The expressivity of grammar

As an alternative we claim that the meaning of the possessive dative is the indication of the speaker's attachment to the person referred to by the dative and substantiate the claim by a

The Expressivity of Grammar

W DE

G

Topics in English Linguistics

The Expressivity of Grammar Grammatical Devices

Expressing Emotion across Time

Axel Hübler

Mouton de Gruyter

Berlin · New York 1998

is a Division of Walter de Gruyter & Co., Berlin

© Printed on acid-free paper which falls within the guidelines of the

ANSI to ensure permanence and durability

Library of Congress Cataloging-in-Publication-Data

Hübler, Axel

The expressivity of grammar : grammatical devices

expres-sing emotion across time / Axel Hübler

p cm - (Topics in English linguistics ; 25)

Includes bibliographical references (p ) and index

ISBN 3-11-015780-2 (alk paper)

1 English language - Grammar - Psychological aspects

2 Emotions I Title II Series

The expressivity of grammar : grammatical devices expressing

emotion across time / Axel Hübler - Berlin ; New York :

Mou-ton de Gruyter, 1998

(Topics in English linguistics ; 25)

ISBN 3-11-015780-2

© Copyright 1998 by Walter de Gruyter GmbH & Co., D-10785 Berlin

All rights reserved, including those of translation into foreign languages No part of this

book may be reproduced 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: Werner Hildebrand, Berlin - Binding: Lüderitz & Bauer-GmbH, Berlin

Printed in Germany

Contents

Chapter 1

Communication, emotions and expressivity

1.1 Aim and scope 1 1.2 Contextualizations 2

1.2.1 Expressivity 3 1.2.1.1 A pragmatic and interactional perspective 10

1.2.2 Emotions and expressions 11

3.2 The state of the art 41

3.3 The ethic dative as speaker's attachment 43

3.3.1 Signs, meanings and scopes 43

3.3.2 Mode of attachment 46

3.4 Ethic dative versus benefactive dative? 52

3.4.1 Grades of attachment and grades of extension 57

3.5 Synopsis 61

5.4 Toward a non-temporal view of the present perfect 100

5.6 The use of the present perfect in the history of English 113

5.6.2 On the linearity of language change 126

6.3 The i/o-periphrasis in affirmative statements 130

6.4 The i/o-periphrasis in negations and questions 143 6.5 Case study: A Supplicacyon for the Beggers, by S Fish 147

6.5.2 Economizing the use of the i/o-periphrasis 153

Contents vii

6.6 The development of the do-periphrasis toward

present-day English usage 157

6.7 Synopsis 160

Chapter 7

The gei-passive

7.1 Plan of discussion 162

7.2 Main-stream concepts and alternatives 162

7.3 What's passive about get + V-ecH 166

7.4 Instrumental condition of the gef-passive 168

7.5 Propositional condition of the gef-passive 171

7.6 Attachment 174

7.6.1 Attachment and describability 177

7.7 Case study: S Terkel, Working People talk about

what they do all day and how they feel about what they do 180

7.8 Synopsis 185

Chapter 8

Putting the results in perspective: Grammar and the Self

8.1 The major results 187

8.2 Subjectification 190

8.3 A cultural-historical frame 192

8.3.1 The twelfth century renascence - before and after 192

8.3.2 The Renaissance and after 195

8.3.3 The present time 201

Notes 204 References 222 Index of authors 244

1 Communication, emotions and expressivity

1.1 Aim and scope

Expressivity is one of the dimensions defining human communication It has its basis in the personal setting of every communicative event in that it originates in a person (sender, addresser) and is directed at some other person (receiver, addressee) It is also related to a person's self-expression, the self comprising her/his emotions, no matter whether they relate to inner dispositions or to evaluative attitudes, no matter whether they are rooted in her/his individual biography (cf Koch 1960) or stimulated by a given situation Expressivity, however, presupposes that there are means available through which such emotions can be expressed

The usual view would follow the principle of the division of labor and assign expressivity to the body (and paralinguistics), tackling language proper from the (sole) perspective of how it serves the needs to express thoughts and perform intellectually controlled actions This accords with a general overestimation of the intellectual part of communication, concomitant with a low regard for emotion and subjectivity (cf Lyons 1982:

101 and Birdwhistell 1971: 66).1 This distribution of scholarly concern,2 however, is in contrast to phylogenetic insights, according to which language originates in the transformation of personal experiences into expressions in order to preserve them (cf Langer 1957: 118-119 and Schmidt 1989: 47-52).3 It is, furthermore, inconsistent with cognitivist stances and related neuro-biological and bio-psychological views such as those advanced by the adherents to the paradigm of radical constructivism (cf McNeill 1992)

It is the aim of this book to modify and correct the undue emphasis of linguistics on the intellectual dimension of language by recovering some of the expressive possibilities which, I contend, language has at its disposal and which, moreover, human beings constantly take recourse to The principle outlook taken here overlaps to a large extent with ideas advocated by Parret (1986) In the words of one of his reviewers,

At the center is the idea that the subject of discourse is a 'passional being', as opposed to the more traditional view which conceives of the speaker as being involved in a strictly cognitive activity the emotion

is not considered as a meaning that is somehow expressed, but as an

operator modifying all meaning (Pezzini 1991: 95)

Such an extreme subjectivist view is still well beyond what cognitive linguistics currently appears to be mainly concerned with, but it is gaining ground, the present treatment itself being an instance of this trend

In order to achieve my objective with greater force and clarity, I will radicalize my approach in two ways First, I will not pay much attention to the explicit form of expressing emotions through identifying and describing them and thus making them the content of a proposition Rather, I will concentrate on the forms of expressivity implicit in grammatical devices, such as the expanded form, the present perfect or the gei-passive On the one hand, these grammatical devices seem, at first sight, to be the means least expected to be expressive;4 on the other hand, they seem to provide the best possibilities, due to their high text frequency, for continually expressing emotions concomitant with "objective" propositional contents Second, I will not aim at completeness and balance, but restrict myself to such devices that illustrate the argument at issue best In effect, six devices will be tackled and their use in specific periods of English will be examined and, where possible, compared with present-day English

The opposition to current objectivist views, which either totally neglect the functional load of the grammatical devices under inspection or intellectualize them, determines the make-up of the individual chapters in at least two (interrelated) respects: (a) They will show a design that will constantly contrast my own subjectivist views with those advanced by objectivists (b) The arguments will mainly utilize language data that objectivists have worked with before and submit them to a qualitative method of reasoning; it

is, I hold, the best solution in face of the only other valid, but rather impracticable alternative, i.e., pursuing for as many as six different devices

a quantitative method based on large corpora

The recent academic interest in emotion (cf Kövecses 1986, inter alii)

Contextualizations 3

does not yet really mark a turning point because it documents an interest in

a better understanding of emotions as a cognitive phenomenon rather than an interest in emotional elements within linguistic structure Ochs and Schieffelin (1989), however, move more along the linguistic line; the claims they make and the theoretical position they define are supported by an impressive amount of cross-linguistic data, though these data are not really shown to be pertinent.5 From my point of view, the fairest treatment that the emotional dimension of language has received so far in linguistic literature

is to be found in a paper by Labov (1984), even if it is no more than a programmatic statement

As long as we can locate a plausible interpretation in cognitive [i.e., intellectual/objectivist, A.H.] terms, oriented to the processing of information, it is not a vital matter to recognize emotional meanings like 'intensive' or social meanings like 'moral indignation' They play no more important role than any other redundant features that cluster about grammatical structure It is quite otherwise when no cognitive or referential meaning appears ~ a cognitive zero ~ or when the context is inconsistent with the cognitive meanings usually recognized ~ a cognitive contradiction We then have no choice but to recognize social and emotional meanings as an integral part of the central grammatical system (Labov 1984: 47)

Sections 1.2.1 - 1.3 will provide a brief survey on some ideas and notions which I consider basic to the topic of language and emotion

on the utilitarian functions, i.e., the representative/referential or appellative/conative, as is the case with Joos (1950), for example

The basics of communication itself can also be seen as interwoven with

expressivity In as much as communication as such follows what Grice (1975) refers to as the "cooperative principle", any communicative act can

be expected to meet the "sincerity condition", which ensures that the act is performed only if the communicator really wants to do what s/he does when performing the act For instance, if a communicator makes a promise, s/he should intend to do what s/he promises; if s/he thanks someone, s/he should

be grateful or appreciative for what the person concerned did; or if s/he makes an assertion, s/he should believe it to be true and consequently be willing to accept some kind of responsibility (cf Searle 1969: 54-71, particularly for speech acts) In other words, it is the honest intention that represents another instance of a general kind of expressivity extant in any communicative act/speech act, no matter what the particular intention is (cf also Sbisä 1990).6

The particular communicative intention is another case of self-expression insofar as it is not externally preconditioned but internally motivated by some personal interest which results from or originates in an act of choice by the speaker/communicator from a set of possible communicative acts

Restricting our interest to the verbal means of expressivity, we can easily identify additional self-expressive choices of a more general yet not necessarily emotional type A brief account of some of the most relevant linguistic devices of self-expression difficult to avoid is provided by Benveniste (1966) There are, first of all, the first person and the second

person singular of the personal pronoun (I mAyou)

C'est dans une realite dialectique englobant les deux termes et les definissant par relation mutuelle qu'on decouvre le fondement linguistique de la subjectivite (Benveniste 1966: 260)

[It is due to the dialectic reality encompassing the two terms and due to defining them through mutual relationship that one discovers the linguistic base of subjectivity, (translation A.H.)]

Then there are the deictic spatial and deictic temporal expressions and all the speaker-dependent expressions of modality, of which Benveniste tackles the verbs that perform intellectual operations in the first person singular

present - je crois -1 think, je suppose -1 suppose

Finally, he considers expressions like je jure, je promets, the enunciation of

which he characterizes as accomplishment, an act, not just the description of

an act

Another instance of expressive choice inherent in any speech act pertains

to the social meaning of the code used From the codes available to her/him, the speaker has to select some particular subcode and within this subcode the particular segment s/he actually uses Since a speaker has to choose

Contextualizations 5

constantly, s/he constantly provides the hearer with some detail about her/himself Her or his selection is expressive in that we can infer from it some personal details concerning, for example, her/his sociocultural background, her/his knowledge of the subject, her/his interests and concerns.7 This outlook fits into the more comprehensive approach advocated, among others, by Halliday and known as a "system network"

A system network is a theory of language as choice It represents a

language, or any part of a language, as a resource for making meaning by

choosing (Halliday 1985: XXVII)

It follows then that in some way or other a speaker's utterance may always allow an expressive interpretation Therefore, Watzlawick et al's (1967: 51)

famous bon mot that one cannot not communicate (in the loose sense, that is) can, in principle, be well supplemented by a further bon mot stating that one

in greater detail

Si le langage n'est pas une creation logique, c'est que la vie dont il est

l'expression n'a que faire des idees pures Si Ton me dit que la vie est

courte, cet axiome ne m'interesse pas en lui-meme, tant que je ne le sens

pas, tant qu'il n'est pas vecu; cette idee generale ne penetre reellement

en moi que par une modification subjective accompagnee d'une vibration

affective, si legere soit-elle, et cela n'est possible que si, par des

associations simples ou complexes, peu importe, je pense ä ma vie ou ä

celle d'autres personnes impliquees dans mon existence (Bally [1965]:

15)

[If language is not a logical creation, it is due to the fact that life, whose

expression it represents, does not simply produce pure thoughts: If I am

told that life is short, I am not interested in this idea unless I feel it,

unless I experience it This general idea takes hold of me only insomuch

as I am subjectively affected by it; it is accompanied by an emotional

vibration of whatever intensity, and this is only possible if I - no matter whether by simple or complex association - think of my own life or other lives which are intimately related to mine, (translation A.H.)]

Among his illustrations we find the following:

Le jugement intellectuel La terre tourne se change en jugement de valeur dans la bouche de Galilee s'ecriant devant ses juges: Ε pur si muove\ Ce

n'est plus une verite scientifique, c'est l'affirmation d'une valeur attachee a cette verite: eile paratt si precieuse ä celui qui 1'emet, qu'il risque sa vie pour eile (Bally [1965]: 15)

[The logical judgement The globe revolves becomes an evaluative

judgement in the mouth of Galileo screaming into the face of his judges

Ε pur si tnuove! Then it is no longer a simple scientific truth, rather it is

the affirmation of a value attached to this truth: to him it appears so precious that he risks his life for it (translation A.H.)]

And he elaborates the idea further:

Celles de mes pensees qui germent en pleine vie ne sont jamais d'ordre essentiellement intellectuel; ce sont des mouvements accompagnes d'emotion, qui tantöt me poussent vers Faction, tantot m'en detournent;

ce sont des epanouissements ou des repliements de desirs, de volitions, d'impulsions vitales Sans doute, c'est par l'intellect que je prends conscience de ces mouvements multiples, mais il n'en forme pas l'essence, il n'est que le vehicule, le metteur en scene et le metteur en oeuvre (Bally [1965]: 15-16)

[Those thoughts which sprout in the fullness of life are never of an essentially intellectual make-up; they are movements accompanied by emotions which sometimes lead me to and sometimes detain me from actions; they are releases or contractions of desires, volitions, vital impulses Admittedly, it is only through my intellect that I become aware

of these multifarious movements, yet it does not represent the essence but is only the vehicle, the stage director, and the mechanic, (translation A.H.)]

Such a body of thought finds a contemporary echo in Lyons' (1982: 107) distinction "between die subjective experiencing self and the objective observing self' and, in general, has been vigorously revived within the newly established cognitive paradigm

Neurological research has well established the possibility of the brain's functioning in either of the two modes (cf., for example, Damasio (1994)

Contextualizations 7

with his case studies on brain-damaged people as well as McNeill (1992) and the literature he refers to there) In the field of linguistics, Langacker's theory is framed by a similar distinction applying to what he calls

"perspective", a term which subsumes "such factors as 'orientation' and 'vantage point'" (Langacker 1990: 315) His key terms are "subjectivity" and "objectivity" and pertain to the way in "which the conceptualizer construes a particular entity or situation" (Langacker 1990: 316) These notions are understood in a narrow sense and do not embrace meaning constituents such as PERSONAL and IDIOSYNCRATIC or IMPARTIAL and BASED ON EVIDENCE respectively

The values that I will henceforth impute to them are best illustrated by a

simple perceptual example Consider the glasses I normally wear If I

take my glasses off, hold them in front of me, and examine them, their

construal is maximally objective, as I will understand the term: they

function solely and prominently as the OBJECT OF PERCEPTION, and

not at all as part of the perceptual apparatus itself By contrast, my

construal of the glasses is maximally subjective when I am wearing them

and examining another object, so that they fade from my conscious

awareness despite their role in determining the nature of my perceptual

experience The glasses then function exclusively as part of the SUBJECT OF PERCEPTION - they are one component of the perceiving apparatus, but are not themselves perceived (Langacker 1990: 316)

Though derived from perceptual aspects of cognition - and in this respect similar to Werner's approach -, Langacker believes "that the notion of subjectivity/objectivity is applicable well beyond the perceptual sphere, narrowly interpreted" (Langacker 1990: 317) Since Langacker's theory does not center, however, on this dichotomy, I will use Bally's terms instead

Bally's dichotomy applies to verbal behavior in general It can therefore be linked to the functional typologies by Bühler and Jakobson, resulting in a cross-classification This allows us to treat utterances more adequately on two different levels: by defining their communicative function and by determining the mode of realizing it For the three functions which the proposals by Bühler and Jakobson have in common, the interaction with mode can be illustrated by table 1.1 represented on the following page

The appellative/conative function, pertaining to some state of affairs still to

be achieved, only conveys the pursuit of an action goal when realized in the

mode pur, as in

(1) Stop talking

Table 1.1 A systematics concerning modes and functions

When the representative/referential function, pertaining to some given state

of affairs, is realized in the mode pur, it acquires the character of a report (3) It was extremely hot

When realized in the mode vecu, it is experiential in that the speaker at the

moment of utterance emotionally "re-lives", as it were, the heat referred to

(4) It was so hot!

In the expressive/emotive function, we can also expect both modes to apply

While the mode pur consists of topicalizing an inner state through

referentially appropriate descriptions resulting in a report of one's inner

sensations, as in (5), the mode vicu results in a live performance of the

(7) He couldn't get the handle fixed, which was maddening

(8) Unfortunately, Irene and Arthur separated

reduced it appears, the more it moves toward the mode vecu And yet, even

the most reduced form, i.e., the attitudinal disjunct, as, for example,

unfortunately in (8), preserves an element of identifying explicitness that

excludes it from being considered a mode vecu expression

I have elaborated and illustrated the two modes on a purely verbal level; and I will continue to apply it to verbal utterances only In natural face-to-face interaction, the nonverbal resources used or available would have to be taken into consideration as well in order to arrive at a valid assessment of a natural utterance in terms of Bally's dichotomy But even if it may turn out

in many cases that the nonverbal constituents of an utterance play a crucial

part in that they add a mode vicu dimension to an utterance linguistically made up in the mode pur, it makes sense, nonetheless, to apply the notion of

mode to the verbal part only; after all, the written form of communication also calls for means to meet similar expressive needs

In conclusion, the following examples illustrate the scope of emotional expressivity taken into account It embraces all a- and b-versions, but

excludes the c-versions The a-versions represent mode pur forms of

expressivity in that they either topicalize an emotion (10) or identify an emotional attitude toward a state of affairs given or envisioned (11/12); the

b-versions of (10) - (12) represent the mode vecu variants of the same

emotion and emotional attitude respectively; the c-versions of (11) and (12),

on the other hand, show no emotion at all, even if - in line with previous considerations (cf p 4) - an emotion may be presupposed, as in (12c)

(10) a I could burst with rage

b Damn it!

(11) a It was maddening, but he could not get the handle fixed

b He couldn't get that damn fucking handle fixed

c He couldn't get the handle fixed

(12) a Fix that handle; that's a serious order

b Will you fix that damn fucking handle immediately, bastard!

c Fix that handle!

1.2.1.1 A pragmatic and interactional perspective

Expressing one's emotions is linked to the pragmatic and interactional dimension Utterances that topicalize emotions (10a) or express them for their own sake (10b) can certainly count as speech acts Whether emotional

attitudes toward propositional state of affairs - be they expressed in the mode

pur (11a and 12a) or the mode νέοι ( l i b and 12b) - can equally be

characterized in speech act terms is dubious Austin himself was undecided

as to how the "evincing" of an emotion should be accommodated in his speech act theory ([1975]: 104) Expressing one's emotional attitude toward

a propositional state of affairs is certainly something that a speaker "does", but s/he does this in connection with something else, he says, and therefore,

it does not appertain, intuitively at least, exactly to either the locutionary (concerning the content of an utterance), the illocutionary (concerning the communicative function of an utterance), or the perlocutionary (concerning the communicative effect of an utterance) aspect of a speech act.10

Another difficulty is more disconcerting It concerns a related concept, namely the intentional base of emotional expressions At least the expression

of emotional attitudes in the mode vecu can be spontaneous and unconscious,

very much like the greater part of nonverbal behavior and not unlike Freudian slips One may, however, argue that under normal circumstances expressive behavior can hardly be conceived of as dissociated from the interpersonal dimension, extant wherever somebody else is present, and its regulative impact In other words, even behavior, when observed or observable, should in principle be understood as having been subjected to (social) mediation, a conclusion which would assign to it a communicative base half-way between the involuntary (unconscious) and the voluntary (intentional) - should we call it "social instinct"?

Even such a compromise might not settle the issue, though What can be held against it is the following line of thought Living one's emotional attitudes, as I have said, is (like any other behavior in the sense qualified above) not conceivable under normal circumstances without assuming that it

is socially mediated, which means that it is in some way checked as to its social acceptability This checking can be described as "face work", a socio-psychological concept promoted by Goffinan (1959) and successfully introduced into linguistic analysis by Brown - Levinson (1978) Applying this concept to the topic currently under discussion, it is important to stress the following points "Face" refers basically to an individual's public self-image (delineated in terms of widely approved social attributes), while "face work" refers to the efforts of the participants in a social situation directed toward sustaining their own as well as the others' faces Take, for example, the nonverbal expressive behavior of a pain-ridden patient at the dentist's

Contextualizations 11

office when s/he groans though s/he might have wanted to scream Once on this track, we might wonder whether one should not interpret even screaming, unchecked as this behavior may appear, in the same vein; screaming, it could be argued, is ultimately not really unchecked but results from discarding existing social norms The face-concept is based on the assumption that a speaker avoids direct insults or confrontations by employing a wide range of "politeness strategies" Expressivity certainly does not represent such a strategy; on the contrary, it often runs counter to this idea A speaker who decides nonetheless to live her/his emotions therefore decides against such politeness strategies It is due to this element

of decision that expressive behavior appears to be reconcilable with the concept of intentionality

1.2.2 Emotions and expressions

Emotions abound; but what are the possibilities of naming or expressing them? What emotions is a language able to take account of? Psychologists interested in the phenomenology of emotions are usually content to identify

a number of basic emotional states comprising all emotions.11 Hastorf et al (1970: 22), for example, reproduced a diagram by Schlosberg (1952) attempting such a comprehensive classification Diagram 1.1 on the following page is a reproduction in a slightly adapted version

The circle contains different classes of emotional states, while the two intersecting axes represent their underlying basic dimensions Thus REJECTION is identified as the underlying dimension of CONTEMPT and DISGUST, and its reverse value, ATTENTION, as the underlying dimension of SURPRISE on the one hand and of FEAR and SUFFERING

on the other Similarly, ANGER and DETERMINATION have as their underlying emotional dimension UNPLEASANTNESS, while the underlying reverse value of PLEASANTNESS is coordinated with LOVE, MIRTH and HAPPINESS

Among (language) philosophers, we discover a similar tendency toward global reductionism Aristotle's writings on the subject are still relatively elaborate, proposing the following scheme (cf Eggs 1990: 51):

A Fields of emotional stirrings (valid for all beings; inarticulate/diffuse, non- conscious)

1 anxiety/confidence

2 being repelled/being attracted

3 pain/lust

Diagram 1.1 Dimensions of emotions

B Basic affects/Fields of affects

1 fear courage(human variant of A 1)

2 antipathy sympathy (human variant of A 2)

3 grief joy (human variant of A 3)

C Social primary affects

sind ihrer Qualität nach stets entweder ein Lieben (eine Hinneigung)

oder ein Hassen (ein Abgestossenwerden) resp ein Vorziehen oder Nachsetzen, das aber verschieden modifiziert erscheint je nach der

Materie, auf welche das Eine und Andere sich bezieht (Marty [1965]:

199-200)

[ are, qualitatively speaking, always either loving (being attracted) or hating (being repelled), elevating or subordinating, which, however, appears to be differently modified depending on the matter to which either one pertains, (translation A.H.)]

Contextualizations 13

Among the latest contributions to a systematization of emotions that has been offered from the linguistic side is Fries (1996) He proposes a three-

dimensional model, synthesized from mainly psychological literature, which

is supposed to accommodate all emotions imaginable The three dimensions are (a) the positivity/negativity of affects, (b) their intensity and (c) the emotional proximity which an individual expresses toward the object or state

of affairs (cf Fries 1996: 56-59)

Talking about the wealth of emotions in as global a manner as Marty does

is not always a well deliberated measure to efficiently abbreviate or summarize, but sometimes rather the necessary consequence of a very mundane fact: there is no language which fully provides the lexical means necessary for denoting the innumerable shades of all possible emotions.12 As Langer (1955) puts it:

We have words for general kinds of emotion such as anger, fear, erotic

passion, regret; but there are hundreds of angry reactions, hundreds of

different loves Some sad experiences and some happy ones are

amazingly alike in feeling- pattern The unfolding of even so simple a

phenomenon as an emotion defies any verbal account, as I think the

brave efforts of the phenomenologists show We usually sum up the

character of an emotion by describing some condition in which it would

occur, or some overt response it might incur, for instance: "It was as

though the bottom had dropped out of everything", or "I could have

wrung her neck" (Langer 1955: 6)

It is the limits of the lexicon that assign a superiority to the expressive

resources of the mode vecu (That it may also be tedious if emotion were continually topicalized in full goes without saying.) The mode vecu,

therefore, attracts particular attention

In natural face to face interaction, a speaker can "live" in this mode whatever emotional nuance s/he is psychically capable of, thanks to the rich array of nonverbal and paralinguistic means (cf Knapp 1972, Ekman 1984; Crystal 1969).13 Within language proper, however, the (non-conceptual)

devices available for expressing emotions in the mode vecu are more

restricted Besides Fries (1996) and Quirk et al (1985: 1414-1419), Volek (1987: 15-25) provides a particularly useful survey Apart from phonetic/phonological means, she refers to morphological tools (e.g.,

diminutives, as in German Mann - Männlein 'little man' and augmentatives,

as in Italian bacio - bacione - 'big kiss'), to lexical signs such as inteijections

(alas!), lexical metaphorizations (son of a bitch) or connotation-loaded

lexemes (What are you jabbering/talking about?), to syntactic devices such

as subjective word order (Where did that bitch disappear to, I would like to

know instead of I would like to know where that bitch disappeared to),

characteristic sentence-types (such as nominal sentences, e.g., A man!), or

interruption of utterances, and finally to features of a textual character (consisting mainly of various types of juxtaposition between two contrasting phenomena, e.g., bookish words in colloquial discourse)

The means mentioned above do not all have the same standing as expressive devices Some may be said to regularly express an emotional attitude; others, however, do so only occasionally Lack of regularity seems

to occur mainly when the means cannot be claimed to be primarily expressive To illustrate this point, reference shall be made to some means of intensification tackled as expressive devices by Mathesius (1964 and 1975)

Under this cover term he subsumes comparisons (e.g., Er hat wie ein Klotz

geschlafen 'He slept like a log'), repetitions (asyndetic as in She is very very sweet or syndetic as in She cried and cried as well as reduplicative forma-

tions such as day-daily), certain compounds, for example zuckersüß sweet', the grading suffixes -er and -est (as in bigger - biggest), grading adverbs (as in schrecklich heißes Wasser 'terribly hot water'), and graded lexemes (e.g., to roar vs to shout) It seems, however, that at least the

'sugar-lexematic, adverbial and suffixal means of intensification cannot generally

be interpreted as expressive It is certainly true that all biased qualifications, those deviating from what Bolinger (1977: 28-29) calls the "expected average", are somewhat bound to the speaker's subjective evaluation and thus likely to evoke an emotional attitude, but it seems at least doubtful to equate qualitatively biased expressions with emotive expressions; the biased quality expressed can just as well stand for a proper rational/intellectual act Which interpretation is to be preferred depends on whether the biased expression at issue can be considered to be the result of cognitive reasoning

or not (cf Toulmin 1958 and Frankena 1958).14 Even less reliable are findings based on investigations that appertain to the field of stylistics, be it

on the level of a language system (cf Aronstein 1924) or on the level of use (cf Charleston 1960) The characteristics that, in my view, qualify a language phenomenon as a "real" means for expressing an emotion in the

mode vecu will be expounded in the section to follow

1.3 Leading concepts

It will not be surprising, after what has been said above, that the grammatical

devices to be tackled in the following chapters will turn out to be mode vecu

devices of expressivity Furthermore, they will all be shown to express

Leading concepts 15

emotional attitudes toward propositional states of affairs The expressions

involvement and attachment will function as terminological variants

The way in which mode νέοι devices express a speaker's emotional

involvement in (or attachment to) a propositional state of affairs differs from

how a lexeme like essay expresses the notion ESSAY Whereas in this latter

case the meaning is symbolically signaled, the emotive meaning of a grammatical device is only accessible through a reasoning induced by a (grammatical) form that, due to certain circumstances, acquires the character

of an index

What then are the circumstances that may make a grammatical device function as an indexical sign pointing toward an emotive meaning? For the cases under consideration it is the fact that in each instance there exists an

alternative; this gives the speaker an option; s/he has got the free choice, free

in the sense that the choice does not have an influence on the propositional content Here, a phenomenon seems to be effective that Clark (1990) has called the pragmatic "Principle of Contrast"

The PRINCIPLE OF CONTRAST, in its most succinct form, goes as

follows: Every two forms contrast in meaning This principle has been

assumed, tacitly or explicitly, at least since Breal (1892) and Paul (1898)

in their work on language change, as well as by de Saussure (1919/1968)

in his synchronic analysis of language use This principle captures the

insight that when speakers choose an expression, they do so because they

mean something that they wouldn't mean by choosing some alternative

expression Speaker choices in any domain mean what they do in part

because they contrast with other options both in that domain and in the

language as a whole As a result, speakers do not tolerate synonyms in

language This principle applies to words, affixes, grammatical functors,

and even constructions (Clark 1990: 417)

Moreover, the forms under investigation are "costlier", than their alternative, i.e., the forms literally consist of more language material Now, if there is any truth at all to the motto that more form/substance implies more meaning,

we have to assume a meaning surplus for the forms in question At the same time, the costlier form appears as marked in comparison to the other one.15

Both characteristics, in my view, help to establish the "indexicalness" of the forms; we are required to look for the meaning surplus "outside"

What will turn out to be the constant feature throughout all text environments in which the grammatical devices at issue are used is that the

propositional content can in one way or another be considered remarkable

Remarkableness then would (plausibly) provide the cause for a speaker to become emotionally involved Consequently, the choice of some such marked form is interpreted as pointing to the speaker's attachment.16

We will distinguish between two types of attachment depending on the pragmatic context in which a speaker gets involved Talking about events or states for their own sake will represent one prototypical context, discussing events or states for current communicative or practical purposes the other; these two contexts more or less coincide with what Weinrich ([1985]) calls

Erzählen and Besprechen In the first context the speaker's attachment can

be considered adaptive, because s/he empathizes and sympathizes with whom- or whatever s/he is talking about, in the second context it is adoptive,

because her/his involvement is mediated through and utilized for her/his own interests

The reinterpretation of grammatical categories and constructions in emotive, non-propositional terms is not isolated from recent theoretical developments in the field of linguistics and with their stimulating findings

Of particular significance is Traugott's (1989) theory of grammaticalization

It tries to establish diachronically the view that grammatical elements often evolve from lexical sources, bringing about a shift from propositional to textual and attitudinal meaning; since this facilitates a more personal way of talking, it has become known as the process of subjectification The analyses

to follow will basically substantiate this view, but cast some doubt on Traugott's (1990) claim as to the unilinear (and thus irreversible) order of this developmental pattern

Since the six grammatical phenomena are taken from various periods in the history of the English language, they may eventually turn out to be bench marks in an outline of expressivity/subjectivity as a cultural-historical phenomenon

2 The so-called possessive dative

2.1 Plan of discussion

In this chapter the Old English phenomenon of the possessive dative will be discussed with regard to its meaning potential As compared with two almost synonymous forms, the possessive pronoun and the possessive genitive, its meaning is taken to go beyond the mere expression of possession This so-called meaning surplus has been topicalized by a number of historical linguists We will show that their assumptions cannot sufficiently explain and motivate the occurrences of this Old English construction As an alternative we claim that the meaning of the possessive dative is the indication of the speaker's attachment to the person referred to by the dative and substantiate the claim by an analysis of the possessive datives occurring

in Beowulf

2.2 Introduction

In present-day English, we encounter various types of indirect objects They can be described as external and internal dative, a distinction that pertains to their syntactic function, as in (la) and (lb), as well as their semantic function, as in (2a) and (2b) (cf Wierzbicka 1988)

(1) a She gave a book to her friend

b She gave her friend a book,

(2) a Sally showed a picture to Mary

b Sally showed Mary a picture

Hence, syntactically, the internal dative refers to the indirect object, if this is positioned between verb and direct object (as in (lb) and (2b)) The external dative refers to the indirect object in postposition to the direct object; i.e., to the prepositional phrase in (la) and (2a) Semantically, the internal dative construction expresses that the speaker's attention is focussed on the effect

of the action produced on the target person, i.e., the person designated by the internal dative (Wierzbicka describes an example such as (2b) to express that

"the agent does something to an object to enable another person to see it"

(Wierzbicka 1988: 372)) The external dative, on the other hand, does not place more importance on the target person than on the patient (cf Wierzbicka 1988: 359-387)

A further but somewhat disputable type of indirect object can be encountered in phrases like

(3) She hit him on the head

From a strictly contemporary point of view, this example may appear out of

place; all too readily will the personal pronoun him be interpreted as direct

object, analogously to (3a), with the prepositional phrase functioning as a specification as to where she hit him

(3) a She hit him

And yet, from an historical point of view, it could also reflect an old use of the dative, which has subsequently become obsolete and is nowadays hardly recoverable, namely what is known as the "possessive dative" A brief reference to present-day German may help to get this point across Sentence (3) would have two equivalents:

( 3 ' ) Sie schlug ihm (dative) auf den Kopf

(3'') Sie schlug ihn (accusative) auf den Kopf

The formal unambiguity in present-day German, which has not undergone the tremendous changes that can be observed in the development of English, still reflects the situation we find in Old English With the dative and accusative having had distinct forms (which during the Middle English period were neutralized), we find side by side:

(4) He sloh hine (acc.) on paet heafod

He sloh him (dat.) on paet heafod

It may be due to the licencing by some special verbs of these two alternative constructions that the dative, otherwise lost, survived in a few instances where it could readily be apprehended as an accusative/direct object (cf Ahlgren 1946: 218) In other surviving cases, it is the idiomaticity of certain

phrases like to laugh (a person) to scorn = to deride or to run (a person)

through the body = to pierce which facilitates the interpretation of the

indirect object as direct object, in accordance with current usage Characteristically enough, the surviving cases all involve prepositional

Introduction 19

phrases; the variant with the dative in combination with an accusative object

is not permitted in present-day English:

(5) * He kissed her the cheeks

(6) * She knocked him the teeth in

Both constructions were equally common in Old English as shows in example (7)-(10)

(7) Him of hrceöre gewät \ säwöl

'The soul went from his breast'

Beowulf mo (quoted in Havers 1911: 275)1 7

(8) Him stod stincende steam of dam mude

'Stinking odour came forth from his mouth'

Aelfric Catholic Homilies p 86 (quoted in Ahlgren 1946: 199)

(9) hcefde Abrahame \ metod moncynnes breost geblissad

'the destiny of mankind had delighted Abraham's breast'

Caedmon Genesis 2922 (quoted in Havers 1911: 278)

(10) him pcet heafod wees adune gewended

'his head was turned downward'

Blickling Homilies p 173 (quoted in Ahlgren 1946: 200)

The examples given have one feature in common, which turns out to be class-constitutive: The dative is always (partly) co-referential with the accusative or the prepositional noun in that it refers to the person who possesses the object referred to by the accusative or the prepositional noun

As a possession indicating device, the dative stands in a relationship of semantic equivalence to the possessive pronoun or genitive; possessive dative and possessive pronoun/genitive are mutually interchangeable without affecting the propositional content of the sentence Thus, examples (11) and (12) can reasonably be considered equivalent

(11) feol him to fotum

'(he) fell at his feet'

Bede's Ecclesiastical History p 228 (quoted in Ahlgren 1946: 197)

(12) da feol he to his fotum

Bede's Ecclesiastical History p 372 (quoted in Ahlgren 1946: 197)

Still, upon comparing both versions one may wonder whether the dative, in addition to its possession indicating function, does not exert something else, something which is not conveyed by the possessive pronoun or the genitive One may acknowledge this and still be content with describing

the dative in purely possessive terms, considering the "rest" to be something accidental to the individual text and delegating its treatment to individual text analysis Or one may try to define the meaning surplus and establish it as systematically given The main proposals that have been advanced in this latter direction either ascribe an experiential value to the dative or subsume

it under the notion of beneficiary My own proposal aims at establishing the dative in question as a device by which a speaker expresses her/his (emotional) attachment to the person referred to by the dative and to the person's state of affairs But before expounding these ideas any further, I would like to discuss alternative positions first The data that these alternative positions use in their arguments will, at the same time, provide the main body of the data that I will use for expounding my own proposal

2.3 Dative and possession

The approach to the Old English dative under consideration in mainly possessive terms has found its most prominent advocate in Mitchell (1985).19 Having written a grammar rather than an argumentative discourse, Mitchell simply lists the various possibilities in which the dative can function as a means for expressing possession, with brief descriptive comments, but without any notable interpretative effort to confront the dative with the possessive pronoun/genitive A detailed discussion of this sort, however, can be found in Seiler (1983) Though Seller's treatise pays only occasionally attention to Old English, its general and universalistic make-up encourages an application to the specific problem under scrutiny After a comparative analysis of numerous ancient and contemporary languages from various language groups, Seiler proposes a scale of possession that summarizes the findings in a way such that the "observables" appear as "traces" of a "constructional process which took place and endlessly takes place in the human mind" (Seiler 1983: 72) The scale has the format shown in diagram 2.1 below

NN and Ν case Ν comprise the alternatives under consideration, NN

accommodating the use of the possessive pronoun (his house) and Ν case Ν the genitive (John's house) and the (Old English) dative (cf German Ich

habe mir das Bein gebrochen Ί have broken my leg' instead of Ich habe mein Bein gebrochen)

Dative and possession 21

Diagram 2.1 Scale of possessivity

to attribute predicative function to them The more predicative the manner is

in which possession is expressed, the higher is the degree to which this attribution appears to be informative and marked Conversely, the more determinatively the concept of possession is expressed, the less marked the expression becomes and the less informative (which means it becomes ever more semantically unspecific and empty) These characteristics work toward the functional principles of establishing possession versus positing it The difference is basically aspectual; in the first case, the distribution of property

is being propositionally established (This book is mine, that one is yours), whereas in the second case it is an established propositional fact (My book

was cheap, yours expensive) The more syntacticized, predicative,

informative and marked the means of expressing possession are, the more possession appears to be established; the more grammaticalized, determinate, unmarked and informatively empty they are, the more possession appears to

more syntacticized , more predicative , and more marked means The other force or tendency starts out from maximally explicit and syntacticized structures and works toward ever more grammaticalized, obligatory, morphological means of expression (Seiler 1983: 79)

Synchronically, the opposing forces come to the fore wherever there is a choice between different means of expressing possession, the best example being the temporary option between the possessive dative and the possessive pronoun in Old English

As to the diachronic perspective, we have to go farther afield In its historical dimension, Seller's view has certain aspects in common with the theories of grammaticalization and subjectification as advanced by Traugott, Langacker and others Reading his scale from right to left enfocuses grammaticalization as well, and Seller's interpretations of how possessivity

is conceptualized accordingly appear in part to be a specified version of Langacker's concept of subjectification (e.g., Langacker 1990) Seller's approach, however, seems more generally conceived than what is currently under consideration He takes into account different types of linguistic material when talking about grammaticalization, whereas Traugott - König (1991) focus on the process a single linguistic element undergoes in the course of time It is in this context that they emphasize the unidirectionality

of historical processes:

"Grammaticalization" refers primarily to the dynamic, unidirectional historical process whereby lexical items in the course of time acquire a new status as grammatical, morpho-syntactic forms, and in the process come to code relations that either were not coded before or were coded differently (Traugott - König 1991: 189)

Seiler would not commit himself to any unidirectionality It is certainly Seller's approach that fits the devices found in Early English on the diachronic axis

The eventual extinction of the possessive dative and its full substitution by the possessive pronoun/genitive exemplifies the second force referred to by Seiler in the quotation given above;20 the historically subsequent development to occasionally use, instead of the possessive pronoun, the

analytic form of + the personal pronoun exemplifies his first mentioned

force,21 for instance the use of me instead of my as in the fare of me (cf

Dative and possession 23

such as (a) the house which I own, (b) the house where I live or (c) the house

which I like, etc In accordance with Seiler (1983: 40), we consider my house

to be a possessive expression only in the meaning of (a)

Traditional grammars usually draw a tight distinction between possessive nouns that are inalienable from the possessor and those that are alienable Seiler uses the term "inherency" and interprets it not statically, but dynamically, proposing a "scale or continuum of bondedness between possessor and possessum" (Seiler 1983: 68)

Kin terms and/or body terms seem to range on top, albeit for different

reasons Part/whole seems to be a derivative or echo of body part, and

thus of weaker bondedness Cultural manifestations [mental cultural concepts, such as WORD, THOUGHT, CHARACTER etc., A.H.] seem

to be more inherent than cultural implements [physical products of cultural activities, such as ships, shields, houses, A.H.] (Seiler 1983: 68)

Although the Old English material I am using covers the whole continuum

of bondedness, it seems advisable to restrict the attention to the best instances, i.e., to examples where the noun shows an inherent relationship to the possessum In effect, we shall limit ourselves to examples where the dative or the possessive pronoun/genitive is in construction with a noun designating a body part or some cultural manifestation, such as is the case in the examples (13) - (16)

(13) pä bendas pä him on banum lägon

shackles were on his limbs'

Caedmon Daniel 435 (quoted in Havers 1911: 278)

(14) pcet he on Biowulfes bearm Megde

'which he laid in Beowulf s lap'

Beowulf 2195 (also quoted in Havers 1911: 275)

(15) weard me on hyge swä leohte

'it got so light in my mind'

Caedmon Genesis 676 (quoted in Havers 1911: 278)

(16) pä his möd ahlög

'then his heart laughed'

Beowulf 731 (quoted in Havers 1911: 276)

How much can Seller's rich account contribute to our approach to a rather limited topic? Apart from a very insightful contextualization of our problem into a general framework, the main benefit seems to consist in enabling us to describe in structural terms some of the differences that hold between the possessive expressions under consideration: Whereas the possessive pronoun

is unequivocally adnominal - i.e., it contracts a grammatical relationship with the noun to which it also enters into a possessive relationship - the dative is ambiguous - it is adverbal and adnominal at the same time, i.e., it contracts

a grammatical relationship with the verb and enters into a relationship of inherent possession with the noun

At first sight, the possessive genitive does not seem to fit in well, because

as a case it should be assigned an adverbal status analogous to the dative whereas intuition would rather group it together with the possessive pronoun

-as adnominal But the genitive is particular -as a c-ase In line with Jakobson (1971: 37-44),22 Seiler characterizes it as a (predominantly) adnominal case (1983: 40), and interprets it as a form "marked for quantification, i.e., it focuses upon the extent to which the entity takes part in the message, by implying that the extent is not total" (Seiler 1983: 39) In other words, the

genitive in the boy's toys functions similarly to the quantifier thousand in the

thousand toys since both reduce the quantitative extent to less than the total

We could enhance Seller's reference to Jakobson by adding that at one point, Jakobson talks about a specifically adnominal genitive where the noun extracts from the entity set in the genitive something that, for example, may

be a property such as BEAUTY Thus in a phrase like the girl's beauty,

beauty refers to the property that is extracted from the entity GIRL

Such structural characterizations, however, do not provide a good motivation for the co-existence of these two form classes for expressing possessivity, viz adnominal (possessive pronoun and genitive) and adverbal (dative) That they co-exist for a certain period simply because it usually takes some time before a new form fully substitutes for its predecessor goes without saying Rather, the question concerns the speaker who had the choice between two form classes that appear to be functionally so similar that one may wonder what may have induced her/him to use - as the literature amply evidences - both, in alternation, at times even together, as in the following example:

(17) & let him pytan ut his eagan & ceorfan of his handa

'and let his eyes gouge out and his hands cut off

Saxon Chronicle 796 F (quoted in Ahlgren 1946: 200)

Is such a use simply arbitrariness or is there a hidden consistency? If the latter option holds to be true, we must then ask: What is the principle that renders this usage consistent? Assuming the former option, we are not confronted with easier questions: Why should a speaker choose a more complex expression of possession (dative) - or even a double form -, if the simple possessive pronoun/genitive does exactly the same job? Or (applying other concepts from Seller's analysis), why does a speaker use a marked form and a semantically explicit one, if the inherent possessive relationship

Meaning surplus 25

between possessor and possessum does not need to be established because it can be taken for granted? All of these questions remain unanswered if we stop at possession

2.4 Meaning surplus

2.4.1 Experience or interest?

To do justice to Seiler, we must add that he does go beyond possession, although rather ephemerally, because it is outside his sphere of interest Characterizing the dative as contracting a twofold relationship, he states, in

reference to the German example Ich habe mir das Bein gebrochen Ί have

broken my leg', that the relationship with the verb "indicates that the agent

is at the same time the experiencer" (Seiler 1983: 44) Yet, it is not clear whether this interpretation is based merely on the reflexive structure of the

example - what about Ich habe ihm das Bein gebrochen Ί have broken his leg'?, where ich Ί ' can actually be interpreted as the agent, which is not

necessarily and perhaps not at all the case with the reflexive structure - or whether it can be generalized

Be that as it may, a similar stance, yet with a decidedly general outlook, is advocated by Havers (1911) He suggests that the dative in question

ist subjektiver, wärmer, innerlicher als der Genitiv [and as we may add,

the possessive pronoun], der einfach objektiv ein Besitzverhältnis konstatiert (Havers 1911: 2)

[ is more subjective, warmer and more inward than the genitive which

simply affirms an objectively given possessive relationship, (translation A.H.)]

Taking this account as a point of departure, he calls the dative "dativus sympathicus" and maintains that it expresses " the inner participation of the

person affected by the verbal concept" (Havers 1911: 2, translation A.H.) I

have typographically emphasized the part of the quotation that shows that it

is obviously the person (referred to by the dative) who is said to be inwardly participating, thus bringing to the fore her/his role as experiencer The inner participation forms part of the content of a sentence just as her/his role as possessor does.23

Havers, whose interest in the dative is not confined to Old English alone but extends to Indogermanic languages in general, uses a classification that,

in part, is compatible with Seller's continuum of inherency From his ample list of Old English quotations referring to BODY (equivalent to Seller's body-parts) and SOUL (equivalent to Seller's cultural manifestations), I will draw upon three examples for each of the two groups (see (18) - (20) and (21) - (23) respectively)

A (18) him hilde-gräp bän-h üs gebrcec

'my warlike grip broke open his bone-house'

Beowulf2508 (quoted in Havers 1911: 274)

(19) midpcere gyrde pe he him on honda hcefde

'with the rod he had in his hand'

Bede's Ecclesiastical History p 228 (quoted in Ahlgren 1946:

197)

(20) Him of eagum stöd leoht unfceger

'From his eyes came a light not fair'

Beowulf727 (quoted in Havers 1911: 274)

Β (21) him wees geomor sefa

'his mind was sad'

Cynewulfs Christ 499 (quoted in Havers 1911: 279)

(22) him wJf-lufan cölran weoröaö

'his wife-love will become cooler'

Beowulf2066 (quoted in Havers 1911: 275)

(23) leetpe äslüpan sorge ofbreostum

'let your sorrow vanish from the breast'

Caedmon Genesis 2796 (quoted in Havers 1911: 278)

Though the meaning category EXPERIENCER certainly applies to the datives in (18) and (21), its applicability becomes increasingly more problematic as we move to the second and third examples within each group How could the experience that would have to correspond with having a rod

in one's hand (example (19)) be elucidated? And for (20), we would have to assume that the person has a self-awareness of his facial expression that is psychologically too refined for the Old English period A similar argument can be made against the experiential24 interpretation in (22), whereas in (23), the proposed meaning does not apply at all, due to the imperative which projects, but does not describe, a state of affairs It is only within the descrip-tive context that the category of experiential meaning seems to make sense

We therefore have to dismiss the claim that the dative's relationship to the verb could be indicative of the experiential involvement of the person that

forms part of the proposition - as Mitchell (1985: 565-568) does vis-a-vis

some other examples not permitting such an interpretation Instead, Mitchell speaks of the "dative of interest", but he is very careful not to fully commit

Meaning surplus 27

himself to this view Rather, he presents it as "the original function of the 'dative proper'" (Mitchell 1985: 565) Also Traugott states that in Old English with a number of verbs the dative is commonly assigned to NPs expressing the affected or interested person (cf Traugott 1992: 203) For the construction under discussion she uses the term "dative of interest", "where possession is metaphorical rather than actual" (Traugott 1992: 206)

(24) Lcecedemonia hcefdon him to ladteowe cenne wisne mon

The Lacaedemonians had a wise man as their leader

' The Lacaedemonians' leader was wise.'

Or 3 1.96.27 (quoted in Traugott 1992: 206)

One could arrive at a similar stance by drawing on Jakobson (1971) and his analysis of the dative According to him, the dative, like the accusative, indicates that the entity is affected by the action of the verb; but in contrast

to the accusative, which is a central case, the dative is peripheral As a result,

we see a comparatively less intimate connection between the dative object and the action at issue (Jakobson 1971: 52)

Jakobson's characterization can well be applied to the concept of the dative

of interest, where interest is considered to denote a weaker form of being involved in/affected by the action Thus the dative in

(3) She hit him on the head

can indeed be interpreted as assigning to the person referred to (besides the role of a possessor) the marginal role of a beneficiary (if this expression can serve as cover term for either positive or negative readings) - marginal because the action is directed toward one of his body parts

It is therefore not very convincing that Havers (1911: 277) and Seiler (1983: 44) seem to exclude interest even as a possibly valid reading for the meaning surplus of the possessive dative Undoubtedly, the dative of interest has a wide scope, and one may see a better example of it in a sentence like (25):

(25) She gave him a present

But putting (25) side by side with

(26) She gave him a blow

and comparing this example in turn with (3), we realize that the distinction between the various readings of the dative is a matter of degree with no

clear-cut dividing line

For the verification of this proposal on a wider scale, I will return to the same set of examples with which I tested the hypothesis of a dative of experience (19) and (23), both of which could not be interpreted in experiential terms, are easily adaptable to the present proposal

(19) mid pcere gyrde pe he him on honda hcefde

'with the rod he had in his hand'

Bede's Ecclesiastical History p 228 (quoted in Ahlgren 1946:

197)

(23) Icet pe äslüpan sorge ofbreostum

'let your sorrow vanish from the breast'

Caedmon Genesis 2796 (quoted in Havers 1911: 278)

(18) and (21) are likewise adaptable

(18) him hilde-gräp bän-h us gebrcec

'my warlike grip broke open his bone-house'

Beowulf250% (quoted in Havers 1911: 274)

(21) him wees geomor sefa

'his mind was sad'

Cynewulfs Christ 499 (quoted in Havers 1911: 279)

In (22), however, it is difficult to decide whether there is an interest implied and, if so, which kind (positive or negative) it may be

(22) h im wJf-lufan c ölran weoröaö

'his wife-love will become cooler'

Beowulf2066 (quoted in Havers 1911: 275)

And in (20) some such interpretation is altogether impossible

(20) Him of eagum stöd leoht unfieger

'From his eyes came a light not fair'

Beowulf 111 (quoted in Havers 1911: 274)

Thus, we find that Mitchell's cautious way of proposing INTEREST as the meaning element associated with the dative is all too justified: The notion of interest cannot be extended to the dative that contracts a possessive relationship with the noun

Meaning surplus 29

2.4.2 Attachment

The proposals so far discussed and ultimately discarded have one feature in common that my own proposal (cf p 20) does not contain The meaning categories of experience and interest are propositionally founded, whereas a speaker's attachment is a meaning category outside the proposition, though operating on it

It should be recalled that - in simplified terms - the proposition is only a part of the speaker's utterance (i.e., the naturally occurring speech unit of whatever size) The proposition is the abstract content of a sentence toward which the speaker takes an attitude Our concept of the speaker's attachment

is such an attitude It is "emotional" and thus different from Traugott's (1989) reading in epistemic terms; it is closer to the speaker's attitude in Halliday's (1985) and Traugott's (1982) sense of an "interpersonal" or

"expressive" component In contrast to this, the person's role as experiencer (cf Seiler 1983: 44) or beneficiary (cf p 27) forms part of the propositional content

My proposal starts out from the double relationship that the dative contracts, i.e., the relationship to both the verb and the noun (referring to the body-part); the sole meaning the approach assigns to the adverbial relationship is the function of setting a focus on the possessor as person.25

With the dative focus, everything the sentence communicates is conceived of

as concerning the person referred to by the dative

The bias toward the person referred to by the possessive dative is also evidenced by the fact that a sentence with a possessive dative is an apt form

of response to general questions that primarily express the inquirer's interest

in the person, such as What's the matter with him?, What happened to herl

or simply And hel Taking the modern German translation of example (18)

as a case for illustration, we indeed see that it is less acceptable to respond to some such questions, as given above, by uttering (18i) than by stating (18ii)

(18i) Seinen Leib zerbrach die Kampffaust

(18ii) Ihm zerbrach die Kampffaust den Leib

The focus that the dative marks is in fact considered a focus of interest set

by the speaker and thus different from Mitchell's concept of dative of interest, which relates to one of the propositionally given participants Moreover, it is not to be mistaken for the focus in the sense in which the theory of Functional Sentence Perspective (FSP) uses it In FSP, the focus is the informatively most important element of a message/ proposition In our context, the focus is the element that the speaker chooses for no other reason than a personal one.26

The reason for setting a focus is, I contend, that the speaker is attached to the possessor as person I have chosen the odd formulation "possessor as person" to convey the following idea: the attachment expressed by the dative pertains not to the person as such, but to the person who, as the possessor of

an entity that is directly affected by an action, is indirectly affected as well Not every action, though, in which the possessor/person may be affected is

of a kind that could trigger the speaker's attachment From this follows that

the dative of attachment - as we may now call it as well - is restricted to

states of affairs that, as such or in a given context, are in a positive or negative sense remarkable enough to arouse the speaker's interest and attachment.27 However, it does not imply, of course, that wherever a situation is remarkable (or rather, is described as remarkable) the speaker must necessarily articulate attachment to the person concerned

If s/he chooses to do so and does so by using the dative, s/he does not express it in explicit terms The dative, as a device for expressing attachment, rather functions, semiotically speaking, indexically Its use leaves nothing but a trace of the speaker's attachment, which in any given instance has, of course, a specific value, be it admiration, compassion, disgust, or whatever On the receiver's part, this requires an interpretative reconstruction: starting out from the conspicuousness of the dative in comparison to the possessive pronoun/genitive, working through the

"processing" of the data which the sentence and its context provide, and ending with an hypothesis on the speaker's specific emotional attitude

This task also represents the last step for fully establishing the proposal itself But before doing so (see section 2.4.2.1.), a few theoretical considerations seem to be appropriate

The preceding discourse aligns well with some ideas put forward by Traugott - König (1991) The authors link up the notion of indexicality with metonymy (1991: 210), extending the notion of the latter

from traditional concrete and overt contexts to cognitive and covert contexts, specifically the pragmatic contexts of conversational and conventional inference (Traugott - König 1991: 211)

They describe the process of metonymizing as "pointing to meanings that might otherwise be only covert" (Traugott - König 1991: 212) and define it subsequently, in analogy to metaphorizing, as "specifying one meaning in

terms of another that is present, even if only covertly, in the context"

(Traugott - König 1991: 212,1 have used italics for the part representing the feature specific for metonymy)

Traugott - König do not stop here, though Linking the concept up with grammaticalization, they establish, on the basis of their own analyses,

Meaning surplus 31

metonymy as another pattern (besides metaphor) by which semantic changes can be understood "Metonymy", they claim, "is largely correlated with shifts to meanings situated in the subjective belief-state or attitude toward the situation" (1991: 213) This also seems to hold true for the hypothesis advanced here, i.e., the rise of the usage of the possessive dative as expressive device when the possessive pronoun became a readily available means of expressing possessivity.28 The claim of the unidirectionality of such a process, which is connected with this approach, is not disproved by the possessive dative's extinction with the end of the Old English period This is different from what we could find for other phenomena of attachment

2.4.2.1 The use of the possessive dative in Beowulf

The choice of Beowulf? 9 no doubt, implies certain disadvantages for a purely linguistic approach, e.g., the limited generalizability of the findings or the constraints that the epic form may have imposed at times on the choice of one or the other form of expressing - in the first place! - possession But I do think that such drawbacks should not be over-emphasized The fact that medieval literature in general exhibits many features typical of oral literature (cf Ong 1984 and Brinton 1990) should alleviate the problem to some degree; and as far as the constraints inherent in the verse-form are concerned,

we may find some reassurance by presuming that - though a poet's artistry may succumb occasionally to some formal constraint - such constraints are unlikely to have been at work continuously - especially where consistency in usage emerges

This said, we may even claim that the analysis of a single document may also carry definite advantages with it I do not so much think of the purely quantitative load that an analysis of a representative number of items implies and that, of course, is reduced if the items derive from a single source I am rather alluding to a qualitative aspect: the confinement to a single document

of some length and particularly to one that contains a story, offers the opportunity not only to verify an hypothesis by showing that it indeed applies, but also to widen the insights into the topic and to study how the means can function in their wider context and to what ends The data from

Beowulf offer various opportunities for connecting their analysis with some

such more comprehensive perspectives

My approach presupposes that the dative is not simply another way of expressing possession My stance may in part be challenged by the