X-SAMPA (Extended SAMPA) is John C. Wells's 1995 proposal for a keyboard-compatible ASCII coding of the entire IPA — "everything on the 1993 IPA Chart, including diacritics and tone marks" — so that IPA-transcribed material could travel over e-mail and other 7-bit channels.1 Where SAMPA is a family of per-language tables, X-SAMPA is one universal mapping. It remains the most widely used ASCII-IPA in linguistics-adjacent computing and hobbyist communities.
The machinery: two special characters
- The backslash \ is the "universal diacritic": it means the preceding character is to be read in a special way, roughly doubling the symbol space. G = voiced velar fricative ɣ, but G\ = voiced uvular plosive ɢ; R = voiced uvular fricative ʁ, but R\ = uvular trill ʀ. The backslash binds to exactly one preceding character.1
- The underscore _ introduces diacritics: t_w labialized, t_h aspirated, d_n nasal release. Where \ and _ co-occur, \ binds first: t_?\ parses as t + pharyngealized, because ?\ (the pharyngeal fricative) is resolved before the underscore attaches.1
- Two SAMPA marks were redefined: the apostrophe ' became palatalization (it had marked rising tone), and the backtick ` became rhoticity/retroflexion — t` is retroflex t, @` is r-coloured schwa. Ejectives are _> and implosives _<; the retroflex approximant is the composite r\`.1
The scheme's defining property, stated outright in the proposal: symbol strings remain uniquely parsable even with no spaces between successive characters. Tones ride either on _-diacritics (_T extra-high … _B extra-low; _R rising, _F falling) or on a separate tier in angle brackets, following the companion SAMPROSA work.1
A worked example
| X-SAMPA | IPA | English |
|---|---|---|
| "fVn@tIks | ˈfʌnətɪks | phonetics |
| eks"s{mp@ | ɛksˈsæmpə | X-SAMPA (approximately!) |
| t` | ʈ | retroflex t |
| G\ | ɢ | voiced uvular plosive |
| s_j | sʲ | palatalized s |
| b_< | ɓ | voiced bilabial implosive |
Status: a proposal that became a de-facto standard
Wells framed the paper explicitly as a proposal soliciting colleagues' reactions; no body ever formally ratified it. Its standardhood is by adoption. With Unicode IPA now ubiquitous, X-SAMPA's living roles are input convenience, legacy data, and 7-bit contexts.12
Errors to avoid
- Mixing X-SAMPA with per-language SAMPA values — the same ASCII glyph can mean different things in a SAMPA table and in X-SAMPA.1
- Assuming any ASCII-IPA you meet online is X-SAMPA: Kirshenbaum and WorldBet are distinct schemes with conflicting assignments (the letter Q famously means ɒ in X-SAMPA-adjacent schemes, ɣ in Kirshenbaum, and ɢ elsewhere).
- Citing "CXS" (Conlang X-SAMPA) as a standard — it appears nowhere in Wells's defining document and has no authoritative specification; treat it as an unstandardized community variant.1
- Reversing the _/\ precedence, or attaching \ to more than one preceding character.1
Learn more
- Wells (1995), "Computer-coding the IPA: a proposed extension of SAMPA" (PDF) — The defining document — read in full for this guide.
- X-SAMPA summary page (UCL) — IPA-free HTML listing, with Unicode code points added in 2000.
- SAMPA — our companion guide — The per-language parent system.
Notes & Bibliography
- Wells, J. C. "Computer-coding the IPA: a proposed extension of SAMPA." Revised draft 1995-04-28, UCL — read in full: §1 scope (1993 chart), §3/§27 base rule, §5 apostrophe, §6 backtick, §12 backslash, §13 underscore + precedence, §14/§18 ejectives/implosives, §20–24 tones, §26 unique parsability; proposal framing. [source] ↩
- X-SAMPA summary page, UCL (last revised 2000-05-03, adding Unicode code points) — case significance, composite r\` example. [source] ↩