I think one of the great ironies of the discipline of communication is that while many of the scholars we turn towards – most notably Burke – would deny that a word can ever represent a concept, the linkage between word and think/thought is central to how we teach language in our basic courses. I found the explanation of Lyotard on page 204 of Anti-Oedipus (both on the top of the page, and in the footnote) as a useful way to help explain how language does/cannot serve a representative function.
Specifically, I think the idea that applying a word to a thing reveals a hidden concept useful for also linking explain the power function of language. When the label dog is applied to the thing curled next to me on the couch, the term fails to either describe everything that is a dog, nor does it encapsulate the unique character of my dog. Rather it speaks to the hidden agreement of the social value and use of dogs. In this way, the word does not convey meaning so much as affirm/enact a specific definition of what ought to be. While this may be a simple concept, and a simple explanation, I think the work done to break the intrinsic link between the signifier and the signified is one that we need to take more to heart. It is easy to fall back on a representational understanding of language, but this is not a particularly useful exercise as representation is truly impossible. So in my basic courses, I will try to talk less about what words mean, and more about what they do.