The Cambridge Handbook of Language Learning

Abstract

Cambridge Core - Applied Linguistics - The Cambridge Handbook of Language Learning - edited by John W. Schwieter

Full text

15.1 Introduction
Working memory (WM) can be thought of as the limited memory capacity
that allows us to hold a very small amount of information (e.g., the sev-
eral digits of our telephone numbers) in our mind and to simultaneously
manipulate this information for completing some cognitive tasks in our
daily life (Baddeley, 1986; Cowan, 1988). Since the inception of the sem-
inal WM model by the British psychologists Baddeley and Hitch (1974), the
concept has received considerable enthusiasm from multiple subfields of
cognitive sciences, spanning from such disciplines as psychology, linguis-
tics, neuroscience, biology, and computer science, to anthropology and
philosophy (Miller, 2003; Carruthers, 2013; 2015). Concerted efforts are
being poured in continuously from diverse research camps, which sub-
sequently give rise to the propagation of a dozen theoretical models of
WM (Miyake & Shah, 1999), notwithstanding lingering debates and con-
troversies over the nature and structure of this key construct of human
cognition (e.g., Baddeley, 2012; 2017; Cowan, 2014; 2017).
Amid these waves of research endeavours, a noticeable movement has
recently gained increasing credence, particularly among researchers who
are more concerned with the applications of the WM concept in more
practical areas of human cognition. One emerging trend is the attempt to
integrate and incorporate multiple perspectives of WM models in cogni-
tive science towards a more reconciled and comprehensive understand-
ing of the nature, structure, and functions of WM, so as to interpret its
important implications for such specific areas as cognitive development
and academic learning (Dehn, 2008). For example, Fenesi et al. (2015) have
recently integrated three currently dominant models of WM in cognitive
psychology, namely, Baddeley’s (1986; 2003; 2012) multiple-component
Working Memory in L2
Learning and Processing
Zhisheng (Edward) Wen and Shaofeng Li
15
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

366 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
model, Cowan’s (1999; 2005) embedded-processes model, and Engle and
Kane’s (2004) executive control model, with a view to reconceptualizing
the WM construct and re-evaluating its implications for various academic
learning domains of language comprehension and production, mathe-
matics, and multimedia learning.
Following a similar approach, in this chapter we aim to develop an inte-
grative perspective on the relationship between WM and second language
acquisition (SLA). Within language, one key domain of human cognition,
WM plays a pivotal role in many essential aspects (Gathercole, 2007;
Baddeley, 2017). Previously, considerable research has demonstrated
that individual differences in WM fairly accurately predict the process
and products of a whole range of native language learning domains and
activities (Gathercole & Baddeley, 1993; Baddeley, 2003; Cowan, 2011)
as well as second language (L2) acquisition and processing (Wen, Mota,
& McNeil, 2013, 2015; Linck et al., 2014; Grundy & Timmer, 2017; Li, in
press). Remarkably, despite recognition of the close relationship between
WM and language, not to mention the growing body of empirical stud-
ies conducted (separately) in cognitive psychology, psycholinguistics, and
applied linguistics, the exact links between WM and language still remain
very much unspecified. This problem is particularly acute in the field of
SLA, in which WM is presumably playing a bigger if not equal role (Wen,
2012; 2016). Overall, research in both fields should benefit from a viable
framework for implementing WM in second language research.
As such, the present chapter serves to fill up this niche. Specifically, we
aim to summarize and further expand on insights from previous works
(e.g., Wen, 2015; 2016; Li, in press) to offer an integrative perspective on
WM and SLA. Towards this end, we begin with a theoretical overview of
the general characterizations of the WM construct regarding its nature
and structure. Following this theoretical overview, we move to further
synthesize results and findings of empirical studies that have investigated
distinctive effects of phonological WM and executive WM on specific L2
domains and skills. We conclude the chapter by highlighting the theo-
retical and methodological ramifications of this integrative account and
its possible implications for the interdisciplinary research agenda of WM
and SLA.
15.2 The Nature and Structure of WM: Three
Characterizations
As mentioned in the above section, scholars on the applied side of WM are
increasingly calling for a unified understanding of the construct so that
such an important concept reaches its full potential for implementation
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 367
in more practical areas of human cognition. Indeed, notwithstanding
lingering controversies and debates, three general characterizations can
be derived from multiple perspectives on WM regarding its nature and
structure (e.g., Miyake & Shah, 1999; Baddeley, 2012; Cowan, 2014; 2017).
These include, namely, conceiving WM as (a) a limited capacity of human
cognition, (b) a fractionated construct subsuming multiple components
and functions, and (c) an activated portion of long-term memory (LTM).
Each of these will be further elaborated below, as well as their possible
implications for language learning and processing in general.
15.2.1 The Limited Capacity of WM
The first and the most important design feature of WM that can be teased
out from multiple perspectives in cognitive sciences relates to its core
nature of being a limited capacity of human cognition (Carruthers, 2013;
2015). This signature feature of WM is generally reflected in the immedi-
ately accessible cognitive resources that are available for holding a certain
amount of information (e.g., modality-specific information of phonolog-
ical-, visual-, spatial-nature, etc.) during the process of executing cogni-
tive activities. More specifically, such a limited capacity translates into a
fundamental (memory) constraint impacting on a wide range of human
cognitive activities in two ostensible fashions.
First of all, WM constrains the limited units of information that can be
held temporarily for manipulation in our heads in the service of task exe-
cution (or to use the term by Cowan, 2005, to “focus attention” upon). This
limited mental capacity of WM is best captured by the most well-known
concept of the “magical number seven plus or minus two” units of informa-
tion that is often attributed to its precursor of short-term memory (Miller,
1956). More recently, though, Cowan (1988, 2001; 2005) contends by cit-
ing both anecdotal facts and empirical evidence that the capacity of WM
is unlikely to surpass “four plus or minus one” chunks of information. Be it
either way (seven or four), the holding power of information in WM is
severely restricted (well, within the range of four to seven!). Given this lim-
ited capacity and its consequential bottleneck effects that are permeating
so many facets of our daily life (e.g., doing arithmetic, language compre-
hension, decision-making and planning, etc.), the concept of “WM limits”
itself has been rendered as the “centered cognition” in the truest philosoph-
ical sense (Klingberg, 2008; Carruthers, 2015).
Another aspect of the limited capacity of WM lies in the transient
duration of the simultaneous storage and processing of information in
our head when carrying out a certain cognitive task. That is to say, the
information being held usually remains accessible to our (immediate)
consciousness only for just a few seconds or so (between five and twenty;
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

368 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
Waugh & Norman, 1965), and it will then gradually fade away unless
rehearsed timely and repeatedly. That being said, the direct cause of this
fleeting nature of WM is still a debatable issue among the different WM
theorists (Conway et al., 2007; cf. the “Now or Never” Bottleneck Effect
by Christiansen & Chater, 2016). For example, there is first of all the pro-
posal of the memory decay hypothesis held by such cognitive psychologists as
Barrouillet and Camos (2012; 2015) in their time-based resource-sharing
(TBRS) view of WM, which ascribes this to the fading of memory traces
for later recall. Then, in contrast, there is also the interference view held
by Oberauer and Lewandowsky (2013, 2014, 2016), who argue that such
loss of information mainly arises from the interference resulting from
distraction.
Putting aside these debates over the exact amount of its holding power
and the exact sources causing its limitation, most (if not all) WM models
and theorists are subscribing to this signature feature of WM (Klingberg,
2008; Carruthers, 2013; 2015). More relevantly for language, a significant
proportion of (if not all) aspects of language acquisition and processing
activities are subject to this WM constraint. For example, WM is postu-
lated to underlie not just phonological representations (e.g., Pierce et al.,
2017), but to also shape a wide range of grammatical phenomena such as
the typology of word order (Hawkins, 2004), the minimization of depend-
ent distance (Gibson, 1998; Futrell, Mahowald, & Gibson, 2015; Liu, Xu, &
Liang, 2017), the interpretation of pronouns, and the permissible cases of
contraction (O’Grady, 2017).
15.2.2 WM Components and Functions
The second important feature of WM that can be identified from nomo-
thetic theories in contemporary cognitive sciences relates to its overall
design and structure. For a long time, Baddeley’s (2003, 2012; Baddeley &
Hitch, 1974) structural view of WM has dominated applications of this con-
cept in both theoretical and practical areas of human cognition (D’Esposito
& Postle, 2015). In terms of WM structure then, in the classic standard
model, Baddeley (2012, 2015; 2017) and colleagues have postulated four
key components. These are: (a) the phonological short-term memory (or
phonological loop in the original tripartite model) which handles sound-
based linguistic materials; (b) the visuo-spatial sketchpad that deals with
visual and spatial information; (c) an episodic buffer that serves to inte-
grate episodic information from all modality sources and connects closely
with LTM; and finally (d) an executive component (executive WM) that is
equivalent to the “central executive” component postulated in the classic
model by Baddeley, which encapsulates the various executive operations
and functions responsible for supervising and coordinating of attention
allocated to the three modality-based buffer systems.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 369
Among these four putative components of WM, two of them have
been researched quite thoroughly in relation to language acquisition and
processing (Baddeley, 2003, 2015, 2017). On the one hand, considerable
research has been conducted by the British and/or European WM camps
(e.g., initiated by Baddeley and colleagues), which points to a close link
between phonological WM (usually measured by a simple and storage-
only version of memory span task, such as the digit span or the non-
word repetition span tasks) and significant aspects of language learning
domains and activities (e.g., Gathercole & Baddeley, 1993; Baddeley, 2003).
In particular, empirical evidence is converging on the instrumental role of
phonological WM in acquiring novel phonological forms that are part and
parcel of vocabulary acquisition and development (Baddeley, Gathercole,
& Papagno, 1998).
Notwithstanding, an alternative interpretation of WM effects on lan-
guage is emerging and gaining more credence, namely, the attention-
oriented and executive control views of WM, as advocated by most North
America-based cognitive psychologists such as Daneman and Carpenter,
King and Just, Cowan, Engle, and Kane, among many others. The change
in WM conception is accompanied by a shift in research paradigms on lan-
guage-focused inquiries. Instead of looking into the individual contribu-
tions of each component of WM, a growing body of studies are conducted
in North America that tap into the individual differences of WM and their
consequence for language processing activities such as reading compre-
hension and production.
Within these North America-based WM research camps, two research
paradigms are becoming quite well-established. On the one hand,
Daneman and Carpenter (1980) have devised the more complex version
of memory span task that taps into both storage and processing aspects
of WM, i.e., the reading span task (1980) that consists of reading aloud an
growing set of sentences and recall of the final words. Later, Waters and
Caplan (1996) further refined this version, modifying the scoring proce-
dures by taking into account the recall, the reaction time, and the accu-
racy. An alternative complex version of memory span task was designed
by Engle, Cantor, and Carullo (1992) which consists of operation and final
word recall (i.e., the operation span task). In terms of overall results, this
growing body of empirical studies has pointed to the significant role of the
executive aspect of WM (i.e., executive WM) in a whole range of secondary
and post-interpretive processes during language comprehension and pro-
duction (Daneman & Merikle, 1996; Waters & Caplan, 1996; Cowan, 2011).
Putting aside the differences in research focus, the two research tradi-
tions, with these WM camps on either side of the Atlantic (Andrade, 2001),
are producing empirical evidence converging on the two distinct compo-
nents of WM regarding their distinctive effects on language acquisition
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

370 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
and processing. For the European (particularly the British) camp, research
points to the close link between phonological WM and vocabulary learn-
ing and development, while for the North American camp, a strong link
is thus forged between executive and control aspects of WM in relation
to the processing activities of language comprehension and production.
Combined, these two research traditions are lending theoretical and
methodological support to the integrative perspective that we are propos-
ing in the next section.
15.2.3 The Interactions Between WM and LTM
The third and final characterization of the WM construct that can be
derived readily from nomothetic theories of WM concerns its relation-
ship with long-term memory (LTM). LTM distinguishes itself from WM in
that it is of potentially unlimited capacity and stores all the knowledge
permanently. In terms of contents in LTM, it can be postulated to con-
sist of two categories (following Ullman, 2001; 2016), which include (1)
a declarative memory component that subserves fact-based knowledge
(i.e., knowledge of “what”, such as lexical knowledge); and (2) a procedural
memory component, that subserves skill-based or rule-based knowledge
(i.e., knowledge of “how”, such as grammatical rules). Regarding the
interactions between WM and LTM then, WM, as conceptualized in
Baddeley’s (2012) standard model, serves as a gateway to the LTM knowl-
edge base. In other words, a bi-directional (i.e., two-way) flow of informa-
tion is envisaged to be taking place between multiple WM components
(which tends to be domain specific, especially those of a phonological
and visuo-spatial nature) and the LTM warehouse. In contrast, there is
also the alternative view adopted by most North America-based cog-
nitive psychologists who conceive WM as the activated portion of LTM
(e.g., in Cowan’s, 1999; 2005, embedded processes model). In this second
and now more dominant view, the boundary between WM and LTM is
becoming blurred (e.g., Hasson, Chen, & Honey, 2015; Jones & Macken,
2015). Increasingly, such a view is more compatible with evidence from
modern cognitive sciences, and particularly cognitive neuroscience
(Conway, Moore, & Kane, 2009; D’Esposito & Postle, 2015; Constantinidis
& Klingberg, 2017).
15.2.4 Summing-Up
To sum up this section, when the differences in epistemic stance and
research focus among the multiple theoretical perspectives in cognitive
psychology are sorted out and further reconciled, a more balanced con-
ception of WM can be derived. To recapitulate, WM is best characterized
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 371
as the primary memory (as opposed to its sibling of secondary memory, aka
LTM; James, 1890; Waugh & Norman, 1965) that consists of multiple com-
ponents and executive mechanisms or functions (accommodating both
the structural view of Baddeley and colleagues as well as the functional
view of North America-based researchers), with WM being an integral
partof LTM. As we shall argue later, by drawing on insights from these
multiple WM research camps (across the Atlantic), a more balanced
account of the WM construct should be in place that will in turn help
facilitate its implementation in more practical research, such as second
language learning and processing, a topic that shall figure in the next
section.
15.3 Empirical Studies of WM and SLA: A Synthesis of
Results and Findings
Whereas the above section deals with the theoretical aspects of WM,
this section synthesizes the findings of the research on the associations
between WM and the process and product aspects of L2 learning. We
start with an overview of the research by reporting the findings of two
recent meta-analyses and then review the findings on the associations
between WM and specific aspects of SLA. In this chapter, we mainly focus
our review on the related research on vocabulary, grammar, reading, and
speaking and leave out writing and listening due to the lack of research
on the role of WM in affecting the learning of these two language skills. As
discussed in the above section, given the theoretical distinction between
phonological WM and executive WM (EWM), and researchers’ interest in
the differential effects of these two WM components on various aspects of
SLA, we will discuss the results separately where it is necessary.
Furthermore, we distinguish two types of research based on their
research design: predictive and experimental. In predictive research, no
instructional treatment is implemented, and the interest is in investigat-
ing WM as a determinant of learning success. Typically, the researcher
obtains two sets of scores—one for WM and one for language proficiency—
and conducts analyses of a correlational nature to determine the pre-
dictive power of WM for learning outcomes. In experimental research,
by contrast, the researcher explores the effects of WM under carefully
manipulated treatment conditions. We argue that experimental WM stud-
ies, especially those investigating whether WM plays different roles under
different learning conditions, are of more theoretical and pedagogical sig-
nificance. Results of this kind are more revealing about the mechanisms
through which WM influences the process and outcome of L2 learning.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

372 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
Therefore, they are likely to provide more valuable implications for ways
to adjust instruction to cater to learners’ individual differences in WM.
15.3.1 Overall Effects of WM on SLA
Recently, two meta-analyses have been conducted on the associations
between WM and aspects of L2 learning and processing. One, which is
conducted by Linck et al. (2014), aggregated 748 correlation coefficients
reported in 79 studies. The meta-analysis showed an overall weak cor-
relation between all measures of WM and language proficiency, r = .25;
executive WM was found to be more predictive of learning than pho-
nological WM, r = .27 and .17, respectively; and verbal measures (i.e.,
language-related) demonstrated stronger predictive power than nonver-
bal measures, r = .26 versus r = .18. However, the meta-analysts took a
coarse-grained approach without distinguishing predictive and experi-
mental studies, and they investigated whether different components of
WM had differential effects on specific aspects of learning such as vocab-
ulary or reading comprehension. Also, the meta-analysis did not seem
to exercise strict quality control, and it left out some relevant studies.
Notwithstanding, the results were based on a large number of studies
and are indicative of the overall effects of WM on SLA.
The overall weak predictive power of WM is confirmed by Li’s (in press)
meta-analysis of the research on the associations between two cognitive
aptitudes—WM and language aptitude—and the process and product
aspects of second language interaction. The process aspects of interaction
refer to what happens during interaction or features that may be condu-
cive to learning, such as the noticing of corrective feedback and modified
output; the product of interaction refers to the effects of instructional
treatments measured through pre-tests and post-tests. The meta-analysis
showed that the predictive power of WM for the process dimensions of
interaction was variable and inconsistent and that its correlations with the
effects of corrective feedback were weak: r = .23 for immediate effects, and
r = .08 for delayed effects. Conversely, language aptitude was found to be a
much stronger predictor of treatment effects: r = .42 for immediate effects
and r = .32 for delayed effects. Based on the results of this and Linck et al.’s
(2014) meta-analyses, Li argues that as a domain-general cognitive device
implicated in almost all areas of academic learning, WM does not appear
to be as important as language aptitude, a domain-specific construct that
is exclusive to language learning. In addition, Li also coded the different
types of instructional treatments as explicit and implicit and found that
both WM and language aptitude were more strongly correlated with the
effects of explicit treatments than implicit treatments. These results sug-
gest that both cognitive factors are essential for conscious learning.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 373
15.3.2 WM on L2 Vocabulary and Grammar Learning
15.3.2.1 Vocabulary
The studies on vocabulary can be divided into two broad types: those
examining vocabulary size as the outcome variable (predictive research)
and those investigating treatment effects on learning rate (experimen-
tal research). Within each category, the research can be further divided
according to whether the predictor variable is phonological short-term
memory (PSTM) or executive WM. Phonological short-term memory has
been found to be a predictor of vocabulary size, and the finding was
obtained for both adults (Hummel, 2009) and young children (Farnia
& Geva, 2011; Masoura & Gathercole, 2005). Executive WM, however,
has been found to be a weak and unstable predictor of vocabulary size
(D’Angiulli, Siegel, & Serra, 2004; Engel de Abreu & Gathercole, 2012;
Jean & Geva, 2009). Experimental studies typically include one or more
treatment sessions where learners were engaged in so-called “paired
associate” tasks in which learners were presented with L2 words and
their first language (L1) translations. This research shows that phonolog-
ical short-term memory was important for learning new vocabulary at
initial stages of learning (Atkins & Baddeley, 1998; Martin & Ellis, 2012;
Speciale, Ellis, & Bywater, 2004). However, at more advanced stages, where
learners had an extensive learning experience, phonological short-term
memory stopped being a significant predictor, and previous vocabulary
knowledge emerged as a more important factor (Cheung, 1996; French
& O’Brien, 2008; Masoura & Gathercole, 2005). Executive WM was also
found to be predictive of learning rate in studies where learners had no
background in the target language (Kempe, Brooks, & Kharkhurin, 2010;
Martin & Ellis, 2012), but because of the lack of relevant research, it
remains unknown whether it is less predictive of vocabulary learning at
more advanced stages of learning—as has been found for phonological
short-term memory.
We would like to highlight the fact that in almost all experimental stud-
ies on the role of WM in vocabulary learning, treatment tasks take the
form of discrete mechanical practice where learners were asked to memo-
rize target words and their meanings. This type of instruction emphasizes
rote learning, which is characteristic of traditional audiolingual classes.
It deviates from the currently popular meaning-oriented approach where
linguistic forms are learned and practised through meaning-primary com-
prehension and production tasks. There is, therefore, an urgent need for
investigating how WM fares in such tasks. Such tasks are of particular
relevance to executive WM, which is posited to be important for the learn-
ing that happens in real-time tasks where learners need to engage in the
constant shift between form and meaning.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

374 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
15.3.2.2 Grammar
Predictive research has show that phonological short-term memory and
executive WM were both correlated with grammar learning (Engel de
Abreu & Gathercole, 2012; Hummel, 2009), even after partialling out the
influence of learners’ previous grammar knowledge (French & O’Brien,
2008). However, Serafini and Sanz (2016) reported that WM was more
relevant for initial grammar learning, and as learners moved to more
advanced levels, its effects weakened—a pattern similar to vocabulary
learning. In this study, L2 Spanish learners at beginning, intermediate,
and advanced levels were tested on ten grammatical features at three time
points during and after a semester of instruction. It was found that pho-
nological short-term memory and executive WM were only predictive of
the beginners’ and intermediate learners’, but not the advanced learners’,
grammar knowledge.
While the predictive studies simply showed significant correlations
between the predictor and criterion variables, experimental studies
revealed whether the effects of WM had to do with the characteristics of
the instructional treatments. Two experimental studies where learners
had no previous knowledge about the target language showed that both
types of WM were predictive of the effects of instructional treatments
involving discrete item-based learning (Kempe, Brooks, & Kharkhurin,
2010; Martin & Ellis, 2012). In both studies, the instruction was computer-
ized and learning happened inductively by understanding input materials
and receiving feedback containing the correct linguistic models. It would
seem that WM is relevant when learners are involved in the active process-
ing of input materials and when they have a heavy processing load during
real-time learning tasks. However, the above two studies only examined
one learning condition, and studies investigating the differential roles
of WM in multiple learning conditions allow us to have a clearer picture
about what instructional characteristics may contribute to the presence
and absence of WM’s impact. In this regard, the studies on corrective feed-
back are revealing.
The feedback studies investigated the interaction between WM and the
effectiveness of different types of corrective feedback, which refers to
responses to learners’ erroneous utterances. Goo (2012) found that exec-
utive WM was only predictive of the effects of recasts (the reformulation
of a wrong utterance with the meaning intact), but not those of metalin-
guistic feedback (i.e., rule explanation), on the learning of a complicated
linguistic target—the English that-trace filter. Li (2013a, b) reported that
when learners received metalinguistic correction (rule explanation + cor-
rect form), executive WM had a positive effect on the learning of Chinese
classifiers but a negative effect on the perfective –le. WM was not associ-
ated with the effects of recasts for either linguistic target. Yilmaz (2013)
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 375
reported that executive WM was unrelated to the effects of recasts but was
predictive of the effects of explicit correction, which informed the learner
of the problematic nature of an utterance followed by the provision of the
correct form.
What do these seemingly conflicting results suggest? One notion the
researchers resorted to when interpreting the absence or presence of
the effects of WM is noticing. Goo argued that WM was correlated with the
effects of recasts because recasts were implicit and WM was essential for
the noticing of the corrective force of such feedback. Metalinguistic feed-
back was explicit, which made noticing unimportant, and this, in turn,
made WM irrelevant in this learning condition. Both Li and Yilmaz inter-
preted their results as suggesting that the effects of WM were only evident
under explicit learning conditions. In other words, the effects of WM are
observable only when learners notice the corrective intention of feedback.
However, the arguments are unconvincing if we consider the fact that
in these studies the treatments contained intensive feedback targeting a
single linguistic target throughout, which made recasts very salient, even
though it was intended to be implicit. We interpret the results differently.
First, we argue that the role of WM is attributable to the onerous process-
ing burden imposed on the learner. In Goo’s recast condition and Yilmaz’s
explicit correction, learners received the correction of their wrong utter-
ances repeatedly, which required them to deploy their WM resources
to constantly store and process the received information to induce the
underlying linguistic regularities. In Li’s study, the learning of Chinese
classifiers was more item-based than rule-based, and repeated metalin-
guistic correction of the wrong use of different classifiers posed great chal-
lenges on learners’ WM. Second, noticing may indeed serve as a trigger for
WM effects, but there needs to be a distinction between different levels of
noticing. Noticing at the level of detection, or simply noticing the pres-
ence of the linguistic target as in the recast conditions of Li’s and Yilmaz’s
studies, may not have imposed participatory demands and therefore did
not tax WM. Noticing at the level of processing, as in Yilmaz’s explicit
correction and Li’s metalinguistic correction, obligated learners to analyse
the received input using their WM resources. Third, the metalinguistic
feedback in Goo’s and Li’s studies consisted of rule explanations about
complicated linguistic structures; the feedback may not have imposed a
burden on WM resources and may have implicated language analytic abil-
ity instead, as Li (2013a) showed.
WM, therefore, may not be important for the learning of metalinguistic
knowledge—a claim that has been borne out by further empirical evidence.
For example, Martin and Ellis (2012) found that although phonological
short-term memory was predictive of grammar learning tested through
translation tasks, it was not predictive of learners’ ability to describe
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

376 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
grammar rules after receiving the treatment that encouraged inductive
learning. Roehr and Gánem-Gutiérrez (2009) found that executive WM
was not a significant predictor of L2 metalinguistic knowledge, and that
the two constructs also clustered around different factors. Therefore, we
can conclude, albeit tentatively, that WM, be it phonological or executive
WM, is important for grammar learning, but may not be important for the
learning of metalinguistic knowledge.
A related issue is whether WM facilitates the acquisition of explicit or
implicit knowledge. Ellis (2005) defined explicit knowledge as (1) primar-
ily knowledge of which learners are aware and (2) secondarily knowledge
that learners can verbalize, that is metalinguistic knowledge or, rather,
metalanguage. The former is measurable through tasks that allow learn-
ers to access their conscious knowledge, such as grammaticality judgment,
while the latter is gauged through tasks tapping into their knowledge
about rules and terms about language per se. We have pointed out that
WM may not be relevant for the acquisition of metalanguage. The ques-
tion that remains is whether it leads to conscious knowledge. Implicit
knowledge, Ellis stated, refers to knowledge which learners are not aware
of and cannot verbalize; it is the knowledge accessible in spontaneous
language production. Implicit knowledge has been measured through
elicited imitation or oral production tasks. One problem that prevents us
from reaching firm conclusions about whether WM facilitates explicit or
implicit knowledge is the failure in many studies to distinguish the two
knowledge types in the measurement of treatment effects.
Based on limited evidence, it appears that the effects of executive WM
were evident on tests of explicit (Goo, 2012; Li, 2013a) as well as implicit
knowledge (Kim, Payant, & Pearson, 2015; Li, 2013a; Li, Ellis, & Zhu, in
press). Révész (2012) reported that phonological short-term memory was
only correlated with learners’ scores on an oral test but not on a written
test, while the opposite was true for executive WM. Révész pointed out
that this may suggest that phonological short-term memory was impor-
tant for the acquisition of implicit knowledge while executive WM was
facilitative of explicit knowledge. Obviously, Révész’ speculation needs
further verification. Our view is that the knowledge learned through WM,
whether it is phonological short-term memory or executive WM, is per-
haps partly explicit and partly implicit, although phonological short-term
memory is more likely to lead to implicit knowledge because it taps the
ability to internalize linguistic input as unanalysed chunks.
Thus far, we have shown that the influence of WM surfaces when the
treatment levies processing pressure on the learner. Is there evidence
showing that the role of WM is wiped out by means of alleviating learn-
ers’ processing burden? The findings of several recent studies suggest a
positive answer to this question. Suzuki and DeKeyser (2017) found that
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 377
executive WM was correlated with the effects of massed instruction where
the treatment was intensive, rather than the effects of spaced instruction
where the treatment was spread out. Sanz et al. (2016) conducted two
experiments that differed in one aspect—whether grammar instruction
was provided prior to practise activities—and found that executive WM
was a significant predictor when there was no prior grammar instruction.
The researchers argued that providing prior grammar instruction lessened
learners’ cognitive burden and levelled out the effects of WM. This spec-
ulation was further confirmed by Li, Ellis, & Zhu. (in press), who reported
that executive WM was not a significant predictor of the effects of the
treatment comprised of pre-task instruction followed by communicative
tasks, nor was it predictive of the treatment where learners received cor-
rective feedback after the learners completed the communicative tasks.
However, it was a significant predictor of the effects of the treatment con-
dition where the learners received corrective feedback during their task
performance.
15.3.3 WM on L2 Skills Learning
15.3.3.1 Reading
We would like to start this section by reporting the results of Daneman
and Merikle’s (1996) meta-analysis of the research into the associations
between WM and reading comprehension in L1, learning. The meta-
analysis showed that both executive WM and phonological WM were sig-
nificantly predictive of reading comprehension, but the former showed
stronger predictive power. The finding is not surprising given that exec-
utive WM involves not only information storage but also information
processing, which is critical for reading comprehension. In L2 research, a
similar pattern has been observed. It has been found that executive WM,
but not phonological WM, was predictive of reading comprehension (Geva
& Ryan, 1993; Harrington & Sawyer, 1992).
Further evidence for the importance of the executive component of
WM in reading comprehension comes from Alptekin and Erçetin (2009),
who distinguished literal and inferential reading, with the former refer-
ring to questions whose answers are transparent and could be easily
located in the reading text and the latter to questions whose answers
must be inferred. The researchers found executive WM was only corre-
lated with inferential reading, but not literal reading. In a similar vein,
Andersson (2010) investigated the effects of the central executive or
attention control (aka, executive WM) and phonological WM on dialogue
comprehension—a simple task—and story comprehension—a complex
task. It was found that only executive WM was a significant predictor of
the complex task. Another theme that has emerged is that for L2, but not
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

378 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
L1, WM measures were correlated with L2 reading (Alptekin & Erçetin,
2010; Miyake & Friedman, 1998; Geva & Ryan, 1993), suggesting the
impact of L2 knowledge in the mediating role of WM. Furthermore, simi-
lar to the pattern found in grammar and vocabulary, there is evidence for
the possibility that WM is more important for reading comprehension at
initial rather than advanced stages of L2 learning (Walter, 2004).
Overall the research on reading is predominantly predictive, exploring
whether WM is correlated with reading comprehension. There is a lack
of experimental research investigating whether and how WM is related
to the effects of reading treatments on the development of reading com-
prehension and L2 knowledge. One study that exemplifies how this line of
research can be conducted is by Leeser (2007). It examined the effects of
topic familiarly and WM on meaning comprehension and noticing of a
grammar feature. Leeser identified a very interesting interaction between
the independent and dependent variables. The study showed that topic
familiarity had a positive influence on both comprehension of the con-
tent and learners’ noticing of the linguistic target. However, the impact
of WM on comprehension and form-related outcomes depended on topic
familiarity. Specifically, it was found that, for meaning comprehension,
WM was facilitative for familiar topics, but for form recognition, it had
a favourable effect when the topics were unfamiliar. The author spec-
ulated that for comprehension, WM was relevant for familiar topics
because it was a space where previous knowledge was integrated with
current information of the ongoing text. However, when it comes to the
processing of linguistic forms, high WM may have a harmful effect when
the topic was familiar because additional cognitive resources resulting
from learners’ previous knowledge about the topics may have interfered
with linguistic processing. We are not certain about the plausibility of the
researcher’s interpretations of the results, and we also doubt whether form-
recognition (asking learners to spot forms they saw from the reading texts)
is a reliable measure of learning. However, we deem the study an initiative
demonstrating the fruitfulness of extending the scope of inquiry beyond
comprehension and examining how reading treatments may interact with
WM in affecting both comprehension ability and linguistic development.
15.3.3.2 Speaking
In addition to the distinction between predictive and experimental
research, the research on speaking can be further split into two sub-
categories according to whether speech performance was assessed subjec-
tively or objectively. Subjective measures refer to raters’ impressionistic
judgments, while objective measures refer to indexes of fluency (e.g.,
pause length), accuracy (e.g., the number of errors), and complexity
(e.g., ratio of subordinate clauses), calculated based on the transcripts of
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 379
learners’ speech samples. Predictive research showed that on subjective
measures, positive correlations were found between executive WM and
fluency (Fehringer & Fry, 2007) as well as overall performance (Kormos
& Sáfár, 2008). Phonological WM was uncorrelated with rated accented-
ness (Hu et al., 2013) or comprehensibility (Venkatagiri & Levis, 2007). This
finding suggests that the phonological component may be primarily a
device for storing novel linguistic stimuli and that it may not be important
for sound production or the phonological aspects of speech production
(similar to the argument made by Baddeley, Gathercole, & Papagno, 1998).
With regard to objective measures, O’Brien et al. (2006, 2007) reported
that phonological WM was a significant predictor of the development of
the fluency and accuracy aspects of L2 oral production.
A few experimental studies have been conducted based on the theoret-
ical models of task-based learning (Robinson, 2011; Skehan, 2014), which
claim that manipulating the procedural and conceptual aspects of tasks may
impact task performance in different ways; Robinson called the two types
of variables resource-dispersing and resource-directing variables, respec-
tively. Skehan argued that due to L2 learners’ limited cognitive resources,
tasks should be implemented in a way that eases the burden on learners’
WM, such as by allowing learners to plan before performing an oral task.
Robinson’s theory focuses more on task complexity, holding that complex
tasks direct learners’ attention to complex linguistic forms. Robinson pre-
dicts that the effects of individual differences are more likely to appear
when learners perform complex tasks than simple tasks.
Several studies based on Skehan’s theory have investigated the medi-
ating effects of WM on speech performance in strategic and within-task
planning conditions. In strategic planning, learners are allowed to plan the
content and language prior to task performance, whereas in within-task
planning learners are encouraged to plan during task performance. The
research shows that while executive WM was related to the effects of
within-task planning, it was not related to the effects of strategic plan-
ning (Ahmadian, 2012; Li & Fu, 2017). Li and Fu speculated that this is
because (1) the within-task planners were able to monitor their speech
using their WM resources while the strategic planners were unable to
monitor because they performed the task under time pressure, and (2) in
the strategic planning condition, learners’ processing burden during task
performance was eased in that much of the planning work was shifted
from the during-task stage to the pre-task stage. However, in order to have
a clear picture of how WM affects task performance, it needs to be investi-
gated in more task conditions such as a condition where neither strategic
nor within-task planning is allowed and a condition where both strategic
and within-task planning are allowed. Another resource-dispersing varia-
ble that has been investigated is task structure. Kormos and Trebits (2011)
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

380 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
divided learners into two groups: one group was required to tell a narra-
tive based on pre-sequenced pictures, and the other had to make up a nar-
rative based on unrelated pictures. The study showed that executive WM
was correlated with the performance of the structured task, but not of the
unstructured task. The researchers speculated that this was because when
performing the structured task, the learners had to find linguistic forms
to match the prescribed content. In the unstructured task, however, the
learners had the freedom to select the forms available in their linguistic
repertoire and avoid those with which they were unfamiliar, thus alleviat-
ing the burden on WM.
One study based on Robinson’s theoretical model examined whether
the influence of WM varied as a function of task complexity, operational-
ized as presence or absence of reasoning demands (Crespo, 2011). Two ver-
sions of the same decision-making task were developed. In the complex
task, the learners had to consider more factors and figure out the relation-
ships between more elements, and they had access to fewer resources.
WM was tested in three ways: as executive WM, phonological WM, and
attention control (the Central Executive). Only phonological WM showed
a significant effect, under both task conditions. The study failed to con-
firm Robinson’s prediction that complex tasks are more likely to impli-
cate WM.
As can be seen, there is a lack of research on the potentially com-
plex relationships between WM and the various resource-directing
and resource-dispersing variables in affecting speech production. Also,
although major theoretical models of task-based learning all posit an
important role for WM, there is a lack of fine-grained theoretical elabora-
tion of how the different components of WM are related to the different
aspects of speech production.
15.3.4 Summing-Up
The above research synthesis has provided a clearer (though not yet com-
plete) picture of emerging patterns regarding the distinctive effects of
the two WM components (i.e., phonological WM and executive WM) on
specific SLA learning domains and activities. More specifically, regarding
phonological WM, it has been generally found to be closely related to
some acquisitional and developmental aspects of such mental representa-
tional domains as vocabulary learning (e.g., in acquiring new phonolog-
ical forms, in predicting vocabulary size, etc.), grammar learning (e.g.,
in learning of implicit or unanalysed chunk-based knowledge, particu-
larly among beginning and intermediate learners as opposed to advanced
learners), and development of L2 fluency (by objective assessment). On
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 381
the other hand, the executive WM component has been shown to be
more relevant to a number of cognitively demanding processes mainly
involved in L2 skill learning, including selective processes in reading
and speaking as well as some L2 interactional processes in grammar
learning (explicit learning, noticing of corrective feedback, etc.). Taken
together, these results and findings from most of these current WM–SLA
empirical studies lend support to the basic tenet and hypotheses of the
Phonological/Executive (P/E) Model (Wen, 2016), which point to the dis-
tinct effects of phonological and executive WM on specific SLA domains
and skills.
Notwithstanding, the research synthesis conducted here also revealed
the lack of studies looking into some potentially important areas of
SLA where WM is likely to exert significant impacts. These include, for
example, the acquisition and development of L2 formulaic sequences
or chunks, both of which, as postulated by the P/E Model, are likely to
implicate phonological WM among L2 learners of low proficiency, and
executive WM among more advanced L2 learners (Ellis, 1996, 1997, 2012,
2017; Wen, 2016, 2017; also see Skrzypek, 2009, and Foster, Bolibaugh, &
Kotula, 2014, for initial attempts). In terms of L2 skills learning, very few
studies have probed into WM effects on listening and writing. Finally, in
terms of research design, most of the empirical studies reviewed here are
cross-sectional; there is an obvious lack of studies conducted by incorpo-
rating the longitudinal design that can track down the trajectory growth
of the two WM components as participants’ L2 proficiency progresses.
All these issues and many others should hold great promise for future
investigations towards a more comprehensive picture of the WM–SLA
nexus.
15.4 Conclusion and Implications for Future Research
It should be apparent from the chapter by now that the initial motivation
for developing an integrative framework for implementing WM in SLA is
to work out the common grounds that are shared by both WM research
in cognitive science on the one hand and language learning in applied
linguistics on the other (Wen, 2016). Despite the considerable research on
both sides, almost all of them have been done separately up until now. As
a matter of fact, given the interdisciplinary nature of this line of inquiry, it
should be beneficial if both fields can communicate with each other more
effectively (Wen, Mota, & McNeil, 2015). We thus call for more concerted
efforts from both sides towards this interdisciplinary research agenda for
WM and SLA (Wen, 2016).
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

382 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
In terms of theoretical implications, as we discussed in section 15.2,
recent years have witnessed a paradigm shift from the “structural” view
to the “functional” view of WM conception in current cognitive psy-
chology, particularly within cognitive neuroscience. However, while the
multiple components of WM as conceived in Baddeley’s structural view
are relatively stable, the exact nature and structure of cognitive mech-
anisms and executive functions associated with the two components
respectively, or the WM construct as a whole, are still far from conclu-
sive. So, it is hoped that more efforts can be directed towards applying
state-of-the-art technology (eye-tracking, event-related potentials, func-
tional magnetic resonance imaging, etc.) to pin down the core memory
mechanisms and executive functions underlying the maintenance, access,
and control of the two languages in the bilingual brain (Altarriba & Isurin,
2013).
Such a lack of progress in theoretical advances also gives rise to two
direct consequences for methodology. First of all, despite the shift in both
the theoretical conception and research focus of WM, research into the
assessment procedures of WM functions are even more seriously lagging
behind. Besides, there is also the inherent problem with current WM span
tasks, with their implicit assumption of a monolingual participant. Such
an assumption actually runs anathema to the current social trend, which
indicates that we are facing a growing number of bilinguals (as opposed
to monolinguals). Therefore, future research should not only focus on the
theoretical demarcation of WM functions but also develop reliable and
valid WM span tasks that can tap into these fine-grained executive func-
tions during SLA.
To end the chapter then, it can be concluded that an integrative
perspective of WM and SLA has its theoretical foundation in, and
derives strong empirical support from, the broad theoretical concep-
tualizations of the WM construct in contemporary cognitive sciences
in alignment with the more specific WM-related explorations in both
cognitive psychology and applied linguistics. Thus, it is hoped that com-
bining research insights from both fields to advocate this more prin-
cipled approach to implementing WM in practical SLA research will
have significant theoretical and methodological implications for future
research and practice. Given that the interdisciplinary enterprise of WM
and SLA is still at its infancy stage (Williams, 2015), there is still a wealth
of issues and topics to be explored before we can have a complete pic-
ture of the fine-grained specifications of the intricate links constituting
the “WM–SLA nexus” (Wen, 2012, 2016). Therefore, it constitutes our
ongoing undertaking and our ultimate goal to integrate WM theories
with SLA theories so as to advance developments in both fields towards
synergy effects.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 383
References
Ahmadian, M. J. (2012). The relationship between working memory capac-
ity and L2 oral performance under task-based careful online planning
condition. TESOL Quarterly, 46, 165–175.
Alptekin, C., & Erçetin, G. (2009). Assessing the relationship of working
memory to L2 reading: Does the nature of comprehension process and
reading span task make a difference? System, 37, 627–639.
Alptekin, C., & Erçetin, G. (2010). The role of L1 and L2 working memory
in literal and inferential comprehension in L2 reading. Journal of Research
in Reading, 22, 206–219.
Altarriba, J., & Isurin, L. (2013). Memory, language, and bilingualism: Theoretical
and applied approaches. Cambridge: Cambridge University Press.
Andersson, U. (2010). The contribution of working memory capacity to for-
eign language comprehension in children. Memory, 18, 458–472.
Andrade, J. (2001). Working memory in perspective. Hove: Taylor & Francis.
Atkins, P. W. B., & Baddeley, A. D. (1998). Working memory and distrib-
uted vocabulary learning. Applied Psycholinguistics, 19, 537–552.
Baddeley, A. D. (1986). Working memory. Oxford: Clarendon.
Baddeley, A. D. (2003). Working memory and language: An overview.
Journal of Communication Disorders, 36(3), 189–208.
Baddeley, A. D. (2012). Working memory: Theories, models and controver-
sies. Annual Review of Psychology, 63, 1–30.
Baddeley, A. D. (2015). Working memory in second language learning. In
Z. Wen, M. Mota, & A. McNeill (eds.), Working memory in second language
acquisition and processing (pp. 17–28). Bristol: Multilingual Matters.
Baddeley, A. D. (2017). Modularity, working memory and language acqui-
sition. Second Language Research, 33(3), 299–311.
Baddeley, A. D., Gathercole, S. E., & Papagno, C. (1998). The phonological
loop as a language learning device. Psychological Review, 105, 158–173.
Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. A. Bower (ed.),
The psychology of learning and motivation (Vol. 8). New York: Academic
Press.
Barrouillet, P., & Camos, V. (2012). As time goes by: Temporal constraints in
working memory. Current Directions in Psychological Science, 21(6), 413–419.
Barrouillet, P., & Camos, V. (2015). Working memory: Loss and reconstruction.
Hove: Psychology Press.
Carruthers, P. (2013). The evolution of working memory. Proceedings of the
National Academy of Sciences, 110(S2), 10371–10378.
Carruthers, P. (2015). The centered mind: What the science of working mem-
ory shows us about the nature of human thought. Cambridge: Cambridge
University Press.
Cheung, H. (1996). Nonword span as a unique predictor of second-language
vocabulary learning. Developmental Psychology, 32, 867–873.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

384 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
Christiansen, M. H., & Chater, N. (2016). The now-or-never bottleneck: A
fundamental constraint on language. Behavioral & Brain Sciences, 39, 1–72.
Constantinidis, C., & Klingberg, T. (2017). The neuroscience of working
memory capacity and training. Nature Reviews Neuroscience, 17(7), 438–449.
Conway, A. R. A., Jarrold, C., Kane, M. J., Miyake, A., & Towse, J. N. (eds.)
(2007). Variation in working memory. New York: Oxford University Press.
Conway, A. R. A., Moore, A. B., & Kane, M. J. (2009). Recent trends in the
cognitive neuroscience of working memory. Cortex, 45(2), 262–268.
Cowan, N. (1988). Evolving conceptions of memory storage, selective
attention, and their mutual constraints within the human information-
processing system. Psychological Bulletin, 104(2), 163–191.
Cowan, N. (1999). An embedded-processes model of working memory. In
A. Miyake & P. Shah (eds.), Models of working memory: Mechanisms of active
maintenance and executive control (pp. 62–101). Cambridge: Cambridge
University Press.
Cowan, N. (2001). The magical number 4 in short-term memory: A recon-
sideration of mental storage capacity. Behaviour and Brain Sciences, 24,
87–185.
Cowan, N. (2005). Working memory capacity. New York/Hove: Psychology
Press.
Cowan, N. (2011). Working memory and attention in language use. In J.
Guandouzi, F. Loncke, & M. J. Williams (ed.), The handbook of psycholinguis-
tics and cognitive processes (pp. 75–98). London: Psychology Press.
Cowan, N. (2014). Working memory underpins cognitive development,
learning, and education. Educational Psychology Review, 26, 197–223.
Cowan, N. (2017). The many faces of working memory and short-term stor-
age. Psychonomic Bulletin & Review, 24, 1158–1170.
Crespo, M. (2011). The effects of task complexity on L2 oral production
as mediated by differences in working memory capacity. Unpublished
master’s thesis, University of Barcelona.
Daneman, M., & Carpenter, P. A. (1980). Individual differences in working
memory and reading. Journal of Verbal Learning and Verbal Behaviour, 19,
450–466.
Daneman, M., & Merikle, P. M. (1996). Working memory and language com-
prehension: A meta-analysis. Psychonomic Bulletin & Review, 3, 422–433.
D’Angiulli, A., Siegel, L. S., & Serra, E. (2001). The development of reading
in English and Italian in bilingual children. Applied Psycholinguistics, 22(4),
479–507.
Dehn, M. J. (2008). Working memory and academic learning: Assessment and
intervention. Hoboken, NJ: John Wiley.
D’Esposito, M., & Postle, B. R. (2015). The cognitive neuroscience of work-
ing memory. Annual Review of Psychology, 66, 115–142.
Ellis, N. C. (1996). Sequencing in SLA: Phonological memory, chunking
and points of order. Studies in Second Language Acquisition, 18, 91–126.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 385
Ellis, N. C. (1997). The epigenesis of language: Acquisition as a sequence
learning problem. In A. Ryan and A. Wray (eds.), Evolving models of lan-
guage (pp. 41–57). Clevedon: Multilingual Matters.
Ellis, N. C. (2012). Formulaic language and second language acquisition:
Zipf and the phrasal Teddy Bear. Annual Review of Applied Linguistics, 32,
17–44.
Ellis, N. C. (2017). Chunking. In M. Hundt, S. Mollin, & S. Pfenninger (eds.),
The changing English language: Psycholinguistic perspectives (pp. 113–147).
Cambridge: Cambridge University Press.
Ellis, R. (2005). Measuring implicit and explicit knowledge of a second lan-
guage: A psychometric study. Studies in Second Language Acquisition, 27,
141–172.
Engel de Abreu, P. M., & Gathercole, S. E. (2012). Executive and phono-
logical processes in second-language acquisition. Journal of Educational
Psychology, 104, 974–986.
Engle, R. W., & Kane, M. (2004). Executive attention, working memory
capacity, and a two-factor theory of cognitive control. In B. Ross (ed.),
The psychology of learning and motivation (Vol. 44, pp. 145–199). New York:
Elsevier.
Engle, R. W., Cantor, J., & Carullo, J. J. (1992). Individual differences in
working memory and comprehension: A test of four hypotheses. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 18(5), 972–992.
Farnia, F., & Geva, E. (2011). Cognitive correlates of vocabulary growth in
English language learners. Applied Psycholinguistics, 32, 711–738.
Fehringer, C., & Fry, C. (2007). Hesitation phenomena in the language
production of bilingual speakers: The role of working memory. Folia
Linguistica, 41(1–2), 37–72.
Fenesi, B., Sana, F., Kim, J. A., & Shore, D. (2015). Reconceptualizing work-
ing memory in educational research. Educational Psychology Review, 27(2),
333–351.
Foster, P., Bolibaugh, C., & Kotula, A. (2014). Knowledge of nativelike selec-
tions in an L2: The influence of exposure, memory, age of onset and
motivation in foreign language and immersion settings. Studies in Second
Language Acquisition, 36(1), 101–132.
French, L. M., & O’Brien, I. (2008). Phonological memory and children’s
second language grammar learning. Applied Psycholinguistics, 29, 463–487.
Futrell, R., Mahowald, K., & Gibson, E. (2015). Large-scale evidence of
dependency length minimization in 37 languages. Proceedings of the
National Academy of Sciences, 112(33), 10336–10341.
Gathercole, S. E. (2007). Working memory and language. In G. Gaskell
(ed.), Oxford handbook of psycholinguistics (pp. 757–770). Oxford: Oxford
University Press.
Gathercole, S., & Baddeley, A. (1993). Working memory and language. Hove:
Lawrence Erlbaum.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

386 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
Geva, E., & Ryan, E. (1993). Linguistic and cognitive correlates of academic
skills in first and second languages. Language Learning, 43, 5–42.
Gibson, E. (1998). Linguistic complexity: Locality of syntactic dependen-
cies. Cognition, 68, 1–76.
Goo, J. (2012). Corrective feedback and working memory capacity in
interaction-driven L2 learning. Studies in Second Language Acquisition, 34,
445–474.
Grundy, J. G., & Timmer, K. (2017). Bilingualism and working memory
capacity: A comprehensive meta-analysis. Second Language Research, 33(3),
325–340.
Harrington, M., & Sawyer, M. (1992). L2 working memory capacity and L2
reading skill. Studies in Second Language Acquisition, 14, 25–38.
Hasson, U., Chen J., & Honey, C. J. (2015). Hierarchical process memory:
Memory as an integral component of information processing. Trends in
Cognitive Sciences, 19, 304–313.
Hawkins, J. A. (2004). Efficiency and complexity in grammars. Oxford: Oxford
University Press.
Hu, X., Ackermann, H., Martin, J., Erb, M., Winklere, S., & Reiterer, S.
(2013). Language aptitude for pronunciation in advanced second lan-
guage (L2) learners: Behavioural predictors and neural substrates. Brain
and Language, 127, 366–376.
Hummel, K. (2009). Aptitude, phonological memory, and second language
proficiency in nonnovice adult learners. Applied Psycholinguistics, 30,
225–249.
James, W. (1890). Principles of psychology (Vol. 1). London: McMillan &
Company.
Jean, M., & Geva, E. (2009). The development of vocabulary in English as
a second language and its role in predicting word recognition ability.
Applied Psycholinguistics, 30, 153–185.
Jones, G., & Macken, B. (2015). Questioning short-term memory and its
measurement: Why digit span measures long-term associative learning.
Cognition, 144, 1–13.
Kempe, V., Brooks, P. J., & Kharkhurin, A. (2010). Cognitive predictors
of generalization of Russian grammatical gender categories. Language
Learning, 60, 127–153.
Kim, Y., Payant, C., & Pearson, P. (2015). The intersection of task-based
interaction, task complexity, and working memory. Studies in Second
Language Acquisition, 37, 549–581.
Klingberg, T. (2008). The overflowing brain: Information overload and the limits
of working memory. Oxford: Oxford University Press.
Kormos, J., & Sáfár, A. (2008). Phonological short-term memory, working
memory and foreign language performance in intensive language learn-
ing. Bilingualism: Language and Cognition, 11(2), 261–271.
Kormos, J., & Trebits, A. (2011). Working memory capacity and narrative
task performance. In P. Robinson (ed.), Second language task complexity:
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 387
Researching the Cognition Hypothesis of language learning and performance
(pp. 267–285). Amsterdam: John Benjamins.
Leeser, M. (2007). Learner-based factors in L2 reading comprehension and
processing grammatical form: Topic familiarity and working memory.
Language Learning, 57, 229–270.
Li, S. (2013a). The interactions between the effects of implicit and explicit
feedback and individual differences in language analytic ability and
working memory. The Modern Language Journal, 97, 634–654.
Li, S. (2013b). The differential roles of language analytic ability and work-
ing memory in mediating the effects of two types of feedback on the
acquisition of an opaque linguistic structure. In C. Sanz & B. Lado (eds.),
Individual differences, L2 development & language program administration:
From theory to application (pp. 32–52). Boston, MA: Cengage Learning.
Li, S. Ellis, R., & Zhu, Y. (in press). The associations between cognitive abil-
ity and L2 development under five different instructional conditions.
Applied Psycholinguistics.
Li, S., & Fu, M. (2017). Strategic and unpressured within-task planning and
their associations with working memory. Language Teaching Research, 22,
1–24.
Linck, J., Osthus, P., Koeth, J. T., & Bunting, M. (2014). Working memory
and second language comprehension and production: A meta-analysis.
Psychonomic Bulletin & Review, 21, 861–883.
Liu, H., Xu, C., & Liang, J. (2017). Dependency distance: A new perspec-
tive on syntactic patterns in natural languages. Physics of Life Review, 21,
171–193.
Martin, K. I., & Ellis, N. C. (2012). The roles of phonological short-term
memory and working memory in L2 grammar and vocabulary learning.
Studies in Second Language Acquisition, 34, 179–413.
Masoura, E. V., & Gathercole, S. (2005). Contrasting contributions of pho-
nological short-term memory and long-term knowledge to vocabulary
learning in a foreign language. Memory, 13, 422–429.
Miller, G. (1956). The magical number of seven, plus or minus two: Some
limits on our capacity for processing information. Psychological Review,
63, 81–97.
Miller, G. (2003). The cognitive revolution: A historical perspective. Trends
in Cognitive Science, 7(3), 141–144.
Miyake, A., & Friedman, N. (1998). Individual differences in second lan-
guage proficiency: Working memory as language aptitude. In A. Healy &
L. Bourne (eds.), Foreign language learning: Psycholinguistic studies on training
and retention (pp. 339–364). Mahwah, NJ: Erlbaum.
Miyake, A., & Shah, P. (1999). Models of working memory: Mechanisms of active
maintenance and executive control. New York: Cambridge University Press.
Oberauer, K., & Lewandowsky, S. (2013). Evidence against decay in verbal
working memory. Journal of Experimental Psychology: General, 142, 380–411.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

388 ZHISHENG (EDWARD) WEN AND SHAOFENG LI
Oberauer, K., & Lewandowsky, S. (2014). Further evidence against decay in
working memory. Journal of Memory and Language, 73, 15–30.
Oberauer, K., & Lewandowsky, S. (2016). Control of information in work-
ing memory: Encoding and removal of distractors in the complex-span
paradigm. Cognition, 156, 106–128.
Oberauer, K., Farrell, S., Jarrold, C., & Lewandowsky, S. (2016). What limits
working memory capacity? Psychological Bulletin, 142(7), 758–799.
Oberauer, K., Lewandowsky, S., Farrell, S., Jarrold, C., & Greaves, M. (2012).
Modeling working memory: An interference model of complex span.
Psychonomic Bulletin & Review, 19, 779–819.
O’Brien, I., Segalowitz, N., Collentine, J., & Freed, B. (2006). Phonological
memory and lexical, narrative, and grammatical skills in second lan-
guage oral production by adult learners. Applied Psycholinguistics, 27,
377–402.
O’Brien, I., Segalowitz, N., Collentine, J., & Freed, B. (2007). Phonological
memory predicts second language oral fluency gains in adults. Studies in
Second Language Acquisition, 29, 557–582.
O’Grady, W. (2017). Working memory and language: From phonology to
grammar. Applied Psycholinguistics, 38(6), 1340–1343.
Pierce, L. J., Genesee, F., Delcenserie, A., & Morgan, G. (2017). Variations
in phonological working memory: Linking early language experiences
and language learning outcomes. Applied Psycholinguistics, 38, 1265–1302.
Révész, A. (2012). Working memory and the observed effectiveness of recasts
on different L2 outcome measures. Language Learning, 62(1), 93–132.
Robinson, P. (2011). Second language task complexity, the Cognition
Hypothesis, language learning, and performance. In P. Robinson (ed.),
Second language task complexity: Researching the Cognition Hypothesis of lan-
guage learning and performance (pp. 3–38). Amsterdam: John Benjamins.
Roehr, K., & Gánem-Gutiérrez, G. (2009). The status of metalinguistic
knowledge in instructed adult L2 learning. Language Awareness, 18(2),
165–181.
Sanz, C., Lin, H., Lado, B., Stafford, C., & Bowden, H. (2016). One size fits
all? Learning conditions and working memory capacity in ab initio lan-
guage development. Applied Linguistics, 37, 669–692.
Serafini, E., & Sanz, C. (2016). Evidence for the decreasing impact of cog-
nitive ability on second language development as proficiency increases.
Studies in Second Language Acquisition, 38, 607–646.
Skehan, P. (ed.) (2014). Processing perspectives on task performance. Amsterdam:
John Benjamins.
Skrzypek, A. (2009). Phonological short-term memory and L2 collocational
development in adult learners. EUROSLA Yearbook, 9, 160–184.
Speciale, G., Ellis, N. C., & Bywater, T. (2004). Phonological sequence learn-
ing and short-term store capacity determine second language vocabu-
lary acquisition. Applied Psycholinguistic, 25, 293–321.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,

Working Memory in L2 Learning and Processing 389
Suzuki, Y., & DeKeyser, R. (2017). Exploratory research on second language
practice distribution: An Aptitude × Treatment interaction. Applied
Psycholinguistics, 38, 27–56.
Ullman, M. T. (2001). The neural basis of lexicon and grammar in first
and second language: The declarative/procedural model. Bilingualism:
Language and Cognition, 4(1), 105–122.
Ullman, M. T. (2016). The declarative/procedural model: A neurobiological
model of language learning, knowledge and use. In G. Hickok & S. A.
Small (eds.), The neurobiology of language (pp. 953–968). New York: Elsevier.
Venkatagiri, H. S., & Levis, J. M. (2007). Phonological awareness and
speech comprehensibility: An exploratory study. Language Awareness, 16,
263–277.
Walter, C. (2004). Transfer of reading comprehension skills to L2 linked
to mental representation of text and to L2 working memory. Applied
Linguistics, 25, 315–339.
Waters, G. S., & Caplan, D. (1996). The measurement of verbal working
memory and its relation to reading comprehension. Quarterly Journal of
Experimental Psychology, 49, 51–74.
Waugh, N. C., & Norman, D. A. (1965). Primary memory. Psychological
Review, 72, 89–104.
Wen, Z. (2012). Working memory and second language learning.
International Journal of Applied Linguistics, 22, 1–22.
Wen, Z. (2015) Working memory in second language acquisition and pro-
cessing: The phonological/executive model. In Z. Wen, M. B. Mota, & M.
McNeill (eds.), Working memory in second language acquisition and processing
(pp. 41–62). Bristol: Multilingual Matters.
Wen, Z. (2016). Working memory and second language learning: Towards an inte-
grated approach. Bristol: Multilingual Matters.
Wen, Z. (2017). Using formulaic sequences to measure L2 task-based per-
formance: The role of working memory. In M. Ahmadian & M. P. García
Mayo (eds.), Multiple perspectives on task-based language teaching and learning
(pp. 31–54). Berlin: Mouton de Gruyter.
Wen, Z., Mota, M. B., & McNeill, A. (2013). Working memory and SLA:
Towards an integrated theory. Asian Journal of English Language Teaching,
23, 1–18.
Wen, Z., Mota, M. B., & McNeill, A. (2015). Working memory in second language
acquisition and processing. Bristol: Multilingual Matters.
Williams, J. N. (2015). Working memory in SLA research: Challenges and
prospects. In M. B. Mota & A. McNeill (eds.), Working memory in second
language acquisition and processing (pp. 301–307). Bristol: Multilingual
Matters.
Yilmaz, Y. (2013). Relative effects of explicit and implicit feedback: The
role of working memory capacity and language analytic ability. Applied
Linguistics, 34, 344–368.
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108333603.016
Downloaded from https://www.cambridge.org/core. Florida State University Libraries, on 11 Sep 2019 at 19:33:10, subject to the Cambridge Core terms of use,