Contrastive study of english voiceless stops by native and non-native speakers

FACULTAD DE FILOLOGÍA
GRADO EN ESTUDIOS INGLESES
TRABAJO DE FIN DE GRADO
CURSO: 2024 / 2025
TÍTULO: CONTRASTIVE STUDY OF ENGLISH VOICELESS STOPS BY
NATIVE AND NON-NATIVE SPEAKERS
AUTORA: CLARA MARÍA ROSILLO GARCÍA
FECHA: 28 / 05 / 2025
TUTOR: DARÍO BARRERA PARDO
FIRMA ALUMNA: FIRMA TUTOR:
0. Abs ac
This s udy in es iga es he p oduc ion o he oiceless s ops consonan s o English (/p/,
/ /, /k/) by na i e speake s o Spanish who a e lea ning English as a second language (L2),
compa ing hem wi h na i e speake s o English. Al hough he e is no doub ha hese
phonemes exis in bo h languages, he way in which hey a e a icula ed di e s
signi ican ly be ween he wo languages, speci ically in e ms o he ime in e al be ween
he elease o he sound and he onse o ocal old ib a ion (VOT).
The p esen expe imen analyzed he VOT in 15 English wo ds p oduced by wo g oups:
i e na i e English speake s and i e na i e Spanish lea ne s, who a e aged be ween 20
and 26 yea s. Each pa icipan eco ded he wo d lis a o al o ou imes, esul ing in
app oxima ely 600 samples. The esul s o he eco dings e ealed ha na i e Spanish
speake s p oduced he oiceless s ops wi h sho e VOT alues compa ed o na i e
English speake s. These indings suppo p e ious esea ch wi h conclusions simila o
hose o his s udy, as hey sugges ha second language lea ne s end o ace di icul ies
in accu a ely imi a ing he phone ic pa e ns o he a ge language, due o he in luence
o hei na i e language, e en a e some yea s o exposu e o he a ge language.
Keywo ds: oiceless s ops, VOT, na i e English, nonna i e English
Index
1. In oduc ion .......................................................................................................... 1
2. Me hod .................................................................................................................. 4
2.1. Design ............................................................................................................... 4
2.2. Pa icipan s ....................................................................................................... 4
2.3. Ma e ials ........................................................................................................... 5
2.4. P ocedu e .......................................................................................................... 6
2.5 Analysis ............................................................................................................ 6
3. Resul s .................................................................................................................. 9
3.1. Desc ip i e s a is ics ......................................................................................... 9
3.2. In e en ial s a is ics .......................................................................................... 11
4. Discussion .......................................................................................................... 16
4.1. My expe imen ’s con ibu ion o knowledge in his a ea o phone ics .......... 18
4.2. Faul s and limi a ions ..................................................................................... 20
5. Conclusion .......................................................................................................... 21
6. Acknowledgemen s ............................................................................................ 23
7. Bibliog aphy ....................................................................................................... 23
1
1. In oduc ion
The globally dominan language oday is English, which o e s g ea ad an ages such as
connec ing people ac oss impo an sec o s such as business, science and en e ainmen .
As i is conside ed a lingua anca, i enables ade and uni e sal communica ion in a e y
e ec i e way. I has app oxima ely 375 million speake s as a mo he ongue and o e 750
million speake s as a second language, making English he mos widely used language
wo ldwide (Rao, 2019). Among he numbe o hose lea ning his language as a second
language a e a la ge numbe o na i e Spanish speake s, who s uggle o cope wi h a
numbe o speci ic di icul ies due o linguis ic a ia ions be ween he wo languages.
P onuncia ion is conside ed one o he mos exhaus ing ac o s, because o he
phonological sys em o English, which compa ed o Spanish is conside ed o be mo e
complex. The lea ne 's mo he ongue has a signi ican in luence on hei compe ence in
acqui ing English sounds and s uc u es, which in some si ua ions makes i di icul o
him o he o domina e ce ain p onuncia ions. Acco ding o he Speech Lea ning Model
(Schoonmake -Ga es, 2012), when an indi idual lea ns a new language, hey a e inclined
o associa e he new sounds wi h hose om hei na i e language. Fo example, i in
Spanish and English he e is a sound ha is e y simila be ween he wo languages, he
lea ne may conside i o be exac ly he same, and he e o e no app ecia e he impo an
small de ails ha di e en ia e hem. By unin en ionally igno ing hem, his makes i
di icul o p onounce some English wo ds co ec ly, such as wo ds ha con ain aspi a ion
o wo ds ha di e in VOT o a g ea ex en .
Speaking wi h a p onuncia ion iden ical o ha o a na i e speake is a g ea challenge o
people lea ning a second language, as only 5% manage o achie e i . One o he easons
is ha hey e ain some phone ic ea u es o hei na i e language, and as a esul , e y
ew indi iduals a e capable o ully mas e ing English p onuncia ion (Tho nbu gh &
Ryalls, 1998).
A numbe o ac o s in luence second language p onuncia ion, o which consonan
p oduc ion is alued as one o he mos impo an . Consonan s play a c ucial ole in
speech accu acy, accen and in elligibili y, and hus di ec ly a ec he way in which a
non-na i e speake is app ecia ed. This s udy ocuses on consonan s, speci ically on
oiceless s ops. English and Spanish sha e he phonemes /p/, / / and /k/, howe e hei
2
p onuncia ion is no simila be ween he wo languages. In English, oiceless s ops a e
aspi a ed when hey appea a he beginning o a wo d, meaning he e is a no iceable bu s
o ai (Voice Onse Time o VOT) be o e he owel begins. In con as , Spanish s ops a e
a icula ed di e en ly, wi h egional a ia ions. Gene ally, Spanish speake s p oduce / /
and /d/ as den i-al eola sounds, whe e he ongue makes con ac close o he ee h han
he al eola idge. Unlike English, oiceless s ops in Spanish do no ha e s ong
aspi a ion a he beginning o wo ds (G a elle, 2020).
A li le mo e han i y yea s ago, Liske and Ab amson in oduced a simple way o
measu e acous ic di e ences be ween s op consonan s based on hei oicing, known as
VOT. Since hen, his me hod has been used in hund eds o s udies (Ab amson & Whalen,
2017).
The ime in e al be ween he elease o a oiceless s op consonan (in his case, /p, , k/)
and he s a o ocal old ib a ion o he ollowing owel de ines wha is known as
Voice Onse Time. In e e yday e ms, his gap is o en pe cei ed as a bu s o sligh
aspi a ion when p oducing English /p, , k/. This s udy adop s he VOT de ini ion by Cho
& Lade oged (1999), which speci ies ha VOT begins a he momen he oiceless s op
is eleased and concludes a he i s ze o c ossing ha ma ks he onse o comple e ocal
old ib a ion.
In Spanish, he oiceless s ops /p, , k/ a e p oduced wi h a nea ly simul aneous elease
and onse o ocal co d ib a ion, esul ing in VOT alues ha ange om ze o o 10
milliseconds. In English, oiceless s op consonan s /p, , k/ ha e a no iceable delay
be ween hei elease and he s a o ocal co d ib a ion. This delay, known as VOT,
las s a leas 30 milliseconds and co esponds o he pe iod o aspi a ion (Benkí, 2005).
P e ious s udies such as hose by Bond & Be mejo (1980), ound ha Spanish speake s
lea ning English end o p oduce /p, , k/ wi h VOT alues ha a e no ably sho e han
hose o na i e English speake s, especially i hey de elop "coexis en " phone ic sys ems.
The inabili y o L2 lea ne s o au hen ically p oduce English /p, , k/ in ea lie s udies
may be linked o he age a which he pa icipan s began lea ning English. Fokes &
S einbe g (1985) ound ha A abic-speaking child en be ween he ages o 2 and 11
p oduced /p, / in English wo ds wi h VOT alues simila o hose o na i e English
speake s (82 ms). Al hough hei speech was no di ec ly compa ed wi h ha o na i e
English child en o he same age, his sugges s ha hese A abic-speaking child en may

3
ha e been close o he English VOT pa e n han adul L2 lea ne s in ea lie s udies
(Flege & Ee ing, 1987).
The pu pose o he p esen s udy is o analyse and compa e he VOT o oiceless s ops
(/p, , k/) in English, as p oduced by Spanish-speaking lea ne s and na i e speake s o
English. To ha end, i e Spanish-speaking lea ne s o English as a second language and
i e na i e speake s o English, aged be ween 20 and 26, we e eco ded p onouncing 15
English wo ds, each beginning wi h one o he a o emen ioned phonemes. Once he da a
is collec ed, he esul s will be o ganized and p esen ed in a desc ip i e s a is ical able
o analysis. By examining VOT di e ences be ween he wo g oups, his s udy seeks o
p o ide a de ailed phone ic analysis o how Spanish lea ne s o English a icula e
oiceless s ops in compa ison o na i e speake s, con ibu ing o a be e unde s anding
o L2 phone ic acquisi ion. The cu en esea ch is pa o a p ojec ha desc ibes he
phone ic ea u es o English obs uen s p oduced by na i e speake s and Spanish lea ne s
o English (Ba e a-Pa do, in p epa a ion).
1. Resea ch ques ion:
To wha ex en does he VOT o English oiceless s op di e be ween na i e English
speake s and English lea ne s?
2 Hypo hesis:
Based on p e ious esea ch and indings om o he s udies, my hypo hesis o his s udy
is ha Spanish lea ne s o English will p oduce he oiceless s ops /p, , k/ wi h sho e
VOT alues han na i e English speake s (Flege & Ee ing, 1987). Due o he in luence
o he phone ic ea u es o hei mo he ongue, Spanish speake s a e expec ed o p oduce
VOT alues ha a e in e media e wi h espec o he VOT pa e ns o bo h languages. In
o he wo ds, since hei p oduc ion shows simila i ies o Spanish, hei adap a ion o
English is no ully comple e.
4
2. Me hod
2.1. Design
In e ms o he o mal design o my s udy, I ocused on wo ypes o a iables:
independen and dependen . A a iable is some hing ha can ake on di e en alues. I
is essen ial o dis inguish be ween hese wo ypes o a iables in o de o ob ain an
app op ia e in e p e a ion o da a. An independen a iable is cha ac e ised by he ac
ha i is no a ec ed by o he a iables in he s udy: i is au onomous wi h espec o he
o he a iables and can in luence he dependen a iable. In con as , a dependen a iable
ep esen s he ou come ha esea che s obse e, and is modi ied depending on he
independen a iables.
In he con ex o my esea ch, he dependen a iable co esponds o he VOT, gi en ha
i is al e ed be ween speake s and samples, and is measu ed in milliseconds. The
independen a iables a e s op (wi h h ee le els: /p/, / /, and /k/, which a e oiceless
s ops) and L1 (wi h wo le els: English and Spanish). These wo ac o s a e expec ed o
a ec he way VOT is p oduced in English. The aim is o examine whe he he place o
a icula ion o he oiceless s op, as well as he speake ’s na i e language, in luence he
du a ion o he VOT.
2.2. Pa icipan s
Rega ding he pa icipan de ails, i e na i e English speake s ook pa , consis ing o wo
women and h ee men (some o whom we e om Washing on, Cali o nia o Uni ed
Kingdom), and i e na i e Spanish speake s, all o whom we e women (coming om
Se ille, Spain). As men ioned ea lie , all pa icipan s we e be ween he ages o 20 and 26
yea s. The pa icipan cha ac e is ics a e p esen ed in Table 1:
5
Table 1. Speake s’ cha ac e is ics.
Speake
Gende
Age
L1
1
Female
21
English
2
Female
23
Spanish
3
Female
24
Spanish
4
Female
25
Spanish
5
Female
26
Spanish
6
Female
23
Spanish
7
Male
21
English
11
Female
20
English
12
Male
20
English
13
Male
21
English
2.3. Ma e ials
The s imuli o his expe imen consis ed o 15 English wo ds, each beginning wi h one
o he a ge oiceless s ops: /p/, / /, and /k/. The selec ed wo ds we e: pa , cod, ap, pi ,
kep , king, op, pu , ip, en, po , pe , ca , oo h, and cook. All o hem we e chosen o
ensu e ha he a ge consonan s appea ed in wo d-ini ial posi ion ollowed by a s essed
owel, which is ideal o measu ing VOT. Each pa icipan ead he ull wo d lis ou
imes, wi h he aim o ob aining comple e s a is ics and o a oid small mis akes by he
pa icipan s, gi ing hem mo e oppo uni ies. The s imuli we e p esen ed in w i en o m
on a p in ed lis du ing he eco ding session. The same se o wo ds was used o bo h
he Spanish lea ne s o English and he na i e English speake s, ensu ing consis en
condi ions o compa ison. As a esul , a o al o 600 measu emen s we e analyzed among
all pa icipan s.
6
2.4. P ocedu e
This expe imen was ca ied ou in se e al s ages. Fi s , we enlis ed he 10 pa icipan s
(including mysel ) o eco d he co esponding English wo ds in he eco ding s udio.
Finding he na i e Spanish speake s was ela i ely easy, bu i was mo e challenging o
loca e he na i e English speake s. I ook us longe o ind hem and o ge hem o ag ee
o pa icipa e in he eco dings. Once we had hem, he eco dings ook place in he
phone ics labo a o y a he Uni e si y o Se ille, whe e each speake illed ou a o m
answe ing ques ions such as hei gende , age, place o bi h, and language p o iciency.
The speake s' names we e kep comple ely anonymous. The nex s ep was o en e he
eco ding oom, whe e we pu on headphones and began eading he 15 English wo ds
con aining he sounds /p, , k/. The pa icipan s eco ded he wo ds ou imes in he ca ie
sen ence ________ is he wo d I say in he phone ics lab, and his p oduced 600 okens.
Each eco ding session las ed no mo e han 20 minu es. Once he eco dings we e
comple ed, we collec ed he VOT da a o each speake using he Mic oso Excel ool
(Mic oso Co po a ion, 2018), wi h he assis ance o PRAAT, which allowed us o
analyse and isualize he speech sounds om each o hei epe i ions.
2.5 Analysis
The ollowing sec ion p esen s analyses o he VOT o wo na i e English speake s and
wo Spanish lea ne s, using he so wa e PRAAT.
Figu e 1 shows an example o how VOT was ex ac ed om a speech sample, based on
he me hodology desc ibed in Cho & Lade oged (1999). This igu e shows a wa e o m
and spec og am o a na i e English speake p oducing he oiceless s op /k/ in wo d-
ini ial posi ion, ob ained using he so wa e PRAAT. This acous ic ep esen a ion
illus a es how VOT is measu ed — speci ically, he in e al be ween he elease bu s o
he s op and he onse o oicing, isible as he beginning o egula ocal old ib a ion.
In his case, he speake is saying he wo d cod.
The VOT can be seen in he ed shaded a ea, which is he momen when he elease o ai
occu s du ing he a icula ion o he /k/ sound ( isible as a sha p peak in he wa e o m),
and ends a he onse o ocal old ib a ion ha p oduces he con inuous owel, which
13
p o ides e idence on whe he he obse ed pa e ns in VOT can be ex ended o he
b oade popula ions o L1 English and L1 Spanish speake s.
The esul s o he ANOVA a e as ollows:
Table 3. ANOVA wi h VOT as he dependen a iable and s op and L1 as independen a iables
ANOVA – VOT
p
STOP
< 0.001
L1
< 0.001
STOP ✻ L1
< 0.001
The p in he able s ands o p obabili y, ha is, how likely i is ha he esul s ob ained
a e ep esen a i e o he popula ion (ins ead o he sample). In s a is ics, a p obabili y
smalle han 5% is accep ed as indica ing a easonable amoun o ce ain y ha he esul s
a e ep esen a i e o he popula ion. The esul s o he ANOVA o s op ( he i s
independen a iable) is p < 0.001, which is clea ly below 0.05 ( he 5% men ioned
ea lie ). This means ha he h ee s ops ha e s a is ically di e en VOT alues. The
esul s o he ANOVA o L1 ( he second independen a iable) is p < 0.001, which is also
clea ly below 0.05. This indica es ha he wo L1s ha e s a is ically di e en VOT alues.
Finally, he in e ac ion be ween he independen a iables is examined, which is he mos
ele an pa o he analysis. The wo independen a iables show a p alue o < 0.001,
which again sugges s ha he s ops ac oss bo h L1s exhibi a s a is ically signi ican
di e ence in VOT alues. The p og am hen compa es he in e ac ion o he wo
independen a iables, wi h he esul s plo ed in Table 4.

14
Table 4. Pos Hoc Compa isons – STOP ✻ L
The pos hoc compa isons u he e ealed speci ic di e ences be ween s op-L1
combina ions.
- The di e ence in VOT be ween L1 English /p/ and L1 Spanish /p/ is 29.41 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
- The di e ence in VOT be ween L1 English /p/ and L1 Spanish / / is -11.40 ms, and his
di e ence is s a is ically signi ican , since p = 0.014 (which is below 0.05).
Compa ison
STOP
L1
STOP
L1
Mean Di e ence
p
p
English
p
Spanish
29.41
< 0.001
English
-11.40
0.014
Spanish
-1.47
0.998
Spanish
English
-40.82
< 0.001
Spanish
-30.88
< 0.001
English
Spanish
9.94
0.047
k
English
k
Spanish
11.49
0.011
p
English
8.12
0.183
p
Spanish
37.54
< 0.001
English
-3.28
0.932
Spanish
6.66
0.388
Spanish
p
English
-3.37
0.928
p
Spanish
26.04
< 0.001
English
-14.77
< 0.001
English
Spanish
-4.84
0.728
15
- The di e ence in VOT be ween L1 English /p/ and L1 Spanish / / is -1.47 ms, and his
di e ence is no s a is ically signi ican , since p = 0.998 (which is much highe han 0.05).
- The di e ence in VOT be ween L1 English /p/ and L1 Spanish / / is -40.82 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
- The di e ence in VOT be ween L1 English /p/ and L1 Spanish / / is -30.88 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
- The di e ence in VOT be ween L1 English / / and L1 Spanish / / is 9.94 ms, and his
di e ence is s a is ically signi ican , since p = 0.047 (which is below 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish /k/ is 11.49 ms, and his
di e ence is s a is ically signi ican , since p = 0.011 (which is below 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish /p/ is 8.12 ms, and his
di e ence is no s a is ically signi ican , since p = 0.183 (which is g ea e han 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish /p/ is 37.54 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish / / is -3.28 ms, and his
di e ence is no s a is ically signi ican , since p = 0.932 (which is much highe han 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish / / is 6.66 ms, and his
di e ence is no s a is ically signi ican , since p = 0.388 (which is g ea e han 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish /p/ is -3.37 ms, and his
di e ence is no s a is ically signi ican , since p = 0.928 (which is much highe han 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish /p/ is 26.04 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
16
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish / / is -14.77 ms, and his
di e ence is s a is ically signi ican , since p < 0.001 (which is below 0.05).
- The di e ence in VOT be ween L1 English /k/ and L1 Spanish / / is -4.84 ms, and his
di e ence is no s a is ically signi ican , since p = 0.728 (which is g ea e han 0.05).
As o he di e ences be ween languages, he da a collec ed show ha na i e English
speake s end o p oduce longe VOTs han na i e Spanish speake s wi h espec o he
same sounds, indica ing ha aspi a ion is an aspec mo e commonly used in English. This
is e lec ed in he ollowing examples: he di e ence be ween /p/ in English and /p/ in
Spanish (29.42 ms), as well as be ween /k/ in English and /k/ in Spanish (11.49 ms), bo h
showing a signi ican dis inc ion be ween he wo sounds (p < 0.001 and p = 0.011). These
esul s ea i m ha English oiceless s ops a e p onounced wi h g ea e aspi a ion.
In con as , ocusing on simila i ies, in ce ain cases he VOT alues end o be simila
be ween he wo languages. The eason o his can be ound in sounds ha a e a icula ed
in simila pa s o he mou h, which end o be e y simila acous ically. Fo example, no
signi ican dis inc ion was ound be ween he VOT alues o /k/ in English and / / in
Spanish (p = 0.388), sugges ing ha hese sounds may be pe cei ed as simila .
4. Discussion
The esul s o he s udy e ealed ha Spanish-speaking lea ne s o English p oduced
sho e oice onse imes (VOT) o he oiceless s ops /p/, / / and /k/, compa ed o na i e
English speake s. This di e ence was no ably obse ed o he s ops /p/ and /k/, whe e
na i e speake s showed signi ican ly longe VOTs. Acco ding o Flege & Ee ing (1987),
he pa e n o English as a second language lea ne s e idences he impac o he mo he
ongue (L1) in luence, which leads Spanish-speaking lea ne s o app oach English
oiceless s ops wi h sho e VOTs, simila o hei na i e language, due o he
cha ac e is ics o he Spanish phonological sys em.
17
The da a also indica ed ha bo h he ype o s op and he speake ’s na i e language had a
clea in luence on VOT p oduc ion. In gene al, English speake s p oduced mo e s ongly
aspi a ed s ops, while Spanish speake s’ p oduc ions e lec ed pa e ns ypical o hei
i s language, wi h less aspi a ion. The mos no able di e ence was ound be ween he
English /p/ and he Spanish /p/, wi h a mean di e ence o app oxima ely 30 ms (p <
0.001). This clea ly e eals he in luence o he mo he ongue on he acquisi ion o a
second language. My esul s align wi h he s udy by Benkí (2005), who ound ha
Spanish speake s ecognize a consonan as oiceless wi h a sho e VOT han English
speake s, who need a longe VOT o di e en ia e hem. This shows ha mo he ongue
in luences he way hey pe cei e and p oduce sounds in English. E en hough hey ha e
been exposed o English o a long ime, Spanish-speaking pa icipan s s ill main ained
na i e-like pa e ns, which is why hey make less use o aspi a ion in oiceless
consonan s.
These esul s a e consis en wi h de elopmen al s udies showing ha younge speake s
ini ially p oduce s ops wi h sho lag VOTs and g adually de elop he a icula o y con ol
needed o p oduce longe lags o p e- oicing, as obse ed by Bond & Be mejo (1980),
who ound ha he abili y o p oduce adul -like VOT alues eme ges p og essi ely wi h
age and mo o de elopmen .
These indings di ec ly answe he esea ch ques ion by showing ha he e is indeed a
signi ican di e ence in he VOT o English oiceless s ops be ween na i e English
speake s and Spanish lea ne s. The s a is ical analysis con i med ha bo h s op ype and
L1 backg ound play a c ucial ole in shaping how hese sounds a e p oduced. Mo eo e ,
he o iginal hypo hesis is ully suppo ed by he esul s: Spanish speake s lea ning English
a e inclined o p oduce sho e VOTs han na i e speake s. This idea p oposes ha e en
highe -le el lea ne s may e ain phone ic ea u es o hei L1, especially in e ms o sub le
acous ic aspec s such as aspi a ion.
These esul s can con ibu e o a be e unde s anding o how p onuncia ion is lea ned and
can be imp o ed. As i has been p o en, one o he aspec s ha s ill pe sis s in na i e
Spanish speake s is he lack o ypical English aspi a ion, o example, hey s ill use
sho e VOT alues in English wo ds compa ed o na i e English speake s. One possible
imp o emen would be o each p onuncia ion ocusing mo e on he de ails ha lea ne s
ind mo e di icul , such as VOT di e ences. Teache s could make use o exe cises aimed
di ec ly a ge ing lea ne s o lis en o he ypical English VOT du a ion, and p oduce
18
hose sounds co ec ly. This would lead o mo e di ec and e icien lea ning wi h espec
o p onuncia ion. In addi ion, hey would be mo e awa e o how hei na i e language
in luences hem when lea ning a second language, and hus co ec ce ain common
e o s. Likewise, showing hem he sounds g aphically— o example by means o
spec og ams— hey could deepen hei unde s anding and make a mo e con olled use o
hei p onuncia ion, especially in he aspec s discussed.
4.1. My expe imen ’s con ibu ion o knowledge in his a ea o
phone ics
The indings o his s udy a e consis en wi h hose o p e ious esea ch on VOT
p oduc ion in Spanish speake s lea ning English. Fo example, Maldonado-Valen ín
(2007) ound ha Pue o Rican Spanish speake s wi h mo e ad anced le els o English
a e able o p oduce English oiceless s ops wi h VOT alues e y close o hose o na i e
speake s, while lea ne s wi h lowe p o iciency showed a no iceable in luence om hei
L1, esul ing in sho e VOTs. My esul s suppo his end: al hough all Spanish speake s
in my s udy we e ad anced lea ne s, hei p oduc ion o English /p/, / /, and /k/ s ill
showed conside ably sho e VOT alues han hose o na i e English speake s,
pa icula ly o /p/ and /k/. This con i ms ha L1 in luence can pe sis e en a highe
le els o p o iciency, and suppo s he idea ha L2 phone ic acquisi ion is p og essi e
and o en incomple e. Thus, his s udy suppo s p e ious indings and p o ides addi ional
e idence ha Spanish lea ne s end o e ain ea u es o hei L1 when p oducing oiceless
s ops in English.
Fu he mo e, he p esen s udy aligns wi h he indings o Flege & Ee ing (1988), who
ound ha Spanish speake s ended o p oduce s ops wi h sho e VOTs, usually wi hou
eaching he long VOT alues simila o hose o na i e English speake s. Thei esul s
indica e ha , al hough lea ne s may imp o e in p oducing new o ms in he second
language, hei p onuncia ion is s ill likely o be in luenced. My da a e lec his as well:
Al hough he Spanish speake s we e able o app oxima e he aspi a ed sounds in English,
hey did no exac ly eplica e he ypical phone ic pa e ns o na i e speake s. This
sugges s ha hei English p onuncia ion s ill e ains Spanish phone ic ea u es, which

19
may be because hey do no ye ha e a comple e unde s anding o how hese sounds
should be p onounced in he a ge language.
In addi ion, hese indings ein o ce he da a epo ed by Flege & Schmid (1995), who
analyzed how na i e English speake s and na i e Spanish speake s pe cei ed he
p onuncia ion o /p/ in English when p oduced wi h a ying ypes o aspi a ion, di e ing
in VOT. They ound ha Spanish lea ne s o English ga e highe a ings o s ops wi h
VOT alues simila o hose used by na i e speake s (a ound 50 ms), and also a ed
sounds wi h sho e aspi a ion imes mo e posi i ely han na i e speake s did. This
sugges s ha , al hough ad anced second language lea ne s ha e an easie ime pe cei ing
sounds simila ly o na i e speake s, hei e alua ion o hose sounds is s ill in luenced by
hei na i e language. My s udy shows simila esul s wi h espec o ha conclusion:
al hough Spanish-speaking pa icipan s had a highe le el o p o iciency, he VOTs hey
p oduced we e s ill sho e han hose o na i e speake s. This is e idence ha al hough
lea ne s acqui e new phone ic pa e ns speci ic o he second language, phone ic ea u es
o hei na i e language emain p esen in hei p onuncia ion.
Finally, hese esul s a e in line wi h he analysis by Shimizu (2011, Augus ), who ca ied
ou a compa a i e s udy in o de o analyze how Ko ean, Thai and Manda in Chinese
speake s make use o VOT o di e en ia e s op consonan s in hei na i e languages and
adap hose na i e pa e ns when lea ning English. Th ough he s udy, i was shown ha
hese lea ne s a e inclined o ela e hei na i e s op pa e ns o hose o he a ge language
based on VOT alues ha a e simila . Fo example, Ko ean speake s ela ed hei ense
s ops o English oiceless s ops —sounds p oduced wi hou ocal old ib a ion—e en
hough Ko ean ense s ops a e also cha ac e ized by g ea e a icula o y ension and a lack
o aspi a ion. A he same ime, Manda in and Thai speake s associa ed hei aspi a ed
s ops wi h English oiceless s ops, p obably because bo h languages use aspi a ion.
Al hough my s udy is ocused on Spanish speake s, he esul s show a simila pa e n: he
VOT alues o English oiceless s ops p oduced by he pa icipan s we e sho e , which
may ha e been in luenced by he sho e VOT du a ions ypical o hei L1. These da a
ein o ce he idea ha lea ne s employ he sound pa e ns o hei na i e language ha
esemble he sounds o he new language.
20
4.2. Faul s and limi a ions
As p e iously men ioned, his s udy o e s an analysis o VOT alues o English
oiceless s ops by na i e speake s and Spanish lea ne s. While he indings suppo he
in luence o he mo he ongue, se e al limi a ions mus be acknowledged.
One o he main limi a ions has o do wi h he s udy sample, which can be conside ed
small. This could educe he alidi y o he indings when ex ended o a la ge popula ion.
The da ase includes a simila numbe o examples o each oiceless consonan .
Howe e , he e is no consis en e idence o con i m ha hese da a ep esen a su icien ly
di e se sample o speake s, wi h espec o aspec s such as gende , dialec , and language
p o iciency le el. This could lead o doub s abou he eliabili y o he esul s ob ained in
he VOT p oduc ion.
Ano he limi a ion is he lack o a en ion o ce ain aspec s o speech, such as sen ence
s ess and speaking speed, which may in luence VOT. Howe e , conside able a en ion
was gi en o he posi ion o he a ge wo ds, which appea ed a he beginning o he
sen ence (e.g., ‘Ca is he wo d I say’). E en so, he uncon olled a iables may ha e
in oduced inconsis encies ha hide mo e sub le pa e ns in he da a.
In addi ion, he p esen s udy only elici ed a ge wo ds in u e ance-ini ial posi ion, while
p io esea ch (e.g., Flege, 1991) has shown ha VOT pa e ns may a y depending on
phone ic con ex . Including he a ge wo ds in o he posi ions (e.g., in he middle o he
u e ance) migh acili a e a be e unde s anding o how VOT p oduc ion a ies in
bilinguals depending on he wo d’s posi ion in he sen ence.
Fu he mo e, pa icipan s om bo h he Uni ed S a es and he Uni ed Kingdom we e
included in he na i e English speake g oup. Placing hem in he same g oup migh ha e
in oduced some a ia ion in o he esul s, since he e a e known di e ences in he
p oduc ion o oiceless s ops ac oss hese language a ie ies o English. Simila
limi a ions ha e been no ed in p e ious esea ch. Schoonmake -Ga es (2012) o ins ance,
ound ha VOT explained only a small pa o o eign accen pe cep ion, and emphasized
he impo ance o o he linguis ic cues and lis ene - ela ed ac o s. As in he
a o emen ioned s udy, his wo k does no conside pa icipan s' possible exposu e o he
di e en na i e a ie ies o English, no does i conside whe he hey ha e e e been
21
amilia wi h VOT as a phone ic cue. Analyzing hese ac o s would be qui e use ul in
u u e esea ch, as pa icipan s' p onuncia ion could be in luenced by hem.
To y o emedy hese limi a ions, u u e s udies should wo k wi h a la ge sample size
o ensu e an e icien and p opo ional ep esen a ion o speake a iables, such as
egional dialec s and assessing speake s’ p o iciency le els assessed h ough o icial
language p o iciency es s. Ano he op ion would be o dis inguish be ween dialec al
a ie ies such as Ame ican and B i ish English. This could educe a iabili y ha can
in e e e wi h he esul s and acili a e a be e unde s anding o how L1 in luences VOT
p oduc ion.
Ano he in e es ing esea ch app oach could be o examine phone ic elemen s in mo e
dep h beyond VOT, such as he aspi a ion noise o he VOT o he cha ac e is ics o he
oice when he ollowing owel is ini ia ed. These aspec s could help o ob ain a b oade
analysis on he p oduc ion and pe cep ion o oiceless s ops, as well as p o ide u he
insigh in o why accen pe cep ion a ies, as his is some hing ha canno be explained
by VOT du a ion alone.
Finally, i would be use ul o u u e esea ch o conduc long- e m s udies o gain a be e
and mo e in-dep h unde s anding o how VOT p oduc ion de elops as lea ne s each
highe le els o English. Resea che s could obse e he same g oup o lea ne s o e a
long pe iod o ime, pe haps mon hs o e en yea s, o see whe he hei p onuncia ion
becomes mo e like ha o na i e speake s as hey use he language egula ly. This would
make i possible o see whe he he pa e ns obse ed in his s udy change as he lea ne s'
le el imp o es wi h espec o hei con inued exposu e o he a ge language, o whe he
hose pa e ns a e only empo a y.
5. Conclusion
This s udy ocused on he analysis o he di e en alues o VOT du a ion o he English
oiceless s ops /p/, / / and /k/ p oduced by na i e speake s o English and Spanish-
speaking lea ne s o English wi h a high le el o p o iciency. The indings e lec ed ha
22
he VOTs p oduced by he Spanish lea ne s we e clea ly sho e in du a ion especially o
/p/ and /k/, u he sugges ing a s ong in luence o hei mo he ongue. Th ough
s a is ical analysis (ANOVA), i was co obo a ed ha he ype o consonan has a ele an
in luence on he p oduc ion o VOT, and also ha ea u es o he mo he ongue
pe sis en ly appea in he VOT pa e ns. These da a suppo he idea ha he phone ic
sys em o he mo he ongue has a s ong in luence on he acquisi ion o a second
language, e en in highly p o icien lea ne s.
Doing his esea ch pe sonally has been a e y consis en and en iching wo k, ull o new
lea ning and de ailed p ac ice in ela ion o phone ics. Lea ning so much abou he
di e ences in VOT p oduc ion be ween na i e speake s and mysel (as a Spanish-
speaking lea ne ) has been e y in e es ing, especially since I had no idea ha so much
ype o da a was collec ed and, abo e all, ha he e was such a ia ion in aspi a ion a he
onse o oiceless s op. I eally enjoyed disco e ing how much ou mo he ongue
in luences us, because wi hou his s udy, I wouldn' be awa e o all he de ails in ol ed
in second language p oduc ion.
I was pa icula ly cu ious when analyzing each eco ding and obse ing he VOT o each
oiceless s op, e en how sligh ly i a ied in he di e en eco dings o he same
indi idual. Al hough he e we e qui e a ew eco dings o analyze, being able o
ho oughly examine he p onuncia ions o na i e speake s was inc edible, one o he bes
expe iences I ha e had in my academic jou ney.
As o he di icul ies I ha e had, being able o ind na i e English speake s was a
somewha challenging ask. I spen days p epa ing pos e s and dis ibu ing hem in many
places, as well as walking a ound he s ee s and he uni e si y, asking people in pe son
i by any chance hey we e na i e English speake s. Some we e a li le su p ised a i s ,
al hough hey we e e y kind a e wa d.
Ano he challenge I aced du ing he s udy was ha i was some imes di icul o collec
he spec og am da a, as he VOT egion whe e he oiceless s op occu ed was no
always clea ly isible a he endpoin , so I hesi a ed wi h some da a, especially om he
lea ne s.