Text,
Discourse,
and
Process
Toward
a Multidisciplinary
Science
of Texts
ROBERT
DE BEAUGRANDE
University
of Florida
Re-reading
one’s earliest “mature” work after 25 years is not without risk. I could
list many in my field who would hardly recognise theirs from their current
standpoint. But this book is one I might facetiously call “the blockbuster
that started it all” — at least for me.
As far as I knew, no book like it had been published before, and I was fortunate
indeed to have the full support of Walter Johnson at his newly founded Ablex in
quest of ‘new ‘ideas’.
Most
of the ideas propounded here were related to, or anticipated, those more
developed in my later books and papers: the triple vision of discourse as
linguistic, cognitive and social; the insistence on functionalism over formalism
and thus on textuality over grammaticality; the search for inter- or
multi-disciplinary integration; the concern for social and educational
usefulness; the preference for authentic (and if possible, interesting) texts
over the then oh-so-fashionable ‘John-and-Mary’ sentences… However, the
later ones are certainly not re-hashes of the same materials, and have always
used other data than those analysed here. Indeed, the data analyses in Chs.
VI-VIII remain unsurpassed in all my later work.
True,
the reliance on network graphs may appear excessive in 2004, but I (and people
who looked at mine, as different as Marshal McLuhan and Bob Simmons,) found them
useful as a non-linear (and non sentence-based!) representation of meaning and
topic, as did many working in ‘artificial intelligence’.
This book was composed (and “pasted” literally, with glue) by means of an IBM Selectric, where you had to change balls for italics or bold, so I opted mainly for UPPER CASE to get emphasis; I caught plenty of flak from stodgy reviewers for that. The graphics were done manually with the aid of a ruler for lines and paper clips for circles. Not all of them scanned very well, but if you increase image size on your computer, they will be reasonably legible. Footnotes had to be moved next to where they are marked, since I have no ‘foots’. I have made mostly just minor changes in the wording, and removed a chunk at the end of Chapter IV (about a TV-addicted gorilla) which was mysteriously stuck in from the 1981 Introduction to Text Linguistics.
This completes my serious of complete book uploadings.
Contents
Acknowledgements
Evolution
and Development
1.
Systems and Models
2.
Levels in Models of Language
3.
Text Versus Sentence
4.
Textuality
5.
Textual Competence
6.
Text Utilization as Model-Building
7.
Overview of the Discussion
1.
Transformational Sentence Grammars
2.
Sequencing Operations
1.
Meaning and Philosophy
2.
Meaning as Feature Clusters
3.
Meaning as Process
4.
Building the Text-World Model
1.
Modifying Information Theory
2.
Apperceptual Informativity
3.
Informativity within the Sentence
4.
A Newspaper Article
1.
Motives for Efficiency
2.
Recurrence
3.
Definiteness
4.
Co-Reference via Pro-Forms
5.
Exophoric Reference
6.
Ellipsis
7.
Junction
VI. Frames, Schemes, and Plans
1.
Global Perspectives on Knowledge
2.
Frame Attachment
3.
Schema Attachment
4.
Plan Attachment
VII.
Further Issues in Text Processes
1.
Text Types
2.
The Production of Texts
3.
Recalling Textual Content
VIII.
Conversation and Narration
1.
Conversation
2.
Narration
IX.
Applications for a Science of Texts
.
1.
The Educational Enterprise
2.
Traditional Grammar Versus Applied Linguistics
3.
The Teaching of Reading
4.
The Teaching of Writing
5.
The Teaching of Foreign Languages
6.
Translation Studies
7.
Literary Studies
8.
A Final Word
Acknowledgments
This
book has demonstrated, to my own satisfaction at least, the truth of my argument
in Chapter VII.2.7 that text production is a process with no built-in point of
completion. Sooner or later, the text producer is satisfied and terminates the
process. I cannot enumerate the total number of versions the book, or parts of
it, underwent in almost three years of constant work. But I can express my
gratitude to those whose expertise, advice, and assistance guided my travels
through this vast and often poorly mapped territory. I profited immensely from
the chance to discuss various issues with Alan Baddeley, Bill Brewer, Jaime
Carbonnel Jr., Wallace Chafe, Phil Cohen, Nick Colby, Max Cresswell, Teun van
Dijk, Wolfgang Dressier, Charles Fillmore, Ken and Yetta Goodman, Paul Grice,
Norbert Groeben, Peter Hartmann, Roland Harweg, Don Hirsch, Hans Hörmann,
Wolfgang Iser, Waiter Kintsch, Waiter A. Koch, Bill Mann, Jim Meehan, Dieter
Metzing, Bonnic Meyer, Maria Nowakowska, Barbara Partee, Dany Paul, János Petöfi,
Wolfgang Prinz, Hannes Rieser, Roland Posner, Gert Rickheit, Siegfried J.
Schmidt, Bob Shafer, Bob Simmons, Berhard Sowarka, Paul Weingartner, Harald
Weinrich, Manfred Wettler, and Bill Woods. Professors Dressier, Hartmann, and
Petöfi studied and commented on the entire penultimate version I circulated as
a mimeograph. Special thanks are due to Mike Smith for bitching at me to replace
my old numbering with mnemonic labels for my networks; and to Genevieve Miller
and Zofia Solczak for discovering and eradicating hosts of typographical errors
and inconsistencies in the manuscript (my own processing resources do not extend
to accurate typing). Vivian Felix prepared the index of names. My chairpersons
Dave Benseler of Ohio State and Ward Hellstrom and Mel New of the University of
Florida were very considerate in not loading my teaching schedule. And finally,
I learned immeasurably from the discussions following my guest lectures at the
departments and institutes of Linguistics, Psychology, Communication Science,
Education, and Computer Science in the Universities of Arizona, Berlin,
Bielefeld, Bochum, Colorado, Florida, Munich, Saarbrücken, Texas, Trier, and
Vienna.
0.
Evolution and Development
1.
Late in 1976, I embarked on the production of a new “introduction to text
linguistics” in co-operation with Wolfgang Dressler, whose (1972a) Introduction
had been well received. I gradually came to appreciate the peculiar nature of
the task: introducing people to a field which had not in fact been constituted
and consolidated as a field at all. Professor Dressler’s solution had been to
extend the usual linguistic methodology to the domain of texts. But this
approach seems too narrow from the perspective of 1980. It is now fairly clear
that we cannot treat texts simply as units larger than sentences, or as
sequences of sentences. The prime characteristic of texts is rather their
occurrence in communication, and they might consist of a single word, a
sentence, a set of fragments, or a mixture of surface structures. It follows
that extending sentence studies to texts must miss a number of vital issues and
raise serious empirical problems (cf. Beaugrande 1979k).
2.
The present volume is the outcome of my attempts to define and lay out the field
of text studies from the standpoint of human activities. I have essayed to
integrate relevant research from such disciplines as cognitive psychology,
sociology of language, and computer science (with its offshoot domain of
artificial intelligence). The requirement of such a large-scale synthesis may
have led me to make some unexpected uses of available research.1 [I
use “cf.” to cover many instances of citing the approximate ideas of
research contributions.] I must take full responsibility for the proposals and
conclusions drawn here.
3.
I can provide here only a bare outline of the evolution of text linguistics (for
more details: cf. Petöfi 1971a; Dressler 1972a; Hartmann 1972, 1975; Fries
1972; Schmidt 1973; Dressler & Schmidt [eds.] 1973; Kalimeyer, Klein,
Meyer-Hermann, Netzer, & Siebert, 1974; Harweg 1974; Gülich & Raible
1977; Rieser 1978; Beaugrande & Dressier 1980). I would identify three main
phases of the field with indistinct chronological boundaries (for a different
division, see Rieser 1976). In the earliest phase, lasting into the later
1960’s, we find only a few astute proposals that the text or the discourse
ought to be the main object of linguistic study (e.g. Ingarden 1931; Bühler
1934; Hjelmslev 1943; Harris 1952; Pike 1954;2
[2. Pike (1954) is now a part of Pike (1967).]
Firth 1957a; Coseriu 1955-56; Uldall 1957; Karlsen 1959; Slama-Cazacu
1961; Hartmann 1964; Weinrich 1966a). These arguments did not affect the
mainstream of conventional linguistics, doubtless because the available
methodologies pointed in contrary directions.The preoccupation with minimal
units or isolated sentences naturally distracts from the study of whole texts.
4.
Around 1968, linguists working mostly independently of each other converged on
the notion of “linguistics beyond the sentence” (e.g. Heidolph 1966; Pike
1967; Crymes 1968; Dik 1968; Harweg 1968a; Hasan 1968; Palek 1968; Isenberg
1971; Koeh 1971).3 [Isenberg (1971) is a revision of a paper
originally published in Replik 2, 1968. Koch (1971) is a habilitation
dissertation which was essentially complete as early as 1966.]
Attention focused on issues that could be stated in terms of sentence
linguistics, but not satisfactorily resolved. The tendency was naturally to look
at texts as sentence sequences (e.g. Waterhouse 1963; Harper 1965; Heidolph
1966; Wheeler 1967; Harper & Su 1969; Isenberg 1971). However, it has been
pointed out later on that this perspective allows us to see only a fraction of
the interesting properties of texts (cf. van Dijk 1972a: v; Harweg 1974: 100f.;
Kintsch 1974: 79; Weinrich 1976: 148; van Dijk 1977a: 3). The greatest obstacle
is that the unity of the text is left obscure.
5.
The year 1968 also witnessed the rise of dissension among representa-tives of
the then dominant “transformational” paradigm. It became evident that even
the restricted range of issues under consideration could not be conclusively
encompassed with the prevailing methodology. The counter-movements of “case
grammar” (Fillmore 1968) and “generative semantics” (Lakoff 1968a, 1968b,
1968c; McCawley 1968a, 1968b) attested to the dissatisfaction with the usual
treatment of meaning in grammar. But the basic tenets were retained. As Howard
Maclay (1971: 178) remarks, the whole debate was carried on in terms of the
general assumptions of Chomskian linguistics. The point of contention was the
“autonomy” of syntax and its corollary of a syntactic “deep structure”
(cf. 11.1.17). But many other key issues, such as the expansion of study from
sentences to texts, were hardly raised. For this reason, the situation has
remained unsettled. A growing awareness of the inadequacies of transformational
grammar has not led to anything but
continual revisions, which try to preserve as much of the old theory as possible
(e.g. the “extended standard theory”). Rumblings of discontent have become
too noticeable to be overlooked. Robert P. Stockwell (1977: 196) concludes a
book on transformational syntax with the ominous observation that “scholars
will cling tenaciously to an explanation, or a principle, or a ‘law’ that
they know to be wrong, because they do not have in hand an alternative
explanation.”
6.
The year 1972 ushered in a new stage of research toward alternative theories of
language rather than toward revisions of older theories. The new works were
manifestly critical of the foundations of sentence grammars, leading to
proposals for new departures (cf. Petöfi 1971a; van Dijk 1972a; Dressler 1972a;
Gindin 1972; Kuno 1972; Schmidt 1973). Sociologists voiced their opposition to
the context-free abstractness of older methods, and pointed out the importance
of social interaction in language groups (cf. Gumperz & Hymes [eds.] 1972;
Labov 1972a, 1972b). Computer scientists faced the exigencies of simulating
human language processes on the computer (cf. Woods 1970; Simmons & Slocum
1971; Charniak 1972; Collins & Quillian 1972; Winograd 1972; Schank & K.
Colby [eds.] 1973). Psychologists embarked on studies of memory (cf. Kintsch
1972; Tulving & Donaldson [eds.] 1972; Kintsch & Keenan 1973) that would
eventually lead to concern for texts (cf. Kintsch 1974; Frederiksen 1975; Meyer
1975). This interdisciplinary demand for theories and models has been a major
impetus in the development of text linguistics. It is clear that these
disciplines want to do more than describe the structures of sentences; they are
concerned with the processes by which language is utilized by human beings. My
own discussions in this volume are intended to reflect this trend.
7.
It would be distracting to offer a historical survey of individual projects and
proposals, because it would remain unclear how all these might fit together. I
recommend consulting surveys of this kind, however. For text linguistics proper,
Dressler (1972a), Fries (1972), Schmidt (1973), Dressier & Schmidt (eds.)
(1973), Kallmeyer et al. (1974), Harweg (1974, 1978), Hartmann (1975), Coulthard
(1977), Gülich and Raible (1977), Jones (1977), Dressler (ed.) (1978), Gindin
(1978), Grosse (1978), Kuno (1978), Nöth (1978), Rieser (1978), Beaugrande
(1980a), and Beaugrande and Dressler (1980) can be consulted. For the psychology
of language, I recommend Kintsch (1974, 1977a), Hörmann (1976, 1977), G. Loftus
and E. Loftus (1976), H. Clark and E. Clark (1977), Freedle (ed.) (1977, 1979),
Rosenberg (ed.) (1977). On language sociology, see Gumperz and Hymnes (eds.)
(1972), Dittmar (1976), and Viereck,(ed.) (1976). On computer simulation of
natural language processing, I recommend Schank and Colby (eds.) (1973), Minsky
and Papert (1974), Bobrow and Collins (eds.) (1975), Charniak and Wilks (eds.)
(1976), Goldstein and Papert (1977), Rumelhart (1977a), Wilks (1977b), Winston
(1977), Walker (ed.) (1978), Findler (ed.) (1979), and the proceedings of
TINLAP-2. On philosophy and language, one can consult Cresswell (1973),
Grewendorf and Meggle (eds.) (1974), Rittenauer (ed.) (1974), and van Dijk (ed.)
(1976).
8.
In order to encompass the field at all, I was obliged to postpone or exclude
some major research trends. I did not deal with the elaborate methodology of
Hans Glinz (1952, 1973). I could not determine the role of
“deconstructionist” work as exemplified by Jacques Derrida (1967a, 1967b,
1972, 1974) (a critical survey is found in Hempfer 1976). I could not go into
the “universal grammar” derived from the writings of the late Richard
Montague (1974) (an outline is presented in Löbner 1976).
9.
Two further deficiencies must be mentioned. First, I restricted my explorations
to English to avoid creating difficulties for readers who are not linguists.
Many intriguing aspects of textuality in other languages, some of which are
totally different from English in their organization, are discussed in Grimes
(1975), Li (ed.) (1976), Longacre (1976), and Grimes (ed.) (1978). Second, I
have not dealt in any concerted way with the undeniably crucial role of text
intonation (cf. Halliday 1967c; Lehiste 1970; Brazil 1975). I surmise that text
intonation would require the introduction of at least two further levels of
networks. One would reflect the progression of intonation curves (cf. Halliday
1967c; Brazil 1975; Takefuta 1975). The other would capture the stress markings
that arise from the need to show priorities of knowledge or contrasts (cf. Chafe
1976). I hope to return to these issues in later work, but for the time being, I
have been experimenting and analysing mostly with regard to written texts.
10.
Today, text research is in progress literally all around the world. It is
prominently represented in the appropriate faculties and departments of the
Universities of Bielefeld, Konstanz, Amsterdam, and Bochum, with further chairs
elsewhere (e.g. West Berlin, Abo Academy). Extensive contributions to
“discourse analysis” have been forthcoming from the Cornell group directed
by Joseph Grimes, the Lancaster group headed by John Sinclair, and the Summer
Institute of Linguistics guided by Kenneth Pike and Robert Longacre. Text
processing is being studied in the psychology departments of such institutions
as the Universities of Colorado, Stanford, California (La Jolla), Illinois,
Yale, Cambridge, and Carnegie-Mellon; in the computer science departments of
such institutions as the Universities of Texas, Toronto, Massachusetts Institute
of Technology, Yale, Carnegie-Mellon, Rochester, Southern California, California
(Berkeley), California (Los Angeles), Illinois, Maryland, and Stanford. The work
carried on at the Stanford Research Institute, the Xerox Palo Alto Research
Center, the Center for the Study of Reading, and Bolt, Beranek, and Newman Inc.
has been extremely enlightening.
11.
The diversities among these various schools, and those between their work and
the older methods of linguistics, should not imply that we do not all share a
commitment to the joint scientific enterprise of exploring human language. I
would look to the conviction eminently expressed by Roman Jakobson (1973: 12):
Like
any age of innovative experimentation, the present stage of reflections on
language has been marked by intensive contentions and tumultuous controversies.
Yet a careful, unprejudiced examination of all these sectarian creeds and
vehement polemics reveals an essentially monolithic whole behind the striking
divergencies in terms, slogans, and technical contrivances.
ROBERT
DE BEAUGRANDE
University
of Florida