Jou nal o Memo y and Language 128 (2023) 104386
A ailable online 9 No embe 2022
0749-596X/© 2022 The Au ho (s). Published by Else ie Inc. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/).
In e e ence be ween non-na i e languages du ing ilingual
language p oduc ion
Angela de B uin
a
,
*
, Li J. Ho e s en
b
, Cla a D. Ma in
c
,
d
a
Depa men o Psychology, Uni e si y o Yo k, Yo k, UK
b
Depa men o Psychology, Uni e si y o Cali o nia, San a C uz, USA
c
Basque Cen e on Cogni ion, B ain and Language (BCBL), Donos ia-San Sebas i´
an, Spain
d
Ike basque, Basque Founda ion o Science, Bilbao, Spain
ARTICLE INFO
Keywo ds:
T ilingualism
Language p oduc ion
Language in e e ence
Inhibi ion
ABSTRACT
Mos esea ch on mul ilingual language con ol has ocused on a bilingual’s i s (L1) and second (L2) languages.
S udies on hi d language (L3) acquisi ion sugges ha , despi e he L1 being mo e p o icien , L3 lea ne s
expe ience mo e L2 han L1 in e e ence. Howe e , li le is known abou how a ilingual’s L2 and L3 in e ac
a e ini ial s ages o language lea ning. In he cu en s udy (Expe imen 1: 30 Spanish-Basque-English i-
linguals; Expe imen 2: 50 English-F ench-Spanish ilinguals), pa icipan s comple ed a speeded naming ask o
assess c oss-language in usions (e.g., using he Spanish “pe o” ins ead o he F ench “chien”). Bo h expe imen s
showed mo e L3 han L1 in usions du ing L2 naming. Fu he mo e, using wo di e en asks, we assessed i his
c oss-language in e e ence was ela ed o language inhibi ion. Bo h expe imen s sugges ed ha ilinguals
inhibi ed hei L1 mo e s ongly han hei L3. Toge he , his sugges s ha a ilingual’s non-na i e language
migh expe ience mo e in e e ence om ano he non-na i e language han om hei L1, possibly because
ilinguals apply mo e inhibi ion o e hei L1.
In oduc ion
While mos esea ch on mul ilingualism ocuses on a bilingual’s i s
(L1) and second language (L2), a la ge p opo ion o he wo ld popu-
la ion can hold a con e sa ion in mo e han wo languages (e.g., Eu o-
pean Commission, Special Eu oba ome e 386, 2012). S udies looking a
hi d language (L3) acquisi ion sugges ha while lea ning a new lan-
guage, he e migh be mo e in e e ence om he o he non-na i e
language (L2) han om he na i e language (L1), despi e he L1
being mo e p o icien (e.g., Puig-Mayenco, Gonz´
alez Alonso, and
Ro hman, 2020). Howe e , i emains la gely unknown how he lan-
guages o a ilingual in luence each o he once luency is eached in all
h ee languages (i.e., a e ini ial s ages o language lea ning). Ac oss
wo expe imen s, his s udy he e o e in es iga ed how a ilingual’s
na i e and non-na i e languages in e ac and compe e wi h each o he
du ing language p oduc ion.
In e ac ions be ween non-na i e languages
Resea ch looking a in e ac ions be ween na i e and mul iple non-
na i e languages has mainly ocused on people acqui ing a new lan-
guage (L3). Despi e he L1 ha ing a highe p o iciency han he L2,
se e al s udies ha e sugges ed L3 acqui e s migh be in luenced mo e (e.
g., using syn ac ic s uc u es om a known language in he L3) by he L2
han L1 (e.g., Ba del & Falk, 2007; Falk & Ba del, 2011; Ro hman &
Cab elli Ama o, 2010). In a ecen sys ema ic e iew, Puig-Mayenco
e al. (2020) examined he in luence om he L1 e sus L2 on hi d
language acquisi ion ac oss 71 s udies, ocusing on mo phosyn ax.
Twen y o he e iewed s udies showed in luence exclusi ely om he
L2 while his was he case o he L1 in only en s udies. While he e a e
many a iables ha could ha e an impac on L3 acquisi ion (wi h
ypological p oximi y being one o he key ac o s), his sugges s ha L3
acquisi ion migh expe ience mo e in luence om ano he non-na i e
language (L2) han om he na i e L1.
While mo phosyn ax has been he ocus o much o his esea ch,
he e is also some expe imen al e idence showing ha compe i ion
be ween non-na i e languages is also p esen a he lexical le el. Mic-
kan, McQueen, and Lemh¨
o e (2020) asked Du ch-English bilinguals o
lea n a se o Spanish wo ds. The nex day, pa icipan s we e asked o
name he same pic u es in ei he Du ch o in English. Pa icipan s hen
* Co esponding au o a : Depa men o Psychology, Uni e si y o Yo k, Yo k YO10 5DD, UK.
E-mail add ess: [email p o ec ed] (A. de B uin).
Con en s lis s a ailable a ScienceDi ec
Jou nal o Memo y and Language
jou nal homepage: www.else ie .com/loca e/jml
h ps://doi.o g/10.1016/j.jml.2022.104386
Recei ed 11 Ap il 2022; Recei ed in e ised o m 26 Oc obe 2022; Accep ed 28 Oc obe 2022
Jou nal o Memo y and Language 128 (2023) 104386
2
comple ed ano he naming ask o assess Spanish ecall immedia ely
a e he Du ch/English naming ask and one week la e . Spanish
naming (in e ms o accu acy and eac ion imes) was a ec ed nega-
i ely by Du ch and English naming. Howe e , his in luence was la ges
o he wo ds ha had been p e iously named in English (L2). This
sugges s ha a he lexical le el oo, a non-na i e language migh
expe ience mo e in luence om ano he non-na i e language han om
he na i e language.
One explana ion o hese indings du ing L3 acquisi ion has been
gi en in he o m o he “L2 s a us” hypo hesis (e.g., Ba del & Falk, 2007;
2012), which a gues ha he L2 and L3 a e mo e cogni i ely simila han
he L1 and L2/L3. T ilinguals migh acqui e he L2 and L3 in simila
en i onmen s (e.g., in a class oom as opposed o a home o in he
communi y) and in simila li e s ages (e.g., in la e childhood o adul -
hood as opposed o om bi h). The L1 and L2/L3 migh also di e in
hei eliance on p ocedu al e sus decla a i e memo y p ocesses, wi h
he la e po en ially being mo e impo an o languages acqui ed la e
in li e (e.g., Ba del & S´
anchez, 2017). Toge he , hese simila i ies in he
way non-na i e languages a e acqui ed (and po en ially used, i lan-
guage use is es ic ed o e.g., he class oom) could explain why he L2
migh in luence L3 acquisi ion mo e han he L1.
Mos o his esea ch has ocused on L3 acquisi ion, howe e , and
lea es open he ques ion how he languages o a ilingual in e ac once
a ce ain le el o luency is achie ed in all languages. The e is some
anecdo al e idence o sugges ha in e e ence con inues o exis be-
ween he non-na i e languages beyond ini ial s ages o acquisi ion. Fo
example, Williams and Hamma be g (1998) epo a case s udy wi h an
English-Ge man-Swedish ilingual who, when swi ching ou o hei L3
Swedish, swi ched mo e o en o hei L2 Ge man han o hei L1 En-
glish. Mo e ecen ly, Tomoschuk, Duyck, Ha suike , Fe ei a, and
Gollan (2021) p o ided expe imen al e idence showing simila pa -
e ns, wi h mo e in e e ence be ween non-na i e languages han be-
ween a na i e and non-na i e language. In hei Expe imen 1, hey
asked Du ch-English-F ench ilinguals (wi h a high p o iciency in hei
L1 and L2 bu a lowe p o iciency in hei L3) o comple e a phoneme
de ec ion ask. Pa icipan s had o indica e whe he a phoneme was
p esen in he wo d co esponding o a pic u e (e.g., /g/ when p esen ed
wi h a pic u e o a “gi l”). Impo an ly, pa icipan s we e also p esen ed
wi h phonemes ha we e no pa o he wo d in he a ge language bu
ha we e pa o he ansla ion equi alen (e.g., in he case o a pic u e
o a gi l, he phoneme /m/ is p esen in he L1 ansla ion equi alen
“meisje”). I he e is mo e in e e ence om a na i e language, L3 blocks
should see mo e alse ala ms om he L1 (i.e., saying ha he phoneme
is p esen in he a ge wo d because i is p esen in he L1 ansla ion
equi alen ) han om he L2. Howe e , i he e is mo e in e e ence
be ween non-na i e languages, he e should be mo e L2 han L1 alse
ala ms. The la e pa e n was ound, wi h mo e alse ala ms om L2
han L1 phonemes du ing he L3 ask, sugges ing ha he L3 expe ienced
mo e in e e ence om he L2 han om he na i e L1.
Language compe i ion and inhibi ion
Tomoschuk e al. (2021) sugges ha he inc eased in e e ence om
he L2 migh be ela ed o inhibi ion, wi h ilinguals inhibi ing hei L1
mo e s ongly o success ully han hei L2. This in e p e a ion is based
on G een’s Inhibi o y Con ol Hypo hesis (1998), which posi s ha bi-
linguals use inhibi ion o a oid in e e ence om he non- a ge lan-
guage(s) o allow o success ul p oduc ion in he a ge language.
Impo an ly, his hypo hesis a gues ha he amoun o inhibi ion
applied is ela i e o he p o iciency o he language, wi h bi-/ ilinguals
supp essing mo e p o icien languages mo e s ongly han less p o icien
languages.
E idence ha bilinguals migh inhibi hei L1 (mo e so han hei
L2) has been ound in a ange o pa adigms and echniques. Language
swi ching s udies (e.g., Meu e & Allpo , 1999) ha e sugges ed ha
unbalanced bilinguals (wi h a highe p o iciency in hei L1 han L2) can
show asymme ical swi ch cos s, wi h la ge cos s when swi ching back
o hei L1 han L2. This migh be he consequence o bilinguals applying
ela i ely la ge amoun s o L1 inhibi ion du ing L2 naming, hus
equi ing mo e ime o elease his inhibi ion when swi ching back o he
L1, leading o la ge swi ching cos s. The inding ha mo e inhibi ion is
applied o e he L1 when swi ching o he L2 han ice e sa is also
suppo ed by neu oimaging s udies showing inc eased ac i a ion in
b ain egions and ERP ma ke s associa ed wi h inhibi ion when
swi ching o an L2 (i.e., he momen L1 inhibi ion would need o be
applied) han when swi ching o an L1 (i.e., when less (L2) inhibi ion is
needed; de B uin, Roelo s, Dijks a, & Fi zPa ick, 2014; Jackson,
Swainson, Cunning on, & Jackson, 2001). O he s udies no using
language-swi ching asks ha e also sugges ed ha L1 wo ds a e less
accessible a e L2 naming due o L1 inhibi ion du ing L2 naming. Fo
example, Mis a, Guo, Bobb, and K oll (2012) assessed po en ial e ec s o
epe i ion p iming (i.e., as e naming when i ems ha e been named
be o e) when naming in he L2 a e he L1 and in he L1 a e he L2. As
expec ed, epe i ion p iming was obse ed when he L2 was used a e
he L1. Howe e , using he L1 second showed no beha iou al epe i ion
bene i and an inc eased N2 (an ERP componen associa ed wi h op-
down language con ol), sugges ing ha he L1 was supp essed du ing
L2 naming.
In e e ence and inhibi ion ha e also been s udied h ough pa a-
digms elici ing language in usions. Fo example, Gollan, Scho e ,
Gomez, Mu illo, and Rayne (2014) asked Spanish-English bilinguals o
ead pa ag aphs in one language o mixing bo h languages aloud. In he
ex s wi h bo h Spanish and English wo ds, pa icipan s made c oss-
language in usions (e.g., saying an English wo d ins ead o he Span-
ish wo d), in pa icula when ying o ead wo ds in he mo e dominan
language. This again sugges s ha he mo e dominan /p o icien lan-
guage migh be supp essed and consequen ly less accessible and mo e
p one o in usions. O he pa adigms oo ha e sugges ed ha c oss-
language in usions du ing p oduc ion a e sensi i e o language con-
ol and in e e ence (e.g., Decle ck, G ainge , & Ha suike , 2021;
Zheng, Roelo s, & Lemh¨
o e , 2020), al hough hey do no always e eal
di e ences be ween languages (e.g., Decle ck e al., 2021).
Le y, McVeigh, Ma ul, & Ande son (2007) also examined he
accessibili y o L1 e sus L2 wo d o ms, by using a e ie al-induced
o ge ing pa adigm, a ask we used in Expe imen 1 oo and which
we will e e o as he “ hyme ask”. In Le y e al.’s s udy, English-
Spanish bilinguals we e asked o gene a e English hyme wo ds. Pa -
icipan s saw a p obe (e.g., “spoon”) and had o gene a e one wo d ha
hymed wi h he p obe (e.g., “moon”). P io o he hyme ask, pa ic-
ipan s named pic u es in English o in Spanish. The names o hose
pic u es could be used as hyme wo ds in he hyme ask (e.g., pa ici-
pan s named a pic u e o a moon, wi h “spoon” being he p obe in he
hyme ask). Naming hose pic u es in English inc eased he chance o
using hose wo ds in he English hyme ask (i.e., pic u e names p o-
duced in English we e mo e likely o be used as hyme esponses i hose
wo ds had been epea ed mo e o en in he pic u e-naming ask). In
con as , he e e se pa e n was ound o wo ds p oduced in Spanish
(L2): he mo e o en a pic u e had been named in Spanish p io o he
hyme ask, he less likely pa icipan s we e o use he English (L1)
ansla ion equi alen in he hyme ask. This sugges s ha pa icipan s
had supp essed he L1 equi alen s du ing L2 naming, hus making hose
L1 ansla ion equi alen s less accessible in he ollowing hyme ask
(bu c . Runnq is & Cos a, 2012 o di e ging indings).
Cu en esea ch
Following hese indings, unbalanced ilinguals migh supp ess hei
mo e p o icien na i e language mo e s ongly han hei less p o icien
non-na i e languages. This, in u n, could be p edic ed o educe in e -
e ence om a na i e language. This p edic ion is also suppo ed by
s udies sugges ing ha L3 lea ning is mo e success ul h ough L1 in-
s uc ion han h ough L2 ins uc ion, po en ially because o be e
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
3
egula ion and inhibi ion o he L1 (e.g., Bogulski, Bice, & K oll, 2019;
Hi osh & Degani, 2021). Howe e , while he e is some e idence o
sugges ha ilinguals migh expe ience mo e in e e ence be ween
non-na i e languages han om a na i e language (Tomoschuk e al.,
2021), mos esea ch is based on L3 acquisi ion. Fu he mo e, Tomo-
schuk e al. (2021) showed in e e ence in a phoneme de ec ion ask bu
i is unknown how po en ial in e e ence be ween non-na i e languages
can ac ually in luence language p oduc ion. Las ly, while L1 inhibi ion
has been p oposed as a po en ial unde lying mechanism o explain
educed L1 in e e ence ela i e o L2/L3 in e e ence (Tomoschuk
e al., 2021), he ole o language inhibi ion has no been assessed
di ec ly and we do no know i and how language inhibi ion explains
po en ial in e e ence be ween languages du ing ilingual p oduc ion.
The cu en esea ch he e o e i s ly aimed o examine how he wo
non-na i e languages o a ilingual in e ac wi h each o he beyond
ini ial s ages o language acquisi ion, when ilinguals ha e al eady
achie ed an in e media e p o iciency in hei non-na i e languages.
Speci ically, we add essed in e e ence be ween languages in he o m o
c oss-language in usions, such as saying he Spanish wo d “caballo”
when he Basque “zaldi” is needed. Second, we examined why ilinguals
migh expe ience in e e ence be ween non-na i e languages by
assessing he ole o L1 e sus L3 inhibi ion. Ac oss wo expe imen s
wi h ilinguals wi h di e en language combina ions, we he e o e used
a speeded pic u e-naming ask o s udy in e e ence om he L1 e sus
L3 on he L2 du ing ilingual language p oduc ion. We also examined
how ilinguals inhibi ed hei L1 e sus L3 du ing L2 p oduc ion
h ough a hyme ask (Expe imen 1) and an n-2 swi ching ask
(Expe imen 2).
Expe imen 1
In oduc ion
To assess language in e e ence, Expe imen 1 asked pa icipan s o
comple e a language-swi ching ask in h ee languages (Spanish-L1,
Basque-L2, and English-L3). We we e especially in e es ed in elici ing
language in usions (e.g., using a Spanish o English wo d when a Bas-
que wo d was in ended). C oss-language in usions can be a mo e di ec
and conc e e measu e o language in e e ence as hey e eal no iceable
mis akes in ilingual language selec ion. As such, his allowed us o
s udy which language (L1 o L3) in e e ed mo e wi h L2 p oduc ion. We
he e o e de eloped a speeded pic u e-naming pa adigm in which pa -
icipan s had o al e na e languages in a ask p esen ing each pic u e o
less han one second. I ilinguals expe ience mo e in e e ence be ween
non-na i e languages, despi e a highe p o iciency le el in hei na i e
language, we hypo hesised ha hey should p oduce mo e L3 han L1
in usions du ing L2 a ge naming.
As a second aim, we wan ed o examine he ole o inhibi ion du ing
ilingual language con ol. To his end, we adjus ed he e ie al-
induced o ge ing pa adigm ( hyme ask) used by Le y e al. (2007).
In his ask, pa icipan s gene a ed a hyme wo d in esponse o a p obe
(e.g., “sheep” in esponse o he p obe “jeep”). They comple ed his ask
in Spanish (L1) and in English (L3) be o e and a e a ilingual naming
ask (a di e en naming ask han he speeded swi ching ask). Du ing
he ilingual naming ask, hey named some pic u es in hei L1, some in
hei L2, and some in hei L3. Du ing he hyme ask, he p obe hymed
wi h wo ds used in he naming ask (wi hin-language ials) o hei
ansla ion equi alen s (ac oss-language ials, see Fig. 1 and he
Me hods sec ion o u he in o ma ion). Fo example, pa icipan s
could see he English p obe “cake” in he hyme ask p e- and pos -
naming. Du ing he naming ask, hey would be asked o name pic-
u es o a snake in English, hus inc easing he likelihood o pa icipan s
saying “snake” in he pos -naming hyme ask (compa ed o he baseline
p e-naming hyme ask). These so-called “wi hin-language” con ol i-
als we e used o bo h English-L3 and Spanish-L1 (e.g., in Spanish, a
hyme p obe could be “ ama”, wi h pa icipan s using “cama” o name
he pic u e o a bed in he naming ask). We expec ed wi hin-language
a ge s (i.e., “snake” and “cama”) o be used mo e o en in he hyme
ask pos - han p e-naming due o hei ecen use in he naming ask.
Howe e , o main in e es , we also included “ac oss-language” a ge s.
These we e L1/L3 esponse a ge s o he hyme p obe ha we e
ansla ion equi alen s o wo ds ha had jus been named in he L2 in
he naming ask. Fo example, pa icipan s would need o name a pic u e
o a glass in Basque (“edalon zi”) in he naming ask and would see he
p obe “mass” in English ( hyming wi h “glass”) and “paso” in Spanish
( hyming wi h “ aso”, meaning glass). In he p e-naming hyme ask we
did no expec di e ences be ween languages gi en ha he wo ds had
no ye been used in he naming ask. Howe e , in he pos -naming
Fig. 1. O e iew o he asks in each session in Expe imen 1.
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
4
hyme ask we expec ed language di e ences on he ac oss-language
a ge s. Speci ically, i ilinguals inhibi L1 equi alen s mo e han L3
equi alen s du ing L2 naming, we would expec hem o use hose L1
equi alen s less o en han he L3 equi alen s du ing he pos -naming
hyme ask. Fo example, i a ilingual names a pic u e o a glass in
L2 (“edalon zi”) and concu en ly supp esses L1 “ aso” mo e han L3
“glass”, we would expec hem o use “ aso” less o en in esponse o he
L1 hyme p obe “paso” han hey would use “glass” o he L3 p obe
“mass”.
Expe imen 1 ook place in he Basque Coun y. Pa icipan s had a
lowe p o iciency in hei L2/L3 han L1 and acqui ed only hei L1 om
bi h. Howe e , hei language en i onmen di e s om he ype o
ilinguals/L3 acqui e s es ed in p e ious s udies. Fo example, in many
o he s udies (e.g., Tomoschuk e al., 2021), he non-na i e languages
we e bo h acqui ed as class oom languages. Acco ding o he L2 s a us
hypo hesis, in e e ence be ween non-na i e languages could be
explained by cogni i e simila i y as a consequence o he L2 and L3
being acqui ed in a class oom. In he Basque Coun y, howe e , he
pa icipan s’ L2 (Basque) is a communi y language while he pa ici-
pan s’ L3 (English) is ea ed as a o eign language ha is la gely
es ic ed o he class oom. I he ilinguals in Expe imen 1 show mo e
L3 han L1 in usions, his would sugges ha his in e e ence is no
pu ely due o he way he L2 and L3 we e acqui ed and a e used in
simila (class oom) en i onmen s.
Fu he mo e, mos pa icipan s ecei ed hei (p ima y and sec-
onda y) educa ion in Basque (L2) o a dual Basque-Spanish sys em.
Tomoschuk e al. (2021, Expe imen 2) sugges ed ha language o in-
s uc ion can modula e he amoun o non-na i e language in e e ence.
In e e ence di e ences be ween he L1 and L2 we e only ound i new
L3 wo ds we e augh in he L1, bu no when he ins uc ion language
was L2. This aises he ques ion whe he ilinguals indeed expe ience
mo e in e e ence om hei non-na i e language in gene al o whe he
hese indings a e speci ic o he L1 being used as he language o in-
s uc ion in he class oom. I ilinguals in ou Expe imen 1 show mo e
L3 han L1 in usions (despi e he L3 being augh in an L2 o bilingual
school en i onmen ), his would sugges ha hey a e be e a egu-
la ing in e e ence om hei na i e language e en i he L1 was no he
main/only language o ins uc ion.
In addi ion, al hough hei L2 Basque had a much lowe p o iciency
and use han hei L1 Spanish, mos pa icipan s acqui ed Basque wi hin
he i s h ee yea s o li e, making i less likely ha he L2 was acqui ed
mos ly h ough decla a i e memo y p ocesses. Acco ding o he L2
s a us hypo hesis (e.g., Ba del & S´
anchez, 2017), eliance on mo e
decla a i e han p ocedu al memo y sys ems o he L2 and L3 migh
explain he in e e ence be ween he wo. Howe e , due o he way
Basque is acqui ed du ing ea ly childhood and is a communi y language
while English is la gely used as a class oom language, he L2 and L3 in
his s udy a e unlikely o be bo h elying p edominan ly on decla a i e
memo y.
Finally, ypological p oximi y has been a gued o be a key a iable in
language in e e ence, wi h mo e in e ac ions be ween languages ha
a e mo e simila (e.g., Puig-Mayenco e al, 2020). Basque (L2) di e s
subs an ially om bo h he L1 and L3 in aspec s such as ocabula y and
mo phosyn ax. Howe e , i any hing, i is mo e simila o he L1 han L3
in e ms o o hog aphy and phonology. Thus, i ou ilingual pa ici-
pan s expe ience mo e in e e ence om he L3 han L1, his is unlikely
o be due o ypological p oximi y be ween he L2 and L3.
Taking in o conside a ion his di e en language p o ile compa ed o
p e ious L3 (acquisi ion) s udies, Expe imen 1 i s ly aimed o assess
how much in e e ence ilinguals expe ience be ween wo non-na i e
languages as compa ed o be ween a na i e and non-na i e language
using he speeded naming ask. To c ea e he la ges di e ence be ween
he na i e and non-na i e language (in e ms o p o iciency, use, mode
o acquisi ion, e ce e a), we compa ed L1 e sus L3 in usions du ing L2
p oduc ion. Nex , we assessed how inhibi ion migh be in ol ed when
ilinguals manage in e e ence be ween hei h ee languages using he
hyme ask.
Da a a ailabili y
The da a and analysis sc ip ( o bo h Expe imen s) a e a ailable on
h ps://os .io/wmehd/. The s imuli a e p o ided in he appendices.
Me hods
Pa icipan s
The inal da ase included hi y Spanish-Basque-English ilinguals
(23 emale, Mage =23.3; SDage =5.6). We speci ically ec ui ed pa -
icipan s who acqui ed only one language om bi h (Spanish) and who
had an in e media e p o iciency le el in Basque and English. We ensu ed
ha pa icipan s we e p o icien in all languages ( hus ocusing on i-
linguals a he han L3 acqui e s) bu had a much lowe use o and
p o iciency in hei L2/L3 han hei L1 ( hus c ea ing a clea dis inc ion
in p o iciency and use be ween he na i e L1 and non-na i e L2 and L3).
Two addi ional pa icipan s we e es ed bu excluded om da a analysis.
One pa icipan was excluded because hey did no know o e hal o he
English a ge wo ds (assessed in a pos -expe imen su ey). The o he
pa icipan was excluded because hey did no p oduce any hyme wo ds
on mo e han hal o he ials in he baseline hyme ask. The use o
no el asks (a speeded swi ching ask o elici in usions and a ilingual
Table 1
Summa y o he objec i e and subjec i e measu es o language p o iciency, language exposu e, and language use o Spanish, Basque, and English.
Spanish Basque English
Mean SD Range Mean SD Range Mean SD Range
Age o Acquisi ion 0 0 0–0 2.9 0.8 2–6 6.1 2.0 2–11
Pic u e naming (0–65) 64.8 0.5 64–65 42.6 6.7 30–51 43.8 8.0 31–56
LexTALE (% wo ds co ec )
1
88.7 8.0 83–100 67.4 16.4 28–94 43.1 22.2 3–100
In e iew (1–5)
2
5 0 5–5 3.3 0.4 3–4 3.0 0.6 2–4
Sel - a ed p o iciency
2
(0–10)
Speaking 9.6 0.7x 7–10 6.9 1.5 4–10 6.0 1.7 1–9
Unde s anding 9.7 0.7 8–10 8.3 1.3 6–10 7.0 1.5 3–9
W i ing 9.3 1.0 7–10 7.3 1.7 4–10 6.3 1.9 2–9
Reading 9.4 1.2 5–10 7.9 1.7 5–10 7.0 1.7 3–9
Gene al 9.4 0.8 8–10 7.0 1.4 4–10 6.3 1.6 2–9
%exposu e
(0–100) 64.3 13.3 50–90 24.6 14.3 0–40 10.4 7.4 0–30
%speaking
(0–100) 74.3 16.5 50–100 19.3 14.6 0–40 15.2 12.4 0–40
1
Da a om one pa icipan missing o Basque and English. LexTALE sco e is calcula ed as he pe cen age o wo ds iden i ied co ec ly.
2
Da a a e missing o wo pa icipan s.
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
5
adap a ion o a hyme ask) mean ha we did no ha e exis ing da a on
which o base a powe analysis. Fu he mo e, he speci ic language
p o ile es ic ed us in he numbe o pa icipan s we could ec ui .
Thi y pa icipan s was he e o e se as he a ge size o ensu e
ec ui men easibili y while aiming o ha e as much powe as possible.
All pa icipan s had no mal o co ec ed- o-no mal ision and no known
neu ological, eading, o hea ing impai men s. They p o ided w i en
in o med consen and he s udy was app o ed by he BCBL E hics Re-
iew Boa d and complied wi h he guidelines o he Helsinki
Decla a ion.
All pa icipan s acqui ed Spanish as hei i s language, Basque as
hei second language, and English as hei hi d language. They we e
li ing in a bilingual socie y in which bo h Spanish and Basque a e
commonly used. Pa icipan s had an in e media e p o iciency in bo h
Basque and English and p edominan ly used Spanish on a daily basis
(see Table 1). Thei language p o ile was assessed h ough a se o
objec i e and subjec i e language measu es in Spanish, Basque, and
English when hey signed up o he da abase (de B uin, Ca ei as, &
Du˜
nabei ia, 2017). The objec i e p o iciency measu es include a 65-
i em pic u e naming ask, he LexTALE (a sho lexical decision ask;
Lemh¨
o e & B oe sma, 2012) and an in e iew. In addi ion, pa icipan s
p o ided sel - a ings o language p o iciency, use, and exposu e. The
esul s om hese asks and measu es a e epo ed in Table 1. Pa ici-
pan s also comple ed a sho su ey a he end o he s udy asking abou
hei language(s) o educa ion and ime spen in English-speaking
coun ies. Mos pa icipan s (18) a ended a bilingual Spanish-Basque
school sys em; 10 pa icipan s comple ed hei educa ion in Basque;
wo pa icipan s comple ed hei educa ion in Spanish. Mos pa ici-
pan s (19) had no spen any signi ican ime in an English-speaking
coun y apa om holidays. Ele en pa icipan s indica ed ha ing
spen some ime in an English-speaking coun y (M numbe o mon hs =
4.2, SD =4.2).
Design o he wo main asks
Speeded ilingual naming ask. In he speeded ilingual naming ask,
pa icipan s we e asked o name pic u es in Spanish (L1), Basque (L2),
o English (L3) in esponse o a coun y lag. A speeded ask (in which
pa icipan s saw each pic u e o only 900 ms) was used o c ea e mo e
ime p essu e and hus o induce mo e e o s. We ocused on he numbe
o Spanish (L1) e sus English (L3) in usions du ing non-swi ch ials
ha we e supposed o be named in Basque (L2).
Rhyme ask. In he hyme ask, pa icipan s we e asked o gene a e
hyme wo ds in esponse o Spanish (L1) and English (L3) p obes. This
ask was comple ed wice: once a he beginning o he s udy (as he
baseline, we will e e o his as “p e-naming”) and once a he end o he
s udy a e a naming ask (“pos -naming”; see Fig. 1). In he in e ening
naming ask (a di e en one han he speeded naming ask) be o e he
hyme ask, people named wo ds in ei he L1, L2, o L3. Wo ds named in
he L1 and L3 in ha ask we e possible hyme esponses (i.e., a ge s) in
he L1/L3 hyme ask (e.g., “snake” had o be named in English in he
naming ask and “cake” was a p obe in he hyme ask). We e e o his
condi ion as “wi hin-language”. The wo ds named in L2 we e ansla ion
equi alen s o possible hyme esponses in he L1/L3 hyme ask (e.g.,
“duck” named in Basque in he naming ask; p obe “ uck”; see Fig. 2).
We e e o his condi ion as “ac oss-language”. The e we e hus h ee
wi hin-subjec independen a iables in he hyme ask: Language (L1/
L3); Session (p e-naming/pos -naming); Condi ion (wo ds named in he
same language “wi hin-language”/wo ds named in Basque “ac oss-
language”).
Session 1:
Ma e ials
We selec ed six y pic u es om he Mul iPic da abase (Du˜
nabei ia
e al., 2018). Eigh o hese pic u es we e used in he speeded ilingual
naming ask and had o be named in he h ee languages in e change-
ably. We used a small se o pic u es in his ask o inc ease ac i a ion o
each wo d, wi h he o e all aim o inc easing compe i ion be ween
wo ds and elici ing mo e in usions.
All six y pic u es we e used in he slow naming ask p eceding he
pos -naming hyme ask ( wen y named in Spanish se ing as L1 wi hin-
language a ge s in he hyme ask; wen y named in English se ing as
L3 wi hin-language a ge s in he hyme ask, and wen y named in
Basque se ing as bo h L1 and L3 ac oss-language a ge s in he hyme
ask). All pic u e names we e non-cogna es and had a high equency.
The eigh wo ds used in he speeded naming ask we e ma ched on
English and Spanish equency and on numbe o le e s and phonemes;
Spanish wo ds had mo e syllables han English wo ds (see Appendix A).
We did no ma ch he Basque wo ds gi en ha we we e ocusing on he
numbe o English (L3) e sus Spanish (L1) in usions bu did ma ch he
English and Spanish wo ds on hei Le ensh ein dis ance o Basque
wo ds (i.e., he numbe o le e s ha would need o be changed o o m
he Basque wo d). Those eigh wo ds we e chosen ( om he ac oss-
language i ems) because hey di e ed in hei onse in he h ee lan-
guages. This allowed us o sco e c oss-language in usions e en i only
one phoneme was p oduced (e.g., i he ini ial esponse was co ec ed
a e he i s phoneme).
Fo he hyme ask, we ma ched he di e en condi ions (wi hin-
Fig. 2. O e iew o he condi ions in he hyme ask.
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
6
English e sus wi hin-Spanish; ac oss-English e sus ac oss-Spanish;
wi hin-English e sus ac oss-English; wi hin-Spanish e sus ac oss-
Spanish) on a ange o measu es including a ge equency, he num-
be o po en ial hyme wo ds o each p obe, and he numbe o al e -
na i e hyme wo ds wi h a highe equency han he a ge (see
Appendix A). Due o gene al wo d leng h di e ences be ween English
and Spanish wo ds, he English and Spanish wo ds could no be ma ched
in e ms o numbe o syllables, le e s, and phonemes. Howe e , we did
ensu e ha he Spanish-wi hin and Spanish-ac oss as well as he English-
wi hin and English-ac oss condi ions we e ma ched on wo d leng h. We
selec ed a ge wo ds ha had a ew high- equency hyme compe i o s
bu no oo many ( o make su e ha a ge wo ds we e ela i ely likely
o be p oduced wi hou he a ge being he only op ion). Fu he de ails
abou he p obes a e gi en in Appendix A.
P ocedu e
The s udy consis ed o wo in-pe son sessions (see Fig. 1), sepa a ed
by an in e al o app oxima ely one o wo weeks (M =12 days). We
op ed o use wo sepa a e sessions o sepa a e he baseline and pos -
naming hyme asks ( o a oid pa icipan s emembe ing exac ly which
answe s hey ga e in he baseline ask) and o sepa a e he wo naming
asks. In he i s session, pa icipan s s a ed wi h he baseline p e-
naming hyme condi ion. Nex , hey comple ed he speeded ilingual
naming ask assessing c oss-language in usions du ing L2 p oduc ion.
In he second session, pa icipan s i s comple ed he slow ilingual
naming ask ha was ela ed o he hyme ask. Nex , hey comple ed he
pos -naming hyme ask. A e he second session, pa icipan s we e
asked o indica e whe he he e we e any English wo ds hey did no
know be o e he s a o he s udy. On a e age, pa icipan s indica ed
knowing 38 o he 40 English hyme a ge s ( ange 36–40; da a missing
om h ee pa icipan s).
P e-naming hyme ask (baseline)
In he hyme ask, pa icipan s we e p esen ed isually and au ally
wi h a p obe wo d. They we e asked o gene a e a wo d ha hymed
wi h he p obe. Fo example, a pa icipan could be p esen ed wi h ‘wig’
and could espond wi h ‘pig’, ‘dig’, e c. Pa icipan s we e ins uc ed o
say he i s hyme wo d ha came o mind, bu no o use any p ope
names o names o coun ies o places. They we e gi en en seconds o
each p obe wo d. The ask was comple ed in Spanish and English in wo
sepa a e language blocks, wi h he o de o languages coun e balanced
ac oss pa icipan s. Each language included wen y wi hin-language
ials and wen y ac oss-language ials (see “slow naming ask”). The
hyme ask was comple ed a he s a o he i s session as he baseline
o he hyme ask a he end o he second session.
Speeded ilingual naming ask
Pa icipan s we e i s amilia ised wi h he 8 pic u es and wo ds and
we e asked o ead each wo d aloud. Pa icipan s we e hen p esen ed
wi h a speeded naming ask showing pic u es accompanied by he
Spanish, Basque, o B i ish lag. Each pic u e was p esen ed below he
coun y lag o 900 ms and pa icipan s we e ins uc ed o name he
pic u e in he indica ed language wi hin ha ime ame. A blank sc een
was shown o one second be o e he nex pic u e was p esen ed. We
eco ded he ull 1900 ms and sco ed esponses ha we e gi en du ing
he en i e in e al, e en hough pa icipan s we e ins uc ed o name he
pic u e while he pic u e emained on he sc een.
In o al, pa icipan s named 720 expe imen al ials (240 Basque
non-swi ch ials, 160 Basque swi ch ials, 40 English non-swi ch ials,
120 English swi ch ials, 40 Spanish non-swi ch ials, and 120 Spanish
swi ch ials). Pa icipan s named an addi ional en ials ha we e
p eceded by a b eak and we e no included in he analysis. We ocused
on he 240 Basque non-swi ch ials o examine he numbe o L1
(Spanish) e sus L3 (English) in usions. C ucially, hese L2 non-swi ch
ials we e always p eceded by ano he L2 ial, hus emo ing he im-
media e in luence o jus ha ing used ano he language. L2 swi ch ials
we e p eceded an equal numbe o imes by each o he languages (e.g.,
80 Basque swi ch ials we e p eceded by a Spanish ial and 80 Basque
swi ch ials we e p eceded by an English ial). L1 and L3 swi ch ials
we e p eceded 80 imes by a Basque ial and 40 imes by an L3 o L1
ial espec i ely.
Session 2:
Slow naming ask:
The slow naming ask a he s a o he second session was used o
pu hyme a ge s in he “wi hin-language” o “ac oss-language” con-
di ion o he pos -naming hyme ask. Be o e s a ing he naming ask,
pa icipan s we e exposed o all six y pic u es and hei names in he
h ee languages o ensu e pa icipan s we e amilia wi h he esponses.
In he naming ask, each pic u e had o be named eigh imes and was
shown on he sc een o 2 s below he coun y lag. Twen y o hese
pic u es always had o be named in Basque (L2 – ac oss-language ials
o he L1 and L3 hyme ask), wen y always had o be named in English
(L3 – wi hin-language ials o he L3 hyme ask), and wen y always
had o be named in Spanish (L1 – wi hin-language ials o he L1 hyme
ask). The h ee languages had o be used in e changeably o inc ease
compe i ion be ween he languages. The e we e 488 ials (8 ials
p eceded by a b eak; he 480 ials we e dis ibu ed equally ac oss
languages and swi ch ype). Each pic u e was p esen ed ou imes as a
swi ch ial ( wice p eceded by each o he wo languages) and ou
imes as a non-swi ch ial.
Pos -naming hyme ask:
The hyme ask om he s a o he i s session was comple ed
again, using he same p obes and s uc u e.
Analysis
The da a a e a ailable on h ps://os .io/wmehd/ and we e analysed
using gene alised linea mixed-e ec s models using lme4 package
e sion 1.1–21 in R 3.6.1.
Speeded ilingual naming ask. In he speeded ilingual naming ask, we
sco ed accu acy on each ial as A) no esponse, B) a c oss-language
in usion (e.g., English ins ead o Basque wo d), C) a wi hin-language
in usion (e.g., ‘cloud’ ins ead o ‘moon’), o D) ano he esponse ha
was no he in ended a ge bu ha had a simila meaning (e.g., ‘pa ed’
ins ead o ‘mu o’, wi h bo h being Spanish wo ds o ‘wall’). We only
sco ed he i s esponse. Fo example, i mul iple c oss-language in-
usions we e made (e.g., English and hen Spanish whe e a Basque wo d
was equi ed), he i s was sco ed. Simila ly, a esponse coun ed as an
in usion ega dless o how much o he in usion was p oduced (e.g., i
jus he i s phoneme o he in usion was p oduced and hen co ec ed,
i s ill coun ed as an in usion).
We we e mainly in e es ed in he numbe o L1 Spanish e sus L3
English in usions p oduced du ing L2 Basque non-swi ch ials. Ou
main analysis he e o e only included L2 non-swi ch ials. Using
gene alised linea mixed-e ec models (glme ; pa icipan and i em in-
e cep s included as andom e ec s), we compa ed he numbe o L1
e sus L3 in usions by using numbe o c oss-language in usions (i.e.,
accu acy ype B) as he dependen a iable (DV) and he language o
in usion (Spanish =-0.5; English =0.5) as he ixed e ec . We used he
“poisson” dis ibu ion wi hin he glme , which is sui able o coun da a
(numbe o in usions in Spanish o English) and is based on he un-
de lying da a being dicho omous (an in usion happening o no ). We
also examined he same ques ion o L2 swi ch ials, now also including
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
7
he language o he p e ious ial (i.e., L1-L2 o L3-L2 swi ch) as a ixed
e ec (swi ching om L3 =0.5; swi ching om L1 =-0.5).
Rhyme ask. In he hyme ask, we sco ed whe he he esponse o he
hyme p obe was he a ge wo d (i.e., he wo d used in he naming ask)
o no (i.e., a di e en wo d was p oduced o he e was no esponse).
The DV used was Ta ge esponse (0 =no, 1 =yes). The h ee inde-
penden a iables (IVs) we e Language (L1 −0.5; L3 0.5); Session (p e-
naming −0.5; pos -naming 0.5); and Condi ion (ac oss language −0.5;
wi hin language 0.5). The model con e ged wi h in e cep s o pa ici-
pan s and i ems and he by-pa icipan slope o condi ion.
Resul s
Speeded ilingual naming ask
Ac oss he 720 ials in he speeded ilingual naming ask (excluding
he en ials a e he b eaks), pa icipan s on a e age answe ed 70.5 %
co ec ly (SD =13.3, ange =37–91 % co ec ). Pa icipan s did no
espond on 13.9 % o ials (SD =6.9). C oss-language in usions (ac oss
all languages) we e made on 14.2 % o all ials (SD =12.2) and wi hin-
language in usions on 1.1 % o ials (SD =0.8); 0.3 % o ials (SD =
0.6) we e bo h c oss- and wi hin-language in usions ( o example
saying “apple” in Spanish when “ho se” had o be named in English). On
0.6 % o ials (SD =0.9), a co ec esponse was gi en ha was no he
in ended a ge wo d (e.g., “mu o” ins ead o “pa ed”).
All pa icipan s p oduced c oss-language in usions, he e o ype o
in e es . We ocused on he numbe o c oss-language in usions du ing
Basque (L2) non-swi ch ials (see Fig. 3). Du ing hese ials, pa ici-
pan s p oduced signi ican ly mo e L3 (M =8.2 % o L2 non-swi ch ials
showed an L3 in usion, SD =8.6 %) han L1 (M =5.7 % o L2 non-
swi ch ials showed an L1 in usion, SD =6.9; β =0.366, SE =0.064,
z =5.719, p <0.001) in usions.
Basque (L2) swi ch ials showed simila indings. Pa icipan s
p oduced mo e L3 (M =9.2 %, SD =8.6) han L1 (M =6.5 %, SD =6.8)
in usions (β =0.344, SE =0.073, z =4.695, p <0.001). The e was no
signi ican e ec o he language used on he p e ious ial (β =0.012,
SE =0.073, z =0.160, p =0.873) no a signi ican in e ac ion be ween
language o p e ious ial and language o in usion (β =0.267, SE =
0.147, z =1.818, p =0.069). Pa icipan s no only made mo e L3 (M =
9.9 %, SD =8.4) han L1 (M =6.1 %, SD =7.1) in usions when an L2
ial was p eceded by an L3 ial bu also when i was p eceded by an L1
ial (L3 in usions M =8.5 %, SD =9.3; L1 in usions M =6.9 %, SD =
7.1).
Gi en he lowe numbe o ials o be named in L3 (160 ac oss ial
ypes) o in L1 (160 ac oss ial ypes), and gi en ou ocus on L2 p o-
duc ion, we did no u he analyse he numbe o in usions du ing L1
o L3 ials. Howe e , nume ically mo e L3 in usions we e made du ing
L1 ials (M =5.3 %, SD =7.2) han L1 in usions du ing L3 ials (M =
4.0 %, SD =6.9). Gi en ha he ask equi ed pa icipan s o use he L2
mo e han he o he languages, he la ges numbe o in usions du ing
L1/L3 ials came om he L2 (du ing L1 ials M =8.2 %, SD =5.7;
du ing L3 ials M =9.7 %, SD =7.2).
In summa y, he speeded naming ask shows ha ilinguals we e
mo e likely o make L3 han L1 in usions when ha ing o name pic u es
in L2.
Rhyme ask
The hyme ask included Session (p e-/pos -naming), Language (L1/
L3), and Condi ion (ac oss-/wi hin-language naming) as he a iables.
Al hough he naming ask sepa a ing he p e- and pos -naming
hyme asks was no o main in e es , we examined accu acy o make
su e pa icipan s named he pic u es co ec ly. Accu acy was high in all
h ee languages (Spanish M =96.6 %, SD =4.0; Basque M =89.5 %, SD
=7.7; English M =89.9 %, SD =8.3). The majo i y o e o s (M =6.2 %
o all ials; SD =4.3) we e no o la e esponses; c oss-language in-
usions (M =0.6 % o ials, SD =0.9) and non- a ge wo d choice (e.g.,
“pa ed” ins ead o “mu o”, M =1.2 % o ials, SD =1.5) we e a e. This
con i ms ha pa icipan s used he a ge wo ds in he naming ask as
in ended.
Fig. 4 shows he esul s om he hyme ask. Pa icipan s p oduced
signi ican ly mo e hyme a ge s in he pos -naming ask (M =36.4 %,
SD =7.5) han in he baseline p e-naming hyme ask (M =20.0 %, SD
=5.5; β =1.052, SE =0.078, z =13.561, p <0.001). The e we e no
main e ec s o language (β =0.145, SE =0.307, z =0.472, p =0.637) o
condi ion (β =0.306, SE =0.313, z =0.978, p =0.328). These wo
a iables did no in e ac wi h each o he (β =-0.841, SE =0.615, z =
-1.368, p =0.171) bu impo an ly hey did in e ac wi h session. The
in e ac ion be ween session and language (β =0.701, SE =0.155, z =
4.537, p <0.001) e lec ed ha he inc ease be ween p e- and pos -
naming was la ge o L3 (English) hyme esponses (Mp e =18.3 %,
SD =6.0; Mpos =40.3 %, SD =9.7) han o L1 (Spanish) hyme e-
sponses (Mp e =21.8 %, SD =7.2; Mpos =32.5 %, SD =7.4). Session
also in e ac ed wi h condi ion (β =0.421, SE =0.155, z =2.723, p =
Fig. 3. Boxplo s showing he pe cen age o language in usions (L1 e sus L3)
du ing L2 non-swi ch ials. Each black squa e shows an indi idual pa icipan
(ji e ed). The ho izon al line shows he median while he black iangle shows
he mean.
Fig. 4. Boxplo s showing he pe cen age o
a ge hyme esponses in he p e-naming
ask (le , baseline) and he pos -naming
ask ( igh ). Wi hin each plo , he le panel
ep esen s he ‘ac oss-language’ condi ion
( a ge s named in L2 Basque) and he igh
panel he ‘wi hin-language’ condi ion ( a -
ge s named in Spanish/English, i.e., in he
same language as he hyme ask). Each
black do shows an indi idual pa icipan .
The ho izon al line shows he median while
he whi e squa e shows he mean.
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
8
0.006), e lec ing ha he inc ease be ween p e- and pos -naming was
la ge o wi hin-language ials (i.e., ials p e iously named in he
same language as he hyme ask; Mp e =19.9 %, SD =6.3; Mpos =
39.9 %, SD =8.5) han o ac oss-language ials (i.e., ials p e iously
named in L2; Mp e =20.2 %, SD =7.3; Mpos =32.8 %, SD =9.7).
Impo an ly, he e was a h ee-way in e ac ion be ween language, ses-
sion, and condi ion (β =-0.744, SE =0.309, z =-2.408, p =0.016). As
Fig. 4 shows, he e was a compa able inc ease ela i e o he baseline in
a ge hyme esponses o L1 and L3 wi hin-language ials. Howe e ,
he e was a la ge inc ease compa ed o baseline in L3 ac oss-language
han L1 ac oss-language ials, sugges ing ha L1 ac oss-language a -
ge s we e less accessible han L3 ac oss-language a ge s.
Follow-up analyses o he p e- and pos - ask sepa a ely i s showed
ha du ing he p e-naming hyme ask (i.e., baseline), c ucially, he e
was no in e ac ion be ween language and condi ion (β =-0.230, SE =
0.653, z =-0.353, p =0.724). This was expec ed gi en ha he ac oss-/
wi hin-language condi ion was only in oduced in he naming ask and
was he e o e i ele an in he p e-naming baseline. The pos -naming
hyme ask, howe e , did show an in e ac ion be ween language and
condi ion (β =-1.208, SE =0.611, z =-1.978, p =0.048). While he e
was a simila numbe o L3 ac oss- and wi hin-language hyme a ge s
(β =-0.089, SE =0.412, z =-0.215, p =0.830), he e we e mo e L1
wi hin- han ac oss-language hyme a ge s (β =1.119, SE =0.453, z =
2.467, p =0.014). This sugges s ha while L3 equi alen s p e iously
named in L2 we e no less accessible han hose named in he L3 i sel ,
L1 ansla ion equi alen s p e iously named in L2 we e less accessible
han wo ds named in he L1 i sel .
Follow-up analyses o he wo condi ions sepa a ely i s showed an
inc ease in a ge esponses pos - compa ed o p e-naming in he wi hin-
language condi ion (β =1.247, SE =0.107, z =11.659, p <0.001).
C ucially, his did no in e ac wi h language (β =0.325, SE =0.213, z =
1.527, p =0.127), demons a ing ha he inc ease in a ge esponses
a e using hose wo ds in he L1/L3 naming ask did no di e o he L1
and L3. In con as , he ac oss-language ials showed an in e ac ion
be ween session and language (β =1.090, SE =0.225, z =4.847, p <
0.001), e lec ing a la ge inc ease in ac oss-language a ge s o he L3
han L1. In he L3, he e was a signi ican inc ease in ac oss-language
hyme a ge s be ween he p e- and pos -naming ask (β =1.395, SE
=0.156, z =8.922, p <0.001). In he L1 he e was only a small inc ease
in ac oss-language hyme a ge s ha did no each signi icance (β =
0.310, SE =0.162, z =1.915, p =0.056).
To summa ise he indings om he hyme ask, all condi ions
showed a simila numbe o a ge esponses in he baseline measu e.
A e he slow naming ask, he e was a signi ican inc ease in a ge
esponses. Fo wo ds p e iously named in he ac ual language (wi hin-
language), his inc ease did no di e o he L1 and L3. Howe e , o
wo ds p e iously named in he L2 (ac oss-language), his inc ease was
la ge o he L3 han L1, sugges ing L1 ansla ion equi alen s we e less
accessible han L3 equi alen s.
Co ela ion be ween in usions and hyme a ge s
We also assessed whe he he e was a co ela ion be ween he ela-
i e numbe o L1 in usions du ing L2 non-swi ch ials (compa ed o L3
in usions) and he numbe o L1 ac oss-language pos -naming hyme
wo ds p oduced ( ela i e o L3). In o he wo ds, we aimed o assess he e
whe he he measu e o inhibi ion ( hyme ask) was ela ed o he in-
usions made. We we e speci ically in e es ed in assessing whe he
people who made ela i ely ew L1 in usions (as compa ed o L3 in-
usions) also p oduced ela i ely ew L1 hyme a ge s (as compa ed o
L3 a ge s). To do his, we used he pe cen age di e ence be ween L1
and L3 in usions and he pe cen age di e ence in hyme a ge wo ds
be ween L1 and L3 in he ac oss-language condi ion pos -naming. The e
was no signi ican co ela ion be ween he in usions and hyme a ge s
( =-0.156, p =0.410; see Supplemen a y Fig. 1). As an addi ional pos -
hoc check, we also examined he co ela ion while aking in o
conside a ion pe o mance du ing he baseline hyme ask ( o exclude
any indi idual di e ences in how likely pa icipan s we e o use a ge
wo ds be o e he naming manipula ion). The co ela ion was no
obse ed ei he when we compu ed he L1-L3 di e ence on he pos -
naming hyme ask ela i e o he p e-naming hyme ask ( =-0.225,
p =0.232; see Supplemen a y Fig. 1). In bo h cases, he di ec ion o he
co ela ion was he opposi e o expec ed (al hough no signi ican ). This
is la gely d i en by one ou lie (see Supplemen a y Fig. 1) who p oduced
a e y la ge numbe o L3 in usions.
Discussion
Expe imen 1 had wo aims. Fi s , we examined he amoun o lan-
guage in e e ence (in he o m o c oss-language in usions) s emming
om he na i e language (L1) e sus om a weake non-na i e language
(L3) du ing L2 p oduc ion. Du ing a speeded pic u e-naming ask pa -
icipan s showed mo e L3 han L1 in usions du ing L2 p oduc ion. This
sugges s ha , despi e he L1 being a mo e p o icien , he e was mo e
in e e ence om he non-na i e language han om he na i e lan-
guage. Second, we aimed o examine whe he ilinguals supp essed he
L1 mo e s ongly han he L3 du ing L2 p oduc ion. Using a hyme ask
a e a pic u e-naming ask showed ha L1 ansla ion equi alen s we e
used less o en han L3 ansla ion equi alen s, possibly because hey
we e supp essed mo e s ongly du ing L2 naming and consequen ly less
accessible.
The la ge amoun o L3 han L1 in usions du ing L2 use is consis-
en wi h he L3 acquisi ion li e a u e (e.g., Ba del & Falk, 2007; Falk &
Ba del, 2011; Mickan e al., 2020; Puig-Mayenco e al., 2020; Ro hman
& Cab elli Ama o, 2010), which shows ha a non-na i e language has a
la ge in luence han he na i e language while acqui ing a new lan-
guage. I is also consis en wi h he limi ed amoun o esea ch (Tomo-
schuk e al., 2021) sugges ing ha his non-na i e language in e e ence
migh pe sis e en a e he ini ial s ages o L3 acquisi ion. He e we show
o he i s ime ha inc eased L3 in e e ence can dis up L2 p oduc-
ion in ilinguals who acqui ed all h ee languages ea ly in li e and who
ha e an in e media e p o iciency in bo h o hei non-na i e languages.
Fu he mo e, mos o he acquisi ion li e a u e has ocused on in e e -
ence s emming om a non-na i e language (L2) acqui ed be o e he
o he non-na i e language (L3). Tomoschuk e al. (2021) only obse ed
inc eased non-na i e in e e ence on an L3 bu no on a (mo e domi-
nan ) L2. He e we show ha his in e e ence be ween non-na i e lan-
guages can also in luence he non-na i e language ha was acqui ed
i s (L2), po en ially p o ided ha his L2 has no eached e y high
le els o p o iciency o use.
One po en ial mechanism leading o inc eased in e e ence be ween
non-na i e languages migh be ha ilinguals apply mo e inhibi ion (o
apply inhibi ion mo e success ully o e icien ly) o e hei na i e lan-
guage. The in e p e a ion ha ilinguals supp ess hei L1 mos s ongly
is in line wi h G een’s (1998) inhibi o y con ol hypo hesis a guing ha
he amoun o inhibi ion applied is ela i e o he p o iciency in ha
language. As a consequence o he inc eased use o inhibi ion, in e e -
ence om he na i e language migh be educed when using a less
p o icien language.
Indeed, ou ilingual pa icipan s had easie access in he hyme ask
o L3 ansla ion equi alen s han L1 equi alen s a e L2 use, sugges ing
ha hey supp essed he L1 mo e s ongly han he L3. These language
di e ences we e no obse ed in he baseline ask, sugges ing ha hey
we e no due o di e ences be ween he L1 and L3 a ge wo ds o be-
ween he p obe- a ge ela ionships. They we e no obse ed on wi hin-
language ials ha we e named in he hyme language ei he . Mul i-
linguals migh apply language con ol no only by inhibi ing a non-
a ge language bu also by o e -ac i a ing he less p o icien lan-
guages (L2/L3, Philipp, Gade, & Koch, 2007). I he obse ed ac oss-
language L1/L3 di e ences in he hyme ask we e due o L3 wo ds
being ac i a ed mo e s ongly han L1 wo ds, language di e ences
should occu on he wi hin-language ials oo. The inding ha only
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
9
ac oss-language ials showed a language di e ence sugges s ha his
language e ec was ela ed o educed L1 accessibili y a he han
inc eased L3 accessibili y.
Va ious o he a iables ha e been sugges ed o explain inc eased
in e e ence be ween non-na i e languages. The L2 and L3 migh be
mo e cogni i ely simila (e.g., Ba del & Falk, 2007) i hey a e acqui ed
in simila ci cums ances. In Expe imen 1, howe e , only he L3 was
acqui ed as a ue “class oom” language. The L2 was, on a e age, ac-
qui ed du ing he i s h ee yea s o li e and is u he mo e pa o he
bilingual socie y o med by he Basque Coun y. Fu he mo e, he sug-
ges ion ha non-na i e language in e e ence migh be ela ed o he
non-na i e languages being augh in he L1 (Tomoschuk e al., 2021)
does no hold he e conside ing ha mos pa icipan s a ended educa-
ion in L2 only o in a combined L1/L2 sys em and used Basque ac oss
hei educa ional p og amme a he han as a class oom language
augh h ough ano he language. Typological p oximi y oo is unlikely
o explain he esul s conside ing ha he L2 di e ed subs an ially om
bo h he L1 and L3 bu , i any hing, shows mo e phonological and
o hog aphic o e lap wi h he L1.
Expe imen 1 hus sugges s ha he in e e ence be ween non-na i e
languages migh no be ela ed o cogni i e simila i y be ween he L2
and L3 bu a he o he way ilinguals apply mo e L1 han L3 inhibi-
ion. Howe e , u he analyses showed no co ela ion be ween he
ela i e numbe o L1 in usions and he ela i e accessibili y o L1
wo ds in he hyme ask. This could sugges ha despi e he same g oup
o pa icipan s showing ewe L1 in usions in he speeded naming ask
and educed L1 access in he hyme ask, he e is no di ec ela ionship
be ween he wo indings. Howe e , he wo asks a e also e y di e en
in many o he aspec s (e.g., he speeded swi ching ask equi es naming
o speci ic wo ds while he hyme ask allows o ee e ie al; he
speeded swi ching ask is a ilingual en i onmen while he hyme ask
was comple ed in single-language blocks). Fu he mo e, he hyme ask
migh elici o he indi idual di e ences beyond language p oduc ion,
including di e ences in he way people gene a e hymes and he size o
hei ocabula y (i.e., he numbe o compe i o s o he a ge hyme
wo ds). These many ask- ela ed di e ences could mask co ela ions
wi hin a ela i ely small sample. The e o e, while he hyme ask in
Expe imen 1 sugges ed ha L1 ansla ion equi alen s we e less
accessible as a consequence o inhibi ion, we did wan o in es iga e he
po en ial ole o inhibi ion and he ela ionship wi h in usions u he .
In Expe imen 2, we he e o e i s ly aimed o eplica e he in usion
indings om he speeded naming ask (in a di e en ype o ilinguals)
and we u he examined he second ques ion ega ding L1 e sus L3
inhibi ion. We used a di e en ask (n-2 swi ching ask, as explained
below) o p o ide a mo e comple e pic u e o he po en ial ole o lan-
guage inhibi ion.
Expe imen 2
In oduc ion
N-2 swi ching ask
Expe imen 2 used he n-2 swi ching ask o u he examine he
ques ion o inhibi ion o he L1 e sus L3. In his ask, pa icipan s a e
asked o swi ch languages on e e y ial. The n-2 ial (i.e., wo ials
be o e he a ge ) ei he has o be p oduced in he same language o in a
di e en language han he cu en ial (e.g., L1-L2-L1 would be an n-2
epe i ion ial while L3-L2-L1 would be a swi ch ial). I he language
used on ial n-2 has o be supp essed when swi ching o he L2 ( ial n-
1), pa icipan s should need mo e ime o swi ch back o ha language
on ial n. In o he wo ds, i language X is supp essed when using he L2,
i should ake ilinguals mo e ime o swi ch back o language X again
(Lx-L2-Lx) han when a di e en language was used wo ials ago (Lz-
L2-Lx). Indeed, se e al ilingual swi ching s udies using a ange o
language combina ions ha e shown hese n-2 cos s, wi h ilinguals
esponding mo e slowly in sequences in which ial n and n-2 ha e o be
named in he same language ( epe i ion) han when hey a e named in
di e en languages (e.g., Decle ck & Philipp, 2018; Decle ck, Thoma,
Koch, & Philipp, 2015; Guo, Liu, Chen, & Li, 2013; Philipp e al., 2007).
These epe i ion cos s a e aken as a e lec ion o pe sis ing inhibi ion
o a p e iously used language. Impo an ly, i he L1 is supp essed mo e
s ongly han he L3 when using he L2, his n2- epe i ion cos should be
la ge o he L1 han o he L3: he di e ence be ween (L1-L2-L1) and
(L3-L2-L1) should be la ge han be ween (L3-L2-L3) and (L1-L2-L3).
Some s udies ha e indeed sugges ed ha n-2 epe i ion cos s a e la ges
o he ilinguals’ mos dominan language(s) (e.g., Decle ck e al.,
2015; Philipp e al., 2007). Ne e heless, i should be no ed ha se e al
s udies ha e no shown la ge epe i ion cos s o he L1 han L2/L3 o
o he L2 han L3 (e.g., Philipp & Koch, 2009; see Decle ck & Koch,
2022, o a e iew showing inconsis encies ac oss s udies). Impo an ly,
howe e , n-2 epe i ion e ec s (and po en ial language di e ences) a e
mo e likely o be explained by inhibi ion accoun s han by o e -
ac i a ion o he a ge language. The la e would p edic epe i ion
p iming a he han a cos (i.e., i he Lx is o e -ac i a ed on ial n-2 and
no supp essed a all, ilinguals should be as e on Lx-L2-Lx ials han
on Lz-L2-Lx ials). Examining L1 e sus L3 epe i ion cos s hus allowed
us o examine po en ial di e ences be ween L1 and L3 supp ession.
Impo an ly, we always kep he L2 as he middle ial (n-1) o speci -
ically examine L1 e sus L3 inhibi ion applied while swi ching o he L2
(con a y o p e ious s udies, in which he middle ial o L1 epe i ion
cos s could be ei he L2 o L3 while he middle ial o L3 cos s could be
ei he L2 o L1). I mo e L1 inhibi ion is applied han L3 inhibi ion
du ing L2 naming, we would expec L1 epe i ion cos s o be la ge han
L3 epe i ion cos s.
The use o an n-2 swi ching ask allowed us o examine supp ession
o L1 e sus L3 h ough a di e en ype o ask ha , in bo h he
language-swi ching as well as he ask-swi ching li e a u e, is equen ly
used as a measu e o dominan language/ ask inhibi ion (e.g., Decle ck
& Philipp, 2018; Philipp e al., 2007). This was in ended o complemen
he, less equen ly used, hyme ask om Expe imen 1. C ucially,
howe e , we designed he n-2 ask di e en ly han p e ious s udies by
always using he L2 as he n-1 ial, which allowed us o speci ically look
a L1 e sus L3 inhibi ion du ing L2 naming. Fu he mo e, we assessed i
and how n-2 cos s we e ela ed o in usions in he speeded naming ask
o elucida e he ela ionship be ween non-na i e language in e e ence
and inhibi ion. The n-2 swi ching ask is mo e simila o he speeded
swi ching ask han he hyme ask, allowing us o examine his co e-
la ion in he absence o la ge ask di e ences (as was he case in
Expe imen 1).
I em-speci ic e sus global-language con ol
Addi ionally, his n-2 ask allowed us o assess global inhibi ion (o
he language as a whole) e sus i em-speci ic inhibi ion. When applying
inhibi ion, mul ilinguals can do his eac i ely by jus supp essing he
ansla ion equi alen o he a ge wo d in he non- a ge languages (e.
g., when naming “caballo” in Spanish, bilinguals migh jus supp ess he
equi alen “ho se” in English bu no o he English wo ds). Al e na i ely
o addi ionally, mul ilinguals migh supp ess all lemmas in he non-
a ge language (e.g., no jus “ho se”, bu also “apple”, “d ess”,
e ce e a). While p e ious s udies (e.g., Philipp e al., 2007) ha e shown
ha n-2 cos s can be used o s udy inhibi ion du ing language p oduc-
ion, i is less clea i his inhibi ion is applied a a global le el o is also
ela ed o speci ic i ems ha ha e o be named. Whole-language and
i em-speci ic inhibi ion a e no mu ually exclusi e: mul ilinguals migh
apply inhibi ion globally bu also inc ease inhibi ion o he ansla ion
equi alen s in an i em-speci ic manne (e.g., Decle ck & Philipp, 2017).
I em-speci ic e ec s ha e o en been s udied by epea ing he same
pic u es in a pic u e-naming ask. Fo example, Mis a e al. (2012)
showed a acili a ion e ec when he same pic u es had o be named in
he L2 a e a block o L1 naming. Howe e , his bene i o epea ing he
same pic u es was no p esen in L1 a e L2 naming, sugges ing inhi-
bi ion o L1 o ms while naming he pic u es in he L2. Howe e , his
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
16
Appendix B
Spanish and English wo ds we e ma ched on equency pe million
(se o 8: English M =66, SD =76; Spanish M =45, SD =47, (7) =
1.592, p =0.155; se o 24: English M =50, SD =63; Spanish M =38, SD
=59, (23) =1.925, p =0.067) and numbe o le e s (se o 8: English
M =4.1, SD =1.1; Spanish M =5.1, SD =1.0, (7) =-2.160, p =0.068;
se o 24: English M =5.0, SD =1.6; Spanish M =5.5, SD =1.1, (23) =
-1.801, p =0.085). Spanish wo ds we e longe in e ms o syllables (se
o 8: English M =1.1, SD =0.4, Spanish M =2.1, SD =0.4, - es could
no be conduc ed as he a iance o di e ence was 0; se o 24: English
M =1.4, SD =0.6; Spanish M =2.4, SD =0.6, (23) =-8.177, p <0.001)
and phonemes (se o 8: English M =3.1, SD =0.4; Spanish M =4.9, SD
=1.1, (7) =-4.249, p =0.004; se o 24: English M =3.8, SD =1.2;
Spanish M =5.3, SD =1.2, (23) =-5.146, p <0.001). Spanish and
English wo ds we e also ma ched in e ms o Le ensh ein Dis ance o he
F ench equi alen , which was compu ed as he numbe o edi s needed
o change one wo d in o he o he , ela i e o he leng h o he wo d
(wi h 1 meaning comple e o e lap and 0 meaning no o e lap: se o 8:
English M =0.06, SD =0.1; Spanish M =0.06, SD =0.1, (7) =0.026, p
=0.980; se o 24: English M =0.1, SD =0.1; Spanish M =0.1, SD =0.2,
(23) =0.936, p =0.359).
(See Table B1).
Appendix C. Supplemen a y ma e ial
Supplemen a y da a o his a icle can be ound online a h ps://doi.
o g/10.1016/j.jml.2022.104386.
Re e ences
Anwyl-I ine, A. L., Massoni´
e, J., Fli on, A., Ki kham, N. Z., & E e shed, J. K. (2020).
Go illa in ou mids : An online beha iou al expe imen builde . Beha io Resea ch
Me hods, 52(1), 388–407.
Ba del, C., & Falk, Y. (2007). The ole o he second language in hi d language
acquisi ion: The case o Ge manic syn ax. Second Language Resea ch, 23(4), 459–484.
Table A3 (con inued)
Condi ion Language Ta ge P obe
Wi hin English B oom Doom
Wi hin English Thie Chie
Wi hin English Bag Tag
Wi hin English Bell Cell
Wi hin English Kni e Wi e
Wi hin English S a Ja
Wi hin English Fa m Cha m
Wi hin English Snake Cake
Wi hin English Wa e Ca e
Ac oss English Ho se Fo ce
Ac oss English Shi Di
Ac oss English Moon Spoon
Ac oss English Honey Money
Ac oss English Wheel S eel
Ac oss English Fox Ox
Ac oss English Duck T uck
Ac oss English Glass Mass
Ac oss English Wall Ball
Ac oss English Neck W eck
Ac oss English Wing Sing
Ac oss English Snow G ow
Ac oss English Egg Peg
Ac oss English Pig Wig
Ac oss English Beach Teach
Ac oss English Pa h La h
Ac oss English Cloud P oud
Ac oss English Sheep Jeep
Ac oss English Ho n To n
Ac oss English Bone Tone
Wi hin Spanish Pue a Hue a
Wi hin Spanish Pico Chico
Wi hin Spanish Flecha B echa
Wi hin Spanish Mu˜
neca Beca
Wi hin Spanish Puen e Cuen e
Wi hin Spanish Conejo Espejo
Wi hin Spanish To o Fo o
Wi hin Spanish A a˜
na Caba˜
na
Wi hin Spanish Ba co Ma co
Wi hin Spanish Nudo Dudo
Wi hin Spanish T igo Digo
Wi hin Spanish Cama Rama
Wi hin Spanish Cla o Pa o
Wi hin Spanish Taza Raza
Wi hin Spanish Vaca Saca
Wi hin Spanish Silla Capilla
Wi hin Spanish Cadena A ena
Wi hin Spanish Boca Loca
Wi hin Spanish Cabeza Empieza
Wi hin Spanish Lla e Na e
Ac oss Spanish Caballo Fallo
Ac oss Spanish Camisa Son isa
Ac oss Spanish Luna Cuna
Ac oss Spanish Miel Piel
Ac oss Spanish Rueda Pueda
Ac oss Spanish Zo o Go o
Ac oss Spanish Pa o T a o
Ac oss Spanish Vaso Paso
Ac oss Spanish Mu o Du o
Ac oss Spanish Cuello Sello
Ac oss Spanish Ala Gala
Ac oss Spanish Nie e A e e
Ac oss Spanish Hue o Lle o
Ac oss Spanish Ce do Izquie do
Ac oss Spanish Playa Vaya
Ac oss Spanish Camino Molino
Ac oss Spanish Nube Tu e
Ac oss Spanish O eja O eja
Ac oss Spanish Cue no Tie no
Ac oss Spanish Hueso Peso
Table B1
S imuli used in Expe imen 2. The i s eigh i ems (in i alics) we e used in he
speeded naming ask; all 24 i ems we e used in he n-2 swi ching ask.
English F ench Spanish
duck cana d pa o
dog chien pe o
cheese omage queso
leg jambe pie na
ski jupe alda
bed li cama
apple pomme manzana
mea iande ca ne
candle bougie ela
doll poup´
ee mu˜
neca
hange cin e pe cha
shoulde ´
epaule homb o
snail esca go ca acol
lea euille hoja
og g enouille ana
abbi lapin conejo
coa man eau ab igo
wa ch mon e eloj
sheep mou on o eja
bi d oiseau p´
aja o
bu e ly papillon ma iposa
ain pluie llu ia
sui case alise male a
ci y ille ciudad
A. de B uin e al.
Jou nal o Memo y and Language 128 (2023) 104386
17
Ba del, C., & Falk, Y. (2012). The L2 s a us ac o and he decla a i e/p ocedu al
dis inc ion. In J. Cab elli Ama o, S. Flynn, & J. Ro hman (Eds.), Thi d language
acquisi ion in adul hood (pp. 61–78). Ams e dam: John Benjamins.
Ba del, C., & S´
anchez, L. (2017). The L2 s a us ac o hypo hesis e isi ed: The ole o
me alinguis ic knowledge, wo king memo y, a en ion and no icing in hi d
language lea ning. In T. Angelo ska, & A. Hahn (Eds.), L3 Syn ac ic T ans e : Models,
new de elopmen s and implica ions (pp. 85–101). Ams e dam: John Benjamins.
Bogulski, C. A., Bice, K., & K oll, J. F. (2019). Bilingualism as a desi able di icul y:
Ad an ages in wo d lea ning depend on egula ion o he dominan language.
Bilingualism: Language and Cogni ion, 22(5), 1052–1067.
B anzi, F. M., Ma in, C. D., Abu alebi, J., & Cos a, A. (2014). The a e -e ec s o
bilingual language p oduc ion. Neu opsychologia, 52, 102–116.
B ysbae , M. (2013). Lex ale_FR a as , ee, and e icien es o measu e language
p o iciency in F ench. Psychologica Belgica, 53(1), 23–37.
De B uin, A., Roelo s, A., Dijks a, T., & Fi zPa ick, I. (2014). Domain-gene al inhibi ion
a eas o he b ain a e in ol ed in language swi ching: FMRI e idence om ilingual
speake s. Neu oImage, 90, 348–359.
De B uin, A., Ca ei as, M., & Du˜
nabei ia, J. A. (2017). The BEST da ase o language
p o iciency. F on ie s in Psychology, 8, 522.
Decle ck, M., G ainge , J., & Ha suike , R. J. (2021). P oac i e language con ol du ing
bilingual sen ence p oduc ion. In e na ional Jou nal o Bilingualism, 25(6),
1813–1824.
Decle ck, M., & Koch, I. (2022). The concep o inhibi ion in bilingual con ol.
Psychological Re iew.Decle ck, M., & Philipp, A. M. (2017). Is he e lemma-based
language con ol? The in luence o language p ac ice and language-speci ic i em
p ac ice on asymme ical swi ch cos s. Language, Cogni ion and Neu oscience,32(4),
488-493.
Decle ck, M., & Philipp, A. M. (2018). Is inhibi ion implemen ed du ing bilingual
p oduc ion and comp ehension? n-2 language epe i ion cos s unchained. Language,
Cogni ion and Neu oscience, 33(5), 608–617.
Decle ck, M., Thoma, A. M., Koch, I., & Philipp, A. M. (2015). Highly p o icien
bilinguals implemen inhibi ion: E idence om n-2 language epe i ion cos s.
Jou nal o Expe imen al Psychology: Lea ning, Memo y, and Cogni ion, 41(6),
1911–1916.
Degani, T., K eine , H., A a ia, H., & Kha eeb, F. (2020). The impac o b ie exposu e o
he second language on na i e language p oduc ion: Global o i em speci ic? Applied
Psycholinguis ics, 41(1), 153–183.
Duchon, A., Pe ea, M., Sebas i´
an-Gall´
es, N., Ma í, A., & Ca ei as, M. (2013). EsPal: One-
s op shopping o Spanish wo d p ope ies. Beha io Resea ch Me hods, 45(4),
1246–1258.
Du˜
nabei ia, J. A., C epaldi, D., Meye , A. S., New, B., Plia sikas, C., Smolka, E., &
B ysbae , M. (2018). Mul iPic: A s anda dized se o 750 d awings wi h no ms o
six Eu opean languages. Qua e ly Jou nal o Expe imen al Psychology, 71(4),
808–816.
Eu opean Commission (2012), Special Eu oba ome e 386: Eu opeans and hei Languages.
[online] A ailable a : h ps://eu opa.eu/eu oba ome e /su eys/de ail/1049.
Falk, Y., & Ba del, C. (2011). Objec p onouns in Ge man L3 syn ax: E idence o he L2
s a us ac o . Second Language Resea ch, 27(1), 59–82.
Gollan, T. H., Scho e , E. R., Gomez, J., Mu illo, M., & Rayne , K. (2014). Mul iple le els
o bilingual language con ol: E idence om language in usions in eading aloud.
Psychological Science, 25(2), 585–595.
G een, D. W. (1998). Men al con ol o he bilingual lexico-seman ic sys em. Bilingualism:
Language and Cogni ion, 1(2), 67–81.
G een, D. W., & Abu alebi, J. (2013). Language con ol in bilinguals: The adap i e
con ol hypo hesis. Jou nal o Cogni i e Psychology, 25(5), 515–530.
Guo, T., Liu, F., Chen, B., & Li, S. (2013). Inhibi ion o non- a ge languages in
mul ilingual wo d p oduc ion: E idence om Uighu –Chinese–English ilinguals.
Ac a Psychologica, 143(3), 277–283.
Hi osh, Z., & Degani, T. (2021). No el Wo d Lea ning Among Bilinguals Can Be Be e
Th ough he (Dominan ) Fi s Language Than Th ough he Second Language.
Language Lea ning, 71(4), 1044–1084.
Izu a, C., Cue os, F., & B ysbae , M. (2014). Lex ale-Esp: A es o apidly and e icien ly
assess he Spanish ocabula y size. Psicol´
ogica, 35(1), 49–66.
Jackson, G. M., Swainson, R., Cunning on, R., & Jackson, S. R. (2001). ERP co ela es o
execu i e con ol du ing epea ed language swi ching. Bilingualism: Language and
Cogni ion, 4(2), 169–178.
Lemh¨
o e , K., & B oe sma, M. (2012). In oducing LexTALE: A quick and alid lexical es
o ad anced lea ne s o English. Beha io Resea ch Me hods, 44(2), 325–343.
Le y, B. J., McVeigh, N. D., Ma ul, A., & Ande son, M. C. (2007). Inhibi ing you na i e
language: The ole o e ie al-induced o ge ing du ing second-language
acquisi ion. Psychological Science, 18(1), 29–34.
Meu e , R. F., & Allpo , A. (1999). Bilingual language swi ching in naming:
Asymme ical cos s o language selec ion. Jou nal o Memo y and Language, 40(1),
25–40.
Mickan, A., McQueen, J. M., & Lemh¨
o e , K. (2020). Be ween-language compe i ion as a
d i ing o ce in o eign language a i ion. Cogni ion, 198, A icle 104218.
Mis a, M., Guo, T., Bobb, S. C., & K oll, J. F. (2012). When bilinguals choose a single
wo d o speak: Elec ophysiological e idence o inhibi ion o he na i e language.
Jou nal o Memo y and Language, 67(1), 224–237.
Philipp, A. M., Gade, M., & Koch, I. (2007). Inhibi o y p ocesses in language swi ching:
E idence om swi ching language-de ined esponse se s. Eu opean Jou nal o
Cogni i e Psychology, 19(3), 395–416.
Philipp, A. M., & Koch, I. (2009). Inhibi ion in language swi ching: Wha is inhibi ed
when swi ching be ween languages in naming asks? Jou nal o Expe imen al
Psychology: Lea ning, Memo y, and Cogni ion, 35(5), 1187–1195.
P o opapas, A. (2007). Check Vocal: A p og am o acili a e checking he accu acy and
esponse ime o ocal esponses om DMDX. Beha io Resea ch Me hods, 39(4),
859–862.
Puig-Mayenco, E., Gonz´
alez Alonso, J., & Ro hman, J. (2020). A sys ema ic e iew o
ans e s udies in hi d language acquisi ion. Second Language Resea ch, 36(1),
31–64.
Ro hman, J., & Cab elli Ama o, J. (2010). Wha a iables condi ion syn ac ic ans e ? A
look a he L3 ini ial s a e. Second Language Resea ch, 26(2), 189–218.
Runnq is , E., & Cos a, A. (2012). Is e ie al-induced o ge ing behind he bilingual
disad an age in wo d p oduc ion? Bilingualism: Language and Cogni ion, 15(2),
365–377.
Tomoschuk, B., Duyck, W., Ha suike , R. J., Fe ei a, V. S., & Gollan, T. H. (2021).
Language o ins uc ion a ec s language in e e ence in he hi d language.
Bilingualism: Language and Cogni ion, 24(4), 707–718.
Van Assche, E., Duyck, W., & Gollan, T. H. (2013). Whole-language and i em-speci ic
con ol in bilingual language p oduc ion. Jou nal o Expe imen al Psychology:
Lea ning, Memo y, and Cogni ion, 39(6), 1781–1792.
Williams, S., & Hamma be g, B. (1998). Language swi ches in L3 p oduc ion:
Implica ions o a polyglo speaking model. Applied Linguis ics, 19(3), 295–333.
Wilson, M. (1988). MRC psycholinguis ic da abase: Machine-usable dic iona y, e sion
2.00. Beha io Resea ch Me hods, Ins umen s, & Compu e s, 20(1), 6–10.
Zheng, X., Roelo s, A., & Lemh¨
o e , K. (2020). Language selec ion con ibu es o
in usion e o s in speaking: E idence om pic u e naming. Bilingualism: Language
and Cogni ion, 23(4), 788–800.
A. de B uin e al.
