SEMIOTICS AND LINGUISTICS

If one were to look at the recent history of the study of signs and sign systems one would see, under the headings of 'semiology' and 'semiotics', a range of scholarly activities. The two terms have in fact been used by various scholars interchangeably to mean the same thing or, sometimes, to mean different things (cf. Barthes's or Derrida's 'semiology', or Kristeva's 'semiotics'—the French sémiologie and sémiotique).

Beyond these terminological differences, which are meaningful only as indices of conceptual distinctions, there are other differences. How semiotics is placed among the sciences, for example, depends, in many ways, upon the definitions of 'science' itself. The Saussurean tradition uses the term semiology in accordance with the well-known passage from the Cpurs de linguistique générale, in which semiology is defined as a part of general psychology and includes the whole of linguistics. In the Peircean framework, semiotic is considered as a specific view of logic; Peirce, unfortunately, made no distinction between logic (as semiotics) and linguistics. If, however, we suppose that natural language is only a small, although important, part of the universe of signs, Peirce's semiotic would embrace many scientific fields, among which linguistics —even when defined in the broadest sense—would have a place.

Such a hierarchy may appear to be the only one possible. Nonetheless, an opposing view has been held by Roland Barthes and is explicated in his stimulating Elements of Semiology. Despite the wave of surprise that Barthes generated, his vantage point is far from isolated, although it had not been expressed before. An entire philosophical tradition has, in fact, come to us from the Greek concept of logos, that is, thought-language as reason, the feature that distinguishes man, the rational animal, from all other living beings (aside from the gods!). The Stoics and philosophers such as St. Augustine later established the principle underlying this tradition: that human language is the only relevant semiotic system insofar as the specificity of anthropos is concerned. Thus, depending upon one's view of the sign and of sign systems in general, and also upon one's view of knowledge as a relationship established by a subject (i.e., man) between himself and an object to be known, one of two hierarchies can be built: one in which language encompasses every other semiotic system (as the universal translator and mediator of knowledge) or one in which a multitude of semiotic systems exist, with language being only one of these systems.

Roman Jakobson, in his essay "Relations entre la science du langage et les autres sciences" chooses a position that is close to Saussure's, except that 'general psychology' is replaced by 'social anthropology' (Jakobson, 1973). Linguistics, then, defined as "the study of communication of verbal messages," is a part of semiotics—"the study of communication of any messages"—and is included in anthropology, the "study of communication," without any qualifications. Going a step further than Saussure, Jakobson joins anthropology to biology, linking the concept of (human) semiotics to zoosemiotics and ultimately to the study of the genetic code. Two problems seem to be raised by these definitions : (1) the central position of communication—rather than the sign—in such a conceptual framework; and (2) the status of social anthropology as the study of the communication of all things and beings (e.g., messages, women, goods ...).

In any event, linguistics is indeed a part of semiotics, insofar as any verbal message is a subset of all messages. The relationship between semiotics and linguistics is therefore a function of how communication is defined.

ON DEFINING SEMIOTICS

How does Barthes's conception of semiotics differ from that of Saussure's, Jakobson's, and almost every other contemporary scholar's? First, what Barthes calls, in his Elements of Semiology, 'translinguistics'¹ is superordinate to semiology since any knowledge of the sign and its functions must be analyzed through the semiotic structures of natural language. Barthes includes under 'translinguistics'—which he does not actually distinguish from semiology—the study of "the great signifying unities of discourse" and "the objects of our culture inasmuch as they are spoken." (This is, in fact, an ambiguous way of defining two different levels in the domain of semiotics, that is, social processes as semiotic processes and their linguistic interprétants; the relationships between these two levels are themselves the objects of knowledge.) The implications stemming from Barthes's theory are numerous. Sebeok, for exampie, has noted, regretfully, the "absolute exclusion of sign processes among the speechless creatures from the semiotic universe." Barthes, it seems, based his analysis on signification, not on communication and coding systems, simply because restricting communication to the exchange of linguistic messages would have been too limiting (cf. the definitions of 'signification' and 'value' in the Saussurean [i.e., linguistic] sense).

Some scholars have tried to draw a line between two "groups of affinities" (e.g., Rulon Wells, 1971) or even between two fields for semiotics : communication, on the one hand; and signification, semiosis, or meaning, on the other. If we can rightly consider that an important difference exists between the two viewpoints and the theories that underlie them (for example, an extension of information theory vs. an extension of functional linguistics), it does not follow that there is any real opposition either in the terminological or in the methodological sense. It seems in fact impossible to accept a dichotomy in which "the science of communication" and the "science of signification"—whether called semiotics or something else—are two different intellectual activities, or even two separate methodological trends. If signification (or semiosis) has been the object of centuries of philosophical thought and discourse, communication as a delimited concept is a new epistemological problem. Of course, communication has been consistently defined in mathematical constructs, such as can be found in information theory, where the concepts of code, message, information, communication, etc., have been quantized. But once these same concepts are used in anthropology, biology, linguistics, and the other social sciences, their identity as single constructs is not so evident. Thus, a French sociologist once wrote: "the role of information theory being only the role of a segmental theory dealing with some necessary and instrumental conditions of communication, we must conclude that there is no such thing as a communication theory" (René Pagès, 1968). It therefore appears that, from the study of "some instrumental conditions of communication" to a general theory of communication, there is much work to be done, and semiotics is precisely the field where the groundwork could be laid. At present, there is nothing that forces us to contrast a semiotics of communication with a semiotics of signs and semiosis.

The idea of a semiotics of expression might prove more fruitful, since 'expression' as a pragmatic concept could divide the discipline with regard to the level of exchange (i.e., communication vs. expression). Even so, semiosis would still occur on another level. In any case, the word 'communication has such a broad meaning that semiotics, if centered on this concept, would be identical to genetics (the exchange of information in genes), to social anthropology (the exchange of women), to economics (the exchange of goods), when, in reality, only a certain portion of such communication would be relevant. The works of LéviStrauss, Keynes, and others, for example, have many semiotic aspects, but they are not semiotic research. When the sign (or semiosis) becomes the underlying concept of semiotics, social anthropology, economics, medical semiology, etc., are semiotically relevant only if communication is described and analyzed according to precise (even if broad) definitions of the elements being communicated as signs. The definitions of the sign and of semiosis would thus replace the definition of communication in the delimitation of any semiotic discipline.

ON SOME CONCEPTIONS OF SEMIOTICS

The narrow conception of semiology in the Saussurean tradition— namely, that of Buyssens, Prieto, and other scholars (e.g., André Martinet), whose conception is akin to a "functional view of language"— is not a result of the fact that such a viewpoint considers communication as the starting point, but rather that the sign itself (and therefore communication) is narrowly defined. Buyssens (1970:12), for instance, states that "communication procedures" are "means used to influence somebody else and recognized as such by the person we want to influence. . . ." Accordingly, any process in which we act upon someone who is unwilling to recognize such an action has nothing to do with semiology. Signs are only signals; 'indices' are considered outside the field. On a much broader level, Charles Morris has divided communication into the following three concepts, each of which encompasses the one after it: (1) communication that covers "any instance of the establishment of a commonage, that is the making common of some property to a number of things"; (2) communication that is restricted to signs, defined as "the arousing of common significata by the production of signs"; (3) communication related to language, "when the signs produced are Ianguage signs" (Morris, 1971:195ff.). Morris's view of semiotics, even when restricted to his second concept of communication, embraces more than Buyssens's since it includes any process in which signs are used without regard to will, conscient knowledge (such a condition, by the way, would make zoosemiotics almost impossible). On the other hand, it should be noted that, since the use of signs gives way to communication (2) even when the commonage is not semantic (e.g., when signs of anger establish a commonage of feeling in another person—this is Morris's example), it seems that Morris does not consider the production of signification (i.e., of new signs) to be a relevant criterion for semiotics.

Morris's description of semiotics in Foundations of the Theory of Signs concerns, rather, a behavioral process in which "something takes account of something else mediately." The first 'something' is defined as the interprétant; the second 'something' as the designatum; and the mediator as the sign vehich (see Peirce's 'interprétant', 'object', and 'sign' or 'representamen'). Morris's account of Peirce's theories of semiosis shows that the two scholars differ in their conceptions of semiotics, despite outward similarities. Morris's critical remarks about Peirce are centered on the latter's lack of precision in the following regard : (1) Peirce's definition of the sign as anything involved in the process of mediation includes, in the concept of sign, such phenomena as a conditioned stimulus; (2) the sign process, similarly, is defined by Peirce in such a way that it can be applied to nonbehavioral situations; and (3) the infinite number of triadic relations, implicit in the fact that one of the elements in such a relation is the starting point of another triadic relation, and so on, is unduly included by Peirce in his definition of the sign.

Let us look briefly at each of these points. Morris's first criticism was intended to restrict the concept of sign processes to a subclass of processes involving mediation, that is, of "restricting sign processes to those in which the factor of mediation is an interprétant" (Morris, 1971:338). His second point "illustrates the difficulties which appear when we leave the ground of behavior situations in attempting to define 'sign'." Morris's final criticism concerns the introduction into the definition of a sign of an "empirical question ... as to whether signs always generate new signs" (1971:339). Even without examining Peirce's writings on these points, we must point out that all of Morris's remarks are grounded in a behavioral approach to science in which the philosophical concepts of Peirce's "phaneroscopy" cannot be accepted without important modifications. Let us consider, for example, the fact that a sign, according to Peirce, implies, in addition to an Object and a Ground, another sign (the Interprétant) that gives way to another sign process in such a way that: Sign 1 Object 1 Interprétant 1 Sign 2 ( Interprétant 1) Object 1 Interprétant 2 Sign 3 ..., ad infinitum. From Peirce's viewpoint, such a sequence of events is not at all an empirical fact, but is of a definitional character that has resulted from the infinite power, linked to Thirdness, and is entailed by a philosophy of thought whose foundations are metaphysical. The concepts of the sign and even of an object (as object-of־a-sign) are therefore defined by Infinity. Thus one must look at the problems raised by Peirce's broad definitions of the sign, semiosis, or the interprétant, within the context of the whole theory before these concepts can be used properly in other contexts (both philosophical and epistemological).

LOOKING FOR UNITY

Considering Saussure's 'semiology', Morris's, Jakobsons, Sebeok's, or Eco's semiotics', Tarskis or Carnap's 'semantics', Hjelmslev's 'glossematics', Barthes's project of 'translinguistics', or Thorn's 'semantics' in topology, it is difficult to conceive of the unification of what would be the major organon to a unified science. This is especially true when methodologies vary, philosophical foundations oppose one another, and when even the focus of inquiry is not the same, ranging from signalling behavior to literature, from the logical structure of mental processes to the social history of ideas, from the genetic code to opera or film. At present, any definition of semiotics that is unable to take into account at least some of the various aspects of current semiotic research is too narrow; and any that is unable to work out a definition in which semiotics is distinct from the adjacent sciences (even if these sciences have distinctively semiotic parameters) is too broad. How narrowly or broadly semiotics is defined ultimately depends upon how the sign, the sign process, the sign system, and so forth, are defined—with communication by means of signs and sign systems being dependent on sign theories.

There is currently no other theory broad enough to cover the entire scope of empirical semiotics except Peirce's. His theory of signs, however, is not the solution to the problems of defining semiotics, but, instead, steers one toward the right questions and offers some powerful tools, such as the classification of signs and the concept of the interprêtant. With semiotics defined as the science of the semeion, Peirce puts communication in his theory, inasmuch as the triadic processes imply the production and use of signs and are therefore implicitly included in the study of codes (even though the latter idea occurred sometime after Peirce) as sign structures. Considering semiosis in the Peircean fashion allows the epistemic "commonage" of science and philosophy in the universe of signs as well as an enlargement of semiotics without an epistemological break from language signs to anthropological signs, or from them to any sign in the living world. It allows, in addition, the inclusion of a diversity of viewpoints, such as Hjelmslev's conception of a hierarchy of semiotic systems.

In short, the disadvantages stemming from a broad definition of the sign—which are clearly evident when we try to look for a contrastive view of science—are compensated for by the power of a general, theoretically based framework. And, if such a frame of reference is paired with a comprehensive view of science—as is Peirce's—its power is difficult to match.

EPISTEMOLOGY FOR SEMIOTICS

What seems more important than asking if a certain theoretical view can supply a satisfactory (for whom?) definition of semiotics is to look at the types of knowledge required to study the different aspects of semiotics. This raises the well-known problem of the nature of knowledge in biology, pathology, psychology, sociology, and other disciplines. That is, how can these sciences be considered as epistemic unities, when their objects of knowledge, the observed data, and the observer (e.g., human vs. nonhominoid biology) may not be of the same nature? The problems raised by the fact that there is a continuum of matter from nonliving to living are of a different nature, since they would concern sciences in which nonhuman objects are always considered (see, for example, Pascual Jordan's Oie Physik und das Geheimnis des organischen Lehens [1948]²).

When a tentative definition of a unified semiotics is discussed, two classes of problems arise: first, how are certain trends in semiotics to be unified and what types of hypotheses, definitions, and rules are required of its domain? Second, if theoretical systems are able to produce consistent definitions that are broad enough, what are the empirical implications of the central hypotheses within each of these systems? The fact that Peirce's framework and philosophy are not only mentalistic (in Morris's terminology) but also idealistic—with the hypothesis of a Supreme Power—and metaphysical, is as important as the existential and "hermeneutic" character of Heidegger's philosophy when Peirce's theories of the sign as a tool, or instrument, are discussed. Yet the empirical nature of their respective definitions cannot be simply deduced from the theoretical structure of their work.

The power of a system is certainly distinct from its empirical and heuristic implications, which can be analyzed and criticized from other systematic viewpoints. What is more relevant here is not Peirce's metaphysical post-Kantian attitude or his pragmaticism', but the possible uses of his definition of the sign and of semiosis as the origin of any semiotic knowledge.

Whatever the system of reference may be, semiotic activities depend upon the epistemic relationship between the seeker of knowledge and his object of study. In a tentative and somewhat naive way, I would like to classify the basic types of relationships between the scientist, the scholar—man, in his own historical and cultural situation, with all of his psychological motives, including the unconscious ones—and the objects, considered here as signs, systems of signs, sign processes, sign behaviors, communication using signs, and so forth. The process of knowledge cannot itself be described without reference to the products of at least a few semiotic systems, either natural language—given that many types of 'discourse' are scientific, philosophical, mythical, political, pedagogical, etc.—and/or 'artificial' languages and formalized systems, with the addition of a few iconic ones. The fact that one of the main objects of semiotic inquiry (and the only one possible for Barthes) is natural language and its products, is relevant to the situation insofar as such a relation defines the only homogeneous epistemic relation: man, using the semiotic system of language, studies the products of his semiotic activities through the same type of semiotic system that he uses in the constitution of his knowledge. This last situation can be empirically observed in philology. Literaturwissenschaft, poetics, rhetoric, textgrammar and text linguistics, and, on a purely theoretical level, in "pure" linguistics.

Next to language, the semiotician may and does study the secondary codes and speech surrogates that can be easily translated into natural language. He can then deal with semiotic behaviors that are linked in an empirical manner to language behavior, with its conditions, its products, and its implications (on psychology, cultural structures, and so forth).

Human semiotic activity, including the use of any sign or sign system —be it symbolic, indexicai, or iconic—is the broadest field in which the object of science is of the same nature as the scientist himself (this sameness is evident when he studies his own semiotic behavior, his own culture, his present time). Anthroposemiotics involves a detailed look at every human phenomenon, since man can be described as a logical, or semiotic, animal (if a broad definition of the sign, such as the one offered by Peirce, is accepted). (This does not imply, however, that other animals are not logical or that man is entirely logical.) We may then study zoosemiotics; here the object, a living organism, is in some respects identical to man, but in other respects different. The latter are of greater relevance where an epistemic relationship is concerned. We are certain of the specific character of anthropos when the semiotics of language is implicated and when historical data, considered as a major element in human social behavior, are involved; we can, in addition, accept or reject the fact that the human unconscious—if considered a major component of human psychology—is unique among bio-psychic structures.

When zoosemiotics is included in the scope of semiotics and not merely associated with it peripherally, all the problems of animal behavior (with respect to ethology, psychology, sociology) are introduced. If the knowledge of animal behavior and its underlying structure are to be considered as possible objects for scientific study, then sign processes —with an adequate definition of the sign—that are accounted for in observed animal behavior are indeed possible areas of inquiry for seience. Any definition of semiotics, either including or excluding sign processes and systems outside the human world, is, of course, analytical since buch a definition would depend upon what meanings are given to the sign and to semiosis. The attitude excluding nonhuman sign processes would require an anthropological definition of the sign, which is contrary to a great part of philosophical tradition and, of course, to Peirce's theories (it would be different with a Husserlian or Heideggerian starting point).

Zoosemiotics, with ethology, points to the continuity from animal to man, but as the relationship between knowledge changes from one object to the other, the methodology and issues raised must be different from those in human semiotics, where a linguistic interprétant can always be elicited even when the object is not language semiotics, or observed (e.g., in philology) when the data are available (which is an empirical, not theoretical, problem). With zoosemiotics, the central object of study is not and cannot be the sign and semiosis as such, but their effects in observable exchanges, in a behavioral approach, that is, in communication. The fact that other approaches, often criticized as mentalistic and introspective, which were rejected as far back as Auguste Comte's early works, are possible only in anthroposemiotics and are still the only possible ones in some respects (e.g., literary semiotics) clearly shows the existence of a gap. A behavioristic approach, with a communication-centered definition for semiotics, would be one answer to the problem. Unfortunately, behaviorism, in its extreme form, is unable to give answers to 'scientific' questions about man and his specific semiotic behavior, be it religious, mythical, poetic, psychotic, economic, political, or even scientific.

In 'external' knowledge patterns (that is to say, essentially not methodologically external), such as those found in the scientific study of animal life, the methods and approaches cannot be identical to those used in 'internal' knowledge patterns, such as in the study of human psychiatry. The hypothesis of a possible 'externalized' unified knowledge seems (at least in my own view of epistemology) (neo)positivist fiction.

Other fields are open to investigation where objects that may be considered signs—according to a specific definition—are at work. What Sebeok calls 'endosemiotics', in which the genetic code plays a central role, or a broad biosemiotical field where all living beings, including plants and protistae, could be studied as communication sources and receptors, give examples of objects quite different from man and his specific semiotic systems. In these fields, it is doubtful if, even with a broad definition of the sign and of semiosis, the underlying concepts of semiotics would still be the same. Thus, with the study of communication and communicative exchanges in non-living matter in the traditional domains of chemistry and physics, the assumed presence of the semeion would be a metaphor, even if an exchange of information could be considered.

The general patterns of mathematics or logic, on the other hand, belong to another type of epistemic relation in which the structures of the human brain are related to the structures of the world as object-tohuman knowledge (which is the only possible meaning of world'). Here, semiotics, as in René Thorn's stimulating work, appears at least in the abstract patterns through which phenomena are represented.

ON THREE EMPIRICAL DEFINITIONS OF SEMIOTICS AND SEMIOLOGY

The opposition between communication-centered and significationcentered semiotics or semiology is thus (a) irrelevant, since the central role of communication seems necessary when the knowledge relation is external and (b) unnecessary, if extremely relevant, when the object is man. The main object for semiotics, however, must be different from the object of the adjacent sciences, such as those concerned with information or communication theory : it must be specific, if semiotics is to be a science at all. The concepts of sign, signification, etc., can be discussed and criticized in their historical genesis, but they cannot be suppressed without suppressing the very idea of semiotics.

Among the various conceptions of semiotics currently in vogue, the following three seem to command the most importance:

(1) An overall study of "the exchange of any messages whatever and of the systems of signs which underlie them" . . . , "the key concept of semiotics remaining] always the sign" (Sebeok, 1974).

(2) A study of semiotic processes in culture, considered as an aspect of social anthropology and/or a critical analysis of human semiotic products especially in the sciences and in semiotics itself. Such a study puts stress on the sign and semiosis as human properties that are linked to history (in a post-Marxian sense) and to human psychology (viewed in a post-Freudian sense).

(3) 'Semiology' as the study of voluntary messages transmitted by means of certain signs that are defined in a restricted way as signals (as opposed to 'indices'); this may include nonhuman systems, but it exeludes many situations in which symptoms, or indices, act, even in anthroposemiotics and in the semiotics of language.

The trend given in (1) corresponds to the Peircean and Morrisian traditions, which, although different in their philosophical foundations, are similar with regard to the scope of semiotics; included in this school of thought is the bulk of American semiotics, and, to a much lesser degree, European. Trend (2) is basically philosophically oriented and is well represented among the scholars in Western Europe, while (3) corresponds to a smaller school of Saussurean linguists, led by Buyssens and Prieto.

Each definition above is linked to a tradition: (1) is connected with structuralism in anthropology, the behavioral sciences, and structural linguistics (in a broad sense, including 'generative' linguistics); (2) comes from post-Husserlian philosophy, Marxist criticism of 'ideologies', and the Freudian analysis of human psychic nature; and (3) is derived from functional linguistics and a communicational approach of socialized codes. The influence of Peirce is stronger on (1), of Hjelmslev on (2).

Each of the three groups uses different definitions of the sign. (1) and (2) give greater importance to the sign process (semiosis) ; (2), in addition, considers semiosis from a genetic point of view (i.e., where does the sign, or significance, come from?) and emphasizes the systematic, theoretical aspects of sign studies. (3) focuses on the transactional aspects of sign studies in terms of functional descriptions. Each claims to be theoretical (or 'pure'); moreover, descriptive, 'applied' semiotics, according to the Morrisian view, is not sympathetic to the view espoused by (2).

With regard to the scope of semiotics, both (2) and (3) exclude nonhuman semiotics, even when animal sign behavior is closely linked to human sign behavior (e.g., facial expressions studied in the primates by Darwin). (2) shows a decided emphasis for natural language, while (3) is methodologically related to functional linguistics, a different view altogether.

Coming, finally, to the question of a unified semiotics, I would argue that Peirce's work, more than Morris's, Hjelmslev's, or Saussure's, gives some means for answering whether or not a definition for semiotics is possible, and with which key concept—semiosis or communication. In my opinion, semiosis, understood as a structural and genetic concept, is the only possible answer and implies communication when social factors are considered. Communication, in addition, is the only empirically centralized concept when methods for studying nonhuman semiosis are necessary. A narrow behaviorism and information theory technologism are proof, a contrario, of the gap between human and nonhuman semiotics; useful, perhaps even necessary in the latter, they are insufficient, if not irrelevant, in the former. The epistemic situation, then, is different when the semiotic object is natural language (and the social, historical phenomenon of 'discourse') and when another semiotic fact, human (such as iconic messages or music) and nonhuman, is considered.

NOTES

1. "Then (since the semiologist finds language in his way) semiology perhaps tends to be absorbed into translinguistics . . "(Barthes, 1965:81).

2. It is interesting to note that Jordan tried elsewhere to apply to human psychology some key concepts of physics ('Verdrängung und Komplementarität_f 1947). His idealistic philosophy is not typical of unitarism in science; Engels's philosophy (with nineteenth-century concepts) shows a similar attitude.

REFERENCES

Barthes, R. 1965. "Elements de sémiologie." In Le Degré zéro de l'écriture. Paris: Editions Gonthier, n.d. [1965], pp.77-176. First published in Communications (Paris: Editions du Seuil, 1964).

Buyssens, E. 1970. La Communication et l'Articulation linguistique. Brussels and Paris: Presses Universitaires.

Jakobson, R. 1973. Essais de linguistique générale, vol.2. Paris: Editions de Minuit, Part I, chap. 1, pp.9-76.

Jordan, P. 1948. Die Physik und das Geheimnis des organischen Lebens, 6th ed. Braunschweig: Friedrich Vieweg und Sohn.

Morris, C. 1971. Writings on the General Theory of Signs. The Hague: Mouton. On Foundations of the Theory of Signs, I and II, see pp. 19-24. On communication, see pp. 195-98; on Signs, Language and Behavior, IV and Glossary, pp.359-68. On Peirce, see "Charles Peirce on Signs," pp. 337-40.

Pagès, R. 1968. "Communication." In Encyclopaedia Universalis, vol.4, pp. 765-68.

Sebeok, T. A. 1974. "Semiotics: A Survey of the State of the Art." In Current Trends in Linguistics, vol.12. The Hague: Mouton, pp.211-64.

Wells, R. 1971. "Distinctively Human Semiotics." In Essays in Semiotics— Essais de sémiotique, J. Kristeva, J. Rey-Debove, and D. J. Umiker, eds. The Hague: Mouton.