AreaAudit

Which languages each generator says it supports, and where

Mutual intelligibility: why a count depends on who is counting

Mutual intelligibility is whether speakers of two varieties understand each other without study. It is often used to decide whether they are two languages or one, and it does not settle the question, because the boundary is political as well as linguistic. As of 2026-09-22.

Why intelligibility does not produce a language countVarieties speakers understand easily are counted as separate languages when they belong to separate states. Varieties that are hard to understand are counted as one when they do not. Standards bodies follow the conventions rather than the linguistics.What decides whether these are one language or two?Separate statesUsually counted as twoEven where speakers understand eachother without study.One stateOften counted as oneEven where the varieties are barelyintelligible to each other.This atlas records codes and names as published, and re-decides nothing
Fig. 1 Any total therefore inherits a boundary somebody else drew, which is fine until the total is treated as a physical quantity.
Why intelligibility does not produce a language count. Recorded 2026-09-22.
SituationCounted as one or twoDecided by
Mutually intelligible, separate statesUsually twoPolitics
Barely intelligible, one stateOften onePolitics
Intelligible in writing onlyVariesConvention
Intelligible one way onlyVariesConvention

Inclusion rule. The situations where intelligibility and the conventional count disagree. The register does not resolve any of them. Order. Most common situation first.

1The count is a convention, not a measurement

Varieties that speakers understand easily are counted as separate languages when they belong to separate states, and varieties that are hard to understand are counted as one when they do not. Standards bodies follow the conventions rather than the linguistics.

So any language total inherits a political boundary somebody else drew. That is fine as long as nobody treats the total as a physical quantity, which is exactly what a ranked coverage column invites.

2Where it matters in this register

Two closely related Southeast Asian languages are declared by overlapping sets of vendors here, and a reader might assume a declaration of one covers the other. Nothing published says it does.

Similarly, one national language appears under two different codes from two vendors, because the older code names the variety the standard was built on. The register merges those by language and prints both codes rather than picking one.

3What this register does about it

Nothing, deliberately. It records the codes a vendor wrote and the names a vendor published, and it counts by language part when a code is written two ways for one market.

Re-deciding any boundary would produce figures no vendor claimed and no reader could reproduce. Where a merge was necessary it is stated on the page where it happened.

A definition, not a measurement: no vendor figure appears here that is not also in the register. The one merged case in this register is printed on its own language page.

Nearby terms: Variant, Writing system. The whole vocabulary is at terms; nothing on this page is a claim about a named product.