Mutual intelligibility is the degree to which speakers of related language varieties can understand each other without prior familiarity or special effort. It is often used to tell languages from dialects, though it is only one factor and can be lopsided.
Key takeaways
- Mutual intelligibility is a matter of degree, not a yes-or-no test.
- It can be asymmetric: speakers of one variety may understand the other better than the reverse.
- Political and social factors often override it when a variety is called a language or a dialect.
- Danish (about 5.6 million speakers, ethnologue.com, 2012) and Swedish are separate languages on the map although they are related.
How mutual intelligibility works
Mutual intelligibility describes related language varieties whose speakers can readily understand each other without prior familiarity or special effort [1]. It is sometimes used to separate languages from dialects, with sociolinguistic factors often considered too [1].
Three things make it messy.
- It comes in degrees. Closely related speakers can often communicate if they try, so intelligibility sits on a spectrum and not in two boxes [1].
- It can be one-sided. Dutch speakers generally understand Afrikaans more easily than Afrikaans speakers understand Dutch. A study cited by the same source found Slovenes understood Croats better than Croats understood Slovenes [1].
- Exposure matters. Swedes in the Öresund region understand Danish better because they live near Danish-speaking areas [1].
Tests have also been run on varieties of one language. Tunisian and Libyan Arabic share a large share of sentences in a functional test, according to the article [1].
Intelligibility can change without the speech changing. In a dialect continuum, the loss of central varieties can turn the two end varieties into separate languages [1][2]. And political and social factors often override the result, as with Chinese, Arabic and the Serbo-Croatian standards [1]. The question is covered on language vs dialect.
An example from our data
Danish, Swedish and Norwegian are related, and the cited article says spoken Danish and Swedish usually have low mutual intelligibility, while Norwegian understanding of Danish is asymmetric because Bokmål grew out of Dano-Norwegian and absorbed Danish vocabulary [1].
On our map all three are separate languages, with the figures about 5.6 million (ethnologue.com, 2012), about 9.2 million (Ethnologue, 2012) and about 5.2 million (Wikidata (unreferenced), 2015) speakers, all approximate and from different sources and years. Dutch and Afrikaans are likewise listed separately, with Afrikaans at about about 7.2 million speakers (Wikidata (unreferenced), 2016).
The map does not measure intelligibility. It shows which languages Glottolog lists and where an atlas places them. Whether two neighbouring languages understand each other has to come from other studies.
The thing people get wrong
People often assume that if two speakers can follow each other they must be speaking one language, and that if they cannot, they must not. The test gives answers that do not match how the varieties are named. Chinese varieties are often called one language although distant speakers usually cannot understand each other [1], while Danish and Swedish are counted as two although they are closely related.
Intelligibility is a useful clue but not a rule. How people feel about their own speech, and the status of standard languages, matter as much.
Frequently asked questions
If two people can understand each other, do they speak the same language?
Not always. Understanding is one test, but related languages can be partly understood, and political and social factors also shape the label.
Can understanding go one way only?
Yes. Dutch speakers generally understand Afrikaans more easily than Afrikaans speakers understand Dutch.
Does the map measure intelligibility?
No. It has no data on how well languages are understood across each other.
Sources
- Mutual intelligibility, Wikipedia.
- Dialect continuum, Wikipedia.
- Every figure from our own map is worked out from the datasets on the data sources page (checked against Glottolog and Glottography, retrieved 13 October 2026).