English month names

The 2026-09-27 month-name extension accepts all 12 full English Gregorian month names in Nepali mode. Each entry has one preferred spelling and no built-in alternatives. These are authorized source-assisted project preferences; independent human linguistic review remains pending. They are separate from the 20-word English loanword pilot and its 34-entry research catalogue.

Mappings

Both lowercase and normal title case are accepted:

English key	Title case	Preferred output
january	January	जनवरी
february	February	फेब्रुअरी
march	March	मार्च
april	April	अप्रिल
may	May	मे
june	June	जुन
july	July	जुलाई
august	August	अगस्ट
september	September	सेप्टेम्बर
october	October	अक्टोबर
november	November	नोभेम्बर
december	December	डिसेम्बर

Try January February March → जनवरी फेब्रुअरी मार्च or September December → सेप्टेम्बर डिसेम्बर. The core convertWord and convertText functions use these entries directly. Browser adapters use them in attached fields while Nepali mode is on; the demo requires no additional setting. The typing reference and API reference describe integration.

Case and custom entries

Ordinary capitalization follows the existing lookup normalization. September and December are explicit built-in aliases because initial S and D otherwise select reserved Shift sounds. Other uses of T, D, S, R and vowel-following H still select their sounds; there is no general all-capital or mixed-case English folding. For example, SEPTEMBER and DECEMBER are not month aliases.

The existing createEngine({ entries }) API can replace any month entry's full candidate list in one engine instance. Replacing lowercase september or december also updates its title-case alias unless an exact September or December override is supplied. The existing Ram and Sita aliases follow the same inheritance rule. See the custom-entry examples.

Meaning and boundaries

• May and may both give मे. The engine cannot distinguish the English modal verb “may” from the month name; disable Nepali conversion or use an English field for literal English. Native Roman input such as maya keeps its existing behavior.
• Only full month names are listed. Abbreviations such as jan, feb, mar and sep do not inherit these entries. may is already the full name of a month.
• Attached forms such as januaryma and decemberko use the existing phonetic fallback. Entries do not infer stems or suffixes.
• Month names do not convert Gregorian dates to Bikram Sambat or substitute Bikram Sambat month names. The later digit rendering policy now displays ASCII digits as Devanagari by default (September 27, 2026 → सेप्टेम्बर २७, २०२६), with punctuation unchanged. The dated month benchmark retains its original ASCII-digit outputs and source identities.
• The later shared mixed-text policy preserves recognizable links, ASCII domain-shaped hosts, and ordinary ASCII email addresses. convertText('may.com') now stays may.com by default, while the standalone month may still gives मे. This policy is separate from the month-name extension and its dated evidence. Arbitrary English or code remains convertible: leave whole English fields unattached, mark them data-sahajlipi-ignore, or disable conversion while typing or pasting literal fragments.

Source evidence

Accessed 2026-09-27. The preferred outputs match the Nepali Gregorian month data in the official Unicode CLDR JSON 48.2.1 source, pinned to commit 26a79cb42bfcc90def764102aa2af126d9ef3108. The full JSON was downloaded and parsed. The selected path is main.ne.dates.calendars.gregorian.months.format.wide; the resolved stand-alone.wide values are byte-identical for all 12 months. Its SHA-256 is cb31c68c4b999191d8eaee6262d8d205f68a7b0a99df6b0cda94e422963e5d9b.

This uses the stable 48.2.1 JSON package release, whose release note describes a time-zone patch over CLDR 48.2. The pinned repository README describes generated JSON with a contributed-or-approved draft threshold. These are locale data used for the project's defaults, not independent human labels for SahajLipi or a claim that every other observed spelling is incorrect.

Government pages provide narrower corroboration: the Ministry of Foreign Affairs disclosure body contains अगस्ट, सेप्टेम्बर and अक्टोबर in dates, and the Department of Postal Service homepage notice title contains नोभेम्बर and डिसेम्बर. The postal notice's linked page was inaccessible, so the claim is limited to its homepage title. The ministry page also contains जुलार्इ; this observed alternate encoding/spelling is recorded for review and is not adopted as an output or candidate.

The machine-readable evidence ledger preserves source URLs, access scopes, context checks, the observed variant and limitations. The month data are available under Unicode License V3; the exact copyright and permission notice is retained in LICENSES/Unicode-3.0.txt. The project code remains MIT licensed.

Verification scope

The dated month benchmark and benchmark guide record the separate month fixture, baseline, post-change checks and provenance. The tests cover lowercase and title case, one-candidate results, custom-entry alias inheritance, reserved Shift controls, text conversion and live field behavior. These verify the intended implementation contracts; they do not establish population-wide typing accuracy or independent linguistic review. Historical loanword research and unrelated dataset metrics retain their original scope.