What ways do we have of measure the mutual intelligibility of two given languages? It's easy enough to realise if there is a degree of mutual intelligibility between your own language variety and another, but what methods do we have to measure it between two specific languages that we don't necessarily speak ourselves?

For example, Standard Ukrainian and Standard Belarusian are said to have a degree of mutual intelligibility and Standard Ukrainian and Standard Russian are also partially mutually intelligibility (but not as much as Ukrainian - Belarusian). So how can we measure the differentness between Ukrainian and Belarusian, and the differentness between Ukrainian and Russian - thus showing which of the two, Belarusian or Russian, is most similar to Ukrainian?

  • 3
    The thing is, this is not an answerable question; this is the same kind of fallacy as IQ. Languages (like "intelligence") do not exist outside the people who speak them, and everybody speaks and learns their own language(s) in their own way; this is especially true with second-language acquisition --some people do it easily, some never do at all. That means there's a multi-dimensional spectrum of variance that can't conceivably be captured by a single number based on Standard (i.e, Abstracted) languages.
    – jlawler
    Commented Jul 9, 2012 at 15:22
  • 1
    Of course, any measure of differentness can't be an exact figure (due to the fact that so many people within a given language speak so many different varieties), but we can look at the standard varieties to give us an indication of how similar or different certain languages are in relation to others, perhaps. Commented Jul 9, 2012 at 15:35
  • 1
    @DangerFourpence To reinforce what jlawler is saying, the concept of a language having a standard form is not linguistically sound. Any "standards" that exist do not represent the actual language in usage. Even the concept of a language is nebulous, in the sense that it cannot be well defined beyond it being "a dialect with an army and a navy": I only speak what's known as English because of political reasons; if things were different, my mothertongue might just be considered some peculiar dialect of Old Frisian. Commented Jul 10, 2012 at 13:06
  • 1
    Ok, what you're both saying about 'standards' is totally right, and I agree with you. But the point is, we can still compare two varieties. I was just using the 'standard' varieties, the varieties that we would be most likely to learn as a second language learner as an example. But I get what you're both saying :) Commented Jul 10, 2012 at 15:22
  • 1
    Cool :) but, as jlawler points out, every person's language is going to be a bit different to another's, even if they have the same name. That being so, it makes it impossible to define what a particular language is and then compare it to another... The best you could hope for is to quantify the mutual intelligibility between two individuals at a particular point in time. Commented Jul 10, 2012 at 22:10

2 Answers 2


Although not a true test of mutual intelligibility, Blair's (1990: 26-34) is one way to quantify the lexical difference between two languages, so as to get a percentage of lexical similarity. The higher the similarity the greater likelihood of mutual intelligence.

The test works by taking a word list, the bigger the better, although I have successfully used it on a mini-corpus of a 100 word Swadesh list before. Each item is gone though, with points assigned for features the same - identical are given full points, differences that are minimal are given less points, and there are no points for completely different items. The great thing about this is that it can be done across multiple languages/dialects at the same time.

There are probably computer programs that can do this now, and more sophisticated analytical processes, but for a small list it's not that hard to do it be hand as long as you have a rudimentary knowledge of phonetics/phonology.

Blair, F. (1990). Survey on a shoestring. Arlington, SIL International and The University of Texas.

  • Thanks very much @LaurenG. I've just had a quick look in the library catalogue and this book isn't available. Will keep searching online for something. Commented Oct 15, 2012 at 20:08
  • @DangerFourpence - I ended up buying it. There may be some papers that cite it that talk through the procedure?
    – LaurenG
    Commented Oct 16, 2012 at 23:51
  • imho, it wouldn't work for the spoken language. Would it? There are several words in English and French that look very similar but sound completely different.
    – Centaurus
    Commented Dec 16, 2015 at 23:13

I guess one could make a list of the n most common words in a language (from a body of texts, news articles can be one source). The number of shared words (ignoring small spelling differences) in the other language, weighted by the commonness of the word, could create a measure of similarity. Similarity of grammatical rules could also be factored in somehow. This could perhaps be accomplished automatically by using google translate?

In reality though, peoples exposure to other languages make a huge difference. The three main Scandinavian languages are asymmetric intelligible due to differences in exposure to the others from early childhood through TV/media.

  • This is only guessing, not an answer, unfortunately. Commented Dec 16, 2015 at 8:53

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.