Communication requires a consensus about language. We have dictionaries to help stabilize that consensus; we have poets to help keep it evolving. I am not much of a poet, but I identify with their part of the task: I use the dictionary words (making up "new" words like quark has always seemed a little on the tachy side to me; why break rules if they are fair?) but I sometimes try to decorate their meanings with a lot of connotations and allusions and specific details in a given context that are not in any dictionary and would be inappropriate in another context. This is a fun ego trip; it is also necessary whenever one is trying to make a point that goes a little beyond where existing language leaves off - which isn't far from where we live daily.
Unlike most poets, however, I will do my best to spoil the mystery of my private terminology: whenever I realize that I am using a word in a specific sense that transcends the dictionary meaning and its colloquial connotations, I will try to call attention to it and explain as much as I can about the differences. Poets don't do this for a very good reason: part of the magic of poetry is its ambiguity. Not just random ambiguity like dictionary words out of context, but coherently ambiguous; a good poet is offended by the question, "What exactly did you mean by that?" because all the possible meanings are intended. Great poetry does not highlight one meaning above all, but rather manipulates the interactions between the several possible interpretations so that each enriches the others and all unite to form a whole greater than the sum of its parts. Unfortunately, the reader/listener can only appreciate this subtlety after mastering the nuances of the language in which the poet writes or speaks. Those who have mastered the language of Physics do indeed rely upon the same sort of "coherent ambiguities" to get their points across, or else no one would be able to discuss quantum mechanics at all (to give the prime example); this is why I have given the subtitle Physics as Poetry to this collection of HyperReferences. But at the beginning we are learning "science as a second language" and it is best to minimize ambiguity where possible.
The first and obvious example is the word PHYSICS. If I mean the (hypothetical) orderly behaviour of the (hypothetical) objective physical universe, I will write "physics." If I mean the sociopolitical human activity, the consensual reality prescribed by a set of conventional paradigms and accepted models about said universe, I will write "Physics." Unlike some deconstructionist sociologists, I believe the former exists independently of the latter. Or at least I have a commitment to that æsthetic....