InTraSAL: words tier
In InTraSAL, the text of utterances is represented on two tiers: the words tier, which represents the text according to the phonological form of the pronunciation, and can be considered a phonemic transcription, and the English tier, which can be considered a semi-morphological gloss.
"Phonemic" means that person-, environment-, or situation-specific variation will not be reflected in the transcription, i.e. free variation and environmentally-conditioned allophony should be suppressed. So, whether a person pronounces the Bengali word পড়তে 'to read' as [poɽt̪e], [poɹt̪e], or [pot̪ːe], it should always be transcribed poṛte (=poRte).
Furthermore, "transcription" rather than "transliteration" means that the silent letters and other idiosyncrasies of spelling will not be represented here. Rather, the pronunciation alone should guide transcription. So, the Bengali word স্বপ্ন 'dream' should be transcribed or šôpno (=SOpno) as it is pronounced [ʃɔpno], not swapna or sbôpnô.
Transcription of consonants
The transcription of consonants is fairly straightforward, with only a few peculiarities:
- Lower-case t, th, d, and dh are used for dental sounds (দন্ত্য বর্ণ) /t̪ t̪ʰ d̪ d̪ʱ/, whereas upper-case T, Th, D, and Dh are used for "cerebral" sounds (মূর্ধন্য বর্ণ). Cerebral sounds can be pronounced as alveolar /t tʰ d dʱ/ or retroflex /ʈ ʈʰ ɖ ɖʱ/ depending on the speaker, the register, and the phonological environment. Note that Bengali has no "cerebral" nasal or fricative in its phoneme inventory, even though the letters ণ and ষ continue to be called মূর্ধন্য ণ and মূর্ধন্য ষ.
- Lower-case s is used for /s/ as in আস্তে /ast̪e/ aste 'softly' whereas upper-case S is used for /ʃ/ as in আসতে /aʃt̪e/ aSte 'to come'. Some speakers (especially those in West and North Bengal) do not consistently distinguish these sounds.
- Lower-case r is used for /ɹ/ as in করা /kɔɹa/ kOra 'doing' whereas upper-case R is used for /ɽ/ as in কড়া /kɔɽa/ kORa 'intense'. Some speakers (especially those in East Bengal) do not consistently distinguish these sounds.
- Lower-case n is used for ন, ণ, and ঞ /n/ as in আন্টি /anti~anʈi/ anTi 'auntie' whereas upper-case G is used for ঙ and ং /ŋ/, as in আংটি /aŋti~aŋʈi/ aGTi 'ring'. Note that the combination /ŋɡ/ as in বঙ্গ /bɔŋɡo/ bOGgo has both the G and the g.
- Lower-case h is used to represent aspiration of a stop or affricate, while upper-case H is used for the consonant হ [h] on its own. This can help distinguish single aspirated consonant ঝ /dʑʱ/ jh in মাঝে /madʑʱe/ majhe from the sequence of two consonants জহ /dʑh/ jH in রাজহাঁস /ɹadʑhãʃ/ rajHaMS 'swan'.
- Not all speakers distinguish /pʰ/ ph from /f/ f, or /bʱ/ bh from /v/ v, but the contrast can arise especially in unscripted speech, where borrowings are frequent. It is up to the discretion of the transcriber whether to show these distinctions.
| Bengali letter(s) |
Pronunciation(s) in IPA (Khan 2010) |
B-ToBI representation(s) |
Example |
| ক | k | k | কাশ kaS 'wild sugarcane' |
| খ | kʰ, x | kh | খাস khaS 'you eat' |
| গ | ɡ | g | গাস gaS 'you sing' |
| ঘ | ɡʱ | gh | ঘাস ghaS 'grass' |
| ঙ, ং | ŋ | G (ṅ) | বাংলা baGla 'Bengali' |
| চ | tɕ~cɕ~tʃ~ts~c s |
c s |
চাল cal 'rice grains' |
| ছ | tɕʰ~cɕʰ~tʃʰ~tsʰ~cʰ s |
ch s |
ছাল chal 'skin' |
| জ, য | dʑ~ɟʑ~dʒ~dz~ɟ z |
j z |
জাল jal 'net' |
| ঝ | dʑʱ~ɟʑʱ~dʒʱ~dzʱ~ɟʱ z |
jh z |
ঝাল jhal 'spicy' |
| ট | t~ʈ | T (ṭ) | টাক Tak 'bald spot' |
| ঠ | tʰ~ʈʰ | Th (ṭh) | ঠাকুর Thakur 'lord' |
| ড | d~ɖ | D (ḍ) | ডাক Dak 'call' |
| ঢ | dʰ~ɖʱ | Dh (ḍh) | ঢাক Dhak 'cover!' |
| ত | t̪ | t | তাক tak 'shelf' |
| থ | t̪ʰ | th | থাক thak 'let it be' |
| দ | d̪ | d | দাগ dag 'stain' |
| ধ | d̪ʱ | dh | ধাক্কা dhakka 'push' |
| ন, ণ, ঞ | n | n | নাক nak 'nose' |
| প | p~ɸ | p | পাটি paTi 'grass mat' |
| ফ | pʰ f |
ph f |
ফাটি phaTi, faTi 'I burst' |
| ব | b | b | বাটি baTi 'bowl' |
| ভ | bʱ v |
bh v |
ভাটি bhaTi 'kiln' |
| ম | m | m | মাটি maTi 'ground' |
| র | ɹ~ɾ | r | হ্রাস raS 'reduction' |
| ল | l | l | লাশ laS 'dead body' |
| শ, স, ষ | ʃ | S (š) | শ্বাস SaS 'breath' |
| শ, স, ষ, চ, ছ | s | s | সাফ saf 'clean' |
| হ, ঃ | h~V̤ | H | হাঁস HaMS 'duck' |
| ড়, ঢ় | ɽ~ɹ~ɾ | R (ṛ) | বড় bORo 'big' |
Transcription of vowels, diphthongs, and nasalization
The transcription of vowels, diphthongs, and nasalization can differ depending on the availability of special diacritics.
- Lower mid vowels অ্যা /ɛ~æ/ ê (=E) and অ /ɔ/ ô (=O) use upper-case letters.
- Nasalization ঁ /Ṽ/, is represented by upper-case M following the nasalized vowel if ṃ is unavailable.
- Note that nasalization can be transcribed even when nasalization is not perceived in the recording. This is up to the discretion of the transcriber, if he/she wants to convey that the vowel is underlyingly nasalized in the speaker's phonology.
- Letters y w ŷ (=Y) ŵ (=W) represent off-glides of i u e o respectively within diphthongs.
| Bengali letter(s) |
Pronunciation(s) in IPA (following Khan 2010) |
B-ToBI representation(s) |
Example |
| অ | ɔ | ô (=O) | দশ dOS 'ten' |
| আ | a | a | দাস daS 'slave' |
| ই, ঈ | i | i | দিস diS 'give-2i' |
| উ, ঊ | u | u |
দু'শ duSo |
| এ | e | e | দেশ deS 'country' |
| অ্যা, এ | ɛ~æ | ê (E) | দেখ্ dEkh 'look!' |
| ও, অ | o | o | দোষ doS 'blame', 'guilt' |
| অয় | ɔe̯ | ôŷ (=OY) | হয় HOY 'become-3' |
| অও | ɔo̯ | ôŵ (=OW) |
হও HOW |
| আই | aj | ay | পাই pay 'get-1' |
| আয় | ae̯ | aŷ (=aY) |
পায় paY |
| আউ | aw | aw | পাউরুটি pawruTi '(Western) bread' |
| আও | ao̯ | aW (aŵ) | পাও paW 'get-2' |
| ইউ | iw | iw | পিউ piw 'Piu' (name) |
| উই | uj | uy | পুঁইশাক puyMSak 'Malabar spinach' |
| এই | ej | ey | নেই ney 'take-1' |
| এউ | ew | ew | ঢেউ Dhew 'wave' |
| ওই, ঐ, অই | oj | oy | বই boy 'book' |
| ওয় | oe̯ | oY (oŷ) | শোয় SoY 'lie.down-3' |
| ওউ, ঔ, অউ | ow | ow | বউ bow 'bride' |
| অ্যায়, এয় | ɛe̯~æe̯ | êŷ (=EY) | দেয় dEY 'give-3' |
| অ্যাও, এও | ɛo̯~æo̯ | êŵ (=EW) | নেও nEW 'take-2' |
| ঁ | ̃ | ṃ (=M) | হাঁস HaMS 'duck' |
Return to InTraSAL
Now that you know how to transcribe Bengali speech for B-ToBI annotation, you can return to the InTraSAL main page.