A macrolanguage is an ISO 639-3 grouping of closely related language varieties that are often treated as one language, such as Arabic, Chinese and Norwegian. It is mainly a bookkeeping device that links the codes in different ISO standards.
Key takeaways
- A macrolanguage groups several individual languages under one code, often a dialect continuum with a shared name.
- It is a bookkeeping device between ISO code sets, not a strict linguistic category.
- Norwegian has the ISO code nor on this map, a macrolanguage covering Bokmål and Nynorsk.
- The map lists Arabic and Chinese by their individual varieties, such as Wu Chinese and Egyptian Arabic.
How macrolanguages work
Every language in the ISO 639-3 standard gets a three-letter code, kept by SIL International as registration authority [2]. A macrolanguage is a code that stands for a group of closely related individual languages, often a dialect continuum, whose speakers commonly share a name [1].
The article on macrolanguages describes them as mainly a mapping device between ISO 639-3 and the older ISO 639-1 and 639-2 code sets, and as a "book-keeping mechanism" rather than a strict linguistic category [1]. That matters for how to read them. Calling something a macrolanguage does not say how close its members are in speech. It says only that the standard groups them under one code.
Three examples from the same source [1]:
- Arabic (ara) covers 28 individual ISO 639-3 languages, including Standard Arabic (arb) and Egyptian Arabic (arz) [1].
- Chinese (zho) covers many varieties that are not all mutually intelligible.
- Norwegian (nor) maps to the individual languages Bokmål (nob) and Nynorsk (nno).
The source notes that ISO 639-2 leans on shared writing systems and literature, while ISO 639-3 focuses on mutual intelligibility and shared vocabulary [1]. Different standards therefore draw the language line in slightly different places, which is one reason counts of languages differ (see language vs dialect).
An example from our data
This map follows Glottolog [3], which gives each language a Glottocode, and shows the ISO 639-3 code where one exists. The entries below show how that plays out for the macrolanguages above.
- Arabic. The map lists Standard Arabic (ISO arb) and Egyptian Arabic (ISO arz) as separate languages. Ethnologue-based figures give about about 335 million first-language speakers for Standard Arabic (ethnologue.com, 2023) and about about 64.6 million for Egyptian Arabic (Ethnologue, 2019), both approximate.
- Chinese. The map lists Mandarin Chinese (cmn), Wu Chinese (wuu) and Yue Chinese (yue) separately.
- Norwegian. The page for Norwegian carries the ISO code nor, the macrolanguage code.
The map does not store macrolanguage membership as a field, so it cannot filter by macrolanguage.
The thing people get wrong
A common assumption is that a macrolanguage is "the real language" and its members are mere dialects. The standard treats the members as individual languages in their own right, and the macrolanguage as a grouping above them [1][2]. Whether the members should be called languages or dialects is a separate question that standards and scholars answer in different ways.
A second mistake is to treat the code as a measure of how alike the members are. Chinese is grouped under one macrolanguage code although its varieties are not all mutually intelligible [1].
Frequently asked questions
Is Chinese one language or many?
ISO 639-3 has a Chinese macrolanguage that covers many varieties, and the article on macrolanguages notes that these are not all mutually intelligible. This map lists the varieties separately, for example Mandarin, Wu and Yue.
Why do macrolanguages exist?
Mainly to map between the older ISO 639-1 and 639-2 codes and the more detailed ISO 639-3 codes.
Sources
- Macrolanguage, Wikipedia.
- ISO 639-3, SIL International (registration authority).
- Glottolog, Max Planck Institute for Evolutionary Anthropology.
- 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).