Recently I've been attempting to think of language as having no inherent relationship to communication. I believe it could be argued that communication and transmission of information are purely accidental and contingent to certain kinds of language (such as so-called natural human languages). Language is, in this tentative opinion of mine, ultimately an issue of relating to alterity, environment, specularity, spacement, death, etc. In this sense, every living being - down to single-cell organisms - would have a linguistic relationship to their (or "the") world. Chemical reactions would eventually be identified as the model linguistic phenomenon. To a molecule or a bacteria, a chemical connection would effect linguistic meaning (I refuse to employ the expression "carry linguistic meaning"), rendering this interaction as a instance of language.
From the little I know about second order systems theory via Cary Wolfe, this formulation does indeed resemble language viewed from such perspective - systems in general would "communicate" to their environments (or constituent parts) via effects of meaning. I like how it seems that Cary Wolfe has finally convinced me that systems theory does resemble Derridean philosophy of language. Derrida's defense of writing as the ultimate model of grammaticality and diacriticity (the conditions of possibility for language, since all language must be built upon schematic differences) reflects precisely this idea that language is, after all, nothing but a relationship to a differential and "grammatical" alterity. Only specific instances of "messages" (we could call them "grammatical") trigger the effects of meaning in chemical compounds, paramecia, systems in general, or in human language. This leads me to the exciting book I'm currently reading - Emanuele Coccia's La Vita Sensibile ("The Sensible Life," or "The Sentient Life").