Paiboon and Paiboon+ schemes¶

Two scheme IDs are available: paiboon for the original system used in Paiboon Publishing's first-edition Thai learner series, and paiboon_plus for the revised system adopted in the 2009 Three-Way dictionary and later titles. The two schemes share every onset, coda, and monophthong. They diverge only in how centring diphthongs spell their length contrast.

from thaiphon import transcribe

transcribe("น้ำ", scheme="paiboon")
# 'náam'

transcribe("สวัสดี", scheme="paiboon")
# 'sà-wàt-dii'

transcribe("เสือ", scheme="paiboon")
# 'sʉ̌a'

transcribe("เสือ", scheme="paiboon_plus")
# 'sʉ̌ʉa'

Format conventions¶

Tone notation¶

Mid tone is not marked. The four other tones use a combining diacritic on the first vowel letter of the syllable. After NFC normalisation, plain Latin vowels precompose with their diacritics; the IPA-letter vowels ʉ ɛ ɔ ə have no precomposed forms and keep the combining codepoint.

Onset map (IPA → Paiboon)¶

The convention is designed for English readers. Aspirated stops take the bare letter, since that matches the English intuition for p t k. Unaspirated voiceless stops — which English has no single-letter way to write distinctively — use digraphs that mimic the clusters in English words like "spa", "stop", "ski".

Aspirated vs. unaspirated: The key asymmetry is that aspirated stops are written with a single letter (p t k) while unaspirated voiceless stops take a digraph (bp dt g). New readers should keep this in mind: p in Paiboon is always the aspirated /pʰ/, and bp is always the unaspirated /p/.

transcribe("พา", scheme="paiboon")   # 'paa'   — aspirated /pʰ/
transcribe("ปา", scheme="paiboon")   # 'bpaa'  — unaspirated /p/
transcribe("ทา", scheme="paiboon")   # 'taa'   — aspirated /tʰ/
transcribe("ตา", scheme="paiboon")   # 'dtaa'  — unaspirated /t/
transcribe("คา", scheme="paiboon")   # 'kaa'   — aspirated /kʰ/
transcribe("กา", scheme="paiboon")   # 'gaa'   — unaspirated /k/

Vowel map (IPA → Paiboon)¶

Long monophthongs are doubled. Four IPA-letter vowels are used: ʉ (U+0289) for /ɯ/, ɛ (U+025B) for /ɛ/, ɔ (U+0254) for /ɔ/, and ə (U+0259) for /ɤ/.

Centring diphthongs — where paiboon and paiboon_plus differ¶

This is the only point of divergence between the two schemes.

paiboon uses the same spelling for both short and long centring diphthongs. paiboon_plus doubles the long form, which lets readers distinguish length in words where the distinction matters.

The /iː/+/w/ coda sequence (เขียว pattern) follows the same logic: paiboon renders it as iao, and paiboon_plus as iiao.

transcribe("เสือ",  scheme="paiboon")       # 'sʉ̌a'
transcribe("เสือ",  scheme="paiboon_plus")  # 'sʉ̌ʉa'

transcribe("เขียว", scheme="paiboon")       # 'kǐao'
transcribe("เขียว", scheme="paiboon_plus")  # 'kǐiao'

Coda map (IPA → Paiboon)¶

Offglide /w/: The coda /w/ writes as o in most contexts, giving combinations like aao, ɛɛo, ɔɔo. Short /i/ + /w/ is the exception — it spells as iu rather than io, matching the English intuition for words like "Matthew".

Offglide /j/: The coda /j/ writes as i, giving combinations like aai, ɔɔi, əəi.

transcribe("ข้าว", scheme="paiboon")    # 'kâao'
transcribe("ลิฟต์", scheme="paiboon")   # 'líf'
transcribe("ฟุตบอล", scheme="paiboon")  # 'fút-bɔɔn'

Loanword coda preservation¶

Under the default everyday profile, the engine preserves foreign codas in lexicon-listed modern loanwords. For words like ลิฟต์ (elevator), เชฟ (chef), and ออฟฟิศ (office), the coda surfaces as f rather than collapsing to p. Some loans — such as กราฟ (graph) and กอล์ฟ (golf) — only preserve under the careful_educated profile.

transcribe("ลิฟต์", scheme="paiboon")                          # 'líf'
transcribe("ลิฟต์", scheme="paiboon", profile="etalon_compat") # 'líp'

transcribe("กราฟ", scheme="paiboon")                                    # 'graap'
transcribe("กราฟ", scheme="paiboon", profile="careful_educated")        # 'graaf'

Pass profile="etalon_compat" to opt back into strict native-coda collapse for every word, regardless of the lexicon.

See Reading profiles for a full description of each profile.

Examples¶

All outputs are verified against the engine.