Read Jean Aitchison's Language Change: Progress or Decay?, especially chapters "I knows how to talks in Readin" and the chapter on Martha's Vinyeard.
Then read Bowra's Tradition and Design in the Iliad the chapter on Homer's language.
Now, he states (as principle behind his conclusion that Homer's dialect was never even close to a spoken one) that one dialect has only one word for one exact meaning.
If Jean Aitchison is correct, that is not quite so. The nuances of words similar in meaning are not only semantic ones (Bowra cites chthôn and gâ as meaning in Lesbian dialect respectively "earth=land-not-sea" and "earth=both-land-and-sea"), but also nuances of speaker's identification or disidentification with rival standards both used in a certain area at a certain time.
Let's not delve into the inaccuracies Bowra commits about Beowulf in the next chapter. Suffice it to say that where he is complacently sure that Beowulf the nephew of Hygelac (Chlochilaicus the Geatish king!) never existed, I am not, I incline to opposite theory. His points about telescoping historic events into one heroic action of epics is however valid./HGL