A Systemic Functional Grammar of French
Also available from Continuum: A Systemic Functional Grammar of Japanese Kazu...
264 downloads
1907 Views
11MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
A Systemic Functional Grammar of French
Also available from Continuum: A Systemic Functional Grammar of Japanese Kazuhiro Teruya A Systemic Functional Grammar of Chinese A Text-based Analysis Eden Sum-hung Li
A Systemic Functional Grammar of French
From Grammar to Discourse Alice Caffarel
With a foreword by M. A. K. Halliday
continuum
Continuum The Tower Building 11 York Road London SE1 7NX
80 Maiden Lane Suite 704 New York, NY 10038
First published 2006 © Alice Caffarel 2006 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN: 0-8264-6632-X (hardback) Typeset by RefineCatch Limited, Bungay, Suffolk Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall
Pour mes enfants, Yannick, Louis et Remi
This page intentionally left blank
Contents
Acknowledgements
viii
Foreword
ix
Introduction
1
1 Systemic functional theory as a metalanguage for description
4
2 The grammar of ideation (1): logical metafunction
20
3 The grammar of ideation (2): experiential metafunction
57
4 The grammar of negotiation: interpersonal metafunction
120
5 The 'enabling' grammar: textual metafunction
165
English—French glossary of terms in systemic functional linguistics
198
References
203
Acknowledgements
Thanks go to John Benjamins Publishing company, for permission to use material which first appeared: (i) in Ruqaiya Hasan and Peter H. Fries (eds), On Subject and Theme, A Discourse Functional Perspective (Amsterdam, 1995), pp. 1-49; (ii) in Anne-Marie Simon-Vandenbergen, Kristin Davidse and Dirk Noel (eds), Reconnecting Language, Morphology and Syntax in Functional Perspectives (Amsterdam, 1997), pp. 249-96.
Thanks also to Gunter Narr Verlag Tubingen, for permission to use material which first appeared in Eija Ventola (ed.), Discourse and Community, Doing Functional Linguistics (Tubingen, 2000), pp. 247-72. Many thanks go to Michael Halliday, whose ideas permeate this book, for his Foreword. I am greatly indebted to my teachers: Christian Matthiessen, who inspired me and continues to inspire me to interpret French grammar in systemic functional terms; Jim Martin for his ongoing support over the years; Claire Painter and Chris Nesbitt for their unique ability to explain complex phenomena. I am also indebted to Robin Fawcett for encouraging me to publish this work. Many thanks to my students and particularly to those who first applied this description of French grammar to analysing the multiple layers of meanings in texts and showed me that it worked. Many thanks also to my first-year French students in 2005 for keeping me going with their wonderful enthusiasm and a year's supply of coffee. My greatest thanks to Kathryn Tuckwell for her editing feat, to Marie Le Rouzic for spotting French oddities and to Charles Humblet for taming the figures. I would also like to thank the Department of French studies, my colleagues and the School of Languages and Cultures at the University of Sydney, as well as my friends and family in France and in Australia, for their support. Many thanks to my three wonderful children, Yannick, Louis and Remi, for their patience and all the cuddles!
Foreword
It is a daunting task for a linguist to present a grammatical sketch of a language that has been as intensively studied, and as comprehensively described, as modern French. There are certain contexts where we might expect to meet up with a descriptive overview of this kind, for example as a non-technical account of a major language written for 'the general reader', or as an introduction to a course for adults studying French as a foreign language; in such contexts we would expect to be offered a description in terms of the familiar categories of the European grammatical tradition. Alternatively the writer might be presenting a theoretical approach, still preserving the same descriptive framework but representing it in a new formalism with its own apparatus of terminology. What Dr Caffarel has produced is something that is different from any of the above. She is, to be sure, adopting a distinctive theoretical orientation, that of systemic functional linguistics; but this is not a new way of organizing and formalizing an already existing description. For Dr Caffarel the theory becomes a resource for conceptualizing the language as a whole, and hence for opening up a new perspective on the descriptive processes and the descriptive categories themselves. The main impetus for her opting to work within this theory is that it enables her to present a grammatical model that is particularly adapted to the analysis and interpretation of discourse, because any text that is under scrutiny is being located within the context of the underlying system of the language. The meaning- or, more often, the meanings - of any portion of the text will be derived from the overall meaning potential that inheres in the linguistic system. In the most general terms, the 'functional' aspect of the theory means that the concept of explanation is defined in functional terms: a category is explained by reference to its value along the various dimensions that characterize a language as a whole. So a grammatical category is validated across the strata, both 'from above' and 'from below': what does it mean? how is it expressed? and, critically, how is the tension between these two perspectives resolved? But this, in turn, involves the 'systemic' aspect of the theory, in which such content—expression pairings are located in their paradigmatic environment: what are the other possibilities? what is the network of opportunities that is being construed at this particular juncture? The
x
FOREWORD
grammarian's job is to define the parameters of the lexicogrammatical space within which any given feature becomes operational in the discourse. Dr Caffarel's strategy for organizing the grammar is as a 'journey through the metafunctional potential of French'. This is the deepest sense in which the grammar is said to be 'functional': the description is mapped onto the fundamental modes of meaning with which human language evolved. These are the 'metafunctions' of systemic functional theory: every language functions to construe human experience, of process and of the relations between processes; to enact human personal and social relationships; and to engender a flow of discourse which fuses and animates the two. In using this as her basic organizing concept, Dr Caffarel has produced a description which is refreshingly new without being forbiddingly revolutionary. Those familiar with systemic functional linguistics, and with descriptions of English or other languages in a systemic functional framework, will find it easy to move along in the traffic flow of her argument, with its clear delineation of the metafunctions and explicit reference to the stratal and instantial dimensions. At the same time, for those knowing French but unfamiliar with systemic functional research, it will serve as a thought-provoking invitation to the theory; in this connection, the glossary of French-English terminology at the end of the book is a particularly valuable addition. Finally, the book can be read in a typological context, as an expansion of Dr Caffarel's 'metafunctional profile' of French in the collection of papers on language typology which she herself edited along with J. R. Martin and Christian M. I. M. Matthiessen (Amsterdam: Benjamins, 2004). It is a pleasure to welcome this volume as a Continuum publication, and especially so if it turns out to be a precursor to comparable volumes in the description of other languages. It takes a great deal of time and energy to produce a systemic functional description; such descriptions demand a penetrating knowledge of the language combined with the trained insight and the native wit of a grammarian. Dr Caffarel has all these qualifications, as a native francophone who is also a theoretical linguist - one moreover with practical experience in teaching her language to semantically recalcitrant foreign learners (i.e. to anglophones). This consistent interplay between theoretical and applied pursuits has always been a defining feature of systemic functional theory, where no clear line is drawn between application and theory and each is a source of positive input to the other. This kind of mutual enrichment is clearly demonstrated in Alice Caffarel's work. The result is a description which penetrates to the heart of the language, revealing it at one and the same time as a specimen of the human semiotic and unique resource for the continuous creation of meaning. M. A. K. Halliday Manly, New South Wales November 2005
Introduction
. . . there can be no such thing as a 'complete' account of the grammar of a language, because a language is inexhaustible. Although there can only be a finite body of text, written or spoken, in any language, the language itself - the system that lies behind the text - is of indefinite extent, so that however many distinctions we introduced into our account, up to whatever degree of fineness or 'delicacy', we would always be able to recognize some more. (Halliday 1994: xiii)
This systemic functional (henceforth SF) interpretation of French grammar is primarily intended as a resource for students of French and French linguistics to explore how meanings are made through grammar in different text types. The aim of this book is to provide a general map of the systemic and structural organization of grammatical regions located at clause rank for use as a tool for discourse analysis (including the analysis of literary texts) and for understanding how French grammar makes meaning in different textual and contextual environments. When we interpret grammar in functional terms as a meaning-making resource, the clause is the natural way 'into the system'. It is the highest grammatical unit on the rank scale (clause/group or phrase/word/ morpheme), which means (i) that it is the 'gateway' to the semantic system and (ii) that it constitutes the immediate environment in which other grammatical units function. It follows that this exploration of the grammar of the French clause takes us beyond the clause, to the text (the highest semantic unit), and below the clause, to the group (the realizing unit of the functional components of the clause). The discourse orientation of this grammar is guided by the two theoretical dimensions of systemic theory: (i) stratification and (ii) instantiation (see Chapter 1 for definitions). The stratification of language in context in the SF model allows us to shunt up from grammar to context via semantics to explore how text as meaning and wording realizes particular contextual configurations of Field, Tenor and Mode. But we can also move in the other direction from context to grammar, and explore the construal of context by language. The relation between the different strata is one of realization, and as Halliday (1998: 15) points out:
2
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
We are able to project this relationship from language on to culture, and show that, in an analogous way, the text 'realizes' the situation. And this is a relationship that can be traversed, or activated, in either direction.
The instantiation dimension (from system to text) also provides us with a bidirectional perspective on language. It gives us the means to account for register variation by moving along the cline of instantiation, that is, from the potential (the system) to the instance (text as wording or meaning). In between the system and the instance we find register (instance type). Approaching register from the instance end (i.e. text) allows us to identify a body of similar instances, a text type. Text types can then be used as data to illustrate how they instantiate a particular set of choices from the potential as well as particular structures. Approaching register from the system end allows us to identify subpotentials of the overall potential. It is precisely the bidirectionality of the stratification and instantiation dimensions that makes SF theory distinct from other theories and this grammar innovative. The multiperspectival approach enables us to foreground both the resources specific to particular registers and the resources general to the language. In addition it provides multiple pathways for exploring how meaning is both construed and constructed by lexicogrammatical patterns in texts. Furthermore, each instance of text can be seen as both realization of the culture in which it is produced and as an instantiation of that culture. In this book, the grammatical system, rather than the semantic system, is taken as the point of departure for the exploration of meaning in different text types. This is because, as pointed out by Halliday and Matthlessen (1999: 17), 'Meanings do not "exist" before the wordings that realize them. They are formed out of the impact between our consciousness and its environment'. In addition, in the ontogenetic development of language (from protolanguage to language), the development of grammar and precisely the stratification of content into semantics and grammar is a semiotic milestone (see Halliday 1975; Painter 2001). Not only does the acquisition of grammar increase our meaning potential, it also opens up for metafunctional diversification, for producing multiple meanings simultaneously as a single structural frame. The stratification of content gives the means to depart from the congruent to arrive at the incongruent through the process of grammatical metaphor (see Section 1.2.1.1). All these aspects of grammar in language increase the power of grammar as a theory, as a tool for discourse analysis and the interpretation of meaning (first and second order) in different text types, as illustrated in Chapters 2 to 5. The rhetorical development of the book is motivated by theoretical, descriptive and educational factors. The relationship that exists between theory and description is also considered to be one of instantiation (see Matthiessen and Nesbitt 1996); it follows that this description of French instantiates SF theory which is itself instantiated in description. Chapter 1 will explore the relationship between theory and description and introduce
INTRODUCTION
3
the metalinguistic properties and categories which are instantiated in the description of French grammar provided in Chapters 2 to 5. The organization of the descriptive chapters reflects both the discourse orientation of this grammar and the 'metafunctional' organizational principle of grammar and semantics identified by SF theory. SF linguists view grammar as a metafunctionally diversified meaning-making potential. The clause is seen as encoding different types of meaning (metafunction) - ideational (logical and experiential), interpersonal and textual - which are realized by different types of structure that map onto one another in the clause. We begin with the grammar of ideation, focusing first on the logical component in Chapter 2 and then on the experiential component in Chapter 3. An account of interpersonal grammar follows in Chapter 4, and an account of textual grammar in Chapter 5. The choice of logical resources as point of departure for this description of French grammar is motivated by its discourse orientation. In a sense, the rhetorical development of this book maps onto the analytical process: the first step in analysing the lexicogrammatical resources of a text consists in dividing that text into clauses before we can proceed to the metafunctional analysis of each clause. Thus, in a discourse-based account of grammar it seems natural to first explore the resources that serve to combine clauses into clause complexes in texts. Furthermore, the principles behind clause complexing give us a means of identifying the clause, that is, the highest ranking grammatical unit. In addition, logical resources are 'fractal' types in the lexicogrammatical system in that they may manifest at other ranks: we may combine words and groups into complexes in the same ways that we combine clauses. This will become important as we move to the metafunctional analysis of the clause and its functional components in Chapters 3 to 5. This journey through the metafunctional potential of French will give the reader a comprehensive understanding of how meanings are created in wordings in French texts through extensive illustrations of text analysis. Also, a bilingual (English—French) glossary of systemic functional terms provided at the end of the book will enable students of French to undertake analysis of French texts in French. The work of many systemic functional linguists has influenced this description of French, and I would like to acknowledge that contribution to this book, and in particular the inspiration provided by the work of Michael Halliday and Christian Matthiessen.
1 Systemic functional theory as a metalanguage for description
. . . theory is a semiotic resource for making meaning in description: description is given value through theory. Since theory is a system of meanings, it gives a higher-level organization to the meanings made in description. The richer the theory we have at our disposal, the richer we can make our description. (Matthiessen and Nesbitt 1996: 67-8, emphasis in original)
1.1 A grammar that makes meaning The choice of systemic functional (SF) theory as a framework for this particular description of French grammar is motivated by the strong orientation towards meaning in the systemic model. SF theory provides a powerful framework for studying the grammar of a particular language as meaning potential, that is, the lexicogrammatical choices available to the speakers of that language to mean in different contexts of use. The descriptive process is an endless process of 'paradigmatic composition', of adding new systems of meaning to arrive at a 'comprehensive' interpretation of the meaning potential (and subpotentials) of that language. One important aspect of SF theory is the interpretation of the relationship between grammar and semantics as a natural one, where grammar is interpreted as construing (in wordings) meaning: Grammatical categories are grammaticalizations of semantic ones; even categories such as Subject which have been claimed to be purely grammatical. (Such claims are a reflection of one's approach to language, not of language itself.) In other words, both semantic and grammatical categories are categories of meaning . . . (Matthiessen 1995: 7-8)
This grammar of French is unlike any other for it realizes a particular approach to language, where language as a whole is interpreted as meaning potential. SF theory is different from other theories in other fundamental respects. On the one hand it has a systemic orientation (developed out of Firth's system-structure theory) - grammar is modelled systemically as a resource for construing meaning in wording; on the other hand it prioritizes functional categories across a spectrum of different modes of meaning over
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION
5
grammatical classes. It follows that functional structure is primary to class or syntagmatic structure. Since functional structure accommodates a spectrum of different modes of meaning, it is multilayered in organization. That is, the different modes of meaning or metafunctions are realized by different layers of structures that map onto each other in the clause. The multilayered account of clause structure is, however, not in itself specific to SF theory but can be found in French functional approaches such as Hagege (1982) and Pettier (1992), which evolved from the Prague School of linguistics, in particular following Danes (1964). What is specific to systemic functional grammar is that all three layers of function structures are internal to the grammar, but are at the same time semantically significant: each layer in the grammar corresponds to one in the semantics. Thus, language constructs and is constructed by different types of meaning which the theory models as metafunctions: ideational (construing our experience of the world around us and inside us as meaning), interpersonal (enacting the world of social roles and relations as meaning) and textual (constructing ideational and interpersonal meaning as a semiotic world of information organized as text in context). In addition, SF theory models language as a tri-stratal semiotic system. Central to this model of language is the notion of realization, which conceptualizes the link between the three strata. Thus, semantics is realized by lexicogrammar and lexicogrammar by phonology. As Halliday (2002 [1996]: 411) points out, the notion of realization explains the huge increase in meaning potential in a tristratal semiotic system (such as adult language) compared to a bistratal system (such as a human infant's protolanguage): . . . in the elements of a primary semiotic (signs), the signifier 'realizes' the signified; but this relationship is unproblematic: although the sign may undergo complex transformations of one kind or another, there is no intermediate structure between the two (no grammar). With a higher order semiotic, where a grammar intervenes, this opens up the possibility of many different types of realization.
The natural relationship between grammar and semantics is reflected in the multilayered organization of the clause, which itself reflects the metafunctional organization of semantics. Stratification and metafunctional diversification of language are modelled in Figure 1.1, using one of the conventional forms of representation of these dimensions in SF theory which will be used throughout this book. That is, the strata (semantics, lexicogrammar, phonology - labelled in bold in Figure 1.1) are represented as nested, co-tangential circles; the metafunctions (textual, interpersonal, ideational) are overlaid across these circles (in Figure 1.1 the notional 'boundaries' between the metafunctions are represented by dashed lines). Although the relationship between grammar and phonology is essentially conventional, we will see that interpersonal meaning is related to TONE systems at the level of phonology (Chapter 4) and that textual meaning can also be realized by phonological systems in a natural way (Chapter 5). Thus,
6
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 1.1 The systemic functional model: metafunctional diversification and stratification it is essentially experiential meaning that is in an arbitrary relationship with the expression stratum. 1.2 An overview of systemic functional theory As seen in Section 1.1, one aspect of the organization of language, and more precisely of the content levels of language, is metafunctional diversification. This is reflected in both the paradigmatic and the syntagmatic organization of the clause. From a paradigmatic or systemic viewpoint, any clause selects options simultaneously from the systems of TRANSITIVITY (experiential), MOOD (interpersonal) and THEME (textual). Features from these systems are realized in three simultaneous layers of functional structuring corresponding to the three metafunctions. While in formal theories structure is generated by rules, in SF theory structure results from the process of choosing systemic features within the systems of system networks and executing the realization statements associated with those options. A description of an example thus includes both a systemic aspect (the features selected, or 'selection expression') and a structural aspect (the structural specifications realizing the systemic features). In addition, for each element of structure, its realization is shown as a set of one or more features characterizing the unit or item that realizes that element. A metafunctional analysis of Examples (1) and (2) is shown in Tables 1.1 and 1.2, respectively. (1)
Yannick a ouvert la porte Yannick opened the door
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION
7
Table 1.1 Metafunctional analysis of Example (1) Yannick
a
ouvert
systemic features of clause:
{indicative: declarative: . . .; material & effective: . . .; unmarked theme & active . . .}
interpersonal: negotiatory structure
Subject
Finite
experiential: transitivity structure
Actor/Agent
Process
textual: thematic structure
Theme
Rheme
features of realizing unit:
{nominal group}
{verbal group}
Predicator
la porte
Complement
Goal/Medium
{nominal group}
Table 1.2 Metafunctional analysis of Example (2) La porte
a ete
ouverte
systemic features of clause:
{indicative: declarative: . . .; material & effective: . . .; unmarked theme & passive: agentive . . .}
interpersonal: negotiatory structure
Subject
experiential: transitivity structure
Goal/Medium Process
textual: thematic structure
Theme
Rheme
features of realizing unit:
{nominal group}
{verbal group: passive}
(2)
Finite
Predicator
par Yannick
Complement
Actor/Agent
{prepositional phrase}
La porte a ete ouverte par Yannick The door was opened by Yannick
The SF descriptions of clauses (1) and (2) differ from other linguistic descriptions with respect to the amount of semantic information that the grammatical analysis provides. In addition, from an SF viewpoint, neither
8
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
clause (1) nor clause (2) is seen as primary. They realize different systemic choices from the general lexicogrammatical system network. Each layer of structuring realizes one type of meaning: the negotiatory (modal) structure realizes interpersonal meanings, the transitivity structure experiential meanings and the thematic structure textual meanings. As pointed out by Matthiessen (1995:21): . . . all structural relations are stated in terms of grammatical functions. Units of particular classes realize functional constituents of other units; for instance, the constituent Theme/Subject/Goal is realized by a group of the class nominal. However, units do not serve directly as constituents of other units; there is always a functional interface. In other words, grammatical structure is not stated in terms of grammatical classes such as nominal group and prepositional phrase, but in terms of the functions they serve.
The boxed representation of the multilayered clause structures clearly highlights the variation in the layering of the simultaneous structures. For example, while in clause (1) the constituent that functions as Subject conflates with the Actor/Agent, in clause (2) it conflates with the Goal/Medium. In addition, while in clause (1) it is the Agent that functions as Theme, i.e. as point of departure of the message, in clause (2) it is the Medium that functions this way. Finally, in a systemic functional description, the example should ideally be taken from natural discourse, so that the properties of the clause can be linked to its contribution to the unfolding discourse, for example in terms of thematic progression (see Chapter 5). SF theory was chosen for this particular description of French because of its metafunctional approach to meaning, its theory of systemic organization and its integration of context to the linguistic system. From a practical point of view, the theory will make it possible to use this grammar in various contexts, including discourse analysis, stylistics and educational linguistics. We will now explore the dimensions of SF theory in more detail, beginning with the dimensions of stratification and instantiation. 1.2.1 Stratification and instantiation Stratification refers to the organization of the linguistic system into strata: it is 'an ordering in symbolic abstraction of the subsystems of language: semantics is realized (expressed, coded, symbolized) by lexicogrammar and lexicogrammar is realized by phonology' (Matthiessen 1995: 5). In contrast 'instantiation is the relationship between the system and the instance; the instance is said to instantiate the system' (Halliday 2002 [1996]: 411). Figure 1.2 shows that stratification organizes not only language but also the whole complex of language in context (see Martin 1992, in particular, for a discussion of this point). The linguistic system is related upwards to context via realization, and related to the instance via instantiation.
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION
9
Figure 1.2 Stratification and instantiation These dimensions give the theoretical basis to the discourse-based approach of this book. (For a stratified model of context, see Martin 1992.) The notion of instantiation is also relevant when we interpret grammar in a discourse perspective and provides the theoretical principle used in the choice of data in this book. As Halliday (2002 [1996]: 412) points out: Instantiation is a cline, with (like lexicogrammar) a complementarity of perspective. I have often drawn an analogy with the climate and the weather: when people ask, as they do, about global warming, is this a blip in the climate, or is it a longterm weather pattern?, what they are asking is: from which standpoint should I observe it: the system end, or the instance end? We see the same problem arising if we raise the question of functional variation in the grammar: is this a cluster of similar instances (a 'text type', like a pattern of semiotic weather), or is it a special alignment of the system (a 'register', like a localized semiotic climate)? The observer can focus at different points along the cline; and, whatever is under focus, the observation can be from either of the two points of vantage.
1.2.1.1 Stratification Systemic functional theory is a tri-stratal model of language in context. The systems to be explored in this book are all located at the lexicogrammatical stratum, as shown in Figure 1.3. We have seen that the semantic orientation of the theory, where language is viewed as a meaning potential, implies that all strata of the linguistic system contribute to the making of meaning. The semantic system semanticizes
10
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
contextual meaning by giving us the resources to enact and construe it as linguistic meaning; the lexicogrammatical system grammaticalizes this meaning, by giving us the resources to create meaning in wording; and the phonological system realizes meaning by sounding the wordings that realize meaning. The three strata of language are grouped into two content strata and one expression stratum. Semantics and lexicogrammar are the levels of content (or the content 'plane') and phonology is the expression plane. Furthermore, as was mentioned in Section 1.1, being the two content strata, semantics and grammar enter in a natural relationship to one another. For example, speech functional semantics stands in a natural relation to MOOD grammar: i.e. the speech function 'question' at the level of semantics is 'congruently' realized by the 'interrogative mood' in the grammar, 'statement' is congruently realized by 'declarative', and so on. However, as will become apparent throughout this book, grammatical realizations of semantic categories are not always congruent. One crucial advantage of the stratification of content is that it gives us the means to account for grammatical metaphor, which in turn provides us with the means to expand our interpretation of the meaning potential, 'as variation in the expression of a given meaning' (Halliday 1994: 342). For instance, whilst the congruent means of realizing a command is the imperative mood as in Ferme la porte (Close the door), we may choose incongruent (or metaphorical) means of realization in different contextual environments, such as the interrogative mood, Pourrais-tu fermer la porte? (Could you close the door?) or the declarative mood,/ voudrais que tu fermes la porte (I would like you to close the door). Metaphorical modes of expression are characteristic of all adult discourse. There is a great deal of variation among different registers in the degree and kind of metaphor that is encountered; but none will be found entirely without it. The only examples of discourse without metaphor that we normally meet with are in young children's speech, and in traditional children's rhymes and songs that seem to survive for that reason: that they lack grammatical metaphors. (Halliday 1994: 342)
Thus the notion of grammatical metaphor is essential to the analysis of most texts and can only be accounted for with a stratified model of language in context. Illustrations of different types of grammatical metaphor will surface throughout this book. 1.2.1.2 Instantiation The relationship that exists between the system and a particular instance of this system is one of instantiation (see Figure 1.4). Such a relationship is intrastratal; there is no move between the strata as there is for realization (see Halliday 2002 [1996]). As mentioned in the Introduction, we will be looking both at grammatical systems as realizations of semantic potential and at texts as instantiations of this potential. We will also consider the region on the cline of instantiation intermediate between system (potential) and text (instance), i.e. that of register.
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION 11
Figure 1.3 Location of this description of French in the overall theoretical landscape
Halliday (2002) models register variation on the cline of instantiation as shown in Figure 1.4. He points out that register may be seen either as variation or similarity depending on which end of the cline of instantiation the observer locates herself. Thus, from the (text) instance end, register is seen as a cluster of similar texts, while from the system end, register is seen as variation in the system. The selection of particular texts in this book will serve to study higher locations on the cline of instantiation, such as registers and, further up, the lexicogrammatical system itself. The discourse-based approach to grammar presented in this book serves to highlight how text instances select particular features (subsystems) from the general grammatical potential. A text does not follow all the paths of a system. Each choice made in the process of creating a text opens and closes other systems. The paths taken by a particular text constitute an instantial system. If we look at the process of creating text, we are adopting a dynamic rather than synoptic view of the system. Most perspectives on language are, however, synoptic. It is easier to analyse and formalize a potential (i.e. a system at rest) or an instance of that potential (i.e. a text seen as a product) than the process of instantiating the potential in an unfolding instance (i.e. a text seen as a process). Most theories of language focus on products rather than processes. Systemic theory is both synoptically and dynamically oriented. However, until recently, most descriptions were essentially synoptic in nature (although computational systemic models in principle had to be both synoptic and dynamic from the start - in practice, even if this was not theorized systemically). Works by Martin (1985), Ventola (1987), Fawcett, Mije and Wissen (1988), Bateman (1989), O'Donnell (1990), Ravelli (1991, 1995), Matthiessen (1993), Fuller (1995), Sefton (1995) and Wu (2000) have contributed to changing the trend. The description presented here will
12
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 1.4 Register variation and instantiation
be essentially synoptic; but it is formulated in such a way that it can be complemented by a dynamic account. 1.2.2 System and structure Since language is a meaning potential, it is the system that provides semiotic environment for structural description in systemic theory. System, as the name of the theory implies, is the primary organizing principle. It refers to the paradigmatic patterning of language, which is seen as more basic than its syntagmatic patterning, i.e. its structure. System and structure are related by realization. System networks are used to represent the paradigmatic mode of organization. A system network is a set of interrelated options or alternative choices called systems. Each system has an entry condition. An entry condition is a feature or a complex of features which serve as choice point in other systems, thus forming networks. A system network consisting of two systems is illustrated in Figure 1.5. This account of French grammar will provide both the systems and the function structures associated with systemic choices. We will arrive at both system and structure by looking at texts which instantiate particular sets of choices from the potential and instantiate particular structures.
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION 13
Figure 1.5 The representational conventions of a system network 1.2.3 The semantic system: metafunctional organization As seen in the previous sections, systemic functional linguists interpret language as a meaning potential diversified functionally into different modes of meaning or metafunctions. It is assumed that the semantic system is organized into three generalized functional components or 'metafunctions': interpersonal, ideational (i.e. experiential and logical) and textual (Halliday 1967, 1968, 1978; Martin 1991; Halliday and Matthiessen 1999). Furthermore, as Halliday (1978: 128) points out: each component of the semantic system specifies its own structures . . . It is the function of the lexicogrammatical stratum to map the structures one onto another so as to form a single integrated structure that represents all components simultaneously.
Figure 1.6 illustrates how the lexicogrammatical stratum realizes the metafunctional component in three structures that map onto one another in the clause. The three metafunctions - interpersonal, ideational and textual - are realized as a tri-layered function structure. Their contribution to the meaning of the clause is as follows: (i) The interpersonal metafunction supplies the resources for enacting (establishing and maintaining) social relations. This component is realized in the clause grammar by MOOD systems (see Chapter 4).
14
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 1.6 The realization of the metafunctions in structure (ii) The ideational metafunction provides the semantic resources for construing (i.e. creating semiotically) our experience of the world around us and inside us as meaning. The experiential metafunction construes phenomena as process configurations, lexicogrammatically realized by TRANSITIVITY systems (Chapter 3). The logical metafunction construes phenomena as chains, for example sequences of process configurations realized by CLAUSE COMPLEX systems (Chapter 2). (iii) The textual metafunction is the 'enabling' or 'relevance' function. It provides the resources for constructing 'text', for presenting ideational and interpersonal meaning as information organized into text in context. Thus, the creation of'text' (a semantic unit) requires simultaneous selection of options from each functional component. At clause rank, the textual metafunction organizes the message into peaks of prominence, realized by THEME systems (Chapter 5). Systemic linguists (Halliday 1979; Matthiessen 1988; Martin 1992) argue that the four metafunctional modes of meaning tend to be realized by four distinct modes of structure. The experiential metafunction has a constituency type of organization, the logical an interdependency type of organization, the interpersonal a prosodic one and the textual a periodic one. This
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION 15
assumption is borne out by this description of French grammar. However, the serviceability of the constituency mode of expression relative to the problems that arise when we are faced with the task of representing prosodic and periodic modes of expression explains the cross-metafunctional use of partwhole structures in Halliday's (1985a/1994/2004) An Introduction to Functional Grammar. This representational uniformity is also useful for showing how the different structures map onto each other. Accordingly, constituency will also be used here to represent experiential, interpersonal and textual structures, while interdependency will represent logical structures. These two basic types of structure are referred to by Halliday (1981 [1965]: 31) as multivariate and univariate: A multivariate structure is one involving more than one variable; a univariate structure is one involving only one variable. The elements of a multivariate structure are thus different variables each occurring once only . . . The elements of a univariate structure are repetitions of the same variable . . .
However, since the grammar is organized systemically, its domain is not defined by reference to types of structures but rather by stratal and metafunctional location. 1.2.4 The lexicogrammatical system: a rank—scale organization
The lexicogrammatical system is modelled as being organized by rank, rather than by immediate constituent structure as in formal theories. Halliday (1966: 111-12) characterizes a rank grammar as follows: By a rank grammar I mean one which specifies and labels a fixed number of layers in the hierarchy of constituents, such that any constituent, and any constitute, can be assigned to one or other of the specified layers, or ranks. . . . [A rank grammar] defines a point of origin for structures and systems, so that the assignment of any item to a given rank, as also the assignment of the structures and systems themselves, becomes an important step in generalization. To show that a system operates at a given rank is the first step in stating its relationship to other systems; likewise to assign an item to a given rank is the first step in stating the systemic and structural relations into which it may enter and those which it may embody within itself.
In French as in English the clause is at the top of the rank-scale hierarchy. It is followed downwards by group rank, then word rank, and morpheme rank as shown in Table 1.3 in relation to Example 3. (3)
Cela te
permettra
de liberer tes
tensions
This you allow-FUTURE to free your tensions This will allow you to get rid of your stress
The grammatical unit within each rank may either be simplex or complex. The latter has a univariate structure (linearly recursive) while the former has a multivariate structure (non-recursive) (see Section 1.2.3). Clause
16
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 1.3 Rank-scale organization in Example (3) Rank
Examples
clause groups words morphemes
|| Cela te permettra de liberer tes te:nsions || || Cela | te | permettra de liberer |;stttensions || || Cela, | te,| permettra, de, liberer, tes, tensions || || ce+la, | te, | permettr+a, de, liberH -er, | tes, tension+s ||
complexes (Chapter 2) and verbal group complexes (see causative and modal expansions in Chapter 4) are examples of univariate structures in French that will be explored in this book. Groups can be seen as organic wholes and also as word complexes: they may be analysed as both multivariate and univariate structures. 1.2.5 The lexicogrammatical system: delicacy
In systemic functional work, the relationship between grammar and lexis is modelled systemically rather than structurally; this leads to a very different view from the traditional one, adopted in formal linguistics, according to which grammar and lexis are different 'components'. In Halliday's view, first discussed in 1961, lexis 'is simply the most delicate grammar. In other words, there is only one network of lexicogrammatical options. And as these become more and more specific, they tend more and more to be realized by the choice of a lexical item than by the choice of a grammatical structure' (Halliday 1978: 43). The delicacy relationship between grammar and lexis means that lexical specifications inherit the more general grammatical specifications. As this book focuses on the least delicate lexicogrammatical systems, we will only be looking at grammatical items and structures. 1.3 Language and context A text is language that is functional in context (Halliday 1978; Halliday and Hasan 1976, 1985). The process of communication cannot be dissociated from the environment (situational, cultural, social) in which it unfolds. 'If we treat both text and context as semiotic phenomena, as "modes of meaning", so to speak, we can get from one to the other in a revealing way' (Halliday and Hasan 1985: 11-12). The systemic functional model, and Martin's (1992) model in particular, does just that by positing language as the expression plane of context. Following Hjelmslev (1943), a semiotic system such as context that finds its expression in another semiotic system (i.e. language) is called a connotative semiotic system; a semiotic system such as language that has its own expression plane (i.e. phonology) is referred to as a denotative semiotic system (see also Barthes 1967).
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION
17
The realizational relationship between context and language is motivated by the reciprocity of their respective organization. The three metafunctions (ideational, interpersonal and textual) that organize the linguistic system reflect respectively the tri-dimensional structure of context formed by the field (social action in which the text lies), tenor (role and status relationships among participants) and mode (role of language) of discourse. Hasan (1989: 237) points out that: Halliday does not claim a simple and absolute correspondence between some specific metafunction and some specific contextual variable, as if one mirrored the other; he does, however, claim that typically the ideational metafunction is constitutive of field, the interpersonal of tenor and the textual of mode.
1.4 Context and register Systemic theory renders explicit the indissociability of language from its context (situational and cultural) by positing a higher-level semiotic plane, context. In systemic work, different models of the organization of context have been explored. In Martin (1992), context consists of three levels of semiosis: ideology (system of coding orientations constituting a culture), genre (context of culture: social processes) and register (context of situation: configuration of situational variables). Such a stratified interpretation of context has been applied and developed in an extensive body of research carried out by Martin and a large number of researchers in linguistics and educational linguistics. Halliday and Hasan (1985) do not represent genre and register as two strata, and Halliday (1978) interprets register as the semantic correlate of a configuration of the situational variables of field, tenor and mode. In other words, Halliday interprets register as part of the linguistic system, where register stands for functional variation (see Matthiessen 1993). For the purpose of this book, which focuses on grammatical systems and on how grammatical choices are constrained in particular contexts and instantiated in particular texts, either type of approach to context can be adopted; I will follow Halliday (e.g. 2002) in foregrounding instantiation (with associated variation) as a dimension separate from stratification (see Figure 1.4). The notion of register motivates the selection of data in this book. As we saw in Section 1.2.1.2 on instantiation, we can approach register from either of two starting points, the system end or the text end of the cline of instantiation. The various lexicogrammatical regions under scrutiny are being explored from both angles. From the text end of the cline, we approached register by selecting texts which are not objects in themselves but representative tokens of more general types of texts or registers. Since context is realized as particular semantic configurations in language, and since register is a semantic construct realized by grammar, we can approach grammatical resources by focusing on particular registers or text types. These
18
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
text types serve to reveal particular lexicogrammatical systems and structures best. For example, we explore the systems of MOOD by analysing dialogues since such text types foreground the lexicogrammatical resources of exchange of information and goods-and-services. From the system end, we explore how different text types instantiate different subparts of a system. I have tried to select varied data to illustrate a broad spectrum of French grammar. In addition to natural texts of various registers that are used to crack the 'code' of the systems and structures of French grammar, sample clauses are also used, either taken from texts or reference grammars and sometimes from my own repertoire as a native speaker to illustrate particular options and structure, as well as to bring out proportionalities in the grammar. 1.5 Locating French systems in the theoretical map Following this introduction of the main theoretical categories of the systemic functional framework, we are now in a position to locate the French lexicogrammatical systems discussed in the book within their metafunctional and rank environment; these are shown in Table 1.4. Each major system is located at an intersection of rank and metafunction. At clause rank, the main systems are MOOD, TRANSITIVITY, THEME and CLAUSE COMPLEXING. The MOOD systems offer the resources for enacting social roles in an exchange; the TRANSITIVITY system, the resources for construing our experience of the world around us and inside us; the THEME system, the resources for presenting interpersonal and ideational meaning as information organized into text. Finally, the CLAUSE COMPLEX systems offer the resources for
Table 1.4 Distribution of the lexicogrammatical systems of French discussed in this book ideational rank
interpersonal
textual
THEME: VOICE
logical
experiential
CLAUSE COMPLEX: PROJECTION EXPANSION
TRANSITIVITY: PROCESS TYPE AGENCY
MOOD: MODALITY & POLARITY
VERBAL GROUP COMPLEX: causative, modal expansion
verbal group: [NUCLEAR TRANSITIVITY] (preselected at clause rank)
nominal group: INTERACTANT TYPE
complexes
simplexes
SYSTEMIC FUNCTIONAL THEORY AS A METALANGUAGE FOR DESCRIPTION
19
construing sequences of experiential configurations by combining clauses into chains. Some group systems, such as INTERACTANT TYPE, will also be discussed (see Table 1.4). As we will see in Chapter 3, transitivity is clearly a clausal system, but its realization spreads across clause and verbal group ranks. Nuclear transitivity, which is realized by clitics in the verbal group, is not a system of the verbal group. The agglutination of participant functions and some circumstantial functions to the verbal group is one of the characteristics of French that has led me to suggest that French is partly like a polysynthetic language (Caffarel 2004a).
2 The grammar of ideation (1): logical metafunction
We shall interpret the relations between clauses in terms of the 'logical' component of the linguistic system: the functional-semantic relations that make up the logic of natural language (Halliday 1994: 216)
2.1 Introduction The grammar of ideation provides us with the resources for construing experience as meaning. There are two complementary modes of construal the logical and the experiential. In the logical mode, experience is modelled as open-ended series or chains; in the experiential mode, it is modelled as organic configurations. The experiential mode will be discussed in Chapter 3. Here we will focus on how the grammar of ideation construes experience as sequences of phenomena realized as clause complexes. The clause complex system provides us with the resources for combining two or more experiential structures. The use of these resources is most prominent in spoken language where the dynamic unfolding of 'talk' is characterized by chains of sequential events. This is displayed in Figure 2.1 by a passage (Text 2.1) which illustrates 'text as process'. In this passage, which is taken from a corpus entitled Poupee (Blanche-Benveniste et al., 1991), a 55-year-old woman from North Africa talks about her childhood and her relationship with her mother. Figure 2.1 illustrates the intricate movement of spoken language that Halliday (1985b: 87) describes as follows: The complexity of spoken language is its intricacy of movement, liquid like that of a rapidly running river. To use a behavioural analogy, the structure of spoken language is of a choreographic kind.
The choreography of Text 2.1 (seen in Figure 2.1) comprises various combinations of logico-semantic relations (expansion and projection) and interdependency relations (parataxis and hypotaxis) selected from the clause complex system which is explored in detail in Section 2.2. Whilst the use of clause complex resources is maximized in spoken language, there are no discrete boundaries between what we characterize
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
21
On etait six enfants a la maison We were six children at home maman travaillait Mum worked mais euh vraiment difficilement but well with great difficulty parce qu'elle lavait le linge pour les gens because she washed clothes for people etj'avais ma sceuret mon frere and I had my sister and my brother ma soeur afnee elle a ete elevee par ma grand-mere my older sister she was raised by my grandmother mais mon frere lui il pensait qu'ajouera courir but my brother him he only wanted to play and run alors pour ainsi dire c'etait presque I'aJnee quej'etais so as it were it was nearly the oldest that I was c'esf moi qui m'occupais de mon grand frere it was me who looked after my older brother etpuis des autres bon bien sur plus petits and also after the others well of course younger et puis alors euh maman elle avail beaucoup confiance en moi voyez and so well mum she trusted me a lot you see
Figure 2.1 The choreography of spoken language as spoken and written language; rather, what we have is a cline, from casual talk to highly technical language, and in between we find text types that have features of both spoken and written language. Either type of language can be graphically or phonically channelled (see Halliday and Hasan 1985). Variation in the construal of logical relations across different text types will be explored in Section 2.3. Logical resources are typically linearly recursive and may potentially be used ad infinitum to realize chains of clauses. They have a univariate or interdependency type of structure, i.e. they consist of repetition of a single variable (see Chapter 1). The recursive potential of logical meaning is typically maximized in texts conveyed via the spoken medium as illustrated in Figure 2.1 above. This feature of spoken language can also be used to convey second-order meanings, that is, meanings that are free from the control of the semantics and, in this particular context, meanings that are free from logical semantics. An illustration of the role played by logical choices in the construal of second-order semiosis is found in Text 2.2, which is the last clause complex in the children's story 'The house that Jack built'. This shows that the power of grammar for creating higher-order semiosis is not limited to literary texts but can be found in any text type. (See Section 2.3 for more examples of second-order semiosis realized by logical grammar.) In Text 2.2, the construal of the text through logical meaning parallels the construal of the first-order phenomenon which is reconstructed by the text, the construction of the house. The clauses themselves do not directly refer to
22
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Text 2.2 (from 'La maison que Jack batit' in Blaireau children's magazine, No 75) la Void lefermier semant son ble Here is the farmer sowing his wheat Ib qui possedait le coq au chant tout guilleret who had the cock with the very lively crow 1 c qui reveilla le cure tout taillade which woke up the cut parish priest Id qui maria Vhomme tout depenaille who married the ragged man le quiembrassa lafille tout esseulee who kissed the forlorn girl If qui tray ait la vache a la corne ondulee who milked the cow with the wavy horn Ig qui fit valser le chien which made the dog waltz Ih qui tourmenta le chat which tormented the cat li quitualerat which killed the rat Ij qui mangea le malt which ate the malt Ik qui se trouve dans la maison which is in the house II que Jack batit. that Jack built
a new step in the building process, but to an event which is embedded, that is, which takes place within the house under construction. Thus we have a story within a story which correlates with the expansion of the house (the material object of the story) which itself correlates with the expansion of the text (the semiotic object). The text starts with one clause, then a clause complex with two clauses, followed by a clause complex with three clauses, and so on until the final clause complex, shown here as Text 2.2, which comprises 12 clauses. Thus, logical resources in this text contribute both to the construal of first-order and second-order meanings. The construal of a second-order semiosis through particular patterns of grammatical selections will be a focus of our discussion throughout this book, and can be seen as epitomizing the power of grammar as a meaning-making resource. In Section 2.2, we explore how French grammar contributes to the semantics of rhetorical relations by constructing sequences of processes (i.e. experiential configurations) as clause complexes. The system of interdependency provides the resources for realizing sequences of processes paratactically or hypotactically and the system of logico-semantic relations provides the resources for construing part of the rhetorical organization of a text, as expansion or projection. The category of parataxis is comparable to the category of coordination in traditional grammar and also includes the notions of apposition (or juxtaposition) and direct speech. Hypotaxis is comparable to subordination, and refers also to non-restrictive relativization
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
23
and indirect speech. Unlike most accounts of subordination, hypotaxis does not include restrictive relative clauses or other embedded clauses (see Matthiessen and Thompson 1989 for a detailed discussion). Section 2.2.1 is concerned with the logico-semantic relation system of expansion and Section 2.2.2 with the logico-semantic system of projection. Section 2.2.3 presents textual and interpersonal criteria for distinguishing hypotaxis from parataxis. In Section 2.3, we explore how clause complex resources are deployed in different text types and foreground some of the differences between texts that have characteristics of spoken language and texts that have characteristics of written language. This leads to a discussion on ideational metaphor. We also discuss instances of second-order semiosis in two literary texts, whereby logical grammar contributes to the overall theme (or secondorder meaning) of the novel. In Section 2.4, we explore the probability of instantiation of features from the interdependency system and the logico-semantic relation system. We will see that certain combinations of logico-semantic relation with parataxis or hypotaxis are more likely to occur than others. 2.2 The CLAUSE COMPLEX system The clause complex system distinguishes between two main types of interdependency relations, one between clauses of equal status - parataxis - and one between clauses of unequal status - hypotaxis. In a hypotactic relation, one clause modifies the other; 'the modifying element is dependent on the modified' (Halliday 1994: 218). In a paratactic relation, one clause is initiating, the other continuing. The category of hypotaxis includes the traditional categories of subordination minus embedding, and indirect speech and thought. The category of parataxis embodies the traditional notions of coordination and juxtaposition plus direct speech and thought (see Halliday, 1985a). The system which offers the choice between parataxis or hypotaxis is named the INTERDEPENDENCY (TAXIS) system. In addition to the two main ways of combining clauses, hypotaxis and parataxis, the clause complex system distinguishes between the different types of logical meaning relations that combine the clauses of a complex. The model presented here allows for the interaction of what will be called logico-semantic relations and parataxis and hypotaxis. The LOGICOSEMANTIC RELATION system distinguishes between two main types of relations: (i) Expansion includes circumstantial relations (discussed in connection with circumstantial subordination) but also apposition and addition. In the present account, circumstantial relations are referred to as enhancing relations, apposition as elaboration and addition as extension. We will see that these meanings may combine with either parataxis or hypotaxis.
24
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(ii) Projection includes the traditional notions of direct and indirect speech. The combination of projection and parataxis corresponds to the traditional category of direct quoted speech (locution) or thought (idea); while the combination of projection and hypotaxis corresponds to the traditional category of indirect quoted speech or thought. The way in which the INTERDEPENDENCE and LOGICO-SEMANTIC RELATIONS systems interact across languages is likely to vary. In French as in English (Halliday 1994) and Tagalog (Martin 1995a), these two main systems are simultaneous. This does not seem to be the case in Japanese, for example (see Matthiessen, Inui and Teruya 1994; Teruya forthcoming), where projection does not seem to differentiate between parataxis and hypotaxis. The correspondences between parataxis, hypotaxis and embedding and traditional categories are tabulated in Figure 2.2. The projection system differentiates between projecting an idea (thought) and a locution (speech). Typically, ideas are projected hypotactically and locutions paratactically, but the reverse patterns are also possible. This is illustrated in Table 2.1. The expansion system differentiates between elaboration, extension and enhancement. These different types of expansion are exemplified in combination with parataxis and hypotaxis in Table 2.2. The grammatical proportionalities illustrated in Table 2.1 for projection and Table 2.2 for expansion motivate the paradigmatic organization of taxis and types of logico-semantic relations as a system (Figure 2.3), The curly bracket { in Figure 2.3 means 'and'; it indicates that the two main systems, INTERDEPENDENCE and LOGICO-SEMANTIC RELATIONS, are
interdependency (tactic) relations
parataxis
hypotaxis
direct speech
indirect speech
elaboration
juxtaposition/ appositions
non-defining relative clauses
extension
coordination
projection
enhancement
embedding
subordination
circumstantial subordination (adverbial clauses)
Figure 2.2 Relationship between the systemic terms for clause complex categories and traditional terms
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
25
Table 2.1 Interaction of projection with parataxis and hypotaxis Projection type
Interdependence type paratactic
hypotactic
locution
"J'ai dormi longtemps " dit-il. "I have slept for a long time" he said.
Edit qu 'il a dormi longtemps. He said that he had slept for a long time.
idea
// pensa: "j 'ai dormi longtemps ". He thought: "I have slept for a long time".
II pensa qu 'il avail dormi longtemps. He thought that he had slept for a long time.
Table 2.2 Interaction of expansion with parataxis and hypotaxis Expansion type
Interdependency type paratactic
hypotactic
elaboration
J'ai dormi longtemps, je vais etre en J'ai dormi longtemps, ce qui va me retard. mettre en retard. I slept for a long time, I'm I slept for a long time, which is going to be late. going to make me late.
extension
J'ai dormi longtemps etje vais etre en retard. I have slept for a long time and I'm going to be late.
J'ai dormi longtemps au lieu de me lever tot. I slept for a long time instead of getting up early.
enhancement
J'ai dormi longtemps doncje vais etre en retard. I slept for a long time so I am going to be late.
J'ai dormi longtemps si bien queje vais etre en retard. I slept for a long time so much so that I am going to be late.
simultaneous. The square brackets mean 'either/or'; they indicate what options are available within each least delicate system. In accordance with Halliday (1985a/1994/2004), the signs next to each feature in Figure 2.3 (1, 2, a, =, etc.) are used to annotate clauses when analysing texts in terms of clause complexing. Arabic numerals (1, 2, 3, . . .) stand for parataxis, and Greek letters (a, (3, y, . . .) for hypotaxis. The logico-semantic relations are symbolized as follows: elaboration by the = sign, extension by +, enhancement by x, projection of an idea by ' and projection of a locution by ". The use of these signs in the analysis of text is exemplified
26
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 2.3 The clause complex system in Figure 2.4 by the analysis of Text 2.1, Poupee, which was introduced in Section 2.1. The clause complex in Figure 2.4 consists of a majority of paratactic relations of expansion. The different semantic relations found in this text can be further subcategorized. Clause Ib does not have a conjunctive marker, but is clearly related to clause la by an 'and' relation, or relation of addition, which is a subtype of extension. Clause Ic extends clause Ib On etait six enfants a la maison We were six children at home maman travaillait Mum worked mais euh vraiment difficilement but well with great difficulty parce qu'elle lavait le linge pourles gens because she washed clothes for people etj'avais ma sceur et mon frere and I had my sister and my brother ma sceur ainee elle a ete elevee par ma grand-mere my older sister she was raised by my grandmother mais mon frere lui il pensait qu'a jouer a courir but my brother him he only wanted to play and run alors pour ainsi dire c'etait presque t'ainee quej'etais so as it were it was nearly the oldest that I was c'est moi qui m'occupais de mon grand frere it was me who looked after my older brother et puis des autres bon bien sur plus petits and also after the others well of course younger et puis alors euh maman elle avait beaucoup confiance en moi voyez and so well mum she trusted me a lot you see
Figure 2.4 Logico-semantic relations in Text 2.1
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
27
paratactically by means of the adversative marker mais. Clause Id is the only example of hypotactic enhancement. It enhances clause Ic by 'embellishing' it through causal expansion. Both clauses If and Ig elaborate on clause le by clarifying what that clause says. Clause Ig extends on If by means of the adversative marker mais. Clause Ih is a paratactic enhancing relation of the consequential type, marked by alors. Clauses li, Ij and Ik elaborate on Ih: clause Ij is in a relation of temporal enhancement with li; clause Ik is in a paratactic consequential relation with clause Ij marked by alors; arguably it is in this type of relation with clause li as well. It also conflates with a temporal relation marked by et puis which also has a continuative function, that of keeping the information flowing. The distribution of the more delicate logico-semantic relations is shown in the network in Figure 2.5. Because of the predominance of paratactic interdependency relations in Text 2.1, and the lack of projections, we do not get a clear picture of the interaction of the two systemic dimensions, types of logico-semantic relation and types of taxis. A more complete paradigm of choices is found in Text 2.3, which is taken from Sartre's short story 'Le mur'. Text 2.3 illustrates a number of new features and interactions of features selected from the clause complex system presented in Figure 2.5. Within parataxis we have another subtype of extension, i.e. variation, realized by ou in clauses lOb and lOd. Within hypo taxis, we have an elaboration in clause 2b and projection of a locution in clause 12b. What may appear as a paratactic
Figure 2.5 Clause complex system: moving towards more delicate options
28
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Text 2.3 From Sartre's 'Le mur' (1939) la 1 On nouspoussa dans unegrande salle blanche, Ib +2 a et mes yeux se mirent a cligner Ic xp pane que la lumiere leurfaisait mal. 2a a Ensuite, je vis une table et quatre types derriere la table, des civils, 2b =P qui regardaient des papiers. 3a 1 On avail masse ks autres prisonniers dans kfond 3b +2 a et il nousfallut traverser toute la piece 3c xp pour ks rejoindre. 4a 1 II y en avail plusieurs [[queje connaissais]] 4b +2 et d'autres [[qui devaient etre etrangers]]. 5a 1 Les deux [[qui etaient devant moi]] etaient blonds avec des cranes ronds, 5b —2 1 its se ressemblaient: 5c =2 des Francais, j'imagine. 6a 1 Le plus petit remontait tout k temps son pantalon: 6b =2 c'etait nerveux. 7a 1 (?« dura pres de trois heures: 7b =2 1 j'etais abruti 7c +2 etj'avais la tetevide 7d +3 mais la piece etait bien chauffee 7e +4 1 etje trouvais ca plutot agreabk: 7f =2 depuis vingt-quatre heures, nous n 'avions pas cesse de grelotter, 8 Les gardiens amenaient ks prisonniers I'un apres I'autre devant la table. 9 Les quatres types kur demandaient alors kur nom et kur profession. lOa 1 La plupart du temps ils n 'allaientpas plus loin lOb +2 1 ou bien alors ils posaknt une question par-ci par-la: lOc = 2 1 "As-tuprispart a u sabotage d e s munitions?" lOd +2 ou bien: "Oil etais-tu k matin du 9 lOe +3 et quefaisais-tu?" 11 a 1 Ils n 'ecoutaient pas ks reponses lib +2 1 ou du moins ils n'en avaientpas I'air: He =2 1 ils se taisaient un moment lid +2 et regardaient droit devant eux He x3 puts ils se mettaient a ecrire. 12a a Ils demanderent a Tom 12b "(3 1 sic 'etait vrai [[qu 'il servait dans la Brigade Internationale]]: 12c =2 Tom ne pouvait pas dire k contraire a cause des papiers [[qu 'on avail trouves dans sa veste]]. 13a 1 A Juan ils ne demanderent rien, 13b +2 xp mais, apres qu 'il eut dit son nom, 13c a ils ecrivirent longtemps.
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION 29
projection of a locution in clauses lOc to lOe is in fact an elaboration of the 'speech function' question functioning as Verbiage (i.e. the nominalization of what is said) in clause lOb. The brackets [ [ . . . ] ] in Text 2.3 indicate downranked clauses, that is, clauses that do not function as a clause but as a participant within the clause. Downranked clauses are either embedded within a nominal group functioning as participant (as in clauses 4a, 4b and 5a) or they are a participant themselves. This is the case in clause 12b c'etait vrai [[qu'il servait dans la Brigade Internationale]], where [[qu 'il servait dans la Brigade internationals]] has the function of Carrier in an attributive clause. (See Chapter 3 on transitivity functions.) Such embedded clauses can also be put in a thematic position, as in [[qu 'il servait dans la Brigade Internationale]] etait vrai. The importance of recognizing downranked (also known as 'rankshifted') clauses (whether embedded in a nominal group or not) will surface again in the next chapter when we discuss complexity within groups and phrases. It is important to note here, however, that downranked clauses are different from nested clauses which function as clausal elaborations. Whilst downranked (embedded) clauses are expansions of the nominal group and as such modify that nominal group, hypotactic elaboration of a clause by another clause can be nested or sequential. Elaborations within the nominal group are realized by what is traditionally known as restrictive or defining relative clauses, whilst hypotactic elaborations of clauses are typically realized by non-restrictive or non-defining relative clauses. These may either interrupt the main clause, marked by the following bracketing « ...» as in Example 3, or be sequential as in Example 4. The following examples show these distinctions. Restrictive: downranked clauses, embedded in a nominal group (1) Je n 'aime pas les gens [[qui se plaignent tout le temps]]. I don't like people who always complain. (2) Prends le livre [[qui est sur la table]]. Take the book that is on the table. Non-restrictive: hypotactic elaborating clauses, which may interrupt the main clause or may be sequential (3) Le Mont Blanc, «qui s'eleve a plus de 4800 metres,» est le plus haul sommet d 'Europe. Mont Blanc, which is higher than 4800 metres, is the highest peak in Europe. (4) Cet annuaire est valable jusqu'au premier Janvier, (date) a laquelle il sera remplace par une nouvelle edition. This directory is valid until January 1st, when it will be replaced by a new one. (Example 4 taken from Judge and Healey 1985: 352) The following sections will focus in some more detail on how parataxis and hypotaxis interact with the logico-semantic relations: Section 2.2.1 focuses on expansion, Section 2.2.2 focuses on projection and Section 2.2.3 looks at the criteria for distinguishing hypotaxis from parataxis.
30
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
2.2.1 Expansion
Matthiessen (1995) points out that when a clause expands on another, there is no difference in order of abstraction between the two clauses; the order of reality they represent is the same. As will be discussed in Section 2.2.2, pro jecting clauses contrast with expanding clauses in this respect, because the projecting clause and the projected clause represent different orders of abstraction. A clause may expand on another in a number of ways, as shown in Text 2.3: (i) through extension, by adding a new element or offering an alternative (e.g. clause Ib); (ii) through enhancement, by qualifying it with some circumstantial feature of (for example) location, cause or condition (e.g. clause lie); or (iii) through elaboration, by restating in other words, specifying in greater detail, commenting or exemplifying (e.g. clause 5b). These categories are further exemplified below, along with some of their more delicate subcategories. Note that the subcategories of enhancement largely correspond to circumstantial categories in the experiential grammar (see Chapter 3); not all of these subcategories have been exemplified here. Indeed, these parts of the grammar are still being developed; just as Halliday points out with regard to English (see Halliday 1994), the area of circumstantiation needs further exploration, so the subcategories represented on the network in Figure 2.5 do not present an exhaustive picture. All examples in this Section and the following Section 2.2.2 are extract taken from the last short story of Les Fous de Dieu (Danielou 1975: 141-84 This short story, also entitled 'Les fous de Dieu', was selected because it displays many characteristics of spoken language, as it is a French rendition of Indian oral story telling. Parataxis and extension: initiating clause (1) followed by ( A ) continuing clause (+2) (5a) 1 Elles leplacent aupres d'unefemme qui vient d'accoucher dans la campagne They place him with a woman who has just given birth in the country (5b) +2 et volent le nouveau-ne. and steal the new-born. Parataxis and extension and elaboration: 1 A +2 (1 A =2) (6a) 1 Moni etait trouble et surpris Moni was confused and surprised (6b) +2 1 mais il se laissait faire; but he let himself be talked into it (6c) =2 Debou etait son ami. Debou was his friend. 2.2.1.1 Paratactic elaboration
Paratactic elaboration is rarely realized by a conjunction but often by a pause, which is marked graphically by a semi-colon or a comma, as in Example 6c above. Thus, paratactic elaboration corresponds to juxtaposition or apposition in traditional grammars (see Figure 2.2). When interacting with parataxis, the elaboration system offers the choice between reformulating what has been said in the initiating clause by (i) rewording it [exposition],
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION 31
(ii) specifying it further by means of an example [exemplification] or (iii) clarifying it by saying something different but related [clarification]. An elaboration relation of the exposition type may be realized by juxtaposition or by a reformulation marker, such as a savoiror soil (that is). The realization of elaboration by means of an explicit conjunction is more widely found in written texts, while intonational realization, which involves the juxtaposition of the clauses combined, is more common in spoken texts. 2.2.1.2 Hypotactic elaboration
In a hypotactic clause complex, elaboration is typically realized by a nondefining relative clause whose function is to comment on the dominant clause or part of the dominant clause. Hypotactic elaborating clauses may be finite or non-finite, as shown in the next set of examples. Hypotaxis and elaboration ('non-defining relativization'): dominant clause (a) followed by ( A ) relative clause (=P) (7a) a Les Bauls sont des 'fous de dieu', des gens sans famille, sans demeure Bauls are 'crazy about god', people without family, without home (7b) =P qui vivent de la charite du monde. who live from people's charity. Hypotaxis and elaboration: non-finite '"' hypotaxis and elaboration: finite a=P (a=P) (8a) a Its allerent jusqu'a une belle maison neuve They went to a beautiful new house (8b) =(3 a situee un pen en dehors de la ville situated a bit outside town (8c) =P oil vivait seul Mohan un homme tres pieux. . . where lived alone Mohan a very devout man . . . 2.2.1.3 Paratactic extension
As shown on the network in Figure 2.5, there are two main types of paratactic extension, [addition] and [variation]. In paratactic extension of the [addition] type, one meaning is added to another (this is typically realized by the conjunctive marker et (and)), whereas with [variation], one meaning is presented as an alternative to another (typically realized by ou (or)). There is a further distinction in the [addition] type, between [additive] and [adversative]: in the [adversative] type, the meaning of the added clause contrasts with that of the first clause rather than simply adding a similar meaning, and this is typically realized by the conjunctive marker mats (but). The [additive] type can further be divided into [positive] and [negative] types (again shown in Figure 2.5). All of these contrasts are displayed in the following set of examples. Paratactic extension: addition: additive: positive (9a) 1 Quelquefois Debou venait au hameau Sometimes Debou came to the hamlet (9b) +2 et Moni etait heureux. and Moni was happy.
32
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Paratactic extension: addition: additive: negative (lOa) 1 On nepouvait rien faire We could not do anything (1 Ob) +2 ni pour les aider nor to help them (lOc) +3 ni pour survivre soi-meme nor to survive ourselves (lOd) +4 et Suresh entrama Moni loin de la ville vers le Bihar. and Suresh took Moni away from the City towards the Bihar. Paratactic extension: addition: adversative (1 la) 1 Les gens commen^aient a fair People started to run away (lib) +2 mais la terreurserepandait dans les villages. but terror was spreading in the villages. Paratactic extension: variation (12a) 1 Ces gens veulent tuer tons les hindous These people want to kill all the Hindus (12b) +2 ou les forcer a adopter I'Islam. or force them to adopt Islam.
2.2.1.4 Hypotactic extension Hypotactic extension also comprises the relations of [addition] and [variation]. The dependent clause, as in the case of hypotactic elaboration, may be finite or non-finite. Hypotactic extensions of addition are mostly realized by non-finite clauses without any conjunctive markers. The form of the verb in such clauses (an instance of which is shown in Example 13) is traditionally referred to as the 'gerund'. Hypotactic extension: addition: additive (non-finite) (13a) a Lesfilles se tenaient par les bras sur une longue ligne, (13b) +(3
The girls were holding hands in a long line les garfons sur un rang leurfaisant face.
boys in a row facing them. Hypotactic extension: addition: adversative (finite) (14a) a // etait pale comme Shiva He was pale like Shiva (14b) +P alors que Vishnou, Rama, Krishna ont lapeau sombre comme Suresh. whereas Vishnu, Rama and Krishna have dark skin like Suresh. Hypotactic extension: variation (15a) a Toutvabien Everything is well (15b) +P sauf qu'ilfait mauvais temps. except that the weather is bad. (Example 15 taken from Judge and Healey 1985: 362)
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
33
2.2.1.5 Paratactic enhancement The network in Figure 2.5 shows the various options for paratactic enhancement; only two of the most commonly seen features will be exemplified in this subsection, but conjunctive markers that typically realize the other features are shown in Table 2.3. Paratactic enhancement of the temporal type is typically indicated by puts (then), and paratactic enhancement of the causal-conditional type is typically indicated by the conjunction car (for). These are illustrated in the examples below. Paratactic enhancement: temporal: succession (16a) 1 II lava son ami He washed his friend (16b) x2 puis le concha sur la couverture. then laid him on the blanket. Paratactic enhancement: causal-conditional: cause (17a) 1 Moni dut oiler mendier dans les temples Moni had to go begging in the temples (17b) x2 car nul ne pretait grande attention a ses chansons. for nobody paid attention to his songs.
2.2.1.6 Hypotactic enhancement As with paratactic enhancement, just some of the options shown on the network in Figure 2.5 will be illustrated here; Table 2.4 shows further examples of the conjunctive markers that typically realize the different types of hypotactic enhancement. As with other types of hypotactic expansion, these clauses may be non-finite, in which case they typically occur without a conjunctive marker, as shown in Example 19. Hypotactic enhancement: temporal: simultaneity (18a) a Les premieres averses annonfaient la saison des pluies The first downpours announced the rainy season (18b) xp quand ils arriverent a Mymensingh. when they arrived at Mymensingh. Hypotactic enhancement: manner (non-finite) (19a) a Monipartit Moni left (19b) xp serrant I'ekatara contre son cceur. holding the ekatara against his heart. Hypotactic enhancement: causal-conditional: condition (within hypotactic projection; see Section 2.2.2) (20a) a Suresh dit a Debou Suresh said to Debou (20b) 'p a qu 'il voulait bien accepter ce nouveau disciple that he would accept this new disciple (20c) xp si celui-ci etait capable de supporter la dure vie des Bauls errants. if this one was capable of bearing the hard life of the travelling Bauls.
34
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
2.2.1.7 Expansion: summary of options and conjunctive markers To finish this overview of how the expansion type of logico-semantic relation combines with parataxis and hypotaxis, the principal markers of the different types of paratactic expansion and hypotactic expansion are displayed in Tables 2.3 and 2.4, respectively. 2.2.1.8 Expansion and the mode subjonctif As shown in Table 2.4, some enhancing logico-semantic relations motivate the use of the subjunctive mode. (The term 'mode' rather than 'mood' will be used to refer to the subjunctive here, to distinguish it from the mood options that realize speech functional semantics; see Chapter 4.) What is traditionally known as the mode subjonctif is used in bound clauses and is dependent on the modality or modulation of the projecting clause in hypotactic projection (see Section 2.2.2 on projection) or the modality/ modulation inherent in the logico-semantic relation; this latter use of the subjunctive mode is discussed in this subsection. Guillaume (1945, 1984) suggests that events expressed in the indicative (non-subjunctive) in the bound clause have the potential of being actualized (realis), while events expressed in the subjunctive do not (irrealis). Guillaume notes that while notions such as 'probable' and 'certain' govern the indicative, the notion of 'possible' governs the subjunctive. These notions refer to the different values of probability (low, median and high), a
Table 2.3 Markers of paratactic expansion Type of expansion Extension addition: additive: positive addition: additive: negative addition: adversative variation
Elaboration
Enhancement temporal manner causal-conditional: purpose causal-conditional: consequence causal-conditional: concession causal-conditional: condition
Markers
et, ainsi que ni mats ou, soil. . . soil, tantot. . . tantot, ou bien, seulement, non seulement. . . mais punctuation marks such as a comma or colon; soil, a savoir
puis, ensuite, alors comme, ainsi que, aussi bien que, de meme que car, c'est pourquoi, alors done, alors, ainsi, enfin neanmoins, cependant, toutefois, pourtant, du moins, du reste alors, sinon
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
35
Table 2.4 Markers of hypotactic expansion Type of expansion Extension addition: additive: positive addition: adversative variation
Elaboration
Enhancement temporal
manner causal-conditional: cause causal-conditional: purpose
causal-conditional: consequence causal-conditional: concession causal-conditional: condition
Markers
alors que, tandis que, pendant que sans si... c 'est que, si. . . que, au lieu de, a par que, sauf que, outre, hors que, excepte que qui, que, quoi, dont, oil, ce qui, ce dont, lequel, laquelle. . .
quand, lorsque, avant que, en attendant que, comme, pendant que, apres que, aussitot que, des que comme, ainsi que, de meme que, tel que, moins. . . que, aussi. . . que, plus . . . que parce que, puisque, vu que, comme, du moment que, d 'autant que pour, pour que, afin que, de crainte que, a cettefin que, defafon que, de sorte que (all of these take the subjunctive) de maniere que, de sorte que, si bien que (all of these take the indicative) bien que, encore que, quoique (these three all take the subjunctive); meme si a condition que (this takes the subjunctive)
subsystem of the modality potential which will be discussed in Chapter 4. The subjunctive indicates that the event has not yet taken place but is desired, wanted or possible, while the non-subjunctive indicates that the event has taken place, is taking place or will take place. Martin (1992: 193) points out the relationship between modalization and modulation and the basic types of consequential relation in English (manner, consequence, condition and purpose): With [manner] relations, the relationship between events is modulated through 'potentiality'; we won by training hard means that the Cause (preparing well) enabled the Effect (winning). With other consequential relations the connection between events is modulated through 'obligation': we won because we trained hard means that the Cause determined the Effect. . . Alongside being modulated through obligation, causal relations may be modalized. With both condition and purpose the relation between Cause and Effect is a contingent one; and in both cases the effect is irrealis - there is a possibility, a probability or a certainty that it will be determined by the Cause, but as the meanings are made it has not yet ensued.
36
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Purposive relations motivate the use of the subjunctive, as seen in Table 2.4 above, because 'purposives contain an additional modulation of inclination, associated with the Effect (the Effect is desired)' (Martin 1992: 194), as in Example 21. (2la) a Jel'aifait I did this (21b) xp pour que cela lui fosseplaisir. in order to make her happy. (Example 21 taken from Judge and Healey 1985: 146)
Condition relations which express a probability such as en admettant que, a supposer que, pourvu que (provided that) also require the use of the subjunctive. On the other hand, conditions introduced by si (if) do not require the subjunctive. This is because in a clause such as sij'etais riche, j'acheterais un bateau, the hypothesis is imagined as taking place. The same does not apply to a hypothesis introduced by que + subjunctive, hence the possibility of contrast, as seen in Example 22. (22a) xp 1 Si vous kfaites (action seen as taking place)(22a) xp 1 Si vous kfaites (action seen as taking place) If you do it (22b) +2 etqu'un accident s'ensuive (hypothesis) and if an accident happens (22c) a vous le regretterez toute votre vie you will regret it all your life (Example 22 taken from Judge and Healey 1985: 148)
Some conjunctions, such as de sorte que, defafon que and de maniere que, may introduce either a purposive relation or a consequential relation. It is the mode of the dependent clause, whether subjunctive or not, that will mark the difference between the two logico-semantic relation types. Thus, as seen above, purposive relations are modulated and this is indicated by the subjunctive, while consequential relations are not, as shown in Examples 23 and 24, both taken from Judge and Healey (1985: 147). Hypotactic enhancement: causal-conditional: purpose (subjunctive indicating that the effect is wanted, i.e. purposive) (23a) a Ils ontferme la barriere They closed the gate (23b) xp de sorte que les cyclistes ne puissent pas passer. in order to stop the bike riders going through. Hypotactic enhancement: causal-conditional: consequence (indicative indicating that the effect is consequential) (24a) a Ih ont ferme la barriere They closed the gate (24b) xp de sorte que les cyclistes ne peuvent pas passer. therefore the bike riders cannot go through.
Thus the mode of the subordinate clause in contrastive contexts such as in Examples 23 and 24 serves to indicate the type of logico-semantic relation involved.
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
37
In temporal enhancing clauses, where the event of the dominant clause is seen as taking place before the event of the dependent clause, there is a doubt as to whether the event will take place, and the subjunctive is used. On the other hand, when it is the event in the dependent clause that is to occur first, the event in the dependent clause is clearly seen as taking place and the non-subjunctive (indicative) is used, as illustrated in Examples 25 and 26, taken from Judge and Healey (1985: 146). Hypotactic enhancement: temporal: succession (subjunctive indicates element of uncertainty) (25a) a J'arriverai a la maison I will arrive home (25b) x(3 avant que I'orage n'ait eclate before the storm starts Hypotactic enhancement: temporal: succession (non-subjunctive) (26a) a J'irai a la plage I will go to the beach (26b) x|3 des que j'aurai fini man travail as soon as I will have finished my work.
To summarize, then, there are three clause complexal environments in which the subjunctive mode is selected in the bound clause: (i) hypotaxis and enhancement: causal-conditional: purpose; (ii) hypotaxis and enhancement: causal-conditional: condition where the condition is non-hypothetical; and (iii) hypotaxis and enhancement: temporal: succession where the event of the main (a) clause is seen as occurring before the event construed in the dependent ((3) clause. We have seen so far that the subjunctive is a possible option of expanding bound clauses. In the next section we will see that it is also an option for projected clauses. 2.2.2
Projection
The previous section considered the various intersections between interdependency relations (taxis) and the subtypes of one of the main types of logico-semantic relation, expansion. This section turns from expansion to consider the intersections between parataxis and hypotaxis and the other major type of logico-semantic relation, projection. As was noted at the beginning of Section 2.2.1, the semantic difference between expansion and projection is that the two clauses in an expanding relation represent the same order of reality, whereas the projecting relationship represents two orders of abstraction: the projected clause represents the 'symbolic content' (meaning or wording) and the projecting clause the 'symbolic expression' (speech or thought). The potential of a clause to
38
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
project another order of reality in this way is criterial, in French, for distinguishing mental and verbal process clauses (both of which can project) from material and relational process clauses (which cannot). This aspect of projection will be discussed further in Chapter 3. As can be seen in the network in Figure 2.5, the [projection] options are much simpler than the [expansion] options: there are only two further options for [projection], [idea] and [locution], corresponding to projection by mental process clauses and verbal process clauses respectively. The intersection of taxis with these two options therefore produces the simple fourway paradigm that was presented in Table 2.1; that is, there are four basic categories, corresponding to the traditional categories of direct (quoted) and indirect (reported) speech and thought. These categories will be briefly exemplified in Section 2.2.2.1; Section 2.2.2.2 will focus on how the subjunctive mode combines with hypotactic projection and the metaphorical realization of modality through hypotactic mental projections. 2.2.2.1 Examples of projection Examples 27 and 28 illustrate the paratactic projection of a locution and of an idea, while Examples 29 and 30 illustrate the hypotactic projection of a locution and an idea. Parataxis and projection: locution (27a) "1 "Tu es tres beau, Moni" 'You are very beautiful, Moni" (27b) 2 dit-il soudain a voix basse. he said suddenly in a soft voice. Parataxis and projection: idea (28a) 1 Ilpensa He thought (28b) '2 "un dieu est entre en mot". "a god has entered me". Hypotaxis and projection: locution (29a) a Suresh dit a Debou Suresh said to Debou (29b) "P qu 'il voulait bien accepter ce nouveau disciple si. . . that he would accept this new disciple if... Hypotaxis and projection: idea (30a) a Les gens du hameau pensaient tons The people from the village all thought (30b) '(3 que Moni etait un enfant desfees. that Moni was a child of the fairies.
Note that in a paratactic complex, the projected clause may come first, as in Example 27, or second, as in Example 28, while in a hypotactic complex, the projected clause always comes second, as in Examples 29 and 30. (Clause reversibility will be discussed further in Section 2.2.3.)
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
39
2.2.2.2 Hypotactic projection and the mode subjonctif As mentioned in Section 2.2.1, projected clauses can also select for the subjunctive when the projected clause indicates high obligation, as in Example 31. Main (a) clause: high obligation; dependent (P) clause: subjunctive (31a) a Jeveux I want (31b) 'P que tu viennes that you come
On the other hand, when the obligation has median value, the projecting clause selects for the indicative, as in Example 32. Main (a) clause: median obligation; dependent (p) clause: indicative future (32a) a Je suppose I expect (32b) 'p que tu viendras that you will come
These examples show that modality, like other interpersonal resources, is realized prosodically across the clause complex. There is a clear relationship between the system of mood and that of modality in the sense that modality choices allow an expansion of the means of realizing speech functional semantics, as we will see in Chapter 4. Modulation typically correlates with the imperative mood and serves to realize proposals, and modalization correlates with the indicative mood and serves to realize propositions. We could extend this further and posit that the subjunctive also correlates with modulation, while the non-subjunctive correlates with modalization. However, this generalization does not hold in the context of mental projections. Indeed, the subjunctive can be a marker of modulation or modalization. In the context of verbal projection, there is a clear relationship between subjunctive and proposal, and non-subjunctive and proposition. As we will see in Chapter 4, modality can be realized congruently (nonmetaphorically) by modal elements, such as modal adjuncts or modal verbs, or non-congruently (metaphorically) by projecting clauses in hypotactic clause complexes. In these non-congruent realizations, the orientation of the modality may be subjective or objective, i.e. the projecting clause may be personal (je+ mental process) or impersonal (il+ est+ modal adjunct). It is the personal (subjective) projections that are interpreted as interpersonal metaphors, and which are of interest in this discussion on clause complexes. Consider the following examples. Personal projection: modalization (probability: median): subjective and explicit; dependent clause indicative (33a) Je crois I think
40
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(33b) que ma voiture est cassee. that my car is broken. Personal projection (of intention: desire); modulation (obligation: high): subjective and explicit; dependent clause subjunctive (34a) Je veux I want (34b) que ma voiture soil cassee. my car to be broken. Mental projection of desire; (no obligation); dependent clause indicative (35a) Ilespere He hopes (35b) que ma voiture est cassee. that my car is broken.
Examples 33 to 35 illustrate various types of projection, but while the projected clause of Example 34 is in the subjunctive, the projected clauses of Examples 33 and 35 are in the indicative. What motivates the choice between indicative and subjunctive in projected clauses? What kind of meaning differentiates these two dependent 'modes'? The choice of either mode in Examples 33 to 35 seems to depend on 'the value that is set on the modal judgement' (low, median or high) and type of the modality of the projecting clause (see Halliday 1985a; Matthiessen 1995). The subjunctive is used in the context of high obligation while the indicative is used in the context of high/median probability and low/median obligation. Thus, as pointed out by Guillaume (1945, 1984), both 'certain' (high value probability in SFL terms) and 'probable' (median value probability) motivate the choice of the indicative, while 'possible' (low value probability) triggers the subjunctive. In the context of modulation, it is high and median obligation that trigger the choice of the subjunctive mode, while low obligation or no expression of obligation triggers the indicative, as further illustrated by Examples 36 and 37. Modulation: obligation: median; dependent clause subjunctive (36a) J'aimerais I would like (36b) qu'ilparte. him to go. No obligation; dependent clause indicative (37a) fespere I hope (37b) qu'ilpartira. that he will go.
To summarize, then, in the clause complexal environment of mental projection and hypotaxis, there are three types of modality that preselect the subjunctive mode in the bound clause:
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
41
(i) modalization: probability: low: positive; (ii) modalization: probability: high: negative; and (iii) modulation: obligation: high. In the same clause complexal environment, there are two types of modality that preselect indicative mood in the bound clause: (i) modalization: probability: median or high; and (ii) modulation: obligation: median or low. The uses of the subjunctive that have been identified are determined by the modality of the dominant clause, where there is mental projection of the intentional or desiderative subtype. In verbal projections, the choice between subjunctive and indicative does not depend on the modality of the dominant clause but rather on whether the projected clause is a proposition or a proposal. While the actualization of the event of the proposition is potentially high, i.e. is either probable or certain, the actualization of the event of the proposal is possible but not certain. This is illustrated by Examples 38 to 41, which are taken from Judge and Healey (1985). Proposition: reported statement: indicative (38a) Pierre a dit Pierre said (38b) quejean viendrait a huit heures. that Jean would come at eight o'clock. Proposal: reported command: subjunctive (39a) Pierre a demande Pierre asked (39b) queje vienne a huit heures. that I come at eight o'clock. Proposition: statement: indicative (40a) Je lui ai dit I told him (40b) qu'ilest a I'heure. that he is on time. Proposal: command: subjunctive (41 a) Je lui ai dit I told him (41 b) qu'ilsoit a I'heure. to be on time.
We could argue that in verbal projection, the subjunctive is analogous to the imperative in free clauses (see Chapter 4). The analogy between the subjunctive and the imperative is foregrounded by the fact that bound clauses can be used with an elliptical dominant clause to express a command addressed to a third person, as shown in Examples 42 and 43.
42
(42)
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH Qu'ilvienne!
That he come! '(I want) him to come' (43)
Qu 'il m 'attende!
That he me wait! ' (I want) him to wait for me'
We have seen that the subjunctive is a possible option of bound clauses (expansions and projections), as formalized in the system network in Figure 2.6. The discussion on the subjunctive continues in Chapter 4 where we explore different resources for expressing modality (modalization and modulation). 2.2.3 Criteria for distinguishing hypotaxis from parataxis 2.2.3.1 The clause reversibility criterion
There are a number of criteria that can be used to differentiate hypotactic from paratactic clause complexes. Halliday (1985a/1994) suggests textual criteria in relation to English which are also applicable to French: in a hypotactic clause complex the (3 clause may be thematized or Theme predicated; thus the order of the a and [3 clauses may be reversed. In contrast, in a paratactic complex, 1 always precedes 2. This is precisely what the numerals indicate: the order of occurrence of the clauses. Clause 1 is referred to as the initiating clause and clause 2 as the continuing clause. In paratactic expansions the ordering of 1 and 2 cannot be changed for textual purposes. In a hypotactic complex, the P clause can sometimes be thematized, by being positioned clause initially and/or predicated as in Examples 46 and 46. (See
Figure 2.6 Clause types systems
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
43
also Chapter 5 on THEME and THEME PREDICATION.) Example 44 is taken from LesFous deDieu (Danielou 1975). a A P (hypotaxis and enhancement: temporal: successive) (44a) a // arrive que les asparas, des danseuses du del, aient un enfant,
It happens that the asparas, dancers from the sky, have a child, (44b) xp
lorsqu'elles s'eprennent d'un homme. when they fall for a man.
P A a (hypotaxis and enhancement: temporal: successive) (45a) x(3 Lorsque les asparas, des danseuses du del, s'eprennent d'un homme, When the asparas, dancers from the sky, fall for a man, (45b) a il arrive qu'elles aient un enfant.
it may happen that they have a child. Theme-predication (46a) x(3 C'est lorsque les asparas, des danseuses du del, s'eprennent d'un homme, qu' It is when the asparas, dancers from the sky, fall for a man, that (45b) a il arrive qu 'elles aient un enfant.
it happens that they have a child. The thematic flexibility of hypotactic clause complexes applies mostly to enhancing relations, which, as we will see in Section 2.4, are quantitatively the most frequent type of hypotaxis, and one could argue, could be interpreted as the prototypical type of hypotaxis. Thus, the reversibility criterion, which reflects the thematic choice of sequences that one may have in some hypotactic clause complexes and not in paratactic ones, is not a sufficient criterion as it does not apply to all types of expansion, hence the interpersonal criteria proposed in the next subsection. For example, in the context of hypotactic elaboration, as in Example 47, the (3 clause cannot be thematized. a P only (hypotaxis and elaboration) (47a) a // etait I'ami d'un Baul He was the friend of a Baul (47b) =P qu'ilaccompagnait parfois dans ses tournees. whom he accompanied sometimes on his tours.
Just as in English, it is not possible to put the elaborating dependent clause first, i.e. we cannot have Qu'il accompagnait parfois dans ses tournees il etait I'ami d 'un Baul In the context of projection, the reversibility criterion is itself reversed: in a paratactic complex, the projected clause may come first, as in Example 48, or second, as in Example 49, while in a hypotactic complex, the projected clause always comes second, as in Example 50. Parataxis and projection: projected clause A projecting clause (48a) 1 "Je veux bien accepter ce nouveau disciple, " "I will accept this new disciple"
44 (48b) 2
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH dit Suresh a Debou. Suresh says to Debou.
Parataxis and projection: projecting clause A projected clause (49a) 1 Suresh dit a Debou: Suresh says to Debou (49b) 2 je veux bien accepter ce nouveau disciple. I will accept this new disciple. Hypotaxis and projection: a (projecting) A (J (projected) only (50a) a Suresh dit a Debou Suresh said to Debou (50b) '|3 qu'il voulait bien accepter ce nouveau disciple si. . . that he would accept this new disciple i f . . .
Again, the reverse of Example 50 is not possible, i.e. we cannot have Qu'il voulait bien accepter ce nouveau discipk dit Suresh a Debou. These examples show that in the context of projection, the clause reversibility criterion is a consistent means of distinguishing hypotaxis from parataxis. In the next subsection, we explore interpersonal criteria for distinguishing hypotaxis and parataxis in the context of expansion. 2.2.3.2. The mood and modality agreement criteria It has already been noted, in the discussions of the use of the subjunctive in relation to hypotactic expansion and projection, that in hypotactic clause complexes mood and modality are realized prosodically across the clause complex through, for example, the differential selection of subjunctive and indicative in the dependent clause in relation to different projecting verbs (where different projecting verbs construe different values of modality, as in the distinction between 'want' and 'hope'). In addition, paratactic and hypotactic clause complexes constitute different environments for MOOD selections, that is, the selection of resources for enacting speech functions such as command, offer, statement and question. (These resources will be discussed in more detail in Chapter 4.) These different environments are outlined and exemplified below. In a paratactic clause complex, clauses which have equal status select independently for mood. Consequently, the clauses that form the complex do not necessarily select the same mood options. For example, one may be declarative and the other imperative, as in Example 51, where clause 51d is in the imperative, while the initiating complex to which it is related is in the declarative mood. Parataxis and MOOD selection (51a) "1 a Debou a assure [declarative] Debou certified (51b) ' (3 que tu sais chanter quelques chants de Bauls, [declarative] that you know how to sing some Baul songs (51c) 2 1 dit Suresh, [declarative] says Suresh,
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION (51d)
=2
45
chante-moi quelque chose, [imperative] sing me something.
In a hypotactic clause complex, on the other hand, the mood of the dependent clause always agrees with the mood of the dominant clause, as in Examples 51a and 51b above and in Examples 52 and 53 (taken from Robbe-Grillet 1966: 14) below. Hypotaxis and enhancement (52a) xp Et tandis que la mer efface aufur et a mesure les traces des pattes etoilees,
And while the sea is continually obliterating the traces of their (52b) a
a
star-shaped feet, les pas des enfants demeurent inscrits avec nettete dans le sable a peine humide,
the children's footsteps remain clearly inscribed in the barely (52c)
=P
moist sand, oil les trois lignes d'empreintes continuent des'allonger.
where the three lines of prints continue to lengthen. Hypotaxis and elaboration; hypotaxis and enhancement (53a) a La surface plate de I'eau est bordee d'un ourlet subit,
The level surface of the sea is fringed with a sudden little wave, (53b) =P
a
qui eclate aussitdt
which immediately breaks (53c)
xp
pour se repandre en mousse blanche.
and runs away in white foam.
The dependency relation that exists between a (3 clause and its a does not only involve mood but also polarity and modality as well as tense. In Examples 52 and 53 all the clauses are set in the present tense, and there is no expression of modality. As we have seen, in hypotactic clause complexes, interpersonal resources are realized prosodically across clauses and only hypotactically dependent clauses can select for what is called the subjunctive mode, the use of which is motivated by the semantics of the a clause or that of the conjunction. We will now turn to variation in clause complex selection across various text types. 2.3 Clause complexing across text types In the previous section all examples were taken from a gangetic tale, 'Les fous de Dieu' (Danielou 1975). This text was selected because it displays many characteristics of spoken language. Its spoken nature is apparent in the grammatical intricacy with which its information is organized (see Halliday 1985b). In texts conveyed via the spoken medium, information tends to be distributed across a number of clauses, while in texts conveyed via the written medium, more information tends to be packed within one clause. As a consequence of this, one configuration of meanings will tend to be realized by a clause complex of two or more clauses in a spoken text, and as
46
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
one clause in a written text. Thus, the grammatical complexity of spoken language reflected in the logical organization of a text is 'translated' in written language by a different kind of grammatical complexity, namely lexical density. In other words, the high ratio of clauses per clause complex in spoken texts is rendered by a high ratio of lexical items per clause in written texts. Example 54 is the first clause complex of the short story 'Les fous de Dieu'. (54a) xp (54b) a
Lorsque les asparas, des danseuses du del, s'eprennent d'un homme, When the asparas, the dancers from the sky, fall in love with a man, il arrive qu'elles aient un enfant. it may happen that they have a child.
This clause complex could possibly be rendered in a 'written' version as in Example 55. (55) L'engouement des asparas, des danseuses du del, pour un homme pent entratner la naissance d'un enfant. The infatuation of asparas, dancers of the sky, for a man can result in the birth of a child.
Clause 55 illustrates the use of grammatical metaphor within the ideational metafunction; the sequences of processes realized by a clause complex in Example 54 are now realized as one clause. Events in Example 54 are realized in Example 55 as nominal groups, and the logico-semantic relation as a process. Thus, the process s'eprendre (falling in love) is realized by a nominal group engouement (infatuation), the Process avoir un enfant (having a child) by the nominal group naissance (birth of a child) and the logico-semantic relation lorsque (when) is realized by the verbal group, entratner (result). The agnation between congruent and non-congruent realizations in Examples 54 and 55 is shown in Table 2.5. Example 55 is metaphorical in the sense that it does not represent congruent grammatical patterns of selection. Congruently, a sequence of two processes is realized as a clause complex as in Example 54 rather than as a clause simplex as in Example 55. The choice of a metaphorical mode of
Table 2.5 Congruent and non-congruent (metaphorical) realizations of a series of processes semantic entity
sequence of processes process logico-semantic relation
grammatical realizations congruent
non-congruent
clause complex verbal group conjunction
clause nominal group verbal group
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
47
expression in Example 54 entails that the two clauses of the clause complex are downranked as nominal groups. In addition the logical semantic relation which is typically realized as a conjunction as in Example 54 is in Example 55 realized as a Process. While the use of clause complexes is maximized in spoken language, grammatical metaphor is at its densest in scientific writing. In Halliday and Martin (1993: 15), Halliday and Martin note: The birth of science, then (if we may indulge in a well-worn lexical metaphor), from the union of technology with mathematics, is realized semiotically by the birth of grammatical metaphor, from the union of nominalization with recursive modification of the nominal group.
An illustration of the use of simple clauses resulting from the packaging of processes into nouns and relations into processes is found in Text 2.4. As a consequence of the reduction in clause complexing and nominalization, we have an increase in lexical density, i.e. the average number of lexical items per clause in a given text. In Text 2.4, logico-semantic relations (in bold) are realized by processes rather than conjunctive relations. Logical metaphor in scientific writing serves to construct cause and effect between nominal groups within a single clause rather than between two clauses in a clause complex. And as we have seen, the use of logical metaphor goes hand in hand with the use of experiential metaphor. Lexical items in Text 2.4 have been underlined. Lexical items 'are lexical because they function in lexical sets not grammatical systems: that is to say, they enter into open not closed contrasts' (Halliday 1985b: 63). Text 2.4 'Decrire puis expliquer 1'Univers' from Les Dossiers Speciaux du Grand Quid Illustre: Etoiles et Galaxies, (Bottinelli, et al. 1984) 1 Le developpement de I 'astronomic est lie a celui des techniques (lunettes, telescopes, spectroscopie, photographie, radar, exploration spatiale) et des connaissances theoriques. The development of astronomy is linked to that of new techniques (lenses, telescopes, spectroscopy, photography, radar, space exploration) and theoretical knowledge. 2a Les lois de la physique nous permettent non seulement de decrire 1'Univers mais aussi de I 'expliquer; The laws of physics allow us not only to describe the universe but also to explain it; 2b toutefois les phenomenes observes dans I 'Univers peuvent aussi nous amener parfois a modifier certaines lois physiques. however, the phenomena observed in the universe can also lead us to modify certain laws of physics. 3 Certaines limites a la mecanique de Newton ont conduit Einstein, en 1916, a formuler la theorie de la relativite generate. Certain limitations to Newtonian mechanics have led Einstein, in 1916, to formulate his theory of general relativity.
Text 2.4 displays a high lexical density with an average per clause of 8.7 lexical items (35 items divided by 4 clauses) despite the fact that this
48
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
particular scientific text is addressed to a young audience rather than scientists. A scientific text addressed to a specialist readership would at times show even greater density. On the other hand a 'spoken' text would display a high grammatical complexity with many clause complexes resulting in a lower rate of lexical items per clause, as in the first text presented in this chapter, Text 2.1, which is repeated here with the lexical items underlined. Text 2.1 Poupee (Blanche-Benveniste et al 1991) 1 a On etait six enfants a la maison1 a On etait six enfants a la maison We were six children at home Ib maman travaillait Mum worked 1c mais euh vraiment difficilement but well with great difficulty 1d pane qu 'elle lavait le linge pour les gens because she washed clothes for people 1e etj 'avals ma sceur et mon frere and I had my sister and my brother If ma sceur ainee elk a ete eleveepar ma grand-mere my older sister she was raised by my grandmother Ig mais mon frere lui il pensait qu 'a jouer a courir but my brother him he only wanted to play and run 1 h alors pour ainsi dire c 'etait presque I'ainee quej 'etais so as it were it was nearly the oldest that I was 1 i c 'est moi qui m 'occupais de mon grand frere it was me who looked after my older brother Ij etpuis des autres bon bien surplus petits
Ik
and also after the others well of course younger et puis alors euh maman elle avait beaucoup confiance en moi voyez and so well mum she trusted me a lot you see
Here we have 11 clauses and 30 lexical items (many of which are repeated in the text), which gives an average of 2.7 lexical items per clause. (For a detailed discussion on the different complexities found in spoken and written language, see Halliday 1985b.) The tension between the different complexities is often at play in literary texts for stylistic purposes. Caffarel (2004b) argues that Camus' 'ideal absence of style' in L'Etranger (1942) is achieved partly by shifting from one complexity to the other (from grammatical intricacy to lexical density). Furthermore, the sudden occurrence of a simple clause packed with grammatical metaphors (see Text 2.5) following the use of simple clauses and clause complexes with little nominalization is significant in creating a world of successive, isolated events (rather than sequentially connected events), emphasizing both the detachment of the main protagonist (Meursault) from his environment and a world devoid of logic and transcendence. The final clause with its logical metaphor of cause realized by a densely packed thematized circumstance serves to stress causality not as logical connection between experiences but simply as a circumstance bringing about an event in
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
49
the protagonist's existence over which he has no control. The circumstance is detached from the rest of the clause by functioning as an absolute theme, i.e. a theme that is outside the experiential structure, in the same way that Meursault is portrayed as an outsider, detached from the reality and society in which he lives. Text 2.5 From 1 2 3 4a 1 4b +2 4c 5a xp 5b a 6a a 6b x(3 6c 1 8a a 8b xp 9
L'Etranger (Camus 1942: 10) J'aipris Vautobus a deux heures. Ilfaisait tres chaud. J'ai mange au restaurant chez Celeste, comme d'habitude. Ils avaient tons beaucoup depeinepour moi 1 et Celeste m'a dit: "2 "on n 'a qu 'une mere." Quandje suis parti, ils m 'ont accompagne a la porte. J'etais un pen etourdi a parce qu 'il afallu queje monte chez Emmanuel xp lui emprunter une cravate noire et un brassard. II a perdu son oncle, ily a quelques mois. J'ai couru pour ne pas manquer le depart. Cette hate, cette course, c'est a cause de tout cela sans doute, ajoute aux cahots, a I'odeur d'essence, a la reverberation de la route et du ciel, queje me suis assoupi.
I caught the two o'clock bus. It was very hot. I ate at Celeste's restaurant, as usual. They all felt very sorry for me and Celeste told me, 'There's no one like a mother'. When I left, they came to the door with me. I was in a bit of a daze because I had to go up to Emmanuel's place to borrow a black tie and armband. He lost his uncle, a few months ago. I had to run for the bus. It was probably all this dashing about and then the jolting and the smell of petrol and the glare of the sky reflecting off the road that made me doze off. [translated by Joseph Laredo]
Camus' novel shows that particular patterns of logical selections (or nonselections), as well as contrastive sequences of congruent and non-congruent realizations of logical relations, contribute to construe the themes of absurdism - detachment, absence of logic and transcendence - that run through the novel. Another illustration of how the different complexities characteristic of spoken and written language can be played with to foreground a particular theme or style is found in Louis-Rene des Forets Le Bavard (1947). 'Un bavard' is someone who talks a lot, a chatterbox. The novel is composed of a succession of extensive clause complexes which foreground the dynamic and fluid nature of talk which is central to the theme of this text. However, unlike in natural talk the internal structure of each clause in a complex is itself complex, with complex group structures and a refined (not ordinary) vocabulary, making the text highly literary despite a logical structure very
50
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
similar to that of casual conversation, as shown by the two extracts presented as Texts 2.6 and 2.7 below. Text 2.6 From LeBavard (Louis-Rene des Forets 1947: 8) (A translation of the entire passage is provided below.) la a On me demanderapeut-etre Ib "P a sij'ai entrepris de me confesser 1 c xp a pour eprouv pour pen morbide Id =P 1 dot j le +2 a et queje comparerais volontiers a celui [[que recherchent quelques personnes a(2 a) xp 1 Si vous kfaites (action se n as taking place) If =P a qui, avec une lenteur etudiee, caressent du bout de I'index une legere egratignure Ig =P 1 qu 'elles se sontfaite sciemment a la levre inferieure 1h +2 ou qui piquent de la langue la pulpe d'un citron apeine mur. They will maybe ask me || if I began to confess || to experience this kind of pleasure a bit morbid || that I am talking about || and that I would happily compare to the one [[sought by some refined people]] || who, with a carefully designed slowness, caress with the tip of the index finger a slight scratch || that they have knowingly done to their lower lip || or prick with the tip of the tongue the pulp of a lemon hardly ripe, [my translation]
Text 2.7 is particularly pertinent here because the narrator 'talks' about his writing style. Text 2.7 From LeBavard (Louis-Rene des Forets 1947: 9-10) (A translation of the entire passage is provided below.) la 1 Mon gout meporte naturellement vers le style allusif, colore, passionne, sombre et dedaigneux Ib +2 a etj'aipris aujourd'hui, nonsans repugnance, la resolution de laisser de cote toute recherche formelle, le xp a de sorte que je me trouve ecrire avec un style Id =P 1 qui n'estpas kmien; le =2 a c'est a dire que j'ai ecarte tous les charmes derisoires If =P a dont il m 'arrive parfois dejouer, Ig +P 1 tout en sachantbien [[cequ'ils valent]]:
Ih
+2
ils ne sont les fruits que d'une habilete assez ordinaire.
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
51
My taste leads me naturally towards an allusive, coloured, passionate, sombre and contemptuous style || and I have taken today, not without repugnance, the resolution to leave aside all formal research, || so that I find myself writing with a style || which is not mine; || that is, I have brushed away all the derisory charms || with which I sometimes play || knowing well [[what they are worth]]: || they are only the fruits of quite an ordinary skill, [my translation]
This section has briefly illustrated how logical grammar can be manipulated to encode meanings of a second order, that is, meanings that go beyond the realization of interdependency and logico-semantic relations, foregrounding Camus' philosophy of the absurd in L 'Stranger and the dynamic character of unfolding talk in Le Bavard. From the instance end of the cline of instantiation we have seen that selection of logical patterns can vary across texts for stylistic purposes. We now turn to the probabilistic nature of the system by exploring which combinations of choices from the logico-semantic and interdependency systems are more likely to occur than others. 2.4 The probabilistic nature of the clause complex system The relative frequency of occurrence of certain combinations of logicosemantic relation with parataxis or hypotaxis allows us to make some predictions about the probabilistic nature of the system, that is, about what is 'more likely' or 'less likely' to occur. For example, we will see that, in French, extension tends to be paratactic, while enhancement tends to be hypotactic. Nesbitt and Plum's (1988) investigation of what features of the clause complex system of English are more likely to occur in texts will provide a background to my discussion on interactive patterns in French clause complexes. While Nesbitt and Plum focus on the relative frequency of occurrence of each feature in the system, we will focus on the relative frequency of patterns of interaction between features, more precisely of interaction between features of the interdependency system and features of the type of logicosemantic relation system. Section 2.2 looked at the choices or possibilities that are available to the French speaker for clause combining. The present section focuses on patterns of typical choice, i.e. on probabilities. As Nesbitt and Plum (1988: 8-9) write: In summary, the linguistic system as a system of paradigmatic oppositions is a system of possibilities. Choosing a particular feature in a system means what it does because of the features that were not chosen but could have been chosen. This is the qualitative aspect of the system, the system of 'either/or' relations. But the system is not only a system of possibilities, it is also a system of probabilities. The linguistic system as a system of probabilities is also a potential to mean. The choice of a particular feature also means what it does against the background of what are more likely and less likely choices. What is said is not only interpreted against a background of what could have been said but was not; it is also interpreted against the background of expectancies, against the background of what was more likely
52
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
and less likely to be said. The grammar of a language is not only the grammar of what is possible but also a grammar of what is probable.
The probability of instantiation of certain features of the clause complex system in French will be investigated in the context of narrative registers. Consider Text 2.8, again a narrative text taken from Danielou (1975). Text 2.8 From 'Les fous de Dieu' (Danielou 1975) la xp Lorsque les asparas, des danseuses du del, s'eprennent d'un homme, When the asparas, dancers from the sky, fall for a man, Ib a il arrive qu'elks aient un enfant. it happens that they may have a child. 2a xp 1 Incapables de le nourrir Unable to feed him 2b +2 ou de I'emmener dans le monde invisible, or to take him in to the invisible world, 2c a 1 elks le placent aupres d'unefemme [[qui vient d'accoucher dans la campagne]] they place him with a woman who has just given birth in the country 2d +2 a et voknt le nouveau-ne and steal the new-born 2e =p qu'ellesfont mourir. that they make die. 3a a On ne pourrait expliquer autrement One could not explain otherwise 3b "P a qu'ily aitparfois des enfants that there are sometimes children 3c =P 1 qui nefont rim comme les autres which do nothing like the others 3d +2 et que personne ne comprend. and that nobody understands.
The types of interdependency and logico-semantic relation found in Text 2.8 are summarized in Table 2.6. Table 2.6 reveals that there is a tendency in French for extension and parataxis to combine, as well as for enhancement and elaboration to combine with hypotaxis. Although the data presented in Text 2.8 are quite restricted, it is interesting to note that the preliminary predictions we can draw from it are borne out by analysis of a further extract of the same text, which is presented here as Text 2.9. The patterns of combination between taxis and logico-semantic types found in Text 2.9 are summarized in Table 2.7. Table 2.7 confirms the following tendencies: parataxis and extension; hypotaxis and enhancement; and hypotaxis and elaboration. While hypotactic elaboration is marked by relative pronouns, there is often no marker of paratactic elaboration of the exposition type, but simply juxtaposition of the
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
53
Table 2.6 Summary of interdependency and logico-semantic relations in Text 2.8 type of logico- tyPe of taxis semantic relation: parataxis expansion elaboration extension enhancement
hypotaxis
2e, 3c
2b, 2d, 3d
projection idea locution
la, 2a
3b
Text 2.9 From 'Les fous de Dieu' (Danielou 1975) 4a a a Les gens du hameau, «4b», pensaient tons The people from the village, «4b», all thought 4b +p < <sans rien dire a sa mere> > without telling his mother 4c 'P que Moni etait un enfant des fees. 5a
a
5b
xp
5c 6a
6b
a
xp a
'P
6c
1 +2
7a
1
7b
+2
7c
a =P
8a
1
8b
+2
that Moni was a child of the fairies. // restait seul pendant des heures He stayed alone for hours couche au bard dufleuve
lying down on the river bank en murmurant des phrases [[depourvues de sens]]. whispering meaningless phrases. Le maitre [[qui apprenait aux enfants a lire et a compter dans une ecole [[qui se trouvait a plus de quatre lieues au sud]] ]] disait The teacher who taught children how to read and count in a school which was more than four leagues south said que Moni recilail des poemes that Moni recited poetry et qu 'il n 'y avail rien de mal.
and that there was nothing wrong with that. Mais le maitre d 'ecole avail surement lui-meme appris ces chases a Moni But the teacher had most certainly taught those things to Moni et il avail vecu a Calcutta and he had lived in Calcutta oil resident les Mlechas, les demons etrangers. where live the Mlechas, the foreign demons. // etait peut-etre un pen magicien lui-meme He was maybe a bit of a magician himself et on ne pouvait pas trap sefier a ses dires. and one could not trust what he said too much.
54
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
initiating and continuing clauses. It would seem that there is a correlation between the probabilistic nature of the system, its quantitative aspect, and the number of structural markers available for the realization of a particular type of logico-semantic relation. For instance, there are a large number of markers for enhancing hypotactic relations and just a few for enhancing paratactic relations. The same kind of reasoning applies to extension: there are a greater number of paratactic markers than hypotactic markers, and this correlates with a greater use in French of paratactic extension than hypotactic extension. Thus, it would seem that the quantitative aspects of the system are reflected in the quantitative aspects of structural realizations. The counts in Tables 2.6 and 2.7 are combined in Table 2.8. It shows that certain combinations of selections from TAXIS and LOGICO-SEMANTIC TYPE are foregrounded in this particular narrative text. The absence of dialogic parts in this abstract explains the lack of paratactic
Table 2.7 Summary of interdependency and logico-semantic relations in Text 2.9 type of logicosemantic relation:
type of taxis parataxis
expansion elaboration extension enhancement
hypotaxis 4b, 7c
6c, 7b, 8b 5b, 5c
projection idea locution
4c 6b
Table 2.8 Conflation of Tables 2.6 and 2.7 type of logicosemantic relation:
type of taxis parataxis
expansion elaboration extension enhancement projection idea locution
hypotaxis
2e, 3c, 4b, 7c 2b, 2d, 3d, 6c, 7b, 8b la, 2a, 5b, 5c
4c 3b, 6b
THE GRAMMAR OF IDEATION (1): LOGICAL METAFUNCTION
55
locutions; this does not represent a tendency, but rather a surprising gap. In novels, it is common for narrative parts to interact with dialogic parts. This is illustrated by Text 2.10, an abstract taken from Chapter 1, 'Rumba', of a detective novel by Delacorta (1985) entitled Vida. Text 2.10 From 'Rumba' (Delacorta 1985) la a Alba et Serge debarquerent du 747 de la TWA Alba and Serge disembarked from the TWA 747 Ib =P qui venait deseposera los Angeles. which had just landed in Los Angeles. 2 Ils avaient pour tout bagage dome kilos de musique dans une mallette et les vetements [[qu 'ils portaient]]. They had for all luggage twelve kilos of music in a small suitcase and the clothes they were wearing. 3a 1 Apres vingt-deux heures de voyage, Alba avail un sourire gros comme le Ritz After twenty-two hours of travel, Alba had a smile large like the Ritz 3b +2 et Serge etait comme un lion [[qui n 'a pas bouffe depuis quinze jours.]] and Serge was like a lion who had not eaten for fifteen days. 4a 1 Ils allaient devorer la ville, They were going to devour the city, 4b =2 c'etait sur. that was sure. 5 Ils passerent devant le bar du terminal. They passed in front of the bar of the terminal. 6a a Deux photos de Marilyn etaient la Two photos of Marilyn were there 6b x(3 pour leur adresser un sourire de bienvenue. to give them a smile of welcome. 7a "1 Je me desseche, I am drying out, 7b 2 dit Alba. said Alba. 8 Je veux un Coca geant, comme ceux [[que les gamins s'enfilent dans Us feuilletons americains]]. I want a giant Coke, like those that kids guzzle down in American soaps. 9 Alba et Serge s 'installment au bar. Alba and Serge sat at the bar. lOa "1 Je gouterais bien du bourbon, I would not mind tasting some Bourbon, lOb 2 articula Serge dans son anglais haute epoque Berlitz/Assimil. uttered Serge in his upper-class English from the Berlitz/Assimil method.
Although Text 2.10 has very few clause complexes, we still find instantiations of the patterns of combination seen in Texts 2.8 and 2.9, i.e. elaboration and hypotaxis, extension and parataxis and enhancement and hypotaxis, as seen in Table 2.9. Text 2.10 also illustrates the pattern of the combination of parataxis with projection of locution which is likely to occur in novels and narrative texts of various kinds. On the other hand, projection of ideas tends to occur in conjunction with hypotaxis. Although this discussion on the probabilistic nature of the French clause
56
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 2.9 Summary of interdependency and logico-semantic relations in Text 2.10 type of logicosemantic relation:
expansion elaboration extension enhancement projection idea locution
type of taxis parataxis
hypotaxis
4b 3b
Ib 6b
7a, lOa
complex system is to be understood as introductory, the analysis of further texts has, however, shown similar results. Those results which have emerged from looking at the system from a probabilistic point of view help explain why the two systems of TAXIS and LOGICO-SEMANTIC RELATIONS have not been interpreted as independently variable in traditional grammars. 2.5 Conclusion This chapter has been concerned with the logical resources of TAXIS and LOGICO-SEMANTIC RELATIONS, which serve to combine clauses into complexes. This account provides a comprehensive interpretation of logical resources for the clause complex analysis of text. A cross-classification of logicosemantic relation for 'subordination' and 'coordination' is lacking in French grammars. In the model presented here, the two main systemic dimensions of the clause complex system, INTERDEPENDENCY and LOGICOSEMANTIC RELATIONS, were shown to be complementary. Interpersonal criteria, in addition to the textual criteria proposed by Halliday (1994), were given for distinguishing hypotactic clauses from paratactic ones. In addition to looking at the choices or possibilities that are available to French speakers for clause combining, this chapter also looked at how foregrounded logical patterns can be used to create second-order meanings and at probabilities of instantiation. The analysis of texts enabled some generalizations as to the likelihood of instantiation of certain patterns of interaction between taxis and logico-semantic relations. We found that combinations of elaboration and hypotaxis, extension and parataxis and enhancement and hypotaxis were more likely than other combinations to occur in narrative texts. In the next chapter, we move on to the experiential metafunction, where we explore in detail how clauses realize our experience of the internal and external world as configurations of processes, participants and circumstances.
3
The grammar of ideation (2): experiential metafimction
Our most powerful impression of experience is that it consists of 'goings-on' happening, doing, sensing, meaning, and being and becoming. All these goings-on are sorted out in the grammar of the clause. Thus, as well as being a mode of action, of giving and demanding goods-&-services and information, the clause is also a mode of reflection, of imposing order on the endless variation and flow of events. The grammatical system by which this is achieved is TRANSITIVITY. (Halliday 1994: 106)
3.1 Introduction The preceding chapter explored one aspect of the grammar of ideation: the logical metafunction, which provides the resources for clause complexing, i.e. for creating sequences of experiential configurations. The present chapter is concerned with the other aspect of the grammar of ideation: the experiential metafunction, which is the internal functional structuring of experiential configurations within the clause itself. This chapter will consider how the clause grammaticalizes our experience of 'goings-on': material events, relations between things, feelings, thoughts and verbalizations. It focuses on the construal of experiential meanings both at clause rank and across texts. On the one hand, it explores the French clause as a process configuration, as the realization of the experiential systems of TRANSITIVITY (grammatical perspective). On the other hand, it explores how specific texts construe different modes of participation in the process (discourse perspective), particularly with regard to whether a text emphasizes the 'cause-and-effect' aspect of processes (ergative model) or 'the deed-and-extension' one (transitive model) (Halliday 1994: 163). The structural organization of French transitivity reflects its metafunctional location; it is essentially part-whole: a composition or configuration of a process, participants and circumstances. While from a systemic viewpoint transitivity is clearly located at clause rank, from a realizational viewpoint transitivity in French is shared across ranks, more precisely between clause and verbal group ranks, as shown in the Table 3.1. As a result of this division of labour between clause and group, French can at times
58
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 3.1 Realization of transitivity functions nucleus clause rank realization
Je donnerai I give-FUT "I will give this book to Paul."
ce livre this book
a to
group rank realization
Je k I it "I will give it to him."
lui to him
donnerai give-FUT
Paul Paul
present characteristics of polysynthetic languages with an agglutination of functions to the verbal group. The variation between realizing transitivity at clause rank or realizing it at verbal group rank is determined by textual factors: participants or circumstances which are realized as clitics within the verbal group tend to be Given and identifiable, while participants realized as non-clitics at clause rank tend to be New and non-identifiable (see Chapter 5). The exploration of transitivity at group rank will open the discussion to the notion of nuclearity in Section 3.6. We will see that transitivity functions that have the potential to be realized at group rank are more nuclear than those that do not. This is because they are inherent in the nuclear Process, while other functions (such as circumstances) are more or less peripheral and less likely to be inherent in the Process. Before moving on to this discussion, it is important to stress that the issues discussed in this chapter in relation to the ergative and transitive models of participation are different from issues of 'case marking'. We are not concerned in the first instance with whether or not French has a nominativeaccusative or ergative-absolutive or even a split ergative system, based on whether the Subject (S) of an intransitive clause is marked the same as the object (O) [ergative system] or as the subject (A) [transitive system] of a transitive clause. 'Case marking' is part of the overall picture, but only a small part: it is a realizational resource, along with other realizational resources. The ergative and transitive models discussed here are not determined on the basis of 'case marking'1 or any other realizational categories. We move to the grammar from discourse-semantics, rather than from word grammar. This approach reveals that transitivity models are realized by complementary systems in the ideational clause grammar. These grammatical systems may be realized by case marking and/or any other realizational categories. However, as Figure 3.1 shows, the metafunctional layering of clause grammar means that case marking, word order, and so on, are not limited to the realization of experiential grammar but can contribute to the realization of any or all types of meaning. In French, for example, the pronominal case marking system serves to realize a combination of functions within the
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
59
different metafunctions: as we will see in Chapter 4, the marking of French pronominals serves on the one hand to distinguish textually prominent pronominals from textually non-prominent ones, such as moifromjein moi,je le veux (me I want it), and on the other hand, within non-prominent pronominals it serves to simultaneously mark interpersonal and experiential functions. For example, interpersonally, the Subject // (nominative case) is distinguished from the complement le (accusative case) in il le veut (he wants it) and experientially, Medium le (accusative case) is distinguished from Beneficiary lui (dative case) in il le lui donne (he it to him gives). The delicacy continuum in Figure 3.1 shows that the ergative and transitive models are complementary both in the realization of grammatical patterns and in the schematization of lexis into collocational patterns. Thus, according to Martin (1992), the discourse-semantic unit underlying a lexical item, that is, 'a message part', can be realized congruently as one lexical item or metaphorically as a combination of lexical items such as Process + Medium (e.g. la tombee de la nuit (nightfall) = fall + night), Process + Range (e.g. recits d'aventures (adventure stories) = tell + adventures), Process + Circumstance (e.g. apportdel'exterieur (foreign support) = contribution + from outside), and so forth. These kinds of experiential structures construe nuclear relations or collocations. Ample examples of collocational patterning can be found in Matthiessen (1995). The theory of transitivity discussed here is therefore different from theories of case or grammatical relations (see, e.g., Plank 1979; Dixon 1979, 1994) in many respects, one being that it is not a typological theory of primitive grammatical relations. Thus, we are not trying to classify the French language as ergative or transitive on the basis of the realization of its core grammatical (experiential) functions as does Dixon (1979, 1994). Rather, we
Figure 3.1 Interpretation of transitivity and ergativity within the SF framework
60
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
are trying to establish whether the systemic notions of ergative and transitive are theoretical notions that can serve to explore patterns of experiential representation in French or any other language. We will see in Section 3.2 that the experiential grammar of the French clause can be viewed from both a transitive and an ergative perspective and that the lexicogrammar itself displays ergative tendencies in the sense that a majority of processes can be used either non-ergatively (in a middle clause) or ergatively (in an effective clause), as with ouvrir (to open) in Examples 1 and 2 below. It will be made explicit that the ergative and transitive models do not represent different ways of subcategorizing processes, but rather complementary ways of construing experience in the French clause. Grammatical criteria will serve to support the classification of processes into general semantic types (Section 3.3). Abbreviations used in the Examples below are: Med = Medium, Proc = Process, Ag = Agent, Ac = Actor and mat = material. ergative pattern: material and middle (1)
La porte The door Med
s 'ouvre opens Proc
ergative pattern: material and effective (2) Louis ouvre la porte Louis opens the door Ag Proc Med
The semantics of the ergative model is causation: 'is the Process represented as self engendering or as engendered by an external cause?' (Halliday 1985a: 147). Note that in Example 1 the self-engendering nature of the process is explicitly marked on the process by the clitic s'. From an ergative perspective, the variable is presence or absence of Agent. A clause which represents a self-caused process (without an Agent) is middle. The nuclear experiential structure of a middle clause consists of a Process + a Medium. The Medium, which is the most nuclear participant, is 'the one [participant] through which the process is actualized, and without which there would be no process at all' (Halliday 1985a: 146). On the other hand, a clause which represents a process brought about by some external cause (by an Agent) is effective. Very few processes in French display intransitive/transitive tendencies as chasser does in Examples 3 and 4. The semantics of the transitive model is extension: 'does the Process extend beyond the Actor?', 'is the Process directed at another participant?' (see Halliday 1985a: 145). Thus, from a transitive perspective, the variable is presence or absence of Goal. transitive pattern (3)
Le chat The cat Ac
chasse chases Proc: mat
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION (4)
Le chat
chasse
la souris
The cat Ac
chases Proc: mat
the mouse Goal
61
Zribi-Hertz (1987) surveyed some 5,400 so-called transitive verbs of French compiled by the research team of LADL (Laboratoire d'Automatique Documentaire et Linguistique, Universite de Paris 7), and found that among those, 1,700 had the potential to be used with the 50-clitic in a middle clause. She shows that what she calls CRE (construction reflexive ergative) is a 'productive structure in French today' (Zribi-Hertz 1987: 35). Moreover 'the CRE enjoys the productivity of various derivational processes, with the suffixes -iser, and -ifier, applicable to adjectives ending in -aire (tabulaire-(se) tabulariser), -able (malleable-(se) malleabiliser), -ible (flexible-(se) flexibiliser, -al (mineral-(se) mineraliser), -il(e) (fossil-(se) fossiliser), -ique(mecanique-(se) mecaniser, -ier (singulier(se) singulariser,-in (feminin-(se) feminiser . . .' (Zribi-Hertz 1987: 35 [my translation]). Thus, the productive nature of the S£-clitic allows many so-called transitive verbs to be used in a middle clause. This is a very important feature of the French language which can be seen to correlate with a generalized semantic change in the language from transitive to ergative. Indeed, we will see in Section 3.7 that in addition to being a productive lexicogrammatical resource, the ergative model of participation is also becoming more conspicuous in some registers or text types. For example, in the analysis of scientific registers as well as in 'hard news' (i.e. reports of accidents, riots, disasters, etc.), the interpretation of the Process as 'having or not having a causer external to the Process+Medium combination' (the ergative model: effective/middle) provides more insight than the interpretation of the Process as 'extending or not extending beyond the Actor' (the transitive model: transitive/intransitive). Even in narratives, where the transitive model typically surfaces by foregrounding 'who did what', 'who did what to whom' or 'who was affected?', there is also a great deal of variation across literary genres. In the writings of new novelists, for instance, middle material clauses are used in descriptive parts, instead of relational clauses, to bring the settings to life, as in Examples 5 to 7 below. (See Halliday 1985a/1994 and Halliday and Matthiessen 1999 for a more detailed discussion of transitivity models.) The prominent role of the middle clause in French texts will become apparent throughout this chapter. Some preliminary illustrations of the uses of the middle clause across different text types follow. Examples 5 to 7 are taken from Robbe-Grillet's (1957) 'New Novel' La Jalousie, where the background to the story is construed essentially by causeless happenings, realized as middle clauses. In the Examples below, function labels are in brackets and precede the element that hold the function; [Circ] stands for circumstance. The Medium is the function essential to the realization of any process. Although we have not yet introduced the PROCESS TYPE system, we include below both the process type and agency analyses for future reference. For
62
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
example, clause 5a is analysed as being both middle (AGENCY) and material (PROCESS TYPE). clause: material and middle (5a) [Med] Les boucles noires et brillantes [Proc] s'immobilisent, [Circ] dans I'axe du dos, The black and shiny curls stop in line with the back clause: material and effective (5b) que [Proc] materialise [Circ] un pen plus has [Ag] I'etroite fermeture metallique de la robe which is materialized a little further below by the narrow metallic zip of the dress clause: material and middle (6) [Circ] A I'autre bout de cette branche ouest de la terrasse, [Proc] s'auvre [Med] I'office At the other end of the western side of the terrace, opens the office clause: relational and middle (7) [Circ] Plus a droite [Proc] se dessine, [Circ] sur la peinture grise du mur, [Med] I'ombre agrandie etfloue d'une tele d'homme - celle deFranck More on the right becomes apparent on the grey paint of the wall, the enlarged and blurred shadow of a man's head - that of Franck
The use of middle clauses, and especially middle material clauses, in the New Novel gives a dynamic view of the settings which tend to be represented as static in traditional novels. Examples 8 to 10 are taken from a news article on the trial of Klaus Barbie written in Le Monde (1987). In this description of the trial, all human participants are left out: we have a passive clause where the agent is omitted and three middle clauses with inanimate actors. This foregrounds the ritualistic nature of the trial as a procedural activity. The use of middle clauses serves to give a sense of detachment, and foregrounds that no external force can stop the trial. It also serves to represent the stages that lead to the start of the trial as generalizations. clause: material and effective: passive: Agent implicit (8) Historique ou non, un proces est conduit selon uneprocedure. Historic or not, a trial is conducted according to a procedure. clause: relational and middle (9) Aussi bien celui de Klaus Barbie s'est engage sur les mots rituels maintes fois entendus dans unpretoire d'assises. Just as well the one of Klaus Barbie began with the ritual words heard time and time again in a court. clause: relational and middle (lOa) Lorsque lundi 11 mai, a 13h03, la porte a double battant s'est lentement ouverte pour I'entree du president Andre Cerdini, de ses deux assesseurs suppleants, When on Monday, May 11, at 1.03pm, the French door slowly opened for the arrival of His Honour Andre Cerdini, and his two assessors,
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
63
clause: relational and middle (lOb) un enorme silence s'estfait. a great silence happened.
The important role of the middle clause in French is also noted with regard to spoken discourse by Lambrecht (1988: 135), who suggests: the preferred clause of spoken French . . . is of the form V(X). In this preferred clause structure the subject is an incorporated pronoun which is also unmarked topic of the clause. Full lexical NPs [noun phrases] follow the verb and are semantically non-agentive and pragmatically non-topical. Furthermore, normally not more than one lexical NP constituent per clause occurs as a primary grammatical relation.
Lambrecht (1988: 136) goes on to say: The second important fact about the grammar of spoken French - which correlates directly with the first - is that the language provides its speakers with systematic means of preserving this preferred clause type, in the form of ready-made grammatical constructions whose main function is to allow lexical NPs to occur elsewhere than in initial subject position. . . . These preferred-clause-structurepreserving constructions are of two major structural types: so called 'dislocations' and so called 'clefts'.
It is the clause nucleus realized by a Medium-Process combination that Lambrecht refers to as the preferred clause of spoken French. Throughout this chapter, it will become apparent that the middle clause is not only the 'preferred clause' of spoken French but also of many French written registers. The French language provides, in addition to transitivity resources for construing agentless clauses, not 'ready-made grammatical constructions' as Lambrecht suggests but textual resources such as VOICE and THEME which allow the speaker/writer to manipulate the ordering of elements in the clause so that certain elements occur in prominent position, and others are omitted, in particular the Agent. As the textual system of VOICE is dependent on the system of transitivity, we will discuss it in Section 3.4, as well as returning to it in Chapter 5 when discussing textual grammar. Section 3.2 will introduce the two semantic models of transitivity, the ergative and the transitive, in some more detail. In Section 3.3, a systemicfunctional approach to French process types will be presented. Then, in Section 3.4, the generalizing principle behind the ergative model will be illustrated by demonstrating how AGENCY is construed across process types. Section 3.6 will look at the realization of transitivity at verbal group rank and discuss the notion of nuclearity in relation to the various participants and circumstances associated with the process types. Finally, in Section 3.7, we return to discourse by exploring transitivity patterns in different text types, and how they encode different models of experience.
64
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3.2 Models of transitivity: complementary perspectives on macrophenomena The semantic relationship that is construed among processes and their participants may be viewed from various perspectives. The type of participantprocess interaction is dependent on the process type, but it also reflects some general semantic principle that applies to all process types. Halliday (1994: 162) argues: It is true that, from one point of view, all these types of process are different. Material, behavioural, mental, verbal, relational and existential processes each have a grammar of their own. At the same time, looked at from another point of view they are all alike.
These two viewpoints reflect the transitive and ergative models respectively. Traditionally, the transitive perspective is represented by clauses of 'doing', i.e. processes of material action where the basic question is whether or not the process extends to another participant, a Goal, by impacting on it or affecting it in some way. However, our world of experience consists of fields other than those of material 'doing' and 'happening', and thus we can recognize other process types. We do not simply act or act upon things; we also reflect on our experience of the world. The category of mental process serves to represent our consciousness; it embodies the resources that allow us to reflect and to project our inner thoughts, desires and emotions. We also use experiential grammar to represent the relations that exist among phenomena; the grammar of relational processes gives us the resources for relating fragments of experience, through identification and attribution. These three main ways of modelling experience - as processes of doing, projecting or being - are realized in the lexicogrammar by the PROCESS TYPE system, which is explored in detail in Section 3.3. The ergative model, on the other hand, recognizes not the distinction between different realms of experience, but the distinction between processes that are brought about by some cause external to the Process + Medium nucleus and those that are 'self-engendered'. This distinction is reflected by the opposition between effective and middle in the AGENCY system. Figure 3.2 displays the PROCESS TYPE and AGENCY systems. The curly bracket { in Figure 3.2 indicates that the two systems AGENCY and PROCESS TYPE are simultaneous. This simultaneity represents the complementarity of the ergative and transitive perspectives. Thus, any clause may be analysed in terms of both ergative and transitive functions. The latter are process specific, and are indicated on the right of the system network in Figure 3.2. On the other hand, ergative functions such as Medium, Agent and Range can be generalized to all major process types. These will be explored in more detail in Section 3.5. The correspondences between the ergative and transitive function labels are shown in Table 3.2.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
65
Figure 3.2 Least delicate TRANSITIVITY systems (Key: The downward pointing arrow underneath the feature 'effective' means 'realized by', i.e. the feature 'effective' is realized in the structure by the presence of an Agent.)
Table 3.2 The complementarity of the ergative and transitive models ergative model middle
effective
doing
[Ac/Med] Laporte s'ouvre The door opens
[Ac/Ag] Pierre ouvre [Goal/Med] laparte Pierre opens the door
projecting
[Se/Med] Pierre s'alarme Pierre worries
[Phen/Ag] La situation alarme [Se/Med] Pierre The situation worries Pierre
being
[Val/Med] Ette s'appelle [Tok/Range] Marie She is called Marie
[Ass/Ag] Sa mere [Val/Med] I'appelle [Tok/Range] Marie Her mother calls her Marie
Key: ergative functions: Ag = Agent, Med = Medium; transitive functions: Ac = Actor, Se = Senser, Phen = Phenomenon, Val = Value, Tok = Token, Ass = Assigner
Each participant in the examples in Table 3.2 is analysed in terms of both the transitive and ergative functions to highlight the simultaneity of the PROCESS TVPE and AGENCY systems and the complementarity of the transitive and ergative models. For example laportein la porte s'ouvre is both
66
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Actor and Medium. Function labels and their abbreviations (which will be used throughout this chapter) are listed in the key under the table. While the transitive perspective projects a classificatory view of the world which reflects different domains of experience, such as doing and happening, thinking and saying (projecting), and being and having, embodied in the different process types, the ergative perspective projects a generalizing view of the world which does not discriminate between processes but focuses on whether the process is self-engendered or not (i.e. caused by an external Agent). This complementarity is diagrammed in Figure 3.3. Figure 3.3 shows that the transitive model is related to the PROCESS TYPE system in the grammar, while the ergative model is related to the AGENCY system. However, it is not the case that the transitive model with which the specific transitive functions are associated represents a verb-based approach to transitivity, while the ergative model represents a clause-based approach to transitivity. As Figure 3.1 showed, both models are clause based and both are schematic for lexical distinctions and collocational patterns. Thus, specific transitivity functions do not serve to subclassify the general ergative functions; both function together to subclassify the clause (see Martin 1994). The complementarity of the transitive and ergative models of transitivity, as we have seen, implies that both perspectives are simultaneously constructed by the experiential grammar. This was formalized systemically in Figure 3.2 and is now illustrated by the structural analysis of a text segment taken from 'La plage' (Robbe-Grillet 1966) both in terms of its transitive and ergative functions. The following text typifies the literary genre of the New Novel in various respects: first the setting in time is the 'present' rather than the 'past', and
Figure 3.3 Relationship between the complementary transitivity models and the simultaneous transitivity systems
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
67
the processes are essentially middle, while participants are inanimate. The effect is visual and dynamic and one of immediacy. What this text does is to represent the ongoing movement of a wave in contrast with the stillness of the weather, as well as the non-agentive aspect of natural phenomena. The static and dynamic construal of the 'weather' and the 'wave' respectively are emphasized by the choice of process types which reflect the transitive model. The shift from static to dynamic is foregrounded by a shift from relational processes to material processes. The stillness of the weather, lack of wind, cloud and blueness of the sky are constructed by relational clauses, the wave's ongoing movements by material processes. Note also the choice of circumstances of Manner in material clauses as a means of further specifying the movement of the wave. On the other hand, an ergative perspective on Text 3.1 projects a view of the world made up mostly of causeless happenings; when an effect is produced, it is caused by a natural phenomenon (Clause 9a). (In addition to the abbreviations given in the key to Table 3.2, the following abbreviations for function labels are used in Text 3.1: Proc = Process; rel = relational; mat = material; ex = existential; mid = middle; eff = effective; Ca = Carrier; Att = Attribute; Ex = Existent.) Text 3.1 From 'La plage' (Robbe-Grillet 1966) 1 relational & middle [Ca/Med] II [Proc] fait [Alt/Range] tres beau. It is a very fine day. 2 material & effective [Ac/Ag] Le sole.il [Proc] eclaire [Goal/Med] le sable jaune [Manner] d'une lumiere violente, verticale. The sun illuminates the yellow sand with a violent, vertical light. 3 existential & middle [ ] If- [Proc] n'y a pas [Ex/Med] un nuage [Place] dans le del. There is not a cloud in the sky. 4 existential & middle [ ] // [Proc] n'y a pas, [ ] nonplus, [Ex] devent. Neither is there any wind. 5a relational & middle [Ca/Med] L'eau [Proc] est [Att/Range] bleue, calme, sans la moindre ondulation [[venant du large]], The water is blue and calm, without the faintest swell from the open sea, 5b relational & middle [ ] bien que [Ca/Med] la plage [Proc] soit [Att/Range] ouverte [Place] sur la mer libre, jusqu 'a I'horizon. although the beach is completely exposed onto the open sea as far as the horizon. 6a material & middle [ ] Mais, [Manner] a intervalles reguliers, [Ac/Med] une vague soudaine, toujours la meme,«nee a quelques metres du bord», [Proc] s'enfle [Manner] brusquement But, at regular intervals, a sudden wave, always the same, originating a few yards away from the shore, suddenly rises
68
6b
7
8
9a
9b
9c
9d
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
material & middle [ ] et [Proc] deferle aussitot, [Manner] toujours sur la meme ligne. and then immediately breaks, always in the same line. relational & middle [Ca/Med] On [Proc] riapas alors [Att/Range] Vimpression [[quel'eau avance, puis se retire]]. And one does not have the impression that the water is flowing and then ebbing. relational & middle [Ca/Med] C'fProc] est, [ ] aucontraire, [Att/Range] [[commesitoutcemouvement s'executait sur place]]. On the contrary, it is as if the whole movement were being accomplished in the same place. material & effective [Ac/Ag] Legonflement de I'eau [Proc] produit d'abord [Goal/Med] une tigere depression, [Place] du cote de la greve, The swelling of the water at first produces a slight depression on the shore side material & middle: ranged [ ] et [Ac/Med] la vague [Proc] prend [Range] unpeu derecul, [Manner] dans un bruissement de graviers roules; and the wave recedes a little, with a murmur of rolling gravel; material & middle
[ ] puis [Ac/Med] elle [Proc] eclate et se repand, [Manner] laiteuse, [Place] sur la pente, then it bursts and spreads milkily over the slope, material & middle [ ] mats pour [Proc] regagner [ ] seulement [Range] le terrain perdu. but is merely regaining the ground it has lost.
There are a number of processes of movement which are embedded as constituents of relational clauses, as in clauses 5a, 7 and 8. These processes have not been analysed but they are significant in that they are all middle and material, foregrounding further the self-engendering character of natural phenomena such as waves. The transitivity selections of Text 3.1 are listed in Table 3.3. Table 3.4 displays frequency patterns. This highlights the prominence of middle clauses in Text 3.1, both material and relational. A simultaneous reading of the representation of experience in terms of the ergative and transitive perspectives therefore brings out the two levels of ideational meaning that are embodied in the text. On one level the text is about the movements of a phenomenon, the wave (construed as material clauses), in contrast with the stillness of the weather (construed as relational clauses). On another level, it is about how we perceive natural phenomena, as causeless happenings (construed as middle clauses). The participant role patterns found in Text 3.1 are summarized in Table 3.5. This brings out the way in which Robbe-Grillet uses the grammar to construe a fragment of a model of the world. Following this general discussion on transitivity models, we move on to discuss the French systems of transitivity in detail.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
69
Table 3.3 Transitivity patterns in Text 3.1 Clause
Process
transitivity functions
ergative functions
1
relational & middle material & effective existential &; middle existential & middle relational & middle relational & middle material & middle material & middle relational & middle relational & middle material & effective material & middle material & middle material & middle
Carrier, Attribute Actor, Goal Existent Existent Carrier, Attribute Carrier, Attribute Actor (Actor) Carrier, Attribute Carrier, Attribute Actor, Goal Actor, Range Actor (Actor) , Range
Medium, Range Agent, Medium Medium Medium Medium, Range Medium, Range Medium (Medium) Medium, Range Medium, Range Agent, Medium Medium, Range Medium (Medium), Range
2 3 4 5a 5b 6a 6b
7 8 9a 9b 9c 9d
Table 3.4 Quantitative patterns in the transitivity selections of Text 3.1 Voice
Process type material
relational
middle
5 + 3 embedded
7 (existential: 2; attributive: 5)
effective
2
3.3 A systemic-functional approach to French Process types In the introductory sections to this chapter, we saw that the transitivity system of French comprises two simultaneous systems, AGENCY and PROCESS TYPE, which reflect two models of transitivity, the ergative and the transitive, respectively. In the present section, we focus on the PROCESS TYPE potential, while in the next Section, 3.4, we explore AGENCY and how it generalizes across process types. The primary process type system presented in Figure 3.2 - i.e. the distinction between doing, projecting and being processes - has been determined on the basis of a number of grammatical criteria. Needless to say, the grammatical basis of this classification of process types by SFL is a reflection of semantic differences. The different grammatical reactances and probes that serve to differentiate the various process types are discussed in subsections 3.3.1, 3.3.2 and 3.3.3.
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
70
Table 3.5 Summary of participant role patterns in Text 3.1
doing material Actor
middle
effective
Medium
Medium
Agent
wave (mobile):
natural force:
(6a) une vague soudaine (6b) [0] (9b) la vague (9a) elle (= la vague) (9b) [0]
(2) le soleil (9a) le gonflement de 1'eau
sand (immobile):
Goal
(2) le sable jaune (9a) une legere depression (9d) le terrain perdu being
relational: attributive Carrier
existential Existent
natural features:
(1) 11 ('day') (5a) 1'eau (5b) la plage (7) On (8)Ce natural feature: sky
(3) un nuage; (4) du vent
3.3.1 Doing clauses The 'doing' process type covers processes of 'doing' and 'happening'. As seen in Figure 3.4, there are three subtypes of doing process: material, behavioural and meteorological. One grammatical criterion that distinguishes doing processes from projecting and being processes is that they can be probed with the general processes faire (do) or arriver (happen). Another distinguishing criterion relates to the choice of unmarked past tense with the different process types. Consider the following examples taken from Henri Thomas's 'The Offensive' (1966), which will be discussed below. Process: doing: material and effective; TENSE: simple past (lla) [Ac] Claude [Proc] enleva [Goal] son sac [Place] de ses epaules
Claude took down the bag from his shoulders (lib) [Goal] le [Proc] fourra [Place] dans I'un de ces renfoncements shoved it into a recess
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
71
Process: doing: material and middle ( l i e ) [ ] oil [Ac] il [Proc] s'allongea [ ] ensuite [Manner] la tete sur le sac, son fusil contre lui where he lay down then, his head on the kit bag, his rifle by his side
One may probe all three of these clauses with quefit Claude'? (what did Claude do?). This is not the case for Examples 12a and 12b, which are relational and mental process clauses, respectively. Process: relational: attributive; TENSE: imperfect past (12a) [Ca] Lejour, dans lefond du ravin, [Proc] etait [Alt] gris [Comparison] comme un crepuscule Day at the bottom of the ravine was a grey twilight Process: mental: perceptive; TENSE: imperfect past (12b) [ ] et [Se] Claude [Proc] nevoyaitplus [Phen] le del and Claude could no longer see the sky
With these Process types, one cannot use the faire probe: one cannot probe le jour etait gris with qu'est-ce que le jour faisait? or Claude voyait le del with qu'est-ce que Claude faisait? Note that behavioural process clauses can be semantically similar to mental process clauses, as in je regarde la television (I am watching television), which represents a mental action. However, unlike mental processes they do not project (see Section 3.5) and they can be probed by faire; this is why they are interpreted as a subtype of the doing process rather than as a type of the projecting process (see subsection 3.3.1.2 on behavioural processes). The other criterion mentioned above that separates 'doing' processes from both 'being' and 'projecting' processes is their unmarked past tense. Doing clauses typically select for the 'simple past' or 'compound past' while 'being' and 'projecting' clauses tend to select for the 'imperfect past'. (See Caffarel 1992 on French tense.) What distinguishes simple past/compound past and imperfect past is aspect. The imperfect presents the situation from within, 'since it can both look backwards towards the start of the situation, and look forwards to the end of the situation' (Comrie 1976: 4), while the simple past/compound past present the situation from the outside. While qualities, emotions and feelings tend to be extended in time,3 actions tend to be transient, which explains the choices of unmarked past tense for doing processes on the one hand and relational and mental processes on the other hand, as illustrated by Examples 11 and 12 above. A systemic representation of the grammar of doing processes is presented in Figure 3.4. As the system in Figure 3.4 shows, doing clauses cover material processes, behavioural processes and meteorological processes. Material processes are the major type of doing processes. This is determined on the basis that they are not restricted in AGENCY, i.e. they can be either middle or effective. In addition, material processes are realized by a fairly open set of verbs. On the
72
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 3.4 Some more delicate options for doing: material processes other hand, both behavioural and meteorological processes are realized by a closed set of verbs and are typically only middle. This is because they represent happenings rather than doings. However, there may be some atypical or extraordinary circumstance where a behavioural process or meteorological process may be constructed as effective (with an Agent) as in Example 13b below, which is taken from a news article from Le Monde (April 29, 2004) on 'the wall of shame'. Process: material and effective (13a) [Ag/Ac] On [Med/Goal] les [Proc] aligne [Place] lelongdumur They line them up against the wall Process: behavioural and effective; analytic (13b) [ ] en [Med/Goal] les [Proc] obligeant a regarder [Range] le soleil obliging them to watch the sun
As we saw in Chapter 2 in relation to clause complex features, some combinations of features are always more likely to occur than others. However, a high or low probability of occurrence does not of course indicate that a particular combination of features will always or never occur.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
73
The system network in Figure 3.4 reads as follows: one may choose between three types of 'doing' processes: material, behavioural or meteorological. If the option [material] is chosen, then the clause may be either [middle] or [effective]. If the option [middle] is chosen, then one may choose between [ranged] or [non-ranged]. If the option [effective] is chosen, then more delicate options are open, among them processes indicating a [transfer] and processes indicating a [service]. The empty options indicate that there are types of material effective processes other than those expressing a transfer or a service. If the option [transfer] is chosen, then there is contrast between [recipiency] and [non-recipiency]. Similarly, if one choses the option [service], there is a contrast between [cliency] and [non-cliency]. The SERVICE system is also open to clauses which are material and middle: ranged; this is indicated by the marking convention if A then B. Material processes of doing can have a third participant, a Beneficiary, which may be a Client or a Recipient depending on whether the process represents a 'done for', i.e. a service, or a 'done to', i.e. a transfer. Middle ranged clauses may also have a Client, as in elk a joue au basketball pour I'equipe de Vuniversite de Sydney (she played basketball for the Sydney University team). Features from the network are further illustrated in the following subsections. 3.3.1.1 Material clauses
Participants specific to material clauses are Actor, Goal, Range, Client or Recipient. A middle material clause can have just one participant, an Actor, two participants, Actor + Range, or three participants, Actor + Range + Client. An effective material clause can have two participants, Actor + Goal, or three participants, Actor + Goal + Client/Recipient. Consider the following examples. Process: material and middle: ranged (14) Remi joue Remi plays Ac/Med Proc
au tennis tennis Range
The function Range serves to specify the range of the process or to elaborate on the process, but is not impacted upon as the Goal is (see Section 3.4 on Range). Process: material and middle: non-ranged (15) Louis court Louis is running Ac/Med Proc Process: material and middle: ranged; service: cliency (16) Yannick joue au football pour I'equipe de Bordeaux Yannick plays football for the Bordeaux team Ac/Med Proc Range Client
74
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Process: material and effective (17) Louis ecrase Louis squashes Ac/Ag Proc
unescargot a snail Goal/Med
Process: material and effective: transfer: recipiency (18) Remi d u donne Remi gives an apple Ac/ proc Goal
sonfrere to his brother Recipi Beneficiary
Process: material and effective: transfer: non-recipiency (19) Pierre a envoye une lettre recommandee Pierre sent a registered letter Ac/Ag Goal med prc Process: material and effective: service: cliency (20) Yannick aachete un gateau Yannick bought a cake Ac/Ag Proc Goal/Med
pour sa mere for his mother Client/Beneficiary
Process: material and effective: service: non-cliency (21) Yannick aachete un gateau Yannick bought a cake Ac/Ag Proc Goal/Med For attempts at extending material grammar towards lexical delicacy in English, see Hasan 1987; Matthiessen 1995.
3.3.1.2 Behavioural clauses Behavioural clauses represent physiological processes as well as mental or verbal activities. The participant essential to the realization of the Process is termed the Behaver. Behavioural clauses differ from mental and verbal processes in that they cannot project, but are similar to mental ones in that the Behaver is always endowed with consciousness. Behavioural processes that construe a mental activity can have a second participant, the Phenomenon. Unlike the phenomenon of mental clauses (see Section 3.3.3), it cannot be metaphenomenal or macrophenomenal. Process: behavioural: physiological (22) Yannick a eternue Yannick sneezed Behaver/Med proc Process: behavioural: verbal action (23)
Louis Louis Behaver/Med
parle speaks Proc
sans arret non-stop Manner
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION Process: behavioural: mental action (24) Remi ecoute Remi is listening to Behaver/Med Proc
75
la radio the radio Range
3.3.1.3 Meteorological clauses
Meteorological grammar in the field of doing is realized by conflating Actor and Process. The element His not a participant, i.e. it is not representational. Other Romance languages, for example Italian, make this explicit by construing our experience of the weather as a Process alone: the 'dummy' Subject is not overt as in piove (it rains, ilpleut). Process: doing: meteorological and middle (25) // vente It winds (is windy) Proc/Ac/Med
Experience of the weather can also be construed as 'being' clauses of the existential type, as in Examples 26 and 27. Process: being: existential and middle (26) // ya (There) there is there Proc Process: being: existential (27) // fait It makes Proc 'It is windy'
du vent wind Ex/Med duvent wind Ex/Med
3.3.2 Projecting clauses
Projecting clauses include mental clauses (construing internal consciousness) and verbal clauses (construing external consciousness). Projecting clauses construe consciousness as symbolic processing capable of creating symbolic content construed by a separate clause. Projecting processes, as their name implies, may project. As we saw in Chapter 2, projection may be done paratactically, by quoting, or hypotactically, by reporting. This characteristic of mental and verbal processes is one of the criteria that sets them apart from all other process types. This is illustrated in Example 28b, which projects clause 28a paratactically. (This example is taken from Henri Thomas's 'The Offensive' (1966).) New abbreviations in this section are: poss = possessive, Phen = Phenomenon, metaphen = metaphenomenon and ment = mental. Verbal clause: projecting: quote (clause 28b projects clause 28a) (28a) Ils n'ont qu' une trentaine de canons, They' ve only got thirty guns Ca Proc: rel: poss Att
en face over there Place
76
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(28b) dit said Proc: verbal
Fremigacci Fremigacci Sayer
Following Halliday (1985a), a projected clause is not interpreted as a participant serving as a constituent of a mental clause but as a separate clause. Unlike participants, projected clauses cannot be thematized or nominalized. In Example 29a below, the Phenomenon is a clause, but a downranked one that functions as participant. As such it can be thematized, as shown in Example 29b. Process: mental: emotive and middle (29a) // await tellement aime He would so much have loved Se Proc (29b) Rester avec eux, To stay with them Theme4
il he Rheme Se
[[rester avec eux]] to stay with them Phen: metaphen
aurait tellement aime would have really liked
[[faire fa]] to do this
Proc
Phen: metaphen
Example 29 illustrates another characteristic of mental clauses which concerns the nature of the Phenomenon. It may not just be phenomenal (i.e. a thing) as are participants in material clauses; it may instead be macrophenomenal (i.e. a fact) or metaphenomenal (i.e. a non-projected clause), as in Example 29. Example 30 illustrates a phenomenon as fact. (30) // He Se
s'inquietait worried Proc
dufait que son cceur battait trap vite (by the fact) that his heart was beating too fast Phen: fact
The Phenomenon in Example 30 is not a projected clause; it functions as a participant and as such can be thematized as in Example 31. (31)
Le fait que son cceur battait trop viteThe fact that his heart was beating too fast Phen
I' him Se
inquietait worried Proc
Examples 30 and 31 illustrate an additional characteristic of mental clauses, and in particular of emotive mental clauses: the fact that they can be bidirectional. More precisely, the sensing may be construed as the Senser having an emotion 'ranging' over the Phenomenon as in Example 30, or as the Phenomenon causing the emotion as in Example 31. We will see in Section 3.4 that emotive clauses that construe the sensing from Senser to Phenomenon are middle, while clauses that construe the sensing from Phenomenon to Senser are effective in the sense that the Phenomenon is represented as causing, or as being responsible for, the emotion felt by the
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
77
Senser. Further examples illustrating the bidirectionality of mental processes follow. (32) // He Se/Med (33)
La musique Music Phen/Ag
dime likes Proc: ment: emotive and mid passionne impassions Proc: ment: emotive and eff
la musique music Phen/Range Henri Henri Se/Med
A further characteristic of mental processes concerns the nature of the Senser (the participant essential to mental processes), which, unlike the Actor of material clauses, must always be endowed with consciousness. If a normally non-conscious thing is construed as Senser, it is personified. The Sayer of a verbal clause, in contrast, can sometimes be a non-conscious participant, as le journal (the newspaper) is in Example 34. (34)
Le journal rapporte que les affaires de Bond vont mal The newspaper states that Bond's businesses are not going well
However, in such cases, the Sayer typically stands in a lexical relation (meronymic or hyponymic) with a conscious being, in this case the journalist. Verbal clauses, like mental clauses, need not always project, but they have the potential to do so. While mental clauses project ideas, verbal clauses project locutions (referred to as 'direct' and 'indirect' speech in traditional grammars). If they do not project a locution, verbal clauses are ranged, that is, they have a participant which specifies the name of the verbalization - a Verbiage, as in Example 35, which also shows that verbal clauses may have an additional participant, the Addressee, i.e. the beneficiary of the verbalization. (35) // He Sayer
a demande asked Proc
un renseignement advice Verbiage
asonvoisin to (from) his neighbour Addressee
The system network of French projecting processes is represented in Figure 3.5 below. Further illustrations of the features from the network of projecting processes shown in Figure 3.5 follow. Examples 36 to 41 are taken from Sainte-Soline (1972). Process: verbal: verbalization: locution and no addressee and middle (36a) "1 Plus fa va, plus t'arrives de bonne heure You are arriving earlier and earlier
(36b) 2
fit said Proc
ill he Sayer
78
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 3.5 A systemic representation of projecting processes (adapted from Matthiessen 1995) Note that the verbal process in Example 36b is realized by faire, which is typically a 'doing' process but is often used as a verbal process in literary texts. In the field of processes, faire is like a chameleon as it is also used as a relational process to talk about the weather, as in ilfait beau (the weather is nice). This shows that the meaning of a process cannot be analysed independently of the clausal configuration in which it occurs. Process: verbal: verbalization: as locution and middle (37a) a II repondit He answered Sayer Proc (37b) "|3
qu' that
il he Se
nesesouciaitpas did not care Proc: ment: emotive
delabeaute. about beauty. Matter
Process: verbal: verbalization: as name and middle (38) La radio a annonce de I'orage. The radio announced storms. Sayer/Med Proc Verbiage/Range Process: mental: emotive and metaphenomenal: fact and effective (39) [[De se trouver si tot debout au milieu des champs]] la To find herself up so early in the middle of the fields her Phen: metaphen: fact/Ag Se/Med
surprenait surprised Proc Med
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
79
Process: mental: perceptive and phenomenal and middle (40)
II
n'entenditpas
la question.
He Se/Med
did not hear Proc
the question. Phen/Range
Process: mental: cognitive and metaphenomenal: idea and middle (41a) a
Je
pense
I Se
think Proc
(41 b) ' P
qu'ila reve ton laitier
that he dreamt, your milkman
Verbal processes with no verbalization either as verbiage or locution include what Matthiessen (1995: 285) refers to as processes of verbal impact, such as accuser, feliciter, critiquer, louer, etc., where the verbalization can be seen as being conflated with the process. Such verbal processes represent a closed set. They are interpreted as effective in the sense that they impact verbally on a participant, the Target, which is analogous to the Goal of a material clause, as in Example 42. (42) Je
I Sayer/Ag
te
felicite
you Target/Med
congratulate Proc: verbal
These processes cannot project as the verbalization cannot be separated from the process, but the reason behind the verbal impact is often realized as an hypotactic enhancing clause or a circumstance of cause or matter, as in Examples 43 and 44. (43a) a
III'a accuse He accused him (43b) xp de mentir of lying (44) // le critique precisement a cause de (a
He criticizes him precisely because of this
With regard to mental processes, we have so far had examples of perceptive, emotive and cognitive processes but not of desiderative processes. While perceptive and cognitive processes project in the indicative, emotive and desiderative processes tend to project ideas in the subjunctive mode (see Chapter 2) where the event projected is represented as 'irrealis', and rather than having temporal reality carries modal meaning such as low probability or high obligation as in the following examples. mental: desiderative (high obligation) (45a) a
Pierre veut
(45b) 'P
Pierre wants quejeanparte thatJean go
80
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
mental: emotive (low probability) (46a) a Jecrains I fear (46b) P qu 'il ne soil malade that he is ill
With the exception of vouloir, which can be both a mental process and a modal expansion within the verbal group, other modal processes cannot project an idea and only serve to modulate the Process they precede. They are options at verbal group rank rather than clause rank. Consider the following examples. (47a) a (47b) (48) (49) (50) (51) (52)
Pierre veut Pierre wants 'P quej'achete une voiture that I buy a car Pierre veut avoir une voiture Pierre a wants P to have a car Pierre veut reflechir Pierre a wants P to think Pierre doit partir Pierre must go Pierre doitfinir son travail Pierre must finish his work Ces vetements peuvent se laverfrequemment These clothes can be washed often
The Senser of a modal verb is not always a conscious participant as with mental processes, as shown by Example 52. In Example 47 the modal verb truly functions as a mental Process in the sense that it is projecting an idea. In the other examples (48 to 52), the modal verbs modulate the main process. Thus only the verb vouloir, when projecting an idea, can truly be considered as a process functioning at clause rank. In contrast, neither pouvoirnor devoir can project. Rather, they modify the main Process modally (see Chapter 4 on modality) and, as a consequence, can be used in clauses with non-conscious participants, as in cet appartement doit etre cher (this apartment must be expensive) or ce livre pent se lire fadlement (this book can be read easily). 3.3.3 Being clauses
Being processes have two main functions: they either represent a relation between two participants (relational processes) or express the existence of one participant (existential processes). The former type, relational processes, represent relations between two participants either through identification as Token and Value (identifying processes) or attribution as Carrier and Attribute (attributive processes).
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
81
These two main types of relational clause, identifying and attributive, are distinguished from one another by the reversibility criterion: identifying clauses are reversible whilst attributive are not. Consider the following examples. Process: relational: attributive (from Leeman-Bouix 1994: 33) (53) La realite de la langue est inaccessible, inobservable The reality of language is inaccessible, unobservable Ca/Med Proc Att/Range Process: relational: identifying (from Leeman-Bouix 1994: 59) (54) L 'objet [[que se donne le grammairien]] est The object to which the grammarian devotes herself is Val/Range Proc
la langue language Tok/Med
The reversibility criterion shows that the first of these is attributive and the second identifying, because the first is not reversible, i.e. one cannot say Inaccessible est la realite de la langue (inaccessible is the reality of language). In other words, the Attribute function can never serve as Subject. On the other hand, in a Token-Value relationship, the Subject may either be the Value, as in Example 54, or the Token, as in Example 55, which shows that Example 54 has an agnate reversed version. Process: relational: identifying (agnate reversed version of Example 54) (55) La langue, est I'objet [[que se donne le grammairien]] Language is the object to which the grammarian devotes herself Tok/Med Proc Val/Range Let us now explore relational clauses in more detail, beginning with existential ones.
3.3.3.1 Existential clauses Existential clauses, in contrast with other relational clauses, are always middle. The sole participant, the Existent, conflates with the Medium. Such clause types do not necessarily have a process, but simply what is traditionally called a 'presentative', i.e. an item that presents the Existent. Such an item will be referred to here as an 'existential particle'. (56)
[Existential particle] Void [Ex] un escargot Here is a snail (57) [Existential particle/Proc] Soit [Ex] un triangle ABC Let ABC be a triangle (58) Ry [Procrsqb; a [Ex] du brouillard There is fog
On the basis that Examples 56 and 57 do not have any process, we could argue that such clause types are minor, and as such should not be interpreted
82
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
as part of the major clause system. However, many languages do not have copula verbs. Furthermore, from an interpersonal viewpoint, clauses such as Examples 56 and 57 do not perform minor speech functions such as exclamations, calls or greetings, but participate in an exchange as propositions, either as a response statement accompanying an action, or as an initiating statement. Accordingly, such clause types whose function is to 'announce' a new participant will be interpreted as a subtype of existential clauses. Verbless existential clauses will be referred to as [presentative] while existential clauses with a verb will be referred to as [stative]. 3.3.3.2 Relational clauses While in existential clauses there is only one part to the 'being', 'in relational clauses there are two parts to the "being": something is being said to "be" something else. In other words, a relation is being set up between two separate entities' (Halliday 1994: 119). The French relational system consists of two main types of relational clause, identifying and attributive. In an identifying clause, a Value is assigned to a Token or a Token is assigned to a Value. The two functions, Token and Value, can be reversed in the clause, as was just outlined in relation to Examples 54 and 55. In contrast, the functions of an attributive clause, Carrier and Attribute, cannot be reversed: an Attribute is assigned to a Carrier, but a Carrier cannot be assigned to an Attribute. An attributive clause may be either intensive or non-intensive. The nonintensive system offers the choice between circumstantial and possessive. While in English the possessive type of relation cross-classifies attributive and identifying clauses, in French possessives can only be of the attributive type and the possessor Attribute is realized as a prepositional phrase, as are the circumstantial attributes. Compare the following examples from English taken from Halliday (1994) and their French translations: (59) The piano Tok/Possessed (60) Le piano Ca: possessed/Med
is Proc: identifying and intensive est Proc: attributive: intensive
Peter's Val/Possessor a Peter Att: possessor/Range
(61) Peter's Val/Possessor (62) * A Peter
is Proc (passive) est
the piano Tok/Possessed le piano
(63) Peter Tok (64) Pierre Ca
owns Proc: identifying and poss possede Proc: attributive: poss
the piano Val le piano Att
(65) The piano Val (66) * le piano
is owned by Proc (passive) est possede
Peter Tok par Pierre
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
83
We have seen that the grammatical criterion for distinguishing identifying clauses from attributive clauses is that identifying clauses can be reversed, i.e. either the Token or the Value may conflate with the Subject. In contrast, the Attribute of an attributive clause cannot function as Subject. In French, possession can only be realized as an attributive type, either as a participant as in Example 60, or as a possessive process as in Example 64, or both as in Example 67. Process: attributive: possessive (67) Lepiano appartient The piano belongs Ca/Med Proc
a Pierre to Pierre Att/Beneficiary
Clause 66, le piano est possede par Pierre, does not mean that the piano belongs to Pierre but that the piano is possessed by Pierre, and thus expresses some kind of supernatural phenomenon. The systemic representation of French 'being' clauses is schematized in Figure 3.6 and examples of realizations of relational clauses are represented in Table 3.6. Circumstantial attributive clauses are similar to possessive attributive clauses in that they can have the Attribute as Circumstance, as in Jean est a Paris, or the Process as Circumstance as in Example 68. (68) La fete The party Ca
a dure lasted Proc: attributive: circumstantial
toute lajournee all day Att
Sometimes the Attribute can be conflated with the process, as in Examples 69 and 70. (69) Ils They Ca (70) Lesfeuilles The leaves Ca
se ressemblaient look alike Proc/Att Jaunissaeint were becoming yellow Proc/Att
To conclude this section on process types, we will mention one final distinguishing criterion, the kind of auxiliary, avoir or etre, that is used to form complex tenses across clause types. The complex tenses of 'being' clauses, whether they have as relational process avoir or etre, are always formed with the auxiliary avoir as in Examples 71 and 72. (71) Leur chien a ete malade Their dog has been sick (72) Leur chien a eu une maladie Their dog has had an illness
84
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 3.6 A systemic representation of French relational processes Table 3.6 Examples of relational clauses identifying
attributive
intensive [Tok] Paris [Proc] est [Val] la capitate de la France Paris is the capital of France
intensive [Ca] Jean [Proc] est [Att] grand Jean is tall
non-intensive [Tok/Circ] La fete de la Bastille [Proc] est [Val/Circ] le lljuilkt Bastille Day is 14 July
extensive circumstantial [Car] Jean [Proc] est [Att/Circ] a Paris Jean is in Paris possessive [Car] Jean [Proc] a [Att] un piano Jean has a piano
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
85
Projecting clauses also use avoir as their auxiliary, as in Examples 73 and 74. (73) II a pense qu 'il viendrait He has thought that he would come (74) II a dit qu 'il viendrait He has said that he will come
Doing clauses, in contrast, may use avoir or Stress auxiliary depending on the type of process. Within material clauses, processes of movement typically form their complex tenses with the auxiliary etre, as in Example 75, while others typically take avoir, as in Example 76. This can be related to the degree of transitivity, as explained by Hopper and Thompson (1980): verbs with high transitivity select avoir, verbs with low transitivity select etre. This explains why all so-called 'pronominal verbs', whether material or mental, like se lever (get up), se nounrir (feed), s'evanouir (faint) and sefdcher (to get angry), which are used in middle clauses, form their complex tenses with etre as in il s 'est leve. material: movement (75) // est descendu He has come down the stairs material: non-movement (76) II a mange He has eaten
Although all process types select for etre to realize complex tenses in middle agency with se, material processes are the only process types which, without se, can realize complex tenses with the auxiliary etre. This characteristic of material processes is another distinguishing criterion. A summary of the criteria that distinguish the three major subtypes of doing, projecting and being processes, i.e. material, mental and relational processes, and an analysis of a text in terms of process types and transitive functions is presented in Table 3.7. The following extract taken from 'Le mur' (Sartre, 1939) was previously analysed in Chapter 2 in terms of clause complexes. The transitivity analysis of this text illustrates the use of the following process types: material, behavioural, mental, verbal and attributive (intensive, circumstantial and possessive). As so far we have not discussed circumstance types, they are simply marked as [Circ]. This text illustrates the interaction of the features attributive and effective in clauses 2a and 7a, where the Attributor is the Agent in the sense that it brings into existence the relation between the Carrier and the Attribute. A more detailed look at the interaction of the process type and agency systems is presented in the next section. (In Text 3.2, in addition to the abbreviations introduced earlier, the following abbreviations are used: beh = behavioural, Beh = Behaver, Attrib = Attributor, circ = circumstantial, verb = verbal and Verb = Verbiage.)
86
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 3.7 Criteria for distinguishing process types Grammatical probes
pro-verb faire, arriver
Process type material
mental
relational
YES Qu 'est-ce qu 'il lui a
NO
NO
fait? II I'afrappe. What did he do to him? He hit him. Qu'est-ce qui est arrive"? II a plu. What happened? It rained.
Project
NO
YES Elle pense qu 'elle reussira. She thinks that she will succeed.
NO
Metaphenomenon
NO
YES Elle pense [[partir en France]]. She is thinking of going to France.
NO
+ Medium endowed with consciousness
NO Le train arriva en retard. The train arrived late.
YES Elle pensa arriver en retard. She thought of arriving late.
NO Le train etait en panne. The train was broken down.
Auxiliary
Avoir or etre
Avoir
Avoir
Unmarked past
Simple/compound Imperfect past past
Imperfect past
Following this overview of the PROCESS TYPE system and how it is instantiated in text, we will now turn to the complementary transitivity system AGENCY, and illustrate the generalizing property of the latter.
a Text 3.2 From 'Le mur' (Sartre 1939) la [Ac] On [Goal] nous [Proc: mat] poussa [Circ] dans unegrande salle blanche, They pushed us in a large white room, Ib et [Beh] mesyeux [Proc: beh] se mirent a cligner and my eyes started to blink Ic parce que [Attrib] la lumiere [Ca] leur [Proc: att] faisait [Att] mal. because the light made them hurt. 2a Ensuite, [Se] je [Proc: ment] vis [Phen] une table et quatre types derriere la table, des civils, Then I saw a table and four guys behind the table, civilians, 2b [Beh] qui [Proc: beh] regardaient [Phen] despapiers. who were looking at papers. 3a [Ac] On [Proc: mat] avail masse [Goal] les autres prisonniers [Circ] dans lefond They had massed the other prisoners at the back 3b et il [Ac] nous [Proc: mat] fallut traverser [Range] toute la piece and we had to traverse all the room 3c pour [Range] les [Proc: mat] rejoindre. to join them. 4a [Proc: ex] II y en avail [Ex] plusieurs [[queje connaissais]] There were a few I knew 4b et [Ex] d'autres [[qui devaient etre etrangers]]. and other who must have been foreigners. 5a [Ca] Les deux [[qui etaient devant moi]] [Proc: att] etaient [Att] blonds avec des cranes ronds, The two in front of me were blond with round skulls, 5b [Ca] ils [ Proc/Att] se ressemblaient: They looked alike: 5c [Att] des Franc,ais, j 'imagine. (They were) French, I imagine. 6a [Ac] Le plus petit [Proc: mat] remontait tout le temps [Goal] son pantalon: The smallest was pulling his pants up all the time: 6b [Ca] c' [Proc: att] etait [Att] nerveux. It was a nervous reaction. 7a [Ca] Qa [Proc: att: circ] dura [Circ] pres de trois heures; It lasted nearly three hours; 7b [Ca] j' [Proc: att] etais [Att] abruti I was dazed 7c et [Ca] j' [Proc: att: poss] avals [Att] la tete vide and my head was empty 7d mats [Ca] la piece [Proc: att] etait [Att] bien chauffee but the room was well heated 7e et [Attrib] je [Proc: att] trouvais [Ca] fa [Att] plutot agreable: and I found this rather nice: 7f [Circ] depuis vingt-quatreheures, [Beh] nous [Proc: beh] n'avionspascessedegrelotter. For the past twenty four hours we had not stopped shivering. 8 [Ac] Les gardiens [Proc: mat] amenaient [Goal] les prisonniers [Circ] I'un apres I'autre [Circ] devant la table. The guards were bringing the prisoners one after the other in front of the table. 9 [Sayer] Les quatres types [Addressee] leur [Proc: verb] demandaient alors [Verb] leur nom et leur profession. The four men were asking them their name and their profession.
88
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3.4 AGENCY: an ergative perspective Agency embodies the meaning of causation, which is most easily perceived in material clauses because only in the field of doing has the Agent the physical power to impact on things. However, this section will illustrate that the Agent function is not restricted to doing clauses of the material type but is also to be found in projecting clauses of the mental type and being clauses of the relational type. On the other hand, as already mentioned, process types such as behavioural, verbal and existential processes are generally restricted in agency and as such are interpreted as minor process types, as opposed to material, mental and relational, which are interpreted as major process types. The introduction of this chapter suggested that the 'preferred' clause type in French was the middle clause. The nucleus of a French clause may consist of just a Process and a Medium through which the Process is actualized, as in la branche a casse (the branch broke); but it is likely to have the reflexive clitic seattached to the Process as in la branche s'est cassee (the branch broke). In this case, se explicitly indicates that the process is not caused by an external Agent; se does not function as a marker of reflexivity or reciprocity as it does in instances like il se lave (he washes himself) where se, one could argue, has the function of Goal, just as son chien does in il lave son chien (he washes his dog). La branche s'est cassee does not mean that 'the branch broke itself, as 'the branch' cannot impact on itself. Something or someone, left unexpressed, has caused the branch to break. In other words, the se-clitic in the example la branche s'est cassee (the branch broke) does not serve a participant function. Furthermore, it cannot be Subject in a passive variant and cannot be Theme-predicated. Thus, we cannot have c'est elle-meme, la branche qui s 'est cassee or elle-meme a ete cassee par la branche. On the other hand, lui-meme (but not se) in il se lave lui-meme can be Theme-predicated as in C'est lui-meme qui se lave, but it cannot be Subject as in *Lui-meme a ete lave par lui. Guillaume (1984: 133) argues that French has three voices, two analytic, the active and the passive voices, and one synthetic, the reflexive voice, the latter encapsulating all reflexive uses. In his view, the various uses of the reflexive constitute a continuum between the active and passive voices as shown in Figure 3.7. Pierre se deplace has active characteristics in the sense that Pierre is both 'actor of the movement' and 'affected by the movement'. Pierre s'ennuie feels truly middle in that Pierre is not imposing the boredom on himself, but is only affected by it. Ces choses se disent is not very different either semantically or grammatically from ces choses sont dites, which is the passive form, hence the use of middle clauses as passive substitute in the context of generic events. In the AGENCY system middle clauses are opposed to effective clauses. If we come back to Guillaume's examples, we can see that not all three examples have an effective counterpart, as suggested by the made up paradigm shown in Table 3.8.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
89
Figure 3.7 Guillaume's interpretation of French middle clauses, as both a synthesis of, and genesis between, active and passive Table 3.8 The AGENCY and PROCESS TYPE paradigm Process type
Agency middle non-ranged
effective ranged
material
Pierre se deplace Pierre moves
Jean deplace Pierre Jean moves Pierre
mental
Pierre s 'ennuie Pierre is bored
La musique ennuie Pierre Music bores Pierre
verbal
Range/Subject Ces choses se disent These things are said
Sayer/Subject Elles disent ces choses They say these things
The examples in Table 3.8 suggest that while material and mental clauses can clearly be either middle or effective, verbal clauses appear to be restricted in agency. That is, the feature [effective] does not appear to be open to verbal clauses, unless, as we saw in Section 3.3, they are clauses of verbal impact. Effective clauses and middle ranged clauses are both transitive in the traditional sense, in that they have a Process and two participants. However,
90
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
the secondary participant of the respective clause types is quite different. While in an effective clause the secondary participant (the Goal in a material clause and the Senser in a mental clause) is an impacted or affected participant, the secondary participant in a ranged middle clause, the Range, is a non-impacted participant: it elaborates on the Process or defines the scope of the Process, as shown in Table 3.9. As both effective clauses and ranged middle clauses have the structure participant 1 + Process + participant 2, only a functional analysis can bring out the semantic differences between these two types of experiential configuration. A clause with the features effective and analytic contrasts in its experiential semantics with a clause which is ranged and middle. This is not surprising as one clause is effective and the other is middle. On the other hand, a clause with the features effective and analytic is similar in its experiential semantics to a clause with the features effective and synthetic, as both are effective. This may seem obvious; however, it is not so obvious to a traditional or formal grammarian, who is not familiar with the distinction between effective and middle and who would interpret middle: ranged clauses and effective clauses as similar as they are transitive in the traditional sense. Examples 77 and 78 are both effective clauses, but while in Example 77 the feature [effective] is realized synthetically, with a verbal group simplex, in Example 78 the feature [effective] is realized analytically with a verbal group complex of which the a is the causative process faire. Both effective clauses construe the same event, 'Jean cooking potatoes'.
Table 3.9 Distinguishing effective clauses from middle: ranged clauses Process A Goal
Process A Range
[Act] Yannick [Proc] essuie [Goal] le piano Yannick is wiping the piano
[Act] Yannick [Proc] joue [Range] du piano Yannick is playing piano
[Act] Louis [Proc] nettoie [Goal] la baignoire Louis is cleaning the bath
[Act] Louis [Proc] prend [Range] un bain Louis is taking a bath
[Act] Remi [Proc] ecrase [Goal] une limace Remi is squashing a slug
[Beh] Remi [Proc] regarde [Range] une limace Remi is looking at a slug
[Act] Bruno [Proc] ecrit [Goal] une chanson Bruno is writing a song
[Act] Bruno [Proc] chante [Range] une chanson Bruno is singing a song
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
91
(77) Jean cuit les pommes de terre Jean cooks the potatoes (78) Jean fait cuire les pommes de terre Jean makes the potatoes cook
The middle counterpart of Examples 77 and 78 is shown in Example 79. (79) Les pommes de terre cuisent The potatoes are cooking
Thus, in addition to the type of opposition exemplified in Section 3.2, such as la branche s'est cassee versus Pierre a casse la branche, the type of opposition displayed in Examples 80 and 81 will also be analysed as an ergative pair. middle: non-reflexive (80) [Med] Les pommes de terre [Proc] cuisent The potatoes are cooking effective: analytic (81) [Ag] Pierre [Proc: a causative A p material] fait cuire [Med] les pommes de terre Pierre makes the potatoes cook
The AGENCY system shown in Figure 3.2 can thus be extended as shown in Figure 3.8. If we now turn to the ranged middle example derived from Guillaume's (1945) non-ranged middle clause, Ces choses se disent, and re-enter the system to select the feature 'analytic', then, as one would predict, we end up with a very different phenomenon, as shown in Examples 82 and 83.
Figure 3.8 Least delicate AGENCY systems
92
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Process: verbal and middle (82) [Med] EUes [Proc] disent [Range] ces choses They say these things Process: verbal: effective: analytic (83) [Ag] Elks [Proc] font dire [Range] ces choses [Med] a ces gens They make these people say these things
With the selection of [analytic], as shown in Example 83, there is the potential to re-enter the system once more, to realize double agency, as in Example 84. Process: verbal and effective: analytic, with double agency (84) [Agentl] Ellefait dire [Range] ces choses [Med] a ces gens [Agent2] parsonfils. She makes her son make these people say these things.
The ergative model is thus potentially realized logically through recursive selections of the effective feature, which allows the construal of chains of causation in which there is more than one Agent. The realization of such causal constructions is analytic rather than synthetic. Matthiessen (1995) illustrates this logical construal of agency by an analytical causative with the following clause: The president Agentl
had Pra
the general Agent2
make -ocp
the squad AgentS
explode -ess y
the bomb Medium
Matthiessen argues that in English the two models of transitivity embodied in the AGENCY and PROCESS TYPE systems reflect two metafunctional perspectives on phenomena, the logical and the experiential, respectively. This ideational diversification of transitivity models is less marked in French where the logical potential seems generally more restricted than in English (see Caffarel 1992 on serial time). The French clause can have two Agents at most, and the secondary Agent is always realized by a prepositional phrase, thus closing down further options, as in Example 85. (85)
k general the general Agentl
a fait exploser had explode Proc
labombe the bomb Med
par les soldats by the soldiers Agent2
The progression in French from middle to analytically causative with two Agents is illustrated in Table 3.10. Looking further at the Agency system, we can see that if we choose [middle], then we have a choice between [reflexive] and [non-reflexive]. As we have seen throughout this chapter, there are some processes in French that can be used in a middle clause with or without the se-clitic, like the verb casser (to break), as seen in Examples 86 and 87.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
93
Table 3.10 A continuum from middle to analytically effective clauses Agency
Example
middle
[Med] les pommes de terre [Proc] cuisent the potatoes are cooking
effective: synthetic
[Ag] son pere [Proc] cuit [Med] les pommes de terre her father is cooking the potatoes
effective: analytic: one Agent
[Ag] son pere [Proc: acausApmat] fait cuire [Med] les pommes de terre *her father makes cook the potatoes
effective: analytic: two Agents
[Agl] elle [Proc] fait cuire [Med] les pommes de terre [Ag2] par son pere she makes her father cook the potatoes
(86) La branche s 'est cassee (87) La branche a casse The branch broke
Zribi-Hertz (1987: 33) refers to Rothemberg's (1974) hypothesis as to the semantic difference that exists between the two middle clause types. In her view, reflexive middle clauses 'indicate the existence of an external cause (non-expressed) in the realization of the process' while nonreflexive middle clauses 'indicate that the process is a natural process which is realized only thanks to qualities which are inherent to the Subject [Medium]'. In other words, reflexive middle clauses imply a cause external to the Process-Medium nucleus, while non-reflexive middle clauses imply a cause internal to the clause nucleus. This is exemplified in Examples 88 and 89. (88) Le vin s'aigrit (89) Le vin aigrit
The wine is turning sour
Following Rothemberg's hypothesis, Example 88 implies that there is some external phenomenon which causes the wine to turn sour, while clause 89 implies that the wine has intrinsic properties causing it to become sour. This section has explored the AGENCY system in some delicacy. The next section turns to the generalizing nature of the ergative model, and in particular the role of Agent, across process types.
94
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3.5 The ergative model generalizing across process types
The advantage of an ergative analysis is that it provides a general set of functions that theoretically can be extended to all process types. Functions such as Actor and Goal are clearly inadequate to describe participants involved in mental, verbal or relational processes. The Medium function can, on the other hand, be generalized to all process types: 'the Medium is the nodal participant throughout: not the doer, or the causer, but the one that is critically involved, according to the particular nature of the process' (Halliday 1985a: 147). Thus, as we have seen, it is the Actor of a middle and material clause and the Goal in an effective and material clause. The second participant of an effective material clause is referred to as the Agent in accordance with Halliday's ergative functions. In this section we will focus on the Agent function across process types, beginning with doing processes, then moving to projecting, and finishing with being processes. 3.5.1 Doing clauses Among doing clauses, material clauses can be either middle or effective. In contrast, behavioural and meteorological clauses are restricted in agency and are typically middle. 3.5.1.1 Material clauses The interaction between the material system and the agency system is illustrated in Table 3.11. The table shows that middle material clauses may be reflexive or non-reflexive, and that effective material clauses may be synthetic or analytic. The paradigm of clauses outlined in Table 3.11 defines the following grammatical proportionalities: Laporte s'ouvreis related to Pierre ouvre laporte in the same way as Lapierre tombeto Claire fait tomber la pierre. (See Matthiessen 1995 on paradigms and proportionalities.) It could be argued that the
Table 3.11 Instantiations of the AGENCY system in the context of material clauses reflexive
non-reflexive
middle
La porte s 'ouvre The door opens
La pierre tombe The stone is falling
effective
Pierre ouvre la porte Pierre opens the door
Claire fait tomber la pierre Claire makes the stone fall
synthetic
analytic
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
95
proportionalities defined by the above paradigm reflect the ergative and transitive models respectively. Thus, it would seem that the grammar itself may foreground one or the other model. We could say that the opposition between middle reflexive (true middle) and effective synthetic (true effective) forms a true ergative pair, while the opposition between middle non-reflexive (intransitive) and effective analytic forms a surrogate ergative pair which evolved from the transitive model. 3.5.1.2 Behavioural processes and meteorological processes
Behaviours and the weather are construed by the material grammar as a nuclear Process + a Medium (Example 90) or as Subject z7plus a conflation of Process and Medium (Example 91). That is, the kind of activity they represent is generally brought about by the Medium only, and not by an external cause. (90) /' I am sneezing Med (91) II It's raining
eternue Proc: beh pkut Med/Proc: meteorological
But both clauses could be made effective with the use of faire, as in Examples 92 and 93. (92)
(93)
// He made me sneeze Ag // He made rain Ag
m'
a fait eternuer
Med
Proc: beh
a fait pleuvoir Proc: causative A Proc/Med: meteorological
3.5.2 Projecting clauses
Among projecting clauses, it is the verbal type which tends to be restricted in agency, while the mental type, because of its bidirectional characteristic (see 3.5.2.1), may be either middle or effective. 3.5.2.1 Men tal processes
One characteristic of mental processes of reaction is their bidirectionality. The process may be coded from consciousness to phenomenon [middle] as in Example 94, or from phenomenon to consciousness [effective], in which case the Phenomenon functions as Agent rather than as Range, as in Examples 95 and 96.
96
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(94) ]' I Med/Se
aime like Proc: ment
lamusique music Range/Then
(95)
Cette musique This music Ag/Phen
me me Se/Med
plait pleases Proc: ment
(96)
La musique le passionne Music fascinates him (he has a passion for music)
What is interesting in French is that the 'please' type mental clause can also be middle, in which case the Phenomenon functions as Range and late News. In such a mental-type clause the Range element takes a preposition, as in Example 97, which is the middle version of Example 96. Halliday (1985a: 149) points out: . . . the choice of 'plus or minus a preposition' with Agent, Beneficiary and Range is not just random variation; it serves a textual function. . .. The principle [in English] is as follows. If a participant other than the Medium is in a place of prominence [marked Theme or late News] in the message, it tends to take a preposition; otherwise it does not.
In French, it seems that the same principle applies also to the Medium when the former functions as late News in a clause with three participants (see Example 84 above). (97)
// He Med/Se
se passionne passions/has a passion Proc: ment
de musique for music Range/Phen
The Phenomenon/Range may also be absent, as in Example 98. (98)
II He Med/Se
s'ennuie is bored Proc: ment
Examples 99 to 102 are further examples of agnation between effective and middle 'please-type' mental clauses, displaying the different types of preposition groups that may introduce the Phenomenon in the middle versions (the prepositional phrases are highlighted in bold). (99)
[Ag/Phen] La situation [Proc] alarme [Med/Se] les voisins The situation alarms the neighbours (100) [Med/Se] Ils [Proc] s'alarment [Range/Phen] de la situation They are alarmed because of the situation (101) [Ag/Phen] Cette reponse [Proc] etonne [Med/Se] Pierre This answer surprises Pierre (102) [Med/Se] Pierre [Proc] s 'etonne [Range/Phen] de cette reponse Pierre is surprised by this answer
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
97
These examples also show that mental middle clauses may have a reflexive verbal group, just as material clauses may. Thus, the AGENCY system as presented in Figure 3.8 can be generalized to mental processes. Examples of mental clauses displaying the features specified in the AGENCY network in Figure 3.8 are shown in Table 3.12. 3.5.2.2 Verbal processes
In contrast to mental clauses, verbal clauses tend to always be middle. However, in rare contexts, the realization of effective verbal clauses is made possible by means of the causative process faire. (103) [Ag] //[Med] m'[Proc] afaitparler [Matter] demoi He made me speak of myself (104) [Ag] //[Med] m'a fait dire [Verb/Range] des choses [Iquejene voulais pas dire]] He made me say things that I did not want to say
Thus, effective verbal clauses can only be analytic, which indicates that verbal processes themselves are inherently middle, with the exception of processes of verbal impact which are inherently effective, i.e. those that are oriented towards a Target, as in Example 105. (105) On They Ag/Sayer
I' him Med/Target
accuse accuse Proc: verbal impact
d'un crime of a crime Range
3.5.3 Being clauses Among being clauses, relational clauses have the potential to be middle or effective while existential clauses are restricted in agency. Examples of attributive and identifying clauses combined with different selections in the system of AGENCY are discussed in the next two subsections; since existential clauses can only be middle, and examples of these were already given above in Section 3.3.3.1, they will not be discussed further in this section.
Table 3.12 Agency and mental processes reflexive
non-reflexive
middle
// s 'ennuie He is bored
// aime la musique He likes music
effective
La musique lepassionne Music impassions him
98
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3.5.3.1 Relational: attributive clauses
In attributive relational clauses the Agent corresponds to the transitive function of Attributor and its function is therefore to attribute a quality to something as in Examples 106 and 107. Process: relational: attributive and effective (106) [Ag/Attrib] Les talibans [Proc] ont qualifie [Med/Ca] les bombardements [Range/Att] 'd'acte terroriste' The talibans qualified the bombardments as a terrorist act (107) [Ag/Att] Lesfranfais ontelu [Med/Ca] Mitterand [Range/Att] president The French have elected Mitterand president
The middle counterparts of these clauses are shown in Examples 108 and 109. Process: relational: attributive and middle (108) [Med/Ca] Les bombardements [Proc] sont [Range/Att] un acte terroriste The bombardments are a terrorist act (109) [Med/Ca] Mitterand [Proc] est [Range/Att] president Mitterand is president
Relational processes involving a change such as changer and transformer are often used with an Agent, as in Examples 110 and 111. (110) [Ag/Att] La fee [Proc] a change [Med/Ca] le carrosse [Range/Att] en citrouille The fairy changed the coach into a pumpkin (111) [Ag/Att] Lefeu [Proc] a transforme [Med/Ca] la glace [Range/Att] en eau The fire turned the ice into water
Both the above clauses can be made middle with a reflexive verbal group, as shown in Examples 112 and 113. (112) [Med/Ca] Le carrosse [Proc] s'est change [Range/Att] en citrouille The coach changed into a pumpkin (113) [Med/Ca] La glace [Proc] s 'est transformee [Range/Att] en eau The ice turned to water
Table 3.13 Agency and relational attributive processes reflexive
non-reflexive
middle
Le carrosse s 'est change en citrouille The coach changed into a pumpkin
Le carrosse est rond The coach is round
effective
La fee a change le carrosse en citrouille The fairy changed the coach into a pumpkin
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
99
Examples of effective attributive clauses came up in Section 3.3 in Sartre's (1939) short story 'Le mur', and are repeated here as Examples 114 and 115. (114) Lalumiere The light Ag/Att (115) Je I Ag/Att
leur them Med/Ca trouvais found Proc
faisait made Proc (a this Med/Ca
mal hurt Range/Att agreable nice Range/Att
3.5.3.2 Relational: identifying clauses
Identifying clauses and in particular naming clauses may also be middle or effective. Middle naming clauses have a reflexive verbal group and effective naming clauses have an additional participant, the Agent, which is called the Assigner in transitive terms. Examples 116 to 119 show the alternation between middle and effective versions of the same clauses. Process: relational: identifying and middle (116) [Med/Tok] Elk [Proc] s'appelle [Range/Val] Marie She is called Marie (117) [Med/Tok] Ce livre [Proc] s'intitule [Range/Val] L'homme deparoles This book is entitled L'homme de paroles Process: relational: identifying and effective (118) [Ag/Assigner] Sa mere [Med/Tok] /'a [Proc] appelee [Range/Val] Marie Her mother called her Marie (119) [Ag/Assigner] Claude Hagege [Proc] a intitule [Med/Tok] son livre [Range/Val] 'L 'homme de paroles' Claude Hagege entitled his book L'homme de paroles 3.5.4 Summary of the combination of AGENCY and PROCESS TYPE
From the above discussion, it can be seen that the ergative model (the system of AGENCY) generalizes across the PROCESS TYPE system in French to a large extent. The major subtype of each of the main process types (doing, projecting and being) - i.e. material, mental and relational processes respectively - interacts relatively freely with features in the system AGENCY. The minor subtypes, on the other hand, are either typically or obligatorily middle: behavioural, meteorological and verbal processes can be effective but only through the formation of analytic causatives, while existential clauses are only ever middle - they simply consist of the nucleus Medium + Process with the Medium coming last as it is simultaneously introduced as a New participant. To finish this section, Table 3.14 illustrates the complementarity of the two systems PROCESS TYPE and AGENCY. Note again the use of the s^clitic across clause types in middle clauses, which has been discussed in relation to each of the process types in the subsections of this section.
100
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 3.14 Agency across process types DOING
PROJECTING
BEING
middle
La porte s'ouvre The door opens
// s 'ennuie He is bored
Elle s 'appelle Marie She is called Marie
effective
// ouvre la porte La musique I' ennuie He opens the door Music bores him
Sa mere I 'appelle Marie Her mother calls her Marie
Thus far, this section has considered the interaction of the AGENCY and the PROCESS TYPE systems, and in particular the way that the features [middle] and [effective] combine with the different major and minor subtypes of the three main process types, doing, projecting and being. The next subsection will finish this section on AGENCY with a brief discussion of the system of VOICE, which although a textual system, varies depending on whether a clause is middle or effective. 3.5.5
Voice in middle and effective clauses
In middle clause types, the VOICE CULMINATION system offers the choice between introducing either the Process or the Medium as late News; this distinction is shown in Examples 120 and 121. Process culmination (120) Desgens
arrivent
People are arriving
Medium
Process
Theme
Rheme
Given
New
Medium culmination (121) // arrive
des gens
People are arriving Process Theme
Medium
Rheme Given <
> New
We interpret Given as beginning from the Process, as the pronominal il is non-identifiable and has no participant function. Further examples of the use of Medium culmination in both non-ranged (e.g. in Example 123) and ranged middle clauses (e.g. in Example 125) follow. (Examples 125, 127 and 129 are taken from Grimshaw 1982: 127-8.)
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
101
(122) Quelque chose sepreparait (Process culmination) (123) // sepreparait quelque chose (Medium culmination) Something was going to happen (124) [Med] Une idee fantastique [Beneficiary] lui [Proc] a traverse [Range] ['esprit (125) // lui a traverse I'esprit une idee fantastique A fantastic idea crossed his mind (126) [Med] Une grande esperance [Proc] prend [Range] corps (127) II prend corps une grande esperance A great hope is taking shape (128) [Med] Trois mille hommes [Proc] se sont denonces [Time] cemois-ci (129) Us'est denonce trois mille hommes ce mois-ci Three thousand men denounced themselves this month
Medium culmination in Examples 125 and 127, which are examples of ranged middle clauses, confirm that the middle culmination system does not depend on whether the clause is transitive or intransitive but rather on whether it is middle or effective. Ranged middle clauses with process culmination can have a further choice between having the Medium explicit (medio-active) or implicit (mediopassive), in which case the Range is conflated with the Subject and the Medium omitted as in Examples 130 and 131. (130) L 'action The action Range
a ete revendiquee was claimed Proc: verb and medio-passive
(131) Les automations The authorization Range
leur to them Beneficiary
avaient ete donnees had been given Proc: mat and medio-passive
One should note that there is a difference between medio-passive clauses and passive clauses, as the former do not have and cannot have an Agent. The option 'passive' is a feature of effective clauses. In effective clauses, the VOICE system distinguishes between an active verbal group or a passive verbal group as in Examples 132 and 133. (132)
Yannick Yannick Ag
(133)
Remi Remi Med
a souleve lifted Proc: mat and active a ete souleve was lifted Proc: mat and passive
Remi Remi Med par Yannick by Yannick Ag
Due in part to the versatility of middle clauses in French, passive clauses in French are not used as often as in English. Furthermore, the conflation of the Beneficiary and Subject in a passive clause is not a possible option in French as it is in English. However, there is another way of conflating
102 A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Beneficiary and Subject in French which involves the use of a reflexive causative process, as pointed out by Judge and Healey (1985: 205-06): an English passive sentence such as her father was offered a book cannot be passivized in French in normal ways; but it can be translated into French by a passive substitute such as sonperes'est vu offrir un livre. . .
Judge and Healey (1985: 206) go on to cite Price in relation to this: Though this construction is ignored by most grammars, it is first recorded in the seventeenth century and it is now in common use, particularly but not exclusively in journalistic style . . . It is likely that the semantic value of 'voir' is no longer felt in this construction and that the verb has become fully grammaticalized as an auxiliary of the passive . . .
The meaning of s 'est vu is in fact analogous to that of s 'estfait in Example 134. (134) // He Med
s'estfait taper was hit Proc: analytic and passive
parsonfrere by his brother Ag
We can thus distinguish two kinds of passives in effective clauses, those constructed with etre and those constructed with sefaire or se voir. In a sense the synthetic—analytic opposition discussed in relation to active effective clauses can also be generalized to passive effective clauses, where the etre-type can be seen as synthetic whilst thefaire- and voir-types can be seen as analytic. Taking into account the VOICE systems discussed above, a more delicate systemic representation of the AGENCY potential is presented in Figure 3.9. In the last two sections, we have looked at process types, the participants associated with each process type, and the way they interact with the two transitivity systems, PROCESS TYPE and AGENCY. This is referred to by Matthiessen (1995: 206) as nuclear transitivity, as opposed to circumstantial transitivity: Nuclear TRANSITIVITY is concerned with the construal of the Process and the participants involved in it and AGENCY and PROCESS TYPE determine the nature of the process, the participants and the way they interact through the Process in the first instance. Circumstances are less directly affected by choices within these two systems and their more delicate elaborations since circumstances are less directly involved in the Process. However, there are certainly circumstances that are implicated by different types of process; for instance, material processes of motion may have an inherent circumstance of directed Location (as in they departed from Victoria Station; she sent the books to Naples) and verbal processes are likely to have a circumstance of Matter (as in she told her aunt about the horses).
In the next section we will see that circumstances that are inherent in the process have the potential to be cliticized to the Process in French just as participants do; such circumstances are interpreted as nuclear circumstances
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
103
Figure 3.9 Systemic representation of the AGENCY potential as opposed to the more peripheral circumstances which cannot be cliticized. The next section will also discuss circumstance types, and a system network of these will be provided. 3.6 Nuclearity and peripherality
Section 3.1 mentioned the fact that although French transitivity systems are clearly clausal systems, they may be realized at clause or at group rank. The realization at group rank involves cliticization or agglutination of participants and some circumstances to the verbal group, as inje le lui ai donne (I gave it to him). This process of agglutination is a characteristic of polysynthetic languages, with which French shares some features. The choice between cliticizing and not cliticizing participants and circumstances is a textual one: clitics tend to be Given while non-clitics tend to be New. However, the potential of a participant or a circumstance to be cliticized also indicates its nuclearity or peripherality in the transitivity of the clause. Thus, an exploration of transitivity at group rank allows us to determine the different degrees of nuclearity of participants and circumstances in French. The model of nuclearity presented here follows work by tagmemicists (see, for example, Pike and Pike 1982) and by Tesniere (1959). In French, the clause nucleus consists of the obligatory Process + Medium,
104
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
plus other potential participants and some circumstances which may enter into the clause nucleus when they are cliticized to the Process. Transitivity functions that have the potential to be realized at group rank are nuclear in the sense that they are inherent in the nuclear process. The clause nucleus is, typically, textually unmarked and Given, as shown in Examples 135 and 136.
(135) Pierre Pierre
park speaks
a Franfoise to Frangoise
de sa sceur about his sister
Beh
Proc
Beneficiary
Circ: matter
Theme: unmarked
Rheme
Given
New
lui to her
en about her
parle speaks
Beh
Beneficiary
Circ: matter
Proc
Theme: unmarked
Rheme
(136) II He
Given
New
A number of criteria are used to determine the degree of nuclearity of participants and circumstances: (i)
their likelihood of occurrence as clitics;
(ii) their position as clitics vis a vis other clitics; (iii) their realization: nominal group or prepositional phrase; and (iv) their 'staying power' in reference chains. What we find is that participants are more nuclear than circumstances. Circumstances of location and matter are more nuclear than the other circumstances since these can never be cliticized. Within the clause nucleus, circumstance clitics always occur last, as in Examples 137 and 138.
(137) II He (138) Je I
en
lui to him /' him
of this
3there
parle speaks
ai conduit drove
The clitics, en and y, which stand for the two circumstance types that may be cliticized, can also function as participants, as in Examples 139 and 140. (139) J'ai laisse mes papiers sur la table; n y touchez pas I left my papers on the table; don't to it touch (140) Je veux du vin; j'en veux I want some wine; I it want
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
105
The use of y and en as both participant and circumstance can be explained by the grammatical and semantic similarity of their respective referents. Grammatically, the referents are in both cases realized typically by prepositional phrases with the prepositions a and de respectively. Semantically, it can be argued that both dative (giving to) and locative (going to) are similar in that they imply a motion or transfer, either to someone or some place, and that being a part of some thing implies a meaning of belonging, which is closely related to the meaning of provenance expressed in coining from some place. In traditional terms, we could generalize that y marks the dative case, while en marks the genitive. Among the participants, the Medium is the most nuclear one. It is the participant without which the process could not take place. It is the participant which is least likely to be realized as a prepositional phrase and which has the longest 'staying power' in discourse - that is, entities that are construed as Medium tend to be central entities in the text as a whole, not just in one or two clauses. We saw in Section 3.4 that there is only one clausal context where the Medium is realized by a prepositional phrase and this is in an analytically effective ranged clause, such as Example 141. (141) Elles They Ag
font chanter make sing Proc
ces chansons these songs Range
a ces gens to these people Med
In the realization of the transitivity functions of Example 141 at verbal group rank, the Range also precedes the Medium, as shown in Example 142. (142) Elles They Ag
les them Range
leur to them Med
font chanter make sing Proc
The Range is interpreted here as the second most nuclear participant. The Range is an elaboration or an extension of the Process. The Range specifies the Process further, but it is not essential to the realization of the Process, as the Medium is. Consider Examples 143 and 144. (143) Elle She Med
chante sings Proc
(144) Elle She Med
chante sings Proc
une chanson a song Range
Furthermore, the Range can be realized by a prepositional phrase as in Example 145. (145) II joue au football; ilyjoue He plays football; he it plays
106
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
The Range has little staying power in discourse. Its status as process elaboration means that its status as participant is weak. This view thus contrasts with that of Martin (1992) for whom the elaborative nature of the Range suggests that it is more nuclear than the Medium. Martin (1992: 318) argues that the Range: process is more nuclear than the Medium on the basis that the combination Process + Range involves just one meaning, while the combination Medium + Process involves two meanings, an action and the participant involved in the action. We have seen that two circumstance types, Location (i.e. spatial: place) and Matter, can be inherent in certain process types, and therefore cliticized. On this basis, they are interpreted as nuclear circumstances. Most circumstances, however, can never enter into the clause nucleus. Such circumstances, like those of Manner and Cause, are peripheral. A network of circumstance types is presented in Figure 3.10. Table 3.15 illustrates each circumstance type shown in the network in Figure 3.10 with an example, and a probe question which makes explicit the meaning of each circumstance type. The location of participants and circumstances in French on the cline of nuclearity-peripherality is schematized in Figure 3.11. Figure 3.11 shows that in French the Process + Medium combination forms the centre of the clause nucleus. The Range comes next. Then follows the Beneficiary and Agent, which can both be realized by a prepositional phrase, as in Examples 146 and 147 on p. 108.
Figure 3.10 A systemic representation of Circumstance types
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
107
Table 3.15 Examples and probe questions for Circumstance types Circumstance type
Probe question
Example
spatial: time spatial: place cause: reason cause: behalf
when? where? why? (for what reason) why? (on whose behalf)
cause: purpose concession condition
why? (for what purpose) despite what? in which conditions?
accompaniment matter viewpoint
with whom? about what? according to whom?
manner: means manner: quality manner: comparison
by what means? how? like what?
Elle vient en avril Elle part en Espagne Elle part a cause de la pluie Elle s 'est battue au nom de tous les siens Elle ecrit ce livre dans un but utilitaire II est heureux malgre lui Je ne pourrais pas lefaire sans son aide Je pars avec mes amies II m 'a ecrit au sujet de safille D'apres cejournaliste, kfilm de Rohmer est nul II a creuse ce trou avec une pioche II marche lentement II marche comme une taupe
Figure 3.11 Nuclear and peripheral elements of the clause as representation
108
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(146) II He Med
aeteaverti was informed Proc
(147) Elle She Med
a telephone rang Proc
parsanfrere by his brother Ag asonjrere her brother Beneficiary
Circumstances of Location and Matter are at the edge of the clause nucleus, while the other circumstances are outside the clause nucleus. Following this account of clause structure in terms of nuclearity, we will now return to a discourse-based approach to transitivity and explore the use of process types and agency across different text types. 3.7 Transitivity patterns in texts This section looks at patterns of instantiation of the general transitivity potential in different text types in order to explore how the different models of transitivity discussed above are deployed and how transitivity resources are used to construe higher order meanings. Again, it should be noted that one model of participation does not exclude the other, and to underline this, all texts will be analysed simultaneously both in terms of the generalized ergative functions and the process-specific transitive functions. Analysing the functions specific to the Process may appear redundant in the sense that by analysing the process type we implicitly indicate the particularities of the participants. However, it is useful in showing different mappings of ergative and transitive functions depending on whether the clause is middle or effective. Whatever the objectives of the experiential analysis, it is always important to simultaneously identify the process types (the kind of experiential domains that are represented) and whether these processes are self-engendered or not. The first text is taken from an article in Geo (October 1984) entitled 'La Planete des hommes brule-t-elle?' (Is man's planet burning?). Text 3.3 From Geo, October 1984 1 [Med/Ac] Le Galungung, volcan indonesien, [Process: material] s'est reveille. The Galungung, an Indonesian volcano, has woken. 2 [Location: time] En pleine journee, [Med/Ac] le del [Proc: mat] s'obscurcit. In the middle of the day, the sky darkens. 3a [Ag/Ac] Les particules de roches solides projetees dans I'airjusqu 'a quarante kilometres d'altitude [Proc: mat] cachent [Med/Goal] la vue du soldi [Location: time] pendant plusieurs jours, The particles of solid rock projected into the air to a height of forty kilometres hide the sun from view during several days, 3b avant de [Proc: mat] se deposer [Manner: quality] en une couche blanchdtre, semblable a de la neige. before settling in a whitish layer, similar to snow.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION 4a
4b
4c 4d 5a
5b
109
Cependant, [Med/Se] nous [Proc: ment] savons, [Cause: reason] apartirdes recherches effectuees pendant Veruption du Mont Saint-Helens (Etats-Unis) en 1980, However, we know, from research carried out during the eruption of Mount Saint-Helens (United States) in 1980, que [Med/Ca] des quantites importantes depoussieres [Process: attributive] restent [Attribute: circumstantial] en altitude, that important quantities of dust stay at altitude, [Proc: mat] absorbant [Med/Goal] unepartie de I'energie solaire absorbing part of the solar energy et [Proc: mat] reduisant [Med/Goal] la transparence de I'atmosphere. and reducing the transparency of the atmosphere. Ren [Proc: ex] resulterait, cettefois, [Med/Ex] non un rechauffement, maisun refroidissement: The result would be, this time, not in an increase but a drop in the temperature: ainsi, [Ag/Ac] I'activite volcanique, mais aussi les poussieres industrielles, [Proc: mat] pondereraient [Manner: means] par des variations climatiques contraires [Med/Goal] les effets de rechauffement [Cause: reason] dus a I 'augmentation des rejets de gaz carbonique. thus, not only the volcanic activity, but also industrial dusts would counterbalance by contrary climatic variations the warming effect due to the increase of carbon dioxide discharge.
The above text is typical of its genre in terms of its transitivity selections, with a predominance of material processes representing happenings (rather than doings) involving inanimate actors. Either natural phenomena are presented as causeless happenings (see clauses 1, 2, 3b) or as being brought about by some external cause which is realized either as an inanimate Agent (3b, 5b) and/or a circumstance of cause (5b). Clauses 4c and 4d are nonfinite clauses where the Agent/Subject is retrievable from the preceding clause. The other most common process used in this type of text is of the relational attributive type, which is typically middle (clause 4b). The above text also contains a mental process of cognition (clause 4a) which requires a conscious participant, in this case researchers, and a relational/existential process of result (clause 5a). A summary of transitivity selections in Text 3.3 is given in Table 3.16. Text 3.3 is concerned with geological happenings, realized mostly as Medium + Process. When the processes are goal-transitive, they encode naturally occurring cause-and-effect phenomena rather than controlled actions. The Actors/Agents are not participating in actions, but unintentionally causing them to happen. This type of pattern embodied in the AGENCY system is very common in factual writing of various types. Texts 3.4 and 3.5 are further illustrations of this patterning where the variable is presence or absence of external cause.
110
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 3.16 Summary of the transitivity selections in Text 3.3 Clause number Process type
Medium
Agent
1 2 3a 3b 4a 4b 4c 4d 5a
material & middle material & middle material & effective material & middle mental & middle attributive & middle material & effective material & effective existential & middle
le volcan le ciel la vue
les particules
5b
material & effective
nous poussiere energie solaire transparence . . . non un rechauffement, mais un refroidissement 1'activite volcanique les effets . , .
Text 3.4 From Le Nouveau Guide de France (Michaud and Kimmel, 1990: 332) la [Med/Ac] Le vetement masculin [Time] de nos jours [Proc: mat & mid] tend a se simplifier Menswear nowadays tends to simplify itself Ib et [Cause] sous Vinfluence desjeunes [Proc: mat and mid] a se diversifier: and, under young people's influence, to vary: le [Med/Ex] plus grande fantaisie dans les formes, les couleurs, lafacon d'assembler les vetements. . . (there is) more extravagance of form, colour and of the way the garments are made up . . . 2 [Med/Goal] L'elegance masculine [Proc: material and middle] est«[ ] neanmoins» fixee [Time] chaque saison [Ag/Ac] par les promoteurs de la mode. Masculine elegance is nevertheless determined each season by fashion promoters. Text 3.5 Introduction from Arts de Coree (Rousset 1977: 7) la5 [Range/Att] Peninsulerocheuse [[ [Med] qui [Proc: rel and mid] s'enfonce [Place] dans Vocean]] A rocky peninsula [ [which plunges into the ocean] ] Ib [Med/Ca] la Coree [Proc: rel and mid] s'allonge [Place] surun millier et demi de kilometres Korea extends over a thousand and a half kilometres le mais [Proc: rel and circ] n'en excedepas [Range/Att] cinq cents [Place] dans sa partie la plus large but does not exceed five hundred at its widest part Id et [Range/Att] deux cents [Place] dans saplus etroite. and two hundred at its narrowest part. 2 [Med/Goal] Elk [Proc: mat and eff] est separee [Place] de la Mandchourie et de la Siberie [Ag/Ac] par deuxfleuves, le Yalou [[qui court vers la Chine]] et le Tumen [[qui remonte vers Vest pour sejeter dans la mer dujapon]]. It is separated from Manchuria and Siberia by two rivers, the Yalou which runs towards China and the Tumen which goes up towards the East and flows into the Sea of Japan.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION 111
In narrative texts, on the other hand, where Actors are predominantly animate, it can be predicted that they will be presented as having control over their actions and the power to impact on 'things' and 'others'. This is exemplified by the following extract from Dino Buzzati's (1968) tale La Fameuse Invasion des Ours en Sidle (The Famous Invasion of Sicily by the Bears). Text 3.6 From La Fameuse Invasion des Ours en Sidle (Buzzati 1968) la [Time] Bien des annees auparavant, «alors que [Ag/Ac] leRoi des ours, Leonce, [Proc: mat and eff] etait alle ramasser [Med/Goal] des champignons [Accompaniment] avec sonjeunefils», Many years earlier, while the King of the bears, Leonce, had gone to pick mushrooms with his young son, Ib [Ag/Ac] deux chasseurs [Proc: mat and eff] avaient enleve [Med/Goal] I'enfant. two hunters had abducted the child. 2a [Med/Ac] Lepere [Proc: mat and mid] s'etant eloigne [Time] un instant [Place] le long d 'un a-^pic,
2b
2c
2d
The father having wandered away for an instant along a cliff, [Se/Med] ils [Proc: ment and mid] avaient surpris [Range/Phen] I'ourson seul et sans defense, they had caught the cub alone and defenceless, [Med/Goal] /' [Proc: mat and eff] avaient ligote [Manner] comme un vulgaire paquet had tied him up like an ordinary parcel et [Proc: mat and eff] fait descendre, [Place] le long des precipices, [Place] jusqu'au fin fond de la vallee. and made him go down the cliffs till the very end of the valley.
In Text 3.6, all clauses are material clauses of doing with animate actors, with the exception of clause 2b, which is a mental clause of perception. Except for one instance, where the clause is middle (clause 2a: note the se-clitic), all material clauses have an Actor-Process-Goal structure, where the Actor is doing something to the Goal. This kind of transitive organization is typical of 'traditional' narratives and of children's storytelling, as Text 3.7, taken from Le Pare National des Pyrenees Raconte aux Enfants (Pyrenees National Park Stories for Children) further illustrates. Text 3.7 From Le Pare National des Pyrenees Raconte aux Enfants (Birmant 1991) 1 [Time] Pendant des milliers et des milliers d'annees, [Ca/Med] les Pyrenees [Proc: att and mid] ont ete [Att/Range] des montagnes tranquilles. During thousands and thousands of years, the Pyrenees were tranquil mountains. 2 Nous, les animaux [Ac/Med] nous [Place] y [Proc: mat and mid] vivions [Manner] enpaix. We, the animals, we lived there in peace. 3a Mais [Ac/Med] les hommes [Proc: mat and mid] sont venus [Place] deplus en plus pres, But men came closer and closer,
112
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3b
et [Se/Med] nous [Proc: ment] avons vu [Phen/Range] [[arriver d'autres animaux]] and we saw other animals arrive 3c [Ac/Med] qui [Range] les [Proc: mat and mid] fuyaient which were fleeing them (the men). 4 [Ac/Med] Nous nous [Proc: mat and mid] sommes caches. We hid. 5a [Ac/Ag] Les hommes [Proc: mat and eff] faisaient [Goal/Med] du bruit, The men made noise, 5b [Proc: mat and eff] reveillaient [Goal/Med] nospetits, woke up our little ones, 5c [Goal/Med] nous [Proc: mat and eff] poursuivaient [Accompaniment] avec leur fusils, chased us with their guns, 5d [Proc: matl and eff] troublaient [Goal/Med] I'eau des torrents, made the rivers' water dirty, 5e et [Proc: mat and eff] detruisaient [Goal/Med] nos arbres and destroyed our trees. 6a [Senser/Med] Certains d'entre nous [Proc: ment and mid] n'ontpas supporte [Phen] [[de vivre dans ces conditions]]: Some of us could not bear living under these conditions: 6b [Se/Med] Its [Proc: ment and mid] ontprefere [Phen] [[mourir]]. They preferred to die.
From a transitive perspective, Text 3.7 presents a world where animals are actors in material and middle clauses as well as conscious participants in mental clauses. On the other hand, men are represented as actors in material and effective clauses and as affecting the animals and animal world (represented as Goal). From an ergative perspective, the transitivity patterns foreground the ideology behind this text which portrays men as responsible for the destruction of the Pyrenees. While the model of 'extension', where individuals impact on other individuals, is prominent in such traditional stories, it is less apparent in the 'New Novel', as seen in the abstracts presented in the preceding sections, where causeless happenings are foregrounded. This kind of experiential pattern was also found in Camus' L'Etranger (Caffarel 2004b), where the preferred clause type is the middle clause, representing happenings rather than doings. This pattern is broken during the murder scene in the novel, where the use of effective clauses increases. However, counter to 'normal' expectations, the experiential analysis of that passage foregrounds the murderer as the victim (as Goal), while the murdered Arab is never represented as a Goal (as impacted upon). This contributes to an absurd representation of experience where the main protagonist is detached from reality and has no control over his actions. The analysis of the section where the murder of the Arab takes place is reproduced below as Text 3.8; references to the sun as a source of causation (either as Agent or as a causal circumstance) are highlighted in bold.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION 113 Text 3.8 From L'Etranger (Camus 1942, pp. 93-5) 1 [Ca/Med] Le bruit des vagu.es [Proc: rel: att] etait encore plus [Att/Range] paresseux, plus etale qu 'a midi. The waves sounded even longer and lazier than they had been at midday. 2 [Tok] C" [Proc: rel: identifying] etait [Val] le meme soleil, la meme lumiere sur le meme sable [[qui se prolongeait id]]. It was still the same sun, the same light and the same sand as before. 3 IIy [Proc: ex] avait deja [Ex/Med] deux heures que la journee n'avanc,aitplus, deux heures qu 'elk avait jete I'ancre dans un ocean de metal bouillant. For two hours now the day had stood still, for two hours it had been anchored in an ocean of molten metal. 4a 1 A I'horizon, [Ac/Med] un petit vapeur [Proc: material and middle] estpasse Out on the horizon a tiny steamer went by 4b +2 a et [Se/Med] j'en [Proc: ment] ai devine [Phen/Range] la tache noire au bord de mon regard, and I could just see it as a black speck out of the corner of my eye 4c x(3 parce que [Beh/Med] je [Proc: beh] n'avaispas cesse de regarder [Phen/ Range] I'Arabe. because I hadn't stopped looking at the Arab. 5a a [Se/Med]/'[Proc: ment] aipense I realized 5b ' p i que [Ca] je n' [Proc: rel: att] avais [Att] qu'un demi-tour [afaire] that I only had to turn round 5c +2 et [Ca/Med] ce [Proc: rel: att] serait [Att/Range] fini. and it would all be over. 6 Mais [Ac/Med] toute uneplage vibrante de soleil [Proc: mat and mid] se pressait [Place] derriere moi. But the whole beach was reverberating in the sun and pressing against me from behind. 7 [Ac/Med] /' [Proc: mat and mid: ranged] aifait [Range] quelques pas [Place] vers la source. I took a few steps towards the spring. 8 [Ac/Med] L'Arabe n' [Proc: mat and mid] a pas bouge. The Arab didn't move. 9 Malgre tout, [Ca/Med] il [Proc: rel: att] etait encore [Att/Range] assez loin. Even now he was still some distance away. 10 Peut-etre [Cause] a cause des ombres sur son visage, [Ca/Med] il [Proc: rel: att: poss] avait [Att/Range] I'air [[de rire]]. Perhaps because of the shadows on his face, he seemed to be laughing. 11 [Ac/Med] /' [Proc: mat and mid] ai attendu. I waited. 12a 1 [Ac/Ag] La brulure du soleil [Proc: mat and eff] gagnait [Goal/Med] mes joues The sun was beginning to burn my cheeks 12b +2 et [Se/Med] j' [Proc: ment] ai senti [Phen] des gouttes de sueur [[s'amasser dans mes sourcils]]. and I felt drops of sweat gathering in my eyebrows. 13a 1 [Tok] C'[Proc: rel: id] etait [Val] le meme soleil que le jour [[ouj'avals enterre maman]] It was the same sun as on the day of my mother's funeral
114
13b +2
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
et, comme alors, [Att/Ag] le front surtout [Ca/Med] me [Proc: att and eff] faisait [Att/Range] ma/ and again it was my forehead that was hurting me most 13c +3 et [Ac/Med] toutes ses veines [Proc: mat and mid] battaient ensemble [Place] sous la peau. and all the veins were throbbing at once beneath the skin. 14 [Cause] A cause de cette brulure [[quejepouvaisplus supporter]], [Ac/Med] j' [Proc: mat and mid: ranged] at fait [Range] un mouvement [Place] en avant. and because I couldn't stand this burning feeling any longer, I moved forward. 15a a [Se/Med] Je [Proc: ment] savais I knew 15b 'P 1 que [Ca/Med] c' [Proc: rel: att] etait [Att/Range] stupide, that it was stupid 15c +2 a que [Ac/Ag] je [Proc: mat and eff] ne me debarrasserais pas du [Goal/Med] soleil and I wouldn't get out of the sun 15d xp en [Ac/Med] me [Proc: mat and mid] depla^ant d'unpas. with one step. 16 Mais [Ac/Med] j' [Proc: mat and mid] aifait [Range] unpas, un seulpas [Place] en avant. But I took a step, just one step forward. I7a xp Et [Time] cette fois, sans [Proc: mat and mid] se soulever, and this time, without sitting up, I7b a a [Ac/Ag] VArabe [Proc: mat and eff] a tire [Goal/Med] son couteau the Arab drew his knife 17c =P [Goal/Med] qu' [Ac/Ag] il [Benefj m' [Proc: mat and eff] apresente [Place] dans le soleil. and held it out towards me in the sun. 18a 1 [Ac/Med] La lumiere [Proc: mat and mid] agic/e [Place] sur Voder The light leapt up off the steel 18b +2 a et [Ca/Med] c' [Proc: rel: att] etait [Att: circ] comme une longue lame etincelante and it was like a long, flashing sword 18c =P [Ac/Ag] qui [Goal/Med] TO'[Proc: mat and eff] atteignait [Place] au front. lunging at my forehead. 19a 1 [Time] Au meme instant, [Actor/Med] la sueur [[amassee dans mes sourdls]] [Proc: mat and mid] a coule [Manner] d'un coup [Place] sur lespaupieres At the same time all the sweat that had gathered in my eyebrows suddenly ran down over my eyelids, 19b +2 et [Goal/Med] les [Proc: mat and eff] a recouvertes [Accompaniment] d'un voile tiede et epais. covering them with a dense layer of warm moisture. 20 [Ca/Med] Mesyeux [Proc: rel and att] etaient [Att/Range] aveugUs [Place] derriere ce rideau de larmes et de sel. My eyes were blinded by this veil of salty tears. 21a 1 [Se/Med] Je [Proc: ment] ne sentais plus que [Phen/Range] les cymbales du soleil sur mon front All I could feel were the cymbals the sun was clashing against my forehead
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION 21b +2
115
et, [Manner] indistinctement, [Ac/Med] le glaive edatant [Proc: mat and mid] jailli [Place] du couteau toujours en face de moi. And, indistinctly, the dazzling spear still leaping up off the knife in front of me. 22a 1 [Ac/Ag] Cette epee brulante [Proc: mat and eff] rongeait [Goal/Med] mes cils It was like a red-hot blade gnawing at my eyelashes 22b +2 et [Proc: mat and eff] fouillait [Goal/Med] mesyeux douloureux. and gouging out my stinging eyes. 23 C'est [Time] alors que [Ac/Med] tout [Proc: mat and mid] a vacille. That was when everything shook. 24 [Ac/Ag] La mer [Proc: mat and eff] a charrie [Goal/Med] un soufle epais et ardent. The sea swept ashore a great breath of fire. 25a a //[Se/Med] m' [Proc: ment] a semble [Phen/Range] [[queleciels'ouvrait sur toute son etendue]] The sky seemed to be splitting from end to end 25b xp pour [Proc: mat and eff: analytic] laisserpkuvoir [Goal/Med] dufeu. And raining down sheets of flame. 26a 1 [Ac/Med] Toutmon etre [Proc: mat and mid] s'est tendu My whole being went tense 26b +2 et [Ac/Ag] _;' [Proc: mat and eff] ai crispe [Goal/Med] ma main [Place] sur le revolver. And I tightened my grip on the gun. 27a 1 [Ac/Med] La gdchette [Proc: mat and mid] a cede, The trigger gave 27b +2 [Ac/Ag] j' [Proc: mat and eff] ai touche [Goal/Med] le ventrepoli de la crosse I felt the underside of the polished butt 27c +3 et c'est [Place] la, dans le bruit a lafois sec et assourdissant que [Ac/Med] tout [Proc: mat and mid] a commence. and it was there, in that sharp but deafening noise, that it all started. 28 [Ac/Ag] /' [Proc: mat and eff] ai secoue [Goal/Med] la sueur et le soleil. I shook off the sweat and the sun. 29a a [Se/Med] /' [Proc: ment] ai compris I realized 29b 'P que [Ac/Ag] j' [Proc: mat and mid] avais detruit [Goal/Med] I'equilibre du jour, le silence exceptionnel d'uneplage [[ouj'avais ete heureux]]. That I'd destroyed the balance of the day and the perfect silence of this beach where I'd been happy. 30a a Alors, [Ac/Med] j' [Proc: mat and mid] ai tire encore [Time] quatrefois [Place] sur un corps inerte And I fired four more times at a lifeless body 30b x(ia oil [Ac/Med] les balks [Proc: mat and mid] s'enfon^aient and the bullets sank in 30c xp sans qu'ily [Proc: ment] parut. without leaving a mark. 31 Et [Ca/Med] c' [Proc: rel: att] est [Alt: circ] comme quatre coups brefs [[que jefrappais sur la porte du malheur]]. And it was like giving four sharp knocks at the door of unhappiness.
116
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Caffarel (2004b: 564) points out that: Although this passage comprises 15 effective clauses, the cause that leads to the killing of the Arab is detached from the murder itself. The victim is never represented as affected or impacted upon by the murderer or any other participant (it never has the function of Goal as shown in the table below) and the murderer is never represented as having any impact on the victim. However, the murderer is represented as Goal seven times, foregrounding him as a victim. A murder is committed without any connection (physical or other) between murderer and victim. This is the world of the absurd.
The table referred to in the above quotation is reproduced here as Table 3.17. Thus experiential grammar, like logical grammar, can be used to construe second order meanings: philosophical ones, as in the text above, or ideological ones, as in the text on the destruction of the Pyrenees. The analysis of ergative patterns of transitivity in news stories serves to reveal the ideological stance of the journalist or a particular newspaper vis a vis a particular event: for example, is a particular participant constructed as the Agent, i.e. assigned responsibility for the events related, or is the responsible participant purposely omitted (see Fowler et al. 1979; Lukin 2005)? Consider the following text taken from Le Monde newspaper (Text 3.9). This text is about a natural disaster, a storm, that killed and injured several people in Strasbourg during an outdoor performance of Yiddish singers. The overall meaning of the text is causality, foregrounding that the disaster was caused by a violent storm that caused a tree to fall on the spectators. Resources for expressing causality in the text are multiple (in bold in the text): we have five effective clauses, two logical metaphors with causal relations construed as processes, one circumstance of cause and one causal conjunction. The proliferation of causal resources in this text serves both to point to the storm as Agent and absolve the authorities from blame and negligence. At the exception of firefighters who helped find the dead and save the wounded, there are no human Agents. The government and organizers are represented as Sayer/Medium, addressing condolences and stressing that they followed all recommendations.
Table 3.17 Patterns of Agency in Text 3.8 Ergative/Transitive roles
Meursault
The Arab
Medium/Actor Agent/Actor Medium/Senser Medium/Goal
8 5
2 2 0 0
7 7
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
117
Text 3.9 From 'Orage meurtrier a Strasbourg' (Le Monde, 9 July 2001) la
Ib
2a
2b 3 4
5a 5b 6 7a 7b 7c
[Ag/Ac] Les violents orages [[qui s 'abattent sur la France depuis plusieurs jours]] [Proc: mat and eff] ontprovoque [Med/Goal] un drame [Place] pres de Strasbourg: [Med/Goal] lOpersonnes [Proc: mat and eff and passive] ont ete tuees et 85 blessees dont 18 grievement, [Time] vendredi soir 6 juillet, [Ag/Ac] par la chute d'un arbre. [Med/Ac] Unplatane, «. . ..», [Proc: mat and mid] s'est abattu vers 22 heures [Place] sur lepublic d'un spectacle de chant Yiddish, organise au pare du chateau de Pourtales, a une dizaine de kilometres du centre de Strasbourg. «[Proc: mat and eff and passive] deracine [Ag/Ac] par un violent coup de vent», [Cause] En raison de lapluie, [Med/Actor] les spectateurs [Proc: mat and mid] s'etaient refugies [Place] sous une tente. [Med/Goal] L'accident, [Viewpoint] selon leprefetdu Bas-Rhin, PhilippeMarland, [Proc: mat and eff] a eteprovoque par [Ag/Ac] une "tornade soudaine et d'une extraordinaire violence". [Ag/Ac] Les pompiers de Strasbourg [Proc: mat and eff] ont du decouper [Med/ Goal] I'immense platane [Manner] a la tronconneuse, pour [Proc: mat and eff] degager [Med/Goal] les marts et les blesses. [Med/Sayer] LionelJospin et Jacques Chirac [Proc: verb and mid] ont adresse [Range] leurs condoleances [Addressee/Beneficiary] aux families des victim [Med/Sayer] Les organisateurs du spectacle [Proc: verb and mid] soulignent que [Range] toutes les autorisations [Addressee/Beneficiary] leur [Proc: verb] avaient ete donnees et que [Range] toutes les consignes de securite [Proc: ment] avaient ete respectees.
It is interesting to note that the spiralling of causal resources stops when the text shifts from the event itself to the government and the organizers. The last four clauses are all middle and are either verbal or mental. The absence of material clauses at the end of the text foregrounds the authorities as not responsible for the crisis, as outside the situation, as not involved in the disaster. In order not to assign blame to anyone, the last two clauses (which are projections of what the organizers said) are medio-passives, thus leaving implicit who gave authorization for the show to go ahead and who was in charge of security. This section has illustrated how different text types instantiate different features from the transitivity potential. Moving in from the system on the cline of instantiation, we can highlight subpotentials of the general transitivity network, each text type instantiating particular choices from the transitivity potential. Moving in from the instance end of the instantiation cline, each text has been selected, not as token in itself but as representative of texts of a similar type, or register. And the hypothesis is that these registers represent different demands on the transitivity system, foregrounding one or the other of its complementary models of how change takes place - the ergative and the transitive.
118
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3.8 Conclusion This chapter has explored the experiential systems of transitivity, their realization in the clause as process configuration and their instantiation in particular registers. From the perspective of grammar, we saw that the transitivity potential consists of two complementary systems, PROCESS TYPE and AGENCY, which are realized simultaneously in the structure of the clause. From the perspective of discourse, these two simultaneous systems were shown to reflect two distinct but complementary models of participation, the transitive and the ergative. The analysis of transitivity patterns in different text types illustrated how different registers foreground different models of transitivity or, more precisely, how different text types instantiate different subpotentials from the overall transitivity potential. The chapter also illustrated how a simultaneous reading of the representation of experience in terms of the ergative and transitive perspectives can highlight two levels of ideational meanings in a text. The discussion of process types stressed the importance of grammatical criteria as a means of establishing grammatical categories. The discussion of agency illustrated how the features of the AGENCY system generalize across the main process types that are material, mental and relational processes. This account of French transitivity has revealed a number of important features regarding the way French speakers construe their experience of the internal and external world. We saw, for example, that processes are very often represented as causeless by means of middle clauses. Most so-called transitive verbs in French can be used in a middle clause by means of the s0-clitic. The productive nature of the se-clitic in French as a marker of middle clauses was said to correlate with a semantic change from transitive to ergative, which is reflected instantially in the difference between narrative and scientific registers. We also saw that French transitivity can be realized at clause or verbal group rank. The analysis of transitivity at verbal group rank, where functions are realized as clitics, served as a criterion for locating participants and circumstances on the cline of nuclearity. The Medium was shown to be the most nuclear participant, as it is necessary to the realization of the Process. The combination of the verbal group + participants and circumstances which have the potential to be cliticized was referred to as the clause nucleus. The process of agglutination involved in the cliticization of participants is a characteristic of polysynthetic languages, with which French shares some features. Following this discussion of the grammatical realization of ideational meanings at clause rank, in the next Chapter we explore the grammar of negotiation by looking at the clause as a move in dialogue. Notes 1 The term 'case marking' is used here as a general term to refer to the various realizational means that are used to indicate grammatical relations, i.e. case marking, word order and prepositions.
THE GRAMMAR OF IDEATION (2): EXPERIENTIAL METAFUNCTION
119
2 Note that not all constituents have a value in the transitivity structure, for example elements such as il (in clauses 3 and 4) which do not have a participant function, conjunctions such as bien que (in clause 5b), and modal Adjuncts such as non plus (in clause 4); this is indicated in the analysis by an empty pair of square brackets [ ]. 3 Relations are thus still processes, unfolding in time. They contrast with participants, which are construed as permanent in time. 4 Themes of this kind, called absolute Themes, which are outside the transitivity structure of the clause (as shown by the analysis here) were encountered in Chapter 3 and are discussed in detail in Chapter 5. 5 Here I have analysed (la) as a verbless dependent clause: (3 (la) a (lb), but it could also be analysed as the Theme of (lb) (see Chapter 5 on absolute Themes).
4 The grammar of negotiation: interpersonal metafunction
The interpersonal metafunction is concerned with interaction between speaker and listener - with the enactment of intersubjectivity. The interpersonal part of the clause grammar gives the speaker the resources for interacting with the listener by establishing and maintaining an ongoing exchange with him/her - by assuming and assigning speech roles such as questioner and answerer - and by giving or requesting attitudes, comments and evaluations. The notion of exchange is the basis for interpreting the interpersonal semantics of speech functions . . . (Matthiessen 1995: 381-2)
4.1 Introduction In the preceding chapter, we approached the clause from the perspective of the experiential metafunction, which provides the resources for construing our experience of the external and internal world. Simultaneously to representing experience, the clause serves to exchange meaning. In this chapter, we explore the clause from the perspective of the interpersonal metafunction, as the realization of an interactive move in speech exchange. Interpersonally, every language constructs dialogue for exchanging meaning, for at the most abstract level the exchange of meaning consists precisely in giving and demanding information (propositions) or goods-andservices (proposals) (Halliday 1984, 1985a). The interaction of the primary speech roles - giving and demanding - with the commodities to be exchanged - namely, information or goods-and-services - makes up the four primary speech functions of statement, question, command and offer. Because the primary speech functions that are essential to the process of dialogue are potentially general to all languages, we can assume that the primary MOOD options will be similar across languages (this was borne out by Caffarel et al (2004) and Teruya et al. (2005) in their typological overviews of MOOD across different languages). We found that the grammar of MOOD in French resembles that of English and other languages in terms of its primary MOOD options, differentiating between indicative and imperative as well as between the indicative subtypes, declarative and interrogative.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
121
However, as our description of the interpersonal grammar of French becomes more delicate, the MOOD options in the systems will be found to be more specific to French. Further, the similarity of primary systemic choices does not entail that the structural realization of these primary features will be the same in French as it is in English, which according to systemic functional descriptions displays a Mood-Residue structure (Halliday 1985a/1994). Rather than implying a similarity of structural elements and/or their configuration, what the theory predicts is that the mode of expression for the MOOD choices will be prosodic rather than segmental (which is the mode of expression for TRANSITIVITY choices (Halliday 1979; Matthiessen 1990)). So the interpersonal organization of the French clause presented here begins with phenomena that are assumed to be congruent across languages, such as the primary categories of speech functional semantics, MOOD systems and the mode of expression or structure type, to arrive at a particular description which, in terms of both systemic options and structural realization, will on the whole be specific to the French clause as interaction, respecting the particularities of the language. With this goal in mind, in Section 4.2 we begin by examining how the French clause is organized structurally as an interactive move1 in dialogic texts, how exchanges are initiated and responded to, which part of the clause is typically replayed and how propositions and proposals are realized lexicogrammatically at the primary degree of delicacy. Section 4.3 is concerned with the particularities of the MOOD structure of French. Here, we account for the various structural patterns, in an attempt to characterize the most important structural features of the interpersonal organization of the French clause. This will pave the way for an examination of MOOD and MODALITY systems in Section 4.4, where the syntagmatic phenomena already outlined in Section 4.3 will be examined from a systemic or paradigmatic perspective. Attention will be drawn in Section 4.5 to certain aspects of the textual organization of the French clause in a dialogic context. This will allow us to indicate the typical pattern of conflation for the textual and interpersonal element of the clause. The discussion of the textual structure of the clause as interactive move in Section 4.5 will be the basis for exploring a group rank system, the INTERACTANT system. We will see that in dialogues the realization of options from the interactant system is both determined by their role in the dialogue and their textual statuses. 4.2 The French clause as a move in an exchange Let us begin by asking how statements, questions, offers and commands are initiated and responded to in French dialogic text, since an understanding of this will enable us to determine which part of the clause is crucially involved in the dialogic process and how its interpersonal choices are realized lexicogrammatically in French.
122
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
4.2.1 A preliminary exploration of French exchanges Consider Text 4.1, which is taken from Simone de Beauvoir's Les Bouches Inutiles (1945). Text 4.1 From Les Bouches Inutiles (de Beauvoir 1945) 1
A
2
A
3
B
4
C
Hate-toil Hurry! Les cloches sonnent. The bells are ringing.
Est-ce commence? Has it started? C'est commence?
It's started? 5
D
6
E
7
E
C'est commence. It's started. C'est commence!
It's started! C'est commence!
It's started! 8
Ils sortent en courant. . . They run outside .. .
Key: The capital letters here represent different speakers in de Beauvoir's novel: A = Unefemme a une autre (a woman to another woman); B = Une autre femme (another woman); C = Un vieillard (an old man); D = Un homme (a man); E = Voix (voices); clause 8 is not spoken Semantically, this short exchange may be characterized as a macro-proposal. It starts with a command in clause 1, followed by a justifying statement in clause 2, and ends with a non-verbal response in clause 8 to the initial command issued in clause 1. The resolution of the exchange is interrupted by four 'dynamic moves' (Martin, 1992). Clause 3 demands clarification by means of a question as to why they should 'hurry up'. Clause 4 echoes this preceding question, while clause 5 is a response statement to the clarification question and clauses 6 and 7 are exclamations. Because the text is congruently realized, with selections in MOOD harmonizing with those in SPEECH FUNCTION, we note that the command is realized by an imperative, the justifying statement and response statement by a declarative, the clarification question and echo question by an interrogative and the exclamation by an exclamative. Non-congruent realizations of speech function options (i.e. interpersonal grammatical metaphors) will be discussed in relation to modality in Section 4.4. But how is the negotiation in Text 4.1 carried forward? How are the mood options realized? As pointed out above, the resolution of the exchange is prolonged by four moves (clauses 3 to 6). These moves have an important role in negotiating the exchange. An analysis of their modal structure is presented below, where those four clauses are presented as Examples 1 to 4.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION (1)
EstFinite C" Subject
ce Subject est Finite
commence? Predicator commence"? Predicator
(3)
C' Subject
est Finite
commence. Predicator
(4)
C' Subject
est Finite
commence! Predicator
(2)
123
The modal structure and the speech functions of Examples 1 to 4 show that in this dialogue negotiating the resolution of the exchange involves the replaying of the interpersonal functions of Subject, Finite and Predicator. It is suggested here that these three functions are, as a general rule, crucial both to the negotiation process in French and to the realization of MOOD options. In view of this, we refer to the part of the clause that is composed of these three crucial elements as the Negotiator. The Negotiator is the most salient part of the interpersonal structure of the French clause, thus implying the same status to the three functions that comprise it. Ignoring the difference between Example 1 and 2 in the ordering of Subject and Finite as irrelevant to the present discussion, Example 5 summarizes the structural relation between the Negotiator and the three functions which comprise it in Examples 1 to 4. (5)
C'
est
commence
Subject
Finite
Predicator
Negotiator
The different tones associated with these clauses are indicated by the conventional punctuation marks (see Section 4.3 on the different tones in French). Intonation is the prototypical means for prosodic expression; and it is notable that this may be the only resource used to indicate the systemic MOOD contrasts in French, as is clear from a consideration of Examples 1 to 4. In Example 2 the sole realization of a [yes-no interrogative] is by intonation, while Example 1 shows that in addition to tone, this MOOD option can also be realized by ordering Finite before Subject (F A S). This latter type of realization is called grammatical prosody (Matthiessen 1992: 398). It is prosodic in the sense that the systemic option [interrogative] is realized by the ordering of the two functions, so that the realization spreads over more than one constituent of the clause. From this point of view, the concatenations S A F: FA S indicate a contrast analogous to falling versus rising tone. The structure of the Negotiator as outlined above pertains to the indicative mood types. How does it compare in the analysis of the imperative clause? There are some differences, as the analysis in Example 6 shows.
124 (6)
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH Hate-
Predicator
tot
Complement: clitic
So far as the modal structure of the imperative is concerned, it appears to consist simply of Predicator followed by a Complement-clitic (C-clitic). The clitic in Example 6 is coreferential with the Subject; and the function Subject is always implicit in the French imperative mood. (For more detail, see Section 4.3.3.) The verb se hater (to hurry up), like many other such verbs in French, is always constructed with a pronominal clitic and therefore the toi in hdte-toi must be treated as an obligatory element. With the analysis of this imperative clause as an example, we can postulate that the Negotiator in imperative clauses consists of an obligatory function, Predicator; and in addition, a C-clitic must occur if the verb in the Predicator is 'reflexive' as in Example 6. Though these represent some of the most crucial functions in the Negotiator, there is more to this element than is pointed out here; later we shall identify certain optional functions such as polarity and modal Adjuncts, in addition to functions that, depending on their textual status, might occur either inside or outside the Negotiator. 4.3 The modal structure in French: Negotiator and Remainder The account of clauses from the dialogue Text 4.1 has highlighted two important facts about the French clause as interaction. First, the replay of Subject (Subj), Finite (Fin) and Predicator (Pred) is one means of resolving an exchange in French. Second, the MOOD options have a prosodic mode of expression, which either solely employs the phonological means or a combination of both phonological and grammatical ones. We will now probe the interpersonal structure of the French clause by focusing on a range of exchanges which highlight the recurrent linguistic patterns central to negotiation. Section 4.3.1 will examine a series of adjacency pairs concerned with the giving and demanding of information. This will be followed in Section 4.3.3 by an account of adjacency pairs concerned with the giving and demanding of goods-and-services. 4.3.1 The clause as exchange of information: the structure of French propositions
This section is concerned with the description of the structure of clauses functioning in question-response in adjacency pairs. Questions can be further classified as confirmation questions or information questions. The realization of these speech functions is distinct; the former is realized by a polar interrogative, the latter by a nonpolar interrogative. Further, the element picked up in the response statement varies depending on whether the initiating question is a confirmation one or an information one. Typically, it is the Negotiator that is replayed in a response to a confirmation question, and the Remainder in a response to an information question. We
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
125
will begin by examining the type of adjacency pair which is initiated by a confirmation question as exemplified in the initiating move of Examples 7 to 15. These questions and their responses are discussed in subsection 4.3.1.1, whilst those adjacency pairs which are initiated by an information question as in Examples 16 to 19 are discussed in subsection 4.3.1.2. 4.3.1.1 Confirmation questions and their responses
In Section 4.2 we noted two means of realizing polar interrogatives: intonation, and the ordering of Finite "Subject (see Examples 1 and 2). A third means of realizing polar interrogatives is exemplified in the initiating move of Example 7. (Note that in the numbering of examples in this chapter as 7i, 7ii, etc. the lower case roman numerals represent parts of an exchange, in contrast to the lower case letters 7a, 7b, etc., in the numbering of examples in previous chapters, which represent clauses within a clause complex. Note also that more detailed interlinear glosses are given for examples in this chapter, because the presence and ordering of clitics and other grammatical particles is crucial to the realization of MOOD selections.) (7i)
Est-ce que tu (Is it that) you M-int Subj Do you see the moon?
(7ii) Out yes Polarity-marker Yes, I do
je I Subj
vois see Fin/Pred
la lune? the moon Comp
la it C-clitic
vois see Fin/Pred
The feature [polar interrogative] in the initiating move of Example 7 is realized by the Mood interrogator (M-int) est-ce que. Matthiessen (1992: 398) refers to this type of grammatical prosody as juncture prosody, since the element Mood-interrogator can only occur at the boundary of the clause. Although clearly a part of the interpersonal organization of the clause, it falls outside the Negotiatory structure: certainly, M-int is associated with the presence of the option [polar interrogative]; however, it is by no means crucial to the realization of that option, since the rising tone is by itself sufficient for the purpose. M-int has the status of Theme, pointing to the interpersonal role of the clause as a yes-no question; unlike other interpersonal functions such as Subject or Complement (Comp), it does not conflate with an experiential function. The nominal group la lune realizing the element Complement in Example 7i also falls outside of the Negotiatior: it forms the Remainder of the clause. We have thus identified three categories of interpersonal functions in the French clause: (i) those which enter into the Negotiator, e.g. Subject, Finite, Predicator, and the various clitics; (ii) those which enter into the Remainder, namely Complement(s) and/or Adjunct(s); and (iii) functions that remain peripheral to both the Negotiator and the Remainder, e.g. M-int. So far as
126
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
the French clause as an interactive move is concerned, it is the first two categories of interpersonal functions - those which enter into the Negotiator and the Remainder - that are of special interest here. The Negotiator and the Remainder together form a structure, which should, in fact, be viewed as the modal structure of the French clause in the sense that it is the immediate components of this modal structure that are relevant to negotiation in speech exchange. For this reason, it seems appropriate to call the modal structure of the French clause the Negotiatory structure. By implication, then, we are suggesting four layers of interpersonal structure as shown in Figure 4.1. In Section 4.2, it was claimed that the resolution of an exchange in French often revolves around the replay of the Negotiator. This claim is further supported by the responding move in Example 7ii. The response statement here is expressed by means of the Polarity-marker (P-marker) oui followed by the replay of all the interpersonal functions of the initiating move; only the latter - je la vois - enters into the negotiatory structure. Note that in this replay cohesive relations become central; thus the Complement of Example 7i la lune reappears in 7ii simply as la acting as a Complement-clitic. As a cohesive device (Halliday and Hasan 1985) la is given and recoverable. What can be recovered by both interactants from the context and stays a constant in the exchange is realized by a pronominal in the form of a clitic, which in the indicative is prefixed to the verbal group functioning as Fin/Pred, as it is in Example 7ii. However, given and recoverable entities may be indicated in other ways. So, in Example 8ii, it is full clausal ellipsis that marks this shared aspect of dialogue. The textual resources of both REFERENCE and ELLIPSIS are quite commonly employed in the type of exchange where the initial move is realized by a polar interrogative, and the following response by a declarative (see also Examples 21 and 22 in subsection 4.3.2). (8i)
Pierre voisPierre see Vocative Fin/Pred Pierre, do you see the moon?
tu you S-clitic
la lune? the moon Comp
(8ii) Non No P-marker No Est-ce due
tu
vois
la lune
M-int
Subj
Fin/Pred
Comp
M *' t
Remainder
Negotiatory st . . Interpersonal
Figure 4.1 The interpersonal organization of the French clause
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
127
Example 8i resembles Example 1; both realize the polar interrogative by ordering the Finite before Subject (FAS) and, in addition, the Subject in both is realized by a pronominal clitic. In Example 8i, the Subject-clitic (S-clitic) is coreferential with the nominal Pierre, which occurs in a thematic position. This nominal segment, called a Vocative, is outside of the negotiatory structure but functions within the interpersonal structure. This is different from the nominal element Paulin Example 9, which performs only a textual function and has no role in the interpersonal or experiential structures. (9)
Paul,
tu
le
vois?
Paul Absolute Theme Paul, do you see him?
you Subj
him C-clitic
see Fin/Pred
Nominals like Paul in Example 9, which enter neither into the interpersonal nor the experiential structure, are referred to as absolute Theme (Th-abs) (see Chapter 5). Nominals, such as Paul here, are not limited to having just this function; for example, Paul could also have been assigned the function of New, realized phonologically by tonic salience. In that case, it would have had a contrastive meaning ('you cannot see the others but can you see Paul?'). Rothemberg (1989: 153) points out that 'orally, the organization of the clause as message is not dependent on word-order alone. Intonation can assign the role of Rheme2 even to a term which is not in final position' [my translation, AC]. She goes on to say that 'graphically, to convey the information that the one at whom the question points is Paul, not others, the solution is to add a tonic pronominal following the absolute Theme, as in "Paul, lui, tu I'as vu?"' [my translation, AC] where she would interpret Paulas Theme and lui as Rheme. However, following Halliday's distinction between ThemeRheme and Given—New structures (see Chapter 5), we interpret both Paul and lui in Rothemberg's example as absolute Themes, treating Paul as Given and lui as New. (For further discussion of some aspects of the textual organization of the French clause specifically from a dialogic point of view, see Section 4.5 and Chapter 5.) Example 9 also shows that, unlike English, the Finite is not always discretely realized in an interrogative clause, but can be fused in the Predicator depending on tense selection, in which case the Subject is ordered in relation to the Predicator with which the Finite is fused. Thus the generalization holds true that Subject is always ordered in relation to the verbal constituent which specifies direct temporal reference to the speech event, that is, the Finite. This constituent is discrete when the realization of temporal relations is complex, and fused when the tense is simple, i.e. when the tense selection is either simple past, imperfect past or present. From this it follows that whenever Finite is discretely expressed, and the interrogative mood is partly/wholly realized by ordering Finite before Subject, the latter would intervene between Finite and Predicator in the Negotiator of the clause; otherwise, it follows the Predicator.
128 A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Consider Example 10, where Paul is in final position. (101) Tu
le
You it S-clitic C-clitic Do you see it, Paul?
vois,
Paul?
see Fin/Pred
Paul
(lOii) Oui. Yes P-marker Yes. Like the nominal Paul in Example 9, Paul is absolute Theme, having simply a textual function and not conflating with any interpersonal or experiential function in the clause. Such a Theme, when it is clause-final, is here referred to as reprise Theme (Th-rep). In Example 10, the reprise Theme conflates with the function Given, in the sense that it is not contrastive. If instead of Paul, we had the tonic pronominal lui as in tu le vois, lui (but not Pierre), then lui will have been both reprise Theme and New (see Chapter 5 for a more detailed discussion of absolute Themes). If, as we have proposed, the Negotiator realizes the MOOD selections, then clearly this realization is prosodic. According to Halliday (1979) the prosodic mode of expression is not restricted to MOOD selections alone; it extends to all the interpersonal resources as seen in Chapter 2 in the discussion on the subjunctive for example. Interpersonal prosodic patterns of realization can also be seen in Example 11. Here negative polarity is realized as a prosody ne . . . pas, the items occurring in different places in the clause. Negation occurs first following tu as a negative clitic Adjunct (A-neg-clitic), ne, then it occurs after the Finite as negative non-clitic Adjunct (A-neg) pas. Pas may be replaced by other negative elements, for example plus (any more),jamais (never), or by a negated Complement or one that is itself a negative form, e.g. personne (nobody). If there is a negative Subject, then the negative Adjunct pas is not present, as in personne n 'est venu (nobody came) (see Battye and Hintze 1992). Note in passing that the polarity marker in Example llii also shows that the polarity positive response to a negative question is si rather than oui. (Hi) Tu
ne. . .
You NEC Subj A-neg-clitic Didn't you see it?
/'
as
...pas
vu?
him C-clitic
have Fin
NEG A-neg
seen Pred
(llii) Si Yes P-marker Yes, I did The lexicogrammatical realization of negative polarity may vary depending on tenor and mode. Battye and Hintze (1992: 268) note that in 'less formal
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
129
styles of spoken French, it is common for the first element ne not to appear . . .'. They give the following examples: (12) Je vois I see I don't see Marie
pas not
(13) Je vois I see I can't see Marie any more
plus any more
(14) J' ai I have I didn't eat any pie
Marie Marie Marie Marie
mange eaten
aucune no
tarte pie
On the other hand, 'in written French there is a small class of verbs which permit negative structures to be formed by the use of ne alone. These are pouvoir (be able to), savoir (to know), cesser (to stop) and oser (to dare)' (Battye and Hintze 1992: 269). It is obvious from these examples that, in addition to Subject, Finite and Predicator, the Negotiator may have a Complement clitic and/or negative clitic and/or non-clitic Adjuncts. The responding move in Example 15 below illustrates that more than one Complement clitic may be attached to the Finite. (15i) AsHave Fin
tu you Subj
donne given Pred
lelivre the book Comp
a Paul? to Paul Comp
Did you give the book to Paul?
(15ii) Out Yes P-marker
je I S-clitic
le it C-clitic
lui to him C-clitic
ai have Fin
donne given Pred
Yes, I did give it to him
The relative ordering of Complement clitics is governed by a number of semantic variables, which will be discussed further in Section 4.5. Interpersonally the Complement clitics with interactant roles (e.g. me, te) precede non-interactant Complement clitics (e.g. le, Id). If there are two third-person non-interactant clitics as in Example 15, then the accusative pronominals (e.g. le) always precede dative pronominals (e.g. lui), whatever the MOOD choice. Thus they follow the ordering of the nominals they cohere with. This is in contrast with the clitics referring to the interactant roles: here the accusative pronominal clitic does not necessarily have to precede the dative one, as can be seen from Paul me le donne (Paul gives it to me) where the dative me precedes the accusative le. 4.3.1.2 Information questions and their responses
Turning to adjacency pairs where the initiating move constitutes a demand for information rather than for confirmation, such questions are
130
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
congruently realized by nonpolar interrogatives. In Example 16i, the focus is not on polarity, but on a missing element of information; this is what the Qu- element stands for. (16i) Quand When Qu-Adj
est-ce que
tu you
arrives ? arrive
M-int
S-clitic
Fin/Pred
Negotiator
Remainder When are you arriving? (16ii) Demain Tomorrow Adjunct Remainder Tomorrow.
The Qu-element may conflate with the interpersonal functions Subject, Complement or Adjunct. In Example 16 it is conflated with the Adjunct function and followed by the interrogative Mood-interrogator est-ce que. In discussing the structure of the polar interrogative, we identified three different means of realization: presence of Mood-interrogator, tone, and the ordering of the Finite before the Subject. These three means are available to nonpolar interrogatives as well. However, when the Qu-element is conflated with the Subject, the possibility of inversion does not exist, and the conflated Qu-Subject, together with the Finite, functions within the Negotiator, as Example 17 illustrates.
(17i) Qui Who
est is
arrive'? arrived
Qu-Subject
Fin
Pred
Negotiator Who has arrived? (17ii) Ton cousin Your cousin Subject Your cousin
In spoken language, when the Qu-element conflates with either an Adjunct or a Complement that refers to a non-human entity, then it may occur in Rheme position, as Examples 18 and 19 illustrate. In this environment, Subject always precedes Finite. Note that in French this interrogative is not limited to simply realizing echo questions; it can also realize an initiating question, at least in the spoken mode.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
(18i) Tu you
arrives arrive
quand? when
Subj
Fin/Pred
Qu-Adj
Negotiator
131
Remainder
When do you arrive? (18ii) Tu
you
paries speak
quoi ? what
Subj
Fin/Pred
Qu-Comp
Negotiator
Remainder
What do you speak?
In Example 18, quoi is the tonic form of que, the latter is used only in initial position, as in Example 19. (19)
Que what
veuxwant
tu? you
Qu-Comp
Fin/Pred
S-clitic
Remainder
Negotiator
What do you want?
The preceding examples serve to illustrate how exchanges of information progress in French and to exemplify the various prosodic structural patternings that realize MOOD selections. Before turning to a discussion of the French proposals, it might be helpful to say a few more words in section 4.3.2 about the Negotiator. 4.3.2 Negotiator, clitics and cohesion
In Section 4.2, where the Negotiator was first introduced, it was pointed out that this component of the French clause consists of the functions Subject, Finite and Predicator (see Example 5). During the discussion of the adjacency pairs, attention was drawn to the fact that when Complement and Adjunct are realized by a clitic, these form part of the Negotiator, whereas when they are realized by nominals and prepositional phrases, respectively, these functions enter into the make-up of the element we have referred to as Remainder. The clitics - whether they realize polarity, Complements and/or Adjunct - are thus integral to the Negotiator: whenever they occur, they are crucial to negotiation, so that it is not simply Subject, Finite and Predicator but the entire complex consisting of Subject, Finite, Predicator and the clitics that is involved in the negotiation. The Negotiator is, thus, crucial to the arguability status of the French clause. In dialogues it is this part of the clause which is being tossed back and forth, as in the exchange in Example 20.
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
132
(20i) Remi, tu m' Vocative Subj C-clitic Remi, are you listening to me? (20ii) Mais
oui P-marker Well, yes, I am.
(20iii) Non P-marker
ecoutes? Fin/Pred
Subj
tu Subj
ecoute. Fin/Pred
t'
Je
C-clitic
ecoutes Fin/Pred
n
Neg-clitic
nen.
Neg-Adj
No, you are not.
j
(20iv) Si P-marker
ecoute. Fin/Pred
Subj
Yes, I am.
If ellipsis occurs, it is defined in relation to the Negotiator or the Remainder. Thus, with full ellipsis, both Negotiator and Remainder are ellipsed and the clause 'has' only textual functions, for example, a polarity marker and/or absolute and textual Themes. Partial ellipsis involves either the whole of the Negotiator or whole or part of the Remainder, as in these adjacency pairs: Tu ecoutes?- Oui (Are you listening? -Yes); Quand viens-tu?- Demain (When are you coming? - Tomorrow); Tu aimes la cuisine italienne oufranfaisef-Franfaise (Do you like Italian or French cuisine? - French). Further, exchanges of information initiated by a confirmation question tend to be carried forward by the Negotiator, displaying ellipsis of whole or part of the Remainder, while information exchanges initiated by an information question tend to progress around the Remainder and thus display ellipsis of the whole of the Negotiator. Examples 21 and 22 illustrate how the textual systems of ELLIPSIS and REFERENCE are used in negotiation. (21i) Est-ce que tu M-int
you
pars go
en vacances on holidays
demain ? tomorrow
Subj
Fin/Pred
Adj
Adj
Negotiator
Remainder
Are you going on holiday tomorrow? yes
j' I
y
there
g°
P-marker
Subj
A-clitic
Fin/Pred
(21ii) oui
pars
Negotiator Yes, I am.
Here the initiating question in Example 21i is a confirmation question. Its interpersonal structure is Mood-interrogatorANegotiatorARemainder as shown in the analysis. Note that the Negotiator consists of SubjectAFinite/
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
133
Predicator expressed by tu pars. The response in 21ii is a declarative, in whose structure the polarity marker is followed by the Negotiator. This Negotiator is related cohesively to the elements of the Remainder in 21i. Thus, 31 is in anaphoric reference relation to en vacances, and there is ellipsis of demain. Both these cohesive relations are with elements that form part of the Remainder in 21i. Compare this with Example 22, which is initiated by a demand for information realized by a nonpolar interrogative. (22i) Quelle langue Remainder
parles-tu? Negotiator
Which language do you speak? (22ii) Franglais
Remainder Franglish
The response to Example 22i in 22ii is a declarative, itself consisting of the element Remainder and displaying the ellipsis of the Negotiator which consists of Fin/Pred A Subj parles-tu. These functions are presumed from the first part of the adjacency pair, i.e. from 22i. A possible continuation of this exchange is shown in Examples 22iii and 22iv, where 22iii is another question and 22iv its response. Here, the elliptical clauses consist solely of the Remainder; nonetheless, the Negotiator still plays an important part in carrying the exchange forward, being presumed by ellipsis. The Negotiator is thus a constant throughout the dialogue; it is always relevant and recoverable from the cotext. (22iii) Franfais et Anglais? Remainder French and English? (22iv) Non P-marker
Franglais Remainder
No, Franglish.
The fact that interpersonal prosodies other than those indicating MOOD have the Negotiator as their domain of realization further supports the interpretation of this element as interpersonally salient. An example of this has already been given above in Example 11, where we drew attention to the negative prosody. A further example of this is found in the modal
134 A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
prosodies within the Negotiator. Modality can be expressed repeatedly in the Negotiator both in the Finite and Predicator and as a modal Adjunct, as in Example 23.
(23) Jean Jean
pourrait could
peut-etre maybe
faire do
(a this
Subj
Fin-mod
Adj-mod
Pred
Comp
Negotiator
Remainder
Jean could maybe do this Here the meaning of probability is expressed three times. First by the modal verb, pouvoir, second in the Finite (Fin-mod) which makes modal rather than temporal reference to the speech event in pourrait and third by the modal Adjunct (Adj-mod), peut-etre. Modal Adjuncts do not only realize modality but also 'presumption, time, degree and intensity' (Halliday 1985a: 82). Thus the postulate of the Negotiator appears justified, both by its function in the speech exchange and by the fact that its mode of expression is prosodic. It is that part of the clause that must be replayed in the negotiation of speech exchange. Not surprisingly, it is 'always there', either overtly or by cohesive presumption. We turn now to a fuller structural account of the Negotiator. If the analysis of the examples discussed so far is examined, it will be seen that the functions of Subject, Finite and Predicator are obligatory: there is no Negotiator in the indicative clause that does not have these functions either explicitly or implicitly. In addition to these, we find certain other functions that are optional. These include a negative Adjunct clitic and/or a negative Adjunct (both exemplified in Example 11) and/or modal Adjuncts (shown in Example 23), Complement clitics (various examples), and/or an Adjunct clitic (see Example 21ii). The unmarked order in which these functions may occur in the Negotiator varies somewhat depending on whether the Finite is discrete or fused. By unmarked order, we mean Subject before Finite, which is always the case with declaratives, often also with the interrogatives, though in some cases they may be reversed, as for example in Example 2 (see Section 4.2). The unmarked order of the various functions in the Negotiator is as shown in Table 4.1. Examples 24 and 25 illustrate the maximal structures A and B (shown in Table 4.1) respectively.
(24) Je I
S
lui him
ne not
le it
A-neg-clitic
C-ckutuc c-ckutuc
Negotiator
I probably didn't give it to him
ai have
probablement probably
pas not
Fin
A-mod
A-neg Pred
donne given
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
135
Table 4.1 The unmarked order of interpersonal functions in the French clause A
Subj A (A-neg-clitic)A (C-clitic) A (A-clitic) A Fin A (A-mod) • (A-neg)APred
B
SubjA (A-neg-clitic)A (C-clitic) A (A-clitic) A Fin/Pred A (A-mod) • (A-neg)
Key. X A Y = element Y follows element X; X»Y = elements X and Y are unordered with respect to one another; X/Y= elements X and Y are conflated; (X) = element X is optional
Compare the structure of Example 24 with the structure potential of the A variant of the Negotiator shown in Table 4.1. With respect to English, Matthiessen (1995: 397) points out that: In a negative clause, it is the Subject that is outside the negative prosody in the unmarked case precisely because it is the element on which the argument rests. With respect to Subject, the proposition or proposal is negative rather than positive . . .
The same observation appears to apply to the French Subject in Example 24 which also falls outside the negative prosody. The same is however not true of Example 25, which is an example of the maximal structure B shown in Table 4.1.
(25)
Ne not
le it
A-neg-clitic C-clitic
a-(t)
encore
lui him
have
41 he
pas not
yet
donne given
C-clitic
Fin
S-clitic
A-neg
A-mod
Pred
Negotiator Hasn't he given it to him yet?
Examples 24 and 25 show quite clearly that the Negotiatory element has the potential of functioning as a complete clause on its own which has the verbal group as its domain of realization, which maps onto the clause nucleus discussed in Chapter 3. Note that when the Subject is realized as a nominal rather than a clitic it is not attached to the verbal group, but it still remains part of the Negotiator. It is the only participant of the Negotiator which may be realized as either a pronominal clitic or a nominal. The shift from pronominal clitic Subject to nominal Subject correlates with the assignment of modal responsibility to some other function. This other function is usually a Complement, and it is secondary to the negotiation process. The status of the Complement is variable. When non-clitic, the Complement is part of the
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
136
Remainder (compare, for example, Examples 7 and 8 in Section 4.3.1.1) and typically has the textual status of New. The clitic status of the Complement shows that it is at the centre of the negotiation; and once this happens, then quite predictably it becomes Given. Our statement of the structure potential of the unmarked Negotiator has shown that its minimal structure consists of Subject, Finite and Predicator. A minimal Negotiator is still capable of functioning as a complete clause. When the Subject in a minimal Negotiator is a clitic, with the Negotiator functioning as a complete clause, the latter presents itself as a Clause nucleus, in experiential terms (see Chapter 3, Section 3.6). To elaborate, the minimal Negotiator with S-clitic consists of constituents which are in themselves sufficient to support the realization of an experiential structure. In this respect the Negotiator is quite different from the Mood element in English, which by itself cannot be assigned an experiential structure. 4.3.3
The clause as exchange of goods-and-services: the structure of proposals
As in English, so in French, there does not seem to be a special lexicogrammatical structure dedicated specifically to the realization of offers. We turn therefore to the demand for goods and services, i.e. the command type, realized grammatically as the imperative clause, the structure of which was briefly visited in Section 4.2. In this section we ask how the Negotiator in the imperative compares with that of the indicative (as shown in Table 4.1). The first point to note is that imperatives do not have a Finite element. They specify neither modal nor temporal reference to the speech situation. Secondly, although the Subject does not appear overtly, its person, number and social distance (formal vs informal) are realized syncretically in the verb which functions as Predicator. Thus the only obligatory function to appear discretely in the Negotiator of an imperative is the Predicator, as illustrated by Examples 26 to 29.
(26i) Ditestell-2SING-FORM
moi me
la verite the truth
Pred
C-clitic
Comp
Negotiator
Remainder
Tell me the truth (26ii) Non,
No
je I
ne not
vous you
la it
dirai will tell
pas not
P-marker
S-clitic
A-neg-clitic
C-clitic
C-clitic
Fin/Pred
A-neg
Negotiator No, I won't
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
you
distell-2- SING- INF
moi me
la verite the truth
Theme/New
Pred
C-clitic
Comp
(27i) Toi,
137
Remainder
Negotiator
You, tell me the truth! (27ii) C' it S-clitic
est is
quoi? what
Fin
Qu-comp
Negotiator
Remainder
What's truth? a la plage to the beach
(28i) Allans Let's go Pred-S+
Adjunct
Negotiator
Remainder
y
(28ii) AllonsLet's go
there
Pred-S+
Adj-clitic
Negotiator
The first move of Examples 26 to 28 each enacts a command which is realized as an imperative. The responding move in each case consists of the Negotiator alone. In Example 27, the responding move is a challenge rather than a compliance or an initiating question. These examples highlight the respects in which the interpersonal organization of the imperative differs from that of the indicative. To reiterate, the imperative has no Finite, nor does it have an explicit Subject. Rather the person and number of the implicit Subject of each clause is marked on the Predicator, and predictably the implicit Subject of a French imperative is either a second person or 'first person plus', the latter being different from first person plural. The marking of these as well as of formality is indicated clearly in the above analysis. A further feature of imperatives should be noted: clitics, whenever they occur, follow the Predicator, except where the clause has a negative prosody, when the order of the clitics with respect to the Predicator is reversed, as in Example 29. (29) Ne not
le it
A-neg-clitic C-clitic Negotiator Don't give it to him!
lui him
give-2-SING
pas! not
C-clitic
Pred
A-neg
donne
138
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
This closes our discussion of the structure of the French clause as an interactive move in an exchange. The description offered here has highlighted the specific interpersonal functions which occur in the Negotiator. The Negotiator and the Remainder form what has been called the negotiatory structure, which constitutes the essence of the modal structure of French. We have also encountered some functions that are outside the negotiatory structure, such as the Mood-interrogator, Vocative, absolute Theme and Polarity markers. This chapter has suggested that the Mood-interrogator and Vocative were part of the interpersonal structure but outside the Negotiatory structure; the textual function of absolute Theme (initial or reprise) and Polarity markers were seen to have an important role in dialogue. One strong motivation for dividing the clause into the Negotiator and the Remainder was provided from a consideration of the behaviour of ellipsis, which is defined in relation to these two parts. Thus, ellipsis may either be of the entire Negotiator or Remainder or both, when simply a P-marker, e.g. oui or non, might occur. If there is an ellipsis of the Negotiator, all of its functional parts must be ellipsed. Thus, unlike English, a move cannot be expressed by a replay of just Subject and Finite (Halliday and Hasan 1976, 1985; Martin 1992). The fact that Finite and Predicator must always function together in French indicative clauses, whether the two are fused or not, in addition to the fact that both Finite and Predicator may realize modality (see Example 23 in Section 4.3.2), suggests that both elements in conjunction with the Subject make the clause arguable. So far, this description of the structure of the clause as a move in dialogue in French has been provided in terms of functions in the syntagm — what elements must occur in a structure, and in what order. In the following section, we attempt an examination of the system 'behind' the structural organization of the French clause as a move: the system of MOOD, as well as the system of MODALITY. The system of MODALITY refers to the range of meaning between the polarity poles, that is, in a proposition (i.e. a clause that gives or demands information), between 'yes' and 'no', or, in a proposal (i.e. a clause that gives or demands goods-and-services) between 'do' and 'don't' (see Halliday 1985a). 4.4 The interpersonal systems of MOOD and MODALITY In this section we will first explore the systems of MOOD that realize speech functional semantics and then the systems of MODALITY as means of expanding the realizational potential of speech functional semantics through interpersonal metaphors. For example, the speech function of command which is congruently realized as an imperative in the MOOD grammar, as in ouvre laporte 'open the door', can be realized metaphorically by combining options from the MOOD and MODALITY systems, as in tu devrais ouvrir laporte 'you should open the door' (i.e. declarative mood with obligation) or pourrais-tu ouvrir la porte? 'could you open the door?' (interrogative mood with probability).
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
139
4.4.1 The primary MOOD options
As far as specific realizations of MOOD selections are concerned, we saw that French makes use of all three types of realization encountered across languages: tone, mood particle and Subject-Finite ordering. The system itself will have to reflect these realizational differences through more delicate features. The lexicogrammatical network presented below will attempt to balance the semantic contrasts of French with the structural realizations. We saw in Text 4.1 that the three indicative clause types - declarative, exclamative and interrogative - may have the same interpersonal organization with the Subject preceding the Finite, differing in terms of tone contours which are indicated in writing by the use of different punctuation marks. The unmarked intonation of French polar interrogatives is a rising tone, and of the nonpolar interrogatives, a falling one. A declarative clause may have a falling or rising-falling tone depending on how many information units it realizes, but it will always end with a falling tone. Exclamatives too have a falling tone. Battye and Hintze (1992: 143) point out that for declaratives the last syllable of the tone will be on a low pitch falling below the normal speaking range. For nonpolar interrogatives and exclamatives, the first syllable will start off on a note which is slightly above that of a normal speaking range.
Thus, both tone and pitch are crucial to the realization of MOOD features in French, as they may be the only means of manifesting the systemic distinctions. Secondary options are distinguished by means of different modes of prosodic expression and thematic organization. Since we have approached the MOOD grammar of French from dialogue, the options systematized in the networks presented here are representative of spoken language. Some of the resources in the interrogative and exclamative systems are not found in the written mode. Figure 4.2 presents the primary options of the MOOD system, those which are expected to be applicable to most languages. As in such network diagrams in previous chapters, the slanting arrow below the option [indicative] in Figure 4.1 indicates 'is realized by', and the realizational patterns (shown in italics) follow such arrows. In Figure 4.2, this
Figure 4.2 Primary MOOD options
140
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
arrow indicates that the option [indicative] is realized by the insertion of functions Subject, Finite and Predicator into the structure, so that any clause with the feature [indicative] must 'have' these functions. In Figure 4.2, the other three options are followed by boxed information which is a guide to the development of the network, in this case indicating the Figures in which the more delicate options of the network will be displayed in this chapter. Thus the options dependent on the features [informative], [interrogative] and [imperative] are presented in Figures 4.3, 4.4 and 4.5, respectively. Note that the term [informative] is used here in preference to the term 'affirmative', to avoid connotations of positive polarity (Martin 1992). 4.4.1.1 Options of the feature [informative] The option [informative] permits a choice between [declarative] and [exclamative]. In French, the crucial properties of the structure of the declarative clause are very easily stated. All French declaratives must 'have' Subject, Finite and Predicator, in that order. These functions are inserted as a response to the feature [indicative] (see Figure 4.2); the criterial ordering of Subject and Finite is indicated in Figure 4.3. Figure 4.3 shows that the term contrasting with [declarative] is [exclamative]. Grammatically, exclamative clauses may sometimes be similar to interrogatives but semantically they are closer to declaratives in the sense that they give rather than demand information; the information is primarily attitudinal rather than factual. This is the reason for treating [exclamative] as an option dependent on [informative] in Figure 4.3. The Qu-expression that appears in an exclamative such as qu 'il est sage! or qu'est-ce qu'il est sage! superficially resembles the Qu-element in nonpolar interrogatives. However, the two are significantly different. The nonpolar interrogative Qu-element has the realizational resource of a full paradigm consisting of such items as que, qui, quand, oil, etc. Further, each of these, when it occurs in a clause, conflates with some interpersonal and experiential
Figure 4.3 Options of the feature [informative]
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
141
function. The que which functions as an exclamator in the exclamative contrasts only with comme; and neither of these exclamators - que or comme conflate with any experiential function in the clause. The exclamators, que or comme, are not like Qu-elements; rather, their function is comparable to est-ce que in a polar interrogative: both function simply as Mood-markers. The situation is further confounded by the fact that the expression est-ce que may also occur in exclamatives as in our example qu'est-ce qu'il est sage! When the est-ce que expression occurs in an exclamative, it no longer has the function of MOOD marker, for there is no 'interrogativeness' about the exclamative. Rather, the choice of est-ce que in an exclamative is indicative of certain register variables. Instantiations for each possible selection expression applicable to the feature [exclamative] are presented in Table 4.2. The left column in Table 4.2 provides examples of the exclamative clause type, while the systemic features relevant to that type are presented in the right column. The five instances of French exclamatives could be translated by the English clause How good he is! Note that the Subject of a French exclamative clause may be brought into prominence indirectly through its cohesive relation to a nominal functioning as absolute (initial or reprise) Theme and/or as New, as illustrated in Examples 30 to 32. (30)
Get enfant, This child,
qu ' how
il he
est is
sat^e! good
Th-abs
M-marker
Subj
Fin
Comp Remainder
Negotiator
How well-behaved this child is!
(31)
Qu 'est-ce qu ' How
il he
est is
sage, good
cet enfant! this child
M-marker
Subj
Fin
Comp
Th-rep
Negotiator
Remainder
How well-behaved this child
Table 4.2 Exclamative clauses: features and examples example
systemic features
il est sage! est-il sage! comme il est sage! qu 'il est sage! qu 'est-ce qu 'il est sage!
no grammatical prosody grammatical prosody: internal grammatical prosody: juncture: comme grammatical prosody: juncture: que: que-alone grammatical prosody: juncture: que: que-plus
142
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(32) Qu' How M-marker
il he
est is
sage, good
lui! him
Subj
Fin
Comp
Th-rep
Negotiator
Remainder
How well-behaved he is! In Examples 30 and 31, the nominal cet enfant functions as the absolute Theme; it is initial Theme in Example 30 and reprise Theme in 31, and in both cases it is related cohesively to the Subject il - cataphorically in Example 30, anaphorically in 31. In Example 32, lui is both the absolute reprise Theme and the New; further, it is anaphorically related to the Subject il. 4.4.1.2 Options of the feature [interrogative] The interrogative network presented in Figure 4.4 starts with two simultaneous systems: the INTERROGATIVE TYPE system acting as the entry condition to the options [polar] versus [nonpolar], and the INTERROGATIVE MARKING system leading to the options [grammatical prosody] versus [no grammatical prosody]. The option [no grammatical prosody] implies that phonological prosody - i.e. intonation - is the only mode of realization; by contrast, the option [grammatical prosody] means that, in addition to intonation, a grammatical prosodic expression will occur in the clause. This grammatical prosody may occur at clause [juncture] or it may be clause [internal]. With [juncture], the Mood-marker est-ce que is chosen as the clause initial element, whereas with the choice [internal], the prosody takes the form of inversion whereby Finite is ordered before Subject. If a clause 'is' both [polar] and also [no grammatical prosody], then this conjunction permits a choice between [tagged] and [untagged]. The choice between [grammatical prosody] and [no grammatical prosody] is motivated by discourse factors, whether the mode of discourse is spoken or written and furthermore whether the tenor of the relationship is formal or informal. The realization of the option [nonpolar] calls for the insertion of a Quelement, which typically occurs clause-initially. The Qu-element may conflate Subject or other function, as reflected in the systemic options [qu-Subject], and [qu-other]. With the option [qu-Subject], the Qu-element can only occur clause-initially, i.e. the option [internal] is closed to it, and the Subject, qui, always precedes the Finite. The option [qu-other] permits a choice between [qu-Comp] or [qu-Adj]. The [qu-Comp] may be [human], in which case it is realized as qui; if it is [non-human], it is realized as que when [thematic] and as quoi when [non-thematic]. Figure 4.4 shows that the contrast between [thematic] and [non-thematic] is simultaneous with that between [qu-Comp] and [qu-Adj]. If the option [thematic] is selected, the Qu-element is positioned clause initially; with [non-thematic] it is located clause-finally, serving as 'focus'. When the options [qu-Comp] and [thematic] are co-selected, the clause must 'have' a grammatical prosody,
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
143
Note: paired symbols (e.g. # -> and -> #) indicate if/then relationships between the paired features, i.e. if the feature [qu-Subject] is selected, then the feature [juncture] will also be selected
Figure 4.4 Options of the feature [interrogative] either [juncture] or [internal]. If [non-thematic] is selected, then [no grammatical prosody] is the only available choice. Table 4.3 exemplifies the range of polar interrogatives and Table 4.4, the range of nonpolar interrogatives. Note that, unlike the English tag, the French tag does not pick up the Subject and Finite. The 'now moribund' n'est-ce pas (see Coveney 1990), which literally translates into the English 'isn't it', now functions as equivalent to 'don't I', 'haven't you', 'won't they', etc. and the use of hein and eh a tags has quite overrun the use of n'est-ce pas in casual conversation. Coveney (1990) mentions two types of nonpolar interrogatives realized with variants of the interrogative Mood-marker est-ce que which are not shown in Table 4.4. These realizational variants were first identified by Behnstedt (1973) who found them in working-class spoken contexts; examples of these variants are displayed in Examples 33 and 34. (33) qui who qu-Subj
c 'est qui
/' it
a has
mis? put
M-marker
Comp
Fin
Pred
Who put it?
(34) oil where qu-Adj
qu'
on we
va go
M-marker
Subj
Fin/Pred
Where are we going?
144
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 4.3 Polar interrogatives: features and examples examples
systemic features
tu aimes les gateaux'? tu aimes les gateaux, hein ? est-ce que tu aimes les gateaux'? aimes-tu les gateaux?
no grammatical prosody: untagged no grammatical prosody: tagged grammatical prosody: juncture grammatical prosody: internal
Table 4.4 Nonpolar interrogatives: features and examples examples
systemic features
qui a mange ce gateau ? qui est-ce qui a mange ce gateau ? que vois-tu ?
qu-Subject: no grammatical prosody qu-Subject: grammatical prosody: juncture qu-Comp: non-human: thematic: grammatical prosody: internal qu-Comp: human: thematic: grammatical prosody: internal qu-Comp: non-human: non-thematic: no grammatical prosody qu-Comp: human: non-thematic: no grammatical prosody qu-Comp: non-human: thematic: grammatical prosody: juncture qu-Comp: human: thematic: grammatical prosody: juncture qu-Adj: thematic: no grammatical prosody qu-Adj: non-thematic: no grammatical prosody qu-Adj: thematic: grammatical prosody: internal qu-Adj: thematic: grammatical prosody: juncture
qui vois-tu ? tu vois quoi? tu vois qui ? qu'est-ce que tu vois? qui est-ce que tu vois? oil tu vas? tu vas ou?
oil vas-tu? oil est-ce que tu vas?
This subsection has highlighted the various modes of prosodic expression that serve to realize the options dependent on [interrogative] MOOD. It should be noted that the various options in the INTERROGATIVE MARKING system not only indicate variable ways of expressing the same interrogative types, they also embody different valeurs as a result of their use in differing environments. Thus, the feature [no grammatical prosody] tends to be selected in a spoken context where interactants have an informal relationship. The features [grammatical prosody] and [internal] tend to be used in the written mode, and in particular in literary texts, while [grammatical prosody] and [juncture] tend to be used in both the written and spoken mode, having become the most neutral way of asking questions.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
145
4.4.1.3 The options of the feature [imperative]
From the realizational viewpoint, the imperative MOOD differs from the indicative by virtue of the absence of the elements Subject and Finite. However, in the imperative, the features of person and number relevant to Subject are marked on the Predicator, while in the indicative they are marked on the Finite. This marking within the verbal group thus becomes a means of recognizing which nominal has the function of Subject; and the principle applies both in propositions where Subject and Finite are in agreement and in proposals where it is Subject and Predicator that agree in number and person. We shall return to this in Section 4.3. The absence of Finite in imperative clauses is explained by the fact that such clause types do not specify temporal reference to the speech event. The imperative system represented in Figure 4.5 distinguishes between [exclusive] and [inclusive] imperative clauses. The former is oriented towards the addressee (s), held responsible for complying, while the latter is oriented towards both the speaker and addressee. Note that the realization of the systemic options in each case implicates the Predicator; this is because such distinctions as that between [exclusive] and [inclusive] are marked on that element, and this marking is always indicative of the implied Subject. Table 4.5 gives some examples of imperative clauses and their systemic features.
Figure 4.5 Options of the feature [imperative]
Table 4.5 Imperative mood: features and examples examples
systemic features
mangeons! mange, Paul mangez, monsieur mangez, les enfants
inclusive exclusive: singular: informal exclusive: singular: formal exclusive: plural
146
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
The imperative differs from the indicative in that it neither has a Finite nor an explicit Subject. Further, in an indicative clause clitics precede the Predicator, while in an imperative clause they follow it so long as the polarity is positive. Thus in Example 35, which is [indicative: informative: declarative], the clitics precede the Predicator donne, whereas in Example 36, which is [imperative: exclusive: singular: informal], they follow it. (35) Je le I it I give it to him (36) DonneleGive it Give it to me
lui him
donne give moi me
When the final clitic of an imperative clause is first or second person it is realized in its tonic form, i.e. as moi or toi. In some dialects of French, Example 37 is an alternative to 36. (37)
DonnemeGive me Give it to me
le it
Declarative clauses differ lexicogrammatically from exclamative and interrogative clauses in that their MOOD is never realized by a juncture particle, which is comme or que in exclamatives and est-ce que in interrogatives. The increase in the use of est-ce que in both spoken and written discourse to realize interrogative correlates with a lessening of the order Fin/Pred A Subj or Fin A Subj as a means of its realization. Although the relative order of Subject to either Finite and/or Predicator may serve to realize mood selections, in the majority of cases the Subject will precede the Finite whatever the MOOD. It is this ordering that has been referred to as the unmarked one. Thus, obviously, we cannot generalize quite as we can in English, that the order of Subject and Finite realizes MOOD selections. The dominant variables which distinguish the mood features are the realizational resources of intonation and the presence or absence of Mood markers. This section has focused on the options within the system of MOOD. We will now turn, in Section 4.4.2, to the systems of MODALITY. 4.4.2
Modality
Within the systemic functional framework, modality can be of two types depending on whether the clause is an 'information' clause (a proposition) or a 'goods-and-services' clause (a proposal): the two types of modality are known as modalization or modulation, respectively (see Halliday 1985a/ 1994). Modalization covers the resources for assessing the probability or usuality of a proposition. Modulation covers the resources for assessing the
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
147
obligation or readiness of a proposal. Examples of probability and obligation were presented in Chapter 2 in the discussion on the use of the subjunctive mode. As pointed out in that chapter, the subjunctive mode is used in the prosodic expression of low probability and high obligation in dependent clauses and attributive clauses, where the attribute is a modal and the Carrier a downranked clause. Indicative, on the other hand, is used in the expression of median or high probability and low obligation. The following examples show this contrasting selection of the subjunctive and indicative in relation to modality. median probability; indicative (38) Rest probable [[qu 'il viendra]] It is probable that he will come low probability; subjunctive (39) II est possible [[qu 'il vienne]] It is possible that he comes/that he will come a: high obligation; P: subjunctive (40a) a Jeveux I want (40b) (3 que tu viennes that you come a: median obligation; P: indicative (41 a) a Je suppose I suppose/expect (41b) P que tu viendras that you will come
Examples 38 to 41 highlight certain preselection patterns for choosing the subjunctive: low probability or high obligation. These modality options are realized as a prosody across the clause or clause complex. There is a clear relationship between the system of MOOD and that of MODALITY in the sense that modality choices serve to expand the realizational potential of speech functional semantics by means of interpersonal metaphor. The different resources that serve to realize modality in French are discussed below. The modality potential in French is represented systemically in Figure 4.6. As can be seen in Figure 4.6, as well as the contrast between modalization (of propositions) and modulation (of proposals) already outlined, there are a number of other ways in which modality varies. Halliday (1985a) recognizes three degrees of probability, usuality, obligation and inclination: low, median and high; these are represented in the network in two steps, which accounts for the differential use of the subjunctive (with low modality) and the indicative (with non-low, i.e. high or median modality) just discussed (and also discussed in Chapter 2). The orientation of the modality may be subjective or objective, corresponding respectively to the projecting clause being personal (je+ mental process) or impersonal (il+ est+ modal adjunct).
148
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 4.6 System of MODALITY TYPE (from Matthiessen 1995) Modality can be realized congruently (non-rnetaphorically) by modal elements, such as modal adjuncts and/or modal verbs, or non-congruently (metaphorically) by projecting clauses in hypotactic clause complexes. This distinction — congruent versus non-congruent — is not included explicitly in the network, but congruent expressions of modality indicate implicitly that a probability or obligation is subjective, while metaphorical expressions indicate explicitly that a probability or obligation is subjective or objective, hence the final set of options under modality type, [explicit] versus [implicit]. Finally, there is the simultaneous system of POLARITY, i.e. all of the above options can combine with the clause being [positive] or [negative]. Tables 4.6 and 4.7 give a number of examples of modality along with their systemic features; Table 4.6 gives examples of congruent realizations, and Table 4.7 gives non-congruent examples. As mentioned above and discussed in more detail in Chapter 2, the choice of indicative or subjunctive mode in the examples in Table 4.7 depends on 'the value that is set on the modal judgement' (low, median or high) and the type of the modality of the projecting clause or modal adjunct (see Halliday 1985a; Matthiessen 1995). This is one of the prosodic modes of expressing modality across a clause complex or an attributive clause where the attribute is modal. The prosodic mode of expression of modality was also illustrated in section 4.3.2 with Example 23, which is reproduced here as Example 42, where the meaning of probability is expressed three times: in the Finite as -ait, in the modal verb, pouvoir, and in the modal Adjunct peut-etre.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
149
Table 4.6 Congruent realizations of modality: features and examples examples
systemic features
Sa voiture est probablement cassee His car is probably broken Sa voiture doit etre cassee His car must be broken Sa voiture doit siirement etre cassee His car must surely be broken Tu dois reparer ma voiture You must repair my car Peux-tu reparer ma voiture? Can you fix my car?
probability & median & implicit (adjunctival modalization) probability & high & implicit (verbal modalization) probability & high & implicit (adjunctival & verbal modalization) obligation & high & implicit (verbal modulation) readiness & ability & implicit (verbal modulation)
(42) Jean pourrait peut-etre faire fa Jean could maybe do this
Following this brief overview of modality in French, we will now turn to the function of Subject, which in systemic functional terms carries 'modal responsibility'. First we will explore criteria for identifying the Subject in French and then criteria for denning the Subject as the focus of 'modal responsibility'. 4.4.3 French Subjecthood: recognition criteria
In English the specific character of the Mood-tag and its cohesive relation to Subject act as a reliable means of recognizing the segment with the function of Subject, as in you are early, aren't you1? However, French tags are quite unlike the English ones; here no nominal is picked up as it is in English. How, then, can we recognize the Subject of the French clause? Although tagging cannot be used as a criterion to establish Subjecthood of a nominal in French, there are some resemblances between English and French; for example, a polar interrogative may be realized by ordering Finite before Subject. Thus one may probe the validity of the statement elk est malade (she is sick) by the confirmation question I'est-elle? (is she?). The Subject is always the clitic suffixed to the Finite in a polar interrogative of this type, the Subject being the element 'by reference to which the proposition can be affirmed or denied' (Halliday 1994: 76). Another criterion for recognizing the Subject is person and number agreement with the verbal group. Thus the Subject is the function which is marked in person and number in the Finite or Predicator depending on whether the MOOD is indicative or imperative, as the comparison between Example 43 (indicative) and Example 44 (imperative) shows.
150
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 4.7 Non-congruent realizations of modality: features and examples examples
systemic features
personal (mental process) clauses realizing modalization Je crois\\ que ma voiture est cassee I think that my car is broken
probability & median & subjective & explicit P clause: indicative
Je suis certain [[que ma voiture est cassee]] I am sure that my car is broken
probability & high & subjective & explicit indicative
Je ne suis pas sure [[que ma voiture soit cassee]] I am not sure that my car is broken
probability & low & subjective & explicit subjunctive
impersonal (relational process) clauses realizing modalization II est probable [[que ma voiture est cassee]] It is probable that my car is broken
probability & median & objective & explicit indicative
// est possible [[que ma voiture soit cassee]] It is possible that my car is broken
probability & low & objective & explicit subjunctive
personal (mental process) clauses realizing modulation Je veux 1 1 que ma voiture soit cassee I want my car to be broken
obligation & high & subjective & explicit P subjunctive
II espere \ \ que ma voiture est cassee He hopes that my car is broken
obligation & low & subjective & explicit (3 indicative
(43)
Les cloches sonnent the bells-3-PL ring-3-PL Sub Fin/Pred The bells are ringing
(44)
Sonnez ring-2-SING-FORM Pred Ring the bells!
les cloches the bells-3-PL Comp
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
151
Here Example 43 is a declarative with les cloches as its Subject; the Fin/Pred agrees with it in person and number. In Example 44, which is an imperative, les cloches is Complement, but since the marking on the Predicator in an imperative relates to the Subject, the person and number marked on the Predicator signal the addressee's person and number. When the verbal group realizing the Finite and Predicator of a Negotiator has a secondary tense choice with the auxiliary etre, the agreement spreads from the Finite to the Predicator. Thus, the realization of the Subject is itself prosodic. It can be repeated throughout the Negotiator as in Example 45. (45) Lesfourmis sont the ants-3-PL are-3-PL Subj Fin The ants have gone
parties gone-PL-FEM Pred
Note that in addition to person and number, the Predicator also marks the gender of the Subject. This adds further support to the analysis of the Predicator as a Function of the Negotiator rather than of the Remainder. While in French intonation is a primary realizational resource for MOOD options, the Subject together with the Finite and the Predicator are at least equally crucial to its realization, since it is the fall or rise of the tone on the Predicator which is criterial to MOOD choice recognition. Note also that the Predicator is the last function of the Negotiator. These are then fairly substantial recognition criteria for the function Subject in French. But what does the function Subject 'do' in French? This is the question we address below. 4.4.4 French Subjecthood: definition criteria
As pointed out before, it is the Subject of a proposal which is responsible for complying with or rejecting a command, whereas a proposition is affirmed or denied in relation to the Subject. Thus, in both types of speech functions, Subject has what Halliday (1985a) has called 'modal responsibility'. In both French and English, this interpretation is reinforced by several features of the Negotiator and the Mood, respectively. For example, the location of the Subject outside the negative prosody (see discussion of Example 11 in this chapter) is significant. The modal responsibility of the Subject is also made manifest in modulated indicative clauses, where the realization of the Subject spreads prosodically over the modal process through the devices of person and number marking. This clearly highlights that what is at issue is the inclination, willingness, ability, etc. of the Subject, not of any other nominal in the clause, as in Example 46. (46) Nous devons we-i-PL must-1-PL Subj Fin: mod We must leave!
partir leave Pred
152
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Here it is the Subject that is held responsible for 'going', and is under the obligation to leave. This is made explicit through the markings on the modal verb devoir. The main modal verbs of French are devoir, vouloir, pouvoir and falloir. They may take on different modal meanings depending on context. permission (47) Vous you-2-siNG Subj You may leave
pouvez can-2-SING-FORM Fin: mod
obligation (48) // faut it-g-siNG must-g-siNG Subj Fin: mod 'we' have to leave
partir to leave Pred
partir to leave Pred
probability (49) Elk pent She-3-siNG may-3-siNG Subj Fin: mod She may be thirty
avoir to have Pred
ability (50) Tu You-2-siNG Subj You can do it
le it C-clitic
peux can-2-siNG Fin: mod
trente am thirty years Comp
faire do Pred
Thus, although French Subject and English Subject cannot be identified following the same grammatical criteria, they are clearly similar on semantic grounds and perform the same function in discourse. This reflects again the assumption that we are more likely to find congruence across different languages by approaching their linguistic system from discourse semantics rather than in terms of syntagmatic structure. This concludes the exploration of the clause as a move in dialogue. We will now move down a rank below the clause to explore the nominal group from an interpersonal viewpoint, i.e. to explore the nominal group as 'person'. [The nominal group] is part of the (dialogic) exchange between speaker and addressee, where commodities are given and demanded; and it enacts the interpersonal universe with the interactants as part of the dialogic centre. (Matthiessen 1995:687)
In the following section we are thus concerned with the pronominal nominal group as interactant or non-interactant and its relationship with the interpersonal structure of the clause. We will see that the positioning of the pronominal within or outside the Negotiator of the clause is significant both interpersonally and textually.
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
153
4.5 The nominal group as an interactant or non-interactant in dialogue The nominal group, like the clause, can be approached from all three metafunctions. Here, it is from an interpersonal viewpoint that we examine the nominal group, and more specifically, the pronominal nominal group. Pronominals are discussed in terms of the interpersonal structure of the clause discussed in the preceding sections. Interpersonally, pronominals may be an interactant — i.e. speaker or addressee — or non-interactant. They may be presented as being central to the negotiation (i.e. within the Negotiator) or peripheral to the negotiation (outside the Negotiator). 4.5.1 Pronominals and the Negotiator—Remainder structure
We have seen that of the MOOD functions identified as 'capable' of occurring in the Negotiator, Complements and two types of Adjunct have a special status, in that they may function either within the Negotiator or in the Remainder. When Complements or Adjuncts form part of the Negotiator, they are cliticized; and to say that they are cliticized is to say that they are textually Given, identifiable and non-prominent. When these functions form part of the Remainder, they are New, identifiable or non-identifiable and prominent. In this section, we will examine those Complements and Adjuncts which are realized as pronominal nominal groups and therefore always identifiable from context or cotext. Their status as Given or New is interpretable both from their location and function in the clause as well as by their form. Indeed, there are two types of pronominals in French, clitic pronominals and non-clitic pronominals. The latter are either functions of the Remainder, absolute Theme or Vocative, and thus outside the interpersonal structure (see Section 4.3). Uses of these two types of pronominals are illustrated below with three examples (51 to 53) taken from Giono's play Le Cheval fou (1974). Prominent pronominals are in bold, and nonprominent pronominals are underlined. In some contexts, pronominals may be both prominent and cliticized; those are in bold and underlined. Each pronominal item is then described in terms of features from the network in Figure 4.8. (51 i) Antonio: IIy a beaucoup de choses queje n 'aime pas et queje fais, moi. There are many things that I don't like and that I do. (51ii) Tatoue: Iln'en a pas I'habitude, lui. He is not accustomed to it. (51iii) Antonio: // laprendra. He will become. (Giono 1974: Tableau I, p. 12)
In Example 51i, je is [interactant: speaker: non-prominent: nominative], while moiis [interactant: prominent]. In Example 51ii, His [non-interactant: non-prominent: nominative], and lui is [non-interactant: prominent]; en, in
154
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
this particular example, presupposes a downranked clause functioning as a nominal group such as de faire des choses qu 'il n 'aime pas in il n 'a pas I'habitude [defaire des choses qu'il n'aimepas]. In the network in Figure 4.8 this particular pronominal is described in terms of the features [non-interactant: genitive]. (52) Jeune Bouvier: Tatoue, £j comprends quelque-chose, toi? Tatoue, do you understand something of it? (Giono 1974: Tableau I, p. 14)
In Example 52, t' is a variant of tu and carries the features [addressee: non-prominent: nominative: informal], toi is [addressee: prominent] and y is [non-interactant: dative: non-human]. (53) Toussaint: Assieds-toi, Matelot, surle tabouret Sit down Matelot, on the stall (Giono 1974: Tableau II, p. 21)
In imperative clauses such as Example 53, the addressee is always prominent. The speaker and the non-interactant are also prominent when they are the recipients of goods-and-services as in donne-moi lapommeor donne-lui lapomme. Thus, in such contexts, pronominals can be both clitics and prominent. Such pronominals are stressed as they indicate either the one who complies with the command (addressee) or the one who benefits from the command (speaker or non-interactant); in either case the role realized from the pronominal is the focus of the proposal and is presented as New. They always fall at the end of the Negotiator, which corresponds to one information unit realized as one tone group (see Figure 4.9). Further examples of the use of cliticized pronominals from Giono (1974) are given in Examples 54 to 57. (54i) Toussaint: J^ vais I am going there
(54ii) Gina: N'y_ vapas . . . ou vas-y, et tue-k! Don't go there . . . or go there and kill him (Giono 1974: Tableau II, p. 27) (55) Antonio: c'est lui It's him (Giono 1974: Tableau IV, p. 37) (56) Antonio: Alors viens avec moi Then come with me (Giono 1974: Tableau IV, p. 38) (57) Gina: Et elle I'aime, lui And she likes him (Giono 1974: Tableau IV, p. 39)
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
155
Thus, typically, clitics are non-prominent, non-thematic, given and identifiable. The non-prominence of clitics reflected the transitional positioning of the Negotiator at the crossroads of the two textual waves that realize the thematic structure and the structure of the information unit respectively. This is shown in Figure 4.7. The two textually distinct pronominals are represented in the system network in Figure 4.8 by the opposition between prominent and nonprominent. The realizations of the Thing in the structure of the pronominal nominal group are listed in Table 4.8, along with their respective systemic features. Non-prominent pronominals which function as participants within the transitivity structure of the clause carry some vestige of case marking. This is formalized in the opposition between nominative and oblique pronouns,
Figure 4.7 The Negotiator: at the transition of the textual waves
Figure 4.8 Interactant-type potential
156
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 4.8 Realizations of pronominal nominal groups and their features from the system pronom.
systemic features
moi [interactant: speaker: singular: prominent] toi [interactant: addressee: singular: informal: prominent] nous [interactant: speaker-plus: plural: prominent or non-prominent: nominative or oblique] vous [interactant: addressee: singular/formal or plural: prominent or nonprominent: nominative or oblique] lui [non-interactant: prominent: singular/masculine] or [non-prominent: singular/masculine or feminine: dative: human reference] die [non-interactant: prominent: singular/feminine] or [non-prominent: nominative] eux [non-interactant: prominent: plural/masculine] elles [non-interactant: prominent: plural/feminine] or [non-prominent: nominative] je [interactant: speaker: singular: non-prominent: personal: diversified: nominative] tu [interactant: addressee: singular: non-prominent: nominative] on [interactant: speaker-plus: non-prominent: non-specific: nominative] or [non-interactant: generalized] il [non-interactant: non-prominent: personal: singular/masculine: nominative or impersonal] ils [non-interactant: non-prominent: plural/masculine: nominative] se [non-interactant: non-prominent: reflexive] me [interactant: speaker: non-prominent: oblique] te [interactant: addressee: non-prominent: oblique] le [non-interactant: non-prominent: singular/masculine: diversified: nonreflexive: accusative] or [generalized] la [non-interactant: non-prominent: singular: feminine: diversified: nonreflexive: accusative] les [non-interactant: non-prominent: plural: accusative] leur [non-interactant: non-prominent: plural: dative] y [non-interactant: non-prominent: dative: non-human reference] en [non-interactant: non-prominent: genitive]
which function as Subject and Complement respectively in the interpersonal structure of the clause. It could be argued that prominent pronouns display no case marking because they are 'absolute' as Theme and/or New and are coreferential with non-prominent forms that are marked for case, as in moi, ne saispas (I, I don't know). The system network in Figure 4.8 shows that only non-interactant pronominals differentiate between accusative and dative forms. This can be explained by the fact that in the transitivity structure of the clause, and in
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
157
particular in material clauses, there may only be one oblique interactant, typically a Recipient, as in, il [Actor] me [Recipient] donne unepomme [Goal] and sometimes a Goal as in // m'a battu. Therefore there is no need to differentiate between two oblique functions, as clauses such as *il me donne are not in use. The same kind of argumentation applies to the lack of distinct reflexive forms for interactant oblique pronouns, as co-reference with the Subject is the only possible interpretation in examples such asje me lave and tu te laves. However, there is a need for two oblique forms for non-interactants in order to distinguish participant functions in the following example: II [Actor] le [Goal] lui [Recipient] donne. In order to express reflexivity with a non-interactant Subject a specific reflexive form is needed. Thus, il se lave can not be interpreted as il k lave or il lui lave (la main). Oblique clitics are recapitulated in Table 4.9. The dative pronoun y which serves as a third person for referring to nonhuman things, as in il m'a raconte une histoire, mats je n'y ai pas fait attention (He told me a story, but I paid no attention to it) may also function as an Adjunct. The pronoun en is another such pronoun that can serve either as an adverbial or as a non-interactant accusative pronoun referring to nonhuman things modified by a partitive deictic, as shown in the following examples taken from Byrne and Churchill (1986). (58) Avez-vous du pain ? - Oui, j'enai achete. Do you have any bread? -Yes, I have bought some. (59)
Voulez-vous de la biere? - Oui, s'ily en a. Do you want some beer? -Yes, if there is any.
Table 4.9 Oblique pronominals interactant
non-interactant
oblique
me (me) te (you) nous (us) vous (you)
Reflexive: se (himself, herself, itself) Genitive: en (of it)
accusative
le (him, it) la (her, it) les (them)
dative
lui (to him, to her, to it - if 'it' is endowed with personal existence) y (to it, to them, with non-human reference) leur (to them)
158
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(60)
Si vous voulez des billets, je peux vous en donner. If you want [some] tickets, I can give you some.
(61)
Comment! Ils n 'ont pas d'argent! Donnez-leur-en; moi, je n 'en ai pas. What! They have no money! Give them some; I haven't got any.
The above examples illustrate again the important role of the textual resource of REFERENCE in French dialogic text, which is comparable to the role of ELLIPSIS in English dialogues. The use of y and en as both pronominals and adverbials can be explained by the grammatical and semantic similarity of their respective referent, as participant or circumstance. Grammatically, the referents are in both cases realized typically by prepositional phrases of which the prepositions are respectively a and de. Semantically, it can be argued that both dative (giving to) and locative (going to) meanings imply a transfer, either to someone or some place, and that being a part of something implies a meaning of belonging, which is closely related to the meaning of provenance expressed in coining from some place. Prominent nominal groups can be seen as having the function of demarking tone groups and clause boundaries (see Chapter 5). Prominent pronominals are stressed and either precede or follow the main tone group, which is unmarked. A tone group is realized as a syllabic chain. Typically the main tone group consists of the verbal group plus nonprominent pronominal. A new tone group is marked by either a prominent pronominal, a change in the pitch movement or a pause. The demarcations by prominent pronominals of the main tone group in French are shown in Figure 4.9. A tone group in Halliday's description of English (see Halliday, 1985a) is not only a phonological constituent but also the realization of the information unit, which in French dialogic texts often corresponds to the Negotiatory structure. 4.5.2 Ordering of French clitic pronouns The unmarked position of nominative pronouns is clause-initial in relation to other clitics. However, as we saw earlier, they may be preceded by a stressed
Figure 4.9 Potential marked-unmarked boundaries of the French clause
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
159
(prominent) pronominal, which functions as Theme. Thus, either there may be two nominal groups - absolute Theme: moi (absolute in the sense that it is purely textual) and Subject: je - or one nominal group: je, which is both Subject and unmarked Theme (see Halliday 1985a). In a clause type which displays a Subject followed by Finite ordering, as interrogatives do in some registers, the nominative clitics are placed after the verbal group, while other clitics stay in the front-end of the verbal group, as in te le rappelles-tu ? Nominative pronominals which are potentially topical Themes may precede or follow non-Subject pronouns. Their positioning in relation to other clitics is thus both interpersonally and textually motivated. The relative ordering of the non-Subject clitics in non-positive imperative clauses is as follows: the reflexive pronominal always follows the pronominal Subject with which it is co-referential. Note that, unlike in English, reflexivization is realized in the Negotiator, whereas in English it is realized in the Residue, as shown in Example 62. (62) Use lave He washes himself
Next to clitics used reflexively follow the pronominals which fill interactant roles and then those which are non-interactant. Besides the general textual motivation behind the relative ordering of clitics, at a local level the interpersonal resource of INTERACTANT TYPE is thus the primary driving force behind the relative ordering of oblique pronouns. The second driving force is experiential and relates to the degree of involvement of the participants in the process. That is, nuclear participants precede peripheral participants which precede circumstances. The ordering of clitic pronouns in indicative clauses is thus as follows: a topical Theme which conflates with the function Subject comes first, followed by interactant non-Subject pronouns, followed by non-interactant accusative pronouns, followed by non-interactant dative pronouns. The following examples display the ordering principles outlined here, and Figure 4.10 summarizes the tri-metafunctional organizing principle behind the relative ordering of French personal pronouns. Abbreviations used here are: non-int = non-interactive, Benef = Beneficiary, periph = peripheral.
(63)
Moi, me
je I
1' it
abs Th: non-clitic
Subj/Ac: clitic
Comp/Goal: clitic: non-int; nuclear
I, I have done it for him
ai fait
pour lui
have done
for him
Fin A Pred/Proc
Adj /Benef: non-clitic; periph
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
160
Figure 4.10 A metafunctional perspective on the ordering of French personal pronouns (64)
Je I
me myself
ks them
suis lave am washed
Subj/Ac: clitic
reflexive
Comp/ Goal: clitic: non-int; nuclear
Fin A Pred/Proc
I washed them (e.g. my hands) (65)
Elk she
te to you
ks them
y there
a emmenes has taken
Subj/Ac: clitic
Comp: interactant: clitic
Comp: non-int: clitic
Adj/Circ: clitic
Firi A Pred/Proc
She brought them there to you (66)
Elk she
k it
lui to him
a donne has given
Subj/Ac
Comp: nonint; nuclear
Comp: non-int; periph
Fin
A
Pred/Pi~oc
She gave it to him
The circle in Figure 4.10 surrounds the participants that are most likely, because of their non-prominent textual status, to construct major reference chains in dialogic text. Circumstances are typically part of the periphery and textually prominent. Also, the clitics that realize circumstances of location or non-human non-interactant participants, i.e. y and en, never enter into major chains of reference. On the other hand, participants that function as Subject
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
161
and Complement in the Negotiatory structure of the clause tend to occur in the longest reference chains. Thus, typically, it is participants which are textually non-prominent, given and identifiable that construct the major cohesive chains of reference, whilst textually prominent participants and circumstances occur in minor chains. Matthiessen (1995: 331) points out that 'participants have more staying power in discourse' than circumstances. Thus, we can predict that reference chains will run mostly through participants which are realized as clitics, both as a consequence of their textual status and as a result of the salient role of the Negotiator in dialogic discourse. These patterns of harmony between reference chains and the textual statuses of participants are illustrated in Texts 4.2 and 4.3, taken from Beckett's En attendant Godot (Waiting for Godot). The extract in Text 4.2 could be labelled a macroproposal. As such the moves are mostly realized by the imperative MOOD of which the Negotiator consists of Predicator A Complement 1 (clitic) A Complement 2 (clitic). Thus, predictably, major reference chains will run through the complement clitics indicating non-prominence and continuous reference. Text 4.2 From En attendant Godot, Act I (Beckett 1952) V (a Pozzo) Dites-lui de penser. P Donnez-lui son chapeau. V Son chapeau? P D ne peut pas penser sans chapeau. V (a Estragon) Donne-lui son chapeau. E Moil Apres le coup qu'il ma.fait! Jamais! V Je vais le lui donner moi. (II ne bouge pas.) E Qu'il aille le chercher. P II vaut mieux le lui donner. V Je vais le lui donner. (II ramasse le chapeau et le tend a Lucky a bout de bras. Lucky ne bouge pas.) P II faut le lui mettre. V Je vais le lui mettre. (II contourne Lucky avec precaution, s'en approche doucement par derriere, lui met le chapeau sur la tete et recule vivement. Lucky ne bouge pas. Silence.) Key: V = Vladimir; P = Pozzo; E = Estragon; bold = references to Lucky; bold + underline = reference to 'le chapeau'; underline = reference to Vladimir; italic = reference to Estragon; text in parentheses = stage directions
The major chains in Text 4.2 are those referring to (i) Lucky, who in experiential terms is the Beneficiary of the Action, (ii) le chapeau, which is the Goal of the Action and (iii) Vladimir, the Actor of the Action. In interpersonal terms, Vladimir is the Subject, the participant who wants to be responsible for the realization of the command: je vais le lui donner moi. In contrast, Estragon refuses to be modally responsible. His refusal of subjecthood is clearly marked in the discourse, by the prominent pronominal,
162 A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
moi, in initial position, and the negative Adjunct, jamais, in final position (moi! Apres le coup qu'il m'a fait!Jamaisf). Two major reference chains in Text 4.2 identify non-interactant participants, which is a consequence of the predominant mood of the text, which is the imperative. The two most central participants of the macro-action that is requested are the Beneficiary of the action (Lucky) and the Medium (the hat). The interactant participant which forms the third major reference chain is the one who is going to carry out the Action. The nominal groups realized as proper names or common names break the reference chains and indicate the occurrence of narrative passages, which are intertwined with the narrative text. In the extract presented here as Text 4.3, also from the first act of En attendant Godot, most moves are realized by indicative clauses, and it can be predicted that there will be mostly reference chains of Subject participants. Text 4.3 From En attendant Godot, Act I (Beckett 1952) G
Monsieur Godot. . .
V (1'interrompant) Je t'ai deja vu, n'est-ce pas? G Je ne sais pas, Monsieur. V Tu ne me connais pas? G Non Monsieur. V Tu n'es pas venu hier? G Non Monsieur. V C'est la premiere fois que tu viens? G Oui Monsieur. (Silence.) V On dit ca. (Un temps.) Eh bien, continue. G (d'un trait) Monsieur Godot m'a dit de vous dire qu'z7ne viendra pas ce soir mais surement demain. V C'est tout? G Oui Monsieur. V C'est tout? G Oui Monsieur. V Tu travailles pour Monsieur Godot? G Oui Monsieur. V Qu'est-ce que tu fais? G Je garde les chevres, Monsieur. V // est gentil avec toi? G Oui Monsieur. V // ne te bat pas? G Non Monsieur, pas moi. V Qui est-ce qu'«7bat? G // bat mon frere, Monsieur. V Ah! tu as un frere? G Oui Monsieur. V Qu'est-ce qu'i/fait? G fl garde les brebis, Monsieur. V Et pourquoi il ne te bat pas, toi?
THE GRAMMAR OF NEGOTIATION: INTERPERSONAL METAFUNCTION
G V G
163
Je ne sais pas, Monsieur. // doit t'aimer. Je ne sais pas, Monsieur.
Key: V = Vladimir; G = Garcon; bold = references to Garcon; italics = references to Godot; underline = references to Vladimir; bold italic = references to Frere; text in parentheses = stage directions.
References to 'Garcon' and 'Godot' form the longest string in this text, which as opposed to the preceding text could be categorized as a microproposition. The 'Garcon' is mostly interactant Subject throughout, while Godot is non-interactant Subject. In generalized experiential terms, Gargon functions as Medium (most nuclear participant; see Chapter 3), and Godot as Agent (second most nuclear participant). Thus, typically, the length of reference chains and the function of participants involved correlates with the metafunctional ordering of clitics schematized in Figure 4.10. The textual status of the pronouns, whether prominent or non-prominent, is the primary determining factor. Typically, marked pronouns never form long chains on their own but reinforce the clitics. Participants for which the transitivity function conflates with the interpersonal function of Subject are likely to be part of long reference chains in text where the major mood function is informative rather than imperative. Then non-Subject pronouns which refer to human participants are more likely to enter into long chains than those referring to non-human participants, as in Text 4.2 where the 'Lucky' chain is longer than the 'chapeau' chain. This is because people involved in a conversation are more likely to stay constant throughout the dialogue, while things that are being talked about are more likely to change. In the preceding sections, we focused on the textual statuses of pronominals in relation to the interpersonal structure of the clause. We found that while the Negotiator is interpersonally salient, it is textually non-prominent. 4.6 Conclusion This chapter has attempted to examine how the French language constructs dialogue - i.e. how it enables the exchange of meanings. At the outset we drew attention to certain theoretical assumptions. These initial assumptions have been confirmed by the subsequent analysis. Thus, ignoring the details of actual structures, and focusing on more abstract aspects, we find that the semantics of exchange is realized by MOOD systems which have a prosodic mode of expression both in English and in French. The primary options of the French MOOD potential reflect the primary speech functions of statement, question and command assumed to be common to all languages (see Figure 4.2). From a language-specific perspective on structure, the modal structure of French is considerably different from that of English. These structural differences result from both the different means of enacting an exchange and the different means of realizing MOOD selection. We saw that French
164
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
makes use of a wider range of prosodic means of realization, i.e. tone, mood particle and the order of the Subject in relation to the Finite or Finite/ Predicator. Other interpersonal resources, such as negation, modality and Subject, were also found to be realized by prosodic means within the Negotiator of the clause. Nonetheless the French Negotiator-Remainder structure is functionally analogous to the Mood-Residue structure of English. Just as for purposes of negotiating an exchange in English, it is Mood that is crucial, so for negotiating French exchanges it is the Negotiator that is the most crucial clausal component. The French clause as a move was found to have a particular textual organization, making use of tonic pronominals for staging and resolving an exchange, thus creating two textual layers: one unmarked within the interpersonal structure, the other marked, outside this structure (see Figure 4.9). While prominent pronominals serve both to mark contrastive emphasis and to reinforce the local Theme, as well as to mark the boundaries of the interpersonal structure, non-prominent pronominals carry the exchange forward and have a tracking function. This discussion on the interaction of the textual and the interpersonal metafunctions in constructing dialogue leads us to the final chapter of this book, which is concerned with the textual metafunction, the 'enabling' function that provide the resources for presenting experiential and interpersonal meanings as a coherent and cohesive message in context. Notes 1 Following Martin 1992, a move is the rank below that of the exchange rank within the discourse semantics stratum. The SPEECH FUNCTION system is located at move rank. A move is realized in the grammar by a clause selecting independently for MOOD. 2 What Rothemberg calls 'Rheme' is very much like Halliday's notion of 'New' in the sense that it is listener-oriented rather than speaker-oriented (see Halliday 1967, 1968, 1985a).
5 The 'enabling' grammar: textual metafunction
Thematic choice is meaningful. . . . The theme-rheme organization of the sentence forms part of a larger pattern which governs the flow of information in any . . . discourse. (Fries 1981: 144)
5.1 Introduction In the preceding chapters, we explored logical, experiential and interpersonal resources at both clause and group ranks, with the greater part of the discussion focusing on clausal organization. Chapters 3 and 4 gave an account of two types of clause organization, experiential and interpersonal. While the former serves to construe our experience of both the external world and our internal world, the latter contributes to the enactment of social roles in an exchange. To complete the tour of the French clause as a multifunctional unit in which three different types of structures are mapped onto one another (see Chapter 1), we now turn to the resources that organize the clause as a message, the textual systems of THEME. In approaching the French clause from a textual viewpoint, we assume that the French clause as a piece of text (or message) is not organized randomly: its organization serves to guide the speaker/listener through the unfolding text so that s/he can process the information constructed by the interpersonal and ideational meanings. In Halliday's (1978: 112-13) words: The textual component represents the speaker's text-forming potential; it is that which makes language relevant. This is the component which provides the texture; that which makes the difference between language that is suspended in vacua and language that is operational in a context of situation. It expresses the relation of the language to its environment, including both the verbal environment - what has been said or written before - and the non-verbal, situational environment. Hence the textual component has an enabling function with respect to the other two; it is only in combination with textual meanings that ideational and interpersonal meanings are actualized.
The 'enabling' metafunction provides the resources for assigning different textual statuses to clausal elements. The system of THEME in conjunction
166
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
with the INFORMATION FOCUS system are two of the systems that are concerned with the assignment of textual statuses: thematic prominence and newsworthiness respectively. From a discourse perspective, thematic information relates to the method of development, that is, the mode of expansion of the text, while the information that is new typically foregrounds the 'main point' (Fries 1981). As the INFORMATION FOCUS system is a phonological system realized through intonation, and as the scope of this book does not provide space for an exploration of French intonation patterns, we will only try to capture some of the generalizations regarding the typical information structure of spoken texts and the unmarked reading of the New in written texts. As Halliday (1978: 133) points out: Since it is realized through intonation, which is not shown in the writing system, the information structure is a feature of the spoken language only; and any interpretation of the information structure of a written text depends on the 'implication of utterance' which is a feature of written language. There are two aspects to this: (i) the interpretation of the paragraphological signals that the written language employs, such as punctuation, underlining and other forms of emphasis; (ii) the assumption of the 'good reason' principle, namely that the mapping of the information structure onto other structures will take the unmarked form except where there is good reason for it to do otherwise . . .
In French as in English, new information typically falls at the end of the information unit, of which the grammatical correlate is, in the unmarked case, the clause. Thus, the New often maps onto the Rheme of the thematic structure and this can be used as the default case for analysing new information in written texts. However, when the New and Rheme are separated, the grammar of French gives us other means of identifying the New. We have already seen in Chapter 4 that in spoken discourse absolute Themes often map onto New elements, and that prominent pronominals, whether thematic or rhematic, are typically New. These grammatical clues provide a means for interpreting New elements in texts channelled graphically via the spoken or written medium. Although there appears to be a correlation between types of Theme and the mapping of Theme onto Given or New, it is important to stress here that 'the choice of whether to treat some item as Theme is basically independent of the choice of whether to treat that item as given or new information' (Fries 1981: 136). This separating approach to thematic and information structures contrasts with the combining approach of French functionalists, such as that ofCombettes (1975) and Hagege (1982). Following systemic theory's assumption regarding the relationship between different types of meaning and different types of structure (see Chapter 1, Section 1.2.3), we can predict that textual structure in French will have a periodic (wave-like) rather than a prosodic or segmental type of organization. Indeed the intersection of the information and thematic structures forms peaks of prominence and non-prominence which create wave-like patterns of various shape, as shown in Figure 5.1.
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
167
Figure 5.1 Wave-like representations of thematic and information structures The wave-like mode of expression of textual structures poses representational problems, just as the prosodic mode of expression of the negotiatory structure does. To simplify the mapping of the three types of structure - interpersonal, experiential and textual - onto one another, both the thematic and information structures will be modelled as constituency structures as in Halliday (1985a/94), as shown in Example 1. (1)
suis en retard am late
Mot, me
> I
Theme: marked
Theme: unmarked
Rheme
New
Given
New
The Given-New structure is realized through intonation and consequently is carried prosodically. As a result, the boundary between Given and New is not as discrete as that between Theme and Rheme. As Matthiessen (1992: 52) points out: In the case of NEWS PROMINENCE, it is, quite predictably, comparatively harder to draw the boundary between Given and New, since the carrier of the wave is not constituency but prosody. That is, the carrier does not itself provide a constituency structure and the boundary between Given and New is more indeterminate
168
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(Halliday 1985a: 275). Following Halliday, this indeterminacy is indicated by means of the left-facing arrow pointing from New. . . . (Identifying the end of the New is not a problem since it coincides with the end of the intonational or tonic prominence.)
We will come back to the interweaving of thematic structure and information structures throughout this chapter. In Section 5.2, we define the SF descriptive category of Theme further and explore the French THEME systems in detail. In Section 5.3, we interpret Theme from a discourse perspective: the analysis of thematic patterns in three texts (a geographic description, a short literary biography and a news article) reveals the semantics behind thematic distribution in terms of method of development and flow of information. 5.2 Theme in SF theory From the point of view of SF theory, it is assumed that any clause in any language realizes textual meaning, i.e. is organized as a message. The realizations of textual organization vary across languages (see Caffarel et al. 2004): in French as in English, the realization of Theme is positional rather than case-marked, as it is in Japanese, for example (Teruya 2004). Theme has been defined by Halliday (1985a) as the 'point of departure'1 or 'starting point' of the message. In French, some themes (referred to as absolute Themes in previous chapters) may occur both clause initially and clause finally. In languages where there are no overt Theme markers, as in French and English, delimiting the Theme can be an issue. According to Halliday (1994:52-3): .. . the Theme of a clause ends with the first constituent that is either participant, circumstance, or process. Since a participant in thematic function corresponds fairly closely to what is called the 'topic' in a topic-comment analysis, we refer to the experiential element in the Theme as the TOPICAL THEME. . . . . . . The Theme always includes one, and only one, experiential element. This may be preceded by elements which are textual and/or interpersonal in function; if so, these are also part of the Theme. The typical ordering is textual A interpersonal A experiential; in any case, the experiential element (the topical Theme) comes last - anything following this is part of the Rheme.
In this interpretation of French Theme, we suggest that a clause may be analysed as 'having' two layers of thematic organization, which involves the selection of a marked and an unmarked (topical) Theme. We will argue that when a marked Theme is selected, the participant functioning as Subject still has thematic prominence, although to a lesser degree than the marked Theme. The marked and unmarked thematic organization of a sentence can contribute to the creation of two complementary methods of development in a particular text instance. The hypothesis is that a text may either have an unmarked thematic organization which foregrounds the topic as method of development, or have both a marked (most prominent) and unmarked
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
169
thematic organization. The interpretation of thematic structure as a doublelayered structure will be first illustrated at clause rank in Section 5.2.2 and then argued from discourse in Section 5.3. We now explore Theme in more detail, and illustrate its textual function through the analysis of a text. 5.2.1 The function of THEME Theme is most commonly defined as expressing what the clause is about (see Firbas 1987; Hagege 1978). This is because Theme is often equated with Topic. However, as Theme within the systemic framework can include not only topical elements but also textual and experiential ones, a broader definition is required. As Halliday (1994: 38) puts it: ' [Theme] is the ground from which the clause is taking off. In Matthiessen's (1995: 531) words: THEME . . . is the resource for setting up the local context or local semiotic environment in which each clause is to be interpreted. This local context constitutes a particular status in the clause-thematic prominence. While all clauses occur in context, THEME is thus the resource for manipulating the contextualization of the clause . . .
The term 'manipulating' describes precisely the process in which we are involved when choosing particular element(s) as 'local context' of a clause. Consider the following text, where each clause has been divided into Theme and Rheme. The selection or manipulation of thematic elements is not arbitrary but conveys the method of development of the text. What has been selected as local context to orient the reader to the upcoming message is first a temporal location and then participants, 'us' (animals) versus 'them' (men). In other words, following the contextualization of the message in time, the text develops in terms of contrast between 'animal' and 'man'. The notion of contrast is not only realized by the opposition of (pro)nominals referring to animal and man but also by the absolute Theme,2 nous, les animauxin clause 2, and by the textual Theme, the adversative relation, mais, in clause 3a. Each Theme orients the reader to important points (in the Rheme) moving from 'the tranquil life of animals prior to man's arrival', to 'man's destruction of the environment', to 'the death of certain animals'. The gradually more shocking information 'packaged' in the Rheme, which here also maps onto the New, is there to get the reader's attention and give a background to what is to come, man's realization of his mistakes and the consequent creation of a National Park. Text 5.1 From Le Pare National des Pyrenees Raconte aux Enfants (Birmant 1991) Theme Rheme 1 Pendant des milliers et des milliers les Pyrenees ont ete des montagnes tranquilles. d 'annees, the Pyrenees were tranquil mountains. During thousands and thousands of years,
170
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
5b
Theme Nous, les animaux, Us, the animals, Mais les hommes But men et nous and we qui which Nous We Les hommes The men []
5c
[]
5d
[]
5e
et[] and Certains d'entre nous Some of us
2 3a 3b 3c
4 5a
6a
6b
7 8 9
zis they Unjour, les hommes, One day the men Alors, ils So, they Enfin, ils Finally, they
Rheme nous y vivions en paix. we lived there in peace. sont venus de plus en plus pres, came closer and closer, avons vu arriver d'autres animaux saw arrive other animals lesfuyaient. were fleeing them (men). nous sommes caches. hid. faisaient du bruit, made noise, reveillaient nos petits, woke up our little ones, nous poursuivaient avec leurfusils, chased us with their guns, troublaient I'eau des torrents, made the rivers' water dirty detruisaient nos arbres. destroyed our trees. n 'ontpas supporte de vivre dans ces conditions: could not bear living under these conditions: ont prefere mourir. preferred to die. se sont rendus compte de ce qui nous arrivait. realized what was happening to us.
ont beaucoup discute entre eux. talked a lot together. ont decide de creer id un Pare National. decided to create here a National Park.
From clauses 7 to 9, the thematic context is a combination of both textual and ideational components. The textual conjunctive elements, in bold, serve to relate the message to the preceding message. Un jour relates the message in clause 7 temporally to the preceding text, alors indicates that the message in clause 8 is a consequence of the events in 7, and enfin relates the message in clause 9 temporally to 8 but also has a rhetorical function, that of introducing the final point. This general introduction to Theme in the context of text paves the way for a more detailed discussion of textual organization in general, and French THEME in particular. 5.2.2 Metafunctions and THEME
As already mentioned in Section 5.2, the interpretation of the descriptive category presented here, Theme, does not only cover topical (experiential)
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
171
elements: the local context or 'semiotic environment' of the clause may be solely experiential, or metafunctionally diverse, that is, textual and/or interpersonal and experiential, as exemplified below. Textual Themes are typically discourse markers, conjunctives and continuatives. Interpersonal Themes include Vocatives, modal Adjuncts and Mood-markers (see Chapter 4). Abbreviations here are: interp = interpersonal, top = topical, text - textual. (2a)
Vraiment, Really,
c' she
est unefilk bizarre, is a strange girl
Theme: interp
Theme: top (unmarked)
Rheme
(2b) mais, but Theme: text
probablement probably
toutes les filles all girls
ont des idees pas ordinaires have funny ideas
Theme: interp
Theme: top (unmarked)
Rheme
In Example 2 (which is taken from Vian 1972: 72) each topical Theme is annotated with the feature 'unmarked'. This is because in the French declarative clause, the Subject is more likely to have thematic status than the Complement or Adjunct. If a Complement or circumstantial Adjunct has thematic status, then it is interpreted as marked, as in Example 3. (3)
Hier, Yesterday
le capitaine the captain
est arrive arrived
Circ: time
Ac
Proc
Theme: marked
Rheme
In other MOOD types, the unmarked Theme may not necessarily be the Subject. Thus, in the imperative mood, the unmarked Theme is the Predicator, while in the interrogative mood, the unmarked Theme will vary depending on the type of interrogative (see Section 5.2.4 below, and Chapter 4 on interrogative types). In Examples 2 and 3, the Themes that are not solely oriented towards the textual metafunction perform some other function in the interpersonal or transitivity structure of the clause. In terms of their realization, they are all positioned initially. As seen before, there are other Themes in French that do not map onto any transitivity or interpersonal functions. They are referred to as 'absolute'. Absolute Themes may be positioned clause initially or finally (reprise absolute Theme). In writing, they are typically demarked from the transitivity/interpersonal structure of the clause by means of a comma, and in speech by tonic prominence realized as a falling tone. The type of thematic organization where the initial Theme is reinforced in final
172
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
position by a reprise absolute Theme is characteristic of French spoken discourse, as illustrated by Text 5.2, a brief dialogue. The relevant clauses of this text are analysed as Examples 4 to 7 below. Text 5.2 From 'Viandox' (Blanche-Benveniste etal, 1991) 1 SI Elle faisait du pot-au-feu ta mere 2 S2 Oui 3 S3 Toujours elle a toujours fait du pot-au-feu ma mere 4 SI bon et comment elle faisait son pot-au-feu ta mere 5 S3 comme ma femme elle le fait (4)
(5)
(6)
(V)
Elle She
faisait du pot-au-feu ta mere made stew your mother
Th: top (unmarked)
Rheme
reprise Theme
toujours always
elle she
a toujours fait du pot-au-feu ma mere has always made stew my mother
Theme: interp
Theme: top
Rheme
reprise Theme
bon well
et and
comment how
elle she
faisait son pot-au-feu made her stew
ta mere your mother
Theme: text
Theme: text
Theme: text/ top
Theme: top
Rheme
reprise Theme
comme ma femme like my wife
elle lefait she makes it
Theme: top (marked)
Rheme
Typically, absolute Themes are conflated with new information and as such have a marked textual status. Marked peaks of prominence (both thematic ones and those presented as New) can also occur at the beginning of the message as initial absolute Theme, as in Examples 8 and 9. (Example 8, like Text 5.2, is taken from Blanch-Benveniste et al. 1991.) (8)
(9)
ma mere my mother
elle she
absolute Theme
Rheme
La rose, the rose,
c' it
absolute Theme
Rheme
a toujours fait du pot-au-feu always made stew
est une belle fleur is a nice flower
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
173
As suggested in Combettes (1975), clauses such as Examples 8 and 9 can be interpreted as having two layers of thematic structure, as shown in Examples 10 and 11. (10)
ma mere my mother
die she
absolute Theme
Rheme
a toujoursfait du pot-au-feu always made stew
unmarked Theme (11)
La rose, the rose,
c' it
absolute Theme
Rheme Theme
Rheme
est une belle fleur is a nice flower
Rheme
This interpretation is in contrast not to the analyses given in Examples 8 and 9, but to the one-layer analysis given in Examples 12 and 13. (12)
ma mere
elk
a toujoursfait du pot-au-feu
absolute Theme
unmarked Theme
Rheme
(13) La rose, absolute Theme
c'
est une belle fleur
Theme
Rheme
Example 14 below could possibly be interpreted as having three layers of thematic structure, following Halliday's 1994 analysis of the thematic structure of clauses with predicated Theme. However, as the textual semantic content of c' and qui is rather weak, there is no discourse motivation for interpreting a third layer of thematic structure as has been done in Example 14. (14)
Celivre This book
c' it
Theme
Rheme
est Paul is Paul
Theme Theme
qui who
I' a lu it has read
Rheme Rheme
Theme
Rheme
Clause complexes of the hypothetic type, where the dependent clause precedes the dominant clause, can also be analysed as having a doublelayered thematic structure, one at clause complex level, the other within each clause, as illustrated in Example 15, which is taken from Saint-Soline (1972: 14).
174
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
(15) Lorsque when
le reveil sonna, the alarm rang
Theme Theme: text
il it
n 'etaitpas encore cinq heures et demi. was not yet half past five
Rheme Theme: top
Rheme
Theme: top
Rheme
We have seen that Theme in French may be metafunctionally diverse, that is, it may include contributions from the textual, interpersonal and experiential metafunctions. This is represented paradigmatically by three simultaneous THEME systems. We also saw that the thematic organization of a clause may be marked or unmarked. In a declarative clause, the unmarked Theme conflates with the Subject, while Themes that conflate with Adjuncts or Complements are interpreted as marked. It was also mentioned that there are two types of marked Theme, absolute and non-absolute. The paradigmatic organization of French Themes is set out in Section 5.2.3. 5.2.3 The French THEME systems
The system network in Figure 5.2 indicates that in any clause one may choose to have a textual Theme or no textual Theme, and an interpersonal Theme or no interpersonal Theme and an experiential Theme. This experiential Theme may be marked or unmarked. If marked, there are three simultaneous sets of choices: the Theme may be (i) single or multiple, (ii) absolute or non-absolute and (iii) predicated or non-predicated. In this account of French Theme, the notion of 'marked' encompasses not only those Themes which conflate with a function other than Subject but also absolute and predicated Themes. Absolute Themes are marked by being outside the experiential structure and predicated Themes are marked in the sense that
Figure 5.2 Primary THEME systems
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
175
they serve to map the 'New' element onto the Theme (see Halliday, 1994: 59, on predicated Theme). All features and combinations of features from the system are illustrated in Examples 16 to 26. When a clause comprises two potential layers of thematic structure, the two layers are analysed as in Example 16. textual and unmarked topical (16) Cependant, il However, he
text Theme
arriva en retard arrived late
top Theme: unmarked
Rheme
textual and interpersonal and unmarked topical (17) Et malheureusement il
arriva en retard
And
unfortunately
he
arrived late
Theme: text
Theme: interp
Theme: top
Rheme
Downranked clauses, that is, clauses that function as an element realized at a rank below the clause, can also function as unmarked theme, as illustrated by Example 18, which is taken from the Introduction to de Saussure's Cours de Linguistique Generate (1972:1). (18)
[[Menerplus loin les connaissances dans le domaine des sciences naturelles et des sciences exactes]]
est une heredite familiale, acceptee avec un orgueil conscient.
To open further the knowledge in the field of natural sciences and exact sciences
is a familial trait accepted with a conscious pride
unmarked Theme (non-finite downranked clause)
Rheme
The clausal Theme in Example 18 could also be predicated, as in Example 19. (19)
C'est [[mener plus loin les connaissances dans le domaine des sciences naturelles et des sciences exactes]] qui est une heredite familiale.
Examples 20 and 21 show the possible two-layered analysis when there is a marked topical Theme or a predicated marked Theme. layer 1: marked topical: single; layer 2: unmarked Theme (20) En retard il arriva Late he arrived marked top Theme
Rheme unmarked Theme
Rheme
176
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
layer 1: marked Theme: predicated; layer 2: unmarked Theme A Rheme A unmarked Theme A Rheme (21) C' est en retard qu' il arriva
it
is late that
he
marked top Theme: predicated
Rheme
Theme
Theme
Rheme
arrived
Rheme
Examples 22 to 24 show analyses of further instances of marked Themes. Example 22 is again from the introduction to de Saussure's Cours, this time the 1972 edition (Saussure 1972: I); Example 23 is from Morel (1992:64). mark
se succedent
dans la vieillefamilk genevoise des Saussure in the old Genevan family of Saussure
premieres annees du XVIII deck,
generation,
Since the first years of the xviii century
from generation to generation
came one after the other
marked top Theme 1
marked top Theme2
marked top ThemeS
mark*sd Theme: absolute: multiple ce tarif-la (23) moi this rate
je ne I'aipas I don't have it
absolute Theme!
Absolute Theme2
1ttieme
It's him
predicated Theme; Theme/Subject
naturalists, physicians and geographers
Rheme
me
mark*ed Theme: predicated (24) C'est lui
des naturalistes, des physiciens, des geographes
qui arriva en retard who arrived late
Rheme
In Example 24 the predicated Theme, lui, is the prominent form of the pronominal Subject form il. This indicates that the predication of an unmarked Theme makes it marked in the sense that it is also prominent in the information structure. The prominent form of the pronominal, lui, instead of its non-prominent correlate, il, indicates tonic prominence (see Chapter 4).
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
177
absolute Theme: initial: cohere with transitivity role and unmarked (25) Lui il arriva en retard Him he arrived late absolute Theme
Rheme unmarked top Theme
Rheme
absolute Theme: reprise and unmarked topical (26) // arriva en retard He arrived late
lui him
unmarked top Theme
Rheme
reprise Theme
Examples of absolute themes that cohere with the previous context are displayed in bold in Text 5.3 and discussed below. This text is taken from Morel (1992). Text 5.3 From Morel (1992: 63) 1 SI [Theme: text] et [Theme: top] quel [Rheme] est le programme d'experimentale en maitrise? and what is the experimental master's programme? 2 S2 [Theme: top] Je [Rheme] nepeux vous le donner au juste I can't tell you exactly 3 SI [Theme: text] non enfin [Theme: top] c' [Rheme] etait pour savoir a pen pres c'que c'etait, quoi no well it was to knowjust what it was, 'damn it' 4 S2 [absolute Theme] e: le programme au niveau de la maitrise [Rheme] vous avez de toutesfacons une recherche en plus des enseignements um: the master's programme you have anyway research in addition to courses 5 SI [Theme: text] out, bon, [absolute Theme] la formation en maitrise [Rheme] quels certificats faut-il prendre? the master's course which diplomas must one have? 6 S2 [Theme: text] etpuis ensuite [absolute Theme] ergonomie [Rheme] c'est pareil? and then ergonomics, is it the same?
The absolute Themes found in clauses 4, 5 and 6 in the exchange in Text 5.3 are not coreferential with an element of the experiential structure, but cohere with preceding elements in the text. Morel (1992) suggests that 'the role of Theme in 4 is to bring back into focus question 1, of which the answer 2 was considered unsatisfactory 3'. The absolute Theme in clause 4 allows the speaker to reorient the addressee to the point of a previous message and elaborate on it. Morel (1992) points out that an absolute Theme which does not cohere with a transitivity role can also accumulate, or tie together, elements from previous messages as in clause 11 in Text 5.4, which is a continuation of the exchange in Text 5.3. This textual accumulation provides a means of introducing the conclusive stage of the exchange.
178
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Text 5.4 From Morel (1992: 63) 7 [Theme] on [Rheme] dit aussi they say also 8 [Theme: P clause] quepourefaire unD.E.S.S. d'ergonomie [Rheme] y avail trespeu de places that to do a D.E.S.S. in ergonomics there were very few places 9 [Theme] il [Rheme] n'y a trespeu d'places effectivement there are very few places indeed 10 [Theme] la selection [Rheme] s'effectue comment? selection is made how? 11 [Theme: text] alors [absolute Theme] les selections e: au niveau dTentree du D.E.S.S en ergonomie bon [Rheme] ce D.E.S.S est ouvert apres une etude de dossier entry selection for a D.E.S.S. in ergonomics well this D.E.S.S. is open after an examination of the student file
The Theme in clause 11 in Text 5.4 accumulates thematic information provided in clauses 8 and 10. Like the Theme in clause 4 (in Text 5.3), it serves to elaborate on points from the preceding cotext, and it indicates a new turn in the exchange. Bi-layered thematic structures where the absolute Theme is coreferential with the unmarked Theme that functions as Subject in the interpersonal structure are further illustrated in Examples 27 (taken from Vian 1972: 72) and 28 (taken from Vian 1972: 52).
(27) £a This
d'ailleurs anyhow
Absolute Theme
Rheme Theme: text
(28)
Ces deux salauds,
ils they
le comprennent en general understand it in general
Theme: top (unmarked)
Rheme
derriere le coin du mur, Those two bastards behind the corner of the wall Absolute Theme
ils they
avaient une mitrailleuse et des tas de cartouches had a machine gun and heaps of cartridges
Rheme Theme: top (marked)
Rheme
The interpretation of the thematic structure in Example 28 indicates that derriere le coin du mur refers, to the location of the 'machine-gun' rather than the location of the 'two bastards'. Example 29 gives an alternative interpretation; it implies that the Theme is a nominal group complex.
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION (29) Ces deux salauds, a
derriere le coin du mur, 0
ils
179
avaient une mitrailleuse et des tas de cartouches
Rheme
Absolute Theme
unmarked Theme
Rheme
The translation of Example 28 by Rawdon Corbett (in Vian 1972) - 'those two bastards [behind the corner of the wall] had a machine gun and heaps of cartridges' - renders the second interpretation of the thematic structure, as it is shown in Example 29. The absolute Theme of the French clause is rendered as an unmarked Theme in English. The comma between ces deux salauds and derriere le coin du mur, in the original, points to the first interpretation, as shown in Example 28. A more faithful translation would have been: 'behind the corner of the wall, those two bastards had a machine gun and heaps of cartridges'. This short discussion on translation issues stresses the importance of rendering all metafunctional components. It also shows that the misinterpretation of one component of meaning, here textual meaning, may lead to a misinterpretation of the experiential meaning as well. The outer or primary layer of the thematic structure is marked in relation to the inner or secondary layer. The inner or secondary layer is unmarked in the sense that it maps transitivity structure onto the interpersonal structure and onto an unmarked textual structure. This is further exemplified by Example 30 (taken from Gadet 1991: 5).
(30) son enfant, her child
mais but
Absolute Theme
Rheme
Reprise Theme
outer . . .
inner layer
. . . layer
elle she
Theme: text Theme (unmarked)
le deteste, him hates
cette mere this mother
Rheme
In Example 30, both son enfant and cette mere tell the listener what the message is about and at the same time carry a contrastive meaning. The message is about her child and no other child and about this mother and no other mother. Within the inner layer of thematic structure, which maps onto a transitivity and negotiatory structure, we have a textual Theme followed by an unmarked Theme. The textual Theme relates the clause to the previous message through contrast, and the unmarked Theme, realized as a pronominal clitic, and hence Given and identifiable, indicates that the message is about a woman who has already been mentioned in the previous message. While absolute Themes tend to be found in spoken discourse, marked topical Themes tend to occur in written discourse. A marked topical Theme
180
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
may conflate with any transitivity role apart from the participant-Subject, since this would be unmarked Theme. Consider Example 31, which was previously encountered in Section 5.2.3, and which is the first clause of the biography on de Saussure introducing his Cours de Linguistique Generak (1972:1). (31)
Depuisles premieres annees du XVIII sieck,
de generation en generation,
Since the first years of the xviii century marked top Theme 1
se succedent
dans la vieille famille genevoise des Saussure
des naturalistes, des physiciens, des geographes
from came one generation to after the generation other
in the old Genevan family of Saussure
naturalists, physicians and geographers
marked top Theme2
Rheme
marked top ThemeS
Where does the Theme of Example 31 end? We could argue that this clause comprises four marked Themes, the first and second conflating with circumstances of time, the third with the Process and the fourth with a circumstance of place thus forming a continuum of degree of markedness which decreases from left to right. Indeed the whole clause is very much marked in its organization, with the Subject coming last as New. This type of clausal organization is not uncommon in narration. The Process, and in particular material processes of movement, often precede the Subject in narrative texts. Should we just analyse the first element as Theme or all three elements as suggested? Does the Theme of any clause, as suggested by Halliday (1985a), extend up to (and include) the first topical element, that is, the first element that has a transitivity function, or can it include more than one topical Theme? As the realization of thematic structure is not segmental but rather wave-like, one cannot be certain of where Theme stops. However, it would seem that when we look at thematic progression in a whole text, it is the first topical element that seems to contribute most strongly to the method of development of the text while the other ideational Themes have a more local function. On this basis, it can be argued that the first topical element is thematically the most salient, with thematic saliency diminishing as we move from left to right. Matthiessen (1992: 51-2) notes, following his discussion on multiple ideational Themes, that: . . . we get the same kind of diminuendo effect even with single marked Themes in declarative clauses, where the Subject still has some thematic prominence, as indicated by the fact that it may relate to the method of development just as when it is the unmarked Theme of the clause . . .
As we have seen in this section, French experiential Theme systems have two main types of marked Themes: absolute and non-absolute. Absolute Themes
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
181
are characteristic of spoken discourse and are often realized by prominent pronominals (moi, toi, lui/elle, nous, vous, eux/elks) in dialogic texts. These prominent pronominals tend to function as Theme and New or Rheme and New, while non-prominent pronominals (clitics) are typically unmarked Theme and Given or Rheme and Given. In addition, non-prominent pronominals have long staying power in discourse, forming reference chains throughout the dialogue, while prominent pronominals more typically mark contrastive emphasis, and serve to stage the exchange (see Chapter 4). These generalizations regarding French pronominals can be extended to the interpretation of the discourse function of marked Themes in general. Thus, marked Themes, whether prominent pronominals, absolute or not, tend to indicate a change in dialogic text or a new stage in monologic ones. They tend to correlate with generic stages. Unmarked Themes, in contrast, tend to convey thematic continuity. This claim will be supported by the analysis of thematic patterns in written texts in Section 5.3 below. At this stage, we have focused solely on the thematic structure of the declarative clause. Typically, textual Themes were conjuncts, while interpersonal Themes were attitudinal and topical Themes, nominals or circumstantials. The kind of elements functioning as textual, interpersonal or ideational Theme, as well as the unmarked Theme, may vary according to mood type. Thus, for example, the Predicator is unmarked (interpersonal) Theme in the imperative mood, while both Finite and Predicator are unmarked (interpersonal) Theme in some polar interrogatives, and the Qu-element is unmarked in a nonpolar interrogative where there is a conflation of interpersonal function and topical Theme. The Qu-element in a relative clause, on the other hand, is a conflation of a textual and topical Theme. In the next section, we investigate the various Theme options in more detail in relation to mood options. 5.2.4 Theme across MOOD types
What constitutes an unmarked Theme, rather than a marked one, in French is not always easy to determine. Unlike in English (see Halliday 1994; Eggins 1994), the unmarked Theme within one particular mood may vary. This is because of the variation in the realizations of MOOD options and in particular of the interrogative MOOD option (see Chapter 4). Thus determining what is marked or unmarked Theme in French is not such a simple matter. How does one determine how the clause is typically organized within different mood types? Is viens-tu ? marked in relation to tu viensfor the latter marked in relation to the former? The use of viens-tu?may be marked in casual conversation, for example, while the use of tu viens? may be marked in a formal context. However the discourse markedness does not imply that the textual organization is marked. Functional variation which determines the choice of mood option has to be separated from variation in thematic structure. Each variety of interrogative mood type is unmarked in some context, and thus we will argue that each interrogative type has a
182
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 5.1 Unmarked thematic structures in the interrogative mood systemic features
examples (unmarked theme in bold)
polar: no grammatical prosody polar: grammatical prosody: juncture polar: grammatical prosody: internal
tu dimes les gateaux? est-ce que tu aimes les gateaux'? aimes-tu les gateaux? or as-tu aime les gateaux? qui a mange ce gateau ?
nonpolar: qu-Subject: no grammatical prosody nonpolar: grammatical prosody: juncture nonpolar: grammatical prosody: internal: thematic nonpolar: no grammatical prosody: non-thematic
qui est-ce qui a mange ce gateau ? que manges-tu ? tu manges quoi
particular unmarked textual organization. Typical thematic structures across interrogative mood types are shown in Table 5.1, with the unmarked Theme in each clause highlighted in bold. Table 5.1 shows that in polar interrogatives the unmarked topical Theme is always the Subject, as in the declarative mood. However, unlike in the declarative, in the interrogative this unmarked Theme can be preceded by an interpersonal Theme. If the mood is realized by a juncture prosody, the initial Theme must be the interpersonal mood interrogator est-ce que. If the prosody is internal, then the initial interpersonal Theme is either the Finite or the Finite conflated with the Predicator depending on the tense selection. Note that whether the Finite and Predicator are conflated or not we analyse both elements as part of the Theme on the basis that they cannot function independently of the other (see Chapter 4). The interpersonal salience of the Finite and Predicator of French gives them the status of interpersonal Theme, but as they also conflate with the Process, they are a conflation of interpersonal and topical Theme. In a nonpolar interrogative the unmarked topical Theme is typically conflated with the Qu-element, except when the latter occurs in rhematic position. Thus the typical initial Theme in a nonpolar interrogative is an interpersonal Theme conflated with a topical Theme. This topical Theme may be a Subject, a Complement or an Adjunct. It may be followed by the mood interrogator est-ce que. In the imperative mood, the typical point of departure is the Predicator/Process, as in Example 32. (32) Donne-moi Give me
ce gateau this cake
Theme: interp/top
Rheme
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
183
In Example 32, the Predicator/Process + Complement-clitic are together interpreted as both interpersonal and topical Theme because the initial position of the Predicator/Process followed by clitic is an indicator of the mood. In an imperative with negative polarity, the process is preceded by an interpersonal Theme, the negative adjunct clitic, ne, as in Example 33. (33) Ne mange pas Don't eat Theme: interp A interp/top A interp
ce gateau this cake Rheme
The prosodic realization of the negative Adjunct implies that the interpersonal Theme it conflates with is split in a way analogous to the Finite and the Predicator in as-tu dime les gateaux? The notions of 'marked' and 'unmarked', as the network in Figure 5.2 shows, are used systemically only with regard to topical Themes. There is no need for such a distinction for textual and interpersonal Themes. Textual Themes which are conjuncts or continuatives always occur first in the clause. If there is an interpersonal Theme, it will occur typically after the textual Theme and before the topical Theme if not conflated with it. Interpersonal Themes can sometimes precede textual Themes, and it could be argued that in this instance the interpersonal Theme is marked. Following this overview of Theme choices, we will explore theme patterns in texts and how they relate to the method of development of these texts. 5.3 A discourse-based approach to Theme The thematic analysis of texts shows that the choice of Theme is not arbitrary; it serves to create the method of development of the text: The lexical material placed initially within each sentence of a paragraph (i.e. the Themes of each sentence of a paragraph) indicates the point of departure of the message expressed by that sentence, and the information contained within the Themes of all the sentences of a paragraph creates the method of development of that paragraph. (Fries 1981: 135)
Fries' (1981) insightful semantic interpretation of Theme will serve as point of departure to this discussion. We will illustrate on the one hand how thematic choices in a text correlate with the structuring of information and contribute to the global organization of ideational content, and on the other hand how, at a more local level, thematic choices correlate with the local flow of information and orient the reader to important points. In addition, we will see that Theme patterns can serve to realize second-order meanings, as can any grammatical pattern. In subsection 5.3.1, we analyse a factual text, more precisely a geographic description, to illustrate Fries' (1981: 119) argument
184
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
that 'thematic content correlates with the method of development of a text (and with the nature of that text)'. The analysis of a short literary biography in subsection 5.3.2 will serve to support the hypothesis that a text may reveal two methods of development, corresponding to two layers of thematic organization: one typically oriented to the local flow of information (unmarked) and the other typically oriented to the text structure (marked). At first, this text does not appear to display any continuity in its THEME selection. However, a more in-depth analysis of the marked and unmarked thematic patterning reveals that discontinuity in marked thematic patterns can be used as a strategy for organizing a text as a coherent whole, while the unmarked thematic pattern creates a topical continuity. The texts selected represent instance types rather than only particular instances. Thus, the claims that are made about their thematic potential are not based on the thematic analysis of these sample texts alone but reflect patterns found in texts of similar types. In subsection 5.3.3, we analyse a news article about France winning the soccer world cup in 1998. The text foregrounds Chirac as the winner as if he was the one who scored the goals. The text is also about the construal of solidarity, how the game brought all French people ('white', 'black' and 'arabs') together. These two main themes are clearly brought forward by the thematic patterns of this text. The thematic analysis will again display two layers of thematic organization. 5.3.1 Method of development and text organization: a geographic description
The text selected, 'Les beaux quartiers', forms part of a macro-subsection, entitled 'Paris', itself embedded in a macro-section 'Visages de la France', which represents one of six main sections of the macro-genre to which the text under focus belongs, i.e. a guide book entitled Le nouveau Guide de France (Michaud and Kimmel 1990). The organization of this instance of a macrogenre can be represented synoptically in terms of a constituency tree, as in Figure 5.3. The nature of the text (a geographic description of a suburb serving as guide) allows us to predict the kind of method of development it may display. The text's method of development will either be the architectural components of the suburb or the relative locations and arrangement of these components. This coincides with Fries' (1981) conclusions about the types of method of development found in geographic descriptions. Fries (1981: 129) shows that there are two main approaches to the description of a place: The first approach . . . takes relative position as a method of development of the description, and this correlates with the regular appearance of references to relative position in the Themes of the component sentences. The second approach . . . takes reference to component parts as the method of development of the description. In these passages the Themes of each of the sentences refer to component parts of the object being described.
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
185
Figure 5.3 A partial constituency representation of Le Nouveau Guide de France If we take into account the function of our text, which is to orient the reader within the particular place it describes, then the locational arrangement of the architectural parts is more likely than the parts themselves to form the method of development of the core text. This is borne out by the following analysis. However, if we take into account the potential double-layering of the thematic structure, then we can argue that this text has a marked and unmarked thematic organization as does the Zevaco text in the next section. While the marked (primary) thematic organization reflects the temporal structuring of the text, the unmarked (secondary) thematic highlights the component parts as method of development. An analysis of the primary layer of thematic organization of the 'Les beaux quartiers' text follows, with marked Themes highlighted in bold. Text 5.5 'Les beaux quartiers de Paris' from Le Nouveau Guide de France (Michaud and Kimmel 1990: 40) 1 [Theme: marked top] Au nord de la Concorde et de I'Elysee, [Rheme] lefaubourg Saint-Honore est le domaine du commerce de luxe, des parfumeurs et de la haute couture. North of the Concorde and the Elysee, the inner suburb of Saint-Honore is the domain of wealthy traders, perfumers and haute couture. 2a [Theme: marked top] Au-dela, [Rheme] c'est la 'plaine Monceau', Beyond, it is the 'plaine Monceau' 2b [Rheme] bdtie apres 1800 autour du pare de Philippe d'Orleans. built after 1800 around the park of Philippe d'Orleans.
186
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
3
[Theme: unmarked top] L 'avenue des Champs-Elysees, mi-promenade, mi-boulevard, traceepar le Notre au XVIIe S.,apres avoir ete la residence de I'aristocratie, [Rheme] est aujourd'hui, avecses cinemas, sesgrands cafes, sesjournaux, ses stands d'automobiles, le centre du Paris cosmopolite. The Champs-Elysees Avenue, half-esplanade, half-boulevard, designed by Le Notre in the C17, after having been the residential place of the aristocracy, is today, with its cinemas, large cafes, newspapers, motorcar stands, the centre of cosmopolitan Paris. 4a [Theme: marked top] Plus au sud, [Rheme] ce sont d'anciens villages agglomeres pen a pen a la capitate: Further south are old villages piled up gradually to the capital 4b [Rheme] Chaillot, la 'colline aux neuf musees', dominant la Seine, quartier des ambassades et des riches etrangers; Chaillot, the 'hill of the nine museums', overlooking the Seine, the quarter of ambassadors and rich foreigners; 4c [Rheme] Passy, celebre au XVIIIe S., par son chateau et par celui de la Muette, Passy, famous in the C 18, for its castle and that of the 'Muette' 4d [Theme: text/top] oil [Rheme] habita Louis XV; where Louis XV lived; 4e [Rheme] Auteuil, longtemps demeure campagnard, Auteuil, for a long time remained country, 4f [Rheme] attirant depuis plusieurs siecles des generations d'hommes de lettres. attracting for many centuries generations of men of letters. 5 [Theme: marked top] Sur I'autre rive, au fond d'une magnifique esplanade, [Rheme] I'Hotel des Invalides, construitde 1671 a 1676 par Liberal Bruant surl'ordre de Louis XVI pour les soldats blesses, encadre harmonieusement Veglise et son dome, chefd'oeuvre de I'art classique du aJ.Hardouin-Mansard. On the other bank, at the far end of a magnificent esplanade, I'Hotel des Invalides, built from 1671 to 1676 by Liberal Bruant by order of Louis XVI for wounded soldiers, surrounds harmoniously the church and its dome, masterpiece of classical art thanks toj. Hardouin-Mansart. 6 [Theme: marked top: predicated] C'est la [Rheme] que se trouve le tombeau de Napoleon I. It is there that is the tomb of Napoleon I. 7 [Theme: text] Enfin [Theme: marked top] a I'extremite du Champ-de-Mars, ancien champ de manoeuvres devenu le cadre des Expositions Universelles successives, [Rheme] se dresse I'Ecole militaire, edifieepar Gabriel au XVIIIe S., au centre d'un quartier d'aspect soknnel, aux larges avenues, qui abrite le nouveau palais de I'UNESCO. Finally, at the end of the 'Champ-de-Mars', former parade ground turned setting for successive world expositions stands the military school, built by Gabriel in the CIS, at the centre of a quarter with a solemn aspect, large avenues, which accommodate the new UNESCO palace.
The Theme selections in Text 5.5 are summarized in Table 5.2. The primary method of development of Text 5.5 is clearly foregrounded by means of marked Themes referring to locations relative to the centre of the suburb, L'Elysee. The latter is the only element that functions as unmarked Theme; although it refers to a component part of the suburb, it has relevance as 'temporal' Theme as it also stands for the centre of the suburb. The relationship between thematic content and the method of development is schematized in Figure 5.4 below.
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
187
Table 5.2 Theme distribution in Text 5.5 Clause
Textual
1
2 3 4a
4d
ou
Topical
Markedness of topical Theme
Au nord de la Concorde et de 1'Elysee Au-dela L'avenue des Champs-Elysees,mi-promenade, mi-boulevard, tracee par Le Notre au XVIIe S., apres avoir etc la residence de 1'aristocratie Plus au sud
marked marked unmarked
ou
5
Sur 1'autre rive, au fond d'une magnifique esplanade
6
la
7
Enfin
a 1'extremite du Champ-de-Mars, ancien champ de manoeuvres devenu le cadre des Expositions Universelles successives
marked unmarked marked marked: predicated marked
Figure 5.4 Theme selection and method of development
188
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 5.4 represents the layout of the suburb, introduced as unique in the hyperTheme. Clearly this layout serves as method of development for this geographic description. The instantial system of Theme which progresses throughout the text serves as a guide to the reader who wishes to discover 'les beaux quartiers'. This text is similar in its thematic organization to the 'Rugged Coast to Bleak Desert' text in Matthiessen (1994: 39): Here the system of THEME is used to guide the reader on his/her semiotic journey so that s/he can build up an ideational map of part of California. The Themes specify the points of departure for each leg of the journey - the points where the instantial system is to be expanded . . . and the instantial system is built up region by region without backtracking or discontinuity.
Following the suggestion that the selection of a marked Theme does not preclude the selection of an unmarked Theme, 'Les beaux quartiers' can be analysed as having a less prominent method of development realized in the selection of unmarked Theme. The secondary thematic pattern reveals the second approach mentioned by Fries (1981) in relation to Theme patterns in similar English texts: choosing the component parts as method of development. The unmarked Themes of Text 5.5 are summarized in Table 5.3. 5.3.2 Method of development and text organization: a biographic description
The next text is a short literary biography found at the beginning of a paperback novel. The topic is Michel Zevaco, author of historical novels. The method of development in Text 5.6 is not as overt as that of Text 5.5. Consider the thematic analysis of the Zevaco text. (Note that in Text 5.6, the clauses are numbered as usual on the left-hand side, but each topical Theme is also numbered in bold within the text, and these numbers will be used to refer to the Themes in the following discussion and its associated figures and tables.)
Table 5.3 Unmarked Themes in Text 5.5 Clause
unmarked Themes
1 2 3 4 5 6 7
le faubourg Saint-Honore c' (la plaine Monceau) 1'avenue des Champs-Elysees d'anciens villages 1' Hotel des Invalides le tombeau de Napoleon I 1'Ecole militaire
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
189
Text 5.6 Short literary biography of Michel Zevaco 1 [Theme: unmarked top] 1 Michel Zevaco [Rheme] est neen 1860 a Ajaccio (Corse). Michel Zevaco was born in 1860 in Ajaccio (Corsica). 2 [Theme: marked top] 2 Apres de brillantes etudes de lettres, [Rheme] il est nomme professeur au college de Vienne (here). After brilliant Arts studies he was appointed professor to the college of Vienne. 3a [Theme: marked top] 3 Revoque au bout de pen de temps pour 'idees avancees', [Rheme] il part pour Paris Dismissed after a short time for 'progressive ideas' he went to Paris 3b [Theme: text] et [Rheme] devient redacteur du quotidien anarchiste L'egalite. and became editor of the anarchist daily paper L 'egalite. 4a [Theme: marked top] 4 Des 1890 [Rheme] il est connu pour son talent depolemiste As early as 1890 he is known for his talent as a polemicist 4b [Theme: unmarked top/text] 5 qui [Rheme] lui vaut de passer un mois a la prison politique de Sainte-Pelagie. which costs him a month in the political prison of Saint-Pelagie. 5 [Theme: marked top: (3 clause] 6 Le journal disparu, [Rheme] Michel Zevaco entre comme critique artistique et litteraire au Courrier fran^ais. The newspaper gone, Michel Zevaco enters as literary and art critic at the Courrier francais. 6 [Theme: marked top: p clause] 7 Pour elever ses cinq enfants, [Rheme] il se met a ecrire des feuilletons historiques. To raise his five children, he starts writing a historical series. 7 [Theme: marked top] 8 Des le premier - Borgia - [Rheme] le succes est immediat. From thefirstpublication - Borgia - success is instantaneous. 8a [Theme: unmarked top] 9 // [Rheme] sepoursuit en 1901 avec la publication du 'Pont des Soupirs' It continues in 1901 with the publication of 'Pont des Soupirs' 8b [Theme: text] et [Rheme] s 'affermit avec les quelques trente ouvrages qu'ecrira Michel Zevaco en dix-huit ans and strengthens with the thirty or so works that Michel Zevaco will write in 18 years 8c [Theme: unmarked top/text] 10 parmi ksquels les aventures des Pardaillans [Rheme] forment un cycle important. among which the adventures of the Pardaillans form an important cycle. 9 [Theme: unmarked top] 11 Michel Zevaco [Rheme] est mart en 1918. Michel Zevaco died in 1918.
It could be argued that the hyperTheme of this text is 'Zevaco'. This would correlate with the nature of the text as a short account of an author's life. In this kind of text, one would predict that the method of development would either be the author or the stages of the author's life. In other words, the method of development is likely to be topical or temporal. At first glance, the method of development of this text appears to follow neither pattern. The Theme-Rheme distribution encountered in Text 5.6 is summarized in Table 5.4. The Themes in Table 5.4 do not display an explicit temporal method of development as in Text 5.5. However, patterns of thematic flow are not always transparent as in Text 5.5, and can progress in various ways, as shown in
190
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Table 5.4 Theme-Rheme distribution in Text 5.6 Cl
Theme
Rheme
1 2
Michel Zevaco Brilliant studies
3a
Dismissed for progressive ideas
4a 4b 5 6 7 8a
From 1890 which Newspaper gone To raise his five children From first publication - Borgia It (success)
8c 9
The adventures of the Pardaillans Michel Zevaco
birth date and place appointed professor, college of Vienne goes to Paris, becomes anarchist editor known as polemicist causes time in political prison becomes literary and art critic writes historical series immediate success continues with'Pont des Soupirs', and strengthens with 30 more works form an important cycle death date
Figure 5.5. Pattern 1 represents a linear progression where the Theme follows from the preceding Rheme and so on; Pattern 2, a progression where each Theme is derived from a more general Macro-Theme; and Pattern 3, a constant progression where the Theme is kept from clause to clause. We could argue that Text 5.5, 'Les beaux quartiers', displays pattern 2 of thematic progression, where the Macro-Theme Les beaux quartiers is then subdivided into different locations. In the Zevaco text the method of development revolves around patterns 1 and 3 of thematic progression represented in Figure 5.5. Pattern 1 is realized within each generic stage of the Zevaco text. When the pattern breaks down, i.e. when there is discontinuity in the thematic progression, discontinuity in the sense that the Theme does not follow from the previous Rheme, then a new life-stage is introduced. Figure 5.6 shows that the global thematic organization, which contributes to the structuring of information, corresponds more or less to a generic structure based on life stages which can be summarized as
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
191
Figure 5.5 Different patterns of thematic progression (Danes 1964, taken from Fries 1981) Figure 5.6 indicates that each global Theme in Text 5.6 begins a new stage in Zevaco's life. Each stage is surrounded by boxes. Within each stage the local flow of information is realized by a pattern 1 of thematic progression. The initial Theme in each stage functions as both global and local Theme, the others perform only a local function. Thus, if we look at the text, the first global Theme is 'Zevaco'. It introduces the first life-stage, the birth of the author. The next Theme does not follow from the previous Theme or Rheme. A new stage is introduced, that of the author's early adult life (education and first job). This second stage forms a thematic pattern where the second Theme follows from the first Rheme of stage 2. Thus, we understand that the non-finite clause revoque pour idees avancees was chosen as Theme because the preceding Rheme tells us about the author's appointment as professor. The next global Theme is a temporal circumstance and thus overtly introduces a new stage. The thematic organization of the text follows this pattern of discontinuity and continuity throughout. Figure 5.6 shows clearly that the most developed stage is that of the author's literary life, which is not surprising considering the nature of the text. Thus despite the fact that the thematic organization is not overtly temporal in the sense that the marked themes are in most parts not temporal in themselves, we can see
192
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 5.6 Global organization and local flow that the thematic patterns follow a temporal method of development carried forward by a linear thematic progression. It was mentioned earlier that pattern 3 is also foregrounded in the Zevaco text. This pattern, where one Theme is re-selected throughout the text, can be seen to be present if we argue that the text foregrounds two layers of thematic organization. Thus, within the marked pattern that contributes to the structuring of the text, where (3 clauses are thematic within clause complexes, there is an unmarked pattern with Zevaco and his literary success forming a constant thematic progression. This second pattern of thematic organization is represented in Table 5.5. So far Theme-Rheme patterns in the previous texts have shown that marked thematic structure correlates more or less with the rhetorical structure of the text and that unmarked thematic structure provides a continuous flow of information throughout the text. The next text will illustrate how thematic patterning can be used to construe second-order meaning, i.e. the overall theme of the text. 5.3.3 Theme patterns and second-order meaning
The following article, taken from Le Monde, is a representation of how the French president, Chirac, portrayed himself and was seen by many after France won the soccer World Cup in 1998: as 'the star' of the day, as if it was Chirac himself who scored the three winning goals. Furthermore, Chirac was
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
193
Table 5.5 The unmarked Theme-Rheme pattern of Text 5.6 Clause
Unmarked Theme
1
Michel Zevaco il il [H] il Michel Zevaco il le succes il [il] Michel Zevaco
2 3a 3b 4 5 6
7 8a 8b 9
also seen as bringing French people together as a multicultural society, as a reflection of the multicultural background of the players. The first layer of analysis is presented in Text 5.7, with marked Themes again highlighted in bold. Text 5.7 'Jacques Chirac, star a la sortie du parking' (Le Monde, 14 July 1998) 1 [Theme: unmarked top] On [Rheme] se serait cru revenu au soir de la presidentielle del 995. It was like we had been brought back to the night of the 1995 presidential election. 2a [Theme: marked top: P clause] Sortant du Stade de France dans sa voiture officielle, dimanche soir, Coming out of the stadium in his official car, Sunday night, 2b [Theme: marked top: P clause] apres avoir remis la Coupe du monde aux vainqueurs after having handed over the World Cup to the winners 2c [Theme: marked top: P clause] et passe un moment avec eux dans les vestiaires, and spent a moment with them in the changing-room 2d [Rheme] Jacques Chirac a ete assailli, a la sortie du parking, par unefoule dejeunes qui manifestaient leurjoie de la victoire des Bleus. Jacques Chirac was assailed, leaving the car park, by a crowd of young people who were showing their joy in response to the victory of the French team. 3a [Theme: marked top: p clause] Faute d'avoirpu acceder au stade Through not being able to get to the stadium 3b [Theme: marked top: P clause] pour voir leur heros to see their heroes 3c [Rheme] ils se sont rues sur la Safranepresidentielle they pounced on the presidential car 3d [Theme: text] pour [Rheme] en feliciter I'occupant to congratulate its occupant 3e [Theme: text] commes' [Theme: unmarked top] il [Rheme] avail lui-meme marque les trots buts de la France. as if he himself had scored the three goals for France.
194 4a 4b 4c 5
6
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH [Theme: unmarked top] M. Chirac [Rheme] ne s'est pas fait prier, Mr Chirac did not need persuading [Rheme] ouvrant grand la fenetre aux vitres fumees opening wide the tinted window [Theme: text] pour [Rheme] serrer les mains qui se tendaient. to shake the hands that were held out. [Theme: marked top] La France 'black-blanc-beur' et le president de la Republique communiant dans la meme ferveur, comme aux plus beaux jours de la 'generation Mitterrand', [Rheme] il nefallaitpas rater I'aubaine. France 'black-white-arab' and the president united in the same fervour, as in the best days of the 'Mitterrand generation', one could not miss that opportunity. [Theme: text] D'autant que [Theme: unmarked top] premier ministre, Lionel Jospin, [Rheme] s'etait, lui, eclipse sans faire de bruit. More especially as the prime minister, Lionel Jospin, him, had slipped away without a noise.
The text starts with an unmarked Theme on in clause 1 which is a generalized pronoun referring in this context to French people in general rather than designating anyone in particular. A sequence of dependent clauses follows in clause complex 2, functioning as marked theme. This serves to foreground 'happenings' as Themes, where the implicit actor is Jacques Chirac, as indicated in the main clause functioning as Rheme. At a second order of semiosis, the text is itself a match where the players are Chirac and the French people who are foregrounded in the next set of marked Themes in clause complex 3 (also dependent clauses) where the implicit actor is a crowd of young people. Then Chirac is again picked up as theme in clauses 3e and 4a as the 'winner', with both groups of 'players' brought together cohesively in the absolute Theme in clause 5 as in a handshake. An absolute Theme, as mentioned before, also functions as New and in this instance serves to bring to the fore this new apparent unity between French people. The final Theme in clause 6 foregrounds the socialist prime minister, Jospin, as outside this communion between Chirac and 'his people', further underlined by the textual Theme, d'autant que (more especially as). At one level, Text 5.7 foregrounds the dynamic atmosphere of the event, with the text developing around 'happenings' foregrounded in nonfinite dependent clauses, thus forcing processes into initial position. At another level it is about the people involved in these processes, Chirac and the French people, who function in turn as implicit unmarked Theme within these dependent clauses. Furthermore the text is about the construal of solidarity between Chirac and all French people ('black, white and arab'), solidarity which is encapsulated in the absolute Theme in clause 5, while keeping one participant - the socialist prime minister, Jospin - out, further foregrounding Chirac as the only winner. The two layers of thematic patterns that are intertwined in this text are schematized in Table 5.6 with the implicit unmarked Themes in brackets preceding the marked Themes. The unmarked thematic patterns are in italic, whilst the marked thematic patterns are in bold.
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
195
Table 5.6 Marked and unmarked Theme selections in Text 5.7 Clause
Theme
1 On 2a [ Chirac] Sortant du Stade de France 2b [ Chirac] avoir remis la Coupe du monde 2c [ Chirac] passe un moment avec eux 2d Jacques Chirac 3a [Lesjeunes] avoir pu acceder au stade 3b [Lesjeunes] voir leurs heros, 3c Ils (les jeunes) 3d [Lesjeunes] en feliciter 1'occupant 3e // (Chirac) 4a M. Chirac 4b [ Chirac] 4c [ Chirac] 5 La France 'black-blanc-beur' et le president de la Republique communiant 6 le premier ministre, Lionel Jospin
While the marked themes foreground the happenings, reliving the events that united the president and the French people, the unmarked themes foreground each participant in turn as in an exchange building towards the construal of solidarity. While within each stage the thematic progression is constant, the shift from one stage to another is indicated by a linear progression, where what is introduced in the Rheme is picked up in the following Theme, as shown by the diagram of the thematic progression of Text 5.7 in Figure 5.7. The analysis of this text has shown that not only do theme selections contribute to the method of development of the text and the generic scaffolding of that text but also contribute to the construal of second-order semiosis, social and political, with the interweaving of both marked and unmarked thematic progressions building up to elevating Chirac, both as the 'winner' of the cup and as a social/political winner. 5.4 Conclusion This overview of the thematic potential of French concludes this descriptive tour of the French clause as a grammatical unit where the metafunctions are mapped onto one another. This chapter has focused on how the information presented by ideational (Chapters 2 and 3) and interpersonal meanings (Chapter 4) is organized as a message. It has suggested that the French clause is not organized randomly but serves to create the method of development of a text: thematic choices provide directions on how to process the information constructed by the ideational and interpersonal meanings.
196
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Figure 5.7 Thematic progression in Text 5.7 We saw that the notion of Theme in systemic functional theory is broader than the notion of Theme in other functional approaches. It is defined as 'the point of departure of the message' or 'the local context in which each clause is to be interpreted'. Theme may consist of topical elements but also interpersonal and textual ones. Thus it is not equated with 'what the clause is about' (the notion of topic), and it is not equated with 'what we already know' (the notion of given). Systemic functional theory has a separating approach to Theme-Rheme and Given-New structure, the latter being a structure of the information unit rather than the clause. Fries' (1981) discourse-based approach to the two textual structures showed that while the Theme created the method of development of a text, the New foregrounded the main 'points'. In this interpretation of French THEME, we suggest that a clause may have two layers of thematic organization, which relate to two potential methods of development, one marked and the other unmarked. It would appear that
THE 'ENABLING' GRAMMAR: TEXTUAL METAFUNCTION
197
there are always two main approaches to the development of a text. Thus, a geographic description, as Fries (1981) points out, may foreground either relative position or reference to component parts as method of development; a short literary biography, in contrast, may foreground either life stages or reference to the author. We also saw in the news example that choices in thematic progression can contribute to the construal of secondorder meanings. The analysis of Theme patterns in these three registers geographic description, short literary biography and news - illustrated how two layers of thematic structure may function together in creating textual unity. Notes 1 The term 'point of departure' was originally coined by Matthesius. Firbas (1987, p. 140) notes that 'Danes (1964) points out that, discussing the theme, Mathesius used three terms: "vychodiste" (point of departure), "tema" (theme), and "zaklad" (basis, foundation)'. 2 An absolute Theme is a Theme that does not map onto a transitivity or interpersonal function.
English-French glossary of terms in systemic functional linguistics Glossaire anglais-frangais des termes en linguistique systemique fonctionnelle The definitions are translations of Christian Matthiessen's glossary definitions that are found in Appendix 3 of his Lexicogrammatical Cartography: English Systems (1995). Some definitions have been shortened, others slightly changed, while others have been added. I am responsible for any possible misinterpretations of systemic functional terms. Many French terms have been found in Claude Hagege's La Structure des Langues (1982) but are also widely used in today's French linguistics, with the exception of what is uniquely systemic terminology. The terms which have been taken from French linguistics are here often used with some differences in meaning, hence the need for definitions. Les definitions sont traduites du glossaire presente en Appendice 3 de la Lexicogrammatical Cartography: English Systems de Christian Matthiessen (1995). Certaines definitions ont ete abregees, d'autres legerement modifiees et d'autres ajoutees. Toute interpretation erronee des termes de la linguistique systemique fonctionnelle serait de ma responsabilite. Les nombreux termes frangais qui ont ete empruntes a La Structure des Langues de Claude Hagege (1982), sont aussi souvent utilises par la linguistique francaise en general, a 1'exception de ceux qui relevent uniquement de la terminologie systemique. Les termes empruntes a la linguistique francaise sont souvent utilises ici dans un sens different, d'ou la necessite des definitions. Clause. Fr: Phrase. Rang superieur du stratum lexicogrammatical. Clause complex. Fr: Phrase complexe. Combinaison de phrases reliees par parataxe ou hypotaxe et non par enchassement; la combinaison des phrases se fait selon le mode d'organisation logique de la metafonction ideationnelle. Plusieurs phrases peuvent etre, par exemple, reliees par coordination pour former une phrase complexe. Lexicogrammar. Fr: Lexicogrammaire. Combinaison de la grammaire et du lexique (vocabulaire); 1'ensemble des ressources pour signifier verbalement. En dehors de la linguistique systemique, la grammaire et le lexique sont presque toujours traites comme des modules distincts. La theorie systemique
ENGLISH-FRENCH GLOSSARY OF TERMS
199
interprete le lexique comme la grammaire la plus fine (voir 'finesse' ci-dessous) ou specifique. MOOD. Fr: Modes/types de phrase. Region interpersonnelle (interpersonal); la grammaticalisation des fonctions elocutives (speech functions) dans la phrase. Realization. Fr: Realisation. Relation interstratique (entre les strata). Instantiation. Fr: Actualisation. Relation intrinseque (a 1'interieur d'un seul et meme stratum). Stratum. Fr: Stratum. Un systeme ou un ordre d'abstraction particulier au langage: semantique, lexicogrammaire et phonologic sont les trois strata du systeme linguistique dans la theorie systemique de Halliday. Les strata sont relies par realisation (interstratique). Par exemple, le stratum de la semantique est realise par la lexicogrammaire. (Discourse) semantics. Fr: Semantique (de texte). Stratum du plan du contenu qui sert d'interface entre la grammaire et le contexte. L'unite semantique de base en linguistique systemique fonctionnelle est le texte. Text. Fr: Texte. Dans la theorie systemique, le texte est 1'unite semantique. C'est une partie de langage qui fonctionne en contexte. Un texte peut etre parle ou ecrit. Context. Fr: Contexte. Contexte culturel; contexte situationnel (B. Malinowski est a 1'origine de ces notions). Dans la linguistique systemique, elles font reference aux systemes semiotiques d'ordre superieur situes au dessus du systeme linguistique. Le terme contexte embrasse le champ, la teneur et le mode de discours. Register. Fr: Registre. Variete de discours determinee par les variables contextuelles que sont le champ [du discours] (type d'activite sociale), la teneur [de la relation entre locuteur et auditeur] (relations sociales) et le mode (d'organisation symbolique). Tenor. Fr: Teneur. L'un des trois constituants du contexte. La relation sociale qui existe entre les participants dans une situation linguistique. Cela inclut les degres de formalite, de pouvoir, d'affection. La teneur de la relation influence les choix interpersonnels dans le systeme linguistique. Par exemple la strategic que Ton choisit pour realiser un ordre depend largement de la teneur de la relation entre auditeur et locuteur. Field. Fr: Champ. L'un des trois constituants du contexte. Le champ du discours est 1'activite sociale qui a rapport au texte. Mode. Fr: Mode. L'un des constituants du contexte. Represente le role joue par le langage dans la situation linguistique. Cela inclut le moyen d'expression (parle, ecrit, ecrit pour etre parle, etc.) ainsi que le mode rhetorique (instructif, persuasif, etc.). Le Mode est une categoric de second ordre dans le sens ou il est amene a exister par 1'existence meme du langage.
200
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
Delicacy. Fr: Finesse. Gamme qui va du general au specifique. Dans un reseau de systemes, la 'finesse' correspond a 1'agencement des sytemes de la gauche vers la droite au moyen de conditions d'entree (entry conditions). Rank. Fr: Rang. Hierarchic d'unites telle qu'on la trouve au stratum lexicogrammatical: phrase-syntagme/groupe-mot-morpheme. Les fonctions des unites d'un rang sont realisees par des unites du rang inferieur. Par exemple, les fonctions au rang de la phrase sont realisees par des syntagmes (nominaux/verbaux) ou groupes prepositionnels et les fonctions au rang du syntagme sont realisees par des mots. Rankshifted. Fr: Deplace de rang. Lorsque 1'unite d'un rang fonctionne dans la structure d'une autre unite comme si elle etait une unite d'un rang inferieur. Par exemple lorsqu'une phrase relative (defining relative clause) fonctionne comme post-modificateur dans la structure d'un syntagme nominal. Transitivity. Fr: Transitivite. Systeme experientiel de 1'enonce: la ressource pour construire notre experience du monde interne et externe au moyen de configurations de proces, participants et circonstances. Process types. Fr. Types de proces. Material process. Fr: Proces materiel Behavioural process. Fr: Proces comportemental Mental process. Fr: Proces mental Verbal process. Fr: Proces verbal Relational process. Fr: Proces relationnel: equatif (identifying); attributif (attributive) Existential process. Fr: Proces existentiel Transitive functions. Fr: Fonctions transitives. Fonctions experientielles specifiques a chaque proces. Par example, Acteur est une fonction associee au proces materiel, alors que Locuteur est une fonction associee au proces verbal. Actor. Fr: Acteur Goal. Fr: Objectif Senser. Fr: Ressenteur Phenomenon. Fr: Phenomene Sayer. Fr: Locuteur Verbiage. Fr: Acte de parole Addressee. Fr: Allocutaire Attribute. Fr: Attribut Carrier. Fr: Porteur Token. Fr: Symbole Value. Fr: Valeur Ergative functions. Fr: Fonctions ergatives. Fonctions experientielles generalisees a tous les proces.
ENGLISH-FRENCH GLOSSARY OF TERMS
201
Medium. Fr: Vehicule Agent. Fr: Agent Range. Fr: Etendue Beneficiary. Fr: Beneficiaire
Metafunction. Fr: Metafonction. Fonctions generalisees intrinseques au langage et que Ton observe dans son organisation. Halliday (1967/68) identifie trois metafonctions: 1'ideationnelle (ideational), l'interpersonnelle (interpersonal) et la textuelle (textual). De plus, la metafonction ideationnelle etablit une distinction entre 1'organisation experientielle (en constituants) et 1'organisation logique (interdependance). Interpersonal. Fr: Interpersonnelle. L'une des trois metafonctions: la langue en tant que ressource interactive/conversationnelle. Les ressources pour etablir et maintenir les relations entre locuteur et auditeur. Textual. Fr: Textuelle. L'une des trois metafonctions. Les ressources pour presenter 1'information interpersonnelle et ideationnelle comme texte en contexte. Ideational. Fr: Ideationnelle. L'une des trois metafonctions. Elle comprend deux modes pour representer et interpreter 1'experience: 1'experientiel et le logique. Alors qu'en dehors de la linguistique systemique le terme 'ideationnelle' se confond souvent avec celui de semantique; ici il est traite comme une metafonction et s'applique aussi bien a la semantique qu'a la grammaire. Experiential. Fr: Experientielle. L'un des modes d'organisation ideationnelle, en constituants. La ressource pour representer 1'experience. Cela correspond a ce que Ton appelle, dans d'autres theories, denotation, semantique, sens cognitif. Ces termes non-systemiques incluent parfois la metafonction logique, parfois non. Logical. Fr: Logique. L'une des sous-fonctions qui, avec 1'experientielle, constitue la metafonction ideationnelle. (Ce terme est bien connu en dehors de la linguistique systemique dans le sens de ce qui appartient a la logique.) Dans le modele systemique, la metafonction logique fournit les ressources pour creer des enonces complexes, par exemple. Group. Fr: Syntagme/groupe. Au niveau lexicogrammatical, rang inferieur a celui de 1'enonce. Theme. Fr: Theme. Fonction textuelle de 1'enonce: le point de depart de 1'enonce en tant que message. Le Theme etablit le contexte local pour chaque enonce. Rheme. Fr: Rheme. Fonction textuelle de 1'enonce. En general, le Rheme suit le Theme. Deictic. Fr: Deictique. Fonction dans la structure du syntagme nominal inseree comme realisation d'un choix dans le systeme DETERMINATION et
202
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
realisee par un determinant. Le terme 'deictique' est aussi utilise dans les travaux non systemiques, dans son sens general lie a celui de deixis, et non en reference a une fonction. Polarity. Fr: Polarite. Designation du systeme 'affirmatif/negatif. En dehors de la linguistique systemique on 1'appelle souvent simplement 'negation'. Modality. Fr: Modalite. Systeme interpersonnel divise entre le systeme de modalisation (parfois simplement appelle modalite: probabilite et usualite) et celui de modulation (obligation-inclination-capacite). Speech function. Fr: Fonction elocutive. Fonction semantique interpersonnelle (commande, question, declaration) realisee par les modes grammaticaux interpersonnels (imperatif, interrogatif, declaratif). extyemsopm Fr extension lun des termes danses le systeme types d expan SION (enonce complexe), les autres etant 'elaboration' et 'embellissement' (enhancement). Ce terme correspond plus ou moins a la notion traditionnelle de coordination. L'expansion d'un enonce se fait par une relation 'logico-semantique' additive, alternative ou adversative. Elaboration. Fr: Elaboration. Terme du syteme EXPANSION. II correspond a la categoric traditionnelle d'opposition, mais 'elaboration' inclut aussi les relations 'relatives non restrictives'. L'expansion d'un enonce se fait au moyen d'exemplification, de clarification et de repetition. Enhancement. Fr: Qualification. L'un des termes du systeme EXPANSION. II inclut ce que Ton appelle traditionnellement les 'circonstancielles', mais aussi leur equivalents paratactiques. Token. Fr: Symbole. Fonction dans le systeme de la transitivite que Ton trouve dans les enonces relationnels equatifs (identifying relational clauses). Value. Fr: Valeur. Fonction transitive que 1'on trouve dans les enonces relationnels equatifs. Network. Fr: Reseau. Un type relationnel d'organisation. Dans la theorie systemique, un reseau est principalement un reseau de systemes. System. Fr: Systeme. Un systeme est la categoric centrale pour representer toute organisation paradigmatique a tous les strata: phonologique, grammatical, ou semantique. II est constitue (i) d'une specification d'un choix entre deux termes ou plus, qui sont represented par des traits, et (ii) d'une condition d'entree specifiant quand le choix est disponible. La condition d'entree est un trait simple ou un trait complexe/compose; ces traits sont des termes dans les autres systemes. En raison de leur condition d'entree, les systemes forment des reseaux de systemes. Chaque terme dans un systeme peut avoir une ou plusieurs specifications de realisation (realization statements) qui lui sont associees. Les specifications de realisation indiquent des fragments de structure.
References
Barthes, R. (1967) Degree Zero & Elements of Semiology, trans. A. Lavers and C. Smith. London: Jonathan Cape. Bateman, J. A. (1989) 'Dynamic systemic-functional grammar: a new frontier', Word, 40 (1-2) 263-86. Battye, A. and Hintze, M.-A. (1992) The French Language Today. London: Routledge. Beckett, S. (1952) En Attendant Godot. Paris: Les Editions de Minuit. Behnstedt, P. (1973) Viens-tu? Est-ce que tu viens? Tu viens? Formen und Strukturen des direkten Fragesatzes im Franzosischen. Tubingen: Narr. Birmant, L. (1991) Le Pare National des Pyrenees Raconte aux Enfants. Tarbes: Nadine Bruneau-Hulin. Blanche-Benveniste, C., Bilger, M., Rouget, C. and van den Eynde, K. (1991) Toupee', in Le Francais Parle. Paris: Editions du centre national de la recherche scientifique, pp. 259-61. Bottinelli, L., Gerbaldi, M. and Gouguenheim, L. (1984) 'Decrire puis expliquer 1'Univers', Les Dossiers Speciaux du Grand Quid Illustre: Etoiles et Galaxies. Paris: Robert Laffont. Buzzati, D. (1968) La Fameuse Invasion des Ours en Sidle. Paris: Editions Stock. Byrne, L. S. R. and Churchill, E. L. (1986) A Comprehensive French Grammar. Oxford: Blackwell. Caffarel, A. (1992) 'Interacting between a generalized tense semantics and registerspecific semantic tense systems: a bi-stratal exploration of the semantics of French tense', Language Sciences, 14, (4) 385-418. Caffarel, A. (2004a) 'A metafunctional profile of French', in A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 77-137. Caffarel, A. (2004b) 'The construal of a second-order semiosis in Camus' L'Etranger, in D. Banks (ed.) Text and Texture. Paris: L'harmattan, pp. 537-70. Caffarel, A., Martin J. R., and Matthiessen, C. M. I. M. (2004) Language Typology: A Functional Perspective. Amsterdam: Benjamins. Camus, A. (1942) L'Etranger. Paris: Gallimard. Combettes, N. (1975) Pour une Linguistique Textuelle. Nancy: Academic de NancyMetz:crdp. Comrie, B. (1976) Aspect. London: Cambridge University Press. Coveney, A. (1990) 'Variation in interrogatives in spoken French: a preliminary report', in J. N. Green and W. Ayres-Bennett (eds) Variation and Change in French: Essays Presented to Rebecca Posner on the Occasion of her Sixtieth Birthday, New York: Routledge, pp. 116-33.
204
REFERENCES
Danes, F. (1964) 'A three-level approach to syntax' Travaux Linguistiques de Prague, 1,225-40. Danielou, A. (1975) 'Les fous de Dieu', Les Fous de Dieu: Contes Gangetiques. Paris: Buchet/Chastel, pp. 141-84. De Beauvoir, S. (1945) Les Bouches Inutiles. Paris: Gallimard. de Saussure, F. (1972) Cours de Linguistique Generale, ed. T. de Mauro. Paris: Payot. Delacorta (1985) Vida. Paris: Mazarine, des Forets, L.-R. (1947) LeBavard. Paris: Gallimard. Dixon, R. M. W. (1979) 'Ergativity', Language, 55, (1) 59-138. Dixon, R. M. W. (1994) Ergativity. Cambridge: Cambridge University Press. Eggins, S. (1994) An Introduction to Systemic Functional Linguistics. London: Pinter. Fawcett, R. P., van der Mije, A. and van Wissen, A. (1988) 'Towards a Systemic Flowchart Model for Discourse Structure', in R. P. Fawcett and D. J. Young (eds) New Developments in Systemic Linguistics, Vol. 2. London: Pinter, pp. 116-43. Firbas,J. (1987) 'On the delimitation of theme in functional sentence perspective', in R. Dirven and V. Friend (eds) Functionalism in Linguistics. Amsterdam: Benjamins, pp. 137-56. Fowler, R., Hodge, B., Kress, G. and Trew, T. (1979) Language and Control. London: Routledge and Kegan Paul. Fries, P. (1981) 'On the status of Rheme in English: arguments from discourse', Forum Linguisticum, 6, (1) 1-38. Fuller, G. (1995) 'Engaging Cultures: Negotiating Discourse in Popular Science', unpublished PhD thesis, Department of Linguistics, University of Sydney. Gadet, F. (1991) 'Le parle coule dans 1'ecrit: le traitement du detachement par les grammairiens du XXe siecle', LangueFrancaise No. 89. Paris: Larousse. GionoJ. (1974) LeChevalFou. Paris: Gallimard. Grimshaw, J. (1982) 'On the lexical representation of Romance reflexive clitics', in Joan Bresnan (ed.) The Mental Representation of Grammatical Relations. Cambridge, MA: MIT Press, pp. 87-147. Guillaume, G. (1929) Temps et Verbe. Theorie des Aspects, des Modes et des Temps, collection linguistique publiee par la societe de linguistique de Paris, tome XXVII. Paris: Champion. Guillaume, G. (1945) L'Architectonique du Temps dans les Langues Classiques. Copenhagen: Munskgaard. Guillaume, G. (1984 [1943]) Langage et Science du Langage (3rd edn). Paris: Librairie A. G. Nizet/Quebec: Presses de 1'universite Laval. Hagege, C. (1978) 'Du Theme au Theme en passant par le sujet. Pour un theorie cyclique', La Linguistique, 14, (2) 3-38. Hagege, C. (1982) La Structure des Langues. Que sais-je. Paris: Presses Universitaires de France. Hagege, C. (1984) L 'Homme de Paroles, folio essais. Paris: Fayard. Halliday, M. A. K. (1966) 'The concept of rank: a reply', Journal of Linguistics, 2 (1) 110-18. Halliday, M. A. K. (1967) 'Notes on transitivity and theme in English', Parts 1 & 2, Journal of Linguistics, 3, 37-81 and 199-244. Halliday, M. A. K, (1968) 'Notes on transitivity and theme in English', Part 3, Journal of Linguistics, 4, 179-215. Halliday, M. A. K. (1975) Learning How to Mean: Explorations in the Development of Language. London: Edward Arnold.
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH
205
Halliday, M. A. K. (1978) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, M. A. K. (1979) 'Modes of meaning and modes of expression: types of grammatical structure and their determination by different semantic functions', in D. J. Allerton, E. Carney and D. Holdcroft (eds) Function and Context in Linguistics Analysis. Cambridge: Cambridge University Press, pp. 57-79. Halliday, M. A. K. (1981 [1965]) 'Types of structure', in M. A. K. Halliday and J. R. Martin (eds) Readings in Systemic Linguistics. London: Batsford Academic and Educational, pp. 29-41. Halliday, M. A. K. (1984) 'Language as code and language as behaviour: a systemicfunctional interpretation of the nature and ontogenesis of dialogue', in R. P. Fawcett, M. A. K. Halliday, S. Lamb and A. Makkai (eds) The Semiotics of Language and Culture, Volume 1: Language as Social Semiotic, London: Pinter, pp. 3-35. Halliday, M. A. K. (1985a) An Introduction to Functional Grammar. London: Edward Arnold. Halliday, M. A. K. (1985b) Spoken and Written Language. Geelong: Deakin University Press. Halliday, M. A. K. (1992) 'How do you mean?', in M. Davies and L. Ravelli (eds) Advances in Systemic Linguistics: Recent Theory and Practice. London: Pinter, pp. 20-35. Halliday, M. A. K. (1994) An Introduction to Functional Grammar (2nd edn) London: Edward Arnold. Halliday, M. A. K. (1998) 'The notion of "context" in language education', in M. Ghadessy (ed.) Text and Context in Functional Linguistics. Amsterdam: Benjamins, pp. 1-24. Halliday, M. A. K. (2002) 'Computing meanings: some reflections on past experience and present prospects', in G. Huang and Z. Wang (eds) Discourse and Language Functions. Shanghai: Foreign Language Teaching and Research Press, pp. 3-25. Halliday, M. A. K. (2002 [1996]) 'On grammar and grammatics', in J. Webster (ed.) On Grammar. Vol. 1 in the Collected Works ofM. A. K. Halliday, London and New York: Continuum, pp. 384-417. Halliday, M. A. K. (2004) An Introduction to Functional Grammar (3rd edn), revised by C. M. I. M. Matthiessen. London: Edward Arnold. Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English. London: Longman. Halliday, M. A. K. and Hasan, R. (1985) Language, Context and Text: A Social Semiotic Perspective, Geelong: Deakin University Press. Halliday, M. A. K., and Martin, J.R. (1993) Writing Science: Literacy and Discursive Power. London: The Falmer Press. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience through Meaning: A Language-based Approach to Cognition. London: Cassell. Hasan, R. (1985a) 'Meaning, context and text: fifty years after Malinowski', in J. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse, Norwood, NJ:Ablex, pp. 16-50. Hasan, R. (1985b) Linguistics, Language and Verbal Art. Geelong: Deakin University Press. Hasan, R. (1987) 'The grammarian's dream: lexis as most delicate grammar', in M. A. K. Halliday and R. P. Fawcett (eds) New Developments in Systemic Linguistics: Theory qnd Description. London: Pinter, pp. 184—211. Hasan, R. (1989) 'Semantic variation and sociolinguistics', Australian Journal of Linguistics 9, 221-75.
206
REFERENCES
Hjelmslev, L. (1943) Omkring Sprogteoriens Grundloeaeggelse. Copenhagen: Munksgaard. Hopper, P. J., and Thompson, S. A. (1980) 'Transitivity in grammar and discourse', Language, 56, 251-399. Judge, A. and Healey, F. G. (1985) A Reference Grammar of Modern French. London: Edward Arnold. Lambrecht, K. (1988). 'Presentational cleft constructions in spoken French', with J. Haiman in Clause Combining in Grammar and Discourse, in J. Haiman and S. Thompson (eds) Clause Combining in Discourse and Grammar. Amsterdam: Benjamins, pp. 135-79. Leeman-Bouix, D. (1994) LesFautes deFranQaisExistent-elks'?Paris: Seuil. Lukin, A. (2005) 'Mapping media bias: a multidimensional affair', Australian Journalism Review, 27 (1) 139-55. Martin, J. R. (1985) 'Process and Text: two aspects of semiosis', inj. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse, Vol. I: selected theoretical papers from the 9th International Systemic Workshop. Norwood, NJ: Ablex, pp. 248-74. Martin, J. R. (1991) 'Intrinsic functionality: implications for contextual theory', Social Semiotics, 1 (1) 99-162. Martin, J. R. (1992) English Text. Amsterdam: Benjamins. Martin, J. R. (1994) 'Metalinguistic diversity: the case from case', in R. Hasan, C. Cloran and D. Butt (eds) Functional Descriptions: Theory in Practice. Amsterdam: Benjamins, pp. 323-74. Martin, J. R. (1995a) 'Logical meaning, interdependency and the linking particle na/-ng'm Tagalog', Functions of Language, 2 (2) 189-228. Martin, J. R. (1995b) 'Text and clause: fractal resonance', in T. A. van Dijk (ed.) Text. Berlin and New York: Mouton de Gruyter, pp. 5-42. Matthiessen, C. M. I. M. (1988) 'Representational issues in systemic-functional grammar', in J. D. Benson and W. S. Greaves (eds) Systemic-functional Approaches to Discourse. Norwood, NJ: Ablex, pp. 136-75. Matthiessen, C. M. I. M. (1990) 'Metafunctional complementarity and resonance in syntagmatic organization', mimeo, Department of Linguistics, University of Sydney. Matthiessen, C. M. I. M. (1991) 'Language on language: the grammar of semiosis', Social Semiotics, 1 (2) 69-111. Matthiessen, C. M. I. M. (1992) 'Interpreting the textual metafunction', in M. Davies and L. Ravelli (eds) Advances in Systemic Linguistics: Recent Theory and Practice. London: Pinter, pp. 37-81. Matthiessen, C. M. I. M. (1993) 'Register in the round', in M. Ghadessy (ed.) Register Analysis: Theory and Practice. London: Pinter, pp. 221-92. Matthiessen, C. M. I. M. (1994) 'THEME as an enabling resource in ideational 'knowledge', in M. Ghadessy (ed.) Thematic Development in English Texts. London: Pinter, pp. 20-54. Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems. Tokyo: International Language Sciences Publishers. Matthiessen, C. M. I. M. (2004) 'Descriptive motifs and generalizations', in A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 537-673. Matthiessen, C. M. I. M., and Nesbitt, C. (1996) 'On the idea of theory-neutral descriptions', in R. Hasan, C. Cloran and D. Butt (eds) Functional Descriptions: Theory in Practice. Amsterdam: Benjamins, pp. 39-85.
A SYSTEMIC FUNCTIONAL GRAMMAR OF FRENCH 207
Matthiessen, C. M. I. M., and Thompson, S. (1989) The structure of discourse and "subordination"', in J. Haiman and S. Thompson (eds) Clause Combining in Discourse and Grammar. Amsterdam: Benjamins, pp. 275-330. Matthiessen, C. M. I. M., Inui, K. and Teruya, K. (1994) 'Notes on a systemicfunctional interpretation of the grammar of Japanese', manuscript, Macquarie University and the University of NSW. Michaud, G. and Kimmel, A. (1990) LeNouveau Guide de France. Paris: Hachette. Morel, M.-A. (1992) 'L'opposition theme/rheme dans la structuration des dialogues oraux', Journal of French Language Studies, 2, (1) 61-74. Nesbitt, C. and Plum, G. (1988) 'Probabilities in a systemic functional grammar: the clause complex in English', in R. P. Fawcett and D.J. Young (eds) New Developments in Systemic Linguistics, Vol. 2. London: Pinter, pp. 1-38. O'Donnell, M. (1990) 'A dynamic model of exchange', Word, 42 (3) 293-327. Painter, C. (2001) Learning through Language in Early Childhood. London: Continuum. Pike, K. L., and Pike, E. G. (1982) Grammatical Analysis (2nd edn). Arlington, TX: Summer Institute of Linguistics. Plank, F. (ed.) (1979) Ergativity: Towards a Theory of Grammatical Relations. London: Academic Press. Pettier, B. (1992) Theorie et Analyse en Linguistique. Paris: Hachette. Ravelli, L. (1991) 'A dynamic approach to grammar', in F. Christie (ed.) Literacy in Social Processes. Papers from the Inaugural Australian Systemic Functional Linguistics Conference held at Deakin University, January 1990. Darwin: Centre for Studies of Language in Education, pp. 159-68. Ravelli, L. (1995) 'A dynamic perspective: implications for metafunctional interaction and an understanding of rheme', in R. Hasan and P. H. Fries (eds) On Subject and Theme: A Discourse Functional Perspective. Amsterdam: Benjamins, pp. 187-234. Robbe-Grillet, A. (1957) La Jalousie. Paris: Les Editions De Minuit. Robbe-Grillet, A. (1966) 'La plage', trans. B. Wright, in P. Lyon (ed.) Parallel Text: French Short Stories 1. London: Penguin, pp. 11-23. Rothemberg, M. (1989) 'Quelques moyens syntaxiques de rhematisation et de thematisation en Francais', Bulletin de la Societe de linguistique de Paris, 84, 143-61. Rousset, H. (1977) Arts de Coree. Paris: Bibliotheque des Arts. Saint-Soline, C. (1972). 'Le tabac vert', trans P. Newmark, in S. Lee (ed.) Parallel Text: French Short Stones 2. London: Penguin, pp. 13—47. Sartre,J.-P. (1939) 'Le mur', Lemur. Paris: Gallimard, pp. 10-38. Sefton, P. (1995) 'State-potentials and Social Subjects in Systemic-functional Theory: Towards a Computational Socio-semiotics', unpublished PhD thesis, Department of Linguistics, University of Sydney. Teruya, K. (2004) 'Metafunctional profile of the grammar of Japanese', in A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: A Functional Perspective. Amsterdam: Benjamins, pp. 185-254. Teruya, K. (forthcoming) Systemic Functional Grammar of Japanese. London and New York: Continuum. Teruya, K., Akerejola, E., Andersen, T., Caffarel, A., Lavid, J., Matthiessen, C. M. I. M., Petersen, U. H., Patpong, P. and Smedegaard, F. (forthcoming) 'Typology of MOOD: a text-based and system-based functional view', in R. Hasan and C. M. I. M. Matthiessen (eds) Continuing Discourse on Language: A Functional Perspective, Vol. 2. London: Equinox. Tesniere, L. (1959) Elements de Syntaxe Structural^. Paris: Klincksieck.
208
REFERENCES
Thomas, H. (1966) 'L'offensive', trans. K. Thomson, in P. Lyon (ed.) Parallel Text: French Short Stories 1. London: Penguin, pp. 111-33. Ventola, E. (1987) The Structure of Social Interaction: A Systemic Approach to the Semiotics of Service Encounters. London: Pinter. Vian, B. (1972) 'Les fourmis', trans R. Corbett, in S. Lee (ed.) Parallel Text: French Short Stories 2. London: Penguin, pp. 49-75. Wu, C. (2000) 'Modelling Linguistic Resources: A Systemic Functional Approach', unpublished PhD thesis, Dept of Linguistics, Macquarie University. Zribi-Hertz, A. (1987) 'La reflexivite ergative en francais moderne', Le Francais Moderne, 23-53.
Index
This index follows the usual SF spelling conventions, using all capitals for system names (e.g. AGENCY) and initial capitals for function labels (e.g. Actor). Page numbers in bold italic indicate that a whole chapter, section or subsection is devoted to the index term (or a closely related topic); page numbers in bold indicate that a definition or description of the term is to be found on that page; page numbers in italic indicate that there is an entry in the glossary for the term. ability (MODALITY) 146-9 absolute Theme 127, 128, 138, 141-2, 158, 166, 168, 171-3, 174, 177-9 see also reprise Theme absolutive see case marking accompaniment (circumstance) 106-7 accusative see case marking active voice see VOICE Actor 65, 73-4, 200 addition (logico-semantic relation) 27, 30, 31-2, 34-5 additive see addition Addressee 65, 77-80, 200 Adjuncts 124-38 modal 134, 146-9, 171, 174 negative 128, 134 adverbial clause 24 see also enhancement adversative see addition AGENCY 64-9, 88-94 and PROCESS TYPE 99-100 and VOICE CULMINATION 100-3 Agent 60, 64, 88-103, 106-8, 201 analytic causative see effective clauses and VOICE apposition 22, 23, 24, 30 Assigner 99 Attribute 65, 80-1, 82-4, 200 attributive clauses see relational process clauses
Attributor 98-9 auxiliary verbs 83-5 Behaver 65, 74-5 behavioural process clauses 65, 70-3, 74-5, 95, 200 being process clauses 65, 80-5, 97-9 Beneficiary 73-4, 106-8, 201 Carrier 65, 80-1, 82-4, 200 case marking 58-9, 105, 129, 155-7 causal-conditional (logico-semantic relation) 27, 30, 33-7 causation 60 see also AGENCY cause (circumstance) 106-7 see also causal-conditional Circumstance types 30, 58, 103-8 clarification (logico-semantic relation) 27, 30-31 clause 1, 15-16, 198 structure, multilayered 5, 6-8, 58-9 types (network diagram) 42 see also names of clause types (embedded, finite, relative) clause complex 198 CLAUSE COMPLEX 14, 18, 20-56 fractal nature of 3 probabilistic nature of 51-6 traditional terms for categories 24 across text types 45-51
210
INDEX
clause reversibility 38, 42-4 Client 73-4
clitics 60, 61, 88, 92-3,118, 124-38, 153-8 ordering of clitic pronouns 158-63 relationship to Negotiator and cohesion 131-6 relationship to nuclear transitivity 19, 103-5 cognitive mental processes 78, 79 cohesion 126, 131-6 see also REFERENCE and ELLIPSIS command see speech function comparison see manner Complement 124-38 complexity of spoken vs written language 45-51 concession (circumstance) 106-7 see also causal-conditional condition (circumstance) 106-7 see also causal-conditional congruence see grammatical metaphor conjunctives 171, 174 see also expansion and textual Theme consequence see causal-conditional constituency structures 14-15 context 16-18, 199 continuatives 171 coordination 22, 23, 24 dative see case marking declarative 140-2 see also MOOD Deictic 201 delicacy 16, 200 dependency, dependent clause see hypotaxis desiderative mental processes 40, 78, 79-80 dialogue see exchange direct speech and thought 22, 23, 24, 38, 78 see also projection discourse analysis 1-3 see also register, second-order meaning and text types discourse markers 171 discourse semantics see semantic system and semantics
doing process clauses 65, 70-5, 94-5 downranked clause see embedded clause and rankshifted effective clauses 60, 88-103 elaboration (logico-semantic relation)
23, 30-1, 202 ELLIPSIS 126, 132, 133, 138, 158 embedded (downranked) clauses 23, 29, 175 emotive mental processes 76-7, 78-80 enhancement (logico-semantic relation) 23, 33, 202 entry condition 12 ergative functions 64, 200 model of transitivity 57-119, 88-94 generalizing across process types
94-103 and register 61 patterns in texts 66-9 see also case marking exchange (of meaning) 120-64 French clause as a move in 121-4 of goods-and-services see proposals of information see propositions preliminary exploration of 122—4 exclamative 140-2 see also MOOD exemplification (logico-semantic relation) 27, 30-31 Existent 65, 81-2 existential process clauses 65, 80-1, 81-2, 84, 97, 200 existential particle 81 expansion (logico-semantic relation) 23,
30-7 conjunctive markers 34-5, 52-4 interaction with parataxis and hypotaxis 25 and the mode subjonctif 34-7 summary of options 34-5 experiential metafunction 14, 20,
57-119, 201 experiential Theme see topical Theme exposition (logico-semantic relation) 27, 30-1 extension (logico-semantic relation) 23, 31-2, 202
INDEX feature (in system network) 13 field (of con text) 17,199 Finite (interpersonal function) 123-46, 149-52 finite and non-finite clauses 27, 31, 32, 33
211
intonation 158, 166 and MOOD 139, 142, 146, 151 and speech function 123, 124, 125 irrealis 34-7 juxtaposition 22, 23, 24, 30, 31
genitive see case marking gerund 32 Given 100, 127, 128, 136, 158, 166-7, 181 Goal 60, 64, 73-4, 200 grammatical intricacy 45-51 grammatical metaphor 10, 39-42, 46-51, 59, 138, 147, 148, 150 group 16, 207 higher-order meanings/semiosis see second-order meanings hypotaxis 22-23 criteria for distinguishing from parataxis 42-5 and elaboration 31 and enhancement 33 and extension 32 and projection 38 and the mode subjonctif 39—42 idea (projection type) 24, 27, 38, 77 ideational metafunction 14, 20-119, 201 identifying clauses see relational process clauses imperative 136-7, 145-6, 182-3 inclination (MODALITY) 36, 146-9 indicative see MOOD indirect speech and thought 23, 24, 38, 77 see also projection INFORMATION FOCUS 166-8 see also Given and New informative (MOOD selection) 140-2 instantiation 1-2, 10-12, 199 and stratification 8-9 INTERACTANT TYPE 129, 153-63 INTERDEPENDENCE (TAXIS) 20-56 interdependency structures 14-15 interpersonal metafunction 13, 120-64, 201 ordering of interpersonal functions 134-5, 146 interpersonal Theme 171, 174, 182, 183 interrogative 124-41, 142-4, 181-2
levels of language see stratification lexical density 45-8 lexicogrammar 198 lexicogrammatical systems delicacy 16 distribution by rank and metafunction 18 rank-scale organization 15-16 relationship to phonology 5-6 relationship to semantic systems 4-6, 10-12 location (circumstance) 104-8 locution (projection type) 24, 27, 38, 77 logical metafunction 14, 20-56, 201 LOGICO-SEMANTIC RELATION 20-56 logogenesis 9 macrophenomenon 76 manner circumstance 106-7 logico-semantic relation 27, 30, 33-5 marked Theme 168-9, 171, 174-7, 183 material process clauses 64, 65, 70-3, 73-4, 94-5, 200 matter (circumstance) 104—8 meaning relationship with grammar 4—6, 10—12 second-order see second-order meaning see also semantics and semantic system means see manner Medium 60, 61, 64, 88, 94, 105, 201 mental process clauses 64, 65, 75-80, 95-6, 200 metafunction 3, 5, 13-15, 201 realizations in structure 6-8, 14 see also individual metafunctions (experiential, ideational, interpersonal, logical, textual) metaphenomenon 75-80 metaphor see grammatical metaphor metereological process clauses 70-3, 75, 95
212
INDEX
method of development (textual) 166, 183-95 middle clauses 60-3, 88-103 modal (Negotiatery) structure 126, 124-38 modal verbs 134, 146-9, 152 MODALITY 138, 146-9, 202 relationship to expansion 34—7 relationship to MOOD systems 39, 44-5 relationship to projection 39-42,79-80 modalization see MODALITY mode (of context) 17, 199 mode subjonctif 147 relationship to expansion 34-7 relationship to hypotactic projection 39-42 modulation see MODALITY Mood-interrogator 125,130, 132, 138 Mood-markers 141-3, 146, 171 Mood-tag 149 MOOD 14, 18, 120-1, 138-146, 199 relationship to projection 41-2 relationship to taxis 44-5 morpheme 1, 15, 16 multivariate structure 15, 16 negation see POLARITY and Polaritymarker Negotiator 123-46, 149-52 Negotiator-Remainder structure 124-38 and pronominals 153-8 relationship to clitics and cohesion 131-6 Negotiatory (modal) structure 126 network 12, 202 representational conventions of 13, 24-25, 65, 73,143 New 100, 127, 128,136, 141,158, 166-7, 181 nominative see case marking non-congruence see grammatical metaphor non-finite clause see finite and non-finite clauses non-reflexive clauses see middle clauses obligation (MODALITY) 35, 39-42, 146-9
oblique cases see case marking and pronominals offer see speech function parataxis 22-23 criteria for distinguishing from hypotaxis 42-5 and elaboration 30-1 and enhancement 33 and extension 31 and projection 38 Participant types 65 nuclearity and peripherality of 58, 60, 103-8 patterns in text 70 passive voice see VOICE systems perceptive mental processes 71, 78, 79 Phenomenon 65, 74, 75-80, 200 phylogenesis 9 POLARITY 148, 202 negative 128-9, 135 Polarity-marker 126, 128, 133, 136, 138 predicated Theme 174-6 Predicator 124-46, 149-52 probability (MODALITY) 34-7, 39-42, 146-9 PROCESS TYPE 64-9 and AGENCY 99-100 Process types 69-88, 200 criteria for distinguishing 86 see also individual Process type names (behavioural, being, doing, existential, material, mental, projecting, relational, verbal) projecting process clauses 65, 75—80, 95-7 projection (logico-semantic relation) 24, 37-42 interaction with parataxis and hypotaxis 25 see also mental process clauses, projecting process clauses and verbal process clauses pronominals (pronouns) and case marking 58-9, 155-7 clitic, ordering of 129, 158-63 and the Negotiator-Remainder structure 153-8 prominent and non-prominent 153-63, 176, 181
INDEX
proposals 39, 41, 120, 136-8, 146, 151 propositions 39, 41, 120, 124-31, 146, 151 prosody, grammatical 123, 142-4 purpose see causal-conditional Qu-element 130-1, 140-4 Qu-expression 140 questions 125-31 see also speech function and MOOD quoted speech and thought see direct speech and thought Range 64, 73-4, 88-103, 105-6, 201 rank 200 rank scale 15-16 rankshifted 200 see also embedded clauses readiness (MODALITY) 146-9 realis 34-7 realization 5, 199 of metafunctions in structure 14 Recipient 73-4 REFERENCE 126, 132, 158 reference chains 161-3 reflexive clauses see middle clauses verbs see clitics see also VOICE register 2, 10-12, 199 analyses biography 188-192 geographical description 184-8 narrative 52-6, 111-17, 180 news report 192-5 scientific 47-8, 108-10 and context 77-7* spoken see exchange and spoken language relational process clauses 64, 65, 80—1, 82-5, 98-9, 200 relative clause defining (restrictive) 23 non-defining (non-restrictive) 22, 24, 31 Remainder 124-31, 138, 153-8 reported speech and thought see indirect speech and thought reprise Theme 128, 158 see also absolute Theme
213
Rheme 130-1, 168, 201 Sayer 65, 77-80, 200 second-order meanings 21-2 and logical patterns 45-51 and Theme patterns 192—5 and transitivity patterns 108-17 semantic system metafunctional organization 13-15 relationship to lexicogrammatical system 10-12 semantics 199 Senser 65, 75-80, 200 spatial (circumstance) 106-7 speech function 39, 44, 82, 120-38, 202 see also interpersonal metafunction and MOOD spoken language 20-1, 45-51, 63, 130, 139,166,179 see also exchange (of meaning) statement see speech function strata/stratum 199 stratification 1, 6, 8-10 and instantiation 8-9 Subject 123-146 definition criteria 151-3 recognition criteria 149—51 subjunctive mode see mode subjonctif subordination 22-3, 24 synthetic causative see effective clauses and VOICE system 202 dynamic and synoptic views of 11-12 network see network and structure 12 see also lexicogrammatical systems and semantic system systems, French location in theoretical map 18-19 see also individual system names (TRANSITIVITY, MOOD, etc.) tag see Mood-tag Target 79, 97 TAXIS (INTERDEPENDENCE) 20-56 temporal (logico-semantic relation) 27, 30, 33-7 tenor (of context) 17, 199 tense 70-1, 83-5, 127 text 199
214
INDEX
definition of 16
organization (textual) 183-95 text types 2 grammatical patterns across 45-51, 66-9, 70,161-3,108-17,183-95 see also register textual metafunction 14,165-97, 201 textual Theme 171, 174, 183 Theme (textual function) 168, 201 across MOOD types 181-3 across text types 183-95 functions of 169-70 in systemic functional theory 168—83 metafunctions and 170-4 THEME 14, 18, 165-97 French 174-81 theory (systemic functional) ix-x, 1-3 overview of dimensions 6-16 relationship to grammatical description 2, 4-19 Token 65, 80-1, 82-4, 202 topical Theme 168,174 transitive functions 64, 65, 200 model of transitivity 57-119 verbs 60-1
transitivity 200 ergative model of 57-119, 88-94 models of 64—9 patterns across text types 66-9, 70, 108-17 transitive model of 57-119 TRANSITIVITY 14,18, 57-119 nuclearity and peripherality of elements 103-8 realizations in the verbal group 57-8 translation 178-9 univariate structure 15, 16, 21 unmarked Theme see marked Theme usuality (MODALITY) 146-9 Value 65, 80-1,82-4, 202 variation (logico-semantic relation) 27, 30,31-2, 34-5 verbal process clauses 65, 75-80, 97, 200 Verbiage 65, 77-80, 200 viewpoint (circumstance) 106-7 Vocative 127,138,171 VOICE 88-9 in middle and effective clauses 100-3