I

Wittgenstein writes in a letter to Ludwig von Ficker (von Wright 1969:35): “My work consists of two parts: the one presented here plus all that I have not written. And it is precisely the second part that is the important one.” Could this quotation be used successfully in the application of the words TEXT, DISCOURSE, and NARRATIVE, which are current in the research of several disciplines including semiotics, linguistics, and literature? Confronted by this question we immediately appeal to Webster’s Third (1964) for possible clues in the disambiguation of these concepts:

DISCOURSE n 1 archaic a: the act, power, or faculty of thinking consecutively and logically: the process of proceeding from one judgment to another in logical sequence: the reasoning faculty: RATIONALITY b: the capacity of proceeding in an orderly and necessary sequence 2 obs : progression of course esp. of events’: course of arms: COMBAT 3a: verbal interchange of ideas; often: CONVERSATION b: an instance of such exchange 4a: the expression of ideas; esp: formal and orderly expression in speech and writing b: a talk or piece of writing in which a subject is treated at some length usu. in an orderly fashion 5 obs a: power of conversing: conversational ability b: ACCOUNT, NARRATIVE, TALE c: social familiarity; also: familiarity with a subject 6 linguistics : connected speech or writing consisting of more than one sentence (647)

————————

NARRATIVE n 1 Scots law: the part of a document containing the recitals; specif: the part of a deed immediately following the name and designation of the grantor reciting the inducement for making it 2: something that is narrated (as the account of a series of events): STORY, NARRATION 3: the art or study of narrating 4: the representation in painting of an event or story or an example of such a representation (1503)

————————

TEXT n la (1): the original written or printed words and form of a literary work (2): an edited or emended copy of the wording of an original work b: a work containing such text 2a: the main body of printed or written matter on a page exclusive of headings, running title, footnotes, illustrations, or margins b: the principal part of a book exclusive of the front and back matter c: the printed score of a musical composition 3a(1): a verse or passage of scripture chosen esp. for the subject of a sermon or for authoritative support (2): a passage from an unauthoritative source providing an introduction or basis (as for an essay, speech, or lecture) b: something providing a chief source of information or authority c: TEXTBOOK 4a: TEXT HAND b: a type considered suitable for printing running text 5a: a subject on which one writes or speaks: THEME, TOPIC b: the form and substance of something written or spoken 6: the words of something (as a poem, libretto, scriptural passage, folktale) set to music (2365-6)

————————

We may conclude, but not with much conviction, that perhaps Wittgenstein’s written words are primarily text, his unwritten words primarily discourse, and his letter a type of narrative. The immediate effect of this conclusion, however, is a mental jolt into the reality of the terminological pollution in which we find ourselves in the burgeoning field of an act of verbal behavior, regardless of our particular research discipline. In particular, text is interchanged with discourse, and discourse is confused with narrative. The problem of their distinction is not trivial, if it concerns relevant conceptual differences. The purpose of this paper is to ascertain whether, in fact, there exist cogent reasons for maintaining these three terms as independent concepts; if there are such cogent reasons we are witnessing the rapid evolution of a semantic set.

II

It appears undeniable that at least at some time in the history of two of these terms they reflect European versus State-side wording. Certainly the use of text in the “text linguistics” of Wolfgang Dressier (1973) rings with greater familiarity to the European ear, while the use of discourse in Zellig Harris’ “discourse analysis” (1952) possibly accounts for the currency of the latter term in the United States. In brief, there may exist a layer of purely English versus, in particular, German labeling of this discipline called variously text or discourse linguistics or grammar. With that disclaimer, we can turn to the definition of the first of these terms, text. Our working hypothesis is then that there is indeed a terminological distinction to be found in the set text, discourse, narrative.

Isenberg (1970:1) defines text as “eine kohärente Folge von Sätzen, wie sie in der sprachlichen Kommunikation Verwendung finden.” Petöfi (1973:205) writes: “The term next will refer . . . to a sequence of spoken or written verbal elements functioning as a single whole, which is qualified according to some (mostly extralinguistic) criterion as being a ‘text.’ “ These two definitions rest on underlying coherence conditions in determining text. Alternately, Gülich/Raible and Hartmann understand text as a supersign. Gülich/Raible (1977:40-1) write: “Ein Text, allgemein gesagt, ein sprachliches Zeichen, und zwar das primäre (bzw. originäre) sprachliche Zeichen . . . Dieses primäre sprachliche Zeichen ist seinerseits aus weniger umfangreichen, hierarchisch niedrigeren sprachlichen Zeichen aufbebaut . . . Texte [werden] nicht in der Art einer ‘linguistique de la parole’ betrachtet . . ., weil sowohl ihre konstitutiven Elemente als auch ihre konstitutiven Regeln auf der Ebene der Langue erfaßt werden.” Hartmann’s (1971:10) definition reads: “Der Text, verstanden als die grundsätzliche Möglichkeit des Vorkommens von Sprache in manifestierter Erscheinungsform, und folglich jeweils ein bestimmter Text als manifestierte Einzelerscheinung funktionsfähiger Sprache, bildet das originäre sprachliche Zeichen.”

The above definitions underscore two widely held parameters of text: coherence and supersign; however, they make no mention of discourse. In these definitions it is nonexistent as a separate entity. In fact Dressier (1973: 12) defines text as “die höchste sprachliche Einheit” equal to English discourse, French discours, Italian discorso, Russian tekst. We need then to turn to those discussions that utilize both words. Halliday (1978:108-9) defines text as “the instances of linguistic interaction in which people actually engage: whatever is said or written, in an operational context, as distinct from a citational context like that of words listed in a dictionary.” Halliday’s use of discourse appears synonymous with that of text. The text is realized in the context of a theoretical sociolinguistic construct called “situation,” which is identified in three parts: (a) what is actually taking place, (b) who is taking part, (c) what part the language is playing (109-10). In his 1964 The Linguistic Sciences and Language Teaching Halliday, along with Mcintosh and Strevens, calls the three components of the situation “field of discourse,” “style of discourse,” and “mode of discourse,” respectively. In his 1978 book Halliday does not insist on the use of field, style (now tenor), and mode in a collocation with discourse. Instead he emphasizes how they are determinants of the text (122, 110). Aside from the above collocations Halliday has two instances in his 1978 book in which discourse is used in a rather general concrete sense: to explain that text is not a supersentence (109) and to attribute genre to each text (134). Consequently, the most we can perceive in Halliday’s two unqualified uses of the term discourse, is that discourse is the raw material of text.

Van Dijk in his 1977 Text and Context: Explorations in the Semantics and Pragmatics of Discourse quite clearly recognizes two separate entities in text and discourse. He writes (3): “utterances should be reconstructed in terms of a larger unit, viz. that of TEXT [which] denote [s] the abstract theoretical construct underlying what is usually called a DISCOURSE.” Discourse, on the other hand, van Dijk considers (5) “as a sequence of linearly ordered n-tuple of sentences” constrained by semantic and pragmatic rules which differentiate it from composite sentences. For van Dijk then, text is a theoretical unit, while discourse has linguistic flesh and blood— an attitude which somewhat supports the supersign or LANGUE concept of Gülich/Raible and Hartmann above. Instead of text linguistics, Gülich (1970) speaks of macro-syntax in dealing with the text, whereas Halliday (1978: 135) insists that text is a semantic concept, not merely a supersentence.

Evidence from psycholinguists points to no distinction in text and discourse, but to a division between syntax and semantics. Frederiksen in an article entitled “Semantic Processing Units in Understanding Text” (1977:66) distinguishes “between semantic units (such as events) and textual units (such as sentences)”. Frederiksen seems to equate text production with discourse production. Frederiksen, who holds that text structure reflects knowledge structure (57), postulates three decision levels for what he labels alternately the process of text generation or discourse production. From the propositional network or the “message domain” a textual message is chosen. This message or propositional knowledge is selected through memory search and pragmatic decisions; it results in level one, the “message base.” To the message base the speaker or writer applies decisions about sequence, topicalization, reference, and correspondence between semantic and textual units, yielding level two, the “text base.” To the text base the speaker applies decisions of sentence structure to yield “a sequence of grammatical sentences (text)” (67).

Similarly, Hurtig in “Toward a Functional Theory of Discourse” (1977:101) holds unequivocally that “the sentence is the largest linguistic unit whose structure is syntactic.” Hurtig uses the word text only in the equivalence “text or discourse” (99) ; he strongly favors the view that sentence grammars deal primarily with syntactic constituents, while discourse grammars are concerned mainly with semantic, logical, or cognitive constituents (102). The text or discourse itself is “a string of successive sentences . . . [which] can be monologues, dialogues, or multiperson interchanges . . . [and have] topical or logical structure” (90).

To summarize thus far, the two psycholinguistic studies employ text or discourse interchangeably to designate a physical set of related sentences which comprise a linguistic message. They hold that the explanation of text or discourse requires the so-called interactionist theory of linguistic behavior, which is based on the interaction of formal linguistic grammar with the principles of perception and production outside the grammatical system. This position is supported also by linguists such as Bever, Katz, and Labov in opposition to the so-called inclusive theory, which holds that the separation of grammatical facts from linguistic but nongrammatical facts is arbitrary and intuitively unmotivated. Halliday would support the psycholinguistic conventions. Van Dijk’s notion of discourse approximates the psycholinguistic / Halliday attitude toward discourse or text, while his notion of text appears closer to that of Gülich/Raible and Hartmann’s supersign.

The viewpoint of the litterateur opens yet other windows on our discussion of text. In particular our horizons are broadened by theories such as those of the Soviet School (e.g. Lotman, Uspensky), which consider texts as comprising a system of many and varied interconnected modalities that is tantamount to culture (cf. Shukman 1976, Baran 1976). However, within literature narrowing trends exist, for example, that of Marie-Laure Ryan (1979), who prefers to study a literary system in terms of rules instead of in terms of signs, maintaining that the linguistic sign is limited to the lexicon. Ryan is intent, as are most literary theorists, on determining the dichotomy between literary and non-literary texts. She is convinced that the opposition between literature and non-literature should be studied with a focus on typology, which includes genre theory. Her premises are that genre applies only to a text and a text is a self-sufficient linguistic utterance. She illustrates her thesis by maintaining that conversation, for example, is a type of discourse but not a type of text. Thus for Ryan text must be governed by “global requirements and conditions of coherence” (311; cf. Isenberg, Petöfi above). Such coherence is lacking in the case of conversation since, she claims, “real-life dialogues” are comprised merely of “a loose linear . . . concatenation of sentences . . . not preventing speakers from changing topics and interrupting each other” (1979: 311). The distinction here, then, between text and discourse is that the former always falls within genre theory. Observe, however, that Ryan’s view of conversation is in stark contrast with that of Halliday; recall that in one of the two uses of discourse (above) Halliday states (1978:134): “there is generic structure in all discourse, including the most informal spontaneous conversation.” For Halliday conversation can qualify indiscriminately as text or discourse; for Ryan it can only be discourse, not text. In some anthropological circles the term text appears to be shunned altogether. Thus, for example, Butterworth (1978:321) simply speaks of “a piece of conversational behavior” and Mathiot (1978:203, 217) refers to language as but one possible ingredient of “communicative episodes.” In another article, “The Theoretical Status of Discourse” (1979), Mathiot says outright that “Discourse is not a part of language.”

Nevertheless, Halliday does consider the possibility of NON-TEXT. He postulates three factors to designate text in opposition to non-text. Besides generic structure and the commonly held cohesion principle, Halliday speaks of textual structure, that is, thematic and informational patterns which contribute to the well-formedness of the texture. He demonstrates (1978:134) that the thematic structure is scrambled in the following, and consequently it is a non-text:

Now comes the President here. It’s the window he’s stepping through to wave to the crowd. On his victory his opponent congratulates him. What they are shaking now is hands. A speech is going to be made by him. ‘Gentlemen and ladies. That you are confident in me honours me. I shall, hereby pledge I, turn this country into a place, in which what people do safely will be live, and the ones who grow up happily will be able to be their children.’

Ryan (1979), in turn, identifies two sorts of texts in literature which do NOT have coherent logico-semantic representations; the one is nonsense poetry, the second is what litterateurs have come to label simply as “text,” which refers to avant-garde literature well formed at the sentence level, but ill-formed at the sentence concatenation level. If nonsense poetry and avant-garde “text” can qualify as text, then it would seem reasonable that sentence sequence sets of schizophrenics, seniles, or small children should likewise qualify as text. Van Dijk (1972: 308) would exclude these on the basis of lack of semantic coherence. Consider, for example, the following “representative of ‘schizophrenic language’ ” as cited by Lorenz (1961:604):

Contentment? Well uh, contentment, well the word contentment, having a book perhaps, perhaps your having a subject, perhaps you have a chapter of reading, but when you come to the word “men” you wonder if you should be content with men in your life and then you get to the letter T and you wonder if you should be content having tea by yourself or be content with having it with a group or so forth.

The passage appears less grammatically scrambled than Halliday’s “non-text” citation above. Whether we would judge it as necessarily schizophremic, had we seen it in isolation, is an open question. Lorenz would answer the question with a qualified affirmative; she writes (603): “We are faced with the paradox that while we recognize schizophrenic language when we see it, we cannot define it.” But Brown (1973:400) tells us “that only some of the linguistic productions of schizophrenics appear disorganized or deluded; very many do not.” In view of the qualifying of nonsense poetry and avant-garde literature as text and in view of the inability of linguistics to define satisfactorily pathological language, it appears that the coherence condition requires revision to embrace texts which do not pass the standard coherence test.

Returning to Lotman, non-text occurs “the moment when a simple fact of linguistic expression ceases to be perceived as sufficient for a linguistic message to become a text” (cited in Rewar 1976:362-3); in its stead the cultural “text” emerges. Barthes (1971, 1968) prefers to speak of more or less text in which text represents the new. By text he means not so much a positive object with well-marked boundaries, but more a methodological field, a syntagm, a chain of meaning. Again text becomes a very broad concept in the hands not so much of the litterateurs Lotman and Barthes, but of the semiotists Lotman and Barthes.

III

Let us now shift the focus to the association of discourse with narrative as part of this evolving terminological set. In narratology proper, discourse is the second arm of a complex sign function, the first being STORY according to Barthes (1966) and Todorov (1966). Here discourse refers to Hjelmslev’s “expression plane,” Saussure’s “significant.” Thus Greimas considers discourse to be the primary datum in every sentence concatenation; he speaks of scientific discourse, interpretive discourse such as art criticism, persuasive discourse such as advertising (Greimas/Courtés 1976).

However, Greimas is convinced that “every discourse entails a narrative dimension” (Nef 1977:19). He introduces (1977:29) a propositional formula for a minimal narrative utterance, “NU = F (A),” which is reminiscent of van Dijk’s (1972:307) “deep structures . . . similar to the internal structure of the propositions in a modal predicate logic.” Not only do Greimas and van Dijk overlap in their propositional base (function :actant, relation : argument, predicate : subject), both are engaged in identifying global modalities. Thus van Dijk (1977) works with the mental structures of action such as wants, desires, intentions, purposes, while Greimas (Greimas and Courtés 1976:443) postulates three “ modalities of doing”: “volition-to-do,” “cognition-to-do, “ and “power-to-do.” To these fundamental features we may add still another global category which we call “inferential structure” (cf. Rauch forthcoming), structure which emulates the way we think, our basic tripartite thought process consisting of a general proposition, a specific proposition, and a conclusion. Thus the Aristotelian concept of narrative action composed, in essence, of an orientation, complication, and resolution, parallels the basic inference. Van Dijk (1977:245-6) extends the “conventional organization of discourse [to those] . . . of the corresponding speech acts, and of action in general,” while Greimas (1971:793) states: “Narrative structures are distinct from linguistic structures because they can be revealed by languages other than the natural languages (in cinema, dreams, etc.). ”

It must be amply clear from the immediately preceding discussion that narrative is not bound to literary language, written language, or even to language in its ordinary sense (cf. further Rauch forthcoming). Less important is the observation, which is a terminological matter, that van Dijk’s use of discourse approximates Greimas’ use of narrative (for van Dijk narrative is a genre type, 1977:243). However, the extension of narrative beyond verbal language to nonverbal communication certainly bespeaks a conceptual breakthrough. It makes quite palatable Mathiot1 s rather provocative statement (quoted above) that “Discourse is not a part of language.” Indeed, one does not have to look far for concrete applications of this concept: consider, for example, Bouissac’s (1977: 149-50) description of a circus act as a “primarily visual meta-discourse . . . a micro-narrative.”

IV

We return now to our experimental data, the Wittgenstein quotation and the definitions in Webster’s Third of text, discourse, narrative. Faced with having to pit the Wittgenstein quotation against the three dictionary entries our initial intuitions yielded no strong prejudices in assigning Wittgenstein’s written words, unwritten words, and letter. In fact, the original task appeared unfruitful since the dictionary senses of the three terms do not strike individually discriminating tones. As the dictionary compilers warn of their definitions (Webster’s Third 1964:19a): “The system of separating [the sense] by numbers and letters . . . is only a lexical convenience. It does not evaluate senses or establish an enduring hierarchy of importance among them. The best sense is the one that most aptly fits the context of an actual genuine utterance.” Yet the dictionary glosses of “ideas” for discourse may have influenced its assignment to unwritten words, namely, the whole world of Wittgenstein’s thoughts, while text and narrative fell then to the more concrete written words and letter, the former having appeared more general and accordingly more appropriate to the seemingly wider latitude offered by the many dictionary glossings of text.

Such floundering in terminological uncertainty, however, is entirely absent in the evolving usage which we document in the crossdisciplinary applications of text, discourse, narrative for acts of verbal and verbal-like behavior. Albeit the parameters we are seeking are similar, the terms shift considerably. At its simplest the following distinctions hold to bind the emerging semantic set :

Compelling arguments for this evolving triad come from Peircean semiotics. Here all phenomena, including language, action, and thought, are signs? indeed man himself is a sign, to quote Sebeok (1977:181): “man is . . . a process of communication . . ., or, in short, a TEXT” (emphasis mine). Beyond this, however, the glue for cementing this semantic set resides in Peirce’s phenomenological categories of Firstness, Secondness, and Thirdness, into which we can set text as a First, discourse as a Second, and narrative as a Third. Alternately, text is the sign proper, discourse is the object, and narrative is the interprétant within Peirce’s classic definition (1931:§339): “A sign [text] stands FOR something [discourse], TO the idea [narrative] which it produces or modifies.” The qualities of the categories serve to reinforce the assignment of the terms. Thus, a letter is mere possibility and belongs to the past, written words exist in fact and are part of the present, while unwritten words have imputed existence and belong to the future thought. That these qualities interlace is to be expected, since that factor is inherent in all phenomena; nonetheless, their PRIMARY qualities serve to identify them distinctively. What remains of the features reviewed above such as syntax semantics, sign:rule, genre :non-genre ; coherence : incoherence, text : non-text, more text:less text, normal language: pathological, child, and senile language is relegated to connotative senses in the evolving set of definitions within a study which is then most graphically (iconically) termed TEXT GRAMMAR.

REFERENCES

Baran, Henryk, (ed.) 1976. Semiotics and structuralism: Readings from the Soviet Union. White Plains, NY: International Arts and Sciences Press.

Barthes, Roland. 1966. Introduction à l’analyse structurale des récits. Communications 8:1-27.

——. 1968. Elements of semiology. Trans, by A. Lavers and C. Smith. New York: Hill and Wang.

——. 1971. De l’oeuvre au texte. Revue d’Esthétique 24:225-32.

Bouissac, Paul. 1977. Semiotics and spectacles: The circus institution and representations. A perfusion of signs, ed. by T. A. Sebeok, 143-52. Bloomington: Indiana University Press.

Brown, Roger. 1973. Schizophrenia, language, and reality. American Psychologist 28:395-403.

Butterworth, Brian. 1978. Maxims for studying conversations. Semiotica 24:317-39.

Dressler, Wolfgang. 1973. Einführung in die Textlinguistik. 2nd ed. Tübingen: Niemeyer.

Frederiksen, Carl H. 1977. Semantic processing units in understanding text. Discourse production and comprehension, ed. by R. 0. Freedle, 57-87. Norwood, NJ: Ablex.

Greimas, A. J. 1971. Narrative grammar: Units and levels. Modern Language Notes 86:793-806.

——, and Joseph Courtés. 1976. The cognitive dimension of narrative discourse. New Literary History 7:433-47.

Gülich, Elisabeth. 1970. Gliederungssignale in der Makrosyntax des gesprochenen Französisch. Munich: Fink.

——, and Wolfgang Raible. 1977. Linguistische Textmodelle. Grundlagen und Möglichkeiten. Munich: Wilhelm Fink.

Halliday, Michael Alexander Kirkwood. 1978. Language as a social semiotic: The social interpretation of language and meaning. Baltimore: University Park Press.

——; Angus Mcintosh; and Peter Strevens. 1964. The linguistic sciences and language teaching. London: Longman.

Harris, Zellig S. 1952. Discourse analysis. Language 28:1-30.

Hartmann, Peter. 1971. Texte als linguistisches Objekt. Beiträge zur Textlinguistik, ed. by W.-D. Stempel, 9-29. Munich: Wilhelm Fink.

Hurtig, Richard. 1977. Toward a functional theory of discourse. Discourse production and comprehension, ed. by R. 0. Freedle, 89-106. Norwood, NJ: Ablex.

Isenberg, Horst. 1970. Der Begriff “Text” in der Sprachtheorie. Arbeitsstelle Strukturelle Grammatik, Berlin: Deutsche Akademie der Wissenschaften zu Berlin, Bericht 8.

Lorenz, Maria. 1961. Problems posed by schizophrenic language. Archives of General Psychiatry 5:406-10.

Mathiot, Madeleine. 1978. Toward a frame of reference for the analysis of face to face interaction. Semiotica 24:199-220.

——. 1979. The theoretical status of discourse. Paper presented at the Fourth Annual Meeting of the Semiotic Society of America.

Nef, Frederic. 1977. Introduction to the reading of Greimas: Toward a discursive linguistics. Diacritics March 1977:18-22.

Peirce, Charles Sanders. 1931. Collected papers of Charles Sanders Peirce, ed. by. I. C. Hartshorne and P. Weiss. Cambridge, MA: Harvard University Press.

Petöfi, János. S. 1973. Towards an empirically motivated grammatical theory of verbal texts. Studies in text grammar, ed. by J. S. Petöfi and H. Rieser, 205-75. Dordrecht: D. Reidel.

Rauch, Irmengard. Forthcoming. Semiotics in search of method. Semiotica.

Rewar, Walter. 1976. Notes for a typology of culture. Semiotica 18:361-78.

Ryan, Marie-Laure. 1979. Toward a competence theory of genre. Poetics 8:307-37.

Sebeok, Thomas A. 1977. Ecumenicalism in semiotics. A perfusion of signs, ed. by T. A. Sebeok, 180-206. Bloomington: Indiana University Press.

Shukman, Ann. 1976. Literature and semiotics: A study of the writings of Yuri Lotman. Amsterdam: North Holland.

Todorov, Tzvetan. 1966. Les catégories du récit littéraire. Communications 8:125-51.

Webster’s Third New International Dictionary of the English Language Unabridged. 1964. Springfield, MA.

Van Dijk, Teun. 1972. Foundations for typologies of texts. Semiotica 6:297-323.

——. 1977. Text and context: Explorations in the semantics and pragmatics of discourse. London: Longman.

Von Wright, Georg Henrik, and Walter Methlagl. (eds.) Briefe an Ludwig von Ficker I. Salzburg: Otto Müller.