LinguoFranco wrote: ↑29 Jan 2020 18:47What are some tips for avoiding this while generating vocabulary?
A big thing is to look at how non-English languages have divided up the semantic space.
A Conlanger's Thesaurus gives a lot of crosslinguistically common polysemies. With a quick scroll-through, I found a few listed as common that aren't found in English, like using the same word for "arm" and "wing"; "slow" and "cold"; or "bowl" and "gourd". (Note: The thesaurus has weird encoding, so it's not searchable unless you use the replacement scheme in the spoiler at the end of this post.)
In some cases, your language can also make more distinctions than English does. The first examples I think of are from Romance languages. Spanish distinguishes between
ser and
estar, which are both translated as "be" in English. Many Romance languages distinguish a reflex of Latin
sapio from a reflex of
cognosco (Fr. savoir vs. connaître; It. sapere vs. conoscere; etc.); in English, they can both be translated as "know".
Then, there are less binary changes one can make.
Kinship terms are a great place for variance, since languages can make more distinctions, fewer, or just go off in different directions. English follows what's called the Eskimo kinship pattern with its terms (no distinction between maternal/paternal relatives; distinction between nuclear family and extended family). The Crow kinship pattern has about the same number of distinctions, but divvied up differently. For example, the same term refers to one's father, one's paternal uncle and the son of one's paternal aunt. Meanwhile, the same words are used for one's siblings, the children of one's maternal aunt, and the children of one's paternal uncle.