Professional Documents
Culture Documents
Introduction
Dutch and English have a considerable number of phonological commonalities in that they
Nonetheless, Dutch speakers ‘commonly find it difficult to master some of the sounds of
English’ (Wester, Gilbers & Lowie 2007, p. 477). This case study explores English pronunciation
and prosody by a native Dutch speaker, comprising a needs analysis and a multi-model
intervention programme to promote intelligibility and reduce interlocutor load. The learning
activities include drama, haptic, and NLP techniques in order to remediate the four
pronunciation errors that pose the largest barriers to intelligibility. These include vowel and
consonant substitutions, inadequate word stress and insufficient separation of tone units.
Method
The data for this study were collected in an informal setting, whereby the learner recorded
two speaking tasks on a smart phone. The tasks consisted of a reading of a set script and a
free speech assignment. The recording was transcribed into English orthography (Appendix
chart (Appendices B & C). The results predominantly confirm findings of previous research of
the target learner group; however, one unexpected feature was identified.
Participant Background
The participant is a 45-year-old native Dutch EFL speaker, who converses in Standard Dutch
secondary schooling. This required approximately five hours of classroom participation and
self-study per week, which primarily involved literature studies, vocabulary learning and
grammar practice. The national curriculum traditionally places a strong emphasis on reading
skills (European Commission 2012, p. 127), whereas oral communication and pronunciation
are either neglected or omitted from school curricula entirely (Noorman 2008, p. 6).
converses with both NNS and NS interlocutors. Although he feels confident conversing with
the former, he feels less confident talking to the latter. He expressed integrative and
to avoid miscommunications with NNS staff so that company time is used more effectively.
The speech recording was analysed for segmental and suprasegmental features respectively.
unfamiliar phonemes; word-final devoicing; apocope; lack of tone unit boundaries; and,
insufficient word stress and intonation. The findings predominantly confirm the research
literature of Dutch speakers of English (DSE). The discussion below is limited to the most
salient issues.
the voiced dental fricative with /d/, /v/ or /f/, and its voiceless counterpart with /f/ or /t/,
depending on positional variation (Appendix B, pp. 24 - 25). Additionally, the open front vowel
is consistently substituted in that the words ‘accent’, ‘can’ and ‘practice’ are pronounced with
a prominent /ɛ/ sound (Appendix A, line 1, 10 & 11). These findings correlate with the results
of a small-scale research project of 132 DSE, which found that pronunciation of the phonemes
/ð/, /θ/ and /æ/ appeared to be problematic for more than fifty percent of the participants
(Hermans & Sloep 2015, p. 61). These errors may be due to L1 interference in that these
phonemes have no correlate in the native tongue. Considering that language learners filter
the L2 through the sound inventory of the L1, they are likely to hear and produce unfamiliar
Nonetheless, vowels that do have a correlate are also pronounced with a distinct ‘Dutch
flavour’ due to incorrect movement of the articulators. For example, the participant produces
the low central vowel /ʌ/ either as a low back or a mid central vowel (Appendix B, p. 25).
Furthermore, diphthongs lack the characteristic glide quality in that the closing diphthong
/əʊ/ becomes the long pure vowel /ɔ:/ (Appendix B, p. 26). These findings match the results
from the aforementioned study, confirming that incorrect productions of /ʌ/ and /əʊ/ may
be characteristic of DSE.
Contrarily, the participant substitutes /ɪ/ with /i:/ (Appendix B, p. 25), which is not listed in
the inventory of common errors of DSE in the aforementioned study. At present, the cause is
suggesting facilitation rather than interference. However, the error occurs in the
pronunciation of multisyllabic words only, which may indicate this is a suprasegmental rather
than segmental error, in that the speaker applies an L1 word stress pattern by lengthening
word-final consonants tend to be devoiced or deleted. For example, /d/ becomes /t/, /z/
becomes /s/, /v/ becomes /f/, and /ʤ/ becomes /ʒ/ (Appendix B, p. 24). Furthermore, /t/ is
‘difficult’ as /dɪfɪkʌl/ (Appendix A, line 1 & 2). This error is particularly problematic in
negations, in that ‘can’t’ is pronounced as /kɛn/ (Appendix A, line 10), resulting in utterances
that convey the opposite of intended meanings. These segmental errors might be caused by
obligatory phonological rule of devoicing (Kuijpers, Van Donselaar & Cutler 2002, p. 1661).
plosive is usually released in formal speech; however, in informal registers, it may be reduced,
especially in coda clusters, or even entirely deleted (Mitterer & Ernestus 2006, p. 77). The
participant speaks a Dutch dialect in informal settings, in which the word-final /t/ is not
The analysis of suprasegmental features demonstrates that the learner applies prosodic
patterns that research literature confirms as characteristic of DSE. This may be caused by L1
interference in that NNSs ‘tend to rely on their L1 rhythm system . . . and this is likely to
Firstly, the participant does not pause between tone units when expected (Appendix C, p. 30).
An analysis of a multilingual speech database found that English speakers pause eight times
in a fast-spoken text consisting of thirty-five words, whereas Dutch speakers pause only ten
times in a text more than twice in length (Monaghan 2002, Table 20.2 & 20.4). Although
comparison of the data is problematic in that each participant read an L1 text, it might
indicate that Dutch speakers pause less frequently than Anglophone speakers. The participant
may have transferred the L1’s rapid rate of speech to the L2.
Secondly, the participant produces the focus words without sufficient highlights or
distortions. Syllables with secondary-stress and unstressed syllables are produced with nearly
equal vowel quality (Appendix C, p. 29). This matches the findings in a review article, which
reported that Dutch speakers ‘do not segmentally reduce the unstressed vowels of English
words’ (Braun, Lemhöfer & Mani 2010 p. 378). This is caused by L1 interference in that an
unstressed syllable directly adjacent to a stressed syllable most often contains a full vowel in
Thirdly, the participant produces the peak in the stressed syllable without sufficient vowel
that DSE produce smaller pitch movements than NS, produce no precursive rise, and produce
The features discussed above signal that the participant is a non-native speaker of English.
However, accentedness is not inherently problematic since it does not necessarily have a
negative impact on either intelligibility or comprehensibility (Munro & Derwing 1999, p. 303).
On the contrary, a small-scale study found that native English listeners might perceive a
accented (Munro & Derwing 1999, p. 302). Therefore, the identification of the four most
problematic features has been guided by the relative functional load (RFL) principle and the
prosody pyramid theory. The identification has been guided furthermore by the general rule
that a conflation of ‘stress errors with other types of errors can seriously disrupt
The two most serious segmental errors include the substitution of word-final /d/ with /t/ and
the substitution of /æ/ with /ɛ/. These errors have been prioritised based on the notion that
these phoneme contrasts carry a high RFL, namely RFL 72 and 42 respectively (Celce-Murcia
2008, pp. 108 & 260). Therefore, these errors are likely to reduce intelligibility significantly by
conveying incorrect or confusing messages. Especially when the two segmental errors appear
in conjunction, as in the minimal pairs /bæd/, /bɛd/, /bæt/ and /bɛt/, which all sound alike, it
is unlikely that the speaker’s intended meanings will be conveyed correctly. Instead,
inadequate contrastive highlighting and obscuring in the focus word. The continuous flow of
speech punctuated with few and relatively brief boundaries renders it a sea of sounds. This
places a burden on the listener, who – if the burden becomes too great – might stop listening
emphasised thus it is unlikely that listeners grasp the speaker’s intended meanings. The
combined lack of thought group boundaries and lexical stress have a negative impact on
The four aforementioned features should remain the focus in both ESL and EIL contexts in
that intervention promotes listener-friendly speech that is intelligible to both NSs and NNSs.
However, in ESL contexts, the teaching and learning programme could be supplemented with
speaker.
In ESL contexts, accentedness could receive more attention than in EIL contexts based on the
perception that ‘there is intolerance for foreign accents in certain circles, particularly
employers’ (Sato 1991, cited in Munro & Derwing 1999, p. 286). Therefore, out of
consideration for the participant’s socio-economic well-being, instruction should target those
features that most saliently signal the learner as a NNS. Perhaps the intervention programme
speaker as a NNS.
should be intelligible to other NNS speakers. Listening and speaking is complicated in EIL
phonological rules. Therefore, Jenkins (2002, p. 96) argues for pronunciation teaching based
on the Lingua Franca Core (LFC), which focusses on ‘those phonological and phonetic features
three out of the four problem areas identified would be considered non-core features, in that
word-initial and word-medial position only, and vowels can be pronounced with consistent
Contrarily, the LFC guidelines appear counterproductive in relation to the present study.
Firstly, it is unclear how Jenkins envisages a speaker emphasises meanings if not through word
stress. Secondly, consonant clusters in word-final position often carry crucial grammatical and
semantic information thus suggesting that they should be teaching priorities. This is
confirmed by the present study in that the speaker deletes negations in word-final consonant
clusters thus reducing intelligibility. Thirdly, the participant produces segmental errors that
have a high RFL, which, especially when conflated, might result in miscommunication and
confusion. For example, the participant pronounces /bæd/, /bɛd/, /bæt/ and /bɛt/ identically,
thus increasing interlocutor load, in that the listener has to solve a linguistic puzzle before the
intended meanings can be comprehended. The LFC guidelines appear to contradict the
Therefore, the aforementioned teaching priorities should not be changed in an EIL context in
Intervention Programme
The intervention program is sequenced according to the prosody pyramid in that it progresses
from thought group boundaries to lexical stress to segmental errors. It includes controlled,
guided and free practice, including activities based on drama, haptic, and NLP techniques.
Where possible, the activities align with the learner’s vocational and avocational interests in
order to render the programme relevant to the learner’s context and foster engagement with
the activities.
This part of the programme focusses on thought group boundaries. The first activity is an
effective tone unit boundaries. Accordingly, the learner listens to two speeches. The first
speech is delivered rapidly, punctuated with few pauses only. The second speech is more
listener-friendly, characterised by a slower speech rate and appropriate tone unit boundaries.
Each speech is followed by listening comprehension questions. The purpose of this activity is
to demonstrate that slower speech with effective boundaries facilities comprehension of the
message. Once awareness of the relation between tone unit boundaries and intelligibility has
been developed, the prosody pyramid and the concept of thought groups should be
The second activity requires that the learner writes and delivers a work-related speech, which
needs to be recorded. Firstly, the learner needs to deliver the speech as per usual, without
considering appropriate pauses. Thereafter, he needs to listen to the recording and mark on
a transcript where he should have paused to improve intelligibility. After this analysis, the
learner needs to make a second recording incorporating his own findings as well as teacher
feedback. The second recording also needs to be evaluated for improvements in delivering
The following two activities focus on word stress. The first activity requires that the learner
of word stress including schwa needs to be explained, perhaps with the help of visual aids and
cartoons to clarify the concepts. Additionally, phonetic notations of word stress and the
schwa-symbol need to be shown so that the learner can proceed with the first activity. This
requires that the learner identifies work-related terminology and examines these in a
dictionary to discover their stress patterns and possible vowel reductions. This is followed by
controlled practice, whereby the learner pronounces the words with the correct stress
The second activity aims to consolidate the speaking skills introduced above. It includes haptic
techniques, based on the notion that ‘a motor act activates neuronal pathways, classically
an adaptation of Acton’s ‘Rhythm Fight Club Demonstration’, and requires a punch bag and
boxing gloves. This technique is particularly suitable for the learner, who has been a keen
amateur boxer for many years. It proceeds as follows: the learner punches the bag while
uttering the words identified in the activity above. The syllable with primary stress is
gentle jab; and unstressed syllables are accompanied by no punches at all. This activity can
progress in complexity, in that the movements can accompany entire phrases, whereby the
learner uses most strength on the focus words, limited strength on non-prominent words,
The following activities focus on devoicing, and the substitution of /d/ with /t/ specifically.
The purpose of the activities is to teach the learner resyllabification and blending strategies
blending two words, the troublesome voiced consonant moves from word-final to word-
medial position thus no longer posing a pronunciation problem. The first activity involves
controlled pronunciation practice of CC+V sequences including /d/ in word-final position, e.g.
‘blin/d__eye, wil/d__ass and ol/d__age’ (Celce-Murcia et al. 2010, p. 172). The activity can be
During the narration, the learner needs to use regular verbs in the past simple tense thus
position. The learner may need to develop a short vocabulary list beforehand. The free
al. (2010, p. 339). Firstly, the learner has to visualise the event to be narrated. Secondly, he
needs to visualise the action verbs. Thirdly, he needs to imagine the /d/s flying from one word
to the next. Once the learner has a strong mental image of /d/-resyllabification, he can
proceed with the narration. The learner should try to recall the mental imagery created prior
The last part of the intervention programme focusses on the substitution of /æ/ with /ɛ/. The
activities include drama techniques and controlled practice of minimal pairs. The activities
/æ/ versus /ɛ/. The first activity involves listening to a recorded speech during which the
teacher and learner jointly mime the production of /æ/ and /ɛ/ sounds. The miming reinforces
that the vowels require different jaw movements, whereby the jaw is opened further to
The second activity consolidates accurate production of the open front vowel by practising
minimal pairs, such as /bæt/ versus /bɛt/, /mæt/ versus /mɛt/. The difficulty level can be
This project has provided numerous new insights and skills in addition to learning how to
conduct a pronunciation needs analysis and undertake a mini-research project. Three insights
appear particularly significant in relation to this project. Firstly, I have learned that
pronunciation – like grammar – can be segmented into manageable chunks, which renders it
both teachable and learnable. Especially for adult learners, this is inspiring news, in that it
means that no one needs to accept a fossilised accent. Instead, it can be improved upon in a
systemic manner. Secondly, I have learned that mastering English pronunciation is more
multi-faceted than learning how to speak like the natives do. As interactive, communicative,
goal-oriented beings, clear communication appears a more rewarding goal than acquiring a
NS accent. The participant will potentially reap more benefits from developing clear, listener-
speaker accent. Thirdly, I am starting to develop a more ‘perceptive ear’, which seems like the
auditory equivalent of the ability to read body language. Prosody conveys interpersonal
meanings the speaker may not intend to verbalise. By carefully examining the participant’s
speech for prosodic clues, I have become – admittedly only incipiently – more aware of the
boredom, enthusiasm; they can all be detected, if only one knows how to listen.
Braun, B, Lemhöfer, C & Mani, N 2010, ‘Perceiving unstressed vowels in foreign-accented English’,
Journal of the Acoustical Society of America, vol. 129, no. 1, pp. 376 – 387.
Celce-Murcia, M, Brinton, DM, Goodwin, JM & Griner B 2010, Teaching Pronunciation, Cambridge
University Press, New York.
Cooper, N, Cutler, A & Wales, R 2002, ‘Constraints of Lexical Stress on Lexical Access in English:
Evidence from Native and Non-Native Listeners’, Language and Speech, vol. 45, no. 3, pp.
207 – 228.
European Commission 2012, Key Data on Teaching Languages at School in Europe 2012, European
Commission, viewed 16 October 2015,
http://eacea.ec.europa.eu/education/eurydice/documents/key_data_series/143en.pdf.
Gilbert, JB 2008, Teaching Pronunciation Using the Prosody Pyramid, Cambridge University Press,
New York, USA.
Hermans, F & Sloep, P 2015, ‘Teaching the Dutch how to pronounce English’, International Journal of
Language Studies, vol. 9, no. 4, pp. 55 – 80.
Kuijpers, C, Van Donselaar, W & Cutler, A 2002, ‘Perceptual effects of assimilation-induced violation
of final devoicing in Dutch’, paper presented at The 7th International Conference on Spoken
Language Processing, 16 – 20 September, Denver, viewed 16 October 2015,
http://www.researchgate.net/publication/27269181_Perceptual_effects_of_assimilation-
induced_violation_of_final_devoicing_in_Dutch.
Mitterer, H & Ernestus, M 2006, ‘Listeners recover /t/s that speakers reduce: Evidence from /t/-
lenition in Dutch’, Journal of Phonetics, vol. 34, no. 1, pp. 73 – 103.
Monaghan, A 2002, ‘An Auditory Analysis of the Prosody of Fast and Slow Speech Styles in English,
Dutch and German’, Improvements in Speech Analysis, 1st edn, John Wiley & Sons, Hoboken,
USA, pp. 204 – 217.
Noorman, I 2008, ‘ENL, ESL, EFL, EIL or ELF? On the Idea of Teaching the Lingua Franca Core at Dutch
Secondary Schools’, MA Thesis, Utrecht University.
Wester, F, Gilbers, D & Lowie, W 2007, ‘Substitution of dental fricatives in English by Dutch L2
speakers’, Language Sciences, vol. 29, no. 2 – 3, pp. 477 – 491.
Willems, N 2010, English Intonation from a Dutch point of view, Netherlands Phonetic Archives, De
Gruyter Mouton, Berlin, New York.
Explanations
The first text below is an orthographic transcript of the scripted text read aloud by the learner.
This text is the reading passage as provided for the subject ‘English Pronunciation and
Prosody’. The second text is a transcript of an unrehearsed narrative during which the learner
talks about his hobby, namely cross-country horse-drawn carriage racing. Nick refers to this
interest by the Dutch term ‘mennen.’ During the narrative he states that he has become the
‘reservist of the section East’ which means that he is currently the runner-up in regional
The following symbols have been used to transcribe the audio recordings:
Asterisk * for obviously mispronounced words due to the learner’s unfamiliarity with
Single dash immediately after phoneme - for words not pronounced completely,
Square brackets [] for words inserted by the transcriber in order to clarify the meaning
of unrehearsed/unscripted speech,
Triple full stops . . . for omissions of text from the original script.
Is English your native language? If not, your foreign accent may show people that you come
from another country. Why is . . . difficult to speak a foreign language without an accent?
There are a couple of answers to this question. First, age is an important factor of learning in
*pronuncation. We also know that older learners usually have an accent, to some older
Another factor that influences your – influences your *pronuncation is your first language.
English speakers can’t, for example, recognize people from France by their France accents.
They can also identify Spanish or Arabic speakers over the telephone, just by listening
carefully to them. . . . this mean that accents can- be changed? Not at all! But you can change
your pronunciation without a lot of hard work. In the end, improving appears to be a
combination of three things: concentrated hard work, a good ear, and a strong ambition to
You also need accurate information about English sounds, effective strategies for practice,
lots of exposure to spoken English, and patience. Will you make progress, or will you give up?
Only time will tell. I’m afraid. But it’s your decision. You can improve! Good luck, and don’t
And now I’m going to tell something about my hobby. I am – my hobby is a Dutch – uh it’s
called in Dutch ‘mennen’. It’s dri- -- is driving with a horse, but you are not sitting on the horse,
but you are driven on a car [carriage] behind it. In the Netherlands, I am the reserve champion
of the section ‘East.’ What do you have to do for this uh hobby? Now, first of all, you need a
very fast horse. And also, what you need is a good – is a good car [carriage]. What do have to
do? You have to drive as fast as possible through twenty coins. The distance is one kilometre,
it in three or less uh than three minutes, than you can possibly win. If I will be the champion
Segmental Analysis
1 Is English your native language? If not, your foreign accent may show people that you come
2 from another country. Why is . . . difficult to speak a foreign language without an accent?
3 There are a couple of answers to this question. First, age is an important factor of learning
4 in pronounce. We know that young children can learn a second language with a perfect
5 pronuncation. We also know that older learners usually have an accent, though some older
/ɪndi:vi:du:ls/ /wɪvaʊt/
7 Another factor that influences your – influences your pronuncation is your first language.
8 English speakers can’t, for example, recognize people from France by their France accents.
9 They can also identify Spanish or Arabic speakers over the telephone, just by listening
10 carefully to them. . . . this mean that accents can- be changed? Not at all! But you can
11 change your pronunciation without a lot of hard work. In the end, improving appears to
12 be a combination of three things: concentrated hard work, a good ear, and a strong
/neɪtɪf/
14 You also need accurate information about English sounds, effective strategies for practice,
15 lots of exposure to spoken English, and patience. Will you make progress, or will you give
/dɔ:nt/ /hɑ:rt/
18 And now I’m going to tell something about my hobby. I am – my hobby is a Dutch – uh it’s
19 called in Dutch ‘mennen’. It’s dri- -- is driving with a horse, but you are not sitting on the
/dətʃ/ /wɪv/
20 horse, but you are driven on a car behind it. In the Netherlands, I am the reserve
/nɛdərləns/
21 champion of the section ‘East.’ What do you have to do for this uh hobby? Now, first of
/ʧæmpjən ɒf/ /hæf/ /dɪs/ /ɒf/
22 all you need a very fast horse. And also, what you need is a good – is a good car.
/fɛri:/
23 What do you have do? You have to drive as fast as possible through twenty coins. The
25 the coin, and, you must drive it in three or less uh than three minutes, than you can
26 possibly win. If I will be the champion next week on Saturday, I have to be -- I go to the
27 championship of the Netherlands. If I will be the champion? I don’t think so. But, I will try.
NB: Relative functional load (RFL) rankings originate from Celce-Murcia, M, Brinton, DM, Goodwin, JM
& Griner B 2010, Teaching Pronunciation, Cambridge University Press, New York.
Suprasegmental Analysis
Transcript 1: Scripted text
1 || Is English your native language? | If || not, || your foreign accent may show people that
2 you come from another country.| Why is . . . difficult to speak a foreign language without
3 an accent? || There are a couple of answers to this question. | First, | age is an important
4 factor of learning in | pronounce. || We know that young children can learn a second
5 language with a perfect pronuncation. | We also know that older learners | usually have
6 an accent, | though some older individuals | also have learned to speak without an
7 accent.|
8 | Another factor that influences your || influences your || pronunciation | is your first
9 language. | English speakers can’t, for example, recognize people from France by their
10 France accents. | They can also identify Spanish | or Arabic speakers over the telephone,
11 | just by listening carefully to them | . . . this mean | that || accents can- be changed?
12 | Not at all! || But you can change your | pronunciation without a lot of hard work. | In
14 | a good ear, and a strong ambition to see | to sound like a native speaker.||
15 || You also need accurate information about English sounds, | effective strategies | for
16 practice, | lots of exposure to spoken English, and patience. || Will you make progress,
17 | or will you give up? || Only time will tell. | I’m afraid. | But it’s your decision. You can
19 || n now I’m going t tell something bout m hobby. || I am | my hobby is a Dutch | uh it’s
20 called in Dutch |‘mennen’. | It’s dri- -- is driving with a horse, | but you are not sitting on
21 the horse, but you are driven | on a car behind it. || In the Netherlands, I am the reserve
22 champion of the section ‘East.’ || What do you have to do for this | uh hobby? | Now,
23 first of all | you need a very fast horse. || And also, || what you need is a good | is a good
24 car.|| What do you have do? || You have to drive || as fast as possible || through twenty
25 coins. | The distance is one kilometre, | you have to do this in three minutes. || You
26 mustn’t touch one of the coin, | and, | you must drive it in three or less uh than three