Why several languages
Sardar Teja Singh Samundri’s history belongs to Punjab, to Sikh history, to India and the wider subcontinent, and to a diaspora in Canada, the United Kingdom, the United States and Australia. No single language reaches all of those readers. The archive therefore treats language as access: each edition opens the same documented record to a different community, with the same sources, evidence tiers and contested-status rules underneath.
Why English is the control edition
English is the edition in which changes are made first and from which every other edition is kept in step, because the archive’s research record and its change control are organised around it, and because it is the widest shared language of the institutions that use the archive. That is an operational choice, not a claim that English is historically more authoritative. Where an original source is in Punjabi, Urdu or any other language, the original source is the evidence — not the English, and not any translation of it.
Punjabi, in two scripts
Punjabi is the language of the people, institutions and cultural world in which Samundri worked, and Gurmukhi is the script of Punjabi in Indian Punjab and of Sikh communities worldwide. The Gurmukhi edition is part of the archive’s purpose, not an optional extra, and its long-term aim is the whole archive. Today it is a structural draft, marked on every page as awaiting review by a native reader; see its translation status.
Much of Samundri’s life, and of the movement’s history, is tied to places now in Pakistan — Lahore, Nankana Sahib, and the Lyallpur district, today Faisalabad, where his family’s village lay. Shahmukhi is Punjabi written in the Arabic script, and it gives Punjabi readers who use that script a direct way into this history. Gurmukhi and Shahmukhi are two scripts of one language, and the archive presents them that way: two doors onto the same history, never two rival histories.
Hindi, Spanish, French and Urdu
Hindi opens the archive to students, teachers and general readers across India, well beyond the Punjabi-speaking regions. A Hindi edition must not flatten Sikh and Punjabi terms into generic equivalents, so it follows the archive’s governed terminology.
Spanish is by a wide margin the most common language other than English spoken in homes in the United States. A bounded Spanish edition would make the archive usable in American schools, museums and public-history settings, well beyond the South Asian diaspora. It is chosen for its reach, not for any direct historical connection to Samundri.
French is one of Canada’s two official languages, and Canada’s 2021 census counted more than half a million people who speak Punjabi most often at home. English, Punjabi and French together serve Canadian institutions, schools and the public.
Urdu widens access across Pakistan and its institutions. It is offered alongside Punjabi in Shahmukhi, not instead of it: Punjabi is the mother tongue of most of the population of Pakistan’s Punjab province.
Why not every major language
Every translation becomes a lasting editorial obligation. When a claim is corrected, a tier changes, a citation is added or a contested point is resolved, every edition has to follow — and a translation that falls behind would quietly misstate the record. A language is therefore added only when there is a clear audience or heritage purpose and a realistic path to human review and upkeep. German, Portuguese, Arabic, Mandarin and Italian are on a watch list. None of them is promised.
How AI is used
AI may help to prepare drafts, check terminology, and find passages that are missing or out of date. Machine output is not published as the archive’s text: a page in another edition is published only after a named reviewer fluent in its language has checked it, and a page that touches sensitive or contested history must also be checked against the evidence by a second person. Until then a draft is kept out of the public site entirely. One exception predates these rules, and is stated here plainly: the Gurmukhi edition’s structural draft — its interface, framing and tier key, drafted with AI assistance from the English, not a translation of the historical narrative — was public before this system existed, and is marked on every page as awaiting review by a native reader.
What “under review” means
Every translated page carries one status, and moves forward only through review:
- Machine-assisted draft — prepared with AI assistance from the English, and not yet read by a reviewer. Never public.
- Language reviewed — checked by a named reviewer fluent in the language.
- Historical check passed — checked by a second person against the evidence: the same claims, the same tiers, the same hedges, nothing strengthened or weakened.
- Accessibility checked — the page works with a screen reader and a keyboard, and reads correctly in its own script and direction.
- Published — only now does the page appear on the public site, in its sitemap and in search.
- Source changed — the English has changed since the page was reviewed. The page leaves the public site until it is reviewed again; this happens automatically.
The same evidence in every edition
A translator may change the language of a sentence, never its evidential force. Every edition shows the same claims with the same evidence tiers, sources and contested status; family testimony stays labelled as family testimony; and a claim this archive withholds is withheld in every language. None of this is copied by hand. The evidence badges and drawers in every edition are drawn from the archive’s single evidence record, and the site will not build with a translation that drops a claim or loses a hedge such as “according to” or “contested”.
If two editions ever disagree, that is treated as a defect in the translation, never as a choice between versions: the claim, the English control text and the cited evidence are checked, and the translation is corrected and reviewed again before it returns.
The language menu at the top of every page lists each edition in its own language and script. It takes you to the same page in the edition you choose where that page exists, and says plainly when it does not. The archive never changes language on your behalf because of where you appear to be browsing from.
Where the figures come from
These counts measure different things — language spoken at home, main language, mother tongue — in different years, and are not comparable with one another. They are reasons to serve a readership, not a ranking of languages.
- United States — U.S. Census Bureau, American Community Survey 2023 (1-year), table B16001: 43,369,734 people aged five and over speak Spanish at home.
- Canada — Statistics Canada, Census of Population 2021: 520,390 people speak Punjabi predominantly at home. English and French are Canada’s official languages under section 16 of the Canadian Charter of Rights and Freedoms.
- Pakistan — Pakistan Bureau of Statistics, Census 2023, Punjab province, table 11: Punjabi is the mother tongue of 85,309,591 of 127,333,305 people (67.0 per cent).
- India — Census of India 2011, language tables (C-16): 528,347,193 people have Hindi as their first language. These are 2011 figures, from the latest completed language census, not present-day counts.