Lexical Organization and Reorganization in the Multilingual Mind

Full text

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.
Lexical Organization
andReorganization inthe
Multilingual Mind
GARY LIBBEN AND JOHN W. SCHWIETER
1. Introduction
Anyone who speaks and understands two or more languages will have immediate and
personal experience of the ways in which the lexicon of a language differs from the lex-
icon of a person. Lexicons of separate languages can be compiled as separate books or
databases. Such separation, however, is not possible for the lexicon of a person, which is
an integrated knowledge system in a single mind/brain. How, then, can an integrated
lexical system acquire, maintain, couple, and decouple lexical elements of two or more
languages? This is the fundamental question in the understanding of the multilingual
mental lexicon and the topic of this chapter.
The notion of a mental lexicon may have first emerged in Treisman’s (1961) disserta-
tion as being the storehouse, or listing, of words in the mind. However, over a half a
century later, developments in psycholinguistic research suggest that the mental lexicon
may be better explained from a dynamic and integrated perspective (Libben and Goral
2015). Indeed, for speakers of two or more languages, their language systems are contin-
uously interactively evolving and adapting to the needs of the communicative demands
of the environment. Libben etal. (2017a) argue that ‘adequate modeling of the (bilingual)
mental lexicon requires an understanding not only of dynamicity and integration as
individual constructs but also of the manner in which they interact with each other and
with key phenomena in bilingual lexical processing’ (p. 2). Under these assumptions,
researchers are now entertaining the idea that the mental lexicon of the multilingual
mind may be the default rather than the exception to a monolingual mental lexicon
(Vaid and Meuter 2017).
In this chapter, we discuss the manner in which recent developments in mental
lexicon research have created new implications for the understanding of how multilin-
gualism develops and how it is maintained across the lifespan. We consider multilin-
guals to be persons who are able to understand and speak two or more languages.
14

298 Neural Representations
Under this view, multilingualism includes bilingualism, trilingualism, etc. In Section2,
we discuss the notion of a mental lexicon within the context of its historical roots in
generative linguistics and psycholinguistics. This is followed by an examination of two
dominant metaphors in the literature: one in which the multilingual person is character-
ized as possessing interacting lexical stores and the other in which the multilingual
person is characterized as possessing a network of interacting lexical items in a single
multilingual lexical store. Throughout the chapter, we explore the special role that
structurally-complex words may play in the organization of words in the mind and how
the specific nature of a multilingual’s lexical system is shaped by the linguistic– and
specifically, morphological–properties of the languages.
2. The Notion ofaMental Lexicon
The fundamental challenge in understanding the nature of linguistic knowledge in the
mind and brain is that, whereas the external manifestations of language are open to
observation, the mental representations and operations that enable language activity
are not. This may be an inherent challenge. On the other hand, it may simply be a tem-
porary one that is awaiting resolution through a yet‐to‐be‐discovered research tech-
nology. But, whatever the eventual outcome in the study of language representation and
processing will be, our present situation is one in which language researchers have no
direct observational access to putative linguistic structures in the mind or brain. For this
reason, they are reliant on constructs that generate predictions that can be related to
observable language activity. The mental lexicon is exactly such a construct.
The mental lexicon may constitute the key to our understanding of the nature of lan-
guage organization among multilingual persons. The reason for this is that it is the the-
oretical space in which the fundamental tension inherent in multilingual language
ability is most clear. The inherent tension is simply this: multilingualism seems to
require that words of different languages both be separated and be related. Thus, the
multilingual mental lexicon must be a cognitive system whose organization creates both
connection and isolation. Moreover, it must be able to do so under conditions of constant
change. More than any other aspect of language knowledge, the knowledge of words
seems to be in a state of flux over the lifespan. A language user will be acquiring (and
perhaps losing) words and modifying connections among them over time. This dynamic-
ity, while present for all language users, is heightened in the case of multilingualism.
Current evidence suggests that multilingualism requires both the online integration and
differentiation of language‐specific lexical forms. All other things being equal, multilin-
guals will have a larger lexical store than monolinguals and their lexical systems will
undergo more change across the lifespan. Thus, the study of the multilingual mental
lexicon may offer language researchers the greatest access presently available to an
understanding of the fundamental features of linguistic knowledge inaction.
In recent years, substantial progress has been made in the understanding of multilin-
gual lexical representation and processing through the lens of the evolving characteriza-
tion of the mental lexicon. In the next sections, we discuss these developments by first
tracing the roots of the construct in both the linguistic and psycholinguistic literature.

Lexical Organization andReorganization inthe Multilingual Mind 299
2.1. Roots inGenerative Linguistics
The roots of the notion of a lexicon as part of (and indeed the backbone of) a language
date back to antiquity. In order, however, to present the backdrop to more recent devel-
opments in the conceptualization of the multilingual mental lexicon, it may suffice to
begin by focusing on a particular body of grammatical scholarship that emerged with
the publication of Chomsky’s (1957) Syntactic Structures. This approach, first known
under the term ‘transformational generative grammar’ will serve to highlight important
contrasts in the ways in which lexical knowledge was seen from a purely grammatical
vs. a processing perspective in the second half of the twentieth century. As we will see,
those contrasts have been in some ways transcended as researchers have moved from
early formulations to models that take into consideration the structural complexities of
words both within and across languages.
From the outset, the notion of a lexicon that was distinct theoretically from the syn-
tactic component of a language system was a core feature of generative grammar. It was
explicit in Chomsky’s (1970) paper entitled ‘remarks on nominalization’ and became
known within the framework as lexicalism. In the lexicalist framework, the syntactic
component of a native speaker’s grammar is supplied with words that are part of a sep-
arate component, the lexicon. The manner in which this lexicon is structured and the
manner in which new words are created within it were the subject of a number of key
subsequent publications in the 1970s and 1980s. In the framework of Halle (1973), the
fundamental unit of the lexicon was the morpheme. In the seminal work of Aronoff
(1976), it was claimed that words, rather than morphemes, were the fundamental units
of representation and word formation. This work was followed by major contributions
of William (1981) and Di Sciullo and Williams (1987), which tackled the seemingly
simple, but in fact quite thorny, issue of what constitutes a word in the grammar of a
language and in the grammar of an individual.
A key feature of generative approaches to the lexicon was the view that the lexicon
was a list of exceptions, i.e. those aspects of the grammar that could not be derived by
rule. This seems intuitively appealing if one considers that a grammar is a system of
rules, constraints, and/or regularities, but a simple word, almost by definition, is an
arbitrary association of form and meaning. Thus, the lexicon, in the generative frame-
work was essentially a listing of idiosyncratic forms. This view was expressed perhaps
most cleverly by Di Sciullo and Williams (1987): ‘The lexicon is like a prison–it contains
only the lawless, and the only thing its inmates have in common is lawlessness’ (p. 3).
It should be noted that the simple (and likely false) dichotomy between rules and idio-
syncratic words was recognized and addressed by several influential approaches to theoret-
ical morphology beginning with the work of Jackendoff (1975), who captured the relationship
among complex words sharing elements through ‘redundancy rules’ and Booj’s (2010)
Construction Morphology. Anderson’s A‐Morphous Morphology (Anderson 1992), contains a
nuanced treatment of simple, affixed, and compound words. The core of his approach is the
rejection of the view that morphologically complex words can be characterized as contain-
ing morphemes. Also noteworthy in Anderson’s landmark book is the extremely poignant
epigraph with which it begins: ‘Linguistics will become a science when linguists begin
standing on one another’s shoulders instead of on one another ’s toes’ (p. viii).

300 Neural Representations
2.2. Roots inPsycholinguistics
The term ‘mental lexicon’, in its psycholinguistic sense, gained common usage in the
psycholinguistic literature at roughly the same time that the notions of the lexicon and
lexical operations were being elaborated in the generative literature. The psycholin-
guistic perspective was, however, quite different. The source of that difference can be
tied to methodological matters. From the 1970s onward, the lexical processing literature
was dominated by a single methodological paradigm–visual lexical decision. In its
most basic form, the visual lexical decision task is one in which a language user is
presented with a letter string on the screen of a computer and is asked to decide whether
that letter string is or is not a word of his/her language by pressing a key labelled ‘yes’
or a key labelled ‘no’ as quickly as possible. Psycholinguistic experiments using this
paradigm typically measured a participant’s response accuracy (% ‘yes’ responses for
existing words; % ‘no’ responses for nonwords) as well as response latency in millisec-
onds. The lexical decision paradigm proved itself to be sensitive to differences among
word types in both a robust and reliable manner. With it, lexical processing researchers
were able to document processing speed differences associated with a variety of lexical
characteristics, including frequency, length, concreteness, and part of speech. The data
yielded by these experiments were used to develop and test models of the functional
architecture of the mental lexicon.
It is important to note that the conceptualization of the mental lexicon that was devel-
oped and tested with this methodological approach was one that was dominated by
considerations of lexical access. This is in stark contrast to the lexicon in the generative
linguistic framework, which had as its focus word formation and structure. In the theo-
retical linguistic literature, little attention was paid to matters of speed or ease of access.
However, in the psycholinguistic literature, these were exactly the considerations that
were used to create models of the mental lexicon. In the very influential work of Forster
(1976), for example, differences in lexical decision latencies to higher frequency words
vs. lower frequency words were used to claim that the mental lexicon could be seen as
a frequency ordered list. It was claimed that during lexical access, this list was searched
from the top downward, so that lower frequency words (the ones lower down on the
list) took longer to process because to get at them, the mental search process had to go
past the lexical listings that were higher in the list.
The metaphor of the mental lexicon as a list received challenge from a number of
competing approaches in which frequency effects, for example, were captured by rather
different sorts of metaphors. In Morton’s (1969) logogen model, differences in recogni-
tion speed between high‐frequency and low‐frequency words were captured by the
postulation that words essentially were in a random‐access store (modelled after the
approach of Selfridge 1959) but that each lexical representation in that store had an
activation threshold that was dynamically adjusted on the basis of experience. Thus,
higher frequency words, because they had been activated more often, came to be
activated more easily (i.e. had reduced activation thresholds). This approach to the
functional architecture of the mental lexicon was incorporated into models of word
reading such as that proposed by Coltheart etal. (1977). Their approach, in turn, was
challenged by connectionist approaches to the organization of the mental lexicon, which

Lexical Organization andReorganization inthe Multilingual Mind 301
employed distributed representations for words and, thus, claimed to be better able to
capture the spread of association from one recognized word to others in a lexical network
(e.g. Seidenberg and McClelland 1989).
3. Structurally Complex Words astheMeeting Ground
In many ways, our discussion of the linguistic and psycholinguistic roots of research on
the human lexicon might appear to be a narrative of two solitudes. Historically, one
approach was concerned with how lexical structures meet grammatical structures and
how words are formed. The other approach was concerned with how words are accessed.
Indeed, as long as the fundamental units considered in lexical modelling are structur-
ally simple ones, that narrative could very well have remained one of two solitudes. But
it has not, and the reason for this is that a consideration of structurally complex words
has created the meeting ground between the two traditions. This was a development
that was necessary to meet the basic requirements of observational adequacy in the
study of lexical representation and processing. Although it might seem that ‘normal’
words are structurally simple ones (e.g. cat, dog), the fact is that most words of a lan-
guage, even a language as morphologically simple as English, are prefixed, suffixed, or
compounded. This fact has great consequences for the architecture of the multilingual
lexicon and for the relation of morphology to psycholinguistic models of the mental lex-
icon. The reason for this is that words that contain other words (e.g. handy, handstand,
walker, catwalk) have a dual character. On the one hand, they are clearly linked formally
to the simple words hand, stand, walk, and cat from which they were formed. Yet, they
also maintain an idiosyncratic character. Even if a speaker of English already knows the
meanings of these simple words, the meanings and usage of the derived and compound
words must also be learned. It is because of their dual nature that derived and compound
words create the key bridge between the linguistic notion of a lexicon and the psycho-
linguistic notion of a mental lexicon.
In a seminal study of the processing of prefixed words, Taft and Forster (1975) claimed
that words such as refill and uncover are stripped of their prefixes in visual word recog-
nition. This processing claim was associated with the representational claim that such
prefixed forms do not have their own representations in the mental lexicon, but rather
are understood through links between separately stored prefixes (e.g. re‐, un‐) and stems
(e.g. fill, cover). In a paper published the following year, Taft and Forster (1976) made a
similar proposal for the representation and processing of compound words. Compounds
were claimed to be represented in the mental lexicon in a decomposed, morpheme‐
based form and to be decomposed morphologically during online visual processing.
In the four decades since the initial work of Taft and Forster (1975, 1976), a very rich
psycholinguistic literature has developed on how derived and compound words are
represented and processed. Alternatives to full decomposition models such as that pro-
posed by Taft and Forster included so‐called horse‐race models (e.g. Bertram etal. 2000;
Schreuder and Baayen 1995), in which visual lexical processing is characterized as a
competition between whole word access and morphological decomposition, as well as
models such as that proposed by Giraudo and Grainger (2001), in which it is claimed

302 Neural Representations
that access to morphological constituents is achieved through prior activation of the
whole word. Several more recent publications have questioned the assumption that the
notion of a morpheme, as a construct inherited from the theoretical linguistic literature,
is appropriate to the characterization of mental lexical representations (e.g. Baayen 2014;
Libben 2017).
In addition to the overarching questions concerning the representation and processing
of derived and compound words as entire classes, many psycholinguistic studies have
investigated whether subclasses of derived and compound words may be represented
and processed differently. One prominent example of this is the treatment of so‐called
semantically-opaque words (Bell and Schäfer 2016). These words often have been
considered an exotic subclass. However, as can be seen in the examples above, many
(perhaps the majority) of derived and compound words can be considered to be seman-
tically opaque. It would be almost impossible, for example, to deduce the meaning of
handy in English from the meanings of the stem hand and the suffix ‐y. Similarly, it would
be difficult to deduce that in the dominant meanings of catwalk (‘a narrow bridge’ or ‘a
platform used by fashion models’), it is people who are walking, not cats. More extreme
examples include compound words such as cakewalk, for which the etymological relation
to cake has been lost in its current meaning and use.
It turns out, that matters of semantic opacity have profound consequences for models
of the mental lexicon as an integrated knowledge store. They also have important impli-
cations for multilingualism. If it is the case that lexical processing involves the activation
of the lexical elements of all words, including those that are semantically opaque (Marelli
etal. 2017), then the constituent cat will be activated in the processing of catwalk and cake
will be activated during the processing of cakewalk. This will result in a routine state of
lexical interference (Smolka and Libben 2017). Under this view, a core feature of the
cognitive system that we call the mental lexicon is the capacity to manage such interfer-
ence. Libben (2014) has claimed that in a lexical system that has evolved to maximize the
opportunity for meaning creation, interference is a routine feature of lexical activity. In
addition, if it is the case that the multilingual lexicon also shows massive interconnectivity,
then the processing of cakewalk may not only activate cake in English, but will do so in
other languages as well.
A masked visual priming study with Chinese‐English bilinguals, reported by Zhang
etal. (2011) offers an example of this interconnectivity. Their study was built upon the
fact that Chinese has many compounds that are semantically opaque. For example, the
Chinese word for thing, is an exocentric compound composed of the elements east and
west. In their study, Chinese‐English bilinguals participated in a lexical decision task
with masked priming. A prime word was shown on the computer screen for 59 millisec-
onds immediately preceding the target stimulus for which a lexical decision was
required. The key masked primes were constituents of Chinese compounds that were
translation equivalents of the target English word. Thus, the English word thing was
preceded by the English target east (a constituent of east–west, the Chinese word for
thing). The authors found significant facilitation for first constituents such as these. They
also found the reverse effect, namely that for bilingual participants, the English word
thing primed the English target east. The authors conclude from this that participants
rapidly and automatically translated English words into Chinese and then decomposed

Lexical Organization andReorganization inthe Multilingual Mind 303
the translated compounds into their constituents. Irrespective of whether this actually
corresponds to the sequence of cognitive events, the data demonstrate how between‐
language associations that involve subword elements play a role in online lexical
processing.
The effects reported by Zhang etal. (2011) are made more dramatic by the fact that,
visually, Chinese and English words are fully distinct. Therefore, the associations must
be at a form‐independent, lexical level.
In a study that also investigated the processing of compounds among bilinguals
whose languages have very different writing systems (English and Hebrew), Libben
etal. (2017b) found that subword priming (e.g. presentation of the prime gold followed
by the target goldfish) were evidenced for both languages of Hebrew‐English bilinguals.
Moreover, these priming effects were comparable for both within‐language and between‐
language priming. Thus seeing gold facilitated the processing of goldfish. But so did see-
ing the Hebrew translation of gold (i.e. Similarly, seeing English constituents
facilitated processing of compounds in Hebrew, the native language of all participants.
The findings of Libben etal. (2017b) reinforce the conclusion that connections in the
multilingual lexicon do not simply involve whole words. In addition, the fact that sub-
word primes in the second language (L2) facilitated compound recognition in the native
language points to how acquiring lexical knowledge in a new language may force a
reorganization of the lexical system as a whole.
As a first example of potential dynamic multilingual reorganization of lexical struc-
tures, consider the English word turkey. It is represented in French as the word dinde. In
Hebrew, it is represented as the compound (literally ‘rooster Indian’). Through
the explicit compound representation, it is likely that the native speaker of French will
be able to reanalyse the opaque form dinde (which was originally coq d’inde in French) as
the French phrase of India (d’Inde), with just the apostrophe dropped.
As a second example, consider the word dandelion in English, which came into English
from the French dent de lion (literally: ‘tooth of lion’). In many languages, the word for
dandelion is a compound composed of the constituents tooth and lion (e.g. Spanish,
Italian, Portuguese, German, Norwegian, Hebrew). One might imagine that the interac-
tion among these lexical representations would enable the multilingual speaking
English and one or more of these languages to have a structured representation for the
English word dandelion (though recognition that it is the French dent‐de‐lion) that a
monolingual speaker of English would be very unlikely to have.
And, finally, consider the compound system of a multilingual as a whole: in English,
all compound words have their modifiers first and their main element last (so that a
horserace must be a type of race and a racehorse must be a type of horse). In many languages,
it is exactly the opposite. And, in languages such as French, Italian, and Farsi, both
orders are attested. The consequence of this, if indeed the multilingual mental lexicon is
characterized by massive lexical connectivity, is that there is a drive towards the
development of a unified system in which compound ordering is less fixed overall, so
that racehorse and horserace might require some ambiguity resolution for the trilingual in
the online lexical processing of English.
The discussion above highlights the ways in which morphological phenomena within
and across languages create a need for dynamic reorganization within the multilingual

304 Neural Representations
lexicon. The reason for this is that, as has been shown across many psycholinguistic
investigations, the mental lexicon is characterized by great connectivity among both
word and subword units. In Section4, we examine the consequences of this connectivity
for modelling the multilingual lexicon.
4. Modelling Multilingual Connections andthe
Functional Architecture oftheMental Lexicon
Research on the bilingual and multilingual lexicon is best described as coming out of
the psycholinguistic tradition described in the Section2.2. Its primary focus has been on
issues of lexical access and the implications of lexical processing data for hypotheses
concerning the functional architecture of the mental lexicon. But, as we will see below,
by examining whether knowing two or more languages involves having two or more
lexicons, the study of the multilingual lexicon has linked lexical processing research to
the most fundamental questions regarding the relation between linguistic and conceptual
structures.
Two enduring questions in research on the bilingual lexicon have concerned whether
bilinguals’ languages are integrated and whether lexical access is a selective or non‐
selective procedure (see recent reviews by Kroll 2017; Kroll and Ma 2017; Kroll etal.
2013). Overwhelming support for the observation that bilinguals activate both their
languages in parallel has come from a wealth of studies looking at both word recogni-
tion (Dijkstra 2005; Dijkstra and Van Heuven 2002; Shook and Marian 2013; Van Heuven
etal. 1998) and word production (Costa etal. 1999; De Groot and Starreveld 2015; Kroll
and Gollan 2014; Poarch and van Hell 2012; Starreveld etal. 2014).
Perhaps short of some neurological intervention or in extreme cases of language attri-
tion (Schmid 2010), it is highly improbable that one of the bilingual’s languages can be
‘turned off’. Consequently, the presence of activated words in both languages requires
lexical access to be a competitive procedure which takes into account activated alterna-
tives (in both languages) and resolves such competition at higher levels. These non‐
selective, dynamic accounts to lexical access can be seen in models of bilingual word
recognition (Dijkstra and Van Heuven 2002; Shook and Marian 2013) and production
(Costa 2005; Costa etal. 1999; Schwieter and Sunderman 2008) in addition to models of
the bilingual memory (Kroll and Stewart 1994; Pavlenko 2009).
One explanation of the bilingual memory which demonstrates the consequences of
persistent activation and competition between languages is the revised hierarchical
model (RHM; Kroll and Stewart 1994). This developmental account identifies L2 profi-
ciency as a modulating factor of lexical processing. The model argues that at beginning
stages of adult L2 acquisition, learners rely on existing links between first language
(L1) words and their meanings to access a shared conceptual system. As L2 proficiency
increases, so does the ability to directly access meaning without the need to rely on L1
lexical links. While much work has supported the RHM (Sholl etal. 1995; Schwieter
and Sunderman 2009; Sunderman and Kroll 2006; Talamas etal. 1999), there are some
inconsistent findings to still be clarified (see Brysbaert and Duyck 2010; Duyck and
Brysbaert 2004).

Lexical Organization andReorganization inthe Multilingual Mind 305
Researchers have also shown that conceptual restructuring occurs when new words are
learned to accommodate for word‐to‐concept mapping (Athanasopoulos 2009, 2011;
Athanasopoulos and Kasai 2008; Cook etal. 2006). Building on this and the hypotheses in
the RHM, Pavlenko (2009) put forth a multimodal representation of the conceptual system
which consists of revised categories she argued are fully shared, partially overlapping, or
entirely language‐specific. Pavlenko’s modified hierarchical model (MHM) presents a
dynamic account of conceptual and lexical processing that incorporates conceptual and
semantic transfer (Jarvis and Pavlenko 2008). The MHM, like the RHM, maintains the devel-
opmental progression from lexical to conceptual mediation as L2 proficiency increases.
Keeping with the notion that concepts can restructure to become fully shared,
partially shared, or language‐specific during L2 acquisition, Benati and Schwieter
(2017) speculated as to how the MHM could be extended to learning a third language.
In their proposed trilingual modified hierarchical model (TMHM; see Figure1), the
L1
words
L1 transfer
L2 transfer
L3 transfer
Conceptual
links
Conceptual
links
Conceptual
links
Conceptual development and
restructuring
L2
words
L1-specific
categories
L3-specific
categories
L2-specific
categories
L1-L2
L1-L2-L3 Shared Categories
L2-L3 L1-L3
L3
words
Lexical links
Lexical links
Lexical links
Figure 1 The trilingual modified hierarchical model. (Source: Taken from Benati and
Schwieter 2017).

306 Neural Representations
idea of conceptual restructuring and target‐like development of linguistic categories
couple with the argument that both L2 and L3 word learning can trigger conceptual
reorganizing. However, empirical tests of the TMHM are needed as noted by the
authors:
The addition of an L3 implies that lexical mediation from the L3 to the L2 can occur
(although empirical support would be needed to tease apart whether the preferred
path of lexical mediation for the L3 would be via L1 or L2 words and whether
these things are modulated by other factors). For instance, when an English (L1),
Spanish (L2), Italian (L3) language learner is asked to name an L3 word (e.g., gatto),
he/she may have to access its meaning by first associating it with the L2 word
(e.g., gato) rather than with the L1 word cat. L3‐to‐L2 word association may be
sensitive to factors such as cognate status, language typologies, lexical robustness
(i.e., an element of proficiency in which automaticity of word retrieval is due to the
familiarity with and frequency of its access; Schwieter and Sunderman 2008, 2009),
and overall proficiency level. Future studies should test the modulating nature of
these variables while also keeping in mind that lexical transfer may also trigger
syntactic transfer (Benati and Schwieter, p. 267).
The predictions of the TMHM align with many of the findings from L3 studies and
models which have attempted to address the complex issue of syntactic transfer (Alonso
and Rothman 2016; Bardel and Falk 2007, 2012; de Bot and Jaensch 2015; Falk and Bardel
2011; Hammarberg 2009, 2010; Jaensch 2013; Mykhaylyk etal. 2015; Rothman 2011, 2013,
2015; Slabakova 2016; William and Hammarberg 1998).
In addition to syntactic transfer and the lexical‐conceptual restructuring posited in
the TMHM, it is important to consider the effect of parallel language activation on how
bilinguals process their languages. At the lexical level, several studies suggest that lan-
guages converge (Ameel et al. 2005; Benati and Schwieter 2017; Malt et al. 2015).
Consequently, bilinguals must adapt to these dynamic changes as they occur and be
able to apply them in situations in which they are not only merited but also contextually
appropriate (Green and Abutalebi 2013). Under this assumption, bilingual experience
alone could shape the way bilinguals process their languages and adapt to such
processes over time. Some of the most recent research that is being hotly debated is the
possibility that these adaptive changes are exemplary at showcasing bilingualism as a
model of neural plasticity (Baum and Titone 2014; Kroll 2017; Li etal. 2014). For in-
stance, fresh characterizations view the bilingual as a ‘mental juggler who negotiates
and resolves cross‐language competition with the consequence that he or she gains
juggling expertise that spills over into the domain of general cognition functions’ (Kroll
2017, pp. 38–39). Recent research has shown some support for this metaphor, with a few
caveats that identify critical and modulating factors such as how supportive their speech
community is of the two languages (Lauchlan etal. 2013), language immersion (Baus
etal. 2013; Linck etal. 2009) and whether they often code switch in daily lives (Green
and Wei 2014; Schwieter and Ferreira 2016).

Lexical Organization andReorganization inthe Multilingual Mind 307
The research summarized above has demonstrated that understanding of the
functional architecture of the human lexical processing system can be significantly
deepened through a consideration of bilingual and multilingual processing. Moreover,
the TMHM–while currently based on empirical work that informed its bilingual coun-
terpart, the MHM, demonstrates the manner in which the cognitive dynamics of
bilingual processing can generalized to trilingual (and, in principle, n‐lingual) cognitive
architectures. If indeed, as we have noted, it is highly improbable that one of the multi-
lingual’s languages can be ‘turned off’, this suggests a system of extraordinary com-
plexity and dynamic balance. Although cases of simultaneous early bilingualism are
common, the acquisition of subsequent languages typically develops over an individu-
al’s life history. The consequence of this, as hypothesized in the TMH, is that the lexical
system of the multilingual may always be in a dynamic state of readjustment, in which
lexical items may be both recoupled, and decoupled.
5. Concluding Remarks
In this chapter, we have discussed the mental lexicon as the cognitive system that con-
stitutes the capacity for conscious and unconscious lexical activity (Jarema and Libben
2007). We traced the roots of the notion of a lexicon in generative linguistics as well as
its roots in the psycholinguistics of visual lexical processing. We noted how models of
the bilingual lexicon and multilingual lexicon create an opportunity to understand more
deeply how lexical and conceptual knowledge may be intertwined in the mind and how
a consideration of structurally complex words may require that we see the mental lex-
icon as a self‐organizing system whose chief property may be plasticity. This perspec-
tive underlines the fundamental difference between the lexicon of a language and the
lexicon of a person. In essence, both are models. But they are models of very different
things. The lexicon of a language is a model of the external lexical manifestations of the
language activity of a speech community at a specific time or over time. The lexicon of
a person is a model of the cognitive system that both creates and is created by their
language activity.
REFERENCES
Ameel, E., Storms, G., Malt, B., and Sloman, S.
(2005). How bilinguals solve the naming
problem. Journal of Memory and Language
53:60–80.
Anderson, S. (1992). A‐Morphous Morphology.
Cambridge, UK: Cambridge University
Press.
Aronoff, M. (1976). Word Formation in
GenerativeGrammar. Cambridge, MA: MIT
Press.
Athanasopoulos, P. (2009). Cognitive
representation of color in bilinguals: the
case of Greek blues. Bilingualism: Language
and Cognition 12 (1): 83–95.

308 Neural Representations
Athanasopoulos, P. (2011). Cognitive
restructuring in bilingualism. In: Thinking
and Speaking in Two Languages (ed. A.
Pavlenko), 29–65. Bristol, UK: Multilingual
Matters.
Athanasopoulos, P. and Kasai, C. (2008).
Language and thought in bilinguals: the
case of grammatical number and nonverbal
classification preferences. Applied
PsychoLinguistics 29 (1): 105–121.
Baayen, R. (2014). Experimental and
psycholinguistic approaches to studying
derivation. In: The Handbook of Derivational
Morphology (ed. R. Lieber and P. Stekauer),
95–117. Oxford, UK: Oxford University
Press.
Bardel, C. and Falk, Y. (2007). The role of
the second language in third language
acquisition: The case of Germanic
syntax. Second Language Research 23 (4):
459–484.
Bardel, C. and Falk, Y. (2012). The L2 status
factor and the declarative/procedural
distinction. In: Third language acquisition in
adulthood (pp. 61–78). Philadelphia, PA/
Amsterdam (ed. J. Cabrelli Amaro, S. Flynn
and J. Rothman). The Netherlands: John
Benjamins.
Baum, S. and Titone, D. (2014). Moving
toward a neuroplasticity view of
bilingualism, executive control and aging.
Applied PsychoLinguistics 5: 857–894.
Baus, C., Costa, A., and Carreiras, M. (2013).
On the effects of second language
immersion on first language production.
Acta Psychologica 142 (3): 402–409.
Bell, M. and Schäfer, M. (2016). Modelling
semantic transparency. Morphology
26:157–199.
Benati, A. and Schwieter, J.W. (2017). Input
processing and processing instruction:
pedagogical and cognitive considerations
for L3 acquisition. In: L3 Syntactic Transfer:
Models, New Developments, and Implications
(ed. T. Angelovska and A. Hahn), 253–275.
Amsterdam, The Netherlands: Benjamins.
Bertram, R., Schreuder, R., and Baayen, R.
(2000). The balance of storage and
computation in morphological processing:
the role of word formation type, affixal
homonymy, and productivity. Journal of
Experimental Psychology: Learning, Memory,
and Cognition 26: 419–511.
Booj, G. (2010). Construction Morphology.
Oxford, UK: Oxford University Press.
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: 359–371.
Chomsky, N. (1957). Syntactic Structures.
TheHague, The Netherlands: de Gruyter.
Chomsky, N. (1970). Remarks on
nominalization. In: Readings in English
Transformational Grammar (ed. R. Jacobs and P.
Rosenbaum), 184–221. Waltham, MA: Ginn.
Coltheart, M., Davelaar, E., Jonasson, J., and
Besner, D. (1977). Access to the internal
lexicon. In: Attention and Performance
(ed.S.Dornic), 535–556. London, UK:
Academic Press.
Cook, V., Bassetti, B., Kasai, C. etal. (2006).
Dobilinguals have different concepts? The
case of shape and material in Japanese L2
users of English. International Journal of
Bilingualism 10 (2): 137–152.
Costa, A. (2005). Lexical access in bilingual
production. In: The Handbook of Bilingualism:
Psycholinguistic Approaches (ed. J. Kroll and
A. de Groot), 308–325. New York, NY:
Oxford University Press.
Costa, A., Miozzo, M., and Caramazza, A.
(1999). Lexical selection in bilinguals: do
words in the bilingual’s two lexicons
compete for selection? Journal of Memory and
Language 41: 365–397.
De Bot, K. and Jaensch, C. (2015). What is
special about L3 processing? Bilingualism:
Language and Cognition 18 (2): 15.
De Groot, A. and Starreveld, P. (2015). Parallel
language activation in bilinguals’ word
production and its modulating factors: a
review and computer simulations. In: The
Cambridge Handbook of Bilingual Processing
(ed. J.W. Schwieter), 389–415. Cambridge,
UK: Cambridge University Press.

Lexical Organization andReorganization inthe Multilingual Mind 309
Di Sciullo, A.‐M. and Williams, E. (1987). On
the Definition of Word. Cambridge MA: MIT
Press.
Dijkstra, T. (2005). Bilingual word
recognition and lexical access. In: The
Handbook of Bilingualism: Psycholinguistic
Approaches (ed. J. Kroll and A. de Groot),
179–201. New York, NY: Oxford University
Press.
Dijkstra, T. and Van Heuven, W. (2002).
Thearchitecture of the bilingual word
recognition system: from identification to
decision. Bilingualism: Language and
Cognition 5: 175–197.
Duyck, W. and Brysbaert, M. (2004). Forward
and backward number translation requires
conceptual mediation in both balanced and
unbalanced bilinguals. Journal of
Experimental Psychology: Human Perception
and Performance 30 (5): 889–906.
Falk, Y. and Bardel, C. (2011). Object pronouns
in German L3 syntax: Evidence for the L2
status factor. Second Language Research 27 (1):
59–82.
Forster, K. (1976). Accessing the mental
lexicon. In: New Approaches to Language
Mechanisms (ed. R. Wales and E. Walker),
257–287. Amsterdam, The Netherlands:
North‐Holland.
Giraudo, H. and Grainger, J. (2001). Priming
complex words: evidence for supralexical
representation of morphology. Psychonomic
Bulletin and Review 8 (1): 127–131.
Gonzalez Alonso, J. and Rothman, J. (2017).
Coming of age in L3 initial stages transfer
models: deriving developmental
predictions and looking towards the future.
International Journal of Bilingualism 21 (6):
683–697.
Green, D. and Abutalebi, J. (2013). Language
control in bilinguals: the adaptive control
hypothesis. Journal of Cognitive Psychology 25
(5): 515–530.
Green, D. and Wei, L. (2014). A control process
model of code‐switching. Language,
Cognition and Neuroscience 29: 499–511.
Halle, M. (1973). Prolegomena to a theory of
word formation. Linguistic Inquiry 4: 3–16.
Hammarberg, B. (2009). Processes in Third
Language Acquisition. Edinburgh, Scotland:
University of Edinburgh Press.
Hammarberg, B. (2010). The languages of the
multilingual: Some conceptual and
terminological issues. International Review of
Applied Linguistics in Language Teaching 48
(2–3): 91–104.
Jaensch, C. (2013). Third language acquisition:
Where are we now? Linguistic Approaches to
Bilingualism 3 (1): 73–93.
Jackendoff, R. (1975). Morphological and
semantic regularities in the lexicon.
Language 51 (3): 639–671.
Jarema, G. and Libben, G. (2007). Matters of
definitions. In: Core Perspectives on the Mental
Lexicon (ed. G. Jarema and G. Libben), 1–12.
Oxford, UK: Elsevier.
Jarvis, S. and Pavlenko, A. (eds.) (2008).
Crosslinguistic Influence in Language and
Cognition. New York, NY: Routledge.
Kroll, J. (2017). The bilingual lexicon: a
window into language dynamics and
cognition. In: Bilingualism: A Framework
forUnderstanding the Mental Lexicon (ed.
M.Libben, M. Goral and G. Libben), 27–48.
Amsterdam, The Netherlands/Philadelphia,
PA: Benjamins.
Kroll, J. and Gollan, T. (2014). Speech planning
in two languages: what bilinguals tell us
about language production. In: The Oxford
Handbook of Language Production (ed.
V.Ferreira, M. Goldrick and M. Miozzo),
165–181. Oxford, UK: Oxford University
Press.
Kroll, J., Gullifer, J., and Rossi, E. (2013). The
multilingual lexicon: the cognitive and
neural basis of lexical comprehension and
production in two languages. Annual Review
of Applied Linguistics 33: 102–127.
Kroll, J. and Ma, F. (2017). The bilingual
lexicon. In: The Handbook of
Psycholinguistics (ed. H. Cairns and
E.Fernandez), 294–319. Malden, MA/
Oxford, UK: Wiley‐Blackwell.
Kroll, J. and Stewart, E. (1994). Category
interference in translation and picture
naming: evidence for asymmetric

310 Neural Representations
connections between bilingual memory
representations. Journal of Memory and
Language 33: 149–174.
Lauchlan, F., Parisi, M., and Fadda, R. (2013).
Bilingualism in Sardinia and Scotland:
exploring the cognitive benefits of speaking
a ‘minority’ language. International Journal of
Bilingualism 17: 43–56.
Li, P., Legault, J., and Litcofsky, K. (2014).
Neuroplasticity as a function of second
language learning: anatomical changes in
the human brain. Cortex 58: 301–324.
Libben, G. (2014). The nature of compounds: a
psychocentric perspective. Cognitive
Neuropsychology 31: 8–25.
Libben, G. (2017). The quantum metaphor and
the organization of words in the mind.
Cultural Cognitive Science 1: 49–55.
Libben, G. and Goral, M. (2015). How
bilingualism shapes the mental lexicon. In:
The Cambridge Handbook of Bilingual
Processing (ed. J.W. Schwieter), 631–644.
Cambridge, UK: Cambridge University
Press.
Libben, M., Goral, M., and Libben, G. (2017a).
The dynamic lexicon: complex words in
bilingual minds. In: Bilingualism: A
Framework for Understanding the Mental
Lexicon (ed. M. Libben, M. Goral and G.
Libben), 1–7. Amsterdam, The Netherlands/
Philadelphia, PA: Benjamins.
Libben, G., Goral, M., and Baayen, R. (2017b).
Dynamicity and compound processing in
bilinguals. In: Bilingualism: A Framework
forUnderstanding the Mental Lexicon
(ed.M.Libben, M. Goral and G. Libben),
199–218. Amsterdam, The Netherlands/
Philadelphia, PA: Benjamins.
Linck, J., Kroll, J., and Sunderman, G. (2009).
Losing access to the native language while
immersed in a second language: evidence
for the role of inhibition in second
language learning. Psychological Science 20:
1507–1515.
Malt, B., Li, P., Pavlenko, A. etal. (2015).
Bidirectional lexical interaction in late
immersed Mandarin‐English bilinguals.
Journal of Memory and Language 82:
86–104.
Marelli, M., Gagné, C., and Spalding, T. (2017).
Compounding as abstract operation in
semantic space: investigating relational
effects through a large‐scale, data‐driven
computational model. Cognition 166:
207–224.
Morton, J. (1969). Interaction of information in
word recognition. Psychological Review 76:
165–178.
Mykhaylyk, R., Mitrofanova, N., Rodina, Y.,
and Westergaard, M. (2015). The linguistic
proximity model: The case of verb‐second
revisited. In: Proceedings of the 39th Annual
Boston University Conference on Language
Development (ed. E. Grillo and K. Jepson),
337–349. Somerville, MA: Cascadilla.
Pavlenko, A. (2009). Conceptual representation
in the bilingual lexicon and second
language vocabulary learning. In: The
Bilingual Mental Lexicon: Interdisciplinary
Approaches (ed. A. Pavlenko), 125–160.
Bristol, UK: Multilingual Matters.
Poarch, G. and van Hell, J. (2012). Cross‐
language activation in children’s speech
production: evidence from second language
learners, bilinguals, and trilinguals. Journal
of Experimental Child Psychology 111:
419–438.
Rothman, J. (2011). L3 syntactic transfer
selectivity and typological determinacy: the
typological primacy model. Second Language
Research 27 (1): 107–217.
Rothman, J. (2013). Cognitive economy,
non‐redundancy and typological primacy in
L3 acquisition: evidence from initial stages
of L3 romance. In: Romance Languages and
Linguistic Theory, vol. 2011 (ed. S. Baauw, F.
Dirjkoningen and M. Pinto), 217–248.
Philadelphia, PA/Amsterdam, The
Netherlands: John Benjamins.
Rothman, J. (2015). Linguistic and
cognitive motivations for the
typological primacy model (TPM) of
third language (L3) transfer: Timing of
acquisition and proficiency considered.

Lexical Organization andReorganization inthe Multilingual Mind 311
Bilingualism: Language and Cognition 18
(2): 179–190.
Schmid, M. (2010). Languages at play: the
relevance of L1 attrition to the study of
bilingualism. Bilingualism: Language and
Cognition 13: 1–7.
Schreuder, R. and Baayen, H. (1995). Modeling
morphological processing. In: Morphological
Aspects of Language Processing (ed. L.
Feldman), 131–154. Hillsdale, NJ: Erlbaum.
Schwieter, J.W. and Ferreira, A. (2016).
Effects of cognitive control, lexical
robustness, and frequency of
codeswitching on language switching. In:
Cognitive Control and Consequences of
Multilingualism (ed. J.W. Schwieter),
193–216. Amsterdam, The Netherlands/
Philadelphia, PA: Benjamins.
Schwieter, J.W. and Sunderman, G. (2008).
Language switching in bilingual speech
production: in search of the language‐
specific selection mechanism. The Mental
Lexicon 3 (2): 214–238.
Schwieter, J.W. and Sunderman, G. (2009).
Concept selection and developmental effects
in bilingual speech production. Language
Learning 59 (4): 897–927.
Seidenberg, M. and McClelland, J. (1989). A
distributed, developmental model of visual
word recognition. Psychological Review 96:
523–568.
Selfridge, O. (1959). Pandemonium: a
paradigm for learning. In: Proceedings of the
Symposium on Mechanisation of Thought
Processes (ed. D. Blake and A. Uttley),
511–529. Richmond, UK: H.M. Stationery
Office.
Sholl, A., Sankaranarayanan, A., and Kroll, J.
(1995). Transfer between picture naming
and translation: a test of asymmetries in
bilingual memory. Psychological Science 6 (1):
45–49.
Shook, A. and Marian, V. (2013). The
bilingual language interaction network
for comprehension of speech.
Bilingualism: Language and Cognition 16:
304–324.
Slabakova, R. (2016). The scalpel model of
third language acquisition. International
Journal of Bilingualism 21 (6): 651–665.
Smolka, E. and Libben, G. (2017). Semantic
transparency and compounding.
Language, Cognition and Neuroscience 32
(4): 514–531.
Starreveld, P., de Groot, A., Rossmark, B., and
van Hell, J. (2014). Parallel language
activation during word processing in
bilinguals: evidence from word production
in sentence context. Bilingualism: Language
and Cognition 17: 258–276.
Sunderman, G. and Kroll, J. (2006). First
language activation during second language
lexical processing: an investigation of lexical
form meaning and grammatical class.
Studies in Second Language Acquisition 28:
387–422.
Taft, M. and Forster, K. (1975). Lexical storage
and the retrieval of prefixed words. Journal
of Verbal Learning and Verbal Behavior 14:
630–647.
Taft, M. and Forster, K. (1976). Lexical storage
and retrieval of polymorphemic and
polysyllabic words. Journal of Verbal Learning
and Verbal Behavior 15: 607–620.
Talamas, A., Kroll, J., and Dufour, R. (1999).
From form to meaning: stages in the
acquisition of second‐language vocabulary.
Bilingualism: Language and Cognition 2:
45–58.
Treisman, A. (1961). Attention and speech.
Unpublished doctoral dissertation. Oxford,
England: University of Oxford.
Vaid, J. and Meuter, R. (2017). Languages
without borders: reframing the study of
the bilingual mental lexicon. In:
Bilingualism: A Framework for
Understanding the Mental Lexicon (ed.
M.Libben, M. Goral and G. Libben), 8–39.
Amsterdam, The Netherlands/
Philadelphia, PA: Benjamins.
Van Heuven, W., Dijkstra, T., and Grainger, J.
(1998). Orthographic neighborhood effects
in bilingual word recognition. Journal of
Memory and Language 39: 458–483.

312 Neural Representations
William, E. (1981). On the notions ‘lexically
related’ and ‘head of a word’. Linguistic
Inquiry 12: 245–274.
William, S. and Hammarberg, B. (1998).
Language switches in L3 production:
implications for a polyglot speaking model.
Applied Linguistics 19: 295–333.
Zhang, T., Van Heuven, W., and Conklin, K.
(2011). Fast automatic translation and
morphological decomposition in Chinese‐
English bilinguals. Psychological Science
22(10): 1237–1242.