Some states are in English and some are in the language respective to the country

[tjscience]

Issue Type

Other

Location (if applicable)

No response

Bug Description

I am not sure if this is intended or not, but different countries have different languages for the states. For example, if you look at Poland (PL), the states are in English. However, if you look at Czechia (CZ), states are in Czech. Is there a reason for the inconsistency?

Expected Behaviour

Language should be standard, one way or the other. If there is no locale parameter or header for the endpoints, the language of the properties returned should be native to the country. Or, choose one language, like English to default to.

Source (optional)

No response

Additional Context

No response

[dosubot]

Hi @tjscience! I'm Dosu and I’m helping the countries-states-cities-database team.

This is a real inconsistency. The database schema has a three-tier naming system — name (intended for English), native (local language), and translations (JSON with ~19 languages) [1]. The contribution guidelines explicitly state that the name field should use official, commonly recognized names in English, with native names going in the native field.

In practice, this policy isn't uniformly enforced. Poland follows it correctly (e.g. "Greater Poland", "Silesia" in name, Polish equivalents in native) , but Czechia uses Czech names in both name and native . This appears to be a data quality gap from community contributions coming from different sources without automated language validation.

Worth noting: a maintainer response on a similar question about Spain referenced following "local/official conventions" [2], which suggests some flexibility in how the English-first policy is applied — but that does create the exact inconsistency you're describing.

If you need a workaround, the translations field (when populated) provides English under the en key, which may be more consistent across countries than the name field.

To reply, just mention @dosu.

---

Docs are dead. Just use Dosu.

[dr5hn]

Thanks for raising this — flagging the inconsistency is fair. Setting policy:

Convention going forward:

• name field → always English (or the canonical international transliteration)
• native field → the local-script name in the country's primary language

So Mazowieckie (Polish), Středočeský kraj (Czech), etc. — those should live in native. The name field should be the English form (Masovian Voivodeship, Central Bohemian Region).

This matches how every consumer-facing application uses the data: name is the lookup/display key in English; native is offered as an alternative for localized UIs.

The existing data is inconsistent because contributions from different sources used different conventions. The repo will fix forward — newly merged PRs follow this rule (e.g., #1382 was just adjusted to keep English in name for Tibet and Hong Kong SAR). Older inconsistencies will be cleaned up country-by-country in subsequent passes.

If you'd like to contribute a cleanup for a specific country (e.g., Czech Republic), open a PR that updates only the name field on contributions/states/states.json entries with non-English values (and leave native intact). That's the most welcome contribution form.

Closing this as the policy is now documented; reopen if you spot a specific case where the rule is ambiguous.

[dr5hn]

Policy documented above; closing as policy issue. Country-specific cleanups will come as separate PRs.