Mongolian Pronunciation and IPA Transcription Guide
← Back to Mongolian transcription ↑ All Languages
Table of Contents
- Getting Started
- Quick Reference (Glossary)
- How to Read IPA Symbols
- Mongolian Pronunciation Guide
- Initial vs Non-Initial Vowels
- Long Vowels and Diphthongs
- Iotated Vowels (я, ю, ё, е) and Palatalization
- Stops: Aspiration, Not Voicing
- Affricates and Sibilants
- Nasals: н, м, г and the Velar Nasal
- Liquids: л and р
- х, в, ф and щ
- Geminates and Consonant Clusters
- The Soft and Hard Signs (ь, ъ)
- Final Short Vowel Deletion
- Vowel Harmony
- Stress
- Orthography to IPA Mapping
- Using the Tool: Practical Notes
- For Developers
- Related Pronunciation Guides
About This Tool
This Mongolian transcription app uses a modified version of the Wiktionary Mongolian Pronunciation Module to turn Cyrillic Mongolian spelling into a phonemic IPA transcription. IPA (International Phonetic Alphabet) is a standardized system for transcribing speech sounds. This tool helps learners, linguists and developers see how a Mongolian word is pronounced. Phonemic transcription represents the abstract sound system (phonemes) rather than every detail of actual speech. The phoneme inventory and the vowel treatment follow The Phonology of Mongolian by Jan-Olof Svantesson et al. (2005), aiming at the Standard Halh dialect as spoken in Ulaanbaatar.
Unlike German, French or Russian in this app, Mongolian uses no dictionary lexicon. Every word is
generated by the module's rules alone (the app calls window.mn_ipa.to_IPA(cleanText) directly), so
the output is exactly as good as the spelling rules. For example, сайн /saiŋ/ ("good"). Only the Cyrillic script is supported; traditional Mongolian script
is not.
Dialects & Limitations
- Supported: Default = Standard Halh (Khalkha), the official language of Mongolia, written in Cyrillic.
- Not supported: Inner Mongolian varieties (Chahar and others), Buryat, Oirat dialects, and the traditional vertical script. Unlike Inner Mongolian, Halh keeps diphthongs, and the module restores them (e.g. ай /ai/, not a monophthong).
- Edge cases: Russian loanwords with е, ё, ю, я after a consonant, words with the letter ы, and words containing ью are handled poorly (see Common Issues).
- Word-by-word: the app transcribes one space-separated word at a time. Hyphens are deleted before processing, so a hyphenated compound is treated as a single word.
Phonemic vs Phonetic
- Phonemic (the only form offered): a broad transcription showing only the sounds that distinguish meaning. Mongolian stops contrast aspiration, not voicing, but this module writes the non-aspirated (lenis) series with the voiced symbols b, d, ɡ and the aspirated (fortis) series as pʰ, tʰ, kʰ. For instance бага /baɡ/ uses ɡ, but хот /xɔtʰ/ ends in tʰ.
- Phonetic: not available for Mongolian.
Quick Reference (Glossary)
- Phoneme
- The smallest unit of sound that distinguishes meaning. In Mongolian, for example, aspirated /tʰ/ and lenis /d/ are different phonemes: тав /tʰaw̜/ vs. дав /daw̜/.
- Allophone
- A variant pronunciation of a phoneme that does not change meaning.
- Aspiration
- A puff of air after a stop, written ʰ. Mongolian п, т, к, ц, ч are aspirated, e.g. цас /t͡sʰas/ "snow".
- Lenis / Fortis
- The two stop series of Mongolian. Fortis stops (п, т, к, ц, ч) are aspirated; lenis stops (б, д, г, з, ж) are not, and are written here with voiced symbols.
- Long Vowel
- A vowel written double in Cyrillic (аа, оо, уу, үү, өө, ээ, ий, ии) and transcribed with ː, e.g. хоол /xɔːɮ/ "food".
- Diphthong
- A vowel plus glide in one syllable, such as ай, ой, ау: сайн /saiŋ/, ойр /ɔir̥/.
- Schwa
- The neutral vowel ə. Short vowels outside the first syllable become schwa, e.g. ахмад /axməd/.
- Epenthetic vowel
- A vowel with no phonemic weight that is inserted to break consonant clusters. The module treats non-initial short vowels this way (as in Svantesson's analysis).
- Palatalization
- Secondary raising of the tongue toward the hard palate, written ʲ, e.g. аль /aɮʲ/.
- Iotated Vowel
- The Cyrillic letters я, ю, ё, е, which stand for a glide j plus a vowel at the start of a syllable: юм /jʊm/.
- Velar Nasal
- The sound ŋ as in English "sing"; н is pronounced this way at the end of a word and before г, к, х: хүн /xuŋ/.
- Vowel Harmony
- The Mongolian system in which suffix vowels agree with the stem (back / front, rounded / unrounded). The transcription module does not use it to generate IPA; see Vowel Harmony.
- Final Vowel Deletion
- The module drops a word-final short vowel in words of two or more syllables: байна /bain/.
- Devoiced r
- The trill r becomes voiceless r̥ at the end of a word: сар /sar̥/.
- Tie Bar
- The mark ͡ joining the two parts of an affricate: t͡sʰ, d͡z, t͡ɕʰ, d͡ʒ.
- Geminate
- A doubled (long) consonant, written with ː: алла /aɮː/.
- Syllable Nucleus
- The vowel at the core of a syllable. The module splits words into syllables with
mn.syllables()and treats the first one differently from the rest.
How to Read IPA Symbols
The tables below list the symbols this engine actually emits for Mongolian, each with a real word transcribed by the engine.
Vowel Symbols
| IPA | Example | English Approximation | Notes |
|---|---|---|---|
| /a/ | ам /am/ | "father" (short) | Cyrillic а in the first syllable |
| /aː/ | аав /aːw̜/ | "father" | Long, аа |
| /ɔ/ | ном /nɔm/ | "law" (UK, short) | о in the first syllable |
| /ɔː/ | хоол /xɔːɮ/ | "law" (long) | Long, оо |
| /ʊ/ | ус /ʊs/ | "put" | у in the first syllable |
| /ʊː/ | уул /ʊːɮ/ | "food" (lax) | Long, уу |
| /u/ | үг /uɡ/ | "food" (tense) | The front rounded ү is written u here, and у is ʊ |
| /uː/ | хүү /xuː/ | "food" (long) | Long, үү |
| /ɵ/ | хөх /xɵx/ | No English equivalent | Mid central rounded; ө |
| /ɵː/ | өөр /ɵːr̥/ | No English equivalent | Long, өө |
| /i/ | чи /t͡ɕʰi/ | "sit/see" | и, and also initial э: эр /ir̥/ |
| /iː/ | ийм /iːm/ | "see" (long) | ии, ий |
| /e/ | эмэгтэй /iməɡtʰe/ | "say" (no glide) | Only from long ээ or non-initial эй (after final deletion) |
| /eː/ | ширээ /ɕireː/ | "say" (no glide) | Long, ээ |
| /ə/ | ахмад /axməd/ | "about" | Every non-initial short а, о, у, э, ө, ү |
| /ɒː/ | ао /ɒː/ | "lot" (long) | The sequence ао |
| /ɛː/ | яай /jɛː/ | "bear" (non-rhotic) | Only from the sequences яай, уай |
| /ai/ | сайн /saiŋ/ | "eye" | Diphthong ай |
| /ɔi/ | ойр /ɔir̥/ | "boy" | Diphthong ой |
| /ʊi/ | уй /ʊi/ | "phooey" | Diphthong уй |
| /ui/ | үй /ui/ | No English equivalent | Diphthong үй |
| /ɵi/ | өй /ɵi/ | No English equivalent | Diphthong өй |
| /ei/ | эй /ei/ | "day" | Diphthong эй |
| /aʊ̯/ | ау /aʊ̯/ | "cow" | Diphthong ау |
Consonant Symbols
| IPA | Example | English Approximation | Notes |
|---|---|---|---|
| /pʰ/ | пүү /pʰuː/ | "pie" | Aspirated; letter п (rare in native words) |
| /b/ | бүх /bux/ | "bee" | Lenis; letter б |
| /tʰ/ | тав /tʰaw̜/ | "tie" | Aspirated; letter т |
| /d/ | дав /daw̜/ | "die" | Lenis; letter д |
| /kʰ/ | банк /baŋkʰ/ | "kite" | Aspirated; letter к |
| /ɡ/ | гар /ɡar̥/ | "go" | Lenis; letter г |
| /t͡sʰ/ | цас /t͡sʰas/ | "cats" (aspirated) | Letter ц |
| /d͡z/ | зун /d͡zʊŋ/ | "adze" | Lenis; letter з |
| /t͡ɕʰ/ | чих /t͡ɕʰix/ | "cheek" (softer, aspirated) | Letter ч |
| /d͡ʒ/ | жил /d͡ʒiɮ/ | "judge" | Lenis; letter ж |
| /f/ | фото /fɔtʰ/ | "fan" | Loanwords only |
| /w̜/ | тав /tʰaw̜/ | "water" (weakly rounded) | Letter в |
| /s/ | сар /sar̥/ | "sun" | Letter с |
| /ɕ/ | шар /ɕar̥/ | "she" (palatal) | Letters ш and щ |
| /x/ | хот /xɔtʰ/ | Scottish "loch" | Letter х |
| /ɬʰ/ | лхагва /ɬʰaɡw̜/ | Welsh "ll" | The cluster лх |
| /m/ | ном /nɔm/ | "man" | Letter м |
| /n/ | нар /nar̥/ | "no" | Letter н (not final, not before г/к/х) |
| /ŋ/ | хүн /xuŋ/ | "sing" | н at word end or before г, к, х |
| /r/ | гарах /ɡarəx/ | Spanish "pero" | Trill/tap; letter р |
| /r̥/ | сар /sar̥/ | No English equivalent | Voiceless r at the end of a word, and before п, т, к, ц, ч: гэрт /ɡir̥tʰ/ |
| /ɮ/ | улс /ʊɮs/ | No English equivalent | Voiced lateral fricative; letter л |
| /j/ | ямар /jamər̥/ | "yes" | The glide in я, ю, ё, е |
Diacritical Marks
| Symbol | Name | Meaning | Example |
|---|---|---|---|
| ː | Length mark | Long vowel or geminate consonant | аав /aːw̜/, алла /aɮː/ |
| ʰ | Aspiration | Puff of air after the stop | хот /xɔtʰ/ |
| ʲ | Palatalization | Consonant with raised tongue body | аль /aɮʲ/, бя /bʲa/ |
| ̯ | Non-syllabic | Glide part of a diphthong or hiatus | ау /aʊ̯/, радио /radi̯oː/ |
| ͡ | Tie bar | Affricate (single unit) | зун /d͡zʊŋ/, чих /t͡ɕʰix/ |
| ̥ | Voiceless | Normally voiced sound devoiced | сар /sar̥/ |
| ̜ | Less rounded | The labial glide of в is written w̜ | тав /tʰaw̜/ |
| ͉ | Left angle below | Appears only in the output for дд; marks a stop with no audible release, followed by a syllable dot | аддаа /at͉.taː/ |
No stress marks (ˈ, ˌ) are ever produced. Stress is not marked in Mongolian transcription by this module (see Stress).
Interactive Features
- Click words to cycle variants: Mongolian gives only one transcription per word, so clicking does not change anything.
- Audio playback: Click the speaker icon next to any word or line to hear text-to-speech pronunciation (requires browser TTS support).
- Export results: Use PDF or CSV buttons to save transcriptions in your preferred format.
Multiple Pronunciation Variants
For Mongolian each word has exactly one rule-generated output. There is no lexicon and no per-dialect variant, so the cycling feature available for German or Latin does not apply. If a word is transcribed incorrectly (see Common Issues), try an alternative spelling that reaches the intended rule.
Mongolian Pronunciation Guide
This guide explains how the module turns Cyrillic Mongolian into IPA. Spelling is fairly regular, but the Cyrillic letters hide a lot of structure: long vowels are written double, the vowel of any syllable after the first is reduced, and the final vowel of most words is not pronounced.
Initial vs Non-Initial Vowels
The most important rule in the module is that the first syllable keeps its full vowel quality and every short vowel in a later syllable is reduced to schwa ə (the exception is и, й, which stays i).
- ном /nɔm/ keeps ɔ, but in модонд /mɔdənd/ the second о becomes ə.
- ажил /ad͡ʒiɮ/ and багш /baɡɕ/: и is stable, and a consonant cluster without vowel stays as is.
- Initial э and и merge: тэр /tʰir̥/ and тир /tʰir̥/ have the same vowel. Initial ү is u while у is ʊ: үг /uɡ/ vs ус /ʊs/.
- Long vowels and diphthongs are exempt from reduction, wherever they occur: хаалга /xaːɮɡ/, наймаа /naimaː/.
Long Vowels and Diphthongs
Doubled vowel letters are long: аа, оо, уу, үү, өө, ээ. The letters ий and ии also give a long iː.
- аа /aː/, оо /ɔː/, уу /ʊː/, үү /uː/, өө /ɵː/, ээ /eː/.
- Long ээ is the only common source of eː: ширээ /ɕireː/, гэгээн /ɡiɡeːŋ/.
- Diphthongs in й: ай /ai/, ой /ɔi/, уй /ʊi/, үй /ui/, өй /ɵi/, эй /ei/. Halh keeps these as diphthongs.
- Rising diphthong-like sequences: ау /aʊ̯/, оа /ʊ̯a/, уа /u̯a/, ио /i̯oː/: ариун /ari̯ʊːŋ/, радио /radi̯oː/.
- The doubled vowels are matched before single vowels inside each syllable, so long vowels are not reduced: даваа /daw̜aː/, долоо /dɔɮɔː/.
Iotated Vowels (я, ю, ё, е) and Palatalization
At the start of a word or after a vowel, the letters я, ю, ё, е are a glide plus a vowel. In the first syllable they give ja, jʊ, jɔ, jɵ; later in the word the vowel reduces to schwa (jə).
- я /ja/, ю /jʊ/, ё /jɔ/, е /jɵ/.
- ямар /jamər̥/, баяр /bajər̥/, хоёр /xɔjər̥/, ес /jɵs/.
After one of the consonants п, б, т, д, к, г, ф, в, х, м, н, р, л, the letters е, ю, ё, я do not give a glide. Instead the consonant is palatalized and the vowel is rewritten, before any other processing: е → ө, ю → у, ё → о, я → а. Hence:
- бя /bʲa/, ня /nʲa/, ля /ɮʲa/, тю /tʰʲʊ/, дё /dʲɔ/.
- Words with е after a consonant therefore get ɵ (a rounded vowel), e.g. Герман /ɡʲɵrməŋ/. This is a rule for Mongolian-like vowel harmony, not for Russian-style е (see Common Issues).
Palatalization after ж, з, с, ц, ч, ш, щ is not implemented (they are not in the palatalization table).
Stops: Aspiration, Not Voicing
Mongolian stops contrast aspiration, not voicing. The module renders the fortis letters as aspirated and the lenis letters with voiced symbols:
- Aspirated: т → tʰ тав /tʰaw̜/, к → kʰ банк /baŋkʰ/, п → pʰ пүү /pʰuː/, ц and ч likewise.
- Lenis: б → b бүх /bux/, д → d дав /daw̜/, г → ɡ гар /ɡar̥/.
- Intervocalic voicing is not shown: the aspirated letters stay aspirated between vowels: акаа /akʰaː/, апаа /apʰaː/.
Affricates and Sibilants
- ц → t͡sʰ цас /t͡sʰas/; з → d͡z зун /d͡zʊŋ/.
- ч → t͡ɕʰ чих /t͡ɕʰix/; ж → d͡ʒ жил /d͡ʒiɮ/.
- с → s сар /sar̥/; ш and щ → ɕ шар /ɕar̥/, щи /ɕi/.
Nasals: н, м, г and the Velar Nasal
- н at the end of a word is ŋ: хүн /xuŋ/, сайн /saiŋ/.
- н before г, к, х assimilates to ŋ: монгол /mɔŋɡəɮ/, банк /baŋkʰ/, мөнх /mɵŋx/.
- н before б, п, м, ф assimilates to m: дэнбэ /dimp/; the cluster нб is written mp, as in улаанбаатар /ʊɮaːmpaːtʰər̥/.
- Otherwise н stays n: нар /nar̥/, гэнэт /ɡinətʰ/.
Liquids: л and р
- л is the voiced lateral fricative ɮ: улс /ʊɮs/, жил /d͡ʒiɮ/. The cluster лх is voiceless ɬʰ: лхагва /ɬʰaɡw̜/.
- р is r inside the word: гарах /ɡarəx/. At the end of a word it is devoiced to r̥: сар /sar̥/, өөр /ɵːr̥/.
- р before п, т, к, ц, ч is also voiceless: эрт /ir̥tʰ/, цирк /t͡sʰir̥kʰ/.
х, в, ф and щ
- х → x in all positions: хот /xɔtʰ/, ах /ax/.
- в → w̜: тав /tʰaw̜/, авах /aw̜əx/. It is the same symbol in all positions.
- ф → f (loanwords): фото /fɔtʰ/.
- щ is treated as ш: ɕ.
Geminates and Consonant Clusters
- Doubled л, м, н, с are long: алла /aɮː/, аммаа /amːaː/, Анна /anː/, русс /rʊsː/.
- Doubled п, т, к, г, ц, з, ч, ж, ф, в, ш, щ, х, р are simplified to a single consonant: аппа /apʰ/ (final vowel is deleted).
- бб and вб become b: бб /b/, вб /b/.
- дд is an unusual case: аддаа /at͉.taː/, Будда /bʊt͉.t/ (a geminate stop written with a syllable dot).
- No vowel is inserted into clusters; the syllabifier treats permitted final clusters as a single syllable (see Implementation Details): багш /baɡɕ/, улс /ʊɮs/.
The Soft and Hard Signs (ь, ъ)
- ь gives the palatalization mark ʲ on the preceding consonant: аль /aɮʲ/, арьс /arʲs/, сургууль /sʊrɡʊːɮʲ/, чинь /t͡ɕʰinʲ/.
- ъ is silent and only separates syllables: объект /ɔbjəkʰtʰ/, подъезд /pʰɔdjəd͡zd/.
- ью crashes the engine; see Common Issues.
Final Short Vowel Deletion
If the word has more than one syllable, a final short vowel is dropped (a code comment calls it "standard practice for 'n' stems etc."). The check is on the final character of the output, so it removes a, e, i, ɔ, ɵ, u, ʊ or ə (including the second part of ei). Long vowels are not touched. A final i is protected only after a, ɔ, u or ʊ (the diphthongs ai, ɔi, ui, ʊi).
- байна /bain/, хаана /xaːn/, үнэ /un/, мөнгө /mɵŋɡ/, шинэ /ɕin/.
- Preserved: ширээ /ɕireː/ (long), манай /manai/, арай /arai/ (diphthong ai).
- эмэгтэй /iməɡtʰe/ loses the i of ei.
- One-syllable words keep their vowel: чи /t͡ɕʰi/, ва /w̜a/.
Vowel Harmony
Mongolian vowel harmony (back / front, with rounding) is real and affects suffix spelling, but the IPA
function export.to_IPA does not consult it. The helper mn.vowelharmony() in
mn-common.lua exists for the Wiktionary inflection templates; the only time to_IPA
reaches it is through the syllabifier when it meets ью (which is where the engine crashes). The
vowel letter itself carries the harmony in Cyrillic, and the transcription simply maps each letter, so a
word like гэр /ɡir̥/ and гар /ɡar̥/ differ only
by their spelled vowel. Masculine/feminine/neutral classes are not computed anywhere in the output.
Stress
Following Svantesson (2005), word stress is treated as non-phonemic in Mongolian. The stress-marking code of the original Wiktionary module was removed, so no ˈ or ˌ marks are ever produced: монгол /mɔŋɡəɮ/.
Comprehensive Spelling-to-IPA Mapping
These tables reflect the actual lookup tables in mn-pron_wasm.lua.
1. Short Vowels (vowels1 and vowels1_reduced)
| Letter | First syllable | Example (initial) | Later syllable | Example (non-initial) |
|---|---|---|---|---|
| а | a | ам /am/ | ə | ахмад /axməd/ |
| о | ɔ | ор /ɔr̥/ | ə | модонд /mɔdənd/ |
| у | ʊ | ус /ʊs/ | ə | Short у outside the first syllable is rare in native words |
| э | i | эр /ir̥/ | ə | эрдэм /irdəm/ |
| ө | ɵ | өвөл /ɵw̜əɮ/ | ə | өвөл /ɵw̜əɮ/ |
| ү | u | үг /uɡ/ | ə | хүмүүс /xumuːs/ |
| и, й | i | ир /ir̥/ | i | ажил /ad͡ʒiɮ/ |
| я | ja | ямар /jamər̥/ | jə | баяр /bajər̥/ |
| ю | jʊ | юм /jʊm/ | jə | аюул /ajʊːɮ/ |
| ё | jɔ | ёс /jɔs/ | jə | хоёр /xɔjər̥/ |
| е | jɵ | ес /jɵs/ | jə | Вьетнам /w̜jɵtʰnəm/ |
| ә | ə | Not used in Halh Cyrillic; mapped for completeness | ||
| ы | Listed in the tables but excluded from the vowel pattern, so it is never converted (see Common Issues) | |||
2. Long Vowels and Diphthongs (vowels2, vowels3)
| Spelling | IPA | Example |
|---|---|---|
| аа | aː | аав /aːw̜/ |
| оо | ɔː | хоол /xɔːɮ/ |
| уу | ʊː | уул /ʊːɮ/ |
| үү | uː | хүү /xuː/ |
| өө | ɵː | өөр /ɵːr̥/ |
| ээ | eː | ширээ /ɕireː/ |
| ии, ий | iː | ийм /iːm/, нийт /niːtʰ/ |
| ай | ai | сайн /saiŋ/ |
| ой | ɔi | хойд /xɔid/ |
| уй | ʊi | хуйх /xʊix/ |
| үй | ui | үй /ui/ |
| өй | ɵi | өй /ɵi/ |
| эй | ei | эй /ei/ |
| ау | aʊ̯ | ау /aʊ̯/ |
| ао | ɒː | ао /ɒː/ |
| оа | ʊ̯a | оа /ʊ̯a/ |
| уа | u̯a | уа /u̯a/ |
| уи | u̯iː | уи /u̯iː/ |
| еа | ia | еа /ia/ |
| ео | eɔ | ео /eɔ/ |
| еу | iʊ̯ | еу /iʊ̯/ |
| иа, ио, иу, иэ, иө, иү | i̯ + long vowel | иа /i̯aː/, ио /i̯oː/, иу /i̯ʊː/, иэ /i̯eː/, иө /i̯ɵː/, иү /i̯uː/ |
| яа, ёо, еэ | glide + long vowel | яа /jaː/, ёо /jɔː/, еэ /jeː/ |
| яй, ёй, ей, юй | glide + diphthong | яй /jai/, ёй /jɔi/, ей /jei/, юй /jui/ |
| юу, юү | glide + long vowel | юу /jʊː/, юү /juː/ |
яай, уай (three-letter, vowels3) | jɛː, u̯ɛː | яай /jɛː/, уай /u̯ɛː/ |
| ъяа, ъёо, ьеэ, ёий, аоа | other three-letter sequences | ъяа /jaː/, ъёо /jɔː/, ьеэ /jeː/, ёий /jiː/, аоа /aʊ̯a/ |
3. Single Consonants (consonants1)
| Letter | IPA | Example |
|---|---|---|
| п | pʰ | пүү /pʰuː/ |
| б | b | бүх /bux/ |
| т | tʰ | тав /tʰaw̜/ |
| д | d | дав /daw̜/ |
| к | kʰ | банк /baŋkʰ/ |
| г | ɡ | гар /ɡar̥/ |
| ц | t͡sʰ | цас /t͡sʰas/ |
| з | d͡z | зун /d͡zʊŋ/ |
| ч | t͡ɕʰ | чих /t͡ɕʰix/ |
| ж | d͡ʒ | жил /d͡ʒiɮ/ |
| ф | f | фото /fɔtʰ/ |
| в | w̜ | тав /tʰaw̜/ |
| с | s | сар /sar̥/ |
| ш, щ | ɕ | шар /ɕar̥/, щи /ɕi/ |
| х | x | хот /xɔtʰ/ |
| м | m | ном /nɔm/ |
| н | n (→ ŋ, m by context) | нар /nar̥/, хүн /xuŋ/ |
| р | r (→ r̥ finally) | гарах /ɡarəx/, сар /sar̥/ |
| л | ɮ | улс /ʊɮs/ |
| ь | ʲ | аль /aɮʲ/ |
4. Clusters and Geminates (consonants2, double_consonants)
| Spelling | Result | Example |
|---|---|---|
| лл | ɮː | алла /aɮː/ |
| мм | mː | аммаа /amːaː/ |
| нн | nː | Анна /anː/ |
| сс | sː | русс /rʊsː/ |
| лх | ɬʰ | лхагва /ɬʰaɡw̜/ |
| бб, вб | b | бб /b/, вб /b/ |
| нб | mp | нбаа /mpaː/ |
| дд | t͉.t | аддаа /at͉.taː/ |
| рк, рп, рт, рц, рч | voiceless r̥ + aspirated stop/affricate | гэрт /ɡir̥tʰ/, цирк /t͡sʰir̥kʰ/ |
| пп, тт, кк, гг, цц, зз, чч, жж, фф, вв, шш, щщ, хх, рр | single consonant | аппа /apʰ/ |
5. Palatalization Table (palatalisation)
| Consonant + vowel letter | Rewritten as | Example |
|---|---|---|
| С + я | С + ʲ + а | бя /bʲa/, ня /nʲa/, ря /rʲa/, пя /pʰʲa/ |
| С + ю | С + ʲ + у | тю /tʰʲʊ/, гю /ɡʲʊ/, нюс /nʲʊs/ |
| С + ё | С + ʲ + о | дё /dʲɔ/, кё /kʰʲɔ/ |
| С + е | С + ʲ + ө | бе /bʲɵ/, те /tʰʲɵ/, телевиз /tʰʲɵɮʲɵw̜id͡z/ |
С is any of п, б, т, д, к, г, ф, в, х, м, н, р, л. For example хя /xʲa/, мя /mʲa/.
6. Context-Dependent Rules
| Context | Result | Example |
|---|---|---|
| н before г, к, х, or at the end of a word | ŋ | монгол /mɔŋɡəɮ/, хүн /xuŋ/, мөнх /mɵŋx/ |
| н before б, п, м, ф | m | дэнбэ /dimp/, нпа /mpʰa/ |
| р at the end of a word | r̥ | нар /nar̥/ |
| Final short vowel in a word of 2+ syllables | deleted | гэнэт /ɡinətʰ/, гарна /ɡarn/ |
| Hyphen | deleted | сайн-уу /sainʊː/ |
| Capital letters | lowercased | Монгол /mɔŋɡəɮ/, САЙН /saiŋ/ |
Using the Tool: Practical Notes
Do I need to mark the stress?
No. Stress is treated as non-phonemic in Mongolian and the module produces no stress marks, so the input is plain Cyrillic spelling. See Stress.
Why do the vowels and the ending of a word differ from the spelling?
Short vowels after the first syllable are reduced to ə, and a final short vowel is deleted in words of more than one syllable: модонд /mɔdənd/, байна /bain/. Long vowels and diphthongs are exempt. See Initial vs Non-Initial Vowels and Final Short Vowel Deletion.
Does the tool use a dictionary, and why is there only one result?
No. Every word is generated by the rules alone, so each word has exactly one transcription and a spelling error is transcribed as typed. Only a phonemic form is offered; see Phonemic vs Phonetic.
Why is a word with ы or ью transcribed incorrectly?
The letter ы is excluded from the module's vowel pattern and passes through unchanged: газрын /ɡad͡zrыŋ/, whereas газрин /ɡad͡zriŋ/ is converted. The sequence ью makes the Lua call fail, and the input is returned unchanged. Russian loanwords are also affected by the palatalization rewrite. See Common Issues & Limitations.
Implementation Details (for Developers)
The whole pipeline is the single function export.to_IPA(text) in
lua_modules/mn-pron_wasm.lua (about 85 lines). The JavaScript handler in utils.js calls
it only for the Phonemic form: window.mn_ipa.to_IPA(cleanText). It uses the helper module
mn-common.lua only for mn.syllables(), and mn/data.lua only for the
table of permitted word-final clusters. The rest of mn-common.lua (vowel harmony, final vowel removal,
the traditional-script variation selectors, the see and ipasee template entry points) is
not used by to_IPA. The module's show() function is the Wiktionary template entry point and
is not called by the app.
- Normalization: lowercase with
mw.ustring.lower, delete every hyphen. - Double consonant collapse: table
double_consonants. - Palatalization: table
palatalisation. - Syllabification:
mn.syllables()splits the text. - Per-syllable vowels:
vowels3,vowels2, thenvowels1orvowels1_reduced. - Consonant clusters: table
consonants2. - Single consonants: table
consonants1. - Contextual rules: nasal assimilation,
rdevoicing, final vowel deletion and cleanup.
Stage 0: Normalization and the ы marker
| Rule | Example Input | Output | Description |
|---|---|---|---|
mw.ustring.lower | САЙН | saiŋ | Capitals are folded, so case never affects the output. |
gsub(text, '-', '') | сайн-уу | sainʊː | Hyphens vanish, which merges the two parts into one word (the second part then counts as non-initial). |
ы$ → ы~, ы%s → ы~%s | номы | nɔmы | Appends a temporary ~ to word-final ы so the later vowel-deletion step cannot see it; ~ is removed in the last step. Since ы is never converted, this has no visible effect. |
Stage 1: Double Consonants (double_consonants)
| Rule | Example Input | Output | Description |
|---|---|---|---|
| пп, тт, кк, гг, цц, зз, чч, жж, фф, вв, шш, щщ, хх, рр → single | аппа | apʰ | Each key is replaced twice, so triples reduce too. Note that лл, мм, нн, сс, бб, дд are not in this table; they are handled in stage 5. |
Stage 2: Palatalization (palatalisation)
| Rule | Example Input | Output | Description |
|---|---|---|---|
| Cе → Cʲө, Cю → Cʲу, Cё → Cʲо, Cя → Cʲа | бя, ля | bʲa, ɮʲa | For C in п б т д к г ф в х м н р л. The vowel is replaced by its plain counterpart (which the normal vowel stage then converts) and the modifier letter ʲ is inserted directly after the consonant, so т, п, к come out as tʰʲ etc. |
| After other consonants or with no consonant | ямар | jamər̥ | No rewrite; я, ю, ё, е are treated as glide plus vowel in stage 4. |
Stage 3: Syllabification (mn.syllables)
The text is split into a list of syllables; the position of each syllable in this list decides how its vowels are converted. Break points are found by scanning each character:
| Rule | Example Input | Output | Description |
|---|---|---|---|
| Break at start of every word and after punctuation or space | ном | nɔm | The first syllable of the text is index 1. Note that the index continues across words if a string contains several (the app passes one word at a time). |
| Break before a vowel after ъ | подъезд | pʰɔdjəd͡zd | The hard sign forces a break and itself maps to the empty string. |
| Break before consonant + vowel (the consonant begins a new syllable) | модонд | mɔdənd | Scanning backward from the consonant, a break is added if a vowel is found first. |
| Word-final clusters | багш, улс | baɡɕ, ʊɮs | Clusters listed in syll_final_cons (mn/data.lua, sorted by length 1 to 6) stay in the same syllable. An unlisted cluster gets an extra vowelless syllable break (code comments cite буддизм, алгебра). |
| Iotated vowel after a vowel | баяр | bajər̥ | я, е, ё, ю directly after a vowel start a new syllable. |
| е or ю after ь | Вьетнам | w̜jɵtʰnəm | The е branch counts back to find whether the cluster is word-initial. The ю branch calls mn.vowelharmony() to decide front/back, and this call fails (see Common Issues). |
Stage 4: Per-Syllable Vowel Conversion
| Rule | Example Input | Output | Description |
|---|---|---|---|
vowels3: ъяа, ъёо, ьеэ, яай, уай, ёий, аоа | яай, аоа | jɛː, aʊ̯a | Three-letter sequences, tried first. |
vowels2: long vowels, diphthongs and glide sequences | хоол, сайн, ау | xɔːɮ, saiŋ, aʊ̯ | Position independent. Because the table is walked with pairs(), the order inside a syllable is not guaranteed; the entries are written so it rarely matters. |
First syllable: vowels1 on [аоуэөүийяюёеъә] | тэр | tʰir̥ | Full vowels. э and и merge to i; ү is u; у is ʊ. |
Later syllables: vowels1_reduced | ахмад | axməd | а о у э ө ү → ə; и й stay i; я ю ё е → jə; ъ → empty. |
| Letter ы | газрын | ɡad͡zrыŋ | ы has table entries but is missing from the character class [аоуэөүийяюёеъә], so it is never replaced. |
Stage 5: Consonant Clusters (consonants2)
| Rule | Example Input | Output | Description |
|---|---|---|---|
| бб, вб → b | вб | b | Run twice to catch overlaps. |
| дд → t͉.t | Будда | bʊt͉.t | Special geminate output. |
| лл, мм, нн, сс → long | алла, русс | aɮː, rʊsː | ɮː mː nː sː. |
| лх → ɬʰ | лхагва | ɬʰaɡw̜ | Voiceless lateral. |
| нб → mp | нбаа | mpaː | Assimilation and loss of lenis voicing. |
| рк, рп, рт, рц, рч → r̥ + stop | цирк | t͡sʰir̥kʰ | Devoiced r before a fortis stop or affricate. |
Stage 6: Single Consonants (consonants1)
| Rule | Example Input | Output | Description |
|---|---|---|---|
gsub(text, '[бвгджзйклмнпрстфхцчшщь]%*?', consonants1) | цас | t͡sʰas | Every remaining Cyrillic consonant is mapped (see mapping table 3). The optional * is consumed. After this step the text is IPA. |
Stage 7: Contextual Rules and Cleanup
| Rule | Example Input | Output | Description |
|---|---|---|---|
Intervocalic t͡ɕ → d͡ʑ, p → β, k → ɣ | ачаа, апаа, акаа | at͡ɕʰaː, apʰaː, akʰaː | Three gsub calls look for bare t͡ɕ, p, k between vowels. By then stage 6 has already written t͡ɕʰ, pʰ, kʰ, so in these tests the rules did not fire and the output keeps its aspirated forms. The symbols d͡ʑ β ɣ do not appear in any output tested. |
n([%sˈ]*)%f[pbmf] → m | дэнбэ | dimp | н before a labial, including across a space. |
n%f[ɡɢkxʃ%s%p%z] → ŋ | монгол, хүн | mɔŋɡəɮ, xuŋ | Before velars, space, punctuation or end of string. |
r%f[%s%p%z] → r̥ | гар | ɡar̥ | Word-final devoicing. |
Protect final [aɔuʊ]i as ⁱ | манай | manai | Temporarily marks the i of a final diphthong so it survives the next step; restored to i at the end. |
Final vowel deletion when #mn.syllables(orig_text) ~= 1 | байна, чи | bain, t͡ɕʰi | Deletes one final [aeiɔɵuʊə] in multi-syllable words. The syllable count uses the original Cyrillic text. A final vowel is only removed if it is a single character; length marks protect long vowels. |
Cleanup: ʲj → j, ⁱ → i, ~ removed | аль | aɮʲ | Final tidy-up. Latin and other unmapped characters pass through untouched. |
Not Implemented
- No stress (code removed), no allophonic voicing, no intervocalic spirantization of stops (the rule is present but inert for Cyrillic input as above; the Russian loanword spirantization rule is commented out), no vowel harmony computation, no lexicon, no phonetic form.
- Russian-style palatalization after sibilants and affricates, and Russian e as e in loanwords, are not handled.
Common Issues & Limitations
Known Transcription Problems
This table shows known issues where the automatic transcription may be incorrect. The System Output and Correct IPA columns are real engine output; where no spelling can produce the correct sound the column describes it in words.
| Input | System Output | Correct IPA | Cause | What to Do |
|---|---|---|---|---|
| газрын | газрын /ɡad͡zrыŋ/ | Same word typed with и: газрин /ɡad͡zriŋ/ (ы should be /i/) | ы is excluded from the vowel pattern in to_IPA, so it passes through as Cyrillic |
Replace ы with и before transcribing; affects all genitive -ын, -ны endings |
| модны | модны /mɔdnы/ | Typed with и: модни /mɔdn/ | Same cause (word-final ы, Stage 0 marker has no effect) | Same workaround |
| ньюс | No transcription (the Lua call errors and the input comes back unchanged) | Typed without the soft sign: нюс /nʲʊs/ | mn.syllables() calls vowelharmony for ью, and a gsub reference fails |
Drop the ь before ю; same for компьютер → компютер /kʰɔmpʰʲʊtʰʲɵr̥/ |
| Герман | Герман /ɡʲɵrməŋ/ | Initial mid e after a palatalized г (described; no spelling gives it, since initial э is merged to /i/: Гэрман /ɡirməŋ/) | е after a listed consonant is rewritten to ө, so a Russian loanword е gives rounded ɵ | Loanwords with е: treat the vowel as e by hand |
| театр | театр /tʰʲɵatʰr̥/ | Mid e then a (described) | Same palatalization rewrite (т + е → тʲө) | Check loanwords; or spell with э: тэатр /tʰiatʰr̥/ |
| Москва | Москва /mɔskʰw̜/ | The final vowel a is pronounced in the Russian loan (described) | Final short vowel deletion applies to every multi-syllable word, including loans | Native Mongolian stems with deleted vowels are expected; verify loans manually |
| Улаанбаатар-хот | Улаанбаатар-хот /ʊɮaːmpaːtʰərxətʰ/ | Transcribe the parts separately: Улаанбаатар /ʊɮaːmpaːtʰər̥/ and хот /xɔtʰ/ | The hyphen is deleted, so хот becomes a non-initial syllable and its vowel is reduced to ə | Put a space instead of a hyphen |
| эмэгтэй | эмэгтэй /iməɡtʰe/ | The diphthong ei kept (described) | Only the i of ai, ɔi, ui, ʊi is protected from final vowel deletion; the i of ei is not, so it is removed | Known limitation; compare манай /manai/, which keeps its i |
| вокзал | вокзал /w̜ɔkʰd͡zəɮ/ | Russian loan with v and a full second vowel (described) | в is always w̜; non-initial vowels always reduce to schwa; there is no loanword mode | Russian loanwords with в, ф, щ may need manual correction |
Related Pronunciation Guides
Mongolian shares the Cyrillic alphabet with Russian, whose help page is listed below:
For technical issues or suggestions, please visit our GitHub repository.