This article has multiple issues. Please help improve it or discuss these issues on the talk page. (Learn how and when to remove these template messages)
|See also mid central vowel|
In linguistics, specifically phonetics and phonology, schwa (//, rarely // or //; sometimes spelled shwa) is the mid central vowel sound (rounded or unrounded) in the middle of the vowel chart, denoted by the IPA symbol 〈ə〉, or another vowel sound close to that position. An example in English is the vowel sound of the 〈a〉 in the word about. Schwa in English is mainly found in unstressed positions, but in some other languages it occurs more frequently as a stressed vowel.
The term schwa was introduced by German linguists in the 19th century from the Hebrew shva (שְׁוָא IPA: [ʃva], classical pronunciation: shəwāʼ [ʃəˑwɒːʔ]), the name of the niqqud sign used to indicate the phoneme.
It was first used in English texts between 1890 and 1895. The symbol 〈ə〉 was used first by Johann Andreas Schmeller for the reduced vowel at the end of the German language name Gabe. Alexander John Ellis, in his Palaeotype alphabet, used it for the similar English sound in but //.
The origin of the symbol 〈ə〉 is an 〈e〉 turned 180 degrees.
In English, schwa is the most common vowel sound. It is a reduced vowel in many unstressed syllables especially if syllabic consonants are not used. Depending on dialect, it may be written using any of the following letters:
- 〈a〉, as in about [əˈbaʊ̯t]
- 〈e〉, as in taken [ˈtʰeɪ̯kən]
- 〈i〉, as in pencil [ˈpʰɛnsəl]
- 〈o〉, as in memory [ˈmɛməɹi]
- 〈u〉, as in supply [səˈpʰlaɪ̯]
- 〈y〉, as in sibyl [ˈsɪbəl]
- unwritten, as in rhythm [ˈɹɪðəm]
Schwa is a very short neutral vowel sound and, like all other vowels, its precise quality varies depending on the adjacent consonants. In most varieties of English, schwa occurs almost exclusively in unstressed syllables. (There is also an open-mid central unrounded vowel or "long schwa", represented as /ɜː/, which occurs in some non-rhotic dialects' stressed syllables, as in bird and alert.)
In New Zealand English, the high front lax vowel (as in the word bit //) has shifted open and back to sound like schwa, and both stressed and unstressed schwas exist. To a certain extent, that is true for South African English as well.
In General American English, schwa and /ɜː/ are the two vowel sounds that can be r-colored (rhotacized); r-colored schwa is used in words with unstressed 〈er〉 syllables, such as dinner. See also stress and vowel reduction in English.
Welsh uses the letter 〈y〉 to represent schwa, which is a phonemic vowel rather than the realisation of an unstressed vowel. The letter 〈y〉 represents schwa in all positions except in final syllables where it represents /ɪ/ or /i/. For example, the word ysbyty ("hospital") is pronounced /əsˈbəti/.
Quite a few languages have a sound similar to schwa. It is similar to a short French unaccented 〈e〉, which is rounded and less central, more like an open-mid or close-mid front rounded vowel. It is almost always unstressed, but Albanian, Bulgarian, Slovene and Afrikaans are some of the languages that allow stressed schwas.
Many Caucasian languages and some Uralic languages (like Komi) also use phonemic schwa, and allow schwas to be stressed. In the Eastern dialects of Catalan, including the standard variety, based in the dialect spoken in and around Barcelona, an unstressed ⟨a⟩ or ⟨e⟩ is pronounced as a schwa (called vocal neutracode: cat promoted to code: ca , 'neutral vowel'). A stressed schwa can occur in the Catalan dialects spoken in the Balearic Islands, as well as in Romanian, as in mătură [ˈməturə] ('broom').
In European and some African dialects of Portuguese, the schwa occurs in many unstressed syllables that end in ⟨e⟩, such as noite ('night'), tarde ('afternoon'), pêssego ('peach'), and pecado ('sin'). However, that is rare in Brazilian Portuguese except in such areas as Curitiba in Paraná.
In Neapolitan, a final, unstressed ⟨a⟩, and unstressed ⟨e⟩ and ⟨o⟩ are pronounced as a schwa: pìzza ('pizza'), semmàna ('week'), purtuàllo ('orange').
Other characters used to represent this sound include ⟨ը⟩ in Armenian, ⟨ă⟩ in Romanian, and ⟨ë⟩ in Albanian. In Bulgarian Cyrillic, the letter ⟨ъ⟩, which has a very different orthographic function in Modern Russian, is used.
In most Sanskrit-based languages, the schwa ⟨अ⟩ is the implied vowel after every consonant and so has no didactic marks. For example, in Hindi, the character 〈 क 〉 is pronounced /kə/ without marking, but 〈 के 〉 is pronounced /ke/ (like "kay") with a marking.
In Albanian, schwa is represented by the letter ⟨ë⟩, which is also one of the letters of the Albanian alphabet, coming right after the letter ⟨e⟩. It can be stressed like in words i ëmbël /i əmbəl/ and ëndërr /əndər/ ('sweet' and 'dream', respectively).
In Armenian, schwa is represented by the letter 〈 ը 〉 (capital 〈 Ը 〉). It is occasionally word-initial but usually word-final, as a form of the definite article. Unwritten schwa sounds are also inserted to split initial consonant clusters; for example, ճնճղուկ (čnčłuk) [tʃʼəntʃʼə'ʁuk] 'sparrow'.
In Catalan, schwa is represented by the letters 〈a〉 or 〈e〉 in unstressed syllables: pare /ˈpaɾə/ ('father'), Barcelona /bəɾsəˈlonə/. In the Balearic Islands, the sound is sometimes also in stressed vowels, pera /ˈpəɾə/ ('pear').
In Dutch, the digraph ⟨ij⟩ in the suffix -lijk [lək], as in waarschijnlijk [ʋaːrˈsxɛinlək] ('probably'), is pronounced as a schwa. If an ⟨e⟩ falls at the ultimate (or penultimate) place before a consonant in Dutch words and is unstressed, it becomes a schwa, as in the verb ending -en (lopen) and the diminutive suffix -tje(s) (tafeltje(s)). The article "een" ('a[n]') is pronounced using the schwa, [ən], while the number "een" ('one') or "één" is pronounced [e:n].
In German, schwa is represented by the letter 〈e〉 and occurs only in unstressed syllables, as in gegessene.
Schwa is not native to Bavarian dialects of German spoken in Southern Germany and Austria. Vowels that are realized as schwa in Standard German change to /-e/, /-ɐ/, or /-ɛ/.
In compound words, like Fernweh, and borrowed terms, like Effekt, unstressed 〈e〉 is not reduced and retains its usual value of /eː/ (if long) or /ɛ/ (if short).
In Kashubian schwa is represented by the letter ⟨ë⟩, it derives from historical short u and i vowels, and thus may alternate with u and i stemming from historical long vowels in different grammatical forms of a given word. It never appears word initially, except for the word ë (and) and its derivates.
In Madurese, an 〈a〉 in some words, usually in non-final position, would be pronounced as the schwa. When writing Madurese in its traditional abugida, Hanacaraka, such words would not be written with a vowel diacritic denoting a schwa. Nowadays, even after the Madurese people have adopted the Latin alphabet, such writing fashion is still used. Examples are:
- Jhabah (/dʒəbəh/) – Javanese, Java Island
- sagara (/sagərə/) – sea, ocean
- lajar (/lədʒər/) – to sail
- Sorbaja (/sorbədʒə/) – Surabaya
- Madura (/madurə/) – Madurese, Madura Island
- Bulan (/bulən/) – Moon
In the Indonesian variant, schwa is always unstressed except for Jakarta-influenced informal Indonesian, whose schwa can be stressed. In final closed syllables in the formal register, the vowel is 〈a〉 (the final syllable is usually the second syllable since most Indonesian root words consist of two syllables). In some cases, the vowel 〈a〉 is pronounced as a stressed schwa (only when the vowel 〈a〉 is located between two consonants in a syllable), but never in formal speech:
- datang ('come'), pronounced [dɑːˈtʌŋ], and often informally written as dateng.
- kental ('viscous'), pronounced [kənˈtʌl].
- hitam ('black'), pronounced [hiˈtʌm], informally written as item.
- dalam ('deep', 'in'), pronounced [dɑːˈlʌm], often written as dalem.
- malam ('night'), pronounced [mʌˈlʌm], informally written as malem.
Indonesian orthography formerly used unmarked ⟨e⟩ only for the schwa sound, and the full vowel /e/ was written ⟨é⟩. Malaysian orthography, on the other hand, formerly indicated the schwa with ⟨ĕ⟩ (called pĕpĕt), and unmarked ⟨e⟩ stood for /e/.
In the 1972 spelling reform that unified Indonesian and Malaysian spelling conventions (Ejaan yang Disempurnakan, regulated by MABBIM), it was agreed to use neither diacritic. There is no longer an orthographic distinction between /ə/ and /e/; both are spelled with an unmarked ⟨e⟩. For example, the word for 'wheeled vehicle' in Indonesia and Malaysia, which was formerly spelled keréta in Indonesia and kĕreta in Malaysia, is now spelled kereta in both countries. This means that the pronunciation of any given letter ⟨e⟩ in both Indonesian and Malaysian variants is not immediately obvious to the learner and must be learned separately. However, in a number of Indonesian dictionaries and lesson books for foreign learners, the notation is preserved to help learners.
In Southern Malaysian pronunciation, which is predominant in common Malaysian media, the final letter represents schwa, and final 〈-ah〉 stands for /a/. The dialect of Kedah in northern Malaysia, however, pronounces final 〈-a〉 as /a/ also. In loanwords, a non-final short /a/ may become schwa in Malay such as Mekah (<Arabic Makkah, Malay pronunciation [ˈməkah]).
In Norwegian, the schwa is often found in the last syllable of definite, masculine nouns, as in mannen Norwegian pronunciation: ['mɑnən] ('the man'), as well as in infinitive verbs like bite Norwegian pronunciation: [ˈbîːtə] ('bite').
In Romanian, schwa is represented by the letter 〈Ă〉, 〈ă〉, which is considered a letter on its own (the second in the Romanian alphabet). It can be stressed in words in which it is the only vowel such as păr /pər/ ('hair' or 'pear tree') or văd /vəd/ ('I see'). Some words which also contain other vowels can have the stress on 〈ă〉: cărțile /ˈkərt͡sile/ ('the books') and odăi /oˈdəj/ ('rooms').
In Serbo-Croatian, schwa is not a phoneme, but it is often colloquially used to pronounce names of consonants. For example, the official name of the letter ⟨p⟩ is pronounced /pe(ː)/, but in everyday speech, it is often called /pə/.
The schwa is denoted in Welsh by the letter 〈y〉. It is a very common letter as y is the definite article with yr being the definite article if the following word starts with a vowel.
Schwa is normally represented in Yiddish by the Hebrew letter 〈ע〉 (Ayin) and, as in German, occurs only in unstressed syllables, as in געפֿילטע פֿיש (gefilte fish) /ɡəˈfɪltə fɪʃ/ ('stuffed fish'). In words derived from Hebrew, which retain their original orthography but undergo significant phonological change, schwa may be represented by another letter, as in רבי (rebe) /ˈrɛbə/ ('rabbi'), or by no letter at all, as in שבת (shabes) [ˈʃa.bəs] ('Shabbat').
In phonology, syncope is the process of deleting unstressed sounds, particularly unstressed vowels. Across languages, schwa vowels are commonly deleted in some instances, such as in Hindi, North American English, French and Modern Hebrew.
Although the Devanagari script is used as a standard to write Modern Hindi, the schwa (/ə/, sometimes written as /ɑ/) implicit in each consonant of the script is "obligatorily deleted" at the end of words and in certain other contexts. The phenomenon has been termed the "schwa deletion rule" of Hindi. One formalization of the rule has been summarized as ə → ∅ /VC_CV. In other words, when a vowel-preceded consonant is followed by a vowel-succeeded consonant, the schwa inherent in the first consonant is deleted. However, the formalization is inexact and incomplete (it sometimes deletes a schwa that exists, and it fails to delete some schwas that it should) and so can yield errors. Schwa deletion is computationally important because it is essential to building text-to-speech software for Hindi.
As a result of schwa syncope, the correct Hindi pronunciation of many words differs from that expected from a literal rendering of Devanagari. For instance, राम is Rām (expected: Rāma), रचना is Rachnā (expected: Rachanā), वेद is Vēd (expected: Vēda) and नमकीन is Namkīn (expected: Namakīna).
Correct schwa deletion is critical also because the same Devanagari letter sequence can sometimes be pronounced two different ways in Hindi depending on the context: failure to delete the appropriate schwas can then change the meaning. For instance, 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. However, it is pronounced dhadak.ne in the first and dhad.kaneṁ in the second.
While native speakers correctly pronounce the sequence 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.
Some forms of American English have the tendency to delete a schwa when it appears in a mid-word syllable that comes after the stressed syllable. Kenstowicz (1994) states, "American English schwa deletes in medial posttonic syllables". He gives as examples words such as sep(a)rate (as an adjective), choc(o)late, cam(e)ra and elab(o)rate (as an adjective), where the schwa (represented by the letters in parentheses) has a tendency to be deleted. Other examples include fam(i)ly (listen), ev(e)ry (listen), and diff(e)rent (listen).
Schwa is deleted in certain positions in French.
- Sobkowiak, Włodzimierz (2004). English Phonetics for Poles (Third ed.). Poznań: Wydawnictwo Poznańskie. p. 131. ISBN 83-7177-252-1.
- Oxford English Dictionary, under "schwa".
- "schwa". Dictionary.com Unabridged. Random House.
- Harper, Douglas. "schwa". Online Etymology Dictionary.
- Rachael-Anne Knight(2012), Phonetics: A course book, Cambridge University Press, p.71.
- Breza, Edward; Treder, Jerzy (1981). Gramatyka kaszubska. Gdańsk: Zrzeszenie Kaszubsko-Pomorskie. p. 16. ISBN 83-00-00102-6.
- Anderson, Deborah; Everson, Michael (2004-06-07). "L2/04-191: Proposal to encode six Indo-Europeanist phonetic characters in the UCS" (PDF).
- Asmah Haji Omar, "The Malay Spelling Reform". Journal of the Simplified Spelling Society (2): 9–13. 1989. Archived from the original on 2010-07-06.
- Larry M. Hyman; Victoria Fromkin; Charles N. Li (1988), Language, speech, and mind (Volume 1988, Part 2), 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...
- 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...
- Monojit Choudhury, Anupam Basu & Sudeshna Sarkar (July 2004), "A Diachronic Approach for Schwa Deletion in Indo Aryan Languages" (PDF), 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...
- Naim R. Tyson; Ila Nagar (2009), "Prosodic rules for schwa-deletion in Hindi text-to-speech synthesis", International Journal of Speech Technology, (12:15–25): 15–25, doi:10.1007/s10772-009-9040-x,
...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...
- Monojit Choudhury & Anupam Basu (July 2004), "A Rule Based Schwa Deletion Algorithm for Hindi" (PDF), 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...
- Kenstowicz, Michael J. (1994), Phonology in generative grammar, Wiley-Blackwell, ISBN 978-1-55786-426-0