Learning and Memory in the Bilingual Mind and Brain

Abstract

This chapter explores various aspects and assumptions of models of bilingual language processing and organization. A brief overview of theoretical language models is provided to include a discussion of the distinction between compound and coordinate bilingualism, as well as models of connectionism, hierarchical structures, and a recently proposed model of language acquisition. The underlying assumptions of these theories and models are assessed with a focus on inter‐language processing and organization. There are many experimental designs that allow for the analysis of behavioural data. Various experimental results obtained using tasks that provide insight into the processing of language through behavioural responses are discussed in relation to learning and memory. The chapter explores findings related to how languages are processed in terms of their lexical units and the underlying nature of these representations. Although the structure of language can impact its processing across languages, the inherent properties of the representations play a key role, as well.

Full text

Thank you for using the
University at Albany’s
Interlibrary Loan Service
NOTICE WARNING CONCERNING COPYRIGHT RESTRICTIONS
The copyright law of the United States (Title 17, United States Code) governs the
making of photocopies or other reproductions of copyrighted material. Under certain
conditions specified in the law, libraries and archives are authorized to furnish a
photocopy or other reproduction. One of these specific conditions is that the photocopy
or reproduction is not to be "used for any purpose other than private study, scholarship,
or research." If a user makes a request for, or later uses, a photocopy or reproduction for
purposes in excess of "fair use," that user may be liable for copyright infringement. This
institution reserves the right to refuse a copying order if, in its judgment, fulfillment of the
order would involve violation of copyright law.
Questions?
Call 442-3613 from 10:00 ~ 4:30 (weekdays)
or
Send email to libill@albany.edu

The Handbook of the Neuroscience of Multilingualism, First Edition. Edited by John W. Schwieter.
© 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.
Learning andMemory
intheBilingual Mind andBrain
ALLISON M. WILCK, JEANETTE
ALTARRIBA, ROBERTO R. HEREDIA,
ANDJOHN W. SCHWIETER
1. Introduction
Bilingual speakers operate independently in their first (L1) or second language (L2), or
interdependently in which both languages interact simultaneously as in language
mixing or code switching. How do bilingual (or multilingual) speakers organize their
linguistic systems in the brain? Are the languages organized in the same or separate
brain regions? This chapter explores various aspects and assumptions of models of
bilingual language processing and organization. A brief overview of theoretical lan-
guage models is provided to include a discussion of the distinction between compound
and coordinate bilingualism, as well as models of connectionism, hierarchical struc-
tures, and a recently proposed model of language acquisition. The underlying assump-
tions of these theories and models will be assessed with a focus on interlanguage
processing and organization.
One purpose of this chapter is to underscore the importance of bilingual research to
aid in the refinement of existing theoretical constructs that distinguish between
conceptual meaning and lexical representations, for example, and to provide further
evidence for structural assumptions of language processing models (Altarriba and
Soltano 1996). Bilingual research, in addition to helping us understand an individual
with knowledge of two or more languages, provides yet another opportunity to assess,
correct, and expand existing models and theories of language processing. Bilingualism
provides a unique perspective for understanding how words and sentences are learned,
encoded, stored, and retrieved from memory.
Additionally, how meaning is derived from language, both within and between lan-
guages, must be considered from several perspectives to include the nature of the words,
how and when the language has been acquired by the individual, and the consistency
of processing across languages in a given context. Overall, the bilingual and L2 literature
19

390 Functions and Processes
supports the view that there are various layers of information extracted from words
(e.g. physical features, conceptual meaning) that become interlinked with the acquisition
of new languages (e.g. Velan and Frost 2007, 2011; Wong etal. 2011). Within the bilingual
mind, the research largely supports models that incorporate a mixed linguistic
representation: some aspects of language share a common store while others are sepa-
rate with language specificity. By studying language processing in bilinguals and mul-
tilinguals, a better understanding of how information is generally integrated into human
memory can be obtained.
2. Overview ofTheories andModels ofBilingual/
Multilingual Memory
Human memory is a rich and complex system that has been researched in a variety of
capacities for decades from working memory (e.g. Baddeley 2003; Baddeley and Hitch
1974) to long‐term memory (Lynch 2004). However, there is not an undisputed under-
standing of how information is processed, stored, and retrieved in memory (see Baddeley
etal. 2015), and there have been numerous models proposed on how language is pro-
cessed in memory (see French and Jacquet 2004; Kroll and Tokowicz 2005). Research
findings from bilinguals and multilinguals have led to the development of models that
attempt to account for how individuals with knowledge of more than one language store
and extract information within and across languages. In this section, we provide a brief
overview of a number of models addressing bilingual and multilingual language
processing, beginning with the fundamental debate between memory store organizations
and concluding with more recent and well‐defined models (for additional discussions on
models of bilingual language processing, see Heredia and Cieslicka 2018).
2.1. Language Memory Stores
The seminal research into the uniqueness of bilingual language processing and repre-
sentation comes from Ervin and Osgood (1954). The distinction between compound and
coordinate bilingualism provides the basis for today’s models. Compound bilinguals
are those who became proficient in an L2 by relating and matching the to‐be‐learned
concepts back to those in their L1. Connecting the new language concepts to those
already in memory, it was proposed that there would be an overlap of representations
such that all linguistic information would be stored in one location within the brain. On
the other hand, coordinate bilinguals are those who acquired a new language in a con-
text that is distinct from any previous language knowledge. For example, if an individual
learned English in the home through immersion, but was explicitly taught French in a
classroom, then these two languages would come to have separate representations.
Under these conditions, English words and their French translations would not be
interpreted as having identical meanings and associations. Therefore, word translations
would not be processed and stored equivalently, and words from each language would
have their own cognitive store.

Learning andMemory intheBilingual Mind andBrain 391
An important underlying distinction between these types of bilinguals is the implica-
tion that the individual must encode information in a context‐specific manner. In other
words, for each word bilinguals encounter, they must make note of the learning condi-
tions, so as to later be able to retrieve the meaning of the word in accordance with its
appropriate language. Although acquisition may be affected by learning context
(Segalowitz and Freed 2004), this theory does not allow for the possibility of alterations
as a product of experience with the languages after the initial learning or from the nature
of the linguistic codes themselves (see Kopeliovich 2006; cross‐linguistic word features
will be discussed further in Section4 below).
In the present‐day literature, it is a rarity to encounter the compound versus coordi-
nate bilingualism distinction. Although Ervin and Osgood (1954) generated their under-
standing based on how language was acquired, the current debate tends to focus on
how linguistic information is stored and processed in memory. Updating and expand-
ing their seminal proposal, the more recent psycholinguistic literature investigates if
linguistic information in individuals with knowledge of more than one language is
stored within a single mental lexicon, or among multiple, separate stores. If multiple
stores indeed exist, the degree of interaction and interconnectivity between them
remains a topic of discussion.
Theorists supporting a single, interdependent linguistic store largely argue that the
concepts represented by words are stored as language‐free abstract meanings (see
Heredia and Cieslicka 2018, for a review of the literature supporting the shared memory
hypothesis). These concepts are then ‘tagged’ with labels corresponding to the word or
phrase associated with each known language. For example, the conceptual representa-
tion for the place one lives will be tagged in an English‐Spanish bilingual with both
house and casa. Therefore, when this concept becomes activated, both labels would be
accessible for use, and the bilingual can attend to the tag that is most applicable for the
current conversation. This shared memory hypothesis predicts that the underlying con-
cepts of information learned in one language (e.g. 3 + 4=7, or seven in English) will be
able to be retrieved in one’s L2 (siete in Spanish) without the need to relearn the notion
in each tongue. Tasks that emphasize a conceptual focus and accentuate attention
towards the meaning of words, such as recall and recognition tests of semantic meaning
and relational processing, often garner support for an interdependence model of
bilingual and multilingual language processing (see Heredia and Cieslicka 2018).
In contrast to a single memory store is the separate or independence memory hypo-
thesis (Kolers 1963). This stance posits that linguistic information is processed, stored, and
retrieved unimodally, with each learned language having its own distinct memory store.
The separation of lexicons is proposed to emphasize the quantitative differences of words
and concepts learned between languages. These differences can be derived by comparing
linguistic aspects such as word morphology, phonology, orthography, or learnability.
Furthermore, interactions between stores only occur through translation processes, and
therefore, information obtained in one language store is often not available to another.
Support for this hypothesis comes from experiments that utilize tasks that are sensitive to
data‐driven or perceptually-based factors (e.g. lexical decision tasks). Additional evidence
for this hypothesis comes from tasks involving code switching (i.e. alternating between a
bilingual’s two languages within a single sentence). For example, intermixed

392 Functions and Processes
English‐Spanish sentences (e.g. dame una hamburguesa sin LETTUCE por favor) take English‐
Spanish bilinguals longer to process than do the monolingual equivalents (give me a ham-
burger without LETTUCE please; Heredia and Altarriba 2001). When assessed using the
separate language stores hypothesis, these results can be explained with one language
store being fully searched for each word’s meaning before searching another, with search
typically beginning in the speaker’s primary language or the language that is currently
more activated (e.g. Dijkstra and Van Heuven 2002; Soares and Grosjean 1984).
To combine the concepts of both the interdependent and independent memory store
hypotheses, models with separate but interconnected bilingual memory stores have
been proposed. One of the more well‐known models, the bilingual dual‐coding theory
(Paivio and Desrochers 1980), predicts that each language of a bilingual speaker has a
separate store for its verbal code that can function independently. Within each verbal
code, there is information pertaining to the specific language’s word labels and syntax.
Although stored separately, these verbal systems are linked through translation
equivalents. The stronger the labels are between languages for a given concept map
onto each other, the stronger and more accessible the links are. For example, the transla-
tion equivalents of cheese in English and fromage in French produce a stronger association,
as compared to the related concepts of cheese and pain (bread in French).
Importantly, the bilingual dual‐coding theory can also account for many language
type effects with the inclusion of a third, language‐free image store that is linked to the
verbal stores. Language type effects refer to differences in attention, reaction time, or
other responses because of manipulations in word type (e.g. abstract, concrete, emotion
[words describing an affective state, such as joyful], or emotion‐laden [words that evoke
emotion, such as funeral]), or language presentation. There is ample evidence for a
bilingual concreteness effect in which tangible objects (e.g. chair, tiger, bottle) are better
recalled from memory than abstract words (e.g. dream, love, death; Altarriba and Bauer
2004; de Groot 1992; Farley etal. 2012). It is important to note that tangible objects or
nouns score high in imagery accessibility and often have a single translation between
languages, while abstract words are lower in imageability and often map onto multiple
translations (see Altarriba 2003). Therefore, concrete nouns can activate meaning from
all three stores (L1, L2, and imagery) of the bilingual dual‐code model, while abstract
and other word types are often limited to only the verbal encodings.
The search for a structural and functional understanding of how language is processed
in memory continues to be a predominant area of research in the fields of neuroscience
and psycholinguistics. Whether there exists a single mental lexicon that encompasses all
linguistic information, separate stores for each language, or some combination of these
is an ongoing discussion. However, many models have been proposed that incorporate
the acquisition and recall of various lexical properties that can be organized according
to their underlying theoretical assumptions.
2.2. Connectionist Models
Connectionist models of linguistic acquisition come from the field of cognitive psy-
chology. These models function to bridge linguistic knowledge and organization with
the mechanisms that operate them. In general, connectionism has been used to describe

Learning andMemory intheBilingual Mind andBrain 393
behaviours that can be attributed to neuronal connections or other basic units. When
describing language acquisition, it is common to refer to these basic conceptual units as
nodes that interact through weighted connections (see Figure1). Connections are formed
when associations between nodes are created, and their connective strength is increased
as learning occurs. This learning can arise from interactions with the environment that
allow for broad categories and rules to develop to link concepts together. For example,
when learning the meaning of an unfamiliar word, the features of the individual letters
must first be interpreted and then connected to form a meaningful word unit. Once the
unit has been established as a word, meaning can be applied to it and this meaning can
be grouped together with concepts already established in memory.
One of the most cited connectionist word recognition models is the bilingual interac-
tive activation model (BIA; Dijkstra and Van Heuven 2002), currently modified to the
BIA+ model (Chauncey etal. 2008; Grainger etal. 2010). This network model posits that
information about the physical appearance of word features initiates the activation of
letter recognition, followed by word recognition and semantic comprehension. These
various lexical levels interact to identify a meaning for the presented words, with the
underlying assumption that activation is not language specific. Information across all of
an individual’s known languages is integrated in a single lexicon, with a given word
being recognized as belonging to a specific language. When presented within a lan-
guage task, the information from both languages will become activated simultaneously
until enough evidence is gathered to determine the language most appropriate to con-
tinue using for processing. For example, if the letters PAN are presented to an English‐
Spanish bilingual, the possible conceptual meanings for this word will be activated
across both languages so that the English definition of a cooking utensil and the Spanish
definition of bread will be generated. The definitional meaning most relevant or contex-
tually appropriate wins the competition for activation. As connections between an L1
and L2 become strengthened through proficiency, translation equivalency, or semantic
links, the representations from both languages become integrated to form a lateral
inhibitory network. With strong connections, information regarding the more dominant
Figure 1 An example of a simple connectionist model. Conceptual nodes feed into larger, more
generalized nodes. Thicker lines represent greater association strength, which the model
indicates as having a higher degree of connection between the represented concepts.

394 Functions and Processes
and proficient language can be more easily inhibited and suppressed during L2
processing to allow for more efficient communication.
2.3. Hierarchical Models
While connectionist models provide an account for how multiple languages can become
connected within the mind, hierarchical models offer an alternative account to explain
this process using interdependent memory. These models propose the existence of a
single store for each language as well as a common store. In addition to processing
information of a word from its individual features, combined letter form, and conceptual
meaning, hierarchical models also assign a relative weight and location of the L1 and L2
connections between these parts of information.
The revised hierarchical model (Kroll and Stewart 1994) is one of the most cited
models in the bilingual language literature. As common to hierarchical models, it is pro-
posed that the shared concepts for words with equivalent meanings are stored in a
common conceptual system, but the individual words are divided at the lexical level by
language. Importantly, this model suggests that L1 words are directly linked to their
meaning in the conceptual store while subsequent languages must route through the L1
to reach the conceptual store.
To explain lexical development and processing in L2 acquisition, the revised hierar-
chical model predicts that learners initially link the L2 words onto their L1 translations,
which are already connected to the conceptual store (see Figure 2). As proficiency
develops in the L2, a link directly from its language’s lexical store to the conceptual store
is made. Hence a second connection to conceptual meaning is made. However, this
L2–conceptual link is weaker than the L1–conceptual link. This unbalanced connec-
tion creates asymmetrical mappings between words and concepts in bilingual
Conceptual Store
Language 2Language 1
Figure 2 A conceptual illustration of the revised hierarchical model. Associations between the
L1 and the shared conceptual store are stronger than those between the L2 and the conceptual
store. The association from the L2 to the L1 is stronger than in the L1 to the L2 direction,
indicating the assumption that L2 learning occurs by translating words into the native language.

Learning andMemory intheBilingual Mind andBrain 395
memory. Support for the asymmetrical mapping assumption has been provided by
research on L1 and L2 translational speeds that indicates bilinguals tend to be faster at
translating words from an L2 to the L1 (e.g. Tokowicz and Kroll 2007), as the model
would assert the L2–L1 direction can occur via a direct lexical access route as opposed
to an indirect, conceptually-mediated route.
Although the revised hierarchical model has been widely accepted, it also has its
critics. In particular, the concept of asymmetrical mapping has been called into
question. Both novice and expert bilinguals have been shown to demonstrate
conceptual interference effects within and between languages during a Stroop task,
which requires participants to ignore the conceptual meaning of words while only
attending to lexical features (Altarriba and Mathis 1997; but see Kroll and Tokowicz
2005, for a counter‐argument to this finding). In addition, proficient bilinguals dem-
onstrate priming effects of semantic categorization in the L2–L1 direction, indicating
a conceptual link between the languages (Finkbeiner etal. 2004; Wang and Forster
2010). The revised hierarchical model stipulates that newly acquired languages gain
meaning by relying on lexical representations. However, these results indicate that
bilinguals of all skill levels can form both conceptual and lexical links for L2 words.
While this model has been used to guide research on language acquisition, it is impor-
tant to continue to adapt and generate bilingual and multilingual language frame-
works that can encompass the empirical findings in the literature (Brysbaert and
Duyck 2010; Brysbaert etal. 2010).
2.4. Model ofL3/LN Acquisition
While the underlying assumptions that make up the models discussed above could be
expanded to incorporate third language (L3) acquisition and processing, research on
multilinguals has indicated that these populations can be unique. Recently, a new model
of language acquisition has been proposed that specifically encompasses multilingual
(fluency in more than two languages) language processing. This model has been used to
emphasize the uniqueness of language acquisition as stemming from a variety of lexical
representations becoming interconnected.
The scalpel model (Slabakova 2016) provides a new set of ideas on how existing
knowledge of languages can facilitate (or selectively hinder) the learning of a new lan-
guage. Language processing does not occur in isolation. The concepts of integrating
grammars across languages is a primary basis for this model. The model proposes that
transference of information between languages occurs with ‘scalpel‐like’ precision to
extract the relevant grammars that aid in making meaning or translating the present
words. This metaphor is used to encompass the notion that the learning of a new lan-
guage occurs at the level of syntactic and word features by selectively drawing on the
properties of languages already in memory. Aspects of a newly acquired language are
broken down and compared to already stored languages, and thus can be influenced by
linguistic features. Furthermore, this model stipulates that the transfer of knowledge
from one language to another does not necessarily occur during the initial acquisition
phase. Rather, concepts acquired in any language can influence the encoding and
retrieval of other languages, previously stored or to‐be‐learned.

396 Functions and Processes
The scalpel model emphasizes cognitive, experiential, and linguistic influences on new
language acquisition such that languages already in memory can both facilitate and inter-
fere with new language acquisition. In this vein, frequency effects become a consideration
for language learning. Learning a new language that appears to be similar to an already
proficient language in memory can benefit the individual in quickly picking up aspects that
are consistent across both. For example, cognates (i.e. words sharing meaning and form
across languages; English: composition and Spanish: composición) tend to be identified in
fluent bilinguals faster than they are identified by individuals not fluent in both languages.
However, in the case of circumstances when there is a false similarity between languages,
interference from stored linguistic information can hinder learning. For example, Bulgarian
adolescents learning English as an L2 have been shown to overgeneralize the rules of
morphology of their L1 onto their L2 (Harakchiyska 2011). When translating sentences,
Bulgarian speakers often extend the rules of pluralization to uncountable nouns that do not
have a pluralized form in English as they would in their native tongue. Thus, the sen-
tence ‘I do my homework at home’ may be incorrectly written as ‘I do my homeworks [sic] at
home’. Overall, the scalpel model attempts to incorporate the idea of foundational
information transfer between languages as a tool for L3 (or more) acquisition. However, the
model acknowledges that there are practical limitations to this transference that can cause
issues depending on the specifics of the language combinations.
3. Methodological Considerations forL2 Studies
The brief overview of various models of bilingual and multilingual language processing
provided in Section2 generates an understanding for how the focus of these theoretical
models can vary. A main underlying dispute involves the structure of mental lexicons,
namely if there exists a single, interdependent store or multiple, independent language
stores. At issue is whether the conceptual properties of words are extracted and stored
together, independent of language, or are superimposed with specific lexical features.
The use of bilingual research has become a valuable tool for assessing and updating
assumptions of language processing, in general. Bilinguals are able to represent the
meaning of words with multiple lexical codes (e.g. a pet that meows can be labelled as cat
or gato for an English‐Spanish bilingual), while monolinguals cannot. By varying the
degree of similarity of lexical codes and overlap of processing within an individual, each
model’s explanations and predictions will necessarily differ for linguistic aspects such as
language acquisition, production, comprehension, memory, and translation abilities.
Advances in technology have allowed these models and their underlying assump-
tions to be evaluated with greater acuity and precision (see van Heuven and Dijkstra
2010). Section3 reviews the existing empirical data for the models described above and
how these models are able to account for these findings.
3.1. fMRI andERP
The models of language processing that have gained the most popularity are limited in
their predictions of a theoretical structure for how processing occurs. Most do not pro-
vide explicit predictions for where in the brain the processing occurs (but see Indefrey

Learning andMemory intheBilingual Mind andBrain 397
and Levelt 2004, for an attempt to link functional components to theory). However,
technologies do allow for language processing to be localized and compared between
the processing of an L1, L2, or LN (languages acquired beyond the L2).
Functional magnetic resonance imaging (fMRI) detects changes in blood flow as a
proxy to measure brain activity. As blood flow to an area increases, it is taken as an indi-
cation that neuronal activity in that area is occurring above the typical resting rate.
Although fMRI provides high‐quality images that detail where activation is localized, it
produces low temporal resolution. Thus, information provided with fMRI is useful for
identifying where in the brain neuronal activity is occurring, but not for creating a time-
line of activation. Identifying the specific brain structures involved in tasks of reading
and listening can allow for a comparison of the processing between languages, ultimately
providing support for a single language store or multiple stores.
Event‐related potentials (ERPs) are an alternative measure of brain activity in
response to a sensory or cognitive stimulus as recorded by changes in voltage. Substantial
changes indicate greater neuronal activation. In contrast to fMRI, measures from ERP
provide low image resolution but high temporal resolution allowing for the study of
rapidly occurring linguistic processing with precise timing of the activation at various
stages (Morgan‐Short 2014; Morgan‐Short and Tanner 2014; Swaab et al. 2012). For
models that provide sequential stages or differentiation of lexical properties, and there-
fore involve a time course of processing, ERPs can be used to evaluate these assump-
tions. By comparing the results of location activation information from fMRI with the
precise timing of neuronal changes from ERP, researchers can generate a detailed under-
standing for the physical processing of language.
Looking at patterns of brain activity has allowed for the testing of assumptions
corresponding to the theoretical considerations presented in Section2. For example, what
happens at a neurological level when translating words between languages? Palmer etal.
(2010) addressed this question with fully proficient Spanish‐English and English‐Spanish
bilinguals. Participants were presented with Spanish and English concrete and abstract
word pairs with the task of indicating if both words shared a common meaning or not.
ERP recordings were taken during the task to examine lexical‐semantic activation during
word translation. The results showed a larger peak of the N400 (a common marker of
semantic activation; see Kutas and Federmeier 2011; see Lau etal. 2008, for a discussion of
lexical and semantic aspects of the N400) when translating in the L2–L1 direction,
regardless of word type, than in the L1–L2 direction. The authors explained this finding in
terms of an asymmetrical link between the L1–conceptual store and L2–conceptual store
as proposed by the revised hierarchical model of language processing. The greater L2–L1
activation of the N400, the authors argued, demonstrated the additional semantic
activation that needed to occur during a conceptual judgement due to the less direct, and
therefore weaker, connections between the L2 and the conceptual store. When the L1 is
presented first, there is a more direct and fast link to its translation equivalent in the L2.
The path to make a conceptual judgement is primed, thereby reducing the amplitude and
effort required of the neurological response. These results are indeed accounted for by the
revised hierarchical model that proposes a direct link from the L1 to the conceptual store.
Although much information can be garnered from studying the neurological compo-
nents of language processing with fMRI and ERP technologies, these methodologies

398 Functions and Processes
present several limitations that merit acknowledgment for assessing theoretical assump-
tions. First, there are practical constraints for the types of tasks that can be performed
while participants are connected to the brain‐scanning equipment. This physical limita-
tion reduces the number of naturalistic behavioural measures that can be recorded, as
these machines often require the participant to minimize motion. Consequently, com-
paring data obtained about brain activity to behavioural data must be done with cau-
tion as natural reading behaviours typically involve more freedom of bodily movement.
Second, although information about where and when in the brain linguistic information
is processed when performing a particular task can be obtained from these techniques,
neither the underlying causes for brain activation patterns nor an evaluation of the pro-
cess can be observed. Data obtained from fMRI and ERP technologies are used to imply
correlations between task performance and neurological activation. Therefore, neuro-
logical data alone cannot be used for evaluating causality, as indicated by many of the
word identification models. To create a fuller understanding of the data obtained from
fMRI and ERP studies, a combination of results with behavioural data from experimental
paradigms should be considered (but see Grey etal. 2017; McLaughlin etal. 2004).
Moreover, when interpreting fMRI and ERP data, results indicating an overlap or
separation of brain activation do not necessarily imply the absence of an alternative
explanation (Hernandez etal. 2005). By considering where language is processed, an
understanding of its function and physical structure can be obtained. However, the pos-
sibility remains that there is a neural separation of linguistic representations. A localized
brain region can account for the processing of an L2 to a differing degree than an L1, or
at differing time courses. In other words, while the physical processing of two languages
(e.g. letter analysis) may occur in given brain regions with varying degrees of activation,
a higher‐order representation (e.g. semantic meaning) can overlap between all lan-
guages. This disambiguation cannot be clarified using these methods alone.
3.2. Behavioural Methodologies
There are many experimental designs that allow for the analysis of behavioural data.
Various experimental results obtained using tasks that provide insight into the
processing of language through behavioural responses will be discussed in this section,
in relation to learning and memory.
Lexical priming, in which the recent experience of a given language input or previ-
ously presented word influences the processing of a subsequently presented language
or word, has been used to study implicit learning. This learning can occur not only at the
level of individual words (e.g. associating the meaning of a presented word with a con-
cept or related words), but also with syntactic structure (Hartsuiker and Bernolet 2017).
Languages that are learned in an informal or immersive setting, such as by children at
home, often incorporate an implicit understanding for proper syntax by mimicking the
sentence structure produced by others. When attempting to learn an L2, information
about the syntax of an L1 is often applied to the L2. This can occur even when the proper
syntax between languages does not match precisely. Furthermore, bilinguals tend to
process phrases across languages faster if the translation is literal or familiar and the
sentence structures of both languages match, as opposed to unfamiliar or figurative

Learning andMemory intheBilingual Mind andBrain 399
meanings (Carrol and Conklin 2015, 2017). The formulaic structure from one language
can carry over into another language and can facilitate processing if the structures are
equated, but disrupt processing speed if different. Furthermore, the learned structure of
an L2 can become formulaic and, when the formula is not followed, language compre-
hension will require more time and effort. These studies in figurative language, for
example, indicate that priming can occur at the multiword level across languages and
that familiar phrases can be stored in the mental lexicon as both individual words and
as a whole meaningful unit.
A recent study assessing audiovisual speech cues demonstrated how language famil-
iarity and proficiency can moderate speakers’ strategies for processing speech.
Barenholtz etal. (2016) had English monolinguals and English‐Spanish bilinguals watch
videos of balanced (equally proficient in both languages) bilingual speakers in
conversation. Participants watched and listened to clips with speakers in either a lan-
guage in which they were proficient (English) or an unfamiliar language (Spanish or
Icelandic) while an infrared‐based eye‐tracking system recorded their eye movements.
An analysis of eye movements indicated that when asked to monitor the conversation,
the individual spent the most time looking at the mouth of a speaker if she used an unfa-
miliar language, but not if the language was familiar. However, when attention was not
explicitly directed towards the conversation, eye movements were focused on the mouth
as often as they were to the eyes of the speaker, regardless of whether or not the lan-
guage was familiar. The authors concluded that language familiarity acts as a mediator
for attention to enhance perceptual comprehension. This attentional strategy is particu-
larly important when learning to integrate physical speech production (mouth move-
ments) with acoustic information, such as when acquiring a new language (see
Lewkowicz and Hansen‐Tift 2012, for a similar study on infants with congruent results).
Understanding the role of factors that influence language learning and comprehen-
sion, such as familiarity and visual focal points, helps to distinguish the kind of
information most relevant to language learning and comprehension at various stages of
proficiency. Because the behavioural responses between an L2 learner and proficient
speaker differ, consideration for context and degree of experience with a particular lan-
guage should be considered when evaluating how it is being processed.
4. Distinguishing theMemory Models
Although the extant literature presents multiple models to explain language learning
and memory, the underlying assumptions have also been evaluated with scientific
research. At this point, it is still unclear exactly how the mental lexicon is structured and
where information pertaining to known languages is stored and processed. However,
the evidence from cognitive psychology and neuroscience points towards models that
combine both integrated conceptual stores as well as language‐specific feature stores.
Furthermore, how lexical information is engaged within memory is influenced by the
specific properties of the language representations. This section explores findings
related to how languages are processed in terms of their lexical units and the underlying
nature of these representations.

400 Functions and Processes
4.1. Findings Related toLevels ofLanguage Representation
Much of the information discussed thus far has primarily focused on language
processing at the word level. However, in the real world, language is primarily con-
veyed by combining words into phrases or sentences. When assessing sentence
processing, as compared to word processing, it is important to consider models for their
ability to incorporate the enhanced complexity of syntax and grammar, as well as the
increased burden placed on working memory to hold more pieces of information simul-
taneously (Hamrick and Ullman 2016). Data from ERPs have indicated that sentence
processing presented between an individual’s languages can vary based on the manner
of acquisition for the given language (see Morgan‐Short etal. 2012 for a study of adults
learning an artificial language). The L1 is typically learned as a child in immersive
informal environments, such as the home, where sentence structure and phrasings
are spoken by the parents. However, L2 is often acquired through formal schooling
where proper syntax and grammar are explicitly emphasized and taught. Regardless of
the context in which one learns, there tend to be similar behavioural patterns that
demonstrate proficiency. Yet, differing patterns of brain activity between processing
languages learned under formal (i.e. explicit), and informal (i.e. implicit) learning
conditions indicate language processing differences.
Once a language has been acquired, the frequency with which specific words or
phrases are encountered within a language can play a role in linguistic processing or
accessibility. That is, words that are commonly encountered tend to be read faster than
uncommon words. This general effect is referred to as the word frequency effect. This
robust finding has been found in isolated word identification tasks and also word
processing tasks involving words in combination (e.g. Monsell etal. 1989; Segui etal.
1982). This effect has largely been studied in monolingual populations, with frequency
referring to the commonality of a given word’s occurrence within a language. However,
for a fully proficient bilingual, word frequency might depend heavily on the bilingual’s
dominant language. Does the word frequency effect occur within and between the
bilingual’s languages?
To address this question, Cop et al. (2015) analysed the eye movements of English
monolingual and Dutch‐English bilinguals while reading a novel. Monolinguals read a
text in English while the bilingual group read half of the novel in the L1 and half in the
L2. Levels of proficiency were also assessed for each language within the bilingual
group. The results indicated that monolinguals and unbalanced bilinguals (having
unequal levels of proficiency between an L1 and L2) showed similar sized word fre-
quency effects when reading in their respective L1. However, balanced bilinguals
(equally proficient in both languages) showed an even larger effect in both languages.
Cop et al. (2015) also found that bilinguals tended to make more fixations while
reading in their L2 as compared to their L1. The increased rate of fixations resulted in a
reduction of reading speed and overall longer reading times. Surprisingly, the ability to
process high‐frequency words quickly decreased in both L1 and L2 as the level of L1
proficiency increased. However, this effect was not influenced by L2 proficiency. The
authors argued that this finding could be explained in terms of a weaker association
between word form and meaning in an L2 than an L1, which can be affected by

Learning andMemory intheBilingual Mind andBrain 401
proficiency level. That is, the more proficient an individual is in each language, the
larger the size of the mental lexicon for that language. Because the size of the frequency
effect was identical between monolinguals and the unbalanced bilinguals, both of which
were only highly proficient in one language, the authors concluded that the same under-
lying lexical processing mechanisms were operating for both groups. Comparable
results have been found using computer connectionist models that support language
proficiency resulting from an increased lexicon size as effectively reducing frequency
effects for that language (Monaghan etal. 2017). Overall, these results support an inter-
dependent conceptual store in the bilingual lexicon, with size of the mental lexicon as a
main factor for enhancing lexical access and thereby the frequency effect. Furthermore,
individual differences in lexical processing can arise as a factor of multilanguage
processing with respect to language exposure.
4.2. Findings Related totheNature ofLanguage Representations
Although the structure of language can impact its processing across languages, the
inherent properties of the representations play a key role, as well. It has been well‐
established that word type is an important consideration for understanding language
processing in monolinguals to include where attention is directed, processing speed,
and ability to recall (for a recent overview of monolingual versus bilingual emotion
word processing see Wilck & Altarriba, in press). Similar effects have been found in
L2 processing. For example, Sutton etal. (2007) demonstrated that emotion words
(e.g. English: sad, Spanish: triste), as compared to neutral words (e.g. English: table,
Spanish: mesa), capture attention across languages in highly proficient Spanish‐
English bilinguals.
Although word type has been shown to influence processing across languages, and
emotion words increase selective attention over neutral words, it is not necessarily the
case that the degree of emotionality translates well across languages. Altarriba (2003)
collected ratings of concreteness (i.e. tangibility of a word), imageability (i.e. how easy
it is to picture the word), and context availability (i.e. how easy it is to generate exam-
ples of the word) for Spanish concrete (e.g. perro [dog], tijeras [scissors]), abstract (consejo
[advice], verdad [truth]), and emotion (e.g., encantado [delighted], miedo [fear]) words
from Spanish‐English bilinguals. Emotion words produced the lowest ratings of con-
creteness, but the highest ratings of imageability and context availability. However,
when Spanish emotion words were compared to their English translations (e.g. Altarriba
etal. 1999), Spanish words had a higher average context availability rating. In other
words, Spanish emotion words generated more meanings than their English counter-
parts. Altarriba argued that words in one language, especially emotion words, may not
precisely and singularly translate into another language. If a bilingual’s understanding
of concepts in the L1 and L2 differs, then the perceptual representations between an L1
and L2 must also differ.
To investigate the differences in conceptual versus perceptual representations of
information across languages, Altarriba and colleagues employed a variety of experimental
paradigms using bilingual populations. To examine if bilingual language information is
processed according to its higher‐level conceptual information or at its lower‐level (i.e.

402 Functions and Processes
lexical level), a series of eye‐tracking studies was performed (Altarriba et al. 1996).
Fluent Spanish–English bilinguals read mixed‐language sentences while their eye
movements were recorded. When the eye focused on a word in the sentence that had a
language switched, it tended to produce a longer fixation than if the word was in
Spanish (the bilinguals’ L1) and was highly expected based on the contextual cues. For
example, it took longer for participants to process the word dinero in the sentence ‘He
wanted to deposit all of his dinero at the credit union’ than if the translated word was the
expected English equivalent, money, or if there were no contextual cues for the target
word. The authors concluded that sentence context and word expectations were influ-
enced by both the conceptual information provided by the context (prediction ability)
and the lexical information (word in English or Spanish). The increased processing time
required to comprehend sentences that switched from an L2 into the L1, as compared to
a purely monolingual presentation, indicated that presentation mode was a crucial
factor in reading comprehension. Retrieval of semantic information is dependent on the
language in which information is encoded. These findings suggest support for language
processing models with assumptions of separate stores for conceptual and lexical
information.
In another attempt to provide evidence for a semantic or lexical overlap across lan-
guages, Altarriba and Soltano (1996) examined the repetition blindness effect in bilingual
memory. Briefly, the repetition blindness effect refers to an individual’s inability to recall
the presentation of a repeated word, often explored using a rapid serial visual presenta-
tion task, in which words are presented one at a time. Most studies of this effect have
been conducted within a single language, although the few experiments using bilingual
populations have indicated meaningful differences (see Martin and Altarriba 2016 for a
discussion). Altarriba and Soltano were interested in understanding if the conceptual
overlap (word meaning) between equivalent translations would be enough to produce
the repetition blindness effect, or if lexical overlap (similarity of word form) of the
physical features was required. For example, the English word book and the French
word livre both refer to a compilation of words on pages, although the letters that make
up the words are not similar. Spanish‐English bilinguals were presented with sentences
(Experiment 1B) or word lists (Experiment 2) consisting of intermixed Spanish and
English words. On half of the trials, a word was repeated in either the same (e.g. sour–
sour) or the alternative language via its translation (sour‐agrio), and the remaining trials
had no repeated words. Therefore, on trials where the repeated word was in the same
language, there would be repetition of both the semantic and lexical code. However, if
the repeated word was presented in the alternative language, then the lexical code
would differ while the conceptual meaning would remain the same.
The results indicated that a repetition blindness effect occurred when a word was
repeated within the same language, but not if the second presentation was a translation
equivalent. Regardless of whether the words were presented in the context of a sentence
or as seemingly unconnected words in a list, participants were able to recall the occur-
rence of both word presentations with the same conceptual meaning if the lexical fea-
tures were distinct. However, in the absence of grammar and syntactical structure, the
repetition of a non‐cognate word (i.e. translations that share a meaning but not form)
across languages aided in the recall of the repeated word. Altarriba and Soltano (1996)

Learning andMemory intheBilingual Mind andBrain 403
argued that words presented in list form were being processed at a conceptual level, with
semantic overlap facilitating translational recall. When assessing words across languages,
it appears that overlap of physical features plays a vital role in merging the concepts
together as one and producing repetition blindness. In the case of non‐cognate transla-
tions, both items are easily processed as separate entities, thus enabling recallability for
both words even though they share a conceptual meaning. Comparable results indicating
that semantic similarity does not aid in generating useful information when reading to
any greater extent than does orthographic similarity have been found in Spanish‐English
bilinguals in eye movements (see for example, Altarriba et al. 2001). Across multiple
experiments, Altarriba and colleagues concluded that language processing does not occur
solely at the conceptual level and that the physical features of individual words are con-
sidered when encoding and retrieving information from the mental lexicon.
5. Conclusions andSuggestions forFuture Directions
Language learning and memory has been a popular topic across many domains,
including psychology and neuroscience. The research reviewed in this chapter has
integrated the findings of language processing with respect to various models of acqui-
sition, storage, and retrieval. Although there are aspects from each model that can be
supported, the overall evidence has largely supported those with a single, interdepen-
dent language memory store (see French and Jacquet 2004, for further discussion). Data
from various tasks, such as code switching (Heredia and Altarriba 2001) and reading
(Altarriba etal. 1996; Cop etal. 2015), seem to support assumptions of hierarchical level
processing, with the physical features of words being encoded and retrieved separately
from the conceptual meanings.
Bilingual speakers have been identified as a useful resource for studying various
levels of language representation. This population has the unique ability to garner the
same conceptual information from words with orthographically and phonologically
distinct features. The research reviewed throughout this chapter indicates that bilin-
guals can store and utilize two sets of symbols that represent one concept (e.g. Heredia
and Altarriba 2001). Capitalizing on this population’s unique mental representations
can allow for the disambiguation of semantic meaning versus lexical features. In a
monolingual population, these features are inherently overlapping. However, bilin-
guals are able to map multiple lexical labels onto a single conceptual meaning, thus
allowing the possibility to determine the degree of influence on processing from each.
Importantly, these can be studied within the same individual and thus within the same
language processing system. Engaging bilinguals and multilinguals in research can
effectively influence theory on general language processing. The continued use of this
population as a research tool in linguistic research is encouraged to aid in the under-
standing of how language is represented across all individuals and is not limited to
benefiting only those with knowledge of more than one language.
Future investigations of language, memory, and learning should continue to explore
and obtain support for the various underlying assumptions of the proposed models.
Furthermore, the present models should be assessed for reliability and precision with

404 Functions and Processes
multilingual populations. Are all consecutively learned languages stored and repre-
sented in the mind in the same way as an L1 or L2? Future research should be aimed at
providing answers to this and related questions.
REFERENCES
Altarriba, J. (2003). Does cariño equal ‘liking’?
A theoretical approach to conceptual
nonequivalence between languages. The
International Journal of Bilingualism 7 (3):
305–322.
Altarriba, J. and Bauer, L.M. (2004). The
distinctiveness of emotion concepts: a
comparison between emotion, abstract, and
concrete words. The American Journal of
Psychology 117 (3): 389–410.
Altarriba, J., Bauer, L.M., and Benvenuto, C.
(1999). Concreteness, context availability,
and imageability ratings and word
associations for abstract, concrete, and
emotion words. Behavior Research Methods,
Instruments, & Computers 31 (4): 578–602.
Altarriba, J., Kambe, G., Pollatsek, A., and
Rayner, K. (2001). Semantic codes are not
used in integrating information across eye
fixations in reading: evidence from fluent
Spanish‐English bilinguals. Perception &
Psychophysics 63 (5): 875–890.
Altarriba, J., Kroll, J.F., Sholl, A., and Rayner, K.
(1996). The influence of lexical and
conceptual constraints on reading mixed‐
language sentences: evidence from eye
fixations and naming times. Memory &
Cognition 24 (4): 477–492.
Altarriba, J. and Mathis, K.M. (1997).
Conceptual and lexical development in
second language acquisition. Journal of
Memory and Language 36: 550–568.
Altarriba, J. and Soltano, E.G. (1996).
Repetition blindness and bilingual memory:
token individuation for translation
equivalents. Memory & Cognition 24 (6):
700–711.
Baddeley, A. (2003). Working memory: looking
back and looking forward. Nature Reviews
Neuroscience 4 (10): 829–839.
Baddeley, A., Eysenck, M.W., and Anderson,
M.C. (2015). Memory, 2nd. New York, NY:
Psychology Press.
Baddeley, A.D. and Hitch, G. (1974). Working
memory. In: Recent Advances in Learning and
Motivation, vol. 8 (ed. G. Bower), 47–89.
New York, NY: Academic Press.
Barenholtz, E., Mavica, L., and Lewkowicz,
D.J. (2016). Language familiarity modulates
relative attention to the eyes and mouth of a
talker. Cognition 147: 100–105.
Brysbaert, M. and Duyck, W. (2010). Is it time
to leave behind the revised hierarchical
model of bilingual language processing
after fifteen years of service? Bilingualism:
Language and Cognition 13 (3): 359–371.
Brysbaert, M., Verreyt, N., and Duyck, W.
(2010). Models as hypothesis generators and
models as roadmaps. Bilingualism: Language
and Cognition 13 (3): 1–2.
Carrol, G. and Conklin, K. (2015). Eye‐tracking
multi‐word units: some methodological
questions. Journal of Eye Movement Research 7
(5): 1–11.
Carrol, G. and Conklin, K. (2017). Cross
language lexical priming extends to
formulaic units: evidence from eye‐tracking
suggests that this idea ‘has legs’.
Bilingualism: Language and Cognition 20 (2):
299–317.
Chauncey, K., Grainger, J., and Holcomb, P.J.
(2008). Code‐switching effects in bilingual
word recognition: a masked priming study
with event‐related potentials. Brain and
Language 105 (3): 161–174.
Cop, U., Keuleers, E., Drieghe, D., and
Duyck, W. (2015). Frequency effects in
monolingual and bilingual natural reading.
Psychonomic Bulletin & Review 22 (5):
1216–1234.

Learning andMemory intheBilingual Mind andBrain 405
de Groot, A.M.B. (1992). Determinants of word
translation. Journal of Experimental
Psychology: Learning, Memory, and Cognition
18 (5): 1001–1018.
Dijkstra, T. and Van Heuven, W.J. (2002). The
architecture of the bilingual word
recognition system: from identification to
decision. Bilingualism: Language and
Cognition 5 (3): 175–197.
Ervin, S.M. and Osgood, C.E. (1954). Second
language learning and bilingualism. Journal
of Personality and Social Psychology 58:
139–145.
Farley, A.P., Ramonda, K., and Liu, X. (2012).
The concreteness effect and the bilingual
lexicon: the impact of visual stimuli
attachment on meaning recall of abstract L2
words. Language Teaching Research 16 (4):
449–466.
Finkbeiner, M.S., Forster, K., Nicol, J., and
Nakamura, K. (2004). The role of polysemy
in masked semantic and translation
priming. Journal of Memory and Language 51
(1): 1–22.
French, R.M. and Jacquet, M. (2004).
Understanding bilingual memory: models
and data. Trends in Cognitive Sciences 8 (2):
87–93.
Grainger, J., Midgley, K., and Holcomb, P.J.
(2010). Re‐thinking the bilingual interactive‐
activation model from a developmental
perspective (BIA‐d). In: Language Acquisition
across Linguistic and Cognitive Systems (ed.
M. Kail and M. Hickman), 267–283.
Philadelphia, PA: John Benjamins.
Grey, S., Tanner, D., and van Hell, J.G. (2017).
How right is left? Handedness modulates
neural responses during morphosyntactic
processing. Brain Research 1669: 27–43.
Hamrick, P. and Ullman, M.T. (2016). A
neurocognitive perspective on retrieval
interference in L2 sentence processing.
Bilingualism: Language and Cognition 20 (4):
687–688.
Harakchiyska, T. (2011). Overgeneralization as
a strategy for the acquisition of L1 and L2
noun, morphology (the category of
number). Научни Трудове На Русенския
Университет 63 (50): 115–123.
Hartsuiker, R.J. and Bernolet, S. (2017). The
development of shared syntax in second
language learning. Bilingualism: Language
and Cognition 20 (2): 219–234.
Heredia, R.R. and Altarriba, J. (2001). Bilingual
language mixing: why do bilinguals code‐
switch? Current Directions in Psychological
Science 10 (5): 164–168.
Heredia, R.R. and Cieslicka, A.B. (2018).
Bilingual mental models. In: An Introduction
to Bilingualism: Principles and Processes, 2nd
(ed. J. Altarriba and R.R. Heredia), 37–67.
New York, NY: Routledge.
Hernandez, A., Li, P., and MacWhinney, B.
(2005). The emergence of competing
modules in bilingualism. Trends in Cognitive
Sciences 9 (5): 220–225.
Indefrey, P. and Levelt, W.J. (2004). The spatial
and temporal signatures of word production
components. Cognition 92 (1–2): 101–144.
Kolers, P.A. (1963). Interlingual word
associations. Journal of Verbal Learning and
Verbal Behavior 2 (4): 291–300.
Kopeliovich, S. (2006). Reversing language
shift in the immigrant family: A case‐study of
a Russian‐speaking community in Israel
(Doctoral dissertation, Bar‐Ilan
University).
Kroll, J.F. and Stewart, E. (1994). Category
interference in translation and picture
naming: evidence for asymmetric
connections between bilingual memory
representations. Journal of Memory and
Language 33: 149–174.
Kroll, J.F. and Tokowicz, N. (2005). Models of
bilingual representation and processing:
looking back and to the future. In: Handbook
of Bilingualism: Psycholinguistic Approaches
(ed. J.F. Kroll and A.M.B. DeGroot), 531–553.
New York, NY: Oxford University Press.
Kutas, M. and Federmeier, K.D. (2011). Thirty
years and counting: finding meaning in the
N400 component of the event‐related brain
potential (ERP). Annual Review of Psychology
62: 621–647.
Lau, E.F., Phillips, C., and Poeppel, D. (2008).
A cortical network for semantics: (De)
constructing the N400. Nature Reviews
Neuroscience 9: 920–933.

406 Functions and Processes
Lewkowicz, D.J. and Hansen‐Tift, A.M. (2012).
Infants deploy selective attention to the
mouth of a talking face when learning
speech. Proceedings of the National Academy of
Sciences 109 (5): 1431–1436.
Lynch, M.A. (2004). Long‐term potentiation
and memory. Physiological Reviews 84 (1):
87–136.
Martin, J.M. and Altarriba, J. (2016). Rapid
serial visual presentation: bilingual lexical
and attentional processing. In: Methods in
Bilingual Reading Comprehension Research (ed.
R. Heredia, J. Altarriba and A. Cieslicka),
61–98. New York, NY: Springer.
McLaughlin, J., Osterhout, L., and Kim, A.
(2004). Neural correlates of second‐language
word learning: minimal instruction
produces rapid change. Nature Neurosciene 7
(7): 703–704.
Monaghan, P., Chang, Y., Welbourne, S., and
Brysbaert, M. (2017). Exploring the relations
between word frequency, language
exposure, and bilingualism in a
computational model of reading. Journal of
Memory and Language 93: 1–21.
Morgan‐Short, K. (2014). Electrophysiological
approaches to understanding second
language acquisition: a field reaching its
potential. Annual Review of Applied
Linguistics 34: 15–36.
Morgan‐Short, K., Steinhauer, K., Sanz, C., and
Ullman, M.T. (2012). Explicit and implicit
second language training differentially
affect the achievement of native‐like brain
activation patterns. Journal of Cognitive
Neuroscience 24: 933–947.
Morgan‐Short, K. and Tanner, D. (2014).
Event‐related potentials (ERPs). In: Research
Methods in Second Language Psycholinguistics
(ed. J. Jegerski and B. VanPatten), 127–152.
New York, NY: Routledge.
Monsell, S., Doyle, M.C., and Haggard, P.N.
(1989). Effects of frequency on visual word
recognition tasks: where are they? Journal of
Experimental Psychology: General 118: 43–71.
Paivio, A. and Desrochers, A. (1980). A dual‐
coding approach to bilingual memory.
Canadian Journal of Psychology/Revue
Canadienne de Psychologie 34 (4): 388–399.
Palmer, S.D., van Hooff, J.C., and Havelka, J.
(2010). Language representation and
processing in fluent bilinguals:
electrophysiological evidence for
asymmetric mapping in bilingual memory.
Neuropsychologia 48 (5): 1426–1437.
Segalowitz, N. and Freed, B.F. (2004).
Context, contact, and cognition in oral
fluency acquisition: learning Spanish in at
home and study abroad contexts. Studies
in Second Language Acquisition 26 (2):
173–199.
Segui, J., Mehler, J., Frauenfelder, U., and
Morton, J. (1982). The word frequency effect
and lexical access. Neuropsychologia 20 (6):
615–627.
Slabakova, R. (2016). The scalpel model of
third language acquisition. International
Journal of Bilingualism 21 (6): 651–665.
Soares, C. and Grosjean, F. (1984). Bilinguals in
a monolingual and a bilingual speech mode:
the effect on lexical access. Memory &
Cognition 12 (4): 380–386.
Sutton, T.M., Altarriba, J., Gianico, J.L., and
Basnight‐Brown, D.M. (2007). The automatic
access of emotion: emotional Stroop effects
in Spanish–English bilingual speakers.
Cognition and Emotion 21 (5): 1077–1090.
Swaab, T.Y., Ledoux, K., Camblin, C.C., and
Boudewyn, M.A. (2012). Language‐related
ERP components. In: Oxford Handbook of
Event‐Related Potential Components (ed. S.J.
Luck and E.S. Kappenman), 397–440. New
York, NY: Oxford University Press.
Tokowicz, N. and Kroll, J.F. (2007). Number of
meanings and concreteness: consequences
of ambiguity within and across languages.
Language and Cognitive Processes 22 (5):
727–779.
van Heuven, W.J. and Dijkstra, T. (2010).
Language comprehension in the bilingual
brain: fMRI and ERP support for
psycholinguistic models. Brain Research
Reviews 64 (1): 104–122.
Velan, H. and Frost, R. (2007). Cambridge
University versus Hebrew University: the
impact of letter transposition on reading
English and Hebrew. Psychonomic Bulletin &
Review 14 (5): 913–918.

Learning andMemory intheBilingual Mind andBrain 407
Velan, H. and Frost, R. (2011). Words with and
without internal structure: what determines
the nature of orthographic and morphological
processing? Cognition 118 (2): 141–156.
Wang, X. and Forster, K.I. (2010). Masked
translation priming with semantic
categorization: testing the sense model.
Bilingualism: Language and Cognition 13 (3):
327–340.
Wilck, A.M. and Altarriba, J. (in press).
Emotion words in monolingual and
bilingual cognitive psycholinguistic
research. In: Handbook on Language and
Emotion (ed. G.Schiewer, J. Altarriba and
N.B. Chin). deGruyter: Berlin, Germany.
Wong, A.C.N., Qu, Z., McGugin, R.W., and
Gauthier, I. (2011). Interference in character
processing reflects common perceptual
expertise across writing systems. Journal of
Vision 11 (1): 1–13.