Chap e 6
Cons uc ing a clien ele in need: The
ield o Ge man as a Second Language
Jana Gampe a, Ch is oph Sch oede b, Julia
Schlaucha& Do o heé S einbockb
aUni e si y o Giessen bUni e si y o Po sdam
When s udying mul ilingual child en and adolescen s, i is c ucial o ask abou
he age o onse and/o leng h o exposu e o he languages in ques ion be o e
classi ying hem as L2 lea ne s o no . In ou semi-sys ema ic e iew o 138 pa-
pe s om he ield o esea ch on Ge man as a Second Language (GSL), we ask o
wha ex en his dis inc ion is d awn in he ield. GSL is conce ned wi h a leas
wo undamen ally di e en g oups. One g oup is mul ilingual child en and ado-
lescen s who acqui e Ge man as one o hei i s languages o as hei ea ly L2.
The o he g oup is he i s gene a ion o newly immig a ed child en, who clea ly
acqui e Ge man as hei second language. We e iewed a da a sample o n=138
pape s ha we e coded acco ding o explici and implici ea u es ela ed o GSL
speake s. Wi h he help o ela i e equencies and associa ion analyses we show
ha GSL speake s a e mainly concep ualized as “bi- o mul ilingual” and speake s
wi h “languages o he han Ge man” in hei epe oi e wi h “insu icien Ge man
compe ence” expe iencing “educa ional disad an ages” in schools. Common psy-
cholinguis ic L2-c i e ia such as Age o Onse (AoO) occu su p isingly a ely in
ou da a. These indings sugges a de ici -o ien ed concep ualiza ion o GSL speak-
e s and ein o ce endencies o O he ing. In ligh o hese indings, we a gue o
a na owing o he concep o GSL and a be e disclosu e o L2-lea ne -speci ic
me ada a.
1 In oduc ion
We ask he e o wha ex en “O he ing” applies o s udies in he ield o Ge man
as a Second Language (Deu sch als Zwei sp ache, hence o h GSL), wi h a ocus
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock. 2025. Cons uc ing a
clien ele in need: The ield o Ge man as a Second Language. In A emis Alexiadou, Claudio
Sca aglie i, Ch is oph Sch oede & Heike Wiese (eds.), The cons uc ion o mul ilinguals as
O he s: Do we p ac ice wha we p each?, 103–133. Be lin: Language Science P ess. DOI: 10.5281/
zenodo.17132449
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
on s udies ela ed o child and adolescen lea ne s. We combine his wi h a sec-
ond ques ion, asking, how academic s udies in he academic ield o GSL con-
cep ualize hei “clien ele” o lea ne s. We conduc ou esea ch in he o m o
a sys ema ic li e a u e e iew. Ou esea ch ques ions o igina e in he obse a-
ion ha GSL is a a he ague concep , ending o ake a de icien pe spec i e
on i s clien ele. This is ela ed, in ou iew, o an unclea di e en ia ion be ween
he psycholinguis ic no ion o GSL in he ame o second language (hence o h
L2) acquisi ion on he one hand, which has o should ha e (in ou iew) ce ain
speci ic and empi ically based implica ions. On he o he hand, he academic
ield o GSL has a pa icula ly applied adi ion in Ge many ha di e s om (es-
ablished) second language acquisi ion (hence o h SLA) esea ch in pa icula
ways.
Ou chap e is s uc u ed as ollows: In sec ion 2, we subs an ia e ou esea ch
ques ion. In sec ion 3, we de elop and conduc ou e iew. Sec ion 4 discusses
ou indings.
2 Theo e ical conside a ions
To subs an ia e ou esea ch ques ion, we i s b ie ly in oduce ou unde s and-
ing o mul ilingualism and second language acquisi ion (2.1), be o e u ning o
he academic ield o GSL (2.2) and some ini ial quali a i e obse a ions (2.3),
which lead us o poin ou ce ain dange s o “O he ing”, which may o igina e in
he misma ch be ween he scope o he academic ield and he psycholinguis ic
no ion o GSL.
2.1 Mul ilingualism, second language acquisi ion and age o onse
Acqui ing an L2 is one way an indi idual can become mul ilingual, and in he
indi idual-o ien ed use, he e m “mul ilingualism” is an umb ella e m o a -
ious o ms o language acquisi ion in he cou se o an indi idual’s li e, as well
as o linguis ic p ac ices and egula ions in e e yday li e, in wo king li e and
ins i u ions (c . G osjean 2010).
Indi iduals become mul ilingual by acqui ing mo e han one language. The e
is no doub ha he age o onse (AoO) and he leng h o exposu e (LoE) o he
language(s) a e decisi e o he cou se o he acquisi ion p ocess, bu also o
i s speed and o some ex en o i s success (c . Gaga ina e al. 2021, G imm &
C is an e 2022, Meisel 2009). Di e ences in AoO, hen, a e impo an indica ions
o he (expec ed) quali a i e cou se o acquisi ion. The acquisi ion o mo e han
104
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
one language simul aneously as i s languages is commonly called “bilingual
i s language acquisi ion”. The acquisi ion o a second language a e he age
o 3 is called “successi e second language acquisi ion”. Speaking o Ge man as
an L2, his is whe e “Ge man as a Second Language” begins.1Successi e second
language acquisi ion up o he age o 7 is summa ized as “ea ly childhood sec-
ond language acquisi ion”. Finally, a dis inc ion mus be made be ween whe he
he language(s) a e acqui ed as a child o as an adul . Adul acquisi ion o an-
o he language is conside ed o s a wi h he onse o pube y (c . Hyl ens am
& Ab ahamsson 2003). All in all, language de elopmen in Ge man as L2 can be
summa ised well by saying ha he ea lie he acquisi ion o Ge man begins, he
mo e simila i s cou se and dynamics a e o monolingual i s language acquisi-
ion. Wo k by Thoma & T acy (2007) and T acy (2008) makes i qui e clea ha
child en who begin acqui ing Ge man be o e he age o 3-4 go h ough he same
s ages o acquisi ion, a leas in he a ea o syn ac ic de elopmen , as child en
who acqui e Ge man as hei i s language. A i s quali a i e b eak becomes
isible a he age o onse in o second language acquisi ion a ound age 4. In
pa icula , he wo k o Habe ze l (2014) and Gaga ina e al. (2021) show ha
child en who begin second language acquisi ion a his age acqui e he language
di e en ly han child en o whom he p ocess begins ea lie (see also G imm &
C is an e 2022).
These age- ela ed di e en ia ion o lea ne g oups a e no mo e han app ox-
ima e guidelines o he connec ion o onse and cou se o L2 acquisi ion, bu
AoO is a c ucial de ining ac o when i comes o L2 lea ne iden i ica ion. How-
e e , he e a e no eliable igu es o Ge many on how many o he mul ilingual
indi iduals can be classi ied as L2 lea ne s. Fo example, he e a e no igu esa ail-
able on he p opo ion o amilies who speak only Ge man as he amily language
(=monolingual Ge man language acquisi ion), o speak Ge man and ano he lan-
guage as amily languages (=simul aneous bilingual acquisi ion). Childca e a es
in kinde ga ens could p o ide a clue, al hough in o icial s a e epo s only da a
on “mig a ion backg ound” is a ailable,2bu no linguis ic da a. S ill, i we ake
1This dis inc ion is by no means uncon o e sial. The e a e ce ainly app oaches ha place he
beginning o successi e second language acquisi ion (in con as o bilingual i s language
acquisi ion) la e , namely a he age o ou o i e (Meisel 2009, Schulz & G imm 2019). No e
ha he ac ha he exac o mula ion o he c i e ia migh be con o e sial does no diminish
he impo ance o clea -cu c i e ia.
2“Mig a ion backg ound” (Ge m. Mig a ionshin e g und) is a sociodemog aphic cha ac e is ic
used in demog aphic s a is ics in Ge many in o de o desc ibe he mig an communi y in he
coun y. Acco ding o he Fede al S a is ical O ice in Ge many, indi iduals ha e a mig a ion
backg ound i hey hemsel es o a leas one o hei pa en s we e bo n wi h a non-Ge man
ci izenship. See o an o e iew and c i ical e alua ion Will (2020).
105
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
he s a o kinde ga en a endance as he s a o he acquisi ion o Ge man o
child en wi h a “mig a ion backg ound”, hen, he (cau ious) conclusions abou
he age a he s a o acquisi ion a e as ollows: Fo he as majo i y o child en
bo n in Ge many wi h a mig a ion backg ound, we can assume bilingual i s
language acquisi ion o ea ly second language acquisi ion.3
Wha does his say wi h ega d o expec ed compe encies in Ge man? On
he one hand, he idea ha second language acquisi ion is comple ed when he
speake has mas e ed he second language a he same le el as a monolingual
speake is no enable: Bilingualism o mul ilingualism does no mean mul iplied
monolingualism, as he amous ci a ion om G osjean (1989) goes. A he same
ime, he a ia ion in language use and compe ence among monolingual speak-
e s is so g ea ha a no ma i e assump ion o monolingual language use and
compe ence is no possible (Shad o a e al. 2021).
Howe e , he e is a g owing body o esea ch esul s which compa e he aca-
demic language compe encies o child en and adolescen s bo n in Ge many wi h
a mig a ion backg ound and/o wi h a amily language o he han o besides
Ge man wi h hose o child en and adolescen s wi hou a mig a ion backg ound
o wi h a monolingual acquisi ion biog aphy. These allow he ollowing con-
clusion: The as majo i y o child en and adolescen s bo n and aised in Ge -
many wi h a amily language o he han o alongside Ge man do no di e in
hei linguis ic compe encies om monolingual child en and adolescen s when i
comes o academic language pe o mance (see among o he s Habe ze l 2016, Ri-
ca B ede 2020, Pe e sen 2014). I he e was successi e second language acquisi-
ion a all (and no bilingual i s language acquisi ion), hen his p ocess eaches
comple ion in p ima y school (Gaga ina e al. 2021). A he same ime, his does
no necessa ily ha e o mean ha he linguis ic compe encies and p ac ices o
mul ilinguals a e indis inguishable om he linguis ic compe encies o p ac ices
o monolinguals. Howe e , po en ial suppo needs a e no di e en om hose
aced by monolinguals and co ela e p ima ily wi h social ac o s ha limi ac-
cess o he academic egis e o school (Reiss e al. 2019, Rauch 2019).
I we summa ise he discussion abo e o ou esea ch ques ions, wo poin s
become e iden : Fi s , in he s udy o mul ilingual child en and adolescen s, i
3Acco ding o he igu es o he Fede al S a is ical O ice, jus o e 20% o 0-3 yea old
child en wi h a mig a ion backg ound a end kinde ga en and he a e ises o o e 80%
o 3-6 yea olds (wi h ce ain di e ences be ween he ede al s a es), see S a is isches
Bundesam , igu es o Ma ch 2020, a h ps://www.des a is.de/DE/Themen/Gesellscha -
Umwel /Soziales/Kinde agesbe euung/Tabellen/be euungsquo e-mig a ion-un e 6jah en-
nach-laende n.h ml;jsessionid=8F81E7EAE4F5065DB0917E1313242370.li e742 [accessed
2024/04/09].
106
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
is essen ial o ask o biog aphical de ails conce ning he exposu e o he newly
acqui ed language, be o e classi ying hem as L2 lea ne s. On a minimal le el,
his is bes done h ough inqui ing abou he age o onse o acquisi ion (AoO),
and/o he leng h o exposu e (LoE), possibly also he leng h o schooling (LoS)
in he language. And second, o hose we classi y as “ea ly second language
lea ne s”, we can assume ha hey s op being “second language lea ne s” a he
la es a p ima y school age.
2.2 The academic ield o Ge man as a Second Language
The academic ield o Deu sch als Zwei sp ache (DaZ / GSL) is ela i ely new. I
eme ged om he pe iod o he i s “gues wo ke ” immig a ion o Ge many
in he 1970s. Ge many expe ienced an economic boom in he la e 1960s, and
he demand o labo exceeded domes ic esou ces. As a esul , wo ke s om
Sou he n Eu ope and No h A ica we e ec ui ed in la ge numbe s. Ini ially,
hese so-called “gues wo ke s” we e o s ay only o a ew yea s, bu i soon
became clea ha he need o labo con inued, and companies did no wan o
le hei ained wo ke s lea e again. Consequen ly, he wo ke s we e in i ed
o se le o good, oge he wi h hei amilies. Resea ch p ojec s ocussing on
he linguis ic aspec s o his de elopmen (ini ially ca ied ou unde he name
o Deu sch als F emdsp ache (DaF), ha is, “Ge man as a o eign language”, and
only om he 1980s on “DaZ”, c . Bau 2001) concen a ed, on he one hand,
on un u o ed second language acquisi ion o (mos ly young) adul s (Ah enholz
& Ros -Ro h 2021). On he o he hand, a s ong applied o ien a ion de eloped,
which u ned o he socially impo an ques ion o how he school sys em deal
o should deal wi h s uden s whose knowledge o Ge man was ini ially oo low
o be able o ollow he con en o ins uc ion (Menk 2000). In his line o esea ch,
a ocal opic o GSL esea ch became he pa icula school egis e o academic
Ge man (Bildungssp ache, c . Lange 2020) ha is p esumed o open doo s o
social pa icipa ion.
Almos hal a cen u y has passed since he eme gence o GSL as an academic
ield. On he one hand, he o iginal ield o ac i i y o GSL has changed conside -
ably. The s udy o adul second language acquisi ion has di e si ied conside ably
because adul immig a ion has changed g ea ly in o m and scope. A e all, he
new Ge man Immig a ion Ac o 2005 has in oduced ins i u ionalized language
eaching in he o m o in eg a ion cou ses (c . Gampe e al. 2021). A he same
ime, he o me “child en o o eigne s” (Ge m. Auslände kinde ) a e now adul s
“wi h a mig a ion backg ound”; hey a e pa en s o g andpa en s; hei child en
and child en’s child en we e bo n in Ge many, go h ough Ge man educa ional
107
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
ins i u ions and, as long as a leas one pa en was bo n ab oad wi h a o eign
ci izenship, a e again child en o young people “wi h a mig a ion backg ound”.
Le us unde line he ac ha a his o y o amily mig a ion does no ell us any-
hing abou whe he Ge man is an L1 o L2, especially o descendan s o he i s
gene a ion o immig an s (see sec ion 2.1).
On he o he hand, immig a ion o child en and adolescen s con inues. Pa -
icula ly wi h he la ge numbe s o e ugee mig a ion in ecen yea s, i s in
2015 ollowing he Sy ian ci il wa , and hen a e he Russian a ack on Uk aine
in Feb ua y 2022, ques ions abou o ms o schooling o newcome s wi h li le
o no knowledge o Ge man, s udies on he second language acquisi ion o his
g oup a school, and on he condi ions and possibili ies o suppo ha e once
again come o he o e (Gampe e al. 2021).
Thus, he academic ield o GSL in Ge many deals wi h a leas wo undamen-
ally di e en g oups: On he one hand, some mul ilingual child en and adoles-
cen s a e bo n in Ge many and acqui e Ge man as one o hei i s languages o
as hei ea ly L2, and, on he o he hand, he i s gene a ion o immig an s, who
acqui e Ge man as hei L2. I is c ucially impo an o dis inguish be ween hese
wo g oups. The second g oup needs suppo in lea ning Ge man, bu he i s
g oup does no , a leas no in he same way. Suppo he e means ha newly
immig a ed child en and adolescen s need help o acqui e basic knowledge o
Ge man o gain access o he school-speci ic egis e s. Secondly, lumping he
wo g oups oge he is meaningless i one wan s o gene alize he dynamics o
second language acquisi ion. And hi dly, no dis inguishing be ween he g oups
may ha e he consequence ha he i s g oup slips in o a “p oblem g oup” ca -
ego iza ion. This is no only w ong bu i also has disc imina ing endencies, in
o he wo ds, endencies o “O he ing”.4
So a , we ha e seen ha he umb ella e m GSL is ambiguous. On he one
hand, i co e s lea ne s who a e new immig an s and mee he c i e ion o ha -
ing s a ed L2 acquisi ion a e he age o 3. Howe e , due o he his o ical de el-
opmen o he ield o GSL, i also co e s lea ne s o whom i is unclea whe he
Ge man is a (second) L1 o an L2. This concep ual agueness5has consequences
o how L2 lea ne s a e concep ualized and wha a ibu ions a e made o his
(unclea ly de ined) g oup.
4We speak o “O he ing” as he di e en ia ion o a g oup o a pe son om ano he g oup by
desc ibing he o me g oup o pe son as di e en and alien. Language is a possible ool o
di e en ia ion he e (c . Szymczyk e al. 2022).
5See o a simila discussion in G ea B i ain Szymczyk e al. (2022).
108
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
2.3 Obse a ions
Gi en he p e ious obse a ions, we ha e he imp ession ha he academic ield
o GSL does indeed ha e a endency o de elop a ce ain blind spo and ea
bilingual young people who we e bo n and aised in Ge many as GSL lea ne s
and in need o special language suppo . Th ee examples should su ice o a s a ,
and we gi e hem wi hou a e e ence, because we do no wan o pu indi idual
au ho s on a spo :
• A mo e didac ically o ien ed in oduc ion o eaching Ge man in mul i-
lingual class ooms om 2017 equa es being mul ilingual wi h acqui ing
Ge man as he second language.
• One o he mos success ul in oduc ions o Ge man as a second language
(4 h edi ion, 2020) uses “mig a ion backg ound” as he cen al de ining c i-
e ion o L2 lea ne s o Ge man. La e , “insu icien knowledge o Ge man
as a second language” is iden i ied as a decisi e ac o o he educa ional
ailu e o his g oup, in addi ion o “lowe class membe ship”.
• A s udy on he eading beha io o adolescen s, published 2014 by a ec-
ognized publishe in he ield, de ines L2 lea ne s o Ge man on he basis
ha hey ha e “non-Ge man mo he ongue”. A no poin does AoO o LoE
play a ole, nei he a e he linguis ic compe encies o he adolescen s opic
a any poin .
Ou e iew is in ended o answe he ques ion, whe he he abo e examples
ep esen a gene al end in he sense ha ea u es like mig a ion backg ound,
mul ilingualism and L2 lea ne a e mixed and addi ionally a ibu ed wi h insu -
icien academic pe o mance, o whe he hey a e excep ions. Le us make qui e
clea a his poin ha ou s udy is no in ended as jus ano he ac o O he -
ing. The au ho s o his s udy by no means exclude hemsel es om his c i ical
(sel -) e lec ion bu ha e in he pas engaged hemsel es in publica ions which
de ined Ge man as a L2 h ough implici and ague ca ego ies. So when asking
he e, wha kind o discu si e cons uc ion o a clien ele eme ges in schola ly pub-
lica ions on GSL, we also ask, wha kind o discu si e cons uc ions we ou sel es
ha e engaged in.
3 Re iew
The ocus o he empi ical wo k ha ollows is on he ques ion o whe he ou
selec i e obse a ions s and up o empi ical e i ica ion. These obse a ions in-
109
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
clude on he one hand he hypo hesis ha di e en a ge g oups, namely L2
lea ne s in he na ow sense as well as mul ilingual speake s in gene al, a e
mixed unde he umb ella e m GSL. On he o he hand, he cons uc GSL seems
o be associa ed wi h insu icien knowledge o Ge man and he need o sup-
po . To es hese assump ions, we conduc ed a semi-sys ema ic e iew. Ou
esea ch ques ion was: How do academic pape s on GSL published in Ge many,
and w i en in Ge man o a Ge man academic audience, concep ualize child and
adolescen GSL lea ne s?
3.1 Me hods
3.1.1 Sample
The sample o ou e iew includes 138 pape s, o which 92 a e empi ical s udies
and 46 a e non-empi ical (i.e. concep ual). To compile he sample, we conduc ed
an au oma ic li e a u e sea ch in he da abase “FIS Bildung”,6which includes
monog aphs, collec i e wo k con ibu ions and jou nal a icles om di e en
ields o acquisi ion- and lea ning- ela ed esea ch. The da abase includes mo e
han 1 million subjec -speci ic i les. Wi hin his da abase, we sea ched o pape s
ha ca ied he e m “Ge man as a Second Language” (in Ge man) in he i le in
di e en a ian s (i.e., Ge man as a Second Language, Second Language Ge man,
L2 Ge man, Ge man as a Fo eign- and Second Language) and we e published be-
ween 2011 and 2021. The i s sample compiled in his way included 113 i les,
which we e educed in a second s ep excluding wo ks ocusing on ou -o -school
lea ning, eache educa ion, de elopmen o eaching ma e ial, as well as on adul
lea ne s. The inal sample ex ac ed om he FIS da abase comp ised 79 i les.
In a second s ep, we included a book se ies in he sample ha has es ablished
i sel in he Ge man-speaking communi y as ele an o he opics o GSL, mul-
ilingualism, and L2 acquisi ion. The se ies is he esul o he wo kshop Kinde
mi Mig a ionshin e g und (Child en wi h a Mig a ion Backg ound), which has
exis ed since 2009 and was enamed Wo kshop Deu sch als Zwei sp ache, Mig a-
ion und Meh sp achigkei (Ge man as a Second Language, Mig a ion and Mul i-
lingualism Wo kshop) a ound 2015. The wo kshop is held annually and in i es
esea che s conduc ing GSL esea ch. The se ies was included in he e iew be-
cause i e lec s cu en esea ch p ojec s and discou ses and can hus be classi-
ied as discou se-shaping. The e iew includes 14 wo kshop olumes ( olume 1
o double olume 14/15) om he pe iod 2009 o 2021. The olumes comp ise a
6h ps://www. achpo al-paedagogik.de/li e a u /e wei e e_suche.h ml?checkFo mPa ams=
1&he kun []= is [accessed 2024/04/09].
110
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
o al o 189 con ibu ions, o which we only included he empi ical pape s in he
e iew (n = 89). O he 89 empi ical wo kshop pape s, 30 we e excluded because
hey did no mee he abo e men ioned inclusion c i e ia, lea ing 59 i les in he
wo kshop se ies subsample. The eason o including only empi ical pape s om
he book se ies is ha we assume ha mainly empi ical s udies can shed ligh
on how GSL lea ne s a e ope a ionalized. Fo he o e all sample, howe e , his
means ha he e is a su plus o empi ical s udies.
3.1.2 Da a p epa a ion
The sample was coded using a double-coding app oach (c . O’Conno & Jo e
2020). The codes we e i s de eloped deduc i ely om he psycholinguis ic li -
e a u e and hen ex ended on he basis o a smalle da a sample. In his way we
we e able o iden i y a o al o 19 ea u es, o which we conside ed h ee o be
explici and 16 o be implici o seconda y. We eco ded bo h se s o cha ac e -
is ics (i.e., explici s. implici ) once wi h ega d o he heo e ical desc ip ion o
GSL in he en i e sample and once exclusi ely wi h ega d o he ope a ionaliza-
ion o subjec s in empi ical s udies. Table 1 gi es an o e iew o he ea u es we
iden i ied wi hin he sample.
Unde “explici ”, we ha e g ouped ea u es which unambiguously cla i y ha
Ge man is an L2 in he psycholinguis ic sense. The AoO, LoE and LoS ea u es
can be used o measu e he onse and age o acquisi ion.7Unde “implici ” we
ha e g ouped cha ac e is ics ha may be a consequence o (e.g., bilingualism o
mul ilingualism, o he o addi ional language spoken a home)8o a eason o
L2 acquisi ion (e.g., mig a ion, e uge s a us), bu hese a e no su icien c i e ia
o de e mine whe he Ge man is indeed an L2 o he lea ne s desc ibed in he
pape s. Among he implici ea u es, mo eo e , he e a e also nume ous ones
ha ha e no ela ion o GSL in he na ow sense, mos no ably he ea u es “in-
su icien compe ence” and “educa ional disad an age”. Such cha ac e is ics may
apply o all lea ne s in he school, ega dless o hei indi idual language biog a-
phies. The lis also includes ea u es ha sugges an O he ing o GSL speake s
(e.g., speci ic e hnic backg ound o he ea u e “ o eign”). The e m “unspeci ic”
was used when GSL lea ne s we e men ioned bu no desc ibed in de ail. “O he ”
includes cha ac e is ics ha a e oo a e o eco d sepa a ely.
7We included Leng h o Schooling (LoS) unde explici ea u es since we assume ha epo ing
he leng h o schooling allows us o in e he leng h o exposu e and also he age o onse .
8“O he language” e e s o he assump ion ha he speake s speak a di e en language and
he e o e do no speak Ge man a home o as L1. “Addi ional language”, on he o he hand,
means ha Ge man is spoken in addi ion o ano he amily language.
111
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
pa a i ely low o e all occu ence, we ind s udies ha emain unspeci ic when
i comes o ope a ionalizing o GSL.
O e all, he da a e iew so a sugges s a s ong endency o ame GSL in
e ms o implici a he han explici ea u es. Wi hin hese implici ea u es,
he e is a endency owa d e y gene al cha ac e is ics o speake s and hei
biog aphies (e.g., bilingual, mul ilingual, immig a ion) ha allow li le in e ence
abou whe he Ge man is an L2. Likewise, cha ac e is ics such as “insu icien
compe ence” and “educa ionally disad an aged” a e o en ound, which in hem-
sel es do no necessa ily ha e any ela ion o GSL and which in u n a e no
o ha dly e e he subjec o empi ical s udies. Wha has been shown based on
he gene al ea u e equencies can be consolida ed wi h he help o Figu e 2
and Figu e 3 as ollows: Implici ea u es no only domina e he cons uc ion o
GSL, hey a e also ound mo e o en o gene al desc ip ion o speake s and less
o en o ope a ionaliza ion o subjec s in he con ex o empi ical s udies. This
disc epancy is pa icula ly p onounced o hose ea u es ha ha dly allow any
conclusions o be d awn abou whe he Ge man is he L2 (i.e., o he ea u es
“insu icien compe ence”, “immig a ion”, and “educa ionally disad an aged”).
So a , we ha e used ela i e equencies o ge an explo a o y o e iew o
he da a. Wha he equency calcula ions do no ell us, howe e , is how he
indi idual ea u es ela e o each o he and whe he , o example, ce ain ea u es
occu pa icula ly equen ly wi h o he s. To iden i y such o ms o ea u e co-
occu ence, we pe o med an associa ion analysis.
3.2.2 Fea u e associa ions
An associa ion analysis (Ag awal e al. 1993) is a s a is ical me hod o explo a o y
da a analysis (compa able o da a mining echniques) ha makes i possible o
iden i y pa e ns and ela ionships in la ge da a se s. I is pa icula ly use ul o
inding associa ions be ween di e en cha ac e is ics ha a e no di ec ly ela ed.
The goal o an associa ion analysis is o unco e ela ionships be ween di e en
a iables and ind possible ules o pa e ns. In he p e ious sec ion, we ha e seen
ha he majo i y o pape s in ou da a sample use implici ea u es o desc ibe
GSL speake s and ope a ionalize pa icipan s. As s a ed se e al imes, hese im-
plici ea u es a e no sui able o making s a emen s abou whe he Ge man is
an L2 o he speake s in ques ion. We ha e also seen ha he gene al desc ip-
ion o po en ial GSL speake s a e o en based on ea u es ha can be classi-
ied as de icien and ha e no explici e e ence o he psycholinguis ic no ion
o L2 acquisi ion (i.e. ea u es such as “insu icien compe ence” o “educa ion-
ally disad an aged”). Fo he empi ical s udies, we also obse ed a disc epancy
118
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
be ween gene al p ope ies assigned o speake s and he ope a ionaliza ion o
pa icipan s. We a gue ha he cons uc ion o a speake g oup ia implici and
also pa ly de icien ea u es con ibu es o he cons uc ion o O he ing. In he
hi d and inal s ep o ou analysis, we a e he e o e in e es ed in whe he pape s
ha ely on explici and psycholinguis ically mo i a ed ea u es (i.e., AoO, LoE,
LoS) a e mo e likely o e ain om implici ea u es and de ici a ibu ions han
hose ha do no men ion hese explici ea u es a all. To make such po en ial di -
e ences mo e isible, we di ided ou da a sample in o wo new subse s: Fo he
i s subse (subse 1, n=29), we ha e selec ed only hose s udies ha use explici
ea u es o ope a ionalize pa icipan s. This subse he e o e con ains only em-
pi ical pape s. The second subse (subse 2, n=87), on he o he hand, includes all
pape s ha do no use explici ea u es a all – nei he in he gene al desc ip ion
o GSL speake s no o he ope a ionaliza ion o pa icipan s. In his subse , we
included bo h empi ical and non-empi ical pape s. He e we would like o iden-
i y whe he he implici ea u es used o desc ibe and ope a ionalize speake s
o pa icipan s a e linked o each o he in a speci ic way. We use an associa ion
analysis o de e mine whe he implici (and de icien ) a ibu ions a e pa icu-
la ly p e alen whe e explici de ining ea u es a e no used.9
We pe o med an associa ion analysis using he Ap io i algo i hm om he
a ules package in R (Hahsle e al. 2005) . Using he algo i hm, we iden i ied en-
dencies ha desc ibe which ea u es end o occu oge he ac oss s udies. Such
endencies o co-occu ences o m ules (compa able o di e en weigh s) wi hin
he da a se . Two cha ac e is ic alues a e usually calcula ed as pa o he asso-
cia ion analysis: Suppo and Con idence. “Suppo ” is a measu e o how o en a
speci ic combina ion o ea u es occu s in he da a. I suppo is high, i means
ha combina ion occu s equen ly. Fo ins ance, i pape s wi hin ou da a sam-
ple con ain a ea u e A and also B, hen he suppo o he combina ion o A
and B is high. “Con idence” is a measu e o how likely i is o B o occu when
A occu s. Toge he , hese alues help us iden i y pa e ns in he da a and unde -
s and which ea u es o en occu oge he and how s ong hei ela ionship is.
In addi ion o Suppo and Con idence, o he pa ame e s can be conside ed when
e alua ing associa ion ules (e.g., Co e age and Li ). In ou conc e e example, we
es ic ed ou sel es o Suppo and Con idence, because hese pa ame e s ha e
he highes in o ma i e po en ial.
9A mo e common da a mining echnique is clus e analysis. Clus e analyses, howe e , a e
based on iden i ying simila i ies be ween obse a ions. In ou case, a clus e analysis would
ha e iden i ied pape s ha sha e speci ic ea u es. An associa ion analysis, howe e , enables us
o ind co-occu ences ac oss single da a poin s (i.e., pape s in ou case). Also, clus e analyses
usually equi e a p ede ini ion o clus e s o be iden i ied in a da a se . Associa ion analysis is
no es ic ed in such a way.
119
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
Associa ion analysis equi es he de ini ion o a con idence h eshold. We de-
cided on a compa a i ely high h eshold o 0.8 which means ha we ge an ou -
pu o hose ules whose p obabili y o occu ence is a leas 80%. By se ing his,
we ha e ensu ed ha he ules we obse e ha e a high p obabili y o ea u es
co-occu ing. In addi ion, an associa ion analysis equi es he se ing o a sup-
po h eshold which is a alue ha is se o de e mine which combina ions o
elemen s in a da a se should be conside ed ele an . I de ines he minimum
equency wi h which a combina ion mus occu o i o be conside ed. In ou
case, we ini ially se a suppo h eshold o 0.2 meaning ha we a e only in e -
es ed in combina ions o ea u es ha occu a leas 20% o he ime in ou da a.
In o he wo ds, we ocus on pa e ns ha occu ai ly equen ly. Fo subse 2,
howe e , we ound ha he e is only one ule ha mee s he Con idence and
Suppo h esholds. To make mo e ules isible, we he e o e lowe ed he Con i-
dence h eshold o subse 2 o 0.6 and Suppo o 0.1. The wo esul s o subse s
1 and 2 hus di e ed in he equency o he ules we obse ed. In consequence,
he ules obse ed o subse 2 a e weake han hose o subse 1. Below we i s
see he esul s o he iden i ied ules in a able o ma , each o subse s 1 and
2. The “lhs” (= le -hand side) column in he Table 2 and Table 3 lis s he ea-
u es ha we e iden i ied as condi ion- o i -elemen s o a ule. I mo e han one
ea u e is lis ed in his column, i means ha hese ea u es occu oge he as a
condi ion o he ule. The o de o he a iables in he column is no meaning ul
and has no signi icance o he in e p e a ion o he ule. The column “ hs” (=
igh -hand side) e e s o he a iables ha a e included in he ule as in e ence-
o hen-elemen s. The ables hus show he equency (= Suppo ) and he like-
liness (= Con idence) o i - hen-elemen s ( ha oge he o m ules) wi hin he
da a se . The column “coun ” s a es he numbe o pape s in which he obse ed
ule occu s.
As we can see o bo h Table 2 and Table 3, he e is a clea con as be ween he
Suppo and Con idence alues. We ind compa a i ely high Con idence alues,
bu a he low Suppo alues. This means ha he i - hen- ela ions we ound in
he da a a e ai ly s ong in he sense ha he likeliness o co-occu ence is e y
high. A he same ime, he co-occu ences do no occu e y equen ly. This is
especially ue o subse 2: since we had o lowe he Con idence and Suppo
h esholds, he i - hen- ela ions a e gene ally lowe . This, in u n, may imply,
ha he e is a ce ain a bi a iness as o which ea u es a e associa ed wi h GSL
and ha he e is seman ic agueness and ambigui y in he cons uc .
To be e unde s and he nume ic esul s om Table 2 and Table 3, he esul s
o he associa ion analyses a e p esen ed in Figu e 4 and Figu e 5. Figu e 4 ep-
esen s he esul s based on Table 2 and Figu e 5 hose based on Table 3. Bo h
120
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
Table 2: Resul s associa ion analysis subse 1, n=29 pape s which men-
ion explici ea u es
condi ion/i -elemen s in e ence/ hen-elemen s suppo con idence coun
{op_LoS} →{op_LoE} 0.21 1.00 6
{ h_disad ana aged} →{ h_biling} 0.21 1.00 6
{ h_mul iling} →{op_o he L} 0.25 1.00 7
{ h_mul iling op_LoE} →{op_o he L} 0.21 1.00 6
{ h_biling op_o he L} →{op_LoE} 0.25 1.00 7
{ h_immig a ion} →{op_LoE} 0.25 0.87 7
{ h_o he L, op_o he L} →{op_LoE} 0.25 0.87 7
op_o he L op_LoE} →{op_o he L} 0.25 0.87 7
{ h_biling op_LoE} →{op_o he L} 0.25 0.87 7
{ h_mul iling} →{op_LoE} 0.21 0.86 6
{ h_insu icien comp} →{op_LoE} 0.21 0.86 6
{ h_mul iling op_o he L}→{op_LoE} 0.21 0.86 6
Table 3: Resul s associa ion analysis subse 2, n=87 pape s which do
no men ion explici ea u es
condi ion elemen s (= lhs) in e ence elemen (= hs) suppo con idence coun
{ h_disad an aged} →{ h_biling} 0.14 0.81 13
{op_mul iling} →{ h_mul iling} 0.13 0.75 12
{op_mul iling} →{op_o he L} 0.13 0.75 12
{ h_mul iling, op_mul iling}→{op_o he L} 0.10 0.75 9
{op_o he L, op_mul iling} →{op_mul iling} 0.10 0.75 9
{ h_mul iling, op_o he L} →{op_mul iling} 0.10 0.75 9
{ h_biling, op_o he L} →{ h_o he L} 0.12 0.68 11
{ h_mul iling} →{ h_biling} 0.20 0.64 18
{ h_disad an aged} →{ h_o he L} 0.11 0.62 10
{ h_immig a ion} →{ h_o he L} 0.13 0.60 12
121
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
igu es we e d awn manually, bu a e based on a isualiza ion o esul s using
he packages a ulesViz and ggg aph in R (c . o applica ion o packages o as-
socia ion analyses Hahsle & Ka pienko 2017). The i -condi ions a e colo ed in
whi e, and he hen-condi ions in g ey ci cles. A ows show he i - hen ela ions
be ween he ea u es. The wid h o he a ows ep esen s he suppo alues,
which a e also indica ed in he igu e legend.
As can al eady be seen om Table 2, we see in Figu e 4 wo main nodes in
hose empi ical pape s ha use explici ea u es o ope a ionalize pa icipan s:
“o he language” and “leng h o exposu e”. I is s iking ha he wo nodes go
hand in hand wi h each o he : “LoE” is an i -condi ion o “o he language” and
ice e sa. The a he s ong connec ion be ween hese wo ea u es is la gely
plausible: I LoE is eco ded, i also seems logical o include o he language(s)
o pa icipan s o ope a ionaliza ion. Also, he “ h”- ea u es used o desc ibe
GSL speake s wi hin he heo e ical sec ion o he pape s ollow a la gely con-
sis en pa e n: we p edominan ly ind “ h”- ea u es such as “mul ilingual” and
“bilingual” o he gene al desc ip ion o GSL speake s. One excep ion is he “ h”-
ea u e “insu icien compe ence”. In addi ion, he isola ed pai o nodes a he
bo om igh o he igu e is also s iking: comple ely de ached om he ope a-
ionaliza ion ea u es LoE and “o he language”, we ind a connec ion be ween
he “ h”- ea u es “bilingual” and “disad an aged”. Toge he wi h he “insu icien
compe ence” ea u e, his is an indica ion ha e en s udies ha make use o ex-
plici and psycholinguis ically mo i a ed ea u es o he ope a ionaliza ion o
pa icipan s use implici (and de icien -o ien ed) ea u es o desc ibe GSL speak-
e s. Howe e , such a ibu ions seem o be a he a e in hese pape s. Whe e
speake s a e iden i ied as GSL lea ne s based on he ea u e LoE, ea u es ha
a e a logical consequence o L2 acquisi ion (i.e., bi-/mul ilingualism, p esence o
o he languages in addi ion o L2 Ge man) p edomina e. This is di e en in pa-
pe s ha do no men ion he explici ea u es AoO and LoE a all (i.e. nei he
o he gene al desc ip ion o GSL speake s no o ope a ionaliza ion o pa ici-
pan s). In such pape s, we see ou main nodes in Figu e 5: “o he language” (as an
“op”- ea u e), “mul ilingual”, “bilingual” and “o he language” (all h ee as “ h”-
ea u es). Wi h espec o he ope a ionaliza ion o pa icipan s wi h he help o
he ea u e “o he language”, we ind co-occu ences wi h “ h”- ea u es like ‘mul-
ilingual’ and “bilingual”. Al hough hese co-occu ences a e in line wi h he pa-
pe s ha e e o mo e explici ea u es (mainly LoE, see Figu e 4), i emains un-
clea whe he an ope a ionaliza ion o GSL speake s h ough he ea u e “o he
language” is jus i ied. Fu he mo e, we see in Table 3 and Figu e 5 ha he “ h”-
ea u e “educa ionally disad an aged” igge s he ea u es “bilingual” as well as
“mul ilingual”. As al eady poin ed ou in sec ion 3.2.1, he ea u e “educa ionally
122
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
Figu e 4: Associa ions o ea u es in subse 1
123
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
Figu e 5: Associa ions o ea u es in subse 2
124
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
disad an aged” is ne e used o ope a ionaliza ion. Whe he he ela ions o
mul ilingualism, GSL and educa ional disad an age sugges ed by such egula i-
ies a e empi ically e i ied and enable emains open. O e all, he associa ion
analysis shows no pa icula ly la ge di e ences be ween pape s ha use explici
ea u es o ope a ionalize subjec s and hose ha do no men ion he ea u es a
all. In bo h da a subse s, we see ha GSL speake s a e p edominan ly desc ibed
as bi- o mul ilingual and as speake s o an “o he language”. The e is also a en-
dency in bo h subse s o use implici de ici cha ac e is ics such as “insu icien
compe ence” and “educa ionally disad an aged” as pa o gene al desc ip ions
o speake s. Finally, he main di e ence be ween he wo subse s is ha hose
who ely on explici ea u es o ope a ionalize pa icipan s can s a e wi h a high
deg ee o ce ain y ha hei pa icipan s a e eal L2 lea ne s. A he same ime,
his does no jus i y he de icien a ibu es ha we ne e heless ind in he da a.
Fo hose pape s ha do no con ain any explici ea u es, we only know ha he
pa icipan s o speake s a e bi- o mul ilingual and speak a language o he han
Ge man. Whe he hey a e L2 lea ne s in he na owe sense emains comple ely
open.
4 Discussion
Ou e iew s a ed wi h he ques ion o how Ge man as a Second Language is
concep ualized in schola ly publica ions on he subjec , which concen a e on
child and adolescen lea ne s. The esea ch ques ion is based on he obse a ion
ha GSL seems o be a a he ague concep , implying a de icien pe spec i e on
i s clien ele. To es ou esea ch ques ion, we e iewed a da a sample o n=138
pape s ha we e coded acco ding o explici and implici ea u es ela ed o GSL.
In he i s explo a o y analysis, we de e mined which ea u es a e gene ally e-
quen and whe he he e a e di e ences be ween empi ical and non-empi ical
s udies. We also assessed he ela ionship be ween gene al desc ip ions o GSL
speake s and he ope a ionaliza ion o pa icipan s in he s udies. In he second
s ep, we used associa ion analyses o de e mine whe he he e a e egula ela-
ions be ween ea u es in s udies ha use explici ea u es o ope a ionaliza ion
(= subse 1) and hose ha do no men ion explici ea u es a all (= subse 2).
Ou esul s can be summa ized as ollows:
1. Implici ea u es occu no iceably mo e o en han explici ones, ega dless
o he ype o pape (i.e., empi ical s. non-empi ical).
125
Jana Gampe , Ch is oph Sch oede , Julia Schlauch & Do o heé S einbock
2. The mos equen ea u es a e linguis ic ea u es, bu mainly hose ha
do no allow o conclude whe he Ge man is an L2 (i.e., “bilingual”, “mul-
ilingual” as well as “o he language”). These ea u es shape he gene al
desc ip ion o GSL speake s o a la ge ex en and a e also equen ly used
o he ope a ionaliza ion o pa icipan g oups in empi ical s udies.
3. The ea u es “insu icien compe ence” (he e conce ning Ge man) and “ed-
uca ionally disad an aged” a e also equen ly used. These non-linguis ic
pa ame e s a e mainly ound in he ield o heo e ical desc ip ions o GSL
bu a e ha dly e e used o ope a ionaliza ion in empi ical s udies.
4. Bo h pape s ha use explici ea u es o ope a ionaliza ion and hose ha
do no men ion hem a all base hei gene al desc ip ion o GSL speake s
and he ope a ionaliza ion o pa icipan s on he ea u es om poin 2.
Rega dless o whe he he pape s use explici ea u es o ope a ionalize
pa icipan s, we ind de icien a ibu ions using ea u es om poin 3.
Based on hese esul s, he ini ial obse a ion ha GSL seems o be a a he
ague and unspeci ic cons uc can be con i med, a leas o ou da a sample. I
is unce ain whe he Ge man can be conside ed an L2 o he speake s desc ibed
in he non-empi ical pape s and he pa icipan s in empi ical s udies, in mos
cases. This is because only a small numbe o pape s a e based on compa a i ely
s ic acquisi ion- heo e ical c i e ia (i.e., AoO, LoE and LoS). In he majo i y o
he pape s, GSL speake s a e concep ualized as a g oup o speake s who a e bi-
and/o mul ilingual, speak a language o he han Ge man a home, a e educa-
ionally disad an aged, and/o ha e insu icien Ge man language skills. The ac
ha explici c i e ia o he de ini ion o GSL speake s a e a he a e is p oblem-
a ic o wo easons: Fi s , he g oup’s second language p o iciency in Ge man
is unce ain due o he lack o explici c i e ia. Second, he lack o explici c i e-
ia allows o he c ea ion o a cons uc o GSL speake s ha may no exis in
eali y o may be highly biased.
Ou da a analysis sugges s an answe o ou second esea ch ques ion: The
GSL cons uc , as pu sued in he academic ield o GSL in Ge many indeed has
endencies owa d a de ici iew o GSL speake s in pa icula and mul ilingual-
ism in gene al. The ac ha bi- and mul ilingualism, inadequa e Ge man lan-
guage skills, and educa ional disad an age appea o be en angled in ou da a
in such a dominan way sugges s ha a discou se is being ei e a ed he e ha
we o iginally ound in non-expe educa ional policy deba es. La ge-scale s udies
such as PISA o educa ional ends o Ge man, among o he s, ha e sugges ed, a
126
6 Cons uc ing a clien ele in need: The ield o Ge man as a Second Language
leas in hei ea ly esul s, ha mig a ion- ela ed mul ilingualism is p oblema ic
when i comes o achie ing adequa e Ge man compe encies and high educa ional
a ainmen (c . A el e al. 2003: 51 .). This ini ial in e p e a ion has since been
signi ican ly weakened (c . Reiss e al. 2019: 158 .). Ins ead o mig a ion- ela ed
mul ilingualism, social ac o s (independen o s uden s’ linguis ic backg ound)
seem o in luence he a ainmen o ce ain language compe encies and deg ees.
GSL esea ch should co ec his o m o bias wi h he help o accu a e me hod-
ological designs, ins ead o using labels such as GSL o mul ilingualism o sup-
po p oxy deba es (p ecisely because impo an causal ac o s such as social
inequali y seem o be mo e impo an easons o di e en school pe o mance).
The e is eason o be conce ned ha such p oxy deba es s igma ize GSL lea ne s
as well as (ce ain) mul ilingual speake s. Such an in e p e a ion o ou da a does
no mean ha he e a e no s udies ha do no add ess his issue - he e ce ainly
a e. Howe e , in ou da a se , we ind a clea endency owa d me hodological
as well as concep ual imp ecision coupled wi h he a o emen ioned co ela ion
wi h mo e de ici educa ion- ela ed ea u es. Ano he endency ha is e iden in
ou da a is he concep ualisa ion o GSL as linguis ic O he ness. The dominance
o he cha ac e is ic ”o he language” ( han Ge man) in ou da a shows ha he
me e exis ence o a language o he han Ge man in he amily con ex leads o a
classi ica ion in which he mul ilinguals a e no compe en speake s o Ge man -
and hus ’o he ’ om he monolinguals (c . also Szymczyk e al. 2022). Al hough
we did no es whe he he cha ac e is ic co ela es wi h concep s o O he ing, i
would be wo hwhile in es iga ing such co ela ions in u u e s udies. Ou ind-
ings highligh p oblema ic endencies in he GSL esea ch communi y o pa ic-
ipa ing in pa s o a de ici discou se and sha e esponsibili y o discu si ely
ea ing mul ilingual speake s in gene al and GSL lea ne s in pa icula as O he .
This is a ha sh accusa ion. Le us s ess again ha he au ho s o his s udy by
no means exclude hemsel es om his p oblem. Ou s udies, oo, o en de ine
GSL h ough implici and ague ca ego ies and engage in a o m o O he ing.10
The e o e, ou e iew in no way in ends o de ame indi iduals o esea ch ap-
p oaches. Ra he , i aims o make ce ain, in ou iew p oblema ic, app oaches
10We a e awa e o he ac ha i is s a e o he a o include a monolingual con ol g oup
in empi ical s udies. Howe e , he e a e nume ous p oblems o his, one o hem being ha
monolinguals ha dly e e cons i u e a homogeneous g oup o speake s bu end o show high
deg ees o he e ogenei y (see also Alexiadou 2025 [ his olume]). I is he e o e necessa y o
ca e ully e alua e whe he a compa ison be ween mul ilinguals and monolinguals is heo e i-
cally g ounded and pu pose ul. In any case, a con ol g oup should be well jus i ied and should
no be an au oma ism, as his can con ibu e o concep ualizing GSL and mul ilingual speake s
as o he s.
127
