Regional synonyms constitute a “pair” of two or more words identical in sense, known and used by a group in at least one locality at a given moment in time [emphasis in original]. Thus, with regard to regional synonymy as well, the degree to which we accept words as synonyms depends on how they fit in time and place.SOURCE: "Synonymy and dialects" (18.104.22.168) in Probleme de sinonimie, by Onufrie Vinţeler (Bucureşti: Editură Sţiinţifică şi Encliclopedică, 1983) [my translation].
Sever Pop (cf. 1929) used to note that, within the territory of Romania, the following terms can be found to denote the concept of ‘horse trader’: barâşnic, craşcadău, cupeţ, factor, fleşer, geambaş, gheşeftar, ghiambabău, gârgez, făznar, hendler, herghelier, hâmbluitor, liverant, mecler, năstrăpaş, negustor, peţer, pilar, potlogar, precupeţ, precupitor, semsar, sfârnar, sfârnăroiu, şmecher, ţânzar, ţigan, tuşer. No one doubts that all the terms listed denote the same concept. The question that arises is the following: can each and every one of these words be considered synonyms? According to some definitions, still in circulation, all words that express the same notion are considered synonyms. Glancing over the list of words above, we observe that only the word negustor, which is the general term, and to a certain degree the word geambaş, are more widely known and can be considered synonyms; the rest are known only in more or less restricted areas. For the great majority of Romanians, words like barâşnic, gârgez, hendler, mecler, tuşer, and so forth do not mean anything; they are just as unintelligible as any others in a foreign language. Of course, in many places negustor can be a synonym of făznar, and geambaş with herghelier [‘herder’], and so on, but this only happens in certain places and not across the whole territory where Romanian is spoken.
These examples prove once again that for two or more words to be considered synonyms it is not sufficient that they express the same notion. And in cases of regional synonymy, the notion of synonym must be localized and made concrete.
Situations like those discussed above are very common in Romanian; they may be found on almost every linguistic map but are not mentioned except sporadically. So, for example, Marius Sala (1958), after analyzing the distribution of terms for ‘maize’, established that porumb is synonymous with păpuşoi [cf. păpuşă ‘doll, puppet’], cucuruz [usu. ‘corncob’; Russian кукурузы, South Slavic kukuruz], mălai [usu. ‘cornmeal’]. All of these words are fairly widely distributed (cf. map 900 in the Atlas of Romanian Linguistics). At present, porumb, being the general, literary term, is synonymous with păpuşoi, cucuruz, mălai in those areas where the latter are used. But the question arises whether păpuşoi, cucuruz, and mălai can be considered synonymous with each other. In the first place, at their points of intersection, they can be completely synonymous, except in cases where certain semantic differences intervene. Second, because of their wide usage, even in literature by great writers, we can admit that they form a set of synonyms at the level of literary language, if not everywhere, at least widely enough. But what do we do with the term tenchi, borrowed from Magyar and recorded at just one locale on the same map? In every other zone, tenchi is a foreign word, and therefore cannot be synonymous with the other words that denote ‘maize’. (Tenchi may eventually become synonymous with porumb or with mălai to the extent that the latter are known and used in the locality where tenchi was recorded.)
I don't have any problem with considering terms in different languages to be synonyms. That's what translationese, calques, and 直訳 are all about.
UPDATE: In response to Language Hat, I need to clarify that when I accept "synonyms" across separate languages, I'm thinking of communities where nearly everyone speaks at least two languages, and where people switch between them as frequently and as easily as they or others might switch between dialects of the same language. I've spent some time in such communities. In fact, my first published paper in graduate school after returning from fieldwork in a New Guinea village whose unique language I had been sent to describe was not not a paper just about the new , undocumented language but one on "multilingualism and language mixture" among the people of that village. People borrowed and calqued all the time, and even recreated for my benefit "pure" equivalents (rarely used by anyone else) in their own language by calquing backwards out of either the local church lingua franca, whose usage has since faded to the point that young people are no longer likely to recognize the source of those words; or else Tok Pisin, which supplies nearly every community in New Guinea an extra set of very general synonyms for specific words in their own languages. Romanians did much the same a few centuries ago when they borrowed a load of new vocabulary from French, then creating Romanianized shapes for many of the words. Those pairs became synonyms. Sometimes the synonyms carved up semantic space in complementary fashion, and sometimes one form gave way to the other.
I plan to translate several more chunks from Vinţeler's chapter on borrowings and synonymy. He compiled a lot of good examples.