I'm curious about (and also stuck at) the natural formation and ways of construction of polysemies in naturalistic conlangs / natural languages.
More specifically: I'm constructing the vocabularies of a proto-language. When it comes to words with only subtle differences, I failed to come up with a proper method.
To my knowledge words are formed by putting word roots together. The word roots should have all the basic meaning covered. But I frequently see words with same or similar basic meanings, like learn / study, middle / center, get (one of the senses) / become, etc. in English. As for now I can only think of two reasons, but I'm not sure:
These words comes from different sources. But for proto-languages this is simply impossible.
Multiple possible combinations of roots has similar meanings, and eventually the two become polysemies. But I suppose this only works for not-so-basic concepts.
My problems are:
a) How polysemies is formed in natural contexts?
b) How can a proto-language have polysemies? Or it shouldn't have any? (In this case how to express the "subtle" differences between words?)
EDIT Thank you, Draconis and Sir Cornflakes! I would clarify that by proto-language which might be the wrong term to use I actually (somewhat) mean the first language ever evolved in a group of people. So it's at the root of any language tree, thus can have no or little external sources of words.
What I care more is how can I construct polysemies (or just words with similar meanings) in a naturalistic conlang: by using a new word root, or some other construction?