In his groundbreaking work COURSE IN GENERAL LINGUISTICS (1915), Saussure divides the entirety of human linguistic communication into two categories: the syntagmatic, and the paradigmatic or associative, as seen in the passages below, derived from this site's translation:
In discourse, on the one hand, words acquire relations based on the linear nature of language because they are chained together. This rules out the possibility of pronouncing two elements simultaneously (see p. 70). The elements are arranged in sequence on the chain of speaking. Combinations supported by linearity are syntagms. The syntagm is always composed of two or more consecutive units (e.g. French re-lire [reread], contre tous [against everyone], la vie humaine [human life], Dieu est bon [God is good], s'il fait beau temps, nous sortirons [if the weather is nice, we'll go out] etc.). In the syntagm a term acquires its value only because it stands in opposition to everything that precedes or follows it, or to both.
Outside discourse, on the other hand, words acquire relations of a different kind. Those that have something in common are associated in the memory, resulting in groups marked by diverse relations. For instance, the French word enseignement [teaching] will unconsciously call to mind a host of other words (enseigner [teach], renseigner [acquaint] etc.; or armement [armament], changement [amendment] etc.; or education [education] apprentissage [apprenticeship] etc.). All those words are related in some way.
This is a pellucid and unproblematic statement of a sound theory of linguistics. However, Barthes, while complimenting Saussure on his "methodologically exemplary semiotics system," has concerns beyond "just the facts, mam:"
For mythology, since it is the study of a type of speech, is but one fragment of this vast science of signs which Saussure postulated some forty years ago under the name of semiology. Semiology has not yet come into being. But since Saussure himself, and sometimes independently of him, a whole section of contemporary research has constantly been referred to the problem of meaning: psycho-analysis, structuralism, eidetic psychology, some new types of literary criticism of which Bachelard has given the first examples, are no longer concerned with facts except inasmuch as they are endowed with significance. Now to postulate a signification is to have recourse to semiology. I do not mean that semiology could account for all these aspects of research equally well: they have different contents. But they have a common status: they are all sciences dealing with values. They are not content with meeting the facts: they define and explore them as tokens for something else.
On the face of it, this is fair enough. Neither Barthes nor anyone else should have to blindly anyone else's system, be it that of Saussure, Karl Marx or the Emperor of California. But in truth Barthes' "science of values" distorts the clarity of Saussure's analysis without bothering to say just where he Barthes chooses to depart from the pioneer of semiology. I speculate that he did so in order to coast on Saussure's reputation for empirical validity while masking the fact that the so-called "values" Barthes champions are extrapolated from Comrade Karl Marx. And these opinions, be they about myth, materialism or the bourgeoise, Barthes does follow with a sort of blind doggy faithfulness.
Once again, for Saussure, there's the syntagmatic and the associative. But somehow, when Barthes tells the story, somebow aspects that Saussure would've considered "associative" are validated as belonging to Barthes' category of "first-order language:"
Take a bunch of roses: I use it to signify my passion. Do we have here, then, only a signifier and a signified, the roses and my passion? Not even that: to put it accurately, there are here only 'passionified' roses. But on the plane of analysis, we do have three terms; for these roses weighted with passion perfectly and correctly allow themselves to be decomposed into roses and passion: the former and the latter existed before uniting and forming this third object, which is the sign. It is as true to say that on the plane of experience I cannot dissociate the roses from the message they carry, as to say that on the plane of analysis I cannot confuse the roses as signifier and the roses as sign: the signifier is empty, the sign is full, it is a meaning. Or take a black pebble: I can make it signify in several ways, it is a mere signifier; but if I weigh it with a definite signified (a death sentence, for instance, in an anonymous vote), it will become a sign.
Barthes makes one small but salient point here; though Saussure's schema attempts to separate the syntagmatic's ability to denote reality from the paradigmatic's ability to connote reality, the two are never operationally separate, and one can ask, as Barthes does in an essay on photography, how much connotation "arranges" human impressions as to what is being denoted to our eyes in an apparently "naturalistic" fashion. However, at no time does Barthes show any ability to cognize that his own "definition and exploration" of the reputed facts is any less a manipulation that the various social myths he attacks.
Shortly after the "roses" passage, we get one of Barthes' opening volleys against another level of what Saussure would call associative relations. He calls this "myth:"
...myth is a peculiar system, in that it is constructed from a semiological chain which existed before it: it is a second-order semiological system. That which is a sign (namely the associative total of a concept and an image) in the first system, becomes a mere signifier in the second. We must here recall that the materials of mythical speech (the language itself, photography, painting, posters, rituals, objects, etc.), however different at the start, are reduced to a pure signifying function as soon as they are caught by myth. Myth sees in them only the same raw material; their unity is that they all come down to the status of a mere language.
Now, though throughout Barthes continues to use Saussrean terminology, he departs-- without saying outright that he does so-- from Saussure's methodology. He tells us that the association "roses=passion" is of the first-order. But how did the association come about? Did Barthes or any other single person think of it? Obviously not. Obviously the concept "roses=passion" is transmitted through a given culture-- call it Culture A-- whose members agree to accept the arbitrary association of this trope. Another culture, "Culture B," may not agree to validate the trope on its own terms, but to the extent that any one culture understands another, B can certainly understand A's arbitrary cultural construction since B has its own set of such meaning-tropes.
However, at no point in "Myth Today" does Barthes give an adequate reason as to why what he calls "second-order language," or "myth," is so radically set apart from the associational aspect, "roses=passion." He gives various other examples-- using real lions as symbols for people's names, or the famous PARIS MATCH magazine cover that supposedly reveals the magazine's use of French cultural symbolism to validate French imperialism. But at what point would the association "roses=passion" go from being a first-order signification to a second-order one?
I suspect, given the way Barthes' nose remains centered on Karl Marx's literary butt, that the transformation would take place as soon as the passionate roses were used by some authoritarian entity. Yet the very idea that the roses are made passionate would seem to be an idea that is mythic, by the terms Barthes lays out much earlier in the essay, in his fourth paragraph no less:
Every object in the world can pass from a closed, silent existence to an oral state, open to appropriation by society, for there is no law, whether natural or not, which forbids talking about things. A tree is a tree. Yes, of course. But a tree as expressed by Minou Drouet is no longer quite a tree, it is a tree which is decorated, adapted to a certain type of consumption, laden with literary self- indulgence, revolt, images, in short with a type of social usage which is added to pure matter.
In what way does the Barthesian formula "roses=passion" leave roses alone to their "closed, silent existence?" Moreover, even assuming that somehow early man might have conceived the formula "roses=passion" in such a way as to isolate that conception from any other symbolic association-- thus privileging that formula as a "semiological chain" that pre-exists the depredations of the myth-language, does not the formula lend itself to "consumption" and "social usage?" If Caveman Unk gives Cavegirl Unka-a a present of roses as a "sign" of his passion for her, is this not social usage? And doesn't Unk hope he'll get some "exchange-value" for his gift, and that he'll be able to "consume" said value?
But of course Barthes' whole notion of the secondariness of mythic language has no basis in either Saussurean semiotics or in anthropological data, so the idea of isolating one category of associations from another category of associations is merely Barthes trying to "define" the facts in order to reflect a doctrinaire Marxist imperative. It's all about the hegemony, stupid, though Barthes doesn't use that specific Marxian term. He also doesn't stick with the passionate roses very long, preferring to focus on how mythic language robs everything it touches, be it people within human culture or animals outside of it. In MYTH TODAY the two examples he re-visits most are those of a lion, used to illustrate the hegemony of linguistic relations while a "Negro" in French uniform is used to illustrate the hegemony of imperial culture, both of which can somehow be blamed on "myth:"
"...the lion and the Negro are deprived of their history, changed into gestures."
Barthes' use of the term "gesture" is ironic to me given that roughly ten years before MYTHOLOGIES Susanne Langer had used the term to signify any motif or trope that summoned an emotional response from memory, as opposed to prompting an emotion from an event in real-time. Thus, for Langer, the emotion one calls forth by, say, naming a child after a lion connotes something very different from the emotion called forth from one's being in the presence of an actual lion. But why is the association attributed to the lion something that deprived the beast of history, but the roses are not so deprived? It may have something to do with Barthes' attempt to distinguish poetry and the poetic impulse from the cruel sway of myth, though his logic for this separation is no less strained than the rest of the essay.
I can think of few things less interesting than a Barthes biography, but thanks to the Internet, I am duly informed that he may have eventually seen some of the flaws in his own system. From David Chandler's SEMIOTICS FOR BEGINNERS:
Related to connotation is what Roland Barthes refers to as myth. We usually associate myths with classical fables about the exploits of gods and heroes. But for Barthes myths were the dominant ideologies of our time. In a departure from Hjelmslev's model Barthes argues that the orders of signification called denotation and connotation combine to produce ideology - which has been described (though not by Barthes) as a third order of signification
Now, this hypothetical "third order" would have been a slight improvement over Barthes' two orders, since he isn't really able to present any human associations for "roses" that don't act to "deprive them of their history." At least then he might have claimed, with some degeee of plausibility, that the "third order" represented by myth was something more articulated and pernicious than incidental cultural associations.
But he doesn't, and my own sign-reading tells me that it's because Barthes, following in the wake of Marx, was too much in love with a Manichean good-vs-evil dichotomy. Barthes knew that a duality was better suited to portray a radical opposition, especially to other Marxist intellectuals.
Susanne Langer doesn't reference Saussure in the NEW KEY work from which I've been quoting, but I believe the two of them could have found some agreement on her notion that the associative symbol has only a "logical analogy" to its source material. This should be true whether one is describing roses as a symbol for passion, as a symbol of life and/or rebirth (as seen in flowers being placed in Neanderthal graves), or as a symbol of the great flower-god Roseata. Were one to go against Saussure and Langer, and join Barthes in saying that some "significant" connotations always attach themselves to the material denoted, one would have no way of proving, via Barthes' flawed reasoning, that any of the latter two associations has any better claim to Barthes' "first-order" status than the one he does favor.
The most famous dog Barthes reminds me of is the one in a certain Aesop's fable. Walking beside a river with a bone in his mouth, the dog spies his reflection and mistakes it for another dog with another bone. Greedy for the second bone, the canine opens his mouth to bark at the strange dog, and promptly loses his real bone to the waters.
That's Barthes all over; hungry to grasp reality and falling victim to his own illusion. The main difference is that the dog didn't busy himself trying to correct other people's illusions while showing no propensity to correct his own.