[Index] [Prev] [Next]

Comments to Semios-L

Steven Skaggs :
Specificity (long);
Wed, 26 Jun 1996 21:44:11 -0400

The article by Skaggs and Shank sans diagrams (hopefully to follow in a few

Codification and Inference in Visual Communication

Steven Skaggs
University of Louisville

Gary Shank
Northern Illinois University

Submitted to Zed
January  1996
(Running Head:  Codification and Inference)
Send Correspondence to:
Steven Skaggs
Department of Fine Arts
University of Louisville
Louisville, Kentucky 40292


This paper explores the idea that the linguistic phenomena of codification and the logical process of inference are related when specificity is considered as a common ground. A semiotic model drawing equally from the semiotic traditions of Saussure and Peirce is offered, drawing upon concrete examples from visual communication for illustration of important points.

Codification and Inference in Visual Communication

Semiotic theory has been divided between two competing viewpoints. One viewpoint sees meaning as arising out of the handling of arbitrary codes. The other viewpoint sees meaning as a process of logical inference. Visual communication, in particular the work of graphic designers who self-consciously manipulate visual elements in an effort to impart meaning in a very precise way, provides a special laboratory in which the value of these two viewpoints may be compared.

No one has ever succeeded in fully describing and reconciling either of these two traditions in the terms of the other. In every such case, something gets lost- an insight perhaps, or an analytical technique. It certainly is not our purpose to present another such attempt here. It is better to consider each of these viewpoints as different perspectives from which to observe the process of semiosis. We propose, instead, to look at semiotic action from the standpoint of certainty and specificity. We hope to demonstrate that a semiotic perspective based upon the concept of specificity provides a common ground for both codification and inference without becoming entwined in the conceptual networks particular to either of them. Meanwhile, the 'specificity model' provides insight into visual communication that warrants discussion irrespective to any benefits it may hold for reconciling two semiotic traditions.

Ultimately, the model we present here suggests that the manipulation of the viewer's powerful and compulsive need to understand the meaning of a display is a (perhaps the) prime ingredient in the graphic designer's craft. The model thereby promises to be a "player" in graphic design criticism.

Saussure and Peirce

Before we start with our model, we need to say a bit about the two main sources of semiotic theory. If you are familiar with their work, then feel free to go straight to our thought experiment in the next section. If you are not familiar with Saussure (1956) and Peirce (1931; 1955), however, please realize that our brief account is hopelessly oversimplified. We highly recommend turning to the primary sources to get these ideas properly presented and explained.

One branch of semiotic theory is grounded in the European efforts at the turn of the century to reconfigure the study of language. That effort was led by the Swiss linguist, Ferdinand de Saussure. Before Saussure, the science of linguistics was oriented toward the discovery of the historical evolution of languages and the relations of existing languages to each other. Saussure called this the diachronic approach to linguistics. He advocated the development of a synchronic approach, where the linguist looks at how the abstract and universal human tendency toward language is actually exhibited in a given language. In this approach, all elements of language are taken as parts of the larger system of language in use. In fact, all components of language are defined not in terms of some absolute standard, but by their relations to other components within the overall system. In the hands of Saussure's disciples, this concept developed into structuralism.

Saussure believed that language functioned as a symbolic system because the link between the sign and that which the sign signified was always arbitrary. Therefore, the system was free and flexible to account for and create new patterns of representations. The key components to Saussure's model of the sign was the notion of the interdependent link between the signifier and the signified to create the actual sign (figure 1). The signifier was that part of the sign relation that expressed or manifested the relation. In language, the actual word or statement is the signifier. The other component of the sign relation was the signified. For Saussure, the signified was the concept that the signifier pointed to. In his most famous example of this relation, Saussure considered the word 'tree' as the signifier, and the mental concept that the word engendered, including an image of a tree, as the signified (figure 2).

Saussure's basic model describes the dynamics of single reference, but other theoreticians built upon these basic insights. In particular, Barthes (1963) demonstrated that signifier-signified relations can serve as signifiers themselves for more complex signs. When a signifier is a simple reference, then the act of signification is called a 'denotation.' However, when the signifier is a prior sign or even network of signs, the relation to the new signified is much more complex. Such a relation is called a "connotation" (figure 3).

Saussure saw language as the premiere, but certainly not the only, sign system. Using language as a guide, we are able to talk about the 'languages' of such diverse things as kinship patterns, fashion, and food preparation. We can call such systems 'codes.' Within the semiotic literature, this notion of codes is a key part of the legacy of Saussure. In his work, Saussure saw codes as an interrelated set of signs that allow us to explain and understand our world. Some of Saussure's disciples have extended the notion of codes to such areas as myth (Levi-Strauss, 1978), folktales (Propp, 1968), and such diverse aspects of culture as popular culture and cultural icons (Barthes, 1957). The code user taps into a variety of signifying acts to create both connotative and denotative aspects. For example, the more that a given code is universally accepted in a culture, then the more denotative its use, and the more ambiguous a code, the more connotative its use.

The second branch of semiotics is grounded in American pragmatist theory. The founder of pragmatism, and also the founder of this branch of semiotics, is Charles Sanders Peirce. Any foray into Peircean thought is fraught with potential conceptual danger, since: 1) Peirce's thought is wide ranging; 2) each component in his system is related to all other components, making it difficult to isolate and talk about separate aspects of his thought; and 3) he had a tendency to speak in precise and abstract terms, thereby rendering his insights nearly opaque to the casual reader.

Any discussion of his semiotics, no matter how introductory, has to mention not only his model of signs, but his model of reality and his model of logic as well. We will touch on each area in turn.

We need to start with a brief outline of Peirce's theory of reality, since his model of signs and his model of logic reflect his understanding of reality. Building upon the Aristotelian notion of potency and act, Peirce expanded this characterization into a triadic model: potency, act, and relation (figure 4). Those aspects of reality which deal with and characterize pure potency he called Firstness. Firstness deals with issues of possibility.

Those aspects of reality which deal with and characterize pure action he called Secondness. Secondness deals with such things as the pure awareness that something is happening here and now, without knowing or understanding what it is that is happening. In other words, Secondness is the pure action-reaction relationship. But any system of Seconds embeds and contains the qualitative potential of Firstness.

Those aspects of reality which deal with mediation he called Thirdness. Anything symbolic, including language and sign systems in general, are real as Thirds because they necessarily involve mediated relations. Any system of Thirds embeds and contains the action/reaction of Secondness as well as the qualitative potential of Firstness (figure 5).

For Peirce, a sign is a mediator between the thing that it stands for (its "object") and the consequence of its representation (its "interpretant"). A word, in terms of its existence as a mark or a sound, is a sign only inasmuch as it stands for something to someone (or something!). It is the mediational role that is unique to signification. Interpretations are interpretants (though not all interpretants need be interpretations) (figure 6).

The notion of the role of inference in sign systems is the third, and for our purposes final, semiotic legacy of Peirce. Peirce laid out three distinct types of reasoning: deductive, inductive, and abductive. Deductive reasoning deals with drawing conclusions that are logically necessary and further our set of truth claims; inductive reasoning deals with drawing conclusions that are highly probable and which expand our domain of probable truth claims; and abductive reasoning deals with conclusions that are plausible, and which expand the domain of meaningful (if not certain) claims. As we move from abduction to deduction, we move from simple reconciliation of meaning toward the production of certain meaning.

Now that we have raced through a whirlwind look at Saussure and Peirce, we can apply their insights to our issue at hand: the use of these two semiotic models to inform our understanding of the design of visual displays.

Semiotic Analyses of Visual Displays

We begin where Charles Sanders Peirce began with his principles of Pragmatism: the difference between a doubt and a belief (See "The Fixation of Belief" (1877) and "How To Make Our Ideas Clear" (1878) included in Houser 1992.) The key point for the current discussion is this: that every conception is apprehended with a degree of certainty (belief) or uncertainty (doubt). The drive to resolve doubt governs our handling of all messages. There is a need to make sense, to believe, to hold a settled viewpoint. This is a process that is constant, a struggle that is the essence of mental life and is the tap root (as we hope to show) of articulate visual discourse. We will assume that this process of belief/doubt operates at every level of apprehension, from the most elemental functions perception, through rising degrees of unconscious and conscious activity, and ultimately to the construction and maintenance of one's general worldview. Researchers in the fields of visual perception (Marr), cognition (Vickers) and artificial intelligence (Josephson & Josephson) have pointed to mechanisms that allow such sensibilities to arise. The ensuing discussion will be framed within the environment of individual graphic messages but the concept is extendable to all varieties of semiotic action.

A thought experiment: Imagine that you interpret an expression with complete certainty. By saying "interpret with certainty" we do not mean to imply that the interpretation is 'correct', or 'accurate'. We simply mean that you take complete satisfaction that the apprehension of the message is decided, whole, fixed, specific: in short, that it is fully believed. We will call this state of certain apprehension the absolute interpretation. Now, it is possible that no interpretation may ever achieve such a state of pure and simple certainty, but we put it forward here only to function as a hypothetical limit of apprehension - a boundary to be approached, if not reached.

Now conjure another, counter, concept; a state of disbelief in which "the irritation of doubt causes a struggle to attain a state of belief"(Peirce. 1877; in Houser p.114). In our framework within the environment of individual messages, we might say that the interpretation is uncertain, flexible, loose, unspecific. Just as the entirely certain belief can be thought of as a hypothetical limit, so can we imagine a point of complete uncertainty - a place of infinite doubt (figure 7). Here, the interpretation is so nebulous that it remains totally undecided, fragmented, incomplete, slippery, open. Beyond lies complete incoherence, an inchoate land of mere Secondness, where sense collapses from triadic to dyadic relations and where signification fails. This point of complete ambiguity we term the threshold of semiosis.

All interpretation occurs between the absolute and this threshold. The region between these two limits maps the territory of Thirdness (the semiosic domain), the only place where the dance of signification occurs. This is a dance patterned by the varying degrees of certainty and specificity. In our thought experiment, we can imagine an interpretation's position along this semiosic continuum: a specific interpretation would be plotted toward the absolute side, a vague interpretation toward the threshold side.

This continuum of semiosis has a pair of distinguishing qualities. The first is that extremely specific interpretations (i.e: those near the absolute) tend to be clear cut and distinct, whereas extremely unspecific interpretations (i.e. those approaching the threshold) tend to be amorphous, hazy and vague. The other distinctive feature of this pathway through semiosis is that as one becomes less specific one discovers there are more potential relata for any given sign vehicle. At the point of absolute interpretation there is a single definite content. At the threshold of semiosis, there exists an infinite number of possible relata, each a potential interpretant. As the interpretation is revealed in an increasingly clear state, becoming more specific, we say that the interpretation is manifest. Conversely, in the case of a less distinct interpretation in which the number of potential relata increases, we say that the interpretation is manifold (figure 8).

Let us pause to consider a concrete example. Regard the logo illustrated in figure 9. It is doubtful that your interpretation of that logo is very certain or specific. Yet, simply because it is "nothing in particular" to you, it has the potential to mean almost anything. Because it is unspecific, it is manifold in that it suggests a myriad of potential relations. Your active mind, as long as you continue to regard the mark, will begin to supply associative 'guesses' and hinted connections.

Now consider the mark in figure 10. This mark is likely to be interpreted by you in a highly specific way. This clearness of interpretation makes the meaning manifest. But in so doing, the mark's very specificity shuts down the other possible relata.

Considering semiosis from the point of view of specificity accommodates competing semiotic traditions without translating one into terms of the other. The manifest and manifold aspects of interpretation can be shown to apply to semiotic action when considered from the point of view of coded systems, as in Saussurian-based semiology, and also from the standpoint of logical inference, as in Peircian semiotic. From the standpoint of code systems, highly manifest content is referred to as denotation; highly manifold content is refered to as connotation. From the standpoint of logical inference, highly manifest interpretations are referred to as deductive, while manifold interpretations are what Peirce termed the synthetic inferences - inductive and abductive.1

                         Manifest          Manifold

Saussurian (Code)        Denotation        Connotation
Peircian (Inference)     Deduction         Synthetic (Induction, Abduction)

We believe this synthesis forms a common ground upon which the competing semiotic schools may stand.

It also produces an important insight for educators and practitioners of visual communication. It highlights what is, we think, a key factor of design as a communicative act: it points to the potential power of eliciting from the interpreter a more intense engagement with the message. By placing an expression at a certain position along the continuum between absolute denotative clarity and complete connotative ambiguity, the designer enkindles the viewer's desire to achieve understanding by the resolution of doubt. The viewer enters into a kind of discourse with the expression, becoming an active participant in the exchange. A position near the deductive end of the scale produces a very quick resolution of doubt but pays a price in a paucity of triggered connotations. A position near the extreme abductive end of the scale produces a situation in which the near infinite number of possible relata render the possibility of resolution of doubt negligible.

Seen in this light, the designer's primary concern is not necessarily to clearly transfer message content from a sender to a receiver (as some information-theory models of design would say), but rather to select an appropriate position for the sign vehicle along the semiosic axis (figure 11). This in many cases may involve the purposeful excitation of doubt: unspecificity becomes an inherent part of the game.

There are a number of implications which follow from this argument, most of which lie outside the scope of what we can cover here. But we will discuss one that extends our philosophical model a bit further. Consider that as as an interpretation becomes less specific, the interpreter is asked to make a greater contribution of energy, time and work in order to supply the relata for the sign, whereas in situations where the sign is highly specific, logical inference or the code system relieves the interpreter of this effort (of course the interpreter must have already mastered a logical or linguistic system in order to reap the benefit).

For instance, if you are given the sign: WET PAINT, the sign will seem extremely denotative (of course you must understand the English language) and you will not lean on the rail upon which the sign hangs. The "system" the language code -carries the energy load in that action. If on the other hand, you encounter the sign TNIAP TEW in the same context, you must work to make some abductive inferences. Perhaps you would dismiss the sign (deciding that cost of the effort would not be offset by a benefit of continuing the process) to your eventual dismay upon resting against the railing. In connotative, unspecific situations, you must make an energetic series of guesses (abductions), each allowing you to be exposed to additional relata, but requiring an expenditure of your energy to do so.

It is as if the sign has a force attached to it which, if denotative, will focus your energies toward a specific rendering, but when connotative, compels you outward toward other possible relations. In the denotative case, the force or pressure flows inward or converges toward the specific interpretation. In connotative case, the pressure flows outward, diverges, toward other relata.

We consider those signs in which the interpretations are at the specific end of the continuum and in which the interpretation seems to flow inward under the pressure of system or deductive inference toward a manifest denotation to be influent signs. Those at the unspecific end, in which the interpreter's abductive energy creates the pressure necessary to flow outward toward manifold connotations are effluentt signs (figure 12). Influent signs emphasize a specific interpretant. Effluent signs emphasize possible connotations. The proper positioning within this influent-effluent axis is the crux of the graphic designer's task. All else is technique.

A corporate identity example

This has been a fairly abstract discussion. It is not possible to offer here a lengthy example of this model in practice, but with the hope of rendering the concepts more concrete, we would like to offer a very brief reading of one graphic design event in terms of our model.

Skaggs has pointed out in a previous number of Zed that design that is categorized as 'modern' tends to be highly denotative, while design that is of a 'postmodern' style tends to be highly connotative. Our model states that the points of absolute interpretation and semiosic threshold act as limits and therefore no matter how specific something may be it never actually attains a state of absolute interpretation. Also, no matter how manifold the connotations, communicative signs never reach a point of complete ambiguity. Neither influent modernist designs nor effluent postmodernist designs are as close to the extremes as one might think. The precise positioning requires more than a sense of following fashion trends: it requires sound judgement and care. Even so, this critical aspect of design process is a tricky business fraught with danger.

The logo in figure 13 was designed by Teresa Heintzman and Steven Skaggs in 1992 for Merchants, a chain of auto and truck servicing centers along the Atlantic Seaboard. The mark could be considered staunchly modernist in its approach: it hopes to be instantly recognizable, strongly denotative. Upon first viewing, the mark would denote simply an M-in-diamond. That is, it would be extremely recognizable as its basic component parts independently of association with Merchants. After a period of time, due to context and a campaign to aculturate the public, the mark would come to function in a completely denotative manner as an identifying signet signifying 'Merchants'. The goal was to have the mark be as strongly influent as possible, positioned toward the certain, absolute interpretation.

The mark never was released as the Merchants logo. A member of Merchants' Board, a person of strong religious and anti-gambling convictions, felt that the mark reminded him of poker and other games-of-chance involving playing cards. For that particular receiver, the first stage of the process, the simple denotation of an M-in-shape, was still ambiguous enough that he connected it with something that in his worldview was a great evil. The negative connotations that cascaded from this association completely blocked any possibility of proceeding to see the mark as an identifying signet meaning hie company. The mark was withdrawn.

What are the lessons to be learned from this? First, that the location along the semiosic axis is not only variable from one expression to another, but also is variable from one interpreter to another. Second, that one's worldview and experiences strongly influence the degree to which the expression is effluent and which particular connotations arise. Perhaps only one in a thousand people would have connected the proposed Merchants logo with gambling; we were unfortunate that the one person to do so was in a position to decide the fate of the design.


The sad little story of the Merchants logo points to the importance of considering the semiosic axis. It demonstrates how concepts and terms used here are at work in the apprehension of graphic displays. It also points out how much of what designers take for granted in the design process, the kind of conceptual underpinnings that always stay buried in one's subconscious as one works, have a theoretical 'reality' that can be very precisely defined in semiotic terms.

Of course, the story also shows that the semiosic axis that we have discussed in this paper is just a small part of the pattern of events that make up the interpretation of visual communication, and that it goes only a small way toward offering a picture of the whole pattern. No matter how promising semiotics may be for explaining communication events in extremely precise ways, the actual working out of the pattern will require a great deal more work. No matter that this 'specificity model' can form a common ground for certain aspects of two disparate semiotic traditions, the value of this model for graphic designers and others involved in the expression of communicative messages will have to await its proven contributions as a conceptual and lexical background for such creative work. Many more parts of the pattern must be constructed before we arrive at that point. We have yet far to go.


1. An inductive manifold interpretation suggests a state of affairs where we have a good sense of what the interpretation is, and we are holding the process open for further information to make sure we are on the right path. In other words, we are moving toward manifestness. An abductive manifold interpretation, on the other hand, is our attempt to pick one path of interpretation and follow from the myriad choices we have, and to subsequently justify and explain that choice. As you can see, as the interpretive landscape becomes more myriadic, abduction becomes more necessary in order for interpretation to continue.


Steven Skaggs

anyone, practical, software, faulty, defects
(five randomly chosen words from a computer manual)

pardon, brought, rum, falling, fruit
(five randomly chosen words from the Bible)

[Index] [Prev] [Next]

Comments to Semios-L