Schwa deletion in Indo-Aryan languages

From Wikipedia, the free encyclopedia
Jump to: navigation, search
The IPA symbol for the schwa

The schwa deletion or schwa syncope phenomenon plays a crucial role in Hindi, Bengali, Marathi, Nepali, Urdu, Kashmiri, Punjabi, Gujarati, Maithili and several other Indo-Aryan languages, where schwas implicit in the written scripts of those languages are obligatorily deleted for correct pronunciation.[1][2] Schwa syncope is extremely important in these languages for intelligibility and unaccented speech. It also presents a challenge to non-native speakers and speech synthesis software because the scripts, including Devanagari, do not provide indicators of where schwas should be dropped.[3]

Schwa deletion in Sanskrit[edit]

The schwa is not deleted in Sanskrit, in Vedic or in Pāli unlike in modern Indo Aryan languages.

Schwa deletion in Hindi[edit]

Although the Devanagari script is used as a standard to write modern Hindi, the schwa ('ə') implicit in each consonant of the script is "obligatorily deleted" at the end of words and in certain other contexts, unlike in Sanskrit.[1] This phenomenon has been termed the "schwa syncope rule" or the "schwa deletion rule" of Hindi.[1][3] One formalization of this rule has been summarized as ə → ∅ /VC_CV. In other words, when a schwa-succeeded consonant is followed by a vowel-succeeded consonant, the schwa inherent in the first consonant is deleted.[3][4] However, this formalization is inexact and incomplete (i.e. sometimes deletes a schwa when it shouldn't or, at other times, fails to delete it when it should), and can yield errors. The rule is reported to result in correct predictions on schwa deletion 89% of the time.[4] Schwa deletion is computationally important because it is essential to building text-to-speech software for Hindi.[4][5]

As a result of schwa syncope, the Hindi pronunciation of many words differs from that expected from a literal Sanskrit-style rendering of Devanagari. For instance, राम is Rām (not Rāma), रचना is Rachnā (not Rachanā), वेद is Véd (not Véda) and नमकीन is Namkeen (not Namakeena).[4][5] The name of the script itself is pronounced Devnāgrī (not Devanāgarī).[6]

Correct schwa deletion is also critical because, in some cases, the same Devanagari letter-sequence is pronounced two different ways in Hindi depending on context, and failure to delete the appropriate schwas can change the sense of the word.[7] For instance, the letter sequence 'रक' is pronounced differently in हरकत (har.kat, meaning movement or activity) and सरकना (sarak.na, meaning to slide). Similarly, the sequence धड़कने in दिल धड़कने लगा (the heart started beating) and in दिल की धड़कनें (beats of the heart) is identical prior to the nasalization in the second usage. Yet, it is pronounced dhadak.ne in the first and dhad.kane in the second.[7] While native speakers correctly pronounce the sequences differently in different contexts, non-native speakers and voice-synthesis software can make them "sound very unnatural", making it "extremely difficult for the listener" to grasp the intended meaning.[7]

Schwa deletion in other Indo-Aryan languages[edit]

Different Indo-Aryan languages can differ in how they apply schwa deletion. For instance, medial schwas from Sanskrit-origin words are often retained in Bengali where they are deleted in Hindi.[8] An example of this is रचना/রচনা which is pronounced rachana (/rətʃənaː/) in Sanskrit, rachna (/rətʃnaː/) in Hindi and rochona (/rɔtʃonaː/) in Bengali - while the medial schwa is deleted in Hindi (due to the ə → ∅ / VC_CV rule), it is retained in Bengali.[4] On the other hand, the final schwa in वेद /বেদ is deleted in both Hindi and Bengali (Sanskrit: /veːd̪ə/, Hindi: /veːd̪/, Bengali: /beːd̪/).[4]

Gujarati
Gujarati has a strong schwa deletion phenomenon, affecting both medial and final schwas. From an evolutionary perspective, the final schwas appear to have been lost prior to the medial ones.[2]
Nepali
Owing to its Sanskrit tradition, schwas are rarely deleted in Nepali except for in Hindi or Urdu loanwords. Because of this, conjunctions and ligatures are used more extensively in Nepali than in many other Indo-Aryan languages. For example, the word लाग्छ (Lagchha) contains a ligature between ग and छ, and is pronounced with the schwa at the end.
Maithili
Maithili's schwa deletion is similar to Hindustani, and the ə → ∅ / VC_CV also selectively applies to the language. For instance, हमरो, which with schwas is həməro (meaning, even ours) is correctly pronounced həmro.[9] This is akin to the neighboring Bhojpuri in which हमरा (meaning mine) is pronounced həmrā rather than həmərā due to the deletion of a medial schwa.[10]
Kashmiri
In the Dardic subbranch of Indo-Aryan, Kashmiri similarly demonstrates schwa deletion. For instance, drāksha (द्राक्ष, drākshə) is the Sanskrit word for grape, but the final schwa is dropped in the Kashmiri version, which is dāch.[11]
Punjabi
Punjabi, too, has broad schwa deletion rules: several base word forms (e.g. ਕਾਗ਼ਜ਼, کاغز, kāghəz/paper) drop schwas in the plural form (ਕਾਗ਼ਜ਼ਾੰ, کاغزاں, kāghzāṅ/papers) as well as with instrumental (ਕਾਗ਼ਜ਼ੋੰ, کاغزوں, kāghzōṅ/from the paper) and locative (ਕਾਗ਼ਜ਼ੇ, کاغزے, kāghzé/on the paper) suffixes.[12]
Marathi
In Marathi, the schwa at the end of the Sanskrit stems is retained in some cases, such as for a few tatsama words.[13]

Common transcription and diction errors[edit]

Since Devanagari does not provide indications of where schwas should be deleted, it is common for non-native learners/speakers of Hindi, who are otherwise familiar with Devanagari and Sanskrit, to incorrectly pronounce several words in Hindustani and other modern Indo-Aryan languages.[14] Similarly, systems that automate transliteration from Devanagari to Latin and other scripts by hardcoding implicit schwas in every consonant often make systematic errors. This becomes evident when English words are transliterated into Devanagari by Hindi-speakers and then translated back into English by manual or automated processes that don't account for Hindi's schwa deletion rules. For instance, English is transcoded by Hindi speakers into इंगलिश (rather than इंग्लिश्) which may be transcoded back to Ingalisha by automated systems, whereas schwa deletion would result in इंगलिश being correctly pronounced as Inglish by native Hindi-speakers.[15]

Some common examples of errors are shown below -

Word in Devnagri and meaning Correct pronunciation with schwa syncope Incorrect pronunciation(s) Comments
लपट (flame) ləpəṭ ləpəṭə The final schwa must be deleted [16]
लपटें (flames) ləpṭeṅ ləpəṭeṅ The medial schwa, ləpəṭ, which was retained in लपट, must be deleted in लपटें [16]
समझ (understanding) səməjh səməjhə The final schwa must be deleted [17]
समझा (understood, verb masc.) səmjhā səməjhā The medial vowel also needs to be deleted here, which it did not need to be in समझ [17]
भारत (India) bhārət bhārətə Final schwa must be deleted
भारतीय (Indian) bhārtīy bhārətīyə Both the medial and final schwa should be deleted, although the final schwa is sometimes faintly pronounced due to the 'y' glide; when pronounced without this, the word sounds close to 'bhārtī'
देवनागरी (Devanagari, the script) devnāgrī devənāgərī Two medial schwas (after व and after ग) should be deleted
इंगलिश (English, the language) inglish ingəlishə Medial and final schwas (after ग and after श) should be deleted
विमला (Vimla, a proper name) vimlā viməlā Medial schwa should be deleted [18]
सुलोचना (Sulochna, a proper name) sulochnā sulochənā Medial schwa should be deleted [18]

Vowel nasalization[edit]

With some words that contain /n/ or /m/ consonants separated from succeeding consonants by schwas, the schwa deletion process has the effect of nasalizing any preceding vowels.[19] Some examples in Hindustani include -

  • sən.kī (सनकी, سنکی, whimsical), in which a deleted schwa that is pronounced in the root word sənək (सनक, سنک, whimsy) converts the first medial schwa into a nasalized vowel
  • chəm.kīlā (चमकीला, چمکیلا, shiny), in which a deleted schwa that is pronounced in the root word chəmək (चमक, چمک, shine) converts the first medial schwa into a nasalized vowel

See also[edit]

References[edit]

  1. ^ a b c Larry M. Hyman, Victoria Fromkin, Charles N. Li (1988 (Volume 1988, Part 2)), Language, speech, and mind, Taylor & Francis, ISBN 0-415-00311-3, "... The implicit /a/ is not read when the symbol appears in word-final position or in certain other contexts where it is obligatorily deleted (via the so-called schwa-deletion rule which plays a crucial role in Hindi word phonology ..." 
  2. ^ a b Indian linguistics, Volume 37, Linguistic Society of India, 1976, "... the history of the schwa deletion rule in Gujarati has been examined. The historical perspective brings out the fact that schwa deletion is not an isolated phenomenon; the loss of final -a has preceded the loss of medial -a-; ..." 
  3. ^ a b c Tej K. Bhatia (1987), A history of the Hindi grammatical tradition: Hindi-Hindustani grammar, grammarians, history and problems, BRILL, ISBN 90-04-07924-6, "... Hindi literature fails as a reliable indicator of the actual pronunciation because it is written in the Devanagari script ... the schwa syncope rule which operates in Hindi ..." 
  4. ^ a b c d e f Monojit Choudhury, Anupam Basu and Sudeshna Sarkar (July 2004), "A Diachronic Approach for Schwa Deletion in Indo Aryan Languages", Proceedings of the Workshop of the ACL Special Interest Group on Computational Phonology (SIGPHON) (Association for Computations Linguistics), "... schwa deletion is an important issue for grapheme-to-phoneme conversion of IAL, which in turn is required for a good Text-to-Speech synthesizer ... Sanskrit rəcəna, Hindi rəcna, Bengali rɔcona ..." 
  5. ^ a b Naim R. Tyson, Ila Nagar (2009 (12:15–25)), "Prosodic rules for schwa-deletion in hindi text-to-speech synthesis", International Journal of Speech Technology, "... Without the appropriate deletion of schwas, any speech output would sound unnatural. Since the orthographical representation of Devanagari gives little indication of deletion sites, modern TTS systems for Hindi implemented schwa deletion rules based on the segmental context where schwa appears ..." 
  6. ^ Nazir Ali Jairazbhoy, The rāgs of North Indian music: their structure and evolution, Popular Prakashan, 1995, p. xi, ISBN 978-81-7154-395-3, "... The Devnāgrī (Devanāgarī) script is syllabic and all consonants carry the inherent vowel a unless otherwise indicated. The principal difference between modern Hindi and the classical Sanskrit forms is the omission in Hindi of this inherent a when in final position (e.g. rāga in Sanskrit and rāg in Hindi) and frequently in medial position (e.g. Māravā in Sanskrit and Mārvā in Hindi)." 
  7. ^ a b c Monojit Choudhury and Anupam Basu (July 2004), "A Rule Based Schwa Deletion Algorithm for Hindi", Proceedings of the International Conference On Knowledge-Based Computer Systems, "... Without any schwa deletion, not only the two words will sound very unnatural, but it will also be extremely difficult for the listener to distinguish between the two, the only difference being nasalization of the e at the end of the former. However, a native speaker would pronounce the former as dha.D-kan-eM and the later as dha.Dak-ne, which are clearly distinguishable ..." 
  8. ^ Anupam Basu, Udaya Narayana Singh, Proceedings of the Second Symposium on Indian Morphology, Phonology & Language Engineering: Simple'05, February 5th-7th, 2005, Indian Institute of Technology, Kharagpur, Central Institute of Indian Languages, 2005, ISBN 978-81-7342-137-2, "... The compound words derived from native words of Bengali show greater tendency towards {a} deletion than those derived from Sanskrit ..." 
  9. ^ George Cardona, The Indo-Aryan languages, Psychology Press, 2003, ISBN 978-0-7007-1130-7, "... The two morphophonemic alternations that are very productive and regular in Maithili are schwa deletion and replacement of a by schwa. (a) Schwa deletion: ... VCəCV → VC0CV ... Schwa deletion in Maithili occurs ..." 
  10. ^ Manindra K. Verma, Karavannur Puthanvettil Mohanan, Experiencer subjects in South Asian languages, Center for the Study of Language (CSLI), 1990, ISBN 978-0-937073-60-5, "... The paradigm in Bhojpuri ... hamaar in isolaton is genitive and has an oblique form in -aa, which according to the general principle of vowel attenuation (schwa deletion) in this language yields the form hamraa before postpositions ..." 
  11. ^ William Jackson Elmslie, A vocabulary of the Kashmiri language: in two parts : Kashmiri-English, and English-Kashmiri, Church missionary house, 1872, "... dach, nm grape ..." 
  12. ^ Tej K. Bhatia, Punjabi: a cognitive-descriptive grammar, Psychology Press, 1993, ISBN 978-0-415-00320-9, "... nazar 'glance' - nazar të - nazrë. Postposition incorporation is quite productive. The stem-final schwa undergoes deletion before the vocalic postpositional elements ..." 
  13. ^ Rajeshwari Pandharipande (1997). Marathi. Psychology Press. p. 571. ISBN 978-0-415-00319-3. Retrieved 26 December 2012. 
  14. ^ Florian Coulmas, The writing systems of the world, Wiley-Blackwell, 1991, ISBN 978-0-631-18028-9, "... in the Devanagari script, the schwa vowel is not indicated in consonant-initial syllables. This is a well-known problem for those learning to read Hindi ..." 
  15. ^ An example is Google's automated transliteration of: J. P. Singh Ahluwalia, Mohan Singh, Jepī Raipiḍa korasa ṭu sapokana Iṅgalisha: including Ingalisha pronouncing dikashanarī, Jaypee Publications, 2008, "... sapokana Iṅgalisha ..." 
  16. ^ a b Rajendra Singh, Rama Kant Agnihotri, Hindi morphology: a word-based description, Motilal Banarsidass Publ., 1997, ISBN 978-81-208-1446-2, "... For a pair of words eg ləpəṭ ~ ləpəṭen 'flame', one has to apply the following phonomorphological interface rules on the abstract ..." 
  17. ^ a b Colin P. Masica, The Indo-Aryan Languages, Cambridge University Press, 1993, ISBN 978-0-521-29944-2, "... on the suffixation: H. samajhna 'to understand' > samjha 'understood'. This too produces clusters, albeit unstable ones. As noted in Chapter 6, the most recent treatment (synchronic) of this "schwa-deletion" phenomenon in Hindi ..." 
  18. ^ a b Manjari Ohala, The schwa-deletion rule in Hindi: phonetic and non-phonetic determinants of rule application, Indiana University Linguistics Club, 1974, "... [sulochna] ~ [sulochəna] ... schwa is conditionally deleted ..." 
  19. ^ G.C. Narang, Donald A. Becker, "Aspiration and nasalization in the generative phonology of Hindustani", Language, 47, 3, 646-667, September 1971 (Linguistic Society of America), JSTOR 412381, "... nasalized vowels are derived from underlying sequences of vowel plus nasal consonant ..."