How do we understand our world? We represent it, we usually think. We normally think that a word like "tree" or "chair" points at a non-linguistic entity in the world, that gives the word its meaning. Similarly, we might suppose that an image drawn or painted in a certain way points to a non-image, the "tree" in the real world. Plato certainly subscribed to this, and supposed that the images were second class at best.
The crucial first step in understanding semiotics is to understand structuralism, and the first step to understanding structuralism is to understand that the "sign-real world" model of understanding has certain problems with it. For one, we have spent much of this term arguing that we come to the "real" world with a set of expectations, strong enough that we "see" what we expect to see. We automatically move from the familiar to the unfamiliar, and in doing so bring not only our past experience, but our interpretations, ruminations, partial memories, emotions, and a host of other things that get mixed in to our memories of the past experience. We rarely remember "objectively", and there is a good question whether that objectivity actually even exists.
Structuralism substitutes another way of understanding, that which supposes that our words and images (called "signs"), point not at an external reality, but at each other. Our signs make sense in the context of other signs, not in the context of an external governing reality.
Saussure's semiotics depends on a binary system, that is, on the notion that there is a relationship between signifier and signified. This requires that we take seriously the statements (or images) that are made, and the context in which they are made. At the most basic level, the context that is relevant is the language. In fact, Saussure's most basic distinction is between langue and parole, or between the language something is in and the message that is being uttered at any time. "Langue" seems to be a kind of shadowy presence, in that, while no message can be spoken except in English, you can never get to "English" without the messages that are spoken. There's just more and more messages, more and more "parole". On the other hand, langue is what continues, while the messages themselves are uttered, and disappear. This means that for Saussure, the important thing about language is that it is synchronic, that is, it is related to other signs that co-exist with it, rather than diachronic, understandable in terms of history and changing over time.
These signifiers and signifieds exist in a kind of reciprocal relationship; Saussure's image is of a unity of the two:
So, understanding means the transfer of the contents of the mind from one person to the other. The signified, is the mental content, the concept that the signifier points at. The signifier is the material content, the thing in the world (at least the world of discourse), that points to the concept.
Now for images, the question is, what serves the function of "langue"? What is the "grammar" of images? We might suppose that has to do with the history of the images. But there is a resistance to thinking of images this way, because it is so much harder to see that images are themselves a language, or better, a set of utterances in a language. We have a much harder time divorcing the "real" world from the signs in images, than we do in words. So, let's try some examples.
One basic kind of image is the traffic sign. What does a stop sign signify? Is there something in the real world? Not really. It is not an image of something, but rather an attempt to communicate to us what we should do at a particular place. The sign signifies a concept, not an external reality. The sign "open" on the door of a business signifies a concept, the readiness of the shop to do business. Any of these signifieds could have other signifiers attached to them, as we recognize with words of a different language. Stop could just as well be "arret". But that assumes another langue. This could even be the case for the same langue – there is nothing "natural" about "stop" as the signifier, or "dog" as the signifier of the concept of dog. Signifiers are arbitrary.
Signs are not infinitely arbitrary, though. There are only certain kinds of substitutions that can be made, and only certain combinations. If we consider "The cat sat on the mat", we recognize that the order is important. The logical collection of signs produces a syntagm, and the sign "cat: has a syntagmatic relationship with the other signs in the syntagm. As well, there are only certain kinds of substitutions we can make for the word "cat". Feline might do it, in some circumstances, but not in others. Kitty, in some. Pussy, in some, but definitely not others. This is a paradigmatic relationship.
This suggests that there are two axes to every expression, the axis of order and the axis of substitution
So, how do we understand the concept that is signified? By convention, certainly – there is nothing innate about red and octagonality that necessitates stopping. But it is not only convention, but tradition as well. In other words, it is not just that we give agreement to a meaning, in a kind of social contract manner, but also that through time the sign takes on more meanings just because it appears in a variety of contexts. Nothing stays only with the consciously agreed meaning; it also accretes meanings to itself over time. So, after awhile, anyone wanting to use the sign will wittingly or unwittingly bring other meanings as well.
Take for example the swastika. It has come to stand for much more than what the original agreement had it to be, so much so that in many places the very use of the symbol is a statement on beliefs about politics, race, gender, lifestyle, and a whole host of other things.
Now, the stop sign or the swastika are images chosen consciously to represent something else. As I have argued, they never stay within the limits of the original intention. But what about images that were not consciously chosen to represent something? Are there such images? Let's take the one mentioned earlier, the tree.
Trees can be found in painting, photography, design, and all sorts of other places. The important thing is not just that the tree is there, but what the tree looks like. The "look" of the thing is crucial here. If it was just a tree, this might be iconography, the one-to-one correspondence between an image and a use. While these icons take on more symbolic functions (like everything does), often much effort is expended in reining in their extraneous senses. Not so with most other images we use. In fact, C. S. Peirce talks about 3 kinds of reference in semiotics: Symbolic representation, a mode in which the signifier does not resemble the signified but which is arbitrary or conventional (e.g., the word "stop", a national flag, a number); Iconic representation, a mode in which the signifier is perceived as resembling or imitating the signified (e.g., a portrait, a diagram, a model, onomatopoeia, metaphors, "realistic" sounds in music); Indexical representation, a mode in which the signifier is not purely arbitrary, but is directly connected in some way to the signified, by observation or inference (smoke, thermometer, clock, pain, fingerprint, handwriting, echo, knock on the door, etc.). These are listed in decreasing order of conventionality.
So, what about this tree? Well, if it is in an ad for SUVs, it might signify ruggedness and being away from civilization. In a vacation ad, it might signify the exotic (palm trees), or "getting away from it all" (the tree with the hammock on it). It could signify death (gnarled tree, no leaves, like in Friedrich's Tintern Abbey), or life (full, leafy tree in the middle of summer). It could be foreboding or welcoming. It could be a sign of the opposition between nature and the artificial world. It could be a sign of Christmas.
Now, how do we know what it is? Well, obviously some of it has to do with what kind of tree is represented. But it is not only that. The context of the tree is important. But what is context? Only other signs. In other words, the tree points to a concept, but that concept is unavailable to us if we are unaware of the other signs available. Signs work, then, by signifying a difference from other signs. The first taste of coffee, for instance – how do we know what that is like? In part, by distinguishing it from other, similar (and not so similar) experiences. So, the absent, previously experienced signs are crucial in the construction of any given experience.
What signs are those? The most obvious ones are those in the immediate vicinity of the image. But they are not the only ones. There is, as has already been mentioned, the convention and the tradition of signs. That means that any sign brings with it a history of associations, and those signs are important. But that's not all. There are also the signs that are absent. How could they be relevant? They are relevant because signs suggest oppositions. Back to the tree example.
I mentioned that the tree could indicate the opposition between the natural and the artificial. That may be true even if the artificial isn't represented.