English pronunciation

Specifying priorities for Swedish learners of English

I have chosen as my sample a group of Swedish learners of English. The learners are all employed by Fagersta AB steel works in central Sweden.

I have based this study on the accompanying recording of a class discussion. The students were clearly involved in the subject of the discussion and had forgotten that they were being recorded. Although the recording was not made with pronunciation in mind, it contains a spontaneous sample of speech unlike those elicited through picture stories or reading passages in formal examinations.

The choice of a monolingual group of fairly advanced learners was deliberate. Fagersta is by no means a monocultural town. Manual workers in the steel works are predominantly Finns, and there are many East Germans, Poles, Danes, Norwegians and Turks employed in the local service industries. My lower-level classes would contain a high proportion of non-Swedes, especially Finns.

However, serious consideration of priorities for pronunciation teaching has to refer to differences between the learners' native phonological systems and the target system.

In the interests of economy, I have limited my contrastive analysis to Swedish.

The purpose of choosing a fairly advanced group is to allow two levels of analysis. The first is a broader level, which was the main reason behind the phonemic transcription.

In elementary groups of learners, priorities for pronunciation relate closely to intelligibility. Moreover, the test for intelligibility is fairly clear-cut. At higher levels of attainment, priorities can relate to more detailed judgements such as maximum intelligibility and social acceptability. I therefore wish to include a second level of analysis, which is sufficiently narrow to account for both phonemic and phonetic differences.

Assessing pronunciation

For a quick overview of the different areas of analysis and the progress of my students in pronunciation, stress, rhythm and intonation, initial reference can be made to an Appreciation Scale used for grading examination candidates.

ARELS Certificate Oral Examination candidates:

Useful scales for assessing language skills can be found in the English Speaking Union's ESU Framework 1989 by B. J. Carroll and R. West.

My students, as can be heard in the recording, oscillate between 6 and 8 on this scale. In order to diagnose the priorities for future pronunciation teaching, a fuller analysis of the recording must now follow.

The phonemic transcription

Phonological description relates closely to the standard of intelligibility since the phonemes of language systems distinguish between the meanings of words as well as allowing recognition of words.

The transcription revealed phonemes which were not realised as a result of the non-existence of the sounds in L1, differences between Swedish and English orthography – that is, students had written forms in mind – or free, idiosyncratic variations. There were also some lesser allophonic variations which came close to being phonemic.

Vowel sounds

• /i:/ replaced by [i], e.g. /swi:dɪʁ/ – Goran; /li:st/ and /mi:l/ – Bert.
• /ɪ/ replaced by [i], e.g. /sɪtɪŋ/ – Goran.
• /e/ moving towards [æ], e.g. /wet/ and /end/ – Bert.
• /eɪ/ replaced by /e/ and [æ], e.g. /deɪnɪʃ/. Note the contrast between Swedish orthography in dansk and English orthography in Danish.
• /ə/ (schwa) is under-used and sometimes replaced by /ɜ:/, the vowel sound in girl: /kɜ:nt/ instead of /kənt/ or /kɑ:nt/; /wɪ hæv bi:n tɔ:t/ shows failure to use the weak form /həv/; /fɔ:/ instead of /fə/ and /wʊmæn/ instead of /wʊmən/ – Goran.
• /əʊ/ replaced by /ɔʊ/ as in /əʊn/ and /həʊm/ – Bert.
• /u:/ moving towards /ʊ/ as in /fu:d/ – Bert, twice.

Consonant sounds

• /θ/ replaced by /f/, e.g. /θri:/ – Goran.
• /ʤ/ replaced by /dj/, as in /ʤəʊk/ or /ʤentlmen/ – Goran; /sʌbʤekt/ – Anita.
• /ʤ/ replaced by /j/, as in /ʤʌst/ – Goran. This is partly an interference problem arising from Swedish orthography.
• /v/ replaced by /f/ in /əfektɪv/ – Anita.
• /j/ is given excessive friction, as in /jʌŋ/ – Bert.

The differences in sounds rarely threatened intelligibility in their contexts in the transcript. However, some of the substitutions could possibly lead to confusion if duplicated in other contexts. It is therefore worth considering questions such as learnability, teachability and functional load.

There is generally much greater allophonic variation of vowel sounds than of consonant sounds among native speakers of languages.

Consonant sounds play a greater part in signalling meaning in words, so they are usually less redundant and carry a greater functional load.

Vowel sounds, however, may relate more closely to the standard of social acceptability that the advanced learner is also trying to achieve.

Vowel quadrilateral

Vowel quadrilateral showing the 12 English monophthongs and the 16 Swedish monophthongs.

Since Swedish possesses many vowel sounds, the English vowel phonemes should not present too much difficulty for Swedes. Arabic, for example, has far fewer vowel sounds, so Arabic speakers will need considerably more help in this area.

It is probably worth doing some work on the /ɪ/ and /i:/ contrast, which distinguishes a large number of minimal pairs in English. Swedes tend to replace English /i:/ with [i] because their own [i:] is produced with the tongue so close to the hard palate that it ends with a fricative sound [j]. Confusion results when they also replace English /ɪ/, which they do not have, with [i] if the context is ambiguous.

Similar considerations apply to the /e/ and /æ/ contrast, which also bears a relatively high functional load.

Like English, Swedish has a marked difference in the force of articulation between stressed and unstressed syllables. It uses full stress, reduced main stress, weak stress and lack of stress, and it has the neutral vowel, or schwa, [ə].

Swedes will generally not require as much practice of schwa as speakers whose native languages are syllable-timed, such as French speakers. Their under-use of schwa probably arises from the interrelationships between intonation and stress – Swedish is a tone language – and from inadequate teaching.

The placement and production of schwa on unstressed syllables is vital in marking stressed syllables by contrast, thereby signalling both the redundant and the meaningful elements of the utterance. Poor use or non-use of stress will inevitably lead to poor rhythm and intonation.

The substitutions for the back vowels are less likely to affect intelligibility, provided Swedish [o:] is not substituted for English /u:/.

The Swedish pensioners whom I teach some summers in Portslade often ask for [kɪʧ ɪ n so:p], "kitchen soap", in the student canteen when they really want /ʧɪkɪn su:p/, "chicken soup". Note that /u:/ does not exist in Swedish.

Consonant sounds

There is a relatively high degree of correspondence between Swedish and English consonant sounds. [p], [t] and [k] occur with aspiration in both languages. The aspiration is more vigorous in Swedish at the beginning of stressed syllables.

[m], [b], [f] and [v] are pronounced similarly. Note that orthographic Swedish v is pronounced [f] before an inflectional -s or following a shortened vowel.

Swedish [t], [d], [n], [l] and [s] differ slightly from their English counterparts. They normally occur as pure dental sounds produced with the tip of the tongue against the upper teeth.

However, when Swedish [r] is immediately followed by any one of these consonants, an assimilation takes place and they become post-alveolar or retroflex in their articulation.

Although Swedish [k] and [g] occur as fricatives before front vowels, [k], [g] and [ŋ] are otherwise pronounced much the same as in English. Swedish [l] in all positions is a clear front sound.

My students make fairly good substitutions for those consonant phonemes which exist in English but not in Swedish.

Goran's substitution of /f/ for /θ/ and the general tendency of Swedes to substitute [dj] for [ʤ], as all three do, will probably not be noticed by most English ears.

More noticeable is the substitution of [j] for [ʤ], which could easily cause misunderstanding or social irritation. When my Swedish pensioners tell me of all the different "yobs" that they have had, I usually guess that they mean "jobs", though I have often confused the Swedish pronunciation of "just", as spoken by Goran, with the word "yes".

A story often quoted is that of the Swedish couple who go out together for a drink. The wife is acting as chauffeur, so the husband's order to the barman is:

[ʤ], or [dj] if they cannot manage it, is therefore relatively important, as is the voiceless English affricate [ʧ], which is often replaced by the fricative [ʃ] or [ʂ], e.g.:

Of the two affricates, [ʤ] will prove the most difficult to teach because of orthographic interference from Swedish j and the rarity or non-existence in L1 of voiced fricatives ranging from dental to alveolo-palatal articulations.

Swedish orthographic j [j], the voiced palatal fricative, has considerably more friction than English [j]. Its alternative pronunciation [ʂ] in French loan words such as [ʂɞɳɑ:] "journal" does little to lessen the confusion.

To complete this summary of my priorities in teaching English consonant sounds to Swedes, I shall have to include a few additions to the problem areas represented in my sample.

An important addition is the /s/ and /z/ contrast, e.g. piece and peas; police and please; is and was. The Swedish letter z occurs only in words of foreign origin, e.g. zon "zone", and is pronounced [s].

A similarly difficult contrast, but one which bears a lower functional load, occurs between /v/ and /w/. The Swedish letter w is also rare, occurring mainly in loan words borrowed from English, e.g. weekend and whisky. I should probably expect advanced-level students to learn not to substitute [v] for [w].

After that would come more detailed refinements, such as slightly reducing the aspiration on voiceless plosives and the friction on [j]. Although somewhat aesthetic, the latter has the virtue of being easily teachable.

Intonation and stress

A general observation about English spoken by Swedes, which I can confirm from my own experience, is that statements may sound like questions and sentences may sound incomplete. The general prescription recommended is further practice in falling tune.

However, although simple prescriptions may help some learners, they are often exposed by more detailed linguistic analysis.

Swedish has a general similarity with English in its use of rising tone in suspensive groups and falling tone in conclusive groups.

When a word having a single tone occurs in the body of a sentence expressing a categorical statement and is emphatic, the tone rises; when it occurs at the end, the tone falls. Interrogative sentences seeking verification are also marked by a rising tone at the end. Gillian Brown points out that this is by no means always the case.

These general patterns would appear to contradict our general observation. However, unlike English, Swedish can be described as a tone language – that is, it makes considerable use of pitch, accent or differences in tone to distinguish between different words and word forms.

Its two tones are known as the single tone and the double tone.

In monosyllabic words, a similar single tone can be heard in both Swedish and English. bok in Swedish and book in English both have the same falling tone.

But in disyllabic words with one stress, such as boken (Swedish, "the book") and father (English), different patterns occur even though both words have a falling tone.

Whereas the English tone drops to a low level on the first syllable and remains low on the second, the Swedish tone only drops slightly on the first syllable and then drops to a low tone. This is one reason why statements by a Swedish speaker which end in a disyllabic word may sound incomplete or like questions to an English listener expecting a steeper, more conclusive fall on the last stressed syllable.

Like its single counterpart, the Swedish double tone may not be considered conclusive by the English listener when used to round off a statement ending in a disyllabic word. Again, it is only the last syllable of the word that drops to the lowest tone. The last stressed syllable – the first syllable of the word – only drops to the middle area of the pitch range.

In interrogative sentences requiring verification, the double tone produces an effect which is uncommonly heard in English.

There is quite a contrast between the relatively low fall on the last stressed syllable and the rise at the end.

Another sharp change of direction is heard when the double tone occurs in the body of a sentence and is emphatic, for example Swedish flickan.

English is not a tone language in the sense described above for Swedish. The intonation patterns of English are used to mark the functions of sentences and the attitudes of speakers rather than operating as tones used to distinguish between different words or word forms in sentence structure.

The occasions where words require emphatic stress are fewer in English than in Swedish, so the movement of our intonation patterns is generally more gradual.

In the recording, Goran, Anita and Bert transmit many signals which are probably unintended through sharp changes of direction and other borrowings from their native use of tones.

Goran's borrowings involve movements into high key which, to the English ear, sound over-emphatic. These include his use of the "rise" and "rise-fall" tones, which linguists such as David Brazil associate with the assertion of dominance. I have indicated many of these movements by inserting red arrows above the words on the phonemic transcript.

Goran's over-emphasis is not merely a question of force on stressed syllables. His superimposition of Swedish tones leads to the perception of stress on syllables which he probably does not intend to highlight semantically.

It is probably worth teaching advanced Swedish learners of English a little about the function of pitch movements and stress in English speech.

Even Bert, whose pitch range is generally narrower than Goran's and whose use of falling tune is more frequent, needs to be wary of hopping down to a low tone or up to a high tone where stress is unintended.

The general message for Swedes is that any syllable on which the pitch of the voice moves perceptibly, whether the pitch rises or falls, will be perceived in English speech as stressed.

My priorities in the teaching of intonation to advanced Swedish learners of English will steer me towards materials which illustrate the English patterns themselves – the gradual movements of the basic tunes – and which aid recognition of the different functions of intonation in English speech.

Such materials may include books such as Roy Boardman's Over to You, Leo Jones's Functions of English and V. J. Cook's Using Intonation, which is more functional than his earlier work Active Intonation, which mainly consists of structurally based drills.

In defining the functions of intonation, I shall be covering a fairly broad area involving many language skills. Reference can be made to speech function, attitude – for example, affective meaning – the interactional structure of discourse, such as turn-taking, information structure and topic structure.

In terms of what I would present to my students, these references will translate into functions or topics such as:

• polite suggestion, persuasion and showing anger;
• holding the floor and letting other speakers in;
• listening to lectures, particularly the intonation on discourse markers signalling a change of topic or distinguishing important content from material which can be treated as less important.

My students are sufficiently advanced to understand explanations given in these terms. They will equally need talking time while I lend them my English ear, gathering impressions as to their standards of intelligibility and whether the attitudes they are transmitting are those which I think they mean to convey.

I suspect that Goran, Anita and Bert would get better feedback if they held their discussion in a multilingual class.

Transcript of speech sample

Transcript of unscripted discussion by Swedish students of English on 'male chauvinism' based on:
'International Business Topics' by David Cotton (Evans) P143 Equality for women – Sweden shows how.

The participants are from my Business English class in Fagersta, Central Sweden.

• T = Teacher / Ted
• G = Goran
• A = Anita
• B = Bert