The concept of semantic prosody in John Sinclair is similar to that of Pound's logopoeia:
"that is to say, it employs words not only for their direct meaning, but it takes count in a special way of habits of usage, of the context we expect to find with the word... It holds the æsthetic content which is peculiarly the domain of verbal manifestation and can not possibly be contained in plastic or in music."
Semantic prosody, for Sinclair, results from the statistical probability of finding a word close to other words. Take the word "pulular" in Spanish. It refers, often, to the swarming of insects. If we find it with people, instead of with insects, we might envision those people as insects. I guess the word swarm in English works the same way.
intr. 1. Moverse de un lado para otro, bullir en algún lugar personas, animales o cosas.
2. Abundar,multiplicarse insectos y animales semejantes:
las moscas pululaban entre la basura.]
A similar example is "enjambre," a colony or swarm of insects. If used outside of this context, it still suggests insects.
An example Sinclair uses is "budge." The word means to move a slight bit, but the semantic prosody is that of intransigence. It is almost always found in contexts in which someone refuses to budge.
But I think semantic prosody is only one device within the greater category of logopoeia. It seems that it should also encompass other kinds of verbal play, the entire "dance of the intellect among words," not merely one device of using a word against the backdrop of its normal usage. By the same token, semantic prosody itself ought to be reconceived more broadly as the linguistic study of logopoeia in its natural settings (not merely in poetry).