Investigations into semiotic theory typically begin with the fundamental question: what is a sign? A definition is then offered, usually quoted from an established authority (such as C. S. Peirce), to get the argument started. But an a priori definition immediately begs a basic methodological question: how does the author of such a definition know that he or she is right? How did C. S. Peirce for example know that a sign is what his celebrated definition (Peirce 1998: 135) says it is. He may well have been right but he gives us no proof of that. His and similar ex cathedra definitions used in semiotics are often little more than intellectual opinions and intuitions presented to the reader to be accepted on faith, but they are not logical conclusions deriving from clearly stated premises. So how else can we arrive at a logically valid and possibly useful understanding of signs?
I don’t know (yet) what a sign is, but let me postulate a definition of this fundamental semiotic concept by a series of deductive steps and intermediary supporting definitions of related concepts.
Step 1: I assume that whatever a “sign” is has to do with interactions between systems. A system is a set of interrelated elements (e.g. a tree, an animal, a human being), while interaction involves an exchange of information. Information is any difference in the physical states occurring in the environment. A difference between light and darkness is information, as is a difference between an empty page and a written page, between one speech sound and another, between two facial expressions and so on;
Step 2: Exchange of information between interacting systems is called communication. When systems interact remaining within the same spatio-temporal context and within each other’s natural sensory range we talk about contiguous communication, as when two animals fight or mate with one another, or when two people face one another. When interacting systems remain within the same spatial context but not within the same temporal context, we talk about spatially contiguous communication, as when an ancient artefact is examined by an archaeologist. When interacting systems remain within the same temporal context but not within the same spatial context, we talk about simultaneous communication, as when two people talk over a phone. Of the three distinguished types of communicative situations only the last two appear to be available to humans as functions of working memory and of technological extension of the senses;
Step 3: A physical change produced in the environment by a system is called an index of that system, as when an animal leaves behind an odour or a footprint. When index is spatio-temporally co-present with the system that caused it we talk about contiguous indexical communication, as illustrated by a cast shadow or mirror reflection. When systems communicate by means of an index spatially displaced from the system that caused it we talk about simultaneous indexical communication, as in live audio-visual satellite link used during TV news bulletins. When index is spatio-temporally displaced from the system that caused it we talk about displaced indexical communication, as exemplified by a fingerprint, a voice recording or a photograph;
Step 4: When a system produces a change in the environment that perceptually resembles some other system, such a change is called an icon, as exemplified by animal mimicry, linguistic onomatopoeia, and realistic representations of art. When icon is spatio-temporally co-present with the system it resembles we talk about contiguous iconic communication, as when a road sign showing a schematic image of a roundabout is situated at the roundabout itself. When systems communicate by means of an icon spatially displaced from the system it resembles we talk about simultaneous iconic communication, as when a person represented in a painted portrait is situated somewhere else than the portrait. When systems communicate by means of an icon that is spatio-temporally displaced from the system it resembles we talk about displaced iconic communication, as when a person represented in a painted portrait is now correspondingly older or is no longer living;
Step 5: When systems interact by means of a change in the environment that bears no perceptual resemblance to a system it is referring to, such a change is called symbol, as exemplified by human verbal language, phonetic alphabet, or mathematical notation. When symbol is spatio-temporally co-present with the system it is referring to we talk about contiguous symbolic communication, as when we point at a dog saying “Look, it’s a dog.” When systems communicating by means of symbols are spatially displaced we talk about simultaneous symbolic communication, as in a telephone conversation. When systems communicating by means of symbols are separated both in time and space from one another we talk about displaced symbolic communication, as when we are reading someone’s letter, or listen to a voice recording. While animal communication is mainly based on contiguous interactions with occasional indexical displacement, human communication relies heavily on indirect indexical, iconic, and symbolic communication involving spatio-temporally displaced communicators and reference;
Step 6: A system’s internal reaction to perceived information is called parainformation. In this sense parainformation constitutes the meaning of information. Consequently, information with parainformation attached to it is called meaningful information, while information on its own, with no parainformation attached to it, is called meaningless information. For example, a tree falling in the forest without anyone noticing the fact constitutes meaningless information, whereas when a forester finds the fallen tree and ponders what caused its fall the fact constitutes meaningful information;
Step 7: When communicating systems possess the same parainformation in relation to exchanged information we talk about an understanding between those systems, as when conversing people agree on the basic semantic meanings of the words they are using. When one of the communicating systems (sender) possess parainformation relating to exchanged information, but the other communicating system (receiver) possesses no parainformation relating to that information, we talk about incomprehension in the receiver, as when a person is confronted with an unfamiliar language. When the sender does not possess parainformation relating to communicated information but the receiver does, we talk about overinterpretation, or “overstanding,” as when someone interprets someone else’s unintentional gesture or facial expression as carrying some specific intentional meaning. When both communicating systems possess different parainformation relating to the same information we talk about misunderstanding, as when a person interprets a nod of the head as “yes” while a person from a different culture interprets the nod of the head as “no;”
Step 8: Some systems appear capable of reacting not just to perceived information but also to their own reactions (parainformation) to that information. A system’s internal reaction to parainformation is called metainformation. While parainformation always accompanies information as an immediate reaction to it, metainformation does not necessarily require any external stimulus because it is prompted from within the system, and is therefore not as constrained as is parainformation by the physical properties of perceived information. Metainformation thus tends to be less predictable, more elusive and subjective than parainformation. The distinction between parainformation and metainformation is analogous to what is often referred to as literal and metaphorical meaning, explicit and implicit meaning, or denotation and connotation, respectively. Metainformation underlies such cognitive phenomena as humour, irony, fantasizing, understatement, lying and double-talk. It appears that only humans are capable of communicating by means of metainformation;
Step 9: Information exchanged between systems using parainformation alone (as in non-human animals) can be of two kinds depending on the involvement of intentionality. Intentionality presupposes active and manipulative behaviour, motivated either genetically and automatically (as in animals), or both genetically and volitionally (as in humans). Thus unintentional information exchanged between systems using parainformation is called a cue, whereas intentional information exchanged between systems using parainformation is called a signal. For example, animals passively reveal phenotypic cues such as body size or weaponry (antlers, canine teeth and so on), and they also actively send signals in the form of warning calls, aggressive grunts, baring of teeth, bristling up and so on. Among animals cues and signals are exchanged mainly during direct, contiguous communication, and occasionally during indirect, indexical communication (as in leaving behind a scent to mark one’s territory);
Step 10: Information exchanged between systems capable of using both para- and metainformation (as in humans) can be of two kinds depending on the involvement of parainformation alone or of parainformation combined with metainformation. The first situation is described in Step 9. On the other hand information exchanged between systems using both para- and metainformation is called a sign (at last!). Signs can be exchanged during contiguous, indexical, iconic, and symbolic communication, they usually involve spatio-temporally displaced reference, and by virtue of their metainformational character they tend to be semantically elusive (as iconic signs used in art, or in abstract religious symbols).
Human interactions with other people and with the wider non-human environment involve a full suite of distinguished types of information, including cues and signals similarly to other animals, as well as intentional, culturally generated signs: indexes, icons, and symbols. Cues refer to such phenotypic traits as manifest bodily anatomy, including sexual dimorphism, skin colour and other racial characteristics. Signals include physiological reactions such as perspiration, blushing, yawning, coughing, sighing, crying and so on, as well as involuntary gestures and bodily postures. Finally signs include articulated speech, symbolic gestures (as in sign language), bodily adornments and visual arts. Metainformational signs are the stuff of human culture. When interacting with animals humans exchange both cues and signals, whereas when interacting with other people humans exchange both parainformational cues and signals and metainformational signs.
The reader will I hope excuse this rather lengthy lead-up to the proposed definition of a sign, so central to any semiotic theory. My purpose has been to demonstrate how one can arrive at an understanding of signs and communication by means of logical deduction from a premise about interaction between systems. I do not believe that one can credibly start a semiotic inquiry with a ready-made definition of a sign without first explaining the theoretical foundations on which such a definition is based. Understanding what signs are and how they work in the empirical world should be the goal of a semiotic theory, not its starting point. Ex cathedra definitions simply beg too many fundamental methodological questions to be accepted uncritically on the sheer strength of the historical authority of their propounders. While C. S. Peirce’s classic definition of a sign has been an inspiration to generations of semioticians, in the absence of intermediary logical steps explaining how Peirce arrived at his definition its acceptance remains a matter of intellectual conviction rather than of sceptical assessment.
Peirce, Charles Sanders 1998. Collected Papers of Charles Sanders Peirce. C. Hartshorne & P. Weiss (eds). Bristol: Thoemmes Press (1931-58).