E-Papers >Cultura Media
The Portuguese version of this paper can be found in
Santaella, Lucia (2004). Cultura das mídias, 4a. ed. São Paulo: Experimento.
The Computer as a Semiotic Medium
The computer has indiscriminately been referred to as a tool, a tool kit, a device, an instrument, a machine, an equipment, an apparatus, and a medium. All these descriptive terms circumscribe, in fact, aspects of the various ways we make use of the computer and indicate functions that the computer can actually perform. However, it performs these functions in a very special way.
After a brief examination of the descriptive names that have been used to characterize the functions of the computer, I would like to discuss the multiple meanings of the word medium and to analyse in detail what it means to consider the computer as a medium from a semiotic point of view.
1. The descriptive names of the computer
First of all, the computer is a physical object, a very complex kind of object, indeed, but in spite of its complexity, it is as real as any physical object. Because of the variety of operations and the multitude of tasks that the computer can fulfil, we furthermore need the designation of a dynamical system to describe the particular complexity of this physical object.
At its most elementary level, the computer performs the function of a tool, not only in the narrow sense of something that is used or worked by hand, but also in the broader sense of an implement that is useful for doing a certain job or performing an operation, that is, something necessary to a person in the practice of his or her profession. In these senses, a computer can be understood in analogy to tools such as a typewriter, a pencil, a brush, a filing cabinet, etc.
However, the computer as a system is as far from being a single tool as it is from being a tool worked only by hand. A more appropriate designation of the computer is the one of the tool kit (Newell 1980: 178), which pertinently describes its complex physical whole composed by many different and interrelated units. But here again, the functions that the computer can perform go much beyond the potentialities of any tool kit. The mechanisms of a computer are designed to accomplish very specialized tasks, and for this aspect of sophisticated purposefulness and specialization, its characterization as a device is clearly adequate. Equally or even more suitable is its description as an instrument. There are at least two senses of the word instrument which make it applicable to the computer. In one sense, an instrument is a special tool used in a kind of work where fine and trained movements are necessary. In a second sense, an instrument is a special device -- as for recording, regulating, controling -- that functions on data obtained by the device itself. At this level of specialization, the computer fulfils the same function as a tape-recorder, a photographic camera, or a mechanism for measuring heat release, etc. In these cases, the word instrument suggests a certain ready applicability to the matter under consideration rather than only the bare fact of use.
According to Flusser (1985: 25-29), devices and instruments are, above all, technical products. This explains why they became increasingly refined since the industrial revolution. Devices are extensions of the human sense organs, and as such they are able to simulate and extend the functions of these organs. After the industrial revolution, when the design of such devices began to be supported by scientific research, these devices came to be generally referred to as machines.
The idea of a machine, however, is not as recent as the industrial revolution. In its ancient sense, it meant a structure or a construction, whether material or immaterial. This semantic duplicity is also apparent in the modern sense of the word, where machines comprise material and fluid bodies, as well as electricity. As it came to be generally understood, since the industrial revolution, a machine is an assemblage of parts or solid bodies and also of fluid bodies or electricity in conductors that transmit forces, motion, and energy in some predetermined manner and to some desired end. But the most relevant meaning associated with the idea of a machine is that of a complex device for doing work beyond human physical or mental limitations mostly faster and more precise than human hand and mind. In this sense, which extends human physical and mental powers in a connected way, the computer should be the most powerful of all machines, and under such a label, the computer is generally considered.
However, not only the external body of the computer, but also its internal parts and complicated joints of levels testify to its compound nature and weigh against considering it as a single machine. There are indeed, many devices that go together with the computer. First, there are the basic units, the key board, the screen, and the cables to connect these parts together. "In addition to these basic units, the computer can be hooked up to a wide variety of other devices, such as a printer, a scanner, a drawing table, a mouse, a tape drive, a large capacity disk drive, a slide projector or slide camera, and a videotape or videodisc player. Just about anything that can be operated electrically can be turned on and off by computer" (Paulsell 1990: 199). It is worth noticing that all those devices increase remarkably in number and sometimes in size as we move from the personal computer to the workstation and the mainframe.
Internally, the subdivision of the computer into component parts, or what is currently called its mechanics, is no less diverse than its external parts. Brown (1989: 105-106) states that although "there is no such thing as a typical computer, since computers come in lots of kinds, intended for different purposes and incorporating different patterns of construction", it is still possible to identify "certain fundamental components which any computer will contain in some shape or other". These components, together with those which interface with the external parts, are at least five in number: (1) the central processing unit (CPU), (2) the memory, (3) the input devices, (4) the output devices, (5) the communication paths or buses.
The central processing unit is the most important part of the computer. It is contained on a single chip, i.e., a small piece of silicon with an integrated circuit etched onto its surface. This CPU is divided into subcomponets: (1.1) the arithmetic-logic unit (ALU) which performs basic operations such as addition, subtraction, multiplication, etc., (1.2) the accumulators, which hold the numbers currently in use by the ALU, (1.3) the clock, which syncronizes those activities of the computer which need to be carried out in step with others, by emitting a regular pulse, and (1.4) the registers. Registers are furthermore subdivided into seven subcomponents: (1.4.1) a memory address register, (1.4.2) a memory date register, (1.4.3) some status registers, (1.4.4) some general purpose registers, (1.4.5) a program counter, (1.4.6) an instruction register, and (1.4.7) a stack pointer.
According to the description given by Aho and Ullman (1992: 144), the computer together with its software functions in a hierarchy of abstractions called virtual machines or levels. Each level, with the exception of the lowest level, is implemented by translating or interpreting the instructions at that level by means of instructions or facilities from lower levels. There is an increasing degree of abstraction, as follows: (1) electronic circuitry or digital logic, (2) microprogram, (3) machine language, (4) operating system kernel, (5) assembly language, (6) programing language, (7) application program.
All the external and internal parts, components, and subcomponents, outlined so far, justify the description of the computer not merely as a machine, but as an equipment with reference to the physical resources, implements, assets, and machinery that are used whenever someone puts the computer into operation or activity. Even more adequate than equipment, however, is the description of the computer set as an apparatus, a very general term that covers instruments, tools, machines, and appliances, including the idea of a system or process, all of them used specially for scientific or technical purposes.
Although the term apparatus is sufficiently generic to suggest a collection or a set of materials, the complex of instrumentalities and processes involved in any computer system, the term apparatus only emphasizes the materiality of the system and fails to bring into the fore the more abstract aspects of the computer's functioning. More appropriate with respect to these aspects is the description of the computer as a medium.
2. The computer as a medium
The descriptive name for the computer which has been lately employed with some frequency is the term medium (cf. Andersen 1986, Bolz et al. 1992, Andersen et al. 1993, Nake 1994 and Hoppé and Nake 1995).
In its broadest sense, medium is a synonym of "means", a very general term applicable to anything employed in performing or executing to some end. Since the 1960's, a more specific meaning, widely used in the context of mass communication, came to be added to the former sense of the word medium. In the domain of communication theory, the word medium then became much discussed because of the writings of Marshall MacLuhan. In this context, medium began to be used in close association with "vehicle" and "channel". Vehicle indicates a means of conveying or communicating and is more specific and tangible than medium, while channel suggests the idea of a more physical path of transmission or communication than the idea of a means.
In the last three decades, with the enormous development of modern systems of communication, information, and entertainment, medium began to be substituted by its plural form "media", which appears most frequently in the expression mass media. The media now refer both to the systems of communication such as magazine, newspaper, radio, television, etc., and to a vehicle of advertising, such as in radio, TV programs, newspapers, etc.
Although the broad sense of medium as something that is employed as a means to achieve an end is perfectly applicable to the computer, it is in the more specific sense of a means of providing and communicating information for the public that the defining word medium has recently been applied to the computer.
In spite of the fact that there is yet no consensus as to the understanding of the computer as a medium (cf. Hoppé and Nake 1995), it cannot be denied that the processing and communication of data performed by the computer are forms of communication. K. Paulsell (1990: 195), for instance, observes that "even if no human being specifically originates or receives the information and even if the computer automatically processes the data without a specific instruction to do so in each case, communication is still taking place through the medium of the computer."
Besides the level of communicative dynamics which occurs inside the computer, there is more obviously the processes of communication between the computer and its users and the processes among users mediated by the computer (cf. Danet 1995:9). Furthermore, there is also communication among computers through a device which allows the computers to send data via a communication channel, and that is possible even for any microcomputer. Through a modem (a word derived from modulator/demodulator) any computer can be connected and transmit signals via telephone wire to any other computer. Actually, "computer communications can travel in several ways. Between rooms, buildings, cities, or continents, they can travel over ordinary voice telephone lines, over leased computer circuits, or over special computer networks called packet switching networks" (Paulsell 1990: 200).
The new transmission technologies, the new telecommunications channels (satellites, fiber optics, etc.) connected with the computer are creating gigantic computerized networks which instantaneously link any part of the globe to any other (cf. Demac 1990). In view of the planetary proportion of such communicative scenario, it becomes almost impossible to deny that the computer can really function as a communications medium. Indeed, there even exist well-known computer communications applications such as computerized financial transactions, eletronic mail, computer conferencing, and on-line data base services, "where computers equipped with large-capacity disk drives can keep vast amounts of data that can be accessed in a few moments on-line" (Paulsell 1990: 200). Not only the kinds of information stored in computers and accessible over telecommunications links continue to grow, but also the communications devices become increasingly sophisticated.
Other kinds of devices that have recently contributed to creating the idea of the computer as a medium, now in the realm of the software that can run within any personal computer, are the various programs for computer graphics, multimedia, hypertext, and hypermedia. Such programs permit the production of sophisticated kinds of verbal, visual, and audio messages which can also be transmitted through electronic mail creating an entirely new idea of electronic on-line publication. Im sum: the computer's capacity to transform all data, video, and voice information into electronic pulses is evidence that the computer is not only a medium, but also well on the way to turning into the medium of all media. Besides being a medium of communication, or precisely because of its being such a medium, the most comprehensive characterization of the computer is the one that defines it as a semiotic medium (Andersen and Mathiassen 1986, Andersen 1986, 1990a, 1990b, 1991, 1992, 1993, 1995, Andersen and Holmqvist 1990, Meunier 1989).
3. The computer as a semiotic medium
There are two main kinds of investigation that characterize the computer as a semiotic medium: those which are only implicitly semiotic and those which are explicitly semiotic.
3.1 Implicit computer semiotics
In 1972, Newell and Simon put forth the notion of physical symbol systems to focus on how people solve problems, given that they are symbol manipulation systems. Later on, in 1980, Newell restated the fundamentals of physical symbol systems in a more systematic way. Newell defined this concept, which emerged from his experience and analysis of the computer and how to program it to perform intellectual and perceptual tasks, as follows: a physical symbol system is "a broad class of systems that is capable of having and manipulating symbols, yet is also realizable within our physical universe." The hypothesis is that these symbols, which are internal to the concept of a system, "are in fact the same symbols that we humans have and use every day in our lives", which means that "humans are instances of physical symbol systems, and, by virtue of this, mind enters into the physical universe" (Newell 1980: 136). After describing the functioning of a paradigmatic symbol system and after defining its essential nature, Newell (1980: 172-173) considers the digital computers as a key example of the realization of a symbol system in our physical universe. The originality of Newell's thesis, as Meunier (1989: 46) has pointed out, is that "it contrasts with a purely materialistic (if not reductionist) conception of artificial intelligence", since
what çharacterizes the operations of a computer manifesting intelligent behavior are not numerical operations, regardless of how complex they may be, nor to an even greater degree, operations achieving sophisticated mechanical or even electronical manipulations. Rather, an intelligent computer is one that processes a special type of sign --- that is to say, symbolic ones. Hence, an artificial intelligence is a 'machine' whose 'rational' behavior consists of manipulating physical symbols.
Meunier puts emphasis on the radicality of Newell's thesis since it no longer situates artificial intelligence "within a theory related solely to material technology and engeneering. On the contrary, it takes AI out of such a theory and inserts it, whether one likes it or not, within a semiotic theory" (ibid: 46). With this understanding of computational systems as those systems which manipulate interpretable symbols, as early as the beginning of the 70's, even without making use of the term semiotics, Newell was, in an implicit way, already giving birth to the conception of the computer as a semiotic medium.
Also implicitly semiotic are the studies in AI which emphasize the problem of representation (see, for example, Bobrow and Collins 1975, Palmer 1978, Winston 1981, Minsky 1981, Rich 1983, Anderson 1983, Pylyshyn 1984, Winograd and Flores 1986, Jorna 1990). In fact, everything that concerns representation falls entirely into the scope of a semiotic investigation. The book by Winograd and Flores on Understanding computer and cognition (1986) is a good example in the field of AI, which stresses the relevant role performed by representation in programming. In one of the chapters of the book dedicated to the topic of computers and representation, while discussing how programming depends on representation, the authors state that "whenever someone writes a program, it is a program about something (...) there is some subject domain to which the programmer addresses the program" (ibid.: 84). The formal logical systems that are used by the programmers "set up correspondences between formulas in such systems and the things being represented in such a way that the operations achieve the desired veridicality" (ibid.: 85).
The most important way of characterizing the computer as a complex machine of interrelated levels and junctures of representation appears in Winograd and Flores's definition of the unique possibility of the digital computer to construct "systems that cascade levels of representation one on top of another to great depth" (ibid.: 87). When running a typical artificial intelligence program, the computer's behavior can be analyzed at any of the following levels:
(1) the physical machine, where the computer is a complex network of components which operate according to the laws of physics, generating electrical and magnetic activity.
(2) The logical machine, at whose level the components are logical abstractions represented by activity in the physical components.
(3) The abstract machine, which, in most of today's computers, is an abstract single sequential processor that steps through a series of instructions. Each instruction is a simple operation of fetching or storing a symbol or performing a logical or arithmetic operation, such as a comparison, an addition, or a multiplication. This is usually the lowest level at which the programmer has control over the details of activity.
(4) A high level language, which provides elementary operations at a level more suitable for representing real-world domains. Formulas of this level of language are converted by a compiler into a sequence of operations for the abstract machine. (5) A representation scheme for facts refers to the conventions or uniform organization of the symbol structures of the higher level language that represents facts about the world (ibid.: 87-89).
What should be retained about the above tower of levels is the fact that each lower level performs the function of representing the activities prescribed in the upper level, that is, the subject domain of any lower level is the next correspondent higher level. However, there is no transparency or item by item correspondence in these processes of representation. A single high-level language step, for instance, "may compile into code using all of the different machine instructions, and furthermore the determination of what it compiles into will depend on global properties of the higher-level code" (ibid.: 90). In sum: although there is no doubt that the computer is a semiotic machine, its semiotic operations are very intricate and interdependent, constituting a real semiotic network with complicated levels of referentiality.
3.2 Explicit computer semiotics
Since the mid-80's, both proposals of a semiotics of informatics in general and the computer as a semiotic medium in particular have been explicitly developed in the studies of P. B. Andersen (1986, 1990a, 1990b, 1991, 1992, 1993, 1995), some of them published jointly with L. Mathiassen (1986) and B. Holmqvist (1990). In his early article on "Semiotics and informatics: The computers as media" (1986: 64-70), Andersen describes computer systems still metaphorically as media. By analogy with newspapers, books, records, films, videos, or television, the computer is seen as a channel through which humans communicate, particularly according to two main types of transmission between user and computer component, as follows: (1) inputs are data flows from human component to computer component, (2) outputs are data flows in the opposite direction. The originality of the article is not as much in this proposal, as it is in the general theory of semiotics that the author uses as a frame of reference for the description of computers as media, especially through the analysis of computer based semiotic communities (ibid.: 72-82).
A broad perspective on computer semiotics, including the theoretical foundations of this new field and its applicability, can be found in Andersen's book on A theory of computer semiotics (1990a). Choosing the glossematics of Louis Hjelmslev as a basic semiotic framework and defending the two claims that computers are operated by means of signs whose meaning must be interpreted by the users and that computer-based work is sign usage, the author designs a general map of computer semiotics. The second part of the book is entirely dedicated to the study of computers as seen from a semiotic point of view, and the third part to language, work, and design.
Several aspects of a semiotic approach to computer systems have been objects of Andersen's study, including the aesthetics of hypertexts systems (1990b). In his recently published work on interactive systems (1995: 5), he presents the computer as an "elastic medium", that is, a medium whose main characteristic is the user's physical actions on it, where the "hand movements of the user of an interactive system must be an integral part of the meaning of the system." The semiotic understanding of the computer is remarkably extended by the latter point of view. The computer systems are not only seen as semiotic media operated by means of signs to be interpreted by the users, but the action of the user is also part of the computer production of signs in an interchangeable semiosis. But besides describing that level of interaction, Andersen (1995: 24-25) adds that
underneath the interface, in the intestines of the system, we also find signs. The system itself is specified by a program text (which is a sign, since it stands for the set of possible program executions to the programmer). The actual execution involves a compiler or interpreter that controls the computer by means of the program text, and since the compiler is a text standing for the set of permissible program texts, the compiler is also a sign -- in fact it is a meta-sign that in some versions very much resembles an ordinary grammar.
Indeed, any computer system is "a complex network of signs". At each of its various levels there are texts, and "as we change levels, the concepts signified by the texts change. On the lower levels, the meaning of the signs are related to the physical parts of the machine, like registers and storage cells". On the upper levels, the texts have to be interpreted differently, according to new software concepts such as stacks, heaps, and variables (ibid.: 24).
In a retrospective view, it is remarkable to notice how near the description, given by Newell (1980: 173-175), of the physical symbol systems with their series of levels of technology was to the "complex network of signs", mentioned by Andersen. In fact, Newell was not only able to recognize the symbolic reality of the computer systems, but he also foresaw the immense variety of physical ways of realizing any fixed symbol system.
Computer semiotics has also received some attention whithin the larger context of cognitive science (see Ouellet, ed. 1989 and Nöth 1994). According to Meunier (1989: 55), for instance, "AI projects seem so attached to computer technology that we tend to forget that their real originality lies in the complex semiotic system they put to work. AI is in fact an applied semiotic. It studies the functioning of a type of sign called symbols in a constructed or artificial system that is interpretable in cognitive terms."
For Ouellet (1989: 2), if we understand intelligence and knowledge as symbolic systems and processes, semiotics has two tasks related to this understanding.
The first is to establish what sort of syntax, semantics, and pragmatics is implied by the artificial and natural language of thought as instantiated in a machine or in a brain; this means that we have to investigate the nature and the function of the kinds of signs involved in symbolic representation systems and to scrutinize the way those signs (1) are related to each other, (2) can make sense by referring to the "external" world or to some "internal" representations such as intentions, beliefs, knowledge, etc., and (3) are used by an agent (human or mechanical) as a means to reach some specific goal or to undertake some special task. That is the theoretical function of the sign theory in the context of cognitive studies and Artificial Intelligence.
The second task of semiotics, which Ouellet characterizes as practical or empirical, in helping cogniticians in their exploration of the human or mechanical mind, is "to supply specific formal models of semiotic behaviors, such as discourse production and comprehension, story recognition, categorization processes, practical or logical reasoning, understanding of visual signs, etc., all of which are types of sign or information processing for which different fields of semiotic studies have developed metasemiotic representations, models, or grammars."
It is true that among computer scientists and cognitivists there is a consensus about the symbolic nature of the computers, of their containing formal symbols that can be manipulated by rules. More than a decade ago, Pylyshyn (1981: 68) already mentioned the growing "understanding of computational processes and of the digital computer as a general symbol." Polyshyn (1984) dates the consideration of the computer as an intellectual tool back to the early 1950s (cf. Turing 1950 and Shannon 1950).
In fact, any description of the computer is an evidence of its symbolic and cognitive character. Computers deal with data, programs, languages, and instructions which are stored in their main memory, that is, they deal with information at a variety of levels of abstraction, each of them having its own data model, in which the levels above are implemented. A brief list of some of the capabilities of the computer are another evidence of their ability to serve as arbitrary symbols. Abstractions of real world problems can be represented and manipulated inside the computer. "They can be programmed to simulate any physical system we like" (P. H. Winston 1981: 4), and their activities are clearly cognitive: they fetch instructions from main memory, decode them and execute them (Aho and Ullman 1992: 146). In sum: computers can solve difficult problems, can help experts to analyze and design, can understand simple English, can help manufacture products, can learn from examples and precedents, and can also model information processing (P. H. Winston 1981: 6-19).
Although the statements above clearly indicate the semiotic nature of the computer, it is also true that the concepts of the symbol in computer and cognitive science are, in general, rather vague and often even simplistic. Any definition of the symbol requires a semiotic foundation, which often lacks in the discourse on symbols in computer science. Semioticians have, in fact, developed highly complex theories of the symbol, but it is true that only little research has been done concerning the different types and mixtures of signs that occur at the various interrelated levels of computer systems, from the physical device level to the more evidently symbolic level of communication between programmers and computer and between computers and users.
What I am trying to suggest is that many descriptive and conceptual resources for the analysis of computer systems are still available from the semiotics of Charles Sanders Peirce. His definitions and classifications of the sign in all its levels of degeneracy are most pertinent, especially his different degrees of iconicity (see Santaella 1995: 141-155), his typology of indices, and also his complex notion of legi-signs and symbolicity. Symbolicity, by the way, does not necessarily imply only an arbitrary representation of the world out there, nor by necessity does it imply any kind of correspondence to actual states of the world. But this subject is too complex to be discussed any further here since the objective I have in mind is to develop an argument about Peirce's notion of the sign, which is at a more general level than that of the application of his types of signs to describe actual processes or webs of signs as they occur in and among the different levels of the computer system.
According to Peirce's definition, the essential features of the sign relations give rise to several layers of meaning. I will argue that these layers can help us to understand why the computer is a very complex kind of semiotic machine that functions at the same time as a physical object, a tool, a channel or a medium, and as a sign or mediation.
4. The sign as mediation
Whenever we speak about the sign in the context of Peirce's semiotics, according to the logical form of semiosis that he described, there are two senses for the word sign, an extensive and a specific one. The foundation of the extensive sense lies in his phenomenological category of thirdness. According to Peirce's phenomenology, there are, in anything whatsoever, in nature or in thought, three omnipresent or universal categories, that he came to call by the very general names of firtness, secondness, and thirdness.
The most basic characteristics of firstness, which is the monadic category, are chance, originality, spontaneity, possibility, uncertainty, immediacy, presentness, quality, and feeling. In secondness, which is the dyadic category, we find ideas related to polarity such as brute force, action and reaction, effort and resistence, dependence, conflict, surprise. Thirdness, or the triadic category, is linked to the ideas of generality, continuity, growth, evolution, representation, and mediation. Mediation, was finally considered by Peirce to be the most general characteristic of thirdness. "The mediation between Secondness and Firstness" (CP 5.121) was Peirce's definition of thirdness. "A Third," he stated in another passage (CP 8.332), "is something which brings a First into relation to a Second... A Sign is a sort of Third."
In many other passages, Peirce restated that one of the simplest kinds of generality or thirdness is the form of a sign. Here, the word sign is certainly being employed in its extensive sense, wich refers to the complete sign relation, taken as the irreducible triadic process of sign, object, and interpretant. In this sense, the sign is a third, functioning as a general synonym for thirdness or mediation, while quality is a synonym for firstness, and reaction for secondness (cf. CP 4.3). In 1867, Peirce also used representation as a synonym of thirdness, sign, or mediation. In 1898, however, he declared that he did not at that time [in 1867] "know enough about language to see that to attempt to make the word representation serve for an idea so much more general than any it habitually carried, was injurious. The word mediation would be better" than representation (CP 4.3). At this time, representation was for him already one species within the genus of mediation, as I will discuss later on.
In its second and more specific sense, the word sign refers strictly to the mediating term in the triadic relation, a mediating term that occupies the logical position of a first, while the object is a second and the interpretant a third.
In a letter to Jourdain, in the year of 1908, Peirce wrote: "My idea of a sign has been so generalized that I have at length despaired of making anybody comprehend it, so that for the sake of being understood, I now limit it" (cf. Fisch 1986: 342). The limited idea of the sign to which Peirce refers is simplified and less abstract. It defines the sign as something that represents something else to somebody. The insertion of this word "somebody" in the place of his more complex notion of the interpretant is, no doubt, what makes the definition easier to understand, but, at the same time, turns it less interesting for its application to processes of cognition and communication that do not depend on human consciousness such as those which occur, for example, in nature and in biological or in artificial intelligence phenomena.
In order to retain the full potentiality of Peirce's definition of the sign, we have to cover the opposite route which Peirce himself felt obliged to follow, that is, we have to consider the definition of the sign in its most abstract and generic level. What the most generalized definitions bring into the fore is the mediating function of the sign between object and interpretant and the relations of determination, of the sign by the object and of the interpretant by the sign. Since the three elements, sign, object, and interpretant, in themselves, or better, in their existential nature, may belong to various orders of reality, as single objects, general classes, fictions, mental representations, physical impulses, human actions, or natural laws, what constitutes the sign relation in its logical form is the particular way in which this triad is bound together (cf. Parmentier 1985: 26).
Let us begin the discussion of the mediating function with one of the abstract definitions of the sign:
In its genuine form, Thirdness is the triadic relation existing between a sign, its object and the interpreting thought, itself a sign, considered as constituting the mode of being of a sign. A sign mediates between the interpretant sign and its object (CP 8.332).
The action of the sign or semiosis is to function as a mediator between the object and the effect which the sign produces on an actual or potential mind. This effect or interpretant is indirectly due to the object through the sign. The mediation of the sign in relation to the object implies the production of the interpretant which, however much the chain of interpretants may grow, will always be due to the logical action of the object, that is, the action mediated by the sign. In this respect, the reference of the sign to the object does not depend on any personal interpretation. It is an objective property of the sign, a property which gives the sign the power of producing an interpretant whether the interpretant is in fact produced or not. It is for this reason that we cannot accept the expression "an effect produced in a mind" as being explanatory of the interpretant.
However, in order to understand mediation better it is necessary to consider the problem of determination, in the logical sense that Peirce gave to the verb "determine". Let us proceed with another quotation from Peirce:
A sign is a Cognizable that, on the one hand, is so determined (i.e., specialized, bestimmt) by something other than itself, called its Object, while, on the other hand, it so determines some actual or potential Mind, the determination whereof I term the Interpretant created by the Sign, that the interpreting Mind is therein determined mediately by the Object (CP 8.177).
The statement that the sign is determined by the object leads us to think that the object has real primacy over the sign. However, in the logical form of the triadic process, the object comes second in relation to the sign which is a first. The real primacy of the object is thus not to be confused with logical primacy. Although the sign is determined by the object, the latter is logically accessible only through the mediation of the sign. The object is something different from the sign and that explains why the sign cannot substitute the object, but can only stand for it and indicate it to the idea or interpretant that the sign produces or modifies. This means that the action of the sign can only be completed when the sign itself determines the interpretant. When the interpretant is created by the sign, it will be indirectly determined by the same object which determines the sign. This is why Peirce said (cf. Parmentier 1985: 28), that the action of the object upon the interpretant is "mediate determination", and that the interpretant itself is a "mediate representation" of the object, occupying therefore the logical position of a third in the triadic relation.
In sum, the sign determines the interpretant but determines it as a determination of the object. The interpretant as such is determined by the object insofar as the interpretant itself is determined by the sign. Furthermore, this triad implies a constant expansion of the process of semiosis since the interpretant, in turn, determines a further sign, becoming thereby a sign to that further interpretant. Semiosis is, thus, an infinite process or an endless series in a process that operates in two directions, "back toward the object" and "forward toward the interpretant" (cf. MS 599: 32).
Parmentier (1985: 27-29) stresses that "the sign relation is constituted by the interlocking of a vector of representation, pointing from the sign and interpretant toward the object and a vector of determination pointing from the object toward both sign and interpretant." The position of the sign is mediate between the object and the interpretant for both the vector of determination and the vector of representation. The sign itself faces simultaneously in two directions: it faces toward the object in a "passive" relation of being determined, and it faces toward the interpretant in an "active" relation of determining.
As it can be seen, while the mediating function of the sign is a general one, the representing function corresponds only to one of the vectors of the sign's mediating function. That is why representation is only one species within the multifaceted genus of mediation. This term mediation refers both to the triadic sign relation in general and to the middle term of that relation in particular. At the same time, this middle term or sign, which Peirce sometimes also called representamen, occupies the mediating position in the vector of determination and also in the vector of representation. The sign or representamen is thus a synthetic element, and to its mediating position all the semiotic relations converge. The sign is determined by the object, but simultaneously, it represents the object. The sign determines the interpretant, and in determining it, the sign transfers to the interpretant the task of representing the object through the mediation of the sign.
5. The sign as a medium
In the two senses of the term mediation discussed above, the extensive sense referring to the triadic sign relation and the specific one referring to the sign in itself or the middle term of the triad, the word medium is a synonym of mediation. In a passage where Peirce described his three categories, he asked the reader to "observe that a means, or medium, is a third", which is also "a Branching or Mediation" (NEM 4.307). In a letter to Lady Welby (SS: 32), he used again the term medium as a synonym of mediation but now referring to the sign itself as the middle, medium, means, or mediation" linking the object and the interpretant.
In his late writings, Peirce generalized his doctrine of mediation and medium still further by focusing on the notion of communication as an essential feature of all semiosis (cf. Parmentier 1985: 42). In its basic sense, which implies the idea of a medium of communication, mediation can be defined as "any process in which two elements are brought into articulation by means of, or through, the intervention of some third element that serves as the vehicle or the means of communication" (ibid.: 25). In fact, in any process of communication, there must be a medium or means through which a message is conveyed from one cognition to the next. It was the mediating function of the sign that led Peirce to claim that a sign is a species of a "medium of communication" between two ideas or between an object and an idea, or better, between an object and the interpretant idea that the sign produces or modifies. This is clearly stated in the following passage:
A Sign may be defined as a Medium for the communication of a Form. It is not logically necessary that anything possessing consciousness, that is, feeling of the peculiar common quality of our feeling should be concerned. But it is necessary that there should be two, if not three, quasi-minds, meaning things capable of varied determination as to forms of the kind communicated. As a medium, the Sign is essentially in a triadic relation to its Object which determines it, and to the Interpretant which it determines ... That which is communicated from the Object through the Sign to the Interpretant is a Form; that is to say, it is nothing like an existent, but is a power, is the fact that something would happen under certain conditions. This form is really embodied in the object, meaning that the conditional relation which constitutes the form is true of the form as it is in the Object. In the Sign it is embodied only in a representative sense, meaning that by virtue of some real modification of the Sign, or otherwise, the Sign becomes endowed with the power of communicating it to an interpretant (MS 793: 2-4).
As Johansen (1993: 60) pointed out, this definition of the sign as a "Medium for the communication of a Form" implies the sign's dynamic or actional nature. Semiosis is "an action or influence, which is, or involves a cooperation of three subjects, such as a sign, its object and its interpretant" (CP 5.494). "This means," Johansen observes, "that a sign is a dynamic and mediating relationship between at least three interdependent positions, through which it produces meaning."
The dynamic and mediating character of the sign mentioned by Johansen indicate that the sign's function as a medium of communication has two aspects. There is the abstract layer which we have discussed above and also a more concrete one that will be analyzed in the following. In a manuscript, Peirce (MS 283: 128-130) wrote an elucidating passage which differentiates between the two aspects of the sign as a medium:
A medium of communication is something, A, which being acted upon by something else, N, in its turn acts upon something, I, in a manner involving its determination by N, so that I shall thereby, through A and only through A, be acted upon by N. We may purposely select a somewhat imperfect example. Namely, one animal, say a mosquito, is acted upon by the entity of a zymotic disease, and in its turn acts upon another animal, to which it communicates the fever. The reason that this example is not perfect is that the active medium is somewhat of the nature of a vehicle, which differs from a medium of communication, in acting upon the transported object and determining it to a changed location, where, without further interposition of the vehicle, it acts upon, or is acted upon by, the object to which it is conveyed. A sign, on the other hand, just insofar as it fulfills the function of a sign, and none other, perfectly conforms to the definition of a medium of communication. It is determined by the object, but in no other respect than goes to enable it to act upon the interpreting quasi-mind; and the more perfectly it fulfills its function as a sign, the less effect it has upon the quasi-mind other than that of determining it as if the object itself has acted upon it.
Peirce's concept of a medium of communication is the concept of the sign, in the sense of an abstract mediation, and the more abstract the medium is, the more it performs the role of a mediator, something that mediately determines or influences the interpretant by functioning "to deflect the emanation from the object upon the interpreting mind" (MS 634: 24). It must be noted, however, that the abstract function of the sign as a medium of communication does not exclude that it also functions as a vehicle. On the contrary, it implies and includes this function. In order to act as a mediation or medium of communication, the sign has to be embodied, it has to be materialized in a sensible vehicle or expressive form. Actually, Peirce always "insisted on the necessity of studying expressive forms or external representations rather than attempting to examine thought itself through some kind of unmediated introspection" (CP 1.551, cf. Parmentier 1985: 43). The importance of the embodiment of the sign for its action as a medium of communication can be clearly observed in the following quotation:
By a Sign I mean anything whatever, real or fictile, which is capable of a sensible form, is applicable to something other than itself, that is already known, and that is capable of being interpreted in another sign which I call its Interpretant as to communicate something that may not have been previously known about its Object. There is thus a triadic relation between any Sign, an Object, and an Interpretant. (MS 654: 7)
The idea that Peirce wanted to convey when he made the difference between the vehicle and the medium is the idea that the abstract or representational function of a medium is more complex than the one of a vehicle. But that does not mean that the sign may perform its communicative function independently of being embodied in some sensible vehicle. This interdependence of the medium and the vehicle can be better understood when we consider the two interrelated forms of causality which are basic for the comprehension of Peirce's semiotic philosophy.
With the exception of some Peirce scholars such as V. Potter (1967), J. Ransdell (1977, 1981, 1983), T. Short (1981, 1983), L. Santaella (1992, 1994), and H. Pape (1993), besides some passages in Johansen (1993), the role played by Peirce's concept of final causation and its counterpart, efficient causation, for the understanding of sign processes has not received the attention it deserves.
For Peirce, there are two types of forces or actions in the whole universe: (1) dyadic action, which is mechanical or dynamic, and (2) triadic action which is intelligent or signic. Peirce equated dyadic action with efficient causation, also called brute causation, and triadic action with final causation. Peirce's notion of efficient causation is the one of an effectively brute action, blind, nonrational, belonging to the hic et nunc, singular in occasion. Final causation, on the other hand, is the type of causation that is exercized by laws as opposed to "forces". It is logical causation, the causation of mind (CP 1.250). Ransdell (1977: 163) states that final cause is the overall form of a process, it is the tendency toward an end state, and "the general features of such a tendency in whatever medium the process may be realized. (...) The idea that living processes exemplify some such form is widely recognized nowadays under other and more acceptable labels, such as 'cybernetics' and 'homeostasis'." More recently, some new labels such as "teleonomy", "autopoiesis", "self-organization", etc., may be also added to this list.
The aspect of highest originality in Peirce's conception of final cause, however, is that final causality does not preclude efficient causation. On the contrary: both are compatible to the extent that the final causational form of any process can only be realized through efficient causation, which implies that being goal directed does not mean that such a process is separated from the mere physical aspect. It rather means that the former depends on brute and physical force for its realization. Although they are distinct types of action, one is dyadic -- blind, the other is triadic -- intelligent, they are inseparable. Peirce said that "final causality cannot be imagined without efficient causality; but no whit the less on that account are their modes of action polar contraries" (CP 1.213, cf. Santaella 1994: 406-407).
These two inseparable modes of action, by efficient and final causation, are the two types of action that characterize semiosis: the sensible or material form of the sign, its outward expression, that which enables it to act in a process of communication, corresponds to its efficient causation, while its complementary logical and mediating role corresponds to the aspect of final causation. This means that to exercize its logical and mediating power, the sign needs to be physically embodied. The material bodies of the signs are responsible for the communicative processes, for transmitting information from a certain source to some destiny. They function as the physical means, as the vehicle through which information travels. But at the same time, through that active medium, the sign is a representation, playing the role of an abstract mediation conveying meaning from the object to the interpretant.
The discussion of the two senses of sign or of mediation as thirdness and as the middle term or medium in the semiotic chain (see section 4. above), and also the above discussion of the two interrelated aspects of the medium as a physical body and as a representation can now help us to understand why the computer is simultaneously a sign, mediation, medium, and vehicle.
6. The various facets of the computer
There can be no doubt that the computer is a genuine sign in Peirce's sense, that is, it is a thirdness or mediation. Among all the different types of instruments, devices, and machines that have been invented by mankind, the computer is the first that can be semiotically characterized as a genuine third or sign. Although other types of technical machines such as photographic and cinematographic cameras, radio and television sets, sound recording devices, etc., have also the nature of signs, they are degenerate kinds of signs. The semiotic comparison of these different types of machines would carry us away from the objectives of this paper, so that I have to limit myself to the semiotic analysis of the computer. Why can the computer accomplish the most complex function of a sign, while the other kinds of machines cannot?
Similarly to all other types of machines for the recording, registering and transmission of sound and image, the computer is a semiotic machine. In contrast to those, however, the computer's semiosis is the only one that, in itself, independently of the users' processes of interaction and interpretation, can reach the level of the most complex of all signs, the symbol. As a result of its semiotic complexity, the computer can perform the mediating role, the role of thirdness, in its entirety, that is, it can fulfill literally and not only metaphorically the epistemological function of modeling the world.
Real world problems are abstracted by the computer scientists with the help of theories, and then these abstractions, which are also called knowledge, are symbolically represented and manipulated inside the computer. Brown (1989: 111-112) observes that the knowledge any intelligent system needs may be broadly divided into two parts: items of knowledge and knowledge structures. "The items are the individual things that the system is said to know something about" such as "objects, properties of objects, relations between objects, numbers, geometrical figures, and so on. (...) But we also want to know how to put this knowledge together, and this is where knowledge structures come in." The main forms of knowledge structure are: (1) state space, an arrangement of facts which allows the system to know where it can and can't go immediately from the state it is currently considering; (2) procedural representation, which allows the system to find its way around by an hierarchical arrangement of procedures, i.e., small chunks of a program; (3) production systems, which uses a series of productions, which are rules saying that if something is the case, then such and such is to be done; (4) frames, which are, metaphorically, rather like arrangements of little pigeon holes (ibid.: 112).
We have to consider, however, that any knowledge is knowledge about something. Thus, any theory or knowledge is already a representation, a kind of modeling of the world. The notational and logical systems, which are used to translate the facts and theories about the world into a kind of language that is accepted by the computer, are representations of a second level, namely representations of representations. It is important to consider that there are representations from the point of view of the processes defined on them and representations from the point of view of the notations used to express them (Anderson 1983: 46). Besides being representations of facts and theories about the world, these notation and logical systems are also representations of different kinds of reasoning or mental faculties. This explains the variety of logical systems that are used in artificial intelligence for representing and manipulating information such as predicate logic, nonmonotonic logic, probabilistic reasoning, modal and intensional logics, and fuzzy logic (cf. Rich 1983, Charniak and McDermott 1984).
The key role of representation (cf. Winston 1981: 21-24), the explanatory role of representations and the appeal to representations (cf Pylyshyn 1984: 23-32) have been much stressed in artificial intelligence. The modalities of representation are various. There are, for instance, types of representation or codes such as the ones proposed by Anderson (1983: 45-85): (1) a temporal string, which encodes the order of a set of items; (2) a spacial image, which encodes spacial configuration; (3) abstract proposition, which encodes meaning. There are furthermore the two well-known branches of programming languages, also called representations of knowledge: the declarative and the procedural branches, whose distiction is based on two types of knowledge (Winograd 1975: 185-210). The first of them, the declarative knowledge refers to facts we know, and the second, the procedural refers to skills we know how to perform (Anderson 1983: viii). Just as in some languages we have declarative sentences and imperative sentences, so there are two kinds of computer languages. "One sort uses declaratives, telling the computer that something or other is the case, and the other sort uses imperatives telling it to do something" (Brown 1989: 112).
It cannot be denied that representation is a key concept for the theorists of artificial intelligence and cognitive science. Unfortunatelly, however these theorists take as their starting point only a broad and vague notion of "knowledge representation". This obliterates the double metalevel of representation of the formal systems and furthermore the real nature of the computer as a genuine sign or mediation whose semiotic objects are already genuine signs, belonging to the universe of thirdness. However, as the result of a kind of intuition about the complex chain of mediations in which the computer is inserted, computer scientists are able to recognize that respresentations in artificial intelligence are stylized versions of the world (cf. Charniac and McDermott 1984: 8).
This topic is very relevant for the analysis of the computer as a semiotic medium, but to discuss it further would take us too far. For our argument, it is sufficient to stress that all information processed by computers is not a mere purposeless exercise, but, on the contrary, it is always in the service of ends; and this is true to the extent that the power of an artificial intelligence system can be measured by its ability to achieve stated ends in the face of variations, difficulties, and complexities posed by the task environment (cf. Newell and Simon 1988: 37).
Every tool, instrument, and machine is designed for particular purposes to attain particular ends, but while some of them are extensions of our physical force and others extensions of our organs of the senses, computers are extensions of our brain. With their ability to store and manipulate symbols, the computers imitate the mind in its capacity to function as a medium of computation and as a medium of representation (cf. Jorna 1990: 195).
Besides being a medium or mediation in the most abstract sense of mediation, the computer also functions as a sign in the second sense that Peirce conferred to that word, that is, as the middle term in the triadic relation of sign, object, and interpretant. The most basic semiosis in which the computer evinces the position of the sign is the one of the processes and operations that are developed strictly inside the computer. Here the computer cannot be seen as a simple sign but as a very complex one, given the various levels of determination and representation of its internal semioses. There are also two other basic semioses according to which the computer functions as the interpretant or as the object of the sign respectively. These are computer semiosis as seen (1) from the point of view of the programmer and (2) from the point of view of the users.
Winograd and Flores (1986: 84-92) have given us a description of some of the representational aspects involved in programming. This description can help us to map computer semiosis from the point of view of the programmer. The first aspect concerns the referential character of computer programs, which are always programs about something, programs of some subject domain to which the programmer adresses a certain program. The subject domain is the semiotic object that determines the sign, which, in this case, is the program. The sign, in its turn, represents that object, the subject domain, to a certain extent and under certain capacities.
The expression "to a certain extent" refers to the fact that the program does not represent the subject domain in all its aspects, but only in those that are under consideration. The expression "under certain capacities" refers to the abilities of the sign to represent its object. In programming, these abilities are dependent on the logic system that is being used and on the extent to which the formulas of the system set up correspondences to the state of things being represented. The representation and set of operations designed in the program have to be veridical. They are supposed to produce results that are correct relative to the domain.
The program as a sign operates typically under the form of final causality. Its application is directed toward an end state, a goal. The computer has to accomplish certain tasks according to a general design. Programs also have to be effective. They are effective depending on how efficiently the computational operations are carried out. When operating the design of the program, the computer acts as its interpretant. Even when the computer ends up operating successfully within a domain totally unintended by the programmers who constructed its programs, it still acts as the interpretant of these programs -- a creative interpretant, by the way -- as well as it still operates under the form of final causality. The overall form of a final causational process does not imply that its end must be predetermined. There is a tendency toward an end. However, as the process cannot escape the influence of objective chance and pure possibility, the end is never closed. The more complex the subject domains and the formal systems for their representation are, the less deterministic and more open to the interference of chance the end states can be.
In the second type of semiosis, where the computer functions as the sign or the middle term of the triadic relation, the programs designed by the programmers are the semiotic object of the computer, while the outputs generated by the computer operations are the interpretants. From this point of view, there are so many levels of representation, there is such a web of signs operating inside the computer that we are led to view its internal semiosis as a case of compound semiosis. Winograd and Flores (1986: 86-89) present a clear overall description of the cascade levels of representation one on top of another that constitute the internal operations of the computer system. As these levels of representations have already been mentioned above (3.1), Newell and Simon's (1981: 35-66) synthesis of "physical symbol systems" will be taken here as a general frame of reference for a sketch of the compound semiosis occurring inside the computer. According to this view (ibid.: 64):
Symbol systems are collections of patterns and processes, the latter being capable of producing, destroying, and modifying the former. The most important properties of patterns is that they can designate objects, processes or other patterns, and that, when they designate processes, they can be interpreted. Interpretation means carrying out the designated process. The two most significant classes of symbol systems with which we are acquainted are human beings and computers.
The most important feature of this definition is that it stresses the symbolic character of the internal computer semiosis. The hierarchy of abstractions, also called "virtual machines" (cf. Aho and Ullman 1992: 143), beginning with the underlying circuitry and progressing through the machine language to the operating system, the programing languages and eventually the aplication packages run on the machine, are all of them, in reality, symbolic patterns and processes, interrelated by means of internal referentiality, and interpreted in terms of operating rules. Even at the most elementary physical level, the computer already deals with symbols. Any pattern of impulses or electrical states is already a representation of numbers. In sum: any activity which is processed by the computer is a symbolic activity.
Newell and Simon (1981: 40) contribute two central notions to the definition of symbol systems, namely, designation and interpretation. "An expression designates an object if, given the expression, the system can either effect the object itself or behave in ways depending on the object". Thus, the essence of designation lies in the access to the object via the expression. Interpretation means that "the system can interpret an expression if the expression designates a process and if, given the expression, the system can carry out the process."
Actually, designation, the relation of the symbol with the object to which the symbol is applicable, and interpretation, the effect produced by the symbol in a quasi-mind, are the two basic characteristics of Peirce's definition of symbols. Also basic is the arbitrary nature of the symbol. Thus, when Winograd and Flores (1986: 86) observe that there is nothing in the design of the computer machine or in the operations of its programs that depends in any way on the fact that the symbol structures are viewed as representing anything at all, they are precisely confirming the basic arbitrary character of the symbol. That the representation is in the mind of the beholder, as Winograd and Flores (1986: 86) observe, corresponds almost entirely to Peirce's definition of the symbol as that type of sign in which the sign and the represented object are related only because the interpretant represents them as related.
However, according to Peirce, the interpretant can never be restricted to the mind of one beholder. The interpretation produced by a single mind is merely one and not even the most important kind of interpretant. The most relevant interpretant consists of the general or logical rules of interpretation that the interpreting mind actualizes in the act of interpretation. Without the effectiveness of the systems of rules that design the operations of the computer, there could be no processes of designation and interpretation among the layered patterns and structures inside the computer.
Another fundamental aspect of the semiotic definition of the symbol which is also pointed out by Newell and Simon (1981: 64) is its evolutionary character. "A physical symbol system", they state, "is a machine that produces through time an evolving collection of symbol structures. Such a system exists in a world of objects wider than just these symbolic expressions themselves." This idea that symbols grow, and that such growing presents a certain level of autonomy was stressed by Peirce in several occasions.
If it is true that computer scientists disagree little that computers manipulate symbols, it is also true that the concern with the symbolicity of computer manipulated signs has blinded them to other types of signs without which the symbol could never work, namely icons and indices. No symbol can function as such unless it includes indexical levels of referentiality and iconic levels of signification. How can robotics be studied, for instance, if indices are not taken into consideration, and how can modeling and simulation processes be analyzed without the aid of the semiotic notion of the icon? The study of indices and icons, which are also semiotically operative in the internal systems, structures and patterns of the computer processses, is a chapter of its own, that is still waiting to be developed by semioticians, although a move ahead in that direction has been recently given by Nöth (1996).
The third type of semiosis is the one from the point of view of the users of the computer programs and the interpreters of its outputs. In this process, the operations inside the computer are the object of the sign, its outputs function as the sign and the behavior of the interpreters and users are the interpretants of the sign relation. This is the type of semiosis that has attracted more attention and interest from semioticians, most probably as a result of the assumption that there can only be semiotic processes when there are sign users. This is another very pertinent topic for a semiotic discussion which unfortunately goes beyond the scope of this paper.
Let me conclude my argument concerning the nature of the computer as a medium. If we consider the term sign in Peirce's sense as a synonym of medium, the computer is, in fact, a very special type of medium, since all the meanings that Peirce gave to the word sign are applicable to the computer. It is a mediation or thirdness. It is also a sign or medium in the triadic relation of sign, object, and interpretant. Furthermore, the computer can act as the interpretant in another process of semiosis, and finally it can also be the object in still another semiotic process.
The fact that the computer is a sign in the sense of medium, does not preclude, but includes its being a sign vehicle. After all, from a certain point of view, the computer can simply be considered as a mere physical machine with a complex network of components such as wires, integrated circuits, and magnetic disks. These components, which operate according to the laws of physics, can be described in terms of electrical impulses traveling through a network of electronic elements. The computer as such a physical machine and vehicle is, in fact, the sensible form, the material object which embodies a semiotic medium and thus a sign in all its complexity.
Aho, Alfred V. and Jeffrey D. Ullman (1992). Foundations of Computer Science. New York: Computer Science Press, W. H. Freeman and Company.
Andersen, Peter Bogh (1986). Semiotics and Informatics: Computers as media. In Information Technology and Information Use: Towards a Unified View of Information and Information Technology, P. Ingwersen, L. Kajberg and A. Mark Peitersen (eds.), 64-97. London: Taylor Graham.
------------------- (1990a). A Theory of Computer Semiotics: Semiotic Approaches to Construction and Assessment of Computer Systems. Cambridge: Cambridge University Press.
------------------- (1990b). Towards an aesthetics of hypertext systems: A semiotic approach. In Hypertext: Concepts, Systems, and Applications, A. Risk, N. Streitz, and J. André (eds.). Cambridge: Cambridge University Press.
------------------ (1991). A Semiotic approach to construction and assessment of computer systems. In Information Systems Research: Contemporary Approaches & Emergent Traditions, H. E. Nissen, H. K. Klein and R. Hirschheim (eds.), 465-515. Amsterdam: North Holland.
------------------- (1992). Computer Semiotics. Scandinavian Journal of Information Systems 4, 3-30.
------------------- (1993). A semiotic approach to programming. The Computer as Medium, Peter B. Andersen, B. Holmqvist, and J. J. Jensen (eds.). Cambridge: Cambridge University Press.
------------------- (1995). The force dynamics of interactive systems: Toward a computer semiotics. Semiotica 103-1/2, 5-45.
Andersen, Peter Borgh and Holmqvist, Berit (1990). Interactive fiction: Artificial intelligence as a mode of sign production. AI and Society 4, 291-313.
Anderson, John R. (1983). The Architecture of cognition. Harvard University Press.
Bolz, Norbert et al. (orgs.) (1992). Computer als Medium. München: Wilhelm Fink Verlag.
Bobrow, Daniel G. and Allan Collins (eds.) (1975). Representation and understanding. Studies in cognitive science. New York: Academic Press, Inc.
Brown, Geoffrey (1989). Minds, brains and machines. New York: St. Martins Press.
Charniac, Eugene and Drew McDermott (1984). Introduction to Artificial Intelligence. Reading, Massachusetts: Addison- Wesley Publishing Company.
Danet, Brenda (1995). SRB Insights: Computer mediated Communication. The Semiotic Review of Books, volume 6, 9-10.
Demac, Donna A. (1990). New communication technologies: A plug 'n' play world? In Questioning the media: A critical introduction, John D. Downing et al. (eds.). London: Sage.
Fisch, Max (1986). Peirce, semeiotic, and pragmatism: Essays by Max H. Fisch, K. L. Ketner et al. (eds.). Bloomington: Indiana University Press.
Flusser, Vilem (1985). Filosofia da caixa preta. Ensaios para uma futura filosofia da fotografia. São Paulo: Hucitec.
Hoppé, Angelika and Frieder Nake (1995). Das allmähliche Auftauchen des Computers als Medium: Ergebnisse einer Delphi-Studie. Informatik, Bericht Nr 3/95, Universität Bremen.
Johansen, Jorgen Dines (1993). Dialogic semiosis: An essay on signs and meaning. Bloomington & Indianapolis: Indiana University Press.
Jorna, René J. (1990). Knowledge representation and symbols in the mind. Tübingen: Stauffenburg Verlag.
Mathiassen, Lars and Andersen Peter Bogh (1986). Semiotics and informatics: The impact of computer-based systems upon the professional language of nurses. In Readings on cognitive ergonomics -- Mind and computers, van der Veer et al. (eds.), 226-247, Berlin: Springer-Verlag.
Meunier, Jean-Guy (1989). Artificial intelligence and sign theory. Semiotica 77-1/3, 43-63.
Minsky, Marvin (1981). A framework for representing knowledge. In Mind Design, John Haugeland (ed.). Cambridge, Massachusetts: The MIT Press, 95-128.
Nake, Frieder (1994). Zeichen und Gebrauchswert: Beiträge zur Maschinisierung von Kopfarbeit. Informatik, Bericht Nr. 6/94, Universität Bremen.
Newell, Allen and Simon, H. A. (1972). Human Problem Solving. Englewood Cliffs, New Jersey: Prentice Hall.
Newell, Allen (1980). Physical Symbol Systems. Cognitive Science 4, 135-186.
----------------------------- (1981). Computer science as empirical inquiry: Symbols and search. In Mind design, John Haugeland (ed.). Massachusetts: The MIT Press, 35-66.
Nöth, Winfried (1994). Semiotic foundations of the cognitive paradigm. Semiosis 73, 5-16.
-------------- (1996). Representation in Semiotics and in Computer Science. Paper presented in the Seminar on "Semiotics and Informatics", Schloss Dagstuhl, Germany, 19-23 February 1996.
Ouellet, Pierre (1989). Introduction. Semiotica 77-1/3, 1-3.
Palmer, Stephen E. (1983). Fundamental Aspects of Cognition Representation. In Cognition and categorization, Eleanor Rosch and Barbara B. Lloyd (eds.). Hillsdale: Erlbaum, 259-303.
Pape, Helmut (1993). Final causality in Peirce's semiotic and the classification of the sciences. Transactions of the Charles S. Peirce Society 29 (4), 581-607.
Parmentier, Richard J. (1985). Sign's place in medias res: Peirce's concept of semiotic mediation. In Semiotic mediation, Elizabeth Mertz and Richard J. Parmentier (eds.). Orlando: Academic Press, 23-48.
Paulsell, Karen (1990). Computers and communication. In Questioning the media: A critical introduction, John D. Downing et al. (eds.). London: Sage, 195-205.
Peirce, Charles S. (1931-1958). Collected papers of Charles Sanders Peirce, 8 vols., C. Hartshorne, P. Weiss, and A. Burks (eds.). Cambridge, Mass.: Harvard University Press. The Collected Papers are here referred to as CP. [MS refers to Peirce's unpublished manuscripts as paginated by ISP, Texas].
---------------- (1976). The new elements of mathematics, 4 vols., Carolyn Eisele (ed.). The Hague: Mouton. The new elements are here referred to as NEM.
--------------- (1977). Semiotics and significs: The correspondence between Charles Sanders Peirce and Victoria Lady Welby, C. S. Hardwick (ed.). Bloomington: Indiana University Press. Semiotics and significs is here referred to as SS.
Potter, Vincent (1967). Charles S. Peirce. On norms and ideals. The University of Massachusetts Press.
Pylyshyn, Zenon (1981). Complexity and the study of artificial and human intelligence. In Mind design, John Haugeland (ed.). Massachusetts: The MIT Press, 67-94.
--------------- (1984). Computation and cognition. Toward a foundation for cognitive science. Cambridge, Massachusetts: The MIT Press.
Ransdell, Joseph (1977). Some leading ideas of Peirce's semiotics. Semiotica 19-3/4, 157-178.
---------------- (1981). Semiotic causation. A partial explication. Proceedings of the C. S. Peirce Bicentennial Congress. K. L. Ketner et al. (eds.). Lubbock, TX.
---------------- (1983). Peircean Semiotics. Draft of a work in progress.
Rich, Elaine (1983). Artificial Intelligence. MacGraw-Hill Book Company.
Santaella, Lucia (1992). A assinatura das coisas. Peirce e a literatura. Rio de Janeiro: Imago.
---------------- (1994). Peirce's broad concept of mind. S European Journal of Semiotic Studies, 6 (3/4), 399-411.
--------------- (1995). A teoria geral dos signos: Semiose e autogeração. São Paulo: Ática.
Shannon, C. E. (1950). Programming a computer for playing chess. Philosophical Magazine 41, 256-275.
Short, Thomas (1981). Peirce's concept of final causation. Transactions of the Charles S. Peirce Society, Fall 1981, 368- 382.
------------ (1983). Teleology in Nature. American Philosophical Quarterly 20 (4): 311-319.
Turing, A. M. (1950). Computing Machinery and Intelligence. In Mind. Reprinted 1964 in Minds and Machines, A. R. Anderson (ed.). Englewood Cliffs, N.J.: Prentice Hall.
Winograd, Terry (1975). Frame representations and the declarative- procedural controversy. In Language, thought, and culture. Advances in the study of cognition, Daniel G. Bobrow and Allan Collins (eds.). New York: Academic Press, 185-210.
Winograd, Terry and Fernando Flores (1986). Understanding computers and cognition: A new foundation for design. Norwood, New Jersey: Ablex Publishing Corporation.
Winston, Patrick Henry (1981). Artificial Intelligence. Massachusetts: Addison-Weley Publishing Company.