How are dialects formed? Are they always a diverging branch from the main language or can they be the fruit of a converging process between different languages because of cultural pressure?

Also, when does it become a dialect instead of merely an accent or another language altogether?

It's a fairly ample question, I'm glad to simply receive reading material on the general subject if a complete answer is not available

  • Welcome to Linguistics SE! I put a reference-request tag on this and let others who know more about dialects decide if this is too broad. Commented Nov 21, 2015 at 0:31
  • @IvanKapitonov Thanks for the suggestion, gladly taken! Commented Nov 21, 2015 at 1:05

1 Answer 1


There are numerous examples of dialects which result from language contact. A typical case would be Indian English (the dialect of English that has become native to India), which arose from contact between British English and various indigenous languages of India. Similarly, the Highland (northern) Ecuadorian dialect of Spanish differs from other nearby dialects of Spanish in ways that reflect properties of the local indigenous language, Quechua. There are some instances where we can be fairly certain that a dialect difference is not due to this kind of language contact, such as some NE/SW Icelandic dialect differences pertaining to aspiration -- where there is no substratum or differential external influences that could explain the differences. But it's far from a foregone conclusion that the difference between English in NW England versus London are strictly due to random divergences starting from a homogenous older version of English. In most cases, we simply do not have enough historical evidence regarding the development of dialects to be able to confidently say that a given difference can be due to some kind of "external" influence.

Linguists don't put much stock in a strict dichotomy between language and dialect, because there's no objective standard for classifying systems as being dialects vs. languages. The best we hope for is a metric of mutual intelligibility. Swedish and Norwegian are classed as distinct languages, but they are close enough that a speaker of one can understand and be understood by a speaker of the other (there are outlier exceptions). There are numerous problems with the concept of mutual intelligibility, for example asymmetries in intelligibility, where spoken Danish is apparently harder for Norwegians to understand that Norwegian is for Danes; and there are "logical inclusion" problems where all speakers of X also speak Y but the opposite is not true (e.g. all Kerewe speakers speak Jita but the reverse is not true). Once a speech form becomes so different that speakers cannot understand each other, then linguists embrace the notion that they are different languages. However, speakers themselves may have a different attitude. Kurmanji, Sorani and Gorani ~ Hawrami are sufficiently divergent language forms that I think they would have to be termed separate languages, but they are often simply called "Kurdish" and referred to as dialects (emphasizing political and cultural unity).

As for "accent", we put zero stock in that term, which is a term that we abjure, at least when speaking professionally. A "Chicago accent" is the dialect spoken in Chicago (with other social modifiers necessary to pick out the right dialect).

You might want to take a look at William Labov's books Principles of Linguistic Change, in 3 volumes.

  • Thanks so much for such a complete answer! I'll definitely look into the book you have mentioned. Commented Nov 21, 2015 at 8:31
  • Put another way, the terms 'language' and 'dialect' are polysemous, having political meanings (e.g. Chinese 'dialects' Scandinavian 'languages') and technical linguistic meanings (Chinese languages, Scandinavian dialects). Commented Nov 21, 2015 at 21:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.