Content uploaded by Kristin Kersten
Author content
All content in this area was uploaded by Kristin Kersten on Dec 16, 2021
Content may be subject to copyright.
Kersten (2022) – The Proximity of Stimulation Hypothesis
1
Chapter 4: The Proximity of Stimulation Hypothesis: Investigating the interplay
of social and instructional variables with the cognitive-linguistic skills of young
L2 learners
Kristin Kersten (Hildesheim University)
Abstract
Factors shaping human cognitive and linguistic development are intertwined and found within
a nested structure of conceptual levels. Proximal levels contain concrete stimuli to the learner,
while distal variables only exert indirect effects, and often represent container variables made
up of numerous proximal ones. The Proximity of Stimulation Hypothesis holds that effects are
best explained using proximal factors with immediate effects on the learner. This study
examines the impact of exemplar influencing variables at different conceptual levels on L2
lexical and grammar reception, working memory and phonological awareness. Structural
equation modeling with mediator analyses accounted for the hierarchical data structure of 93
L2 learners of English in German conventional and bilingual primary schools. Results
supported the proximity hypothesis in that the effect of both distal variables, SES and L2
program, on internal variables was partially mediated by proximal variables (parental
language/literacy support and teacher’s input quality, respectively). L2 program also predicted
L2 lexicon and phonological awareness without a mediating effect, showing the effectiveness
of bilingual teaching programs. Additionally, parental language/literacy support predicted L2
lexicon, teachers’ patience predicted L2 grammar and phonological awareness, and children’s
degree of multilingualism predicted L2 grammar. Phonological awareness correlated with L2
grammar, corroborating the interconnectedness of cognitive-linguistic development and a
cognitive advantage hypothesis.
Introduction
Research into factors that predict and explain second language acquisition (SLA) has
traditionally focused on the relationship between two or only a few variables (Han, 2016).
Studies have usually been situated within specific theoretical strands within SLA discourse.
Lately, however, researchers have increasingly deplored the lack of a systems approach within
SLA, as exemplified in this quote by Han (2019, Ch. 7, p. 7):
the SLA field as a whole, despite its decades of existence, has come up short of a fuller and more credible
understanding of the evolving nature of learning […] in response to the changing configuration of
interaction between learner-internal and learner-external resources.
Newer approaches call for a broader view on SLA as an interplay of various factors which have
been identified in previous research and are situated in nested relationships. As language
learning is strongly intertwined with cognitive processes, and embedded in and shaped by the
social environment, they point out the necessity to widen the scope of SLA research and go
beyond the boundaries of our own field to include insights from research from adjacent fields
such as psychology and sociology (Mitchell et al., 2019). According to Atkinson (2019, p. 115),
“the continued lack of theoretical integration of the cognitive, the social, and the material is
To appear in: Kersten, K., Winsler, A. (eds.) (2022). Understanding variability in second language
acquisition, bilingualism, and cognition – A multi-layered perspective. London: Routledge.
Kristin Kersten, English Department, Hildesheim University, Germany. E-Mail: kristin.kersten@uni-hildesheim.de
Kersten (2022) – The Proximity of Stimulation Hypothesis
2
the single greatest barrier to transdisciplinarity in SLA studies”, i.e., a barrier to a “viable
transdisciplinary framework for SLA” (p. 119).
Following calls such as these, the last decade has seen an upsurge in interdisciplinary
approaches to make sense of what is increasingly seen as a complex and dynamic system of
factors shaping the (language) learning process (Mitchell et al., 2019). Only rather recently
have attempts been made to develop more overarching frameworks that focus on the
entanglement and interdependence of such factors. These approaches come from very
different strands of SLA including Complex Dynamic Systems Theory (Lenzing et al., 2022, this
volume), the Douglas Fir Group (2016) and the research strand it has stimulated (Duff &
Byrnes, 2019), or the systemic but more individual- and cognition-oriented, Modular Cognition
Framework (Truscott & Sharwood Smith, 2019).
These developments pose numerous theoretical and methodological challenges. Traditionally,
studies including a restricted number of variables rely on the general linear model for data
analysis. Yet a systemic view of influencing variables first of all needs to account theoretically
for their complex relationships, i.e., their different conceptual levels, conceptual ‘distance’
(distal and proximal external factors, internal factors), and their interdependence (i.e.,
container variables). Secondly, the complex nature of these interrelations calls for different
analytical approaches (see Kersten & Greve, 2022, this volume, for a detailed discussion).
This chapter focuses on the interplay between external factors affecting internal linguistic and
cognitive competences. In that sense, it is an empirical attempt to follow a line of thought
outlined in Kersten and Greve (2022, this volume). It argues that effects of external on internal
factors can best be described by proximal rather than distal external factors. This will be
referred to as the Proximity of Stimulation Hypothesis. To that end, external factors need to
be conceptually disentangled. Models on the interplay of external and internal factors will be
introduced and conceptual consequences for the operationalization of hierarchically ordered
proximal and distal variables in empirical SLA studies will be discussed. Methodologically, the
difference between statistical prediction for distal variables and causal explanation for
proximal ones will be addressed.
The empirical study reported in the second part of the paper is constructed, accordingly, to
test predictions of the Proximity Hypothesis. Data on several exemplary distal, proximal
external, and cognitive-linguistic internal variables for which strong predictive effects have
been described in the literature were collected from 93 young German learners of English in
primary school. These variables include socioeconomic status and type of school program
(meso-level), different measures of home parent-child interaction and teachers’ L2 input
quality in the classroom (micro-level), cognitive skills (working memory, phonological
awareness), and degree of multilingualism and L2 competence (nano-level). Data were
analyzed using structural equation modeling to account for the conceptual interplay of the
variables. The goal of the study was, thus, to illustrate exemplary interrelationships of
hierarchically structured variables rather than to give a near complete overview of influencing
external factors.
Kersten (2022) – The Proximity of Stimulation Hypothesis
3
The interplay of external and internal factors in (language) learning
A number of models have been put forth in recent research which address the complex nature
of intertwined factors.
Models
Model of Cognitive Growth and Language Development
In 1991, Paul van Geert pointed out the role of internal and external factors for the learning
process, depicting learning as the growing of a cognitive system in an interplay of internal
(biological) resources and resources in the external social environment. To describe this
process, he differentiated between internal and external systems in his model, with four types
of resources within these systems, respectively, i.e., spatio-temporal, informational,
motivational-energetic, and material resources. More precisely, internal resources comprise:
information processing capacity, time on task (spatio-temporal resources)
prior knowledge and skills (informational resources)
amount of energy, effort, activation during acquisition (motivational-energetic
resources)
physical properties, sensory/nervous system and beyond (material resources)
Analogously, external resources entail:
spatial and temporal freedom in environment (spatio-temporal resources)
type, amount, availability of information for learning (informational resources)
reinforcement / payoffs provided by environment (motivational-energetic resources)
objects, physical materials available to the learner (material resources)
According to this model, these resources constitute the relevant variables to be taken into
account when explaining cognitive-linguistic development. In other words, in cognitive and
linguistic development, learners draw from factors which are individual, i.e., internal to their
own system, and from external factors in their immediate environment. To give an example
relating to the study at hand, individual internal informational resources include, for instance,
the prior world knowledge and the linguistic skills that a learner has already acquired, while
external informational resources include available external linguistic input in the environment
(provided by parents and teachers) that leads to language learning. What is notable and
relevant here is that the model refers to external factors that directly stimulate the individual,
which are referred to here as proximal external factors.
The Modular Cognition Framework
This immediate, proximal interaction of the individual with factors in the environment is also
at the center of Sharwood Smith and Truscott's Modular Cognition Framework (MCF)
(Sharwood Smith & Truscott, 2014, Sharwood Smith, 2017).
1
The MCF describes in detail
memory stores, their respective processors, and their connecting interfaces that form the
internal basis for learning. In the internal context, cognitive systems comprise the auditory,
somato-sensory, visual, gustatory, and olfactory perceptual systems, complemented by the
conceptual and affective systems and two linguistic systems for phonological and syntactic
processing and storage, all of which are richly interconnected. Figure 4.1 depicts how external
Kersten (2022) – The Proximity of Stimulation Hypothesis
4
information that is encountered in the environment reaches the individual exclusively in the
form of the different types of sensory stimuli, which are then further processed within the
different cognitive systems.
Figure 4.1. Interfaces of proximal external stimuli with internal memory stores and processors in the Modular
Cognition Framework; interaction of internal context with external environment is expressed as a dotted line.
(This figure is taken from the homepage of The Modular Cognition Framework as Fig. 7, The MCF [stores, their
respective processors and connecting interfaces between the stores]
https://www.cognitionframework.com/home-1/mogul-architecture-the-basics/. It is slightly adapted from
“Figure 12.1 – The MCF mind: The generator of internal context,” Truscott, J. & Sharwood Smith, M., pp. 275, in
“The Internal Context of Bilingual Processing” (2019) published by John Benjamins Publishing Company
Amsterdam/Philadelphia; see https://benjamins.com/catalog/bpa.8. Reprinted with permission.)
With this focus on the immediate environment as a trigger of internal processes, both van
Geert and Truscott and Sharwood Smith describe a proximal view of resources, i.e., those
internal resources of both the learner and the external environment which are in direct
interaction with each other. Two other influential models are compatible with this view, but
extend it to take into account additional more distal layers of external factors.
Developmental Conceptual View of Human Development
One such model from the field of psychology is Richard Lerner’s (2002) Developmental
Conceptual View of Human Development. His overarching model (p. 211) shows internal
properties of the individual (cognitive, biological, personality, attitude, health and other
factors) depicted in immediate (i.e., proximal) contact with the direct social environment of
family members, peers, teachers (family/social/school network), and in indirect contact with
factors on higher external (distal) levels. These distal levels include different social contexts
encompassing the central level in increasingly removed concentric circles of the community,
the society, the culture, and designed and natural environments. Double-headed arrows
Kersten (2022) – The Proximity of Stimulation Hypothesis
5
indicate that all of these factors are in dynamic interplay with each other, and are all subject
to change over time. Thus, in contrast to the two previous models, Lerner introduces various
distal factors as several layers of higher order factors, in addition to proximal ones that interact
directly with the individual. These are depicted as conceptually further and further remote
from the actual point of interaction between individual and environment.
Context Dimensions in SLA: The Multifaceted Nature of Language Learning and Teaching
More recently, this comprehensive view has also been adopted by a transdisciplinary group
of researchers in SLA which has come to be known as the Douglas Fir Group (2016), whose
intention was “to integrate cognitive and social dimensions of language learning at several
interconnected levels” (Mitchell et al., 2019, p. 373). According to the authors,
Language learning is a complex, ongoing, multifaceted phenomenon that involves the dynamic and
variable interplay among a range of individual neurobiological mechanisms and cognitive capacities and
L2 learners’ diverse experiences in their multilingual worlds occurring over their life spans and along
three interrelated levels of social activity: the micro level of social action and interaction, the meso level
of sociocultural institutions and communities, and the macro level of ideological structures. (The
Douglas Fir Group, 2016, p. 36)
Their graphical illustration of these levels of activity (2016, p. 25) is compatible with the other
views in that the individual neurobiological and cognitive basis represents the level of internal
factors, the micro-level of social activity in which “individuals are engaging with others”
represents the first external proximal level of “multilingual contexts of action and interaction,”
while the meso-level refers to the second, distal, level including families, schools, and social
organizations. The third, distal, macro-level includes, according to the model, the systems of
beliefs and values in which sociocultural institutions are embedded. The model was partly
developed by researchers within Complex Dynamic Systems Theory (Lenzing et al., 2022, this
volume), and is compatible with it.
While tremendously helpful for conceptualizing the interrelations of different variables, the
actual mechanisms of how they affect the learner and lead to changes of internal
representations, i.e., learning, are yet to be described (Han, 2016). The approach outlined in
this paper aims to present one step toward this goal.
Conceptual consequences for the operationalization of proximal and distal variables in
empirical (SLA) studies
These four models, which stem from different decades and different disciplines, have in
common that they differentiate between different conceptual levels of variables, i.e., internal
variables as well as several layers of external variables that are all closely intertwined. Such a
framework to explain language learning processes has important theoretical and
methodological implications. As pointed out before (Kersten & Greve, 2022, this volume), it is
common practice to capture the influence of several variables on others in an additive way in
multivariate analyses. We hold, however,
that these types of analysis are often not sufficient to discover and/or confirm the causal effects of
factors on different hierarchical levels, because they remain “blind”, so to say, to a number of
theoretical problems […] [e.g.,] that factors may be related to each other in different conceptual ways.
If these conceptual relations are disregarded, a prediction or a causal explanation can be incomplete or,
at worst, downright wrong. (Kersten & Greve, 2022, this volume)
Kersten (2022) – The Proximity of Stimulation Hypothesis
6
Such conceptual relations include variables that are interrelated because they are
conceptually ‘contained’ in a higher-order variable on a more distal level. This practice, thus,
raises several questions: What does it actually mean that a school program “influences”
language acquisition, or that socioeconomic status “affects” cognitive development? Who and
what is it exactly that influences the learner? How can this influence be conceived of, both
theoretically and practically? How can researchers methodologically account for effects of
variables on different levels, which may or may not be contained within each other?
It is not reasonable to assume that a distal container variable such as type of school program,
or a family’s socioeconomic status (SES) – variables often used in regression analyses – exerts
a direct influence on the individual. This holds for so-called ‘proxies’ used for statistical
analyses, as well, such as the number of books in a family used as a proxy for SES. The question,
then, arises as to which aspects contained within these distal variables actually have the
power and chance to affect changes within the individual. Good, and frankly the only,
candidates for this are interfaces where the learner experiences direct stimulation from
environmental input. In other words, the immediate influence on a learning process manifests
itself at the intersection of concrete, sensory experience and cognitive engagement, leading
to individual knowledge construction (see Kersten, in press). It is for that reason that Truscott
and Sharwood Smith (2019, p. 10) define ‘input’ in SLA as
sights, including pointing and gesturing, sounds, smells, tastes, etc., in other words everything that
contributes to the interpretation of an utterance and which can lead to further development of an
individual’s linguistic ability, i.e. all the relevant external contexts. This should be included in a
comprehensive understanding of what input is.
It must therefore be assumed that externally induced change at the level of the individual can
only be achieved by direct stimulation via the threshold of the sensory organs, which convey
the stimuli of the outside world to the internal system.
2
Variables that describe these concrete
interactions will be referred to as proximal variables. It follows that anything that does not
describe an immediate interaction of the individual with the outside world is a variable of
higher order, on a higher conceptual level further removed from the learner, i.e., a distal
variable. Following this logic, distal variables can only exert an influence on the individual via
proximal variables. This is a typical mediator relationship (Kersten & Greve, 2022, this volume).
To give an example, a distal variable such as parental education can only exert an influence on
the individual via a proximal variable such as the actual stimulating interaction of a parent
with a child (Figure 4.2).
Distal factors such as parental education or SES are often used as proxies in statistical analyses.
What complicates matters, however, is that such distal variables as parental education and
parental SES are ‘container variables’ which may contain numerous other variables on
different conceptual levels. SES might, for example, be conceptualized as containing home
resources in addition to parental education. Material resources, however, do not ‘exert any
influence’ through their mere presence but only if the learner interacts with them. This means
that the presence of, for example, books in a home might be used as a proxy in an analysis,
i.e., as an indicator of something else that actually impacts the child.
Kersten (2022) – The Proximity of Stimulation Hypothesis
7
Figure 4.2. Variables on different conceptual levels affect the individual in different ways (proximal variables in
direct interaction on the micro-level; distal container variables in indirect interaction on the meso-level)
A container variable such as SES might involve several mediators on different conceptual levels
(Figure 4.2). Consequently, each relationship would explain a part of the whole (partial
mediation). In addition, there might be other proximal variables which are not part of the
container, i.e., which are not predicted by the distal variables within the calculation – for
example if mothers interact intensively with their children no matter what social background
they come from (depicted as parent-child interaction 2 in Figure 4.2).
3
Moreover, variables
within the container might also be responsible for effects from so-called emergent
phenomena (banded arrow) (Noordhof, 2010). Emergence cannot be explained by the sum of
effects of single parts as is assumed in statistical analyses; rather, such effects emerge from a
complex interplay of factors as something completely different, creating “something more”
than just the sum of its parts. An example of emergent phenomena is consciousness. This line
of thought is out of the scope of this paper, but see Kersten and Greve (2022, this volume) for
elaboration.
Methodological consequences for the operationalization of proximal and distal variables in
empirical (SLA) studies
Consequently, one has to be aware that many of the factors that are often measured as
independent predictors in competition with other ones in statistical models cannot be
conceptualized on the same theoretical level. If these relationships are not disentangled prior
to statistical analysis but are used concurrently as if belonging to the same level, this might
have severe consequences for the interpretation of the results. Container variables, which
consist of numerous different, proximal, micro-level predictors, can be assumed to have more
statistical power than single proximal variables, even though the latter are, theoretically, the
ones with the better, or more pure explanatory value. This can have a considerable influence
on effect sizes (Kersten & Greve, 2022, this volume). In other words, if a distal factor instead
of a proximal one is used in a calculation, the actual effect may be overestimated because the
Kersten (2022) – The Proximity of Stimulation Hypothesis
8
distal one may well include other mediators for the same behavior (parallel mediation), but
also various other variables for further proximal influences. Therefore, results could be biased
when higher order distal container variables are calculated in direct competition with proximal
ones in a linear model, because the distal ones potentially contain more options of
unmeasured effects and, therefore, will presumably achieve larger effect sizes. This renders
explanatory interpretations difficult and imprecise because the actual cause is ‘hidden’ within
the large container, so to speak, and it is not apparent exactly which proximal factor is the
determining factor for the effect. In other words, distal variables might function (well) as
statistical predictors, while proximal variables have greater potential for causal explanation.
To give an example from the current study, while (distal) type of L2 program might predict L2
attainment, (proximal) quantity and quality of L2 input the learners receive have greater
potential to explain it.
For reasons such as these, it is helpful for a theoretical understanding of actual explanatory
effects to define conceptual levels of variables involved in an investigation, to include proximal
variables of direct interaction with the individual, and to account for their interrelated
structure using appropriate techniques of analysis such as mediator or moderator analyses,
structural equation models, or multilevel models for a nested structure of variables.
4
The
Proximity of Stimulation Hypothesis, which takes these premises into account, will be
formulated below.
Operationalizing the interplay of variables
Hierarchical model of SLA predictors
To enable the operationalization of variables affecting SLA on different conceptual levels,
Figure 4.3 depicts central internal and external variables which have previously been found to
be related to SLA. The illustration is derived from and compatible with the above-mentioned
models by van Geert (1991), the Douglas Fir Group (2016), Lerner (2002, p. 211), Sharwood
Smith (2017), and Truscott and Sharwood Smith (2019). The selection of variables is intended
to be exemplary, not exhaustive, and can serve as a basis for research designs with different
or additional variables.
Cultural environment is conceptualized on a macro-level in this hierarchy, and social and
institutional environments on the meso-level. The level where actual social interactions take
place with family members, peers, and teachers in the classrooms, is depicted as micro-level.
Learner-internal variables such as personal traits, cognitive skills, affect, and language
competences are seen at the nano-level. The icon of a thermometer represents the perceived
emotional atmosphere in which learning takes place. This can include for example stress-levels
in the family (Lawson et al., 2016) or the perceived emotional climate of the classroom
(Dewaele, 2022, this volume). Across disciplines, all of these aspects have been used as
variables in empirical studies and shown to predict learning and development in general, and
language learning in particular (for an overview, see Kersten, 2020).
Kersten (2022) – The Proximity of Stimulation Hypothesis
9
Figure 4.3. Hierarchical model of SLA predictors – exemplary representation of variables hypothesized to affect
SLA on different conceptual levels (adapted from Kersten, 2020, p. 83), based on the ‘Model of Cognitive Growth
& Language Development’ (van Geert, 1991), the ‘Modular Cognition Framework’ (Sharwood Smith, 2017), the
‘Developmental Contextual View of Human Development’ (Lerner, 2002, p. 211), and the ‘Multifaceted Nature
of Language Learning & Teaching’ (Douglas Fir Group, 2016, p. 25)
From prediction to explanation: Proximity of Stimulation
To sum up, the effect of a container variable on the learning process is mediated by proximal
external variables, which induce a stimulus for the learner-internal system via direct
interaction. It is only possible for an external stimulus to reach the learner if it passes the
internal threshold via sensory organs to enter the internal system, as described in the Modular
Cognition Framework (Sharwood Smith, 2017). As an example, one might expect to find
mediating effects of parent-child interaction for the effect of SES on language acquisition, or
of the form of classroom interaction for the effect of school program on language acquisition.
To be sure, both distal and proximal factors can predict development in empirical studies.
However, it should be the ambition of any theoretically driven empirical research to explain
relationships. Only immediate proximal factors (i.e., as interaction with the environment)
present stimuli for the learner, hence, exert influence, and could therefore be regarded as
directly causal for a change in the internal system. If a distal factor predicts development, this
influence must be mediated or moderated by proximal stimuli. Other proximal factors that do
not depend on a distal one within a specific analysis may have additional independent effects.
Kersten (2022) – The Proximity of Stimulation Hypothesis
10
Following this argumentation, it is assumed that in empirical studies, effects are best explained
using factors that describe the direct point of contact with the learner, i.e., proximal factors. I
call this the Proximity of Stimulation Hypothesis.
This approach predicts that
1. distal and proximal variables can be shown to have a (partial or full) mediator
relationship, in which the proximal variable mediates the effect of the distal one (Vdistal
Vproximal Vdependent)
2. part of the variance within learner variables is likely explained by other, independent
proximal factors that are not contained in the distal variables in a specific calculation.
For this purpose, the conceptual relationships of all variables within a research design first
need to be identified and theoretically justified, which will be attempted in the empirical study
discussed below.
The current study
The study is part of the project Studies on Multilingualism in Language Education (SMILE),
carried out from 2014-2019 at Hildesheim University. The aim of the project was to collect
dense cross-sectional and longitudinal data from young learners of English that included
individual information on all levels as shown in Figure 4.4.
Following the Proximity of Stimulation Hypothesis, structural equation modeling was used to
investigate the interplay of direct and mediator relationships between variables on the meso-
level (SES and L2 program), on the micro-level (parent-child interaction [PCI] and teachers’ L2
input quality), and on the nano-level (cognitive skills, operationalized as working memory and
phonological awareness, and receptive lexical and grammatical L2 attainment [L2A]) in a cross-
sectional data set.
Research questions
The study investigates the following questions:
1. Do proximal variables in the data set (PCI, teachers’ input quality) mediate the
effect of distal container variables (SES, L2 program) on internal variables
(cognition, L2A)?
2. Can part of the variance in learner variables be explained by independent proximal
factors?
In line with the present argumentation, it was expected that main effects of distal SES and L2
program on cognitive and linguistic skills will be found, which will be mediated by some
aspects of parent-child interaction and teachers’ L2 input quality, respectively, while other
aspects predict both internal skill sets independent of SES and L2 program.
Kersten (2022) – The Proximity of Stimulation Hypothesis
11
While evidence for meso-level SES and type of L2 program as predictors of cognitive-linguistic
skills is strong in reported studies (Hackman & Farah, 2009; Kersten et al., 2010; Kishiyama et
al., 2009; Lawson et al., 2016 for SES; Sheridan et al., 2012; Trebits et al., in press for school
program; Wesche, 2002), it is less clear which exact types of micro-level parental and teacher
behavior exert the strongest influence on learner skills. As results in these strands of research
are much more heterogeneous, no hypotheses were formed concerning the effect of
particular micro-level variables.
Method
Participants
Cross-sectional data stem from 93 primary school L2 learners of English (43 female, mean age
9;6, age range 7;6-11;6) who took part in different types of primary foreign language programs
with L2 English in Germany. Learners from three classes (n = 31) in three different schools took
part in conventional foreign language programs [FLP] with typically two hours of L2 English
per week. Students from six classes (n = 62) took part in two partial immersion schools in
which all subjects except for German language arts are taught through L2 English. L2 teachers
in these programs are either native speakers of English or trained German foreign-language
teachers with a very high command of English, while not all FLP teachers were trained as
teachers of English.
Measures and procedures
A questionnaire on home language background, parents’ education and occupation, and
parent-child interaction (Kersten & Ponto, 2016)
5
was administered to the parents through
the teachers. Schools were interviewed concerning the intensity and duration of their L2
programs.
Distal external variables (meso-level)
L2 program was operationalized by the total number of lessons (45 minutes) in the L2
administered to the learners during their time in the primary school program. These values
were calculated based on the number of lessons per week instructed in the L2 for the total of
weeks at school per year for each grade level the students spent in the program (number of
weekly lessons instructed in the L2*number of weeks at school per year*grade level). Total
number of lessons in the FLP schools ranged from 80-160 lessons, and for immersion classes
from 1,296-4,212 lessons in the L2. To account for the fact that data were not normally
distributed due to the two different school programs involved, Bollen-Stine bootstrapping was
used in the structural equation modeling (see below).
Socioeconomic status (SES) was indicated using HISEI (Highest International Socio-Economic
Index of Occupational Status, 2008 version, Ganzeboom, 2010), included in the parental
background questionnaire (Kersten & Ponto, 2016).
Kersten (2022) – The Proximity of Stimulation Hypothesis
12
Proximal external variables (micro-level)
Parent-child interactions (PCI): Items operationalizing PCI stem from the parental
questionnaire. The intention of the questionnaire was to gain differentiated information on
various types of parental behavior in early and later childhood to measure their differential
effects on learner development. Items were adapted from Deutscher Bundesverband für
Logopädie (2020), Hertel et al. (2009), Lange et al. (1998), Linberg et al. (2019), the NICHD
(2005, 2006), Safwat and Sheikhany (2014), Steinmetz and Hommers (2003) (for a detailed
analysis of SMILE project data, see Fleitling, 2021). Item-based correlational analyses between
types of parental behavior and learners’ working memory scores
6
rendered a short scale of
parental language and literacy support (Cronbach’s = .77). Remaining items of general
interactional behavior were subsumed under the scale of parent-child interaction (Cronbach’s
= .84).
Teachers’ L2 input quality
7
was derived from videography of best practice videos of each
teacher analyzed with the Teacher Input Observation Scheme (TIOS; Kersten et al., 2018a;
Kersten, in press). TIOS contains 41 items of L2 teaching techniques as part of four scales:
Cognitively stimulating L2 task characteristics, verbal L2 input, non-verbal input, and support
of L2 output.
8
A previous study (Kersten et al., 2019) found that interrater reliability of two
raters tested on the basis of 18 videos of L2 teaching was high (Krippendorff‘s = .88* based
on 687 cases, 1,374 decisions, item-based IRR Pearson’s r = .69**-1.00**, p < .05), as was
internal consistency (Cronbach’s = .91 for 38 items).
An item-based correlational analysis rendered a short-scale of thirteen L2 teaching techniques
measured with the TIOS that correlated significantly with both L2 lexical and grammatical
reception in the sample (input quality short scale: Cronbach’s = .90).
9
TIOS video analysis
was used for one best practice lesson in the L2 of each L2 teacher (n = 9), i.e., the video with
the highest score of each respective teacher. For one teacher, only one video was available,
which was rated with a high total TIOS score. For the other teachers, up to five lessons could
be videotaped and were selected according to a number of additional criteria to ensure
comparability. Teacher consent had been collected beforehand. Teachers were aware of the
overall project goal of the investigation of learner variables, but had not been introduced to
the specific goal of our classroom observation nor to the observation instrument. Learners
had already been familiarized with videography to reduce observer paradox. All sessions were
general science lessons which were comparable in topic and structure, i.e., they contained an
introductory phase, a main phase with a task or activity, and a final closing or consolidation
phase, and were comparable in the proportion of teacher input in that all lessons included
sections with observable teacher input and group work phases. For all videos, one teacher
camera and another camera facing the students during the session were used. All videos were
technically sound. All lessons were taped during the second half of a school year.
Internal variables (nano-level)
Learner data were elicited during 2016-2017 with individual and group tests at the end of each
school year. Group tests were carried out for L2 tests (BPVS 3 and ELIAS Grammar Test 2)
during 45-minute lessons. Individual cognitive tests (WISC-IV and BAKO) were taken in a quiet
room at school by two interviewers in the ambient language, German. One interviewer carried
Kersten (2022) – The Proximity of Stimulation Hypothesis
13
out the test while the other one recorded the results on a score sheet. Test instruments are
introduced below.
Working memory was assessed using the WISC-IV, with the subscales digit span (digits
forward, digits reversed) and letter-number sequencing (Petermann & Petermann, 2011).
Phonological awareness was assessed using BAKO 1-4 (Basiskompetenzen für Lese-
Rechtschreibleistungen [Basic Competencies 3 for Reading and Writing Skills], Stock et al.,
2003), with the five subscales segmentation of pseudo-words, vowel replacement, phoneme
exchange, vowel length, and reversing words.
L2 attainment: The British Picture Vocabulary Scale (BPVS) 3 (Dunn et al., 2009) was used for
receptive L2 lexicon, and the ELIAS Grammar Test 2 (Kersten et al., 2012, Kersten et al., under
review.) for receptive L2 grammar.
Learners’ degree of multilingualism was operationalized as the sum of proficiency of all home
languages as reported by parents in the parental questionnaire. Parents were asked to name
all home languages of their child, the age of first contact with this language and with the
majority language German if it was not listed among their home languages, and to describe
their child’s language competences in the majority language and up to three other home
languages on a seven-point Likert scale ranging from “only a few words” to “native-like.”
For statistical calculations, structural equation modeling (SEM) using SPSS (IBM SPSS Statistics
26) with the add-on AMOS was used. As not all variables were normally distributed, Bollen-
Stine boostrapping was carried out. To that end, missing data were imputed (single imputation
using the expectation maximization algorithm) as AMOS needs a complete data set for this
procedure.
Results and discussion
Figure 4.4 display the results of the SEM which tests whether hypothesized effects of the two
distal container variables (SES, L2 program) on internal variables (working memory,
phonological awareness, receptive lexical and grammatical L2 skills, degree of multilingualism)
were mediated by proximal factors (parent-child interaction, teacher input).
Structural equation modeling shows the mutual interdependency of numerous variables and
is thus suitable to test the conceptual predictions based on hierarchical ordering of the data
in line with a Proximity of Stimulation focus (Figure 4.3). The model had a good model fit
(Bollen-Stine bootstrap p-value > .05, CFI > .95, RMSEA < .07; Hu & Bentler, 1999), and yielded
a number of other interesting relations between variables. The number of participants is small
for SEM analysis, however.
Kersten (2022) – The Proximity of Stimulation Hypothesis
14
Figure 4.4. Structural equation model of distal meso-level variables, proximal micro-level variables, and internal
nano-level variables (N = 93)
Note. *p < .05; **p < .01; ***p < .001; +p < .10; standardized β regression weights (black single-headed regression
arrows showing directional effects) and covariances (grey double-headed correlational arrows)
Testing Proximity of Stimulation predictions: Mediation effects and additional effects of
proximal variables
With regard to RQ1, Figure 4.4 shows three partial mediations:
(1) The effect of SES on working memory is partially mediated by proximal parental language
and literacy support at home. (The path of SES
parental language and literacy support is
marginally significant at p = .052.) The effect of SES on working memory capacities of the child
(which was also found on phonological awareness) corroborates what has been shown in
numerous studies before (Waters et al., 2021, for WM; McDowell et al., 2007; Zhang et al.,
2013, for PA, Bruhn et al., 2022, this volume). As argued above, however, SES as a higher order
construct cannot be assumed to exert a direct influence on the learner’s mind, and might
encompass a large number of other factors associated with high social class. This is why a large
effect of the distal variable is to be expected. Effects of proximal variables as immediate points
of contact would thus be expected to be much smaller but to contain explanatory value. It is
thus surprising that the effect of parental language/literacy support on WM is comparable to
that of SES on WM, corroborating the assumption that early linguistic interactions with a child
are strong predictors of their working memory development (Valcan et al., 2018). This finding
lends some support to the Proximity approach, but it will need to be confirmed in larger
longitudinal studies with more statistical power since the cross-sectional design does not
allow for a causal interpretation of the direction of the effect.
(2) The effect of the L2 program on grammar comprehension is partially mediated by proximal
teachers’ L2 input quality. The main effect of L2 program (on both language variables) supports
a well-known finding in SLA research (Sun et al., 2015). The mediation, however, shows that
in the more intensive programs, teachers use significantly more stimulating L2 input
techniques, and that this input quality is crucial for L2A. Part of this difference could be
Kersten (2022) – The Proximity of Stimulation Hypothesis
15
explained by the fact that the more intensive L2 programs are content-based immersion
programs in which a meaningful use of language for purposes other than language learning
itself is inherent. Teaching of subject matter in bilingual programs requires a particularly
intensive use of L2 instructional techniques to render content comprehensible to the learners.
Bilingual teachers might therefore make stronger use of such techniques. There might also be
some self-selection of teachers with good L2 knowledge and high methodological competence
and motivation who choose to teach in bilingual programs (Wegner, 2021). However, as I
pointed out elsewhere (Kersten, 2020, in press), bilingual teaching is not based on “special”
teaching approaches. The TIOS was developed based on theories of foreign language teaching
in Instructed Second Language Acquisition. The difference in the use of teaching techniques
can thus not necessarily be deduced from programs’ foreign language teaching approaches.
Even though we showed that trained L2 teachers in conventional programs are as capable as
immersion teachers to score highly on TIOS scales (Kersten, 2019, 2020), the influence of the
program type is apparent. These are crucial questions for teaching practice in conventional
and bilingual programs which need to be investigated further.
This finding confirms, at least in part, that the learning context in which the L2 is encountered,
i.e., the actual proximal stimulation of the learners through activities, input and interactions,
impacts their level of L2 competence. This is in line with models on the effect of input
processing such as Gass et al.’s (2020) Model of Second Language Acquisition, Leow’s (2015)
Model of the L2 Learning Process in ISLA, Kormos’ (2011) Model of Bilingual Speech Production,
Truscott and Sharwood Smith’s (2019) Modular Cognition Framework, and Li and Jeong’s
(2020) Social Brain of Language Learning (Kersten, in press). These results also support the
wide range of studies on the effectiveness of immersion teaching (Wesche, 2002; Kersten &
Rohde, 2015).
(3) The effect of L2 program on phonological awareness is partially mediated by L2 input. This
last mediation has a very strong main effect of program
PA, and of program
L2 input
quality, and a negative regression weight for L2 input quality
PA. A possible explanation
might be that PA is highly susceptible to the amount and frequency of L2 encounters, as
operationalized in the program variable (number of accumulated lessons of L2 contact). In
other words, learners in the bilingual programs who experience the language with a much
higher intensity significantly outperform learners in low-intensity conventional programs in
PA. The additional negative impact of input quality might indicate the phenomenon that
teachers tend to use more scaffolding techniques for learners with lower linguistic capacities
than for more advanced learners – in fact, they reduce scaffolding as a function of learners’
comprehension (“fading,” van de Pol et al., 2010).
This points to a chicken-and-egg relationship in cross-sectional data sets in which the causal
direction of relationships cannot be established. It is, in other words, possible that the data
indicate that within both programs, teachers use fewer L2 input techniques for more
advanced learners with inherently higher PA, and not that (as the data seem to suggest) L2
input quality has a negative effect on learners’ PA (compare van de Pol et al., 2010). A
longitudinal design is needed to shed more light on this question, which we are currently
working on.
Kersten (2022) – The Proximity of Stimulation Hypothesis
16
In addition, parental language/literacy support also predicts L2 lexical skills, but not L2
grammatical reception. While L2 lexicon is, beyond that, only dependent on the school
program, i.e., the intensity in which the L2 is provided, L2 grammar reception seems to be
much more susceptible to input quality and the learner’s degree of multilingualism, which
supports research showing that previous experiences with multiple languages can facilitate
metalinguistic awareness such as pattern recognition in an additional language (Adesope et
al., 2010, Maluch et al., 2015). Grammar reception, which also correlates with PA (see below)
thus seems to be more susceptible to instructional quality and different cognitive abilities than
the L2 lexicon, which is more strongly dependent on the intensity of the input, a finding which
supports the results of Kersten et al. (2018b) in bilingual preschools.
These findings are in line with the Proximity of Stimulation Hypothesis which posits that effects
of higher-order (distal) container variables in a data set are mediated by the proximal variables
they comprise, which affect a learner directly, and that additional variance might be predicted
by proximal variables which are independent of the distal container.
Pertaining to RQ2, however, not all immediate stimuli – here, not all types of parental or
teacher behavior – were either dependent on SES / school program, or, in turn, predictive for
cognitive and linguistic skills. No additional PCI variable was found to predict cognitive skills
independently of SES, but not all PCI variables were predicted by SES, as was found for general
PCI and the single-item variable of discussing problems. This shows that parental behavior has
to be viewed and assessed in a much more fine-grained way than only using SES as a proxy.
Even though SES (measured as education and occupation) might often strongly predict PCI,
this relationship is not obligatory. Behavior underlies numerous factors and intentional
changes. Some parents will interact with and cognitively stimulate their children even though
they do not have a high educational degree or a high-paying job. In order to detect and capture
these effects, it is necessary to look at proximal variables of direct stimulation in much more
detail. If only higher-order variables are included, these actual stimulating effects become
blurred and are reduced to background noise. As specific parent-child interactions seem to be
carried out by parents no matter which social class they belong to, this calls for a differentiated
view on parental behavior and a much more fine-grained analysis of types of parental
behavior, which may translate into item-based rather than cluster-based statistical analyses.
It also stresses that including a container variable such as SES as a proxy in empirical studies
falls short of identifying which actual types of parental stimulation might explain child
development.
Moreover, teachers’ patience to wait for student answers also predicted L2 grammar and
phonological awareness independently of L2 program, which supports the hypothesis that
stimulating proximal (and thus potentially causal) factors are not necessarily included in distal
containers, and should therefore be measured separately.
Additional correlational effects
Finally, a number of correlations between the included variables were found within the
different levels. As is to be expected, variables which belong to the same larger construct
correlate with each other. This holds for the three PCI variables, the two teacher input
Kersten (2022) – The Proximity of Stimulation Hypothesis
17
variables, the two cognitive skills (working memory and phonological awareness), and the two
linguistic variables (L2 grammar and lexical reception).
SES and school program also show a strong positive correlation. This hints at a social selection
effect, which has often been reported in special programs such as immersion. One of the two
immersion schools is a private school which provided certain funding for socially
disadvantaged families but is still prone to attract a majority of more affluent families.
Selection effects are also known in public schools where motivated parents compete for a
place in the bilingual strand for their children. Controlling for this correlation means, however,
that the remaining effects are independent of this selection bias.
In addition, phonological awareness and L2 grammar reception were correlated significantly
and positively, in line with Engel de Abreu and Gathercole (2012). Correlations in cross-
sectional data do not allow the interpretation of causal direction. Individual cognitive abilities
are considered an important factor in the L2 acquisition process (Dörnyei, 2005; Dörnyei &
Ryan, 2015), and thus the relationship between cognitive and L2 skills corresponds to various
studies that find predictor effects in both directions (Kersten, 2020). Research into effects of
bilingual language competence on cognitive skills, on the other hand, has been focused of the
so-called bilingual advantage debate (Festman et al., 2022, this volume), which claims that
high L2 skills and frequent L2 usage, as present in the immersion programs, can lead to an
increase in cognitive skills. Effects of bilingual competences on metalinguistic awareness, of
which phonological awareness is a part, were shown by Bialystok and Barac (2012), Laurent
and Martinot (2010), and Trebits et al. (in press). Effects on working memory were found by
Bialystok et al. (2009), Trebits et al. (in press), as well as several meta-analyses (Adesope et
al., 2010; Grundy & Timmer, 2017), while Lehtonen et al. (2018) do not find working memory
effects in their meta-analytical review. Trebits et al. (in press) also found in a longitudinal study
of 39 fourth graders in conventional and immersion schools that PA is more malleable than
WM, and that effects of degree of bilingualism on PA are found earlier and are stronger. While
advantages of PA were present in third grade and increased in fourth grade, WM scores were
the same in third grade, but at the end of fourth grade, immersion students had developed a
significant advantage. The findings in the current data set thus lend some support to the
bilingual advantage hypothesis with regard to L2 grammar skills and PA and, albeit just as a
tendency, for L2 lexical skills and working memory. This also calls for a detailed view on
differential relationships between specific linguistic and cognitive abilities. However, as
pointed out before, whether degree of bilingualism represents a stronger predictor on
cognitive abilities, or the other way round, cannot be determined with this type of cross-
sectional data. For that purpose, longitudinal data and a different research design (e.g., cross-
lagged panel or growth models) are needed.
This paper mainly focusses on exemplifying the theoretical assumptions of the Proximity of
Stimulation Hypothesis, using selected constructs at the distal, proximal, and internal levels.
To sum up, several mediating relationships identified in the data lend support to the
assumption that the proximal level of direct stimuli provides a more differentiated view on
immediate effects on learners’ cognitive-linguistic skills than variables at the distal level in
which proximal variables are conceptually contained. As argued above, they can provide much
more detailed information about what exactly stimulates and hence shapes the learners’
internal system.
Kersten (2022) – The Proximity of Stimulation Hypothesis
18
Limitations
There are obviously numerous limitations to the study. With respect to the data set, a larger
number of participants is needed to achieve higher statistical power and more reliable results.
As it is, a full set of variables was not available for all participants in the project due to illness
and missing parental responses. This is a general risk of data sets with such dense descriptions
of participants.
Moreover, data were collected cross-sectionally, which means that the direction of the
observed effects cannot be determined. Longitudinal or experimental data are preferable to
investigate causal effects. The project at large includes longitudinal data for a subset of the
participants; however, analysis of these data is still in progress.
The current sample is not homogenous concerning age and school program, which increases
the risk that several factors might be confounded. Only the immersion program included
participants in second grade. In conventional FLT programs in Germany, English is introduced
in grade 3 so no grade 2 data were available. On the other hand, this has an intended positive
effect, in that L2 competence in immersion programs is considerably higher than in
conventional programs so that L2 attainment in lower immersion grades can be assumed to
be more comparable to older learners in conventional programs. Including second grade data
adds to the heterogeneity and variance in the data. The inclusion of age as a covariate was
tested in the SEM but did not contribute to the model.
More reliable scores for teachers’ instructional quality as measured with the TIOS could be
achieved using several classroom videos with comparable lesson structures per teacher. This
was, however, not possible to elicit due to the permissions of the schools and teachers.
The parental questionnaire only elicits parents’ self-reported behavior as an indicator of
parent-child interactions, which might be biased due to social desirability. Actual observations
of parental behavior would be preferable but were out of the scope of this project.
Conclusion
Human cognitive and linguistic development are intrinsically intertwined and take place within
a nested structure of conceptual levels. Proximal levels contain variables which present
concrete stimuli to the learner, while variables at distal conceptual levels only exert indirect
effects and often represent container variables made up of numerous proximal ones. Research
on the combined and differential effects of these variables that takes these different levels
into account is still scarce. Major difficulties lie in the assumed nature of effects and statistical
representations (Kersten & Greve, 2022, this volume). This paper took a Proximity of
Stimulation focus which holds that, since distal variables, logically, exert only indirect effects
mediated by proximal factors, effects are best explained using factors that have an immediate
effect on the learner. Such stimuli induce and condition intake and processing, activate
associative networks, and are thus crucial for storage of the information within the learner’s
internal system (Li & Jeong, 2020, Kersten, in press). This is in line with overarching models on
layers of contextual factors (Lerner, 2002; Douglas Fir Group, 2016).
Kersten (2022) – The Proximity of Stimulation Hypothesis
19
In this vein, this study investigated the impact of an exemplary selection of potentially
influencing variables on different levels on L2 acquisition and cognitive skills. To do so, it
attempted to unpack container variables and to disentangle conceptual and causal
hierarchies, accounting for mediating relationships between the different variables, and
analyzing their differential contributions. Variables included in the analysis were social status
and school programs with different amounts of L2 teaching (distal, meso-level), types of
parent-child interaction and types of teacher behavior/input (proximal, micro-level), and
working memory, phonological awareness, and L2 grammar and lexical reception (internal,
nano-level). Structural equation modeling was used to account for the hierarchical data
structure of 93 L2 learners of English in German conventional and bilingual primary schools.
Results showed that, as hypothesized within the Proximity of Stimulation focus, the effect of
both distal container variables (social status and school program) on the internal cognitive and
linguistic variables was partially mediated by proximal variables, i.e. parental language and
literacy support and teacher’s input quality, respectively. The L2 program also strongly
predicted L2 lexical comprehension and phonological awareness without a mediating effect
of the included proximal variables, pointing to the effectiveness of bilingual teaching programs
as compared to conventional school programs with regard to L2 acquisition and aspects of
cognition.
Additional independent effects which were not predicted by the higher-level distal variables
were found for parental language/literacy support on L2 lexical comprehension, for teachers’
allowing time and space with respect to L2 utterances (‘teachers’ patience’) for L2 grammar
comprehension and phonological awareness, and, internally, children’s degree of
multilingualism on their L2 grammar.
Moreover, correlations were found for SES with school program, pointing to selection effects
of more engaged or affluent parents in the bilingual programs. As expected, further
correlations were shown between variables belonging to the same constructs (PCI variables,
teacher behavior variables, cognitive skills, and L2 skills, respectively). Finally, phonological
awareness correlated with L2 grammar, and working memory showed a correlational
tendency with L2 lexicon, showing that cognitive and linguistic development are intertwined
and lending at least some support to a cognitive advantage hypothesis (Grundy & Timmer,
2017; Laurent & Martinot, 2010).
Lexical and grammar comprehension were thus predicted by different stimuli. Phonological
awareness was predicted much more strongly by numerous external variables in family and
school contexts and seems therefore much more malleable than working memory, which was
mainly related to social variables. These results also confirm the importance of a differential
view on factors which shape different linguistic and cognitive skills.
As the data presented in this paper suggest, to shed light on the interplay of various variables
relevant to SLA, it might be promising for future studies to elicit dense and detailed
information on types of distal and proximal factors for each participant. To best understand
the actual driving forces of SLA and its interconnection with cognitive development in
Kersten (2022) – The Proximity of Stimulation Hypothesis
20
individual learning trajectories, it seems necessary to disentangle conceptual levels of
variables included in the respective analysis, and to pay specific attention to proximal stimuli
for learning processes. It will also be necessary to further develop suitable instruments to elicit
this type of data. As we experienced in this project, eliciting a multitude of variables has
practical consequences both in terms of capacities of the research team and of individuals and
institutions involved in the study. Desired background information has to be balanced with
length of questionnaires, for instance, so as not to reduce response rates. In addition,
disadvantages of self-reports have to be carefully weighed up against the workload necessary
for more objective observations of behavior within classrooms and families. The Teacher Input
Observation Scheme (Kersten et al., 2018a) used in this study for L2 input quality
operationalized observation of proximal instructional variables, while the parental
background questionnaire (Kersten & Ponto, 2016) relied on parents’ self-reports. This type
of elicitation should be refined in future studies. To corroborate both the theoretical views as
well as preliminary findings of the study presented above, longitudinal data are necessary to
determine the causal direction of a relationship between variables. Further methodologies
and more refined statistical analyses need to be explored to account for the complex systemic
nature of (language) learning, including longitudinal growth models. In terms of practical
relevance, findings regarding effective types of behavior in families and instructional contexts
may serve as a basis for recommendations for educational institutions that work closely
together with parents, and thus increase the chance to foster the linguistic and cognitive
development of young learners within families (Niklas et al., 2018) and school systems.
10
References
Adesope, O. O., Lavin, T., Thompson, T., & Ungerleider, C. (2010). A systematic review and meta‐analysis of the
cognitive correlates of bilingualism. Review of Educational Research, 80(2), 207–245.
http://dx.doi.org/10.3102/0034654310368803
Atkinson, D. (2019). Second language acquisition beyond borders? The Douglas Fir Group searches for
transdisciplinary identity. The Modern Language Journal, 103, 113–121.
https://doi.org/10.1111/modl.12530
Bialystok, E., & Barac, R. (2012). Emerging bilingualism: Dissociating advantages for metalinguistic awareness
and executive control. Cognition, 122, 67–73. https://doi.org/10.1016/j.cognition.2011.08.003
Bialystok, E., Craik, F. I. M., Green, D. W., & Gollan, T. H. (2009). Bilingual minds. Psychological Science, 10(3),
89–129. https://doi.org/10.1177/1529100610387084
Bruhn, A.-C., Miller, L., Ponto, K., Kersten, K., & Mähler, C. (2022, this volume). Can type of schooling
compensate for low SES? Investigating effects of instruction and SES on cognitive skills. In K. Kersten, &
A. Winsler (Eds.), Understanding variability in second language acquisition, bilingualism, and cognition –
A multi-layered perspective. Routledge.
Dewaele, J. (2022, this volume). The interplay between learner-internal variables and levels of anxiety and
enjoyment among Spanish EFL learners. In K. Kersten, & A. Winsler (Eds.), Understanding variability in
second language acquisition, bilingualism, and cognition – A multi-layered perspective. Routledge.
Deutscher Bundesverband für Logopädie e.V. (2020). Wie unterstützen Eltern den Spracherwerb ihrer
Kinder? Tipps, die Eltern dabei helfen können, die Sprachentwicklung ihres Kindes zu unterstützen.
Retrieved 15/04/2021 from: https://www.dbl-ev.de/logopaedie/foerderung-der-
sprachentwicklung/sprachfoerderung-beim-spracherwerb/wie-unterstuetzen-eltern-den-spracherwerb-
ihrer-kinder.html/
Dörnyei, Z. (2005). The psychology of the language learner: Individual differences in second sanguage
acquisition. Erlbaum. http://dx.doi.org/10.4324/9781410613349
Kersten (2022) – The Proximity of Stimulation Hypothesis
21
Dörnyei, Z., & Ryan, S. (2015). The psychology of the language learner revisited. Routledge.
https://doi.org/10.4324/9781315779553
Douglas Fir Group (2016). A transdisciplinary framework for SLA in a multilingual world. The Modern Language
Journal, 100(S1), 19–47. https://doi.org/10.1111/modl.12301
Duff, P. A., & Byrnes, H. (2019). SLA across disciplinary borders: Introduction to the special issue. The Modern
Language Journal, 103, 3–5.
Dunn, L. M., Dunn, D. M., Styles, B., & Sewell, J. (2009). The British Picture Vocabulary Scale (3rd ed.). GL
Assessment.
Ellis, P. D. (2010). The essential guide to effect sizes. Statistical power, meta-analysis, and the interpretation of
research results. Cambridge University Press. https://doi.org/10.1017/CBO9780511761676
Engel de Abreu, P.M., and Gathercole, S.E. (2012). Executive and phonological processes in second language
acquisition. Journal of Educational Psychology, 104(4), 974-986. https://doi.org/10.1037/a0028390
Festman, J., Czapka, S., & Winsler, A. (2022, this volume). How many moderators does is take till we know...
that too many bilingual advantage effects have died? In K. Kersten, & A. Winsler (Eds.), Understanding
variability in second language acquisition, bilingualism, and cognition – A multi-layered perspective.
Routledge.
Fleitling, V. (2021). Indicators and consequences of parent-child interaction. Unpublished Master’s Thesis,
Hildesheim University.
Ganzeboom, H. B. G. (2010). A new international socio-economic index (ISEI) of occupational status for the
inter-national standard classification of occupation 2008 (ISCO-08) constructed with data from the ISSP
2002-2007. Paper presented at the Annual Conference of International Social Survey Programme,
Lisbon.
Gass, S. M., Behney, J., & Plonsky, L. (2020). Second language acquisition: An introductory course (5th edition).
Routledge. https://doi.org/10.4324/9781315181752
Grundy, J. G., & Timmer, K. (2017). Bilingualism and working memory capacity: A comprehensive meta-analysis.
Second Language Research, 33(3), 325–340. https://doi.org/10.1177/0267658316678286
Hackman, D. A., & Farah, M. J. (2009). Socioeconomic status and the developing brain. Trends in Cognitive
Sciences, 13(2), 65–73. https://doi.org/10.3389/fnins.2014.00276
Han, Z. (2016). A “reimagined SLA” or an expanded SLA? A rejoinder to The Douglas Fir Group (2016). The
Modern Language Journal, 100(4), 736–740. https://doi.org/10.1111/modl.12378
Han, Z. (2019). Researching CDST: Promises and pitfalls. In Z. Han (Ed.), Profiling learner language as a dynamic
system. Multilingual Matters. Retrieved from http://ebookcentral.proquest.com; created from
teacherscollege-ebooks on 2019-10-10 11:30:17. https://doi.org/10.21832/9781788922807-009
Hertel, S., Hochweber, J., Mildner, D., Steinert, B., & Jude, N. (Eds.). (2009). PISA 2009 Skalenhandbuch.
Waxmann.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.
https://doi.org/10.1080/10705519909540118
Kersten, K. (2019). Einflussfaktoren im bilingualen Fremdsprachenerwerb. In A. Rohde, & A. K. Steinlen (Eds.),
Sprachenvielfalt als Ressource begreifen: Mehrsprachigkeit in bilingualen Kindertagesstätten und
Schulen (Band II), (pp. 35–70). dohrmann.
Kersten, K. (2020). Der Zusammenhang von kognitiven Fähigkeiten und Zweisprachigkeit im frühen L2-Erwerb:
Individuelle und externe Faktoren [The relationship between cognitive abilities and bilingualism in early
L2 acquisition: Individual and external factors]. In H. Böttger, J. Festman, & T. Müller (Eds.), Language
education & acquisition research: Blick auf den frühen Spracherwerb [Focusing early language
acquisition] (pp. 81–118). Klinkhardt.
Kersten (2022) – The Proximity of Stimulation Hypothesis
22
Kersten, K. (in press). L2 input and characteristics of instructional techniques in early foreign language
classrooms – Underlying theory and pedagogical practice. European Journal of Applied Linguistics and
TEFL.
Kersten, K., Rohde, A., Schelletter, C., & Steinlen, A. K. (Eds.). (2010). Bilingual preschools. Vol. I: Learning and
development. WVT.
Kersten, K., Piske, T., Rohde, A., Steinlen, A. K., Weitz, M., Couve de Murville, S., & Kurth, S. (2012). ELIAS
Grammar Test II. Studies on Multilingualism in Language Education, 1, Hildesheim University.
http://dx.doi.org/10.13140/RG.2.2.35027.32806/1
Kersten, K., & Greve, W. (2022, this volume). Investigating influencing factors of cognitive-linguistic
development in SLA: Challenges for empirical research. In K. Kersten, & A. Winsler (Eds.), Interplay of
Variables in Cognitive-Linguistic Development – A Multi-Layered Perspective. Routledge.
Kersten, K., & Ponto, K. (2016). Fragebogen zum familiären und sprachlichen Hintergrund [Parental
questionnaire on socioeconomic and linguistic background of language learners]. Studies on
Multilingualism in Language Education, 3, Hildesheim University.
http://dx.doi.org/10.13140/RG.2.2.20734.61760
Kersten, K., Schelletter, C., Bruhn, A.-C., & Ponto, K. (under review). Quality of L2 input and cognitive skills
predict L2 grammar comprehension in instructed SLA independently. Languages, 6.
Kersten, K., & Rohde, A. (2015). Immersion teaching in English with young learners. In J. Bland (Eds.), Teaching
English language to young learners: Critical issues in language teaching with 3-12 year olds (pp. 71–89).
Bloomsbury.
Kersten, K., Bruhn, A.-C., Ponto, K., Böhnke, J., & Greve, W. (2018a). Teacher Input Observation Scheme (TIOS).
Studies on Multilingualism in Language Education, 4, Hildesheim University.
https://doi.org/10.13140/RG.2.2.22467.35368
Kersten, K., Steinlen, A.-K., & Schüle, C. (2018b). The effect of L2 input and cognitively stimulating tasks on
second language acquisition. Paper presented at EUROSLA 28, University of Münster.
Kersten, K., Bruhn, A.-C., Koch, M., & Schriek, J. (2019). The effect of L2 input and cognitively stimulating tasks
on second language acquisition. Paper presented at EUROSLA 29, Lund University.
Kishiyama, M. M., Boyce, W. T., Jimenez, A. M., Perry, L. M., & Knight, R. T. (2009). Socioeconomic disparities
affect prefrontal function in children. Journal of Cognitive Neuroscience, 21(6), 1106–1115.
https://doi.org/10.1162/jocn.2009.21101
Kormos, J. (2011). Speech production and the Cognition Hypothesis. In P. Robinson (Ed.), Second language task
complexity (pp. 39–60). Benjamins. https://doi.org/10.1075/tblt.2.06ch2
Lange, A., Blonk, R., & Wiers, R. W. (1998). The parent-child interaction questionnaire, PACHIQ. Clinical
Psychology & Psychotherapy: An International Journal of Theory and Practice, 5(3), 187–198.
Larson-Hall, J. (2012). How to run statistical analyses. In A. Mackey, & S. G. Gass (Eds.), Research methods in
second language acquisition (pp. 245–274). Wiley-Blackwell. https://doi.org/10.1002/9781444347340
Laurent, A., & Martinot, C. (2010). Bilingualism and phonological awareness: The case of bilingual (French–
Occitan) children. Reading and Writing, 23(3), 435–452. https://doi.org/10.1007/S11145-009-9209-3
Lawson, G. M., Hook, Cayce J., Hackman, D. A., & Farah, M. J. (2016). Socioeconomic status and the
development of executive functions: Behavioral and neuroscience approaches. In J. A. Griffin, P.
McCardle, & L. S. Freund (Eds.), Executive function in preschool-age children: Integrating measurement,
neurodevelopment, and translational research. American Psychological Association.
http://dx.doi.org/10.1037/14797-012
Leow, R. P. (2015). Explicit learning in the L2 classroom. Routledge. https://doi.org/10.4324/9781315887074
Lehtonen, M., Soveri, A., Laine, A., Järvenpää, J., de Bruin, A., & Antfolk, J. (2018). Is bilingualism associated
with enhanced executive functioning in adults? A meta-analytic review. Psychological Bulletin, 144(4),
394–425. https://doi.org/10.1037/bul0000142
Kersten (2022) – The Proximity of Stimulation Hypothesis
23
Lenzing, A., Pienemann, M., Nicholas, H. (2022, this volume). Lost in Translation? On some key features of
dynamical systems theorizing invoked in SLA research. In K. Kersten, & A. Winsler (Eds.), Understanding
variability in second language acquisition, bilingualism, and cognition – A multi-layered perspective.
Routledge.
Lerner, R. M. (2002). Concepts and theories of human development. Erlbaum.
Li, P., & Jeong, H. (2020). The social brain of language: Grounding second language learning in social
interaction. npj Science of Learning, 5(8). https://doi.org/10.1038/s41539-020-0068-7
Linberg, A., Mann, D., Attig, M., Vogel, F., Weinert, S., & Roßbach, H.-G. (2019). Quality of parent-child-
interactions. Assessment of interactions with the macro-analytic rating system of parent-child-
interactions in the National Educational Panel Study at the child’s age of 7, 17, and 26 months (NEPS
Survey Paper No. 51). Leibniz-Institut für Bildungsverläufe, Nationales Bildungspanel.
Maluch, J. T., Kempert, S., Neumann, M., & Stanat, P. (2015). The effect of speaking a minority language at
home on foreign language learning. Learning and Instruction, 36, 76–85.
McDowell, K. D., Lonigan, C. J., & Goldstein, H. (2007). Relations among socioeconomic status, age, and
predictors of phonological awareness. Journal of Speech, Language, and Hearing Research, 50(4), 1079–
1092. https://doi.org/10.1044/1092-4388(2007/075)
Mitchell, R., Myles, F., & Marsden, E. (2019). Second language learning theories. Routledge.
https://doi.org/10.4324/9781315617046
National Institute of Child Health and Human Development (2005). Child Care and Child Development – Results
from the NICHD Study of Early Child Care and Youth Development. The Guilford Press.
National Institute of Child Health and Human Development (2006). The NICHD Study of Early Child Care and
Youth Development (SECCYD): Findings for Children up to Age 4 ½ Years. U.S. Government Printing
Office.
Niklas, F., Cohrssen, C., & Tayler, C. (2018). Making a difference to children’s reasoning skills before school-
entry: The contribution of the home learning environment. Contemporary Educational Psychology, 54,
79–88. https://doi.org/10.1016/j.cedpsych.2018.06.001
Noordhof, P. (2010). Emergent causation and property causation. In C. Macdonald, C., & G. Macdonald (Eds.),
Emergence in mind (pp. 69-99). Oxford University Press.
https://doi.org/10.1093/acprof:oso/9780199583621.003.0006
Petermann, F., & Petermann, U. (Eds.) (2011). Wechsler Intelligence Scale for Children (WISC-IV). Pearson
Assessment. https://doi.org/10.1007/978-0-387-79061-9_3066
Safwat, R. F., & Sheikhany, A. R. (2014). Effect of parent interaction on language development in children. The
Egyptian Journal of Otolaryngology, 30(3), 255.
Sharwood Smith, M. (2017). Introducing language and cognition: A map of the mind. Cambridge University
Press. https://doi.org/10.1017/9781316591505
Sharwood Smith, M., & Truscott, J. (2014). The multilingual mind: A modular processing perspective. Cambridge
University Press. https://doi.org/10.1017/CBO9781139644044
Sheridan, M. A., Sarsour, K., Jutte, D., D’Esposito, M., & Boyce, W. T. (2012). The impact of social disparity on
prefrontal function in childhood. PloS One, 7(4), e35744. https://doi.org/10.1371/journal.pone.0035744
Steinmetz, M., & Hommers, W. (2003). Das “Parent-Child Relationship Inventory” als deutschsprachiges Eltern-
Diagnostikum [The “Parent-Child Relationship Inventory” as German parent diagnostic]. Diagnostica
49(3), 120–128.
Stock, C., Marx, P., & Schneider, W. (2003). BAKO 1-4. Basiskompetenzen für Lese-Rechtschreibleistungen. Ein
Test zur Erfassung der phonologischen Bewusstheit vom ersten bis vierten Grundschuljahr [Basic
competencies for reading and spelling performance. A test to assess phonological awareness from the
first to the fourth year of primary school]. Beltz Test GmbH.
Kersten (2022) – The Proximity of Stimulation Hypothesis
24
Sun, H., Steinkrauss, R., Tendeiro, J., & de Bot, K. (2015). Individual differences in very young children's English
acquisition in China: Internal and external factors. Bilingualism: Langauge and Cognition, 19(3), 550–
566. https://doi.org/10.1017/S1366728915000243
Trebits, A., Koch, M., Ponto, K., Bruhn, A.-C., Adler, M., Kersten, K. (in press). Cognitive gains and
socioeconomic status in early second language acquisition in immersion and EFL learning settings.
International Journal of Bilingual Education and Bilingualism.
https://doi.org/10.1080/13670050.2021.1943307
Truscott, J., & Smith, M. S. (2019). The internal context of bilingual processing. John Benjamins.
https://doi.org/10.1075/bpa.8
Valcan, D. S., Davis, H., & Pino-Pasternak, D. (2018). Parental behaviours prediciting childhood executive
functions: A meta-analysis. Educational Psychology Review, 30, 607–649.
https://doi.org/10.1007/s10648-017-9411-9
van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher‐student Interaction: A decade of
research. Educational Psychology Review, 22(3), 271–296. https://doi.org/10.1007/S10648-010-9127-6
van Geert, P. (1991). A dynamic systems model of cognitive and language growth. Psychological Review, 98(1),
3–53. https://doi.org/10.1037/0033-295X.98.1.3
Waters, N. E., Ahmed, S. F., Tang, S., Morrison, F., & Davis-Kean, P. (2021). Pathways from socioeconomic
status to early academic achievement: The role of specific executive functions. Early Childhood Research
Quarterly, 54, 321–331. https://doi.org/10.1016/j.ecresq.2020.09.008
Wegner, A. (2021). Die Implementation von CLIL (Content and Language Integrated Learning) in der
Grundschule: Eine methodenintegrierende Untersuchung zu Einflussfaktoren bei der Umsetzung
bilingualer Unterrichtskonzepte [The Implementation of CLIL (Content and Language Integrated
Learning) in Elementary Schools: A Method-Integrating Study of Factors Influencing the Implementation
of Bilingual Teaching Concepts]. Universitätsverlag Hildesheim.
Wesche, M. B. (2002). Early French immersion: How has the original Canadian model stood the test of time? In
P. Burmeister, T. Piske, & A. Rohde (Eds.), An integrated view of language development. Papers in honor
of Henning Wode (pp. 357–379). WVT.
Zhang, Y., Tardif, T., Shu, H., Li, H., Liu, H., McBride-Chang, C., Liang, W., & Zhang, Z. (2013). Phonological skills
and vocabulary knowledge mediate socioeconomic status effects in predicting reading outcomes for
Chinese children. Developmental Psychology, 49(4), 665–671. https://doi.org/10.1037/a0028612
Kersten (2022) – The Proximity of Stimulation Hypothesis
25
1
For recent updates of the framework, see https://www.cognitionframework.com/.
2
There are exceptions to this reasoning which affect the individual as a whole not via sensory organs, such as
gravity, magnetism, but those are beyond the scope of the current effort.
3
Unfortunately, almost all of these variables can, theoretically, be further deconstructed to ever lower levels –
as a system within a system.
4
It has to be pointed out that a causal relationship cannot be inferred even with a mediation analysis from
cross-sectional data. However, a presumed mediator relationship can be refuted. Longitudinal data and
experiments with randomized subjects allow for causal interpretations.
5
The questionnaire elicits data concerning individual factors such as age and gender of the child, as well as
linguistic background/s of the family. The full questionnaire is available at
https://www.researchgate.net/publication/348760402_Elternfragebogen_zum_familiaren_und_sprachlichen_
Hintergrund.
6
Items were included which correlated up to an alpha-level of .10 to avoid a type 2 error and to make sure that
all potential predictors were regarded (Ellis 2010, p. 82; Larson-Hall, 2012, p. 249, see below). The scale
comprised the following types of parental behavior (translated from German): Explaining events to the child
that are difficult to understand, listening to the child, playing with letters in early childhood, playing word
games in early childhood, repeating the young child’s utterances in simple, whole sentences, singing songs in
early childhood, talking about daily events, playing word or language games, and encouraging the child to write
postcards, letters, etc.
7
See Kersten (accepted) and Kersten et al. (subm.) for a definition and background of the construct of input
quality.
8
The instrument can be downloaded at
https://www.researchgate.net/publication/340096869_Teacher_Input_Observation_Scheme_TIOS_and_Manu
al.
9
The short scale on teachers’ input quality includes the following TIOS items: Cognitively Stimulating Tasks /
Activities: 1 (tasks/activities focus on meaningful content goals), 5 (tasks/activities are explicitly linked to their
specific learning goals / learning objectives), 7 (tasks/activities require active problem-solving by the learners), 8
(tasks/activities are based on the prior world knowledge of the learners (i.e., their everyday experiences)), 11
(tasks/activities provide opportunities for genuine output (language use) of the learners), 12 (tasks/activities
are based on authentic materials / realia / texts / auditory displays); Verbal Input: 14 (the teacher has a high
language proficiency in the L2), 15 (the teacher exclusively uses the L2 in class), 16 (the teacher provides a high
amount of L2 input (i.e., uses L2 a lot to accompany all actions)), 17 (the teacher uses varied L2 input), 25 (the
teacher uses comprehension checks); Non-Verbal Input: 26 (the teacher uses body language); Support of
Learners’ Output: 31 (the teacher asks questions which promote open answers).
10
Acknowledgments This paper was funded by a grant from the Ministry for Science and Culture of the State of
Lower Saxony (VW Vorab, grant number 12.3-76251-99-55/14). The study is part of the project Studies on
Multilingualism in Language Education (SMILE) on variables influencing SLA in primary school and conducted in
close cooperation with Werner Greve. The thoughts presented in this chapter are largely the result of numerous
discussions and would not have been possible without his inspiring input and advice. I am exceedingly grateful
to Katharina Ponto and Ann-Christin Bruhn for an extraordinary job of project coordination, Martin Koch for his
invaluable support with statistical analysis, and Laureen Gallwitz and numerous student research assistants for
their help with data collection. This study would not have been possible without them.
