Barthes said that "every use becomes a sign of itself". This locates signs in action and practice (including linguistic practice). What happens to the sign after that is presumably largely influenced by culture and convention (until someone uses it in completely new ways, of course).

I think that the main criticism is that the idea of code (i.e. law) does not resist closer examination. Even our natural languages ain't fully coded and each of us speaks his or her idiolect. Things otherwise would be much easier: I for one, as a non-native speaker of English, would very much like to find its 'code', but all books (or others) purporting to give the code disagree between themselves -- as do also the native speakers, as a matter of fact.

Nobody ever agrees on a code. Learning a language is not learning its code, but becoming acquainted with some of its usages. Even 'convention' would be too much to say: languages are not the result of conventions between their users. A child learning his or her mother tongue receives neither a code nor a convention, and develops a language that differs from that of his/her mother!

A culture does not result from codes or conventions, especially in the sense that there is no mutual agreement about what the culture should be. It more probably results from (more or less) shared usages, from habits.

I don't think that anyone ever uses a sign in "completely new ways" -- it would not be a sign, as it would not be understandable by others. But everybody uses signs in partly new ways, which is the reason why (1) I do not belive in codes; (2) our intelligence may progress.

Personally, I have not read a persuasive rebuttal of this notion and would be interested in hearing critiques of Semiotics of this sort. Up to this point I have found the notion of codes to be almost self-evident. It simply seems to "fit" with our experience of life in our world.

One important critical contribution to the notion of "code" you can find in Dan Sperber's research work; cf., for instance, in the 70ies his critics on symbolism in anthropology, and in the 80ies his important research on relevancy.

However, the notion (or concept) of code is not so evident - at least with respect to its different uses in semiotics. Also its relationship with the concept of "convention" (for instance in David Lewis' sense) is not at all evident in (cultural) semiotics.

Moreover, there are several currents in semiotic research for which the notion of code is not really a central one. This is the case, for instance, of structural semiotics of Greimas who considers "culture" (more or less in the sense of Foucault) as an "episteme", a cognitive and axiological reference frame for social (personal, collective) appearing, doing and behaving ("à la Goffman").

I recommend exploring the Semiotics of Cultures that is being developed by Francois Rastier in France. His criticism of code is clear and succinct. In line with Saussure’s Linguistics of parole (not langue), his methodology is a de-ontologized, praxeological approach to meaning that falls in within the hermeneutic and rhetorical traditions more than the logico-grammatical tradition (for which the notion of code is often key). With this kind of Semiotics of Cultures, meaning, codes, and signs are never given in advance; they are only – and can only be – the results of interpretation. Isolated signs and codes can only be decontextualized models of signification (as opposed to meaning which happens in a specific situation and context, i.e., parole). From this point of view, what we would be more interested in describing how we arrive at the idea of code through interpretation, rather than describing the identity of a code (i.e., a code as identical to itself); in other words, we are looking at semiosis as an action without resorting to any sort of ontology. From the point of view of parole, a code only “exists” (if we can even say that! ) in the act of semiosis (parole), in a particular act of enunciation, in its “manifestation”. It does not have any independent existence beyond that. It is always just a construct, constructed in the very act of parole. Natural Languages, for example, are not codes to be encoded or decoded; that would assume that a code pre-exists any act of parole, which it does not. One text may reproduce a structure or interpretive path that has already occurred in another text; but to call that a code is misleading, because each occurrence is unique (a hapax). A code cannot have any independent existence unless you accept the fact that the code is always changing from one moment to the next (which essentially renders the notion useless) . On top of that, transcoding does not account for the act of interpretation either, which is the key to constructing meaning. Only artificial and formal languages - or the Morse code --can be treated like codes. This doesn’t mean that there are not social norms. But norms should not be confused with codes which comes from information theory. I could probably say more, but this sketches out a certain point of view that could be interesting to explore.

Here are some good readings:

In English:

http://french.chass.utoronto.ca/as-sa/ASSA-No5/Vol2.No5.Rastier.pdf

http://www.revue-texto.net/Parutions/Livres-E/Rastier_SCR/Semantics-and-cognitive-research.pdf

In French:

http://www.revue-texto.net/Inedits/Rastier/Rastier_Transmission.html

http://www.revue-texto.net/docannexe/file/1612/rastier_normes.pdf

this lecture may be useful:

https://www.youtube.com/watch?v=rVOTdheq5eU&list=PLD00D35CBC75941BD&index=25

Discussed is Knapp and Michael`s neo-pragmatist critique of Literary Theory (Against Theory). A lengthy discussion of the relationships between, on the one hand, intention and meaning and, on the other hand, language and speech follows with reference to Saussure, deconstruction, and Russian formalism.

The most powerful criticisms have been articulated by Thomas Sebeok and the school he inspired. Sebeok's seminal paper was:

Sebeok, T. A. (1988). In what sense is language a ‘Primary Modelling System? In B. Henri & K. Rebecca (Eds.), Semiotics of culture, pp. 67–80. Helsinki: Arator.

What you will find there is an argument that culture grows out of nature. Codes alone won't do.

The article by Sebeok is hard to find and (thus?) hardly read today, but it had a huge behind-the-scenes influence. You might therefore consult:

Article Semioticians Make Strange Bedfellows! Or, Once Again: “Is La...

as well as

Article Oxford Bibliographies in Philosophy: Semiotics