Non-native pronunciations of English

Non-native pronunciations of English result from the common linguistic phenomenon in which non-native users of any language tend to carry the intonation, phonological processes and pronunciation rules from their mother tongue into their English speech. They may also create innovative pronunciations for English sounds not found in the speaker's first language.

Overview

The speech of non-native English speakers may exhibit pronunciation characteristics that result from such speakers imperfectly learning the pronunciation of English, either by transferring the phonological rules from their mother tongue into their English speech ("interference") or through implementing strategies similar to those used in primary language acquisition.^[1] They may also create innovative pronunciations for English sounds not found in the speaker's first language.^[1]

The age at which speakers begin to immerse themselves into a language (such as English) is linked to the degree in which native speakers are able to detect a non-native accent; the exact nature of the link is disputed amongst scholars and may be affected by "neurological plasticity, cognitive development, motivation, psychosocial states, formal instruction, language learning aptitude", and the usage of their first (L1) and second (L2) languages.^[2]

English is unusual in that speakers rarely produce an audible release between consonant clusters and often overlap constriction times. Speaking English with a timing pattern that is dramatically different may lead to speech that is difficult to understand.^[3]

More transparently, differing phonological distinctions between a speaker's first language and English create a tendency to neutralize such distinctions in English,^[4] and differences in the inventory or distribution of sounds may cause substitutions of native sounds in the place of difficult English sounds and/or simple deletion.^[5] This is more common when the distinction is subtle between English sounds or between a sound of English and of a speaker's primary language. While there is no evidence to suggest that a simple absence of a sound or sequence in one language's phonological inventory makes it difficult to learn,^[6] several theoretical models have presumed that non-native speech perceptions reflect both the abstract phonological properties and phonetic details of the native language.^[7]

Such characteristics may be transmitted to the children of bilinguals, who will then exhibit a number of the same characteristics even if they are monolingual.^[8]

Examples

Arabic

• Speakers tend to speak with a rhotic accent and pronounce /r/ as [ɾ] or [r].^[9]

Brazilian Portuguese

Various pronunciation mistakes are bound to happen among Brazilian L2 speakers of English, among which:^[10]

Pronunciation of vowels

• Confusion of /ɪ/ and /iː/, usually realized as [i], and of /ʊ/ and /uː/, usually realized as [u].
• Especially in a British context, confusion of /əʊ/ and /ɒ/. The Brazilian /ɔ/ is equivalent to RP English /ɒ/, and English orthography rarely makes a clear demarcation between the phonemes, thus cold (ideally [ˈkɜʊ̯ɫd]) might be homophone with called /ˈkɔːld/. The North American equivalent of British /əʊ/, /oʊ/, may be easier to perceive as it closely resembles the Portuguese diphthong [ow]. Speakers may also have trouble distinguishing between schwa and /ʌ/.
• In a British context, the diphthong /əʊ/ might also be pronounced as the Portuguese diphthong eu, [ew].
• Persistent preference for /æ/ over /ɑː/ (even if the target pronunciation is England's prestige accent), and use of /æ/ within the IPA [ɛ] space (Portuguese /ɛ/ is often [æ], what makes it even more due to confusion in production and perception), so that can't, even in RP, might sound like an American pronunciation of Kent. Some might even go as far as having [le̞st] instead of /læst ~ lɑːst/ for last

Pronunciation of consonants

• Difficulty with dental fricatives /θ/ and /ð/. These may be instead fronted [f v], stopped [t̪ d̪] or hissed [s̻ z̻].
• Speakers may pronounce word-initial r as a guttural ar pronunciations or a trill). These often sound to English speakers as /h/, leading to confusion between ray and hay, red and head, height and right, etc.
• Neutralization of coda /m n ŋ/, giving preference to a multitude of nasal vowels (often forming random diphthongs with [j̃ w̃ ɰ̃], or also randomly losing them, so that sent and saint, and song and sown, are homophonous) originating from their deletion. Vowels are also often strongly nasalized when stressed and succeeded by a nasal consonant, even if said consonant starts a full syllable after it.
• Fluctuation of the levels of aspiration of voiceless stops /p t k/, that might sound like /b d g/.
• Loss of contrast between coronal stops /t d/ and post-alveolar affricates /tʃ dʒ/ due to palatalization of the earlier, before vowels such as /iː/, /ɪ/, /juː/,^[11] and /h/.
• Epenthetic [i]) to break up consonant clusters.
• Palatalization due to epenthetic /ɪ ~ iː/, so that night sounds slightly like nightch ([ˈnajtɕ ~ ˈnajtɕi̥] rather than /ˈnaɪt/) and light sounds like lightchie ([ˈlajtɕi] rather than /laɪt/).
• Loss of unstressed, syllable-final [i ~ ɪ ~ ɨ] to palatalization, so that city sounds slightly like sitch ([ˈsitɕ ~ sitɕi̥] rather than /ˈsɪti/).
• Post-alveolar affricates /tʃ dʒ/ are easily confused with their fricative counterparts /ʃ ʒ/, often merging chip and ship, cheap and sheep, and pledger and pleasure.
• Absence of contrast of voice for coda fricatives. He's, hiss and his are easily homophonous. Spelling pronunciations, with all words with historical schwas left in the orthography being pronounced /z/ even when the usual would be /s/, are also possible.
• English is less prone to perfect liaison-style sandhi than Portuguese, Spanish and French might be. Often, two identical or very similar consonants follow each other within a row, each in a different word, and both should be pronounced. Brazilians might either perform epenthesis or delete one of them. As such, this stop is produced either [ˈdis i̥sˈtɒpi̥ ~ ˈdiz isˈtɒpi̥] or [ˈdi sˈtɒpi̥], instead of the native /ðɪs ˈstɒp/
• In Portuguese, the semivowels [j] and [w] may be vocalized to their corresponding vowels ([i] and [u, respectively).^[12] so that I love you is pronounced [ˈaj ˈlɐviː ˈuː]. These semivowels may also be epenthetically inserted between vowels of very dissimilar qualities.
• To the exception of /s ~ z/ (here represented with a loss of contrast at the end of a word) and /r/, consonants tend to not elide corresponding to or assimilate to the next word's phoneme, even in connected speech. This means, for example, occasional epenthesis even if the following word starts in a vowel, as in their native language (not[ɕi] really).

Catalan

• Devoicing of final consonants:^[13] /b d ɡ v z dʒ ʒ/ to [p t k f s tʃ ʃ].

E.g. phase can be pronounced like face (even though Catalan has both /s/ and /z/ phonemes).^[14]

• Vowel length confusions.^[13]
• Confusion of /æ/ /ɑ(ː)/ /ʌ/, usually realized as [a]^[13]
• Confusion of /ɪ/ /i(ː)/, usually realized as [i].^[13]
• Confusion of /ʊ/ /u(ː)/, usually realized as [u].^[13]
• Confusion of /ɔ(ː)/ /ɒ/, usually realized as [ɔ] or [o].^[13]
• Confusion of /b/ /v/, usually realized as [b~β] (/b/ /v/ are only distinguished in Valencian and Balearic).^[14]
• Rhotic pronunciation, with /r/ pronounced as a trill [r] or a flap [ɾ].^[14]
• Difficulties with word-initial clusters involving /s/, where an epenthetic e is usually added.^[15]

E.g. stop being pronounced estop.^[15]

• Simplification of some word consonant clusters.^[15]

E.g. instant being pronounced instan^[15]

• Narrower pitch range, with emphasis marked with extra length instead of extra pitch variation.^[16]
• Problems with variable stress.^[13]

E.g. the blackbird. vs. the black bird.^[13]

• Problems with contrastive stress.^[13]

E.g. with sugar or without sugar? (the second sugar is more heavily stressed)^[13]

Czech

These are the most common characteristics of the Czech pronunciation of English:^[17]

• Final devoicing of voiced consonants (e.g. "bet" and "bed" are both pronounced [bɛt]), since non-sonorant consonants are always voiceless at the end of words in Czech. Some speakers may pronounce consonant-final English words with a strong vocalic offset, especially in isolated words (e.g. "dog" can be [ˈdɔɡə]).
• Czech /r/ is alveolar trill. There is a tendency to pronounce the trill in English and in all position where <r> is written.
• Final -er (-or) pronounced as syllabic alveolar trill [r̩] (e.g. "water" sounds [ˈvɔːtr̩]). Stressed /ɜː/ tends to be realized as [ɛːr] (e.g. "bird" [bɛːrt].
• Tendency to realize both /v/ and /w/ as [v], since /w/ does not exist in Czech.
• Tendency to pronounce the initial ⟨wr⟩ cluster as [vr] (e.g. "write" [vrajt]).
• Tendency to realize /θ/ as [s] or [f], since [θ] does not exist in Czech.
• Tendency to substitute /ð/ as [d] or [d͡z], since [ð] does not exist in Czech.
• Tendency to pronounce /h/ as voiced (e.g. "how" [ɦau̯]).
• Tendency not to aspirate the stops /p, t, tʃ, k/ (e.g. "keep" sounds [kiːp] instead of [kʰiːp]), since these stop consonants are not aspirated in Czech.
• /æ/ is often realised as [ɛ], so that "had" sounds like "head" [ɦɛt], homophonous with "hat".
• Schwa [ə] does not exist in Czech. Speakers tend to pronounce it as [ɛ] (e.g. "a table" [ɛ ˈtɛjbl̩]) or [a] (e.g. "China" [ˈt͡ʃajna]).
• Tendency to realise /ŋ/ as [ŋk] or [ŋɡ] (e.g. "singing" [ˈsɪŋgɪŋk]), because Czech [ŋ] is an allophone of /n/ before velar stops.
• Tendency to isolate all words in speech, because the liaison is unusual in Czech. For instance, "see it" tends to be pronounced [siː ʔɪt], rather than [siː‿ɪt].
• The melody of the Czech language is not so strong as in English. Czech speakers may sound monotonous to an English ear.

French

• Because of the phonetic differences between English and French rhotics, speakers may perceive English /r/, allophonically labialized to [ɹʷ], as /w/-like and have trouble distinguishing between /r/ and /w/.^[18]
• French speakers have difficulty with /h/ and systematically delete it, as most French dialects do not have this sound.^[19]

German

• Speakers may not velarize /l/ in coda positions as most native speakers do.^[4]
• They have a smaller pitch range, less consonant cluster reduction and less vowel reduction.^[20]

Hebrew

• The lack of discrimination in Hebrew between tense and lax vowels makes correctly pronouncing English words such as hit/heat and cook/kook difficult.^[21]
• The dental fricatives /ð/ (as in "the") and /θ/ (as in "think") are often mispronounced.^[21]
• Hebrew speakers may confuse /w/ and /v/.^[21]
• In Hebrew, word stress is usually on the last (ultimate) or penultimate syllable of a word; speakers may carry their stress system into English, which has a much more varied stress system.^[21] Hebrew speakers may also use Hebrew intonation patterns which mark them as foreign speakers of English.^[21]

Hungarian

• The dental fricatives /θ/ and /ð/ may be realised as [s̻] and [d̪] respectively.^[22]
• Since Hungarian lacks the phoneme /w/, many Hungarian speakers substitute /v/ for /w/ when speaking in English. A less frequent practice is hypercorrection: substituting /w/ for /v/ in instances where the latter is actually correct.^[23]

Italian

A study on Italian children's pronunciation of English revealed the following characteristics:^[24]

• Tendency to realise the English high lax vowels /ɪ/ and /ʊ/ as [i] and [u] (ex: "fill" and "feel" are homophones, as are "cook" and "kook"), since Italian does not have these vowels.
• Tendency to realise /ŋ/ as [ŋɡ] ("singer" rhymes with "finger") or as [n] because Italian [ŋ] is an allophone of /n/ before velar stops.
• Tendency to realise word-initial /sm/ with [zm], e.g. small [zmɔl]. This voicing also applies to /sl/ and /sn/.
• Tendency to add /h/ to some vowel-initial words, due to hypercorrection.
• Tendency to realise /ʌ/ as [a] so that mother is pronounced [ˈmad̪ɛr].
• Italian does not have dental fricatives:
  • Voiceless /θ/ may be realised as [t̪] or [f].
  • Voiced /ð/ may be realised as [d̪].
• Since /t/ and /d/ are typically pronounced as dental stops anyway, words like there and dare can become homophones.
• /æ/ is often realised as [ɛ], so that bag sounds like beg [bɛɡ].
• Tendency to pronounce /k/, /p/, /t/ as unaspirated stops.
• Schwa [ə] does not exist in Italian; speakers tend to give the written vowel its full pronunciation, e.g. lemon [ˈlɛmɔn], television [ˌt̪ɛleˈviʒɔn], parrot [ˈpɛr(ː)ɔt̪], intelligent [in̪ˈt̪ɛl(ː)idʒɛn̪t̪], water [ˈwɔt̪ɛr], sugar [ˈʃuɡar].
• Italian speakers may pronounce consonant-final English words with a strong vocalic offset, especially in isolated words, e.g. dog [ˈdɔɡːə].
• Tendency to realise /r/ as [r]; a trill rather than the native approximant [ɹ]~[ɻ].

In addition, Italians learning English have a tendency to pronounce words as they are spelled, so that walk is [walk], guide is [ɡwid̪], and boiled is [ˈbɔilɛd]. This is also true for loanwords borrowed from English as water, which is pronounced [ˈvat̪ɛr] instead of [ˈwɔːtə(r)]. Related to this is the fact that many Italians produce /r/ wherever it is spelled (e.g. star [star]), resulting in a rhotic accent, even when the dialect of English they are learning is nonrhotic.

Japanese

• Speakers tend to confuse /l/ and /r/ both in perception and production,^[25] since the Japanese language does not distinguish any consonants by laterality. The closest Japanese phoneme to either of these is /z/, though speakers may hear English /r/ as similar to the Japanese /w/.^[26]

Russian

• There is no /w/ in Russian; speakers typically substitute [v]^[27] or [u].
• Native Russian speakers tend to produce an audible release for final consonants and in consonant clusters and are likely to transfer this to English speech, creating inappropriate releases of final bursts that sound overly careful and stilted and even causing native listeners to perceive extra unstressed syllables.^[28]
• There are no dental fricatives (/θ/ and /ð/) in Russian, and native Russian speakers may pronounce them as [s] and [z].^[29]
• Difficulty with English vowels. Russian speakers may have difficulty distinguishing /iː/ and /ɪ/, /æ/ and /ɛ/, and /uː/ and /ʊ/; similarly, speakers' pronunciation of long vowels may sound more like their close counterpart (e.g. /ɑː/ may sound closer to /æ/)^[30]
• Speakers typically realise English /r/ as [r], the native Russian rhotic.^[30]
• Likewise, /h/ may be pronounced like its closest Russian equivalent, [x].^[30][31]

Spanish

• Vowel length confusions.^[13]
• Confusion of /æ/ /ɑ(ː)/ /ʌ/, usually realized as [a]^[13]
• Confusion of /ɪ/ /i(ː)/, usually realized as [i].^[13]
• Confusion of /ʊ/ /u(ː)/, usually realized as [u].^[13]
• Confusion of /ɔ(ː)/ /ɒ/, usually realized as [o].^[13]
• Since Spanish does not make voicing contrasts between its fricatives (and its one affricate), speakers may neutralize contrasts between /s/ and /z/; likewise, fricatives may assimilate the voicing of a following consonant.^[32]
• Rhotic pronunciation, with /r/ pronounced as a trill [r] or a flap [ɾ].^[14]
• Cuban and other Central American speakers tend to merge /tʃ/ with /ʃ/, and /dʒ, ʒ/ with /ɪ/.^[32]
• /ɪ/ and /w/ often have a fluctuating degree of closure.^[32]
• For the most part (especially in colloquial speech), Spanish allows only five (or six) word-final consonants: /θ/, /s/, /n/, /r/ and /l/; speakers may omit word-final consonants other than these, or alter them (for example, by turning /m/ to /n/ or /ŋ/).^[5]
• In Spanish, /s/ must immediately precede or follow a vowel; often a word beginning with [s] + consonant will obtain an epenthetic vowel (typically [e̞]) to make stomp pronounced [e̞sˈto̞mp] rather than [stɒmp].^[5]
• In Spanish, the /θ/ phoneme exists only in (most dialects of) Spain; where this sound appears in English, speakers of other Spanish dialects substitute /θ/ for /t/, /s/ or /f/.^[32]
• Speakers tend to merge /ð/ and /d/, pronouncing both as a plosive unless they occur in intervocalic position, in which case they are pronounced as a fricative.^[33] A similar process occurs with /v/ and /b/.^[32]
• The three nasal phonemes of Spanish neutralize in coda-position; speakers may invariably pronounce nasal consonants as homorganic to a following consonant; if word-final (as in welcome) common realizations include [n], deletion with nasalization of the preceding vowel, or [ŋ].^[32]
• Devoicing of final consonants.^[13]
• Narrower pitch range, with emphasis marked with extra length instead of extra pitch variation.^[16]
• Problems with variable stress.^[13]

E.g. the blackbird. vs. the black bird.^[13]

• Problems with contrastive stress.^[13]

E.g. with sugar or without sugar?

(the second sugar is more heavily stressed)^[13]

Vietnamese

Note: There are three main dialects in Vietnamese, a northern one centered on Hanoi, a central one whose prestige accent is centered on Huế, and a southern one centered on Ho Chi Minh City.

• Speakers may not produce final consonants since there are fewer final consonants in Vietnamese and those that do exist differ in their phonetic quality:^[34]
  • Final /b/ is likely to be confused with /p/.
  • Final /d/ is likely to be confused with /t/.
  • Final /f/ is likely to be confused with /p/.
  • Final /v/ is likely to be confused with /p/.
  • Final /s/ is likely to be omitted.
  • Final /ʃ/ is likely to be omitted.
  • Final /z/ is likely to be omitted.
  • Final /tʃ/ is likely to be omitted.
  • Final /l/ is likely to be confused with /n/, but some Vietnamese pronounce the word bell as [ɓɛu̯]
  • Final /t/ is likely to be confused with /k/ by southern Vietnamese.
• Speakers also have difficulty with English consonant clusters,^[35] with segments being omitted or epenthetic vowels being inserted.^[36]
• Speakers may not aspirate initial /p/, /t/, /k/ and /tʃ/, native English-speakers think that they pronounce as /d/ and /ɡ/. For example, when Vietnamese people pronounced the word tie, native English-speakers think that they say the word die or dye. respectively.^[37]
• Speakers often have difficulty with the following phonemes, which may depend in some cases upon where in Vietnam they are originally from:^[35]
  • /θ/, which is confused with /t/, /s/.
  • /ð/, which is confused with /d/, /z/.
  • /p/, which is confused with /b/ (especially in southern dialects).
  • /ɡ/, which is confused with /k/.
  • /dʒ/, which is confused with /z/.
  • /ʒ/, which is confused with /z/ or /dʒ/.
  • /s/, which is confused with /ʃ/ by northern Vietnamese.
  • /tr/, which is confused with /dʒ/, /tʃ/, or /t/ by northern Vietnamese.
  • /v/, which is confused with /ɪ/ by southern Vietnamese.
  • /ɪ/, which is confused with /iː/.
  • /ʊ/, which is confused with /uː/ or /ʌ/.
  • /æ/, which is confused with /ɑː/.
• Vietnamese is a tonal language and speakers may try to use the Vietnamese tonal system or use a mid tone with English words, but they pronounce with a high tone when the closed syllable is followed by /p, t, k/. They may also associate tones onto the intonational pattern of a sentence and become confused with such inflectional changes.^[36]

See also

• Anglophone pronunciation of foreign languages
• Pronunciation of English ⟨th⟩
• Non-native speech database
• International Dialects of English Archive
• Accent reduction
• Koiné language
• Shibboleth

References

Bibliography

• Ravid, Dorit (1995), Language Change in Child and Adult Hebrew: A Psycholinguistic Perspective, Oxford University Press, ISBN 978-0-19-508893-9 
• Hwa-Froelich, Deborah; Hodson, Barbara W; Edwards, Harold T (2003), "Characteristics of Vietnamese Phonology", American Journal of Speech-Language Pathology, 11 (3): 264–273, doi:10.1044/1058-0360(2002/031) 
• Goldstein, Brian; Fabiano, Leah; Washington, Patricia Swasey (2005), "Phonological Skills in Predominantly English-Speaking, Predominantly Spanish-Speaking, and Spanish–English Bilingual Children", Language, Speech, and Hearing Services in Schools, 36 (3): 201–218, doi:10.1044/0161-1461(2005/021) 
• Goto, Hiromu (1971), "Auditory perception by normal Japanese adults of the sounds "l" and "r""", Neuropsychologia, 9 (3): 317–323, doi:10.1016/0028-3932(71)90027-3, PMID 5149302 
• Gut, Ulrike (2009), Non-native speech. A corpus-based analysis of phonological and phonetic properties of L2 English and German., Frankfurt: Peter Lang 
• Hallé, Pierre A.; Best, Catherine T.; Levitt, Andrea (1999), "Phonetic vs. phonological influences on French listeners' perception of American English approximants", Journal of Phonetics, 27 (3): 281–306, doi:10.1006/jpho.1999.0097  |first4= missing |last4= in Authors list (help)
• Jeffers, Robert J.; Lehiste, Ilse (1979), Principles and Methods for Historical Linguistics, MIT press, ISBN 0-262-60011-0 
• Khattab, Ghada (2002), "/r/ production in English and Arabic bilingual and monolingual speakers", in Nelson, Diane, Leeds Working Papers in Linguistics and Phonetics, 9, pp. 91–129 
• Kovács, János; Siptár, Péter (2006), A-Z angol kiejtés, Budapest: Corvina Kiadó, ISBN 963-13-5557-8 
• LaCharité, Darlene; Prévost, Philippe (1999), Greenhill, Annabel; Tano, Cheryl; Littlefield, Heather, eds., 2, Somerville, Massachusetts: Cascadilla Press, pp. 373–385  Missing or empty |title= (help)
• MacDonald, Marguerite (1989), "The influence of Spanish phonology on the English spoken by United States Hispanics", in Bjarkman, Peter; Hammond, Robert, American Spanish pronunciation: Theoretical and applied perspectives, Washington, D.C.: Georgetown University Press, pp. 215–236, ISBN 9780878404933 
• Melen, Dušan (2010). Výslovnost angličtiny na pozadí češtiny (in Czech). Praha: Big Ben Bookshop Prague. pp. 71–75. ISBN 978-80-904306-2-4. 
• Munro, Miles; Mann, Virginia (2005), "Age of immersion as a predictor of foreign accent", Applied Psycholinguistics, 26 (3): 311–341, doi:10.1017/S0142716405050198 
• Nádasdy, Ádám (2006), Background to English Pronunciation, Budapest: Nemzeti Tankönyvkiadó, ISBN 963-19-5791-8 
• Paradis, Carole; LaCharité, Darlene (2001), "Guttural deletion in loanwords", Phonology, 18 (2): 255–300, doi:10.1017/S0952675701004079 
• Shoebottom, Paul (2007), "The differences between English and Hebrew", Frankfurt International School 
• Swan, Michael (2001). Learner English: A Teacher's Guide to Interference and Other Problems, Volume 1. Cambridge University Press. ISBN 9780521779395. 
• Thompson, Irene (1991), "Foreign Accents Revisited: The English Pronunciation of Russian Immigrants", Language Learning, 41 (2): 177–204, doi:10.1111/j.1467-1770.1991.tb00683.x 

Further reading

• Wiik, K. (1965), Finnish and English Vowels: A comparison with special reference to the learning problems met by native speakers of Finnish learning English, Turku: Annales Universitatis Turkuensis 

External links

• A site collecting recordings of people from different areas reading the same paragraph (most recordings also have an IPA transcription)
• International Dialects of English Archive
• Articles, Determiners, and Quantifiers