a linguistic assessment of emotional features in romances, see Taavitsainen (forthcoming d). Biber—Finegan (1989b: 94) define personal affect as a wide range of personal attitudes, emotions, feelings, moods and general dispositions, and place it under the title of "stance" together with evidentiality. According to Halliday (1978: 33) the relationship between participants includes "the degree of emotional charge". With status and contact it forms the third component of the participant relation, but unlike the other two, affect is not always present: it can be turned on or off, and the "volume" may be adjusted (Martin 1992: 533). (For a discussion, see Taavitsainen 1994b). Recent discussions have largely concentrated on the anthropological study of emotions and discourse (see Lutz—Abu-Lughod 1990) and gender studies (Poynton 1989). Lexical studies of emotions have also received attention (e.g. the first part of Busse 1992 is called "Linguistic aspects of emotions", and it contains several articles on lexis). Of the non-fictional texts, Mowntayne has the highest proportion of direct speech: 2899 words versus 2842 words of narration, total 5740. No other text comes near it. The underlying dimensions were interpreted to reflect (1) involvement versus informational focus, (2) narrative discourse, (3) endophoric versus exophoric reference, (4) persuasion, (5) abstract and technical discourse, (6) informational elaboration under strict real-time constraints, and (7) academic qualification or hedging (Biber 1988: 101-120). For example, present and past tense were also included in the primary analysis but excluded from the final one. Their communalities in the factor analysis were low, which showed that they did not really contribute to any factor; see note 13. It is possible to rotate the factors to simplify the pattern and make the distinctions stand out more clearly; thus in a rotated solution each factor is characterised by a few features that are most representative. The extent to which the clusters of features distinguish among different sets of texts varies, and the factors are given eigenvalues that show their importance in explaining the amount of underlying variation.

260

Irma

Taavitsainen

I acknowledge the expertise of Ms Leena Sadeniemi at the Computing Centre of the University of Helsinki in the statistical analyses carried out for this study. 9. I shall turn to this grouping and the special quality of personal affect in philosophy and education in a future study. 10. It is, of course, evident that the small number of texts both in fiction and adjoining genres is not adequate for definitive conclusions, but this material may give indications of whether the question is appropriate and how to find a possible answer to it. 11. The corpus parameters "interactive versus x" and "informal versus x" are important for the present assessment. They are coded in the corpus, but they have not been systematically applied in research. I wanted to test their influence on the above factors. T-tests were used for this end. Factor scores were counted for each text, and the group means of scores were compared with one another. The results show that the parameter "interactive versus x" is very significant in Factor 1 (p < 0. 0005), significant in Factor 2 (p < 0.026), but insignificant in Factor 3 (p < 0.058). Likewise, the parameter "informal versus x" was significant in Factor 1 (p < 0.001), in Factor 2 (p < 0.013), but insignificant in Factor 3 (p < 0.677). 12. The parameter NI versus NN coincides with the present focus on the difference between fiction and the adjoining genres because plays are excluded from this phase of my study. In Late Middle English, Mandeville 's Travels belongs to imaginative narration although it is a travelogue. 13. Not all features proved significant with respect to the difference between imaginative versus non-imaginative narration. Non-significant features were e.g. so, then, that, here, and past tense versus present tense. 14. My hypothesis was that exclamations act as a marker of the fictional mode. To test my hypothesis I divided the material into two categories: the fictional and the non-fictional, the former group comprising the material from The Canterbury Tales, Gower, Caxton, romances and drama. A t-test was performed and the results proved that the difference in the use of exclamations in the two groups is highly significant. To shift the angle, another ttest was performed with the texts by well-known literary authors, Chaucer, Gower, Caxton and Malory, on the one side, and the anonymous texts and non-literary authors, including the mystical authors of religious treatises, on the other side; drama was excluded from this t-test. The difference was highly significant in this case as well (Taavitsainen forthcoming a). 15. In Chaucer's fabliaux the collocations of interjections are a regular feature in highlighting the stories and bringing turning-points of the plot to the audience's attention: The Miller's Tale provides a case in point. For comparison, the same story in a prose version included in "Jokes and Jests" in the Penny Merryments contains only one interjection, in the only direct quotation.

Genre conventions: Personal affect

261

16. An example of foregrounding can be found in the Squire's Tale: So ferde this tercelet, alias the day! Though he were gentil born, and fressh and gay, And goodlich for to seen, and humble and free, He saugh upon a tyme a kyte flee, And sodeynly he loved this kyte so That al his love is clene fro me ago, .... withouten remedie! (621-629) 17. This is the mechanism of involvement in Middle English mystical writings. It is most frequent in Julian of Norwich's treatise and Richard Rolle uses it in his affective works (see Taavitsainen 1993). 18. A change in the pronouns is used as a dramatic effect in one speech act which reflects a change of attitude when a scholar announces the end of his loyalty to the tyrant (lines 806-810). The passage has a strong emotional loading. 19. A typical second-person oriented text guides the reader in phrases like you shall see/iudge/marke/ you must vnderstand/marke/ you may see/perceave, etc. The interpersonal aspect is often present in phrases like I will show you (Taavitsainen 1994a).

References Adamson, Sylvia 1994 "Subjectivity in narration: Empathy and echo", Subjecthood and subjectivity. Paris—London: Ophrys, 193-208. Ameka, Felix 1992 "Interjections: The universal yet neglected part of speech", Journal of Pragmatics 18(2/3): 101-118. Benveniste, Emile 1966 Problems in general linguistics. Trans, by Mary Elizabeth Meek. [1971] Coral Cables, Flo.; University of Miami Press. Biber, Douglas 1986 "Spoken and written textual dimensions in English: Resolving the contradictory findings", Language 62: 384—414. 1988 Variation across speech and writing. Cambridge: Cambridge University Press. 1989 "A typology of English texts", Linguistics 27:3-43. Biber, Douglas—Edward Finegan 1986 "An initial typology of English text types", in: Jan Aarts—Willem Meijs (eds.), Corpus linguistics II. Amsterdam: Rodopi, 19-46.

262 Irma 1989a

Taavitsainen

"Drift and the evolution of English style: A history of three genres", Language 65/3: 487-517. 1989b "Styles of stance in English: Lexical and grammatical marking of evidentiality and affect", Text 9-1: 93-124. Blake, Norman F. 1992 " Why and what in Shakespeare", in: Toshiyaki Takamiya—Richard Beadle (eds.), Chaucer to Shakespeare: Essays in Honour of Shinsuke Ando. Cambridge: D. S. Brewer, 179-193. Busse, Wilhelm G. (ed.). 1992 Anglistentag 1991: Proceedings, Düsseldorf. Tübingen: Max Niemeyer. Calvo, Clara 1992 "Pronouns of address and social negotiation in As you like it", Language and Literature 1/1: 5—27. Carter, Ronald—Walter Nash 1990 Seeing through language: A guide to styles of English writing. Oxford—Cambridge, Mass.: Blackwell. Chafe, Wallace L. 1975 "Givenness, contrastiveness, deflniteness, subjects, topics and point of view", in: Charles N. Li (ed.), Subject and topic. New York—San Fransisco—London: Academic Press, 25-55. 1985 "Linguistic differences produced by differences between speaking and writing", in: D. R. Olson—N. Torrance—A. Hildyard (eds.), Literature, language and learning: The nature and consequences of reading and writing. Cambridge: Cambridge University Press, 105123. Davis, Walter R. 1969 Idea and act in Elizabethan fiction. Princeton, New Jersey: Princeton University Press. Ehrlich, Susan 1990 Point of view: A linguistic analysis of literary style. London—New York: Routledge. Fludernik, Monika 1993 The fictions of language and the languages offiction. London—New York: Routledge. 1995 "Pronouns of address and 'odd' third person forms: The mechanics of involvement in fiction", in: Keith Green (ed.), New essays on deixis?, 99-129. Fowler, Alastair 1982 Kinds of literature: An introduction to the theory of genres and modes. Oxford: Clarendon Press. Frijda, Nico Η. 1986 The emotions. Cambridge: Cambridge University Press.

Genre conventions: Personal affect

263

Görlach, Manfred 1991 "Text types and the linguistic history of Modern English", in: Claus Uhlig—Rüdiger Zimmermann (eds.), Proceedings of the Conference of the German Association of University Professors of English XII. Tübingen: Max Niemeyer, 195-215. Guillen, Claudio 1986 "Notes toward the study of the Renaissance letter", in: Lewalski (ed.), 70-101. Halliday, Μ. A. K. 1978 Language as social semiotic: The social interpretation of language and meaning. London—New York—Melbourne, Auckland: Edward Arnold. 1989 Spoken and written language. Oxford: Oxford University Press. Hart, Francis R. 1974 "Notes for an anatomy of modern autobiography", in: Ralph Cohen (ed.), New directions of literary history. London: Routledge and Kegan, 221-247. HC = Helsinki Corpus 1991 The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Heine, Berd—Ulrike Claudi—Frederike Hünnemeyer 1991 Grammaticalization: A conceptual framework. Chicago—London: The University of Chicago Press. Hughes, Geoffrey 1991 Swearing: A social history of foul language, oaths and profanity in English. Oxford—Cambridge, Mass.: Blackwell. Imbrie, Ann E. 1986 "Defining nonfiction genres", in: Lewalski, 45-69. Jucker, Andreas H. (ed.) 1995 Historical pragmatics. Amsterdam—Philadelphia: John Benjamins. Jucker, Andreas H. forthc. "The origin of the discourse marker well: A case for historical pragmatics". 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. Kytö, Merja forthc. "Robert Keayne's notebooks: A verbatim record of spoken English in early Boston?", in: Susan Herring—Pieter van Reenen—Lene Schesler (eds.), Textual parameters in older languages.

264 Irma

Taavitsainen

Leech, Geoffrey Ν.—Michael Η. Short 1981 Style in fiction: A linguistic introduction to English fictional prose. London—New York: Longman. Lewalski, Barbara Kiefer (ed.). 1986 Renaissance genres: Essays on theory, history and interpretation. Cambridge, Mass.—London, England: Harvard University Press. Lutz, Catherine A.—Lila Abu-Lughod 1990 Language and the politics of emotion. Cambridge: Cambridge University Press. Margolies, David 1985 Novel and society in Elizabethan England. London— Sydney: Croom Helm. Martin, J. R. 1992 English text: System and structure. Philadelphia—Amsterdam: John Benjamins Philadelphia. McKeon, Michael 1987 The origins of the English novel 1600-1740. Baltimore—London: The Johns Hopkins University Press. Meurman-Solin, Anneli 1993a "Involvement markers in the characterization of early prose genres", in: Anneli Meurman-Solin, Variation and change in early Scottish prose: Studies based on the Helsinki Corpus of Older Scots. (Humanarum Litterarum 65 / Annales Academiae Scientiarum Fennicae. Diss.) Helsinki: Suomalainen Tiedeakatemia, 276-312. 1993b "The author-addressee relationship and the marking of stance in the characterization of sixteenth- and seventeenth-century genre styles", Revue Beige de philologie et d'histoire, Fasc. 3: Langues et litteratures modernes, 733-745. Mühlhäusler, Peter—Rom Harre 1990 Pronouns and people : The linguistic construction of social and personal identity. Oxford: Blackwell. Nevanlinna, Saara—Irma Taavitsainen 1993 St Katherine of Alexandria: The Late Middle English prose legend in Southwell Minster MS 7. Cambridge—Helsinki: Boydell—Brewer— The Finnish Academy of Science and Letters. Poynton, C. 1989 Language and gender: Making the difference. Geelong, Vic.: Deakin University Press. [1985] Oxford: Oxford University Press. Quirk, Randolph—Sidney Greenbaum—Geoffrey Leech—Jan Svartvik 1985 A comprehensive grammar of the English language. London: Longman.

Genre conventions: Personal affect

265

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. Sampson, George 1970 The concise Cambridge history of English literature. (3rd revised edition.) Cambridge: Cambridge University Press. Simpson, Paul 1993 Language, ideology and point of view. London—New York: Routledge. Spufford, Margaret 1981 Small books and pleasant histories: Popular fiction and its readership in seventeenth-century England. Cambridge: Cambridge University Press. Suter, Hans-Jürgen 1993 The wedding report: A prototypical approach to the study of traditional text types. Amsterdam—Philadelphia: John Benjamins. Taavitsainen, Irma 1988 Middle English lunaries: A study of the genre. (Mdmoires de la Societe Ndophilologique de Helsinki XLVII. Diss.) Helsinki: Societ6 N6ophilologique de Helsinki. 1993 "Genre/subgenre styles in Late Middle English?", in: Rissanen et al., 171-200. 1994a "On the evolution of scientific writings between 1375 and 1675: Repertoire of emotive features", in: F. Fernändez et al (eds.), Papers from the 7th International Conference of English Historical Linguistics, Valencia, Sept. 1992. Philadelphia—Amsterdam: John Benjamins, 329-342. 1994b "Subjectivity as a text-type marker in historical stylistics", Language and Literature 3/3: 197-212. 1995a "Interjections in Early Modern English: From imitations of spoken to conventions of written language", in: Andreas H. Jucker (ed.), 419445. 1995b "Narrative patterns of affect in four genres of the Canterbury Tales", The Chaucer Review 30/2: 82-101. forthc. a "Exclamations in Late Middle English", in: Jacek Fisiak (ed.), Studies in Middle English. Berlin—New York: Mouton de Gruyter. forthc. b "By Saint Tanne: Pious oaths or swearing in Middle English? An assessment of genres", in: Raymond Hickey—Stanislav Puppel (eds.), Festschrift in honor of Jacek Fisiak's 60th birthday. Berlin—New York: Mouton de Gruyter. forthc. c "Genres and text types in Medieval and Renaissance English", to appear in Poetica: An International Journal of Linguistic-Literary Studies 47: 49-62 (Tokyo, Japan).

266 Irma

Taavitsainen

forthc. d "Emphatic language and Romantic Prose: Changing functions of interjections in a sociocultural perspective", European Journal of English Studies. Special Issue. Language theory and practice in current literary scholarship, ed. by Monika Fludernik. Todorov, Tzvetan 1990 Genres in discourse. Cambridge: Cambridge University Press. Toolan, Michael J. 1988 Narrative: A critical linguistic introduction. London—New York: Routledge. 1990 The stylistics of fiction: A literary-linguistic approach. London—New York: Routledge. Virtanen, Tuija 1995 '"Then I saw to antique heddes': Discourse strategies in Early Modern English travelogues", in: Andreas H. Jucker (ed.). Amsterdam— Philadelphia: John Benjamins, 499-513. Wales, Katie 1983 "'Thou' and 'you' in Early Modern English: Brown and Gilman reappraised", Studia Linguistica 37/2: 107-125. Werlich, Egon 1983 A text grammar of English. Heidelberg: Quelle and Meyer.

Towards reconstructing a grammar of point of view: Textual roles of adjectives and open-class adverbs in Early Modern English Anneli Meurman-Solin

1. Introduction In this study the general approach to the concept of point of view in texts is based on recent developments in corpus-based sociohistorical linguistics. A text as a linguistic construct is affected by extralinguistic features in various complex ways, and this is reflected in variation and change over time. For example, besides their text-organizing function, linguistic expressions of point of view can be meaningfully related to the authoraddressee relationship (Meurman-Solin 1993b). Studies of features such as the social functions of texts, the social roles and social networks of their authors and addressees, the general narrative framework of texts and the varying narrator roles in them, provide information that may explain linguistic choices on all levels of language use. In the present study the hypothesis is that the above-mentioned extralinguistic features condition textual participant roles. A closer examination of how these roles are explicitly or implicitly expressed in texts may lead not only to the identification of a text's ideology (Simpson 1993: 5) but also to a better understanding of processes towards the establishment of extra-linguistically conditioned linguistic conventions and norms. Additionally, it may shed light on processes towards diversification in the development of genres. It is the latter of these aims, the presumably multidirectional diachronic development of genres, that the present study sets out to explore further. In more specific terms, the more immediate aim of the study is to provide material for the reconstruction of a grammar of point of view, and to identify genre-specific features and genre styles in such a grammar in a

268

Anneli Meurman-Solin

wide range of Early Modern English texts. 1 A further aim is to increase our knowledge of how the social and communicative functions of texts are reflected in linguistic choices and in patterns of information processing resulting in various degrees of syntactic complexity. To achieve the aims defined above, the analysis of genres is viewed as a multi-disciplinary activity (cf. Bhatia 1993: 1). Genre analysis discusses the more or less 'conventionalized communicative purposes' (Bhatia 1993: 1) of genres and the ways in which such purposes link with social and cultural factors, and, as Bhatia (1993: 1-2) succinctly puts it, it shares approaches applied in 'discourse analyses of institutionalized use of language in socio-cultural settings with a heavy emphasis on communication as social action'. To use Hallidayan terminology (Halliday—Hasan 1985 [1990]: 12), the analysis of point of view is here mainly concerned with the tenor of discourse, but the use of particular linguistic features is linked with the field and the mode of discourse by stressing 'the nature of the social action that is taking place' and the way in which explicit linguistic devices are used to organize texts in terms of text categories (such as non-imaginative narration, argumentative and expository texts). I became interested in the social and textual roles of authors and their readers when trying to identify extralinguistic factors conditioning linguistic variation and change in the history of Scottish English, a minority variety where these roles could be assumed to be affected by how that variety related to a majority variety under various social and political pressures. The general assumption was that by describing what was known about the individual authors, their social status and education, their social networks, and, on the other hand, about the audiences and addressees of their texts, it would be possible to identify factors conditioning linguistic variation and change, particularly as regards language standardization, in Scotland the so-called anglicization of Scottish English. Meurman-Solin (1993a) shows that the most important conditioning factors in the standardization of Scots are extralinguistic variables such as dialect, i.e. dialects within Scots, whether the text was printed or not, 2 genre, text category, the author's sex, and, in letters, the participant relationship. Perhaps predictably, the conditioning of the variable 'genre' proved to be more difficult to trace, partly because of problems inherent in compiling a diachronic corpus, where scarcity of texts in earlier time periods may sometimes fatally decrease its representativeness, and partly because of the innate heterogeneity of certain genres or overlappings between them. 3 To suggest solutions to some of these problems, the present study provides more information about texts as representatives of

Reconstructing a grammar of point of view

269

genres and text categories, and about similarities and differences between them. The study isolates a number of linguistic features reflecting point of view in texts, and it attempts to interpret their frequencies and distributions in terms of how they relate to other linguistic and extra-linguistic features of the texts and how they compare with findings in other recently published genre studies (see especially Taavitsainen 1993, 1994 and her chapter in the present volume). Given the special challenges of a diachronic perspective and the innate problems related to a semantic analysis of large amounts of data, further study will be required to deal in more detail with some of the findings presented here. At present we do not know exactly what the normal, basic, least marked or canonical (see Enkvist 1988: 138) features of prototypical representatives of genres are in a given context at a given time (cf. Enkvist 1988: 127), and therefore both the description of one particular genre and comparisons between genres necessarily remain tentative. Moreover, the traditional classification of texts, a framework used as a practical tool also in the Helsinki Corpus, tends to rigidify the interpretation of findings and their correlation patterns with extralinguistic factors. Therefore, instead of restricting the focus to similarities and differences between genres, priority is always given to properties of individual texts. First and foremost, my study tries to identify features that could be claimed to be text-specific stylistic properties from a synchronic point of view. An attempt is then made to suggest what features can be identified as period-bound or petrified generic conventions and what diachronic developments features and their co-occurrence patterns undergo. It is plausible that the conservative or innovative pressures attested can be related to changes in the social functions of texts and the rise of new ideals and ways of thinking. 4 Ultimately, it may be possible to identify not only diagnostic features in individual texts but also features clustering in particular text types, and to build a network of overlappings and cross-cuttings between them resulting in a scalar or multi-dimensional typology. A text is viewed as a network of points of view expressed by markers of attitude such as private verbs, modal auxiliaries, adverbs and adjectives in the roles of modality and degree, expressions conveying evaluations, and so on. In my terminology, point of view is directly related to participant voices in a text, whereas involvement refers to a functional notion in a set of textual dimensions: 'integration' versus 'fragmentation', 'involvement' versus 'detachment' (Chafe 1982, 1985). As these four notions are also directly related to degrees of informational density and syntactic complexity, various aspects of genre styles can be illuminated just by

270

Anneli

Meurman-Solin

considering the frequencies and distributions of adverbs and adjectives and their position and function in sentences. As comparative material, studies based on large corpora, those using a multi-feature approach and those relating linguistic evidence to sociohistorically relevant extralinguistic evidence, are particularly significant here. Degree of involvement as opposed to informational density has been shown to be relevant in describing genre styles (Biber 1988, Biber— Finegan 1992), and especially frequencies and distributions of markers of stance have given interesting evidence for suggesting a typology of texts (Biber—Finegan 1989). In the Helsinki Scots Corpus extralinguistic factors such as parameters describing the author or the audience of a text shed light on the general function of genres in their sociocultural context (Meurman-Solin 1993a and 1993b), and also on the more pragmatic needs reflected in linguistic choices in language contact situations. In addition to these and other studies based on computer-readable corpora, the present study has been motivated by ideas supporting the importance of point of view as a text-organizing agent (for example Simpson 1993) and, on the other hand, by evidence of grammaticalization processes related to expressions of attitude (for example Hopper—Traugott 1993).

2. The text corpus and methodological considerations A corpus of approximately 550,000 words of running text was used, available in the Early Modern English section of the Helsinki Corpus of English Texts. The texts, altogether 79, represent sixteen genres, each type usually illustrated by two texts per each one of the three time periods: 1500-1570, 1570-1640, and 1640-1710. The genres are: law, handbooks, science, educational treatises, philosophy, sermons, trial proceedings, history, travelogue, diaries, biography, fiction, comedies, private letters, official letters and the Bible. Scientific treatises are subdivided into 'science, medicine' and 'science, other', biographies into 'auto-biography' and 'biography, other', and the Bible contains extracts from both the Old Testament and the New Testament. Private and official letters each contain letters by about ten different individual writers per period. As was stated above (p. 268), because of the serious limitations of the material the findings per genre can only be very preliminary, but the corpus can successfully identify similarities and differences between texts

Reconstructing a grammar of point of view

271

that serve to regroup them in various ways. It is shown later (see for example Tables 1 and 8) that, in fact, several texts share practices that confirm the conventional typological labels coded into the corpus, whereas others, as a result of their idiosyncratic practices, do not allow any straight-forward comparisons within the corpus. In addition to date and genre, the following extralinguistic parameters coded into the Helsinki Corpus are taken into consideration in the statistics: relationship to spoken language, sex and social rank of author, audience description, interaction, setting and prototypical text category.

3.

Adverbs

An analysis of all open-class adverbs ending in -ly and variants was done for each text (for information on some high-frequency closed-class adverbs, see note 14). It may seem that, apart from practical considerations, there is no valid justification for restricting the material to open-class adverbs. Yet the exclusion of structural alternatives makes it possible to trace changes from adverbs integrated in sentence structure to those functioning as sentence adverbs (disjuncts), so that formal features do not influence the diversification or restriction of their function and meaning. Moreover, as the position of an adverbial — specified in terms of initial, medial and final — in a sentence may be relevant in the assessment of its semantic role and syntactic function (see pp. 277-284), only structural realizations allowing all these positions provide useful evidence for studying diachronic variation and change. In the present study, both the position within the sentence structure and the position in the text structure were considered. A difference was made between sentence-initial disjuncts and subject-oriented item subjuncts (see Quirk et al. 1985: 572578). Despite the relative infrequency of relevant examples, an attempt was made to trace developments from medially or finally positioned adverbials integrated in the sentence structure to syntactically peripheral disjuncts, typically positioned at the beginning or separated by means of punctuation (cf. Swan 1988). Significantly, such patterns of variation and change can be related to grammatical ization processes from propositional via textual to expressive meaning (Traugott 1982: 257, 1990: 497). In my further studies, I will look at whether the development of textual and expressive meanings and sentence adverbial functions can be traced in terms

272

Anneli Meurman-Solin

of how the changes are located in specific genre styles or text types, the ultimate aim being the identification of subjectifying genre styles. Altogether 3212 open-class adverbs were analysed. Despite the relatively high total number of instances, the breakdown into subcategories often results in small numbers per subcategory in individual texts, which sometimes weakens the statistical significance of the results. This problem cannot be avoided in studies of lexical items in corpora consisting of relatively short text extracts (ranging from 5,000 to 15,000 words). The semantic developments of individual adverbs were taken into account in the categorization of occurrences (see section 6). This categorization was based on an analysis of each occurrence in a context long enough to allow the interpretation of function and meaning. 5 The problem of homomorphy (see for example Quirk et al. 1985: 70-75) was thus solved by means of a manual classification of occurrences. There were approximately 1,400 evaluative, 450 modal, 450 time, 400 descriptive, 300 intensifying and 200 focusing adverbs. The analysis of adverbial scope and orientation was based not only on positional criteria but also on textual and semantic factors. As it was assumed that the use of evaluative expressions realised by means of adverbs significantly correlated with verb category, the semantic category of the verb was defined in each sentence where an adverb expressing evaluation had been attested. The verb categories used as labels in the classification were as follows: verbs of action, speech act verbs, verbs of inert perception and cognition, verbs acting as metalinguistic comments, verbs of volition, copular and relational verbs. It turned out, however, that only the first two were sufficiently frequent to allow a more detailed study; a statistical analysis was performed only of verbs of action (see section 9). Information about whether the verb was used transitively or intransitively, in the active or in the passive voice was also recorded.

4.

Adjectives

Because of the high frequency of adjectives, a random sample of 15,000 adjectives in the WordCruncher word list was analysed; in addition, adjectives derived from numerals and own and whole were ignored in the

Reconstructing a grammar of point of view

273

statistics. Information about the syntactic functions (prenominal modifier, postnominal modifier, subject complement, object complement) was recorded, ignoring, however, most participles.6 No distinction was made between adjectives on the basis of their morphological characteristics. To allow the compound effect of the two features to be assessed in a grammar of point of view, the classification of adjectives reflects that of adverbs (see section 6). Besides their syntactic properties, adjectives were grouped into the following subcategories: temporal (point of time: the present study; habitual: the usual reaction; generic: a common practice), locative (the left ventricle), descriptive (a slow movement), evaluative (an easy task), state of mind (I am sorry that I am late.), modal (epistemic: a possible solution; perception: a manifest lie; deontic: a necessary change), intensifying (a complete victory) and focusing (a particular text, an additional example). A great majority of the adjectives studied were either descriptive (7,000) or evaluative (4,000); in addition, there were approximately 1,000 modal, 1,000 adjectives referring to state of mind, 750 to time, 500 marking focus, 400 indicating place, and 300 with an intensifying function. The decision to focus chiefly on modal and evaluative adverbs and adjectives of course drastically reduces the range of material which is undoubtedly relevant in the reconstruction of point of view. Not only modal auxiliaries and verba sentiendi but also lexical choices of various kinds (see for example Simpson 1993: 108-109) and features related to types of transitivity are of great significance. The findings will, however, be compared with the degree of involvement attested by means of features included in Biber's Factor 1 (see Meurman-Solin 1994 and forthcoming; cf. also Taavitsainen in the present volume). Moreover, the significant correlation patterns between various types of attitudinal expressions in the material included in this study stress their converging roles in a grammar of point of view (see pp. 316-317, 324-328, Appendix 2 and Figure 1, p. 274).

274

Anneli

Meurman-Solin

C/5 ω

α

JI Ρ Cu ώ ω .s

*Ή -2§

6 Ii 11 X Q S o . S £ ω 'Σλ I

§ .Ξ δII Ηπ < s

'S ω IIf