Hindustani phonology
This article is about the phonology of the Hindi and Urdu language standards, which are nearly the same thing.
Note: Information specific to Urdu has not yet been integrated into this article.
There are approximately 11 vowels and 35 consonants in Standard Hindī. They are shown below:
Vowels
The vowels of Hindi with their word-initial devanagari symbol, diacritical mark with the consonant प (p), pronunciation (of the vowel alone and the vowel following /p/) in IPA, equivalent in IAST and (approximate) equivalents in British English are listed below:[1]
![The vowel phonemes of Hindi](http://upload.wikimedia.org/wikipedia/commons/3/3d/Hindi_vowel_chart.png)
Alphabet | Diacritical mark with प | Pronunciation | Pronunciation with /p/ | IAST equiv. | English equivalent |
---|---|---|---|---|---|
अ | प | /ə/ | /pə/ | a | short or long Schwa: as the a in above or ago |
आ | पा | /ɑː/ | /pɑː/ | ā | long Open back unrounded vowel: as the a in father |
इ | पि | /ɪ/ | /pɪ/ | i | short close front unrounded vowel: as i in bit |
ई | पी | /iː/ | /piː/ | ī | long close front unrounded vowel: as i in machine |
उ | पु | /ʊ/ | /pʊ/ | u | short close back rounded vowel: as u in put |
ऊ | पू | /uː/ | /puː/ | ū | long close back rounded vowel: as oo in school |
ए | पे | /eː/ | /peː/ | e | long close-mid front unrounded vowel: as a in game (not a diphthong) |
ऐ | पै | /ɛː/ | /pɛː/ | ai | long open-mid front unrounded vowel: as e in bed, but longer |
ओ | पो | /οː/ | /pοː/ | o | long close-mid back rounded vowel: as o in tone (not a diphthong) |
औ | पौ | /ɔː/ | /pɔː/ | au | long open-mid back rounded vowel: as au in caught |
Additional notes on vowels
- The short open-mid front unrounded vowel (/ɛ/: as e in get) can occur as a conditioned allophone of schwa. Thus, the pronunciation of the vowel अ occurs in two forms. When this vowel is followed by word-middle /h/, or it surrounds word-middle /h/, or is followed by word-ending /h/, it changes allophonically to short /ɛ/. In all other cases it is the mid central vowel schwa. Thus, the following words शहर, रहना, कह are pronounced as /ʃɛhɛr/, /rɛhnɑː/ and /kɛh/ and not as /ʃəhər/, /rəhənɑː/ and /kəh/. It also occurs in loanwords from English, where it is sometimes accorded a new vowel symbol of ऍ (candra: पॅ). e.g., pen: पॅन.
- The short open back rounded vowel (/ɒ/: as o in hot in Received Pronunciation), does not exist in Hindi at all, other than for English loanwords. In orthography, a new symbol has been invented for it: ऑ (पॉ). If included in Hindi phonology, it brings the number of phonemic vowels to 11.
- There are some additional vowels traditionally listed in the Hindi alphabet. They are
- ऋ (originally in Sanskrit a vowel-like syllabic retroflex approximant), pronounced in modern Hindi as /ri/, used only in Sanskrit loanwords (पृ).
- अं (called anusvāra), pronounced as /əŋ/. Its diacritic (the dot above) is used for a variety of purposes, consisting of vocalic nasalization, and the nasal consonants /n/, /m/, /ɳ/, /ɲ/, /ŋ/ before another consonant. This leads to alternative Hindi spellings for the some words, e.g., the word Hindi itself has two spellings: हिन्दी and हिंदी.
- अः (called visarga), pronounced as /əh/. Used only in Sanskrit loanwords (पः).
- The diacritic अँ (called candrabindu), not listed in the alphabet, is used interchangeably with the anusvāra to indicate nasalization of the vowel (पँ).
- If a lone consonant needs to be written without any following vowel, it is given a halanta/virāma diacritic below (प्).
- There is less lip-rounding than in English in the long open-mid back rounded vowel (/ɔː/: as au in caught). The vowel /ɑː/ in Hindi is more central and less back than in English, like /ä/.
- All vowels in Hindi, short or long, can be nasalized, except ऑ. There is disagreement in the scholarly literature over the extent to which nasalization is phonemic.[2] Minimal pairs exist for many vowels: e.g. कहा, /kəhɑː/, "he said", vs. कहाँ, /kəhɑ̃ː/, "where".
- In Sanskrit and in some (eastern) dialects of Hindi (as well as in a few words in Standard Hindi), the vowel ऐ is pronounced as a diphthong /əi̯/ or /ai̯/ rather than /ɛː/. Similarly, the vowel औ is pronounced in some words as the diphthong /əu̯/ or /au̯/ rather than /ɔː/. Other than these, Hindi does not have true diphthongs—two vowels might occur sequentially but then they are pronounced as two syllables (a glide might come in between while speaking).
- The vowel ऐ is used to represent the English vowel /æ/ in loan-words. In these cases, many Hindi speakers pronounce the vowel as [æː] instead of [ɛː], adding an additional vowel phoneme to the Hindi inventory. For example, बैट "cricket bat" may be pronounced [bæːʈ] rather than the expected [bɛːʈ] [3]
- In the Devanagari script used for Sanskrit, whenever a consonant in a word-ending position is without a virāma (ie, freely standing in the orthography: प as opposed to प्), the short neutral vowel schwa (/ə/) is automatically associated with it—this is of course true for the consonant when in any other position in the word. However, in Hindi, even if the word-ending consonant is written without a virāma, the associated schwa is almost never pronounced. The schwa (/ə/) may be pronounced very short only if the absence of schwa would otherwise make the pronunciation of the word very difficult — such a situation arises when there is a consonantal cluster at the end of the word. Thus, for phonological purposes, a word-ending grapheme without a halant or any other vowel-diacritic must be treated as consonant ending. The schwa in Hindi is usually dropped (syncopated) in khariboli even at certain instances in word-middle positions, where the orthography would otherwise dictate so. e.g., रुकना (to stay) is normally pronounced as /ruknɑː/, while according to the orthography, it should have been /rukənɑː/. (Tiwari, [1966] 2004). Schwa is never syncopated in the first syllable, but often syncoped in the second or the penultimate syllable — this of course reduces the number of syllables in the word. The syncopation of schwa is not phonemically contrastive.
The dropping of schwa at the end in Hindi (for Sanskrit loanwords)is illustrated in the examples given below:
Hindi/Sanskrit word | Usual transliteration | Sanskrit pronunciation | Hindi pronunciation | English pronunciation |
---|---|---|---|---|
िशव—a deity | Shiva | /ɕiʋə/ | /ʃiʋ/ | /ʃiːvə/ |
वरुण—a deity | Varuna | /ʋəruɳə/ | /ʋəruɳ/ | /vʌɹuːnə/ |
वेद—a scripture | Veda | /ʋeːd̪ə/ | /ʋeːd̪/ | /veɪdə/ |
राम—a hero | Rama or Rāma | /rɑːmə/ | /rɑːm/ | /ɹɑːmə/ |
कामसूत्र—a love manual | Kamasutra | /kɑːməsuːt̪rə/ | /kɑːmː suːt̪r̩/ or /kɑːm suːt̪rə/ | /kɑːmə suːtɹə/ |
अशोक—an emperor | Ashoka or Asoka | /əɕoːkə/ | /əʃoːk/ | /ʌsəʊkə/ |
The Handbook of the International Phonetic Association also describes the near-close near-front unrounded vowel (/ɪ/) the near-close near-back rounded vowel (/ʊ/) as occurring in Hindi phonology. They respectively occur as free allophones of short /i/ and /u/.
Consonants
Hindi has a large consonant system, with about 38 distinct consonant phonemes. An exact number cannot be given, since the regional varieties of Hindi differ in the details of their consonant repertoire. To what extent certain sounds that appear only in foreign words should be considered part of Standard Hindi is also a matter of debate. The traditional core of the consonant system, inherited from Sanskrit, consists of a matrix of 20 plosives, 5 nasals, and 8 sonorants and fricatives. The system is filled out by 5 sounds that originated in Persian, but are now considered Hindi sounds. The table below shows the phonology of the Hindi consonants.[4] Note that all nasals, trills, flaps, approximants and lateral approximants in Hindi are regarded as voiced consonants, and that many linguists also call the aspirated voiced plosives as breathy voice or murmur stops.
Bilabial | Labio- dental |
Dental | Alveolar | Retroflex | Post-alveolar/ Palatal |
Velar | Uvular | Glottal | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Plosives (unaspirated) Plosives (aspirated) |
p pʰ |
b bʱ |
t̪ t̪ʰ |
d̪ d̪ʱ |
ʈ ʈʰ |
ɖ ɖʱ |
k kʰ |
g gʱ |
q |
|||||||||
Affricates | ʧ or cɕ ʧʰ or cɕʰ |
ʤ or ɟʝ ʤʱ or ɟʝʱ |
||||||||||||||||
Nasals | m | n | (ɳ) | (ɲ) | (ŋ) | |||||||||||||
Fricatives | f | (v) | x | ɣ | (χ) | (ʁ) | h | ɦ | ||||||||||
Sibilants | s | z | ʂ | ʃ | ||||||||||||||
Flaps | ɾ | ɽ ɽʱ |
||||||||||||||||
Trills | (r) | |||||||||||||||||
Approximants | ʋ | j | ||||||||||||||||
Lateral approximant |
l |
The 25 stop consonants occur in five groups, with each group sharing the same position of articulation. These positions in their traditional order are: velar, retroflex, palatal, dental, and bilabial. In each position, there are five varieties of consonant, with four oral stops and one nasal stop. An oral stop may be voiced, aspirated, both, or neither. This four-way opposition is the hardest aspect of Hindi pronunciation for a speaker of English. The table below shows the traditional listing of the Hindi consonants (in the Devanagari script) with the (nearest) equivalents in English/Spanish. Each consonant shown below is deemed to be followed by the neutral vowel schwa (/ə/), and is named in the table as such. The Roman script equivalent that is normally used to transcribe Hindi in casual transliteration is also given in the second line.
Plosives | |||||
Unaspirated Voiceless |
Aspirated Voiceless |
Unaspirated Voiced |
Aspirated Voiced |
Nasals | |
---|---|---|---|---|---|
Velar | क /kə/ k; English: scald |
ख /kʰə/ kh; English called |
ग /gə/ g; English: game |
घ /gʱə/ gh; Aspirated/murmured /g/, somewhat similar to doghouse |
ङ /ŋə/ n; English: ring |
Palatal | च /cɕə / or / tʃə/ ch; English butcher |
छ /cɕʰə / or /tʃʰə/ chh; English: chat |
ज /ɟʝə / or / dʒə/ j; English: jam |
झ /ɟʝʱə / or / dʒʱə/ jh; Aspirated/murmured /ɟʝ/, somewhat similar to hedgehog |
ञ /ɲə/ n; English: hinge |
Retroflex | ट /ʈə/ t; like "t" but with the tongue tip curled back |
ठ /ʈʰə/ th; Aspirated /ʈ/ |
ड /ɖə/ d; like "d" but with the tongue tip curled back |
ढ /ɖʱə/ dh; Aspirated/murmured /ɖ/ |
ण /ɳə/ n; like "n" but with the tongue tip curled back |
Apico-Dental | त /t̪ə/ t; Spanish: tomate |
थ /t̪ʰə/ th; Aspirated /t̪/ |
द /d̪ə/ d; Spanish: donde |
ध /d̪ʱə/ dh; Aspirated/murmured /d̪/ |
न /nə/ n; English: name |
Labial | प /pə/ p; English: spin |
फ /pʰə/ ph; English pin |
ब /bə/ b; English: bone |
भ /bʱə/ bh; Aspirated/murmured /b/, somewhat similar to clubhouse |
म /mə/ m; English: mine |
Non-Plosives/Sonorants | ||||
Palatal | Retroflex | Dental/ Alveolar |
Velar/ Glottal | |
---|---|---|---|---|
Approximant | य /jə/ y; English: you |
र /ɾə/ or /rə/ r; Scottish English: trip |
ल /lə/ l; English: love |
व /ʋə/ or /və/ v; between English "w" and "v" |
Sibilant/ Fricative |
श /ʃə/ sh; English: ship |
ष /ʂə/ sh; Retroflex /ʃ/ |
स /sə/ s; English: same |
ह /ɦə / or / hə/ h; English: behind |
At the end of the traditional table of alphabets, three consonantal clusters are also added: क्ष /kʃə/ (in Hindi), त्र /t̪rə/ and ज्ञ /gjə/ (pronunciation given for Hindi). Other than these, sounds borrowed from the other languages like Persian and Arabic are written with a dot (bindu or nukta) beneath the nearest approximate alphabet. They are not included in the traditional listing. Many native Hindi speakers do not pronounce these sounds (except /ɽ / and / ɽʱ/) and replace them instead with the nearest equivalents, as shown in column 4 in the table below. These are:
Extra sounds | |||
Symbol | IPA Pronunciation and name | Equivalent in other languages | Often replaced with: |
---|---|---|---|
क़ | /qə/ voiceless uvular plosive | Arabic: Qur'an | /k/ |
ख़ | /xə/ voiceless velar fricative | German: doch | /kʰ/ |
ग़ | /ɣə/ voiced velar fricative | Persian: Mughal | /g/ |
ज़ | /zə/ voiced alveolar fricative | English: zoo | /ɟ / or / dʒ/ |
य़ | /ʒ/ voiced postalveolar fricative | English: Measure | /dʒ/ |
ड़ | /ɽə/ unaspirated retroflex flap | Similar to English butter pronounced laxly | |
ढ़ | /ɽʱə/ aspirated retroflex flap | <none> | |
फ़ | /fə/ voiceless labiodental fricative | English: fun | /pʰ/ |
ड़ /ɽə/ and ढ़ /ɽʱə/ are not of Persian/Arabic origin, but they are allophonic variants of simple voiced retroflex stops of Sanskrit.
Additional notes on the consonants
Some additional features of Hindi consonant system are given here, as well as some useful tips to those whose native language is English but are interested in learning Hindi language.
- The distinction between the aspirated and the unaspirated consonants is really very strong, not only in Hindi, but also in Sanskrit and many other languages of India.
- The distinction between the dental plosives and the retroflex plosives is also very stark in all Indo-Aryan and Dravidian languages.
- Aspiration is actually a puff of breath that may follow a plosive consonant. English speakers should try to pronounce the voiceless aspirates by speaking these words in a quick but clear liaison-like fashion:
- "take him", "get him", "ditch him", "slap him".
- One could also try speaking the words "kite", "take", "chip" and "pat" with a greater-than-usual puff of breath after the first consonant. The corrsponding unaspirated plosives must be pronounced with no significant puff of breath at all.
- No nasal consonant except /m/ and /n/ can start a word in Hindi. All the other nasal consonants in modern Hindi are conditioned allophones of /n/.
- Tiwari ([1966] 2004) lists the following additional phonemes in Hindi: the aspirates /nʰ/, /mʰ/ and /lʰ/. They are phonemically contrastive to the corresponding unaspirated consonants, but orthographically appear as consonant clusters न्ह, म्ह, ल्ह.
- The retroflex flaps cannot start a word. They did not exist as such in Sanskrit—they have sprung up as the allophonic flap variants of Sanskrit's simple voiced retroflex plosives. The other Indo-Aryan languages and Dravidian languages tend to use retroflex flaps in their vocabulary even more frequently than Hindi does. Originally, they were purely conditioned allophones in Hindi, coming if and only if the consonants /ɖ/ and /ɖʰ/ came in word-middle positions when surrounded by two vowels or word-ending positions when preceded by a vowel, and never in the word-beginning position or in consonantal clusters (where the plosives would come). But the inclusion of English words, whose alveolar plosives become retroflex in Hindi , have made at least the unaspirated retroflex flap /ɽ/ a separate phoneme. The grapheme ण /ɳ/, when followed by a vowel, is usually flapped.
- Hindi /t̪/ and /d̪/ are (apico-) dental, somewhat similar to Spanish or Italian.
- The retroflex consonants are pronounced by curling the tongue such that its tip touches the roof of the mouth. Much though has been listed that Hindi (being an Indo-Aryan language) has retroflex plosives, certain linguists (who are also the native speakers of Hindi) disagree. According to Tiwari ([1966] 2004), the so-called retroflex plosives in Hindi are at the most (apico-) post-alveolar or pre-palatal, and sometimes even alveolar. The retroflex flaps are pronounced in a similar way, by bringing the tongue's tip to the roof of the mouth or pre-palatal region and giving it a sharp flap downwards. Hence, the so-called retroflex plosives are not phonemically contrastive with the alveolar stops that come through English loanwords, but rather with the dental plosives.
- The palatal affricates of Hindi do not have as much a sharp frictional release as in English. They have more of a plosive component. It is not very clear whether they are clearly palatal or postalveolar (as in English). Tiwari ([1966] 2004) classifies them as palatal, and certainly as affricates and not pure plosives, that Sanskrit used to have. He has also called the sibilant श as a voiceless palatal fricative /ç/, rather than postalveolar. However, Hock (1991) prefers to say that this particular sibilant is phonetically the same, whether articulated from the postalveolar region or from the palatal region.
- The letter व corresponds to /ʋ/, but may also be realized by /v/ or /w/ [5][6]. /ʋ/ is very close to /v/, but does not have a friction or buzzing sound associated with it. Hindi makes no phonemic distinction between these three sounds.
- ष represents the voiceless retroflex fricative /ʂ/ in Sanskrit, but in modern Hindi is usually realized as /ʃ/, and is thus indistinguishable from श.
- It is doubtful whether Hindi has a voiced glottal fricative /ɦ/ (for the alphabet ह) or its unvoiced counterpart. The difference between them, unlike in Sanskrit, is not phonemically contrastive.
- The five new consonants borrowed from Perso-Arabic, /q/, /x/, /ɤ/, /z/, /ʒ/ and /f/, are only peripheral phonemes in Hindi. Most native Hindi-speakers cannot produce them, and substitute /k/, /kʰ/, /g/, /ɟʝ/, /ɟʝ/ and /pʰ/ respectively for these consonants. Because the phonemes /z/ and /f/ come in English loanwords, people of the new generation well-educated in English usually pronounce them correctly, but still use substitutes for the other four. In the Urdu style, all six are pronounced correctly.
- The consonant cluster ज्ञ, found in Sanskrit-derived words, is pronounced /gjə/ in Hindi rather than the expected /dʒɲə/, thus becoming indistinguishable from ग्य. Thus, for example, the man's name ज्ञानेन्द्र is pronounced /gjɑːnend̪̪rə/ instead of /dʒɲɑːnend̪̪rə/. The vowel following ज्ञ is often nasalized, reflecting the original sound of the cluster.[7]
Supra-segmental features
Hindi has a stress accent, but it is not so important as in English. Usually in a multisyllabic Hindi word, the stress falls on the last syllable if all the syllables are equally heavy or equally light. (A light syllable is closed by a short vowel a, i, u, while a medium syllable is closed by a long vowel or diphthong ā, e, ī, o, ū, au, ai or by two consonants, and a heavy syllable is closed by both a long vowel/diphthong and two consonants.) If the word contains a mixture or heavy and light syllables, the stress falls automatically on the penultimate heaviest syllable. (Cf. McGregor, pp. xx-xxi.) Content words in Hindi normally begin on a low pitch, followed by a rise in pitch.[8][9] Strictly speaking, Hindi, like most other Indian languages, is rather a syllable timed language. The schwa /ə/ has a strong tendency to vanish into nothing (syncopated) if its syllable is unaccented. Also note that in written Hindi, many words end in short /u/ or short /i/, but in speech they are often converted to ending in long /uː/ or long /iː/, respectively.
- ^ Ohala, p. 102
- ^ Masica, Colin P (1993). The Indo-Aryan Languages (First paperback edition ed.). Cambridge University Press. p. 117. ISBN 0521299446.
{{cite book}}
:|edition=
has extra text (help) - ^ Orala, p. 102
- ^ Ohala, p. 100
- ^ Tiwari [1966] 2004
- ^ Ohala
- ^ McGregor, R S (1995). Outline of Hindi Grammar (Third edition ed.). New York: Oxford University Press. pp. xxix. ISBN 0198700083.
{{cite book}}
:|edition=
has extra text (help) - ^ http://www.und.nodak.edu/dept/linguistics/theses/2001Dyrud.PDF Dyrud, Lars O. (2001) Hindi-Urdu: Stress Accent or Non-Stress Accent? (University of North Dakota, master's thesis)
- ^ http://www.speech.sri.com/people/rao/papers/icslp96_wbhyp.pdf Ramana Rao, G.V. and Srichand, J. (1996) Word Boundary Detection Using Pitch Variations. (IIT Madras, Dept. of Computer Science and Engineering)