Other Topics

Summary and key points

Structural aspects of sign language
IAPTER 17
Ie Bilingual Brain
* le-,tet[Talizedorgani2ation°fsignt
• Sign languages are complex encoding systems with similar ,n many respects to that o'f snllan9ua9
multiple Levels of organization; hence they have the • Neur°P^ologkalst^^^
same communicative potential as spoken languages.
^often cause Product,o^^an<
As a group, they are characterized by the following
"°nfluentandagramn.atic;whH:Zrne_
structurat properties:
^^reh""on;Incontrast;p^3d
• Phonology: Signs are specified along the parameters
impair comprehension and lead tonLT
of handshape, location, and movement. that is fluent ^;^^;^°|
• Morphology: The internal complexity of signs can
be increased through rute-governed processes of
. Sb;e.uM?:;^X^:^
• ^°naL;euroima9in9st^e;S^J|
compounding, derivation, and inflertion.
normal production of signs engages7haeTn^! far-flung network of relatives, spouses actual and potential,
• Syntax: Signs can be combined in principled ways
to generate an unlimited number of muttifaceted
^wlM:/aso^l
I^ontra?.thenormatperce^'^s
h)duction
ceremonial age-mates and allies, which makes you someone
who counts in the greater world.
messages. irthern Australia there is a dusty road that stretches
the posterior superior temporal 9yrus/suyt'c;sc" (Evans, 2010, pp. 8)
• Nonmanuals: Some signs are articulated not by the
it 200 kilometers from the coastal town of Wityi southward
' HT^Tteral'zed "gan'zation of si9" ^"guage aj jgh tropical wetlands, eucalyptus savannah, and sandstone
arms/hands, but rather by the torso, head, lips,
tongue, cheeks, eyebrows, and eyelids. ?iffe;lTOWays from thato;s^en^2^1 )ps to the inland town of Jabiru. In recounting a trip Although such an extraordinary level of societal

• Untike spoken word production, manuat^T'1 ^g this route, the linguist Nick Evans notes that "in a few multilingualism is not unique to Arnhem Land, it is quite rare,
• Iconicity: Although the majority of signs have
production activates the superior parieta7cort on the road we passed through the territories of nine being confined to just a few places where many Languages
arbitrary form-meaning relationships, many are
Perhaps as a reflection of proprioceptivefeedb3 lis and seven languages from four language families, at least are spoken in close proximity, like Nagaland m northeastern
iconic insofar as the form of the expression
the arms/hands. jifferent from each other as Germanic, Slavic, Indo-Aryan India and the Mandara Mountains of Cameroon. On the other
physically resembles, to some extent, the entity or
• Furthermore, sign language relies on the Romance" (Table 17.1; Evans, 2010, p. 6). In this region, hand, more modest bilingualism is widespread throughout the
event that it designates.
^pramarginal gyrus more than spoken language.1 kh is called Arnhem Land, there are intimate ties between world (Grosjean & Li, 2013). In fact, approximately half the

^region may play a key role in integratin'gThj Iguage and country. For instance, people must directly people on earth frequently use not only their native tongue
Left-hemisphere dominance
different phonobgicat components of signs^e'.' Iress certain natural resources, such as springs, with the (henceforth LI) but also a second language (henceforth L2).
• The left hemisphere is dominant for sign language, just their specifications for handshape, location, red name in order to use them, and when people cross clan Bilingualism is not an aLL-or-nothing affair, however, but
[ike it is for spoken language: ^undaries they usually switch languages too, so as to show is instead a complex capacity that is manifested to different
and movement-during both production and
• Support for this asymmetry comes from Wada comprehension. |e local spirits that they are familiar and have the right to degrees and shaped by numerous factors. Accordingly, some
testing, neuropsychotogical studies with brain- there. Not surprisingly, almost everyone is multilingual, researchers have posited several types of bilingualism that
damaged patients, and functional neuroimaging Right-hemisphere contributions (id not just a little: vary along parameters such as proficiency, age of acquisition,
studies with healthy subjects. and manner of acquisition. When the focus is on proficiency,
• Despite the fact that the left hemisphere 15 dominant
• The ability to use sign language dissodates not Normal members of Arnhem Land society are highly a distinction is sometimes made between "balanced"
only from the ability to perform visuospatial for sign language, the right hemisphere is also involved
multilingual, often speaking half a dozen languages by the bilinguals—i.e., those who have mastered both Languages
cognitive tasks, but also from the ability • Although the data are mixed, the right hemispherel equally well—and "dominant" bilinguals—i.e., those who
time they are adults. This is helped by the fact that you
to produce symbolic gestures, particularly seems to respond more strongly and extensively to •
have to marry outside your clan, which likely means your are more fluent in LI than L2. When the focus is on age of
pantomimes. the linguistic properties of signed than spoken/ acquisition, a distinction is sometimes made between "early"
wife or husband speaks a different language from you.
written sentences.
• These findings suggest that the neural It also means that your parents each speak a different bilinguals—i.e., those who Learned L2 before the age of about
implementation of sign language is driven mainly • The right superior parietaL lobute, in particular,
language, and your grandparents three or four languages seven—and "Late" bilinguals—i.e., those who Learned L2 after
by the lexicaL and grammatical aspects of the contributes to the processing of classifier
between them. The late Charlie Wardaga, my Ilgar teacher, that age. And when the focus is on manner of acquisition, a
signals that are transmitted, as opposed to their constructions, perhaps by capturing the gradient
was typical. Knowledge ofllgar, Manangkardi, Marrku, distinction is sometimes made between "implicit" biLinguals—
purely physical features. spatial relationships that are specified by these
Iwaidja, and Kunwinjku came to him from his grandparents i.e., those who learned L2 through immersion in the relevant
constructions.
and parents. Although he lived mostly on lands where communicative environment—and "explicit" bilinguals—i.e.,

•nfirar, Marrku, and Iwaidja were the locally appropriate those who learned L2 through didactic instruction, typically in
school. These factors, among others, interact to yield a large
Recommended video languages, he married a Kunwinjku-speaking woman from
0 mainland dan and would regularly speak Kunwinjku with multidimensional space of bilingual types. It is clear, however,
/)er and her relatives.... In this system your clan language that some combinations of traits are more common than others.
Karen Emmorey: The^sigmng brain: What sign languages tell us about human I
"your title deed, establishing your claims to your own For instance, the most proficient bilinguals usually acquired
https://www.youtube.com/watch?v=9fp9kGd7TCO
countiy, your spiritual safety and luck in the hunt there. L2 early and implicitly, whereas the least proficient ones usually
Meanwhile the knowledge of other languages gives you the acquired L2 late and explicitly.
 Other Topics

CHAPTER 17
Summary and key points
Structural aspects of sign language

• Sign languages are complex encoding systems with

• The left-lateralized organization of sign language is
similar in many respects to that of spoken language:
• Neuropsychological studies indicate that anterior
The Bilingual Brain
multiple Levels of organization; hence they have the
lesions often cause production to become
same communicative potential as spoken languages.
nonftuent and agrammatic, while mostly sparing
As a group, they are characterized by the following
comprehension. In contrast, posterior tesions often
structural properties:
impair comprehension and lead to production
* Phonology: Signs are specified along the parameters
that is fluent and syntactically weLL-formed but
of handshape, location, and movement.
susceptible to subLexicat phonotogicat errors.
• Morphology: The internal complexity of signs can
• Functional neuroimaging studies indicate that the
be increased through ruLe-governed processes of
normal production of signs engages the inferior Introduction
compounding, derivation, and inflection. far-flung network of relatives, spouses actual and potential.
frontal gyrus and the middle/mferior temporal cortex.
• Syntax: Signs can be combined in principled ways ceremonial age-mates and allies, which makes you someone
In contrast, the normal perception of signs engages In northern Australia there is a dusty road that stretches who counts in the greater world.
to generate an unlimited number of multifaceted
the posterior superior temporal gyrus/sulcus. about 200 kilometers from the coastal town of Wityi southward
messages. (Evans, 2010, pp. 8)
through tropical wetlands, eucatyptus savannah, and sandstone
• Nonmanuals: Some signs are articulated not by the • The left-lateratized organization of sign language also
outcrops to the inland town ofJabiru. In recounting a trip
arms/hands, but rather by the torso, head, lips, differs in some ways from that of spoken language: Although such an extraordinary level of societal
along this route, the linguist Nick Evans notes that "in a few
tongue, cheeks, eyebrows, and eyelids. • Unlike spoken word production, manual sign muttilinguatism is not unique to Arnhem Land, it is quite rare,
hours on the road we passed through the territories of nine
• Iconidty: Although the majority of signs have production activates the superior parietal cortex, being confined to just a few places where many languages
dans and seven languages from four language families, at least
arbitrary form-meaning relationships, many are perhaps as a reflection of proprioceptive feedback are spoken in close proximity, like Nagatand in northeastern
as different from each other as Germanic, Stavic, Indo-Aryan
iconic insofar as the form of the expression from the arms/hands. India and the Mandara Mountains of Cameroon. On the other
and Romance" (Table 17.1; Evans, 2010, p. 6). In this region,
physically resembles, to some extent, the entity or • Furthermore, sign language relies on the hand, more modest bitingualism is widespread throughout the
which is called Arnhem Land, there are intimate ties between
event that it designates. supramarginal gyrus more than spoken language. world (Grosjean & Li, 2013). In fact, approximately half the
language and country. For instance, people must directly
This region may play a key role in integrating the people on earth frequently use not only their native tongue
address certain natural resources, such as springs, with the
Left-hemisphere dominance different phonologicat components of signs—i.e., (henceforth LI) but also a second language (henceforth L2).
correct name in order to use them, and when people cross clan
their specifications for handshape. Location, Bitinguatism is not an att-or-nothing affair, however, but
• The Left hemisphere is dominant for sign language, just boundaries they usually switch languages too, so as to show
and movement—during both production and Is instead a complex capacity that is manifested to different
like it is for spoken Language: the local spirits that they are familiar and have the right to
comprehension. degrees and shaped by numerous factors. Accordingly, some
• Support for this asymmetry comes from Wada be there. Not surprisingly, almost everyone is muttilingual,
researchers have posited several types of bitingualism that
testing, neuropsychological studies with brain- and not just a little:
Right-hemisphere contributions vary along parameters such as proficiency, age of acquisition,
damaged patients, and functional neuroimaging
and manner of acquisition. When the focus is on proficiency,
studies with healthy subjects. • Despite the fact that the Left hemisphere is dominant Normal members of Arnhem Land society are highly
a distinction is sometimes made between "balanced"
• The ability to use sign language dissociates not for sign language, the right hemisphere is also involved: multilingual, often speaking half a dozen languages by the
bitinguats—i.e., those who have mastered both languages
only from the ability to perform visuospatial • Although the data are mixed, the right hemisphere time they are adults. This is helped by the fact that you
equally weLt—and "dominant" bilinguals—i.e., those who
cognitive tasks, but also from the ability seems to respond more strongly and extensively to have to marry outside your clan, which likely means your
are more fluent m LI than L2. When the focus is on age of
to produce symbolic gestures, particularly the linguistic properties of signed than spoken/ w'fe or husband speaks a different language from you.
acquisition, a distinction is sometimes made between "early"
pantomimes. written sentences. It also means that your parents each speak a different
bilinguats—i.e., those who teamed L2 before the age of about
• These findings suggest that the neural • The right superior parietal lobule, in particular, wage, and your grandparents three or four languages
seven—and "late" bilinguats—i.e., those who teamed L2 after
implementation of sign language is driven mainly contributes to the processing of dassifier veen them. The late Charlie Wardaga, my Ilgar teacher,
that age. And when the focus is on manner of acquisition, a
by the texical and grammatical aspects of the constructions, perhaps by capturing the gradient ww typical. Knowledge ofllgar, Manangkardi, Marrku.
distinction is sometimes made between "implicit" bilinguals—
signals that are transmitted, as opposed to their spatial relationships that are specified by these •°]a, and Kunwinjku came to him from his grandparents
i.e., those who learned L2 through immersion in the relevant
purely physical features. constructions. and parents. Although he lived mostly on lands where
commumcative environment—and "explicit" bilinguats—i.e.
W Marrku, and Iwaidja were the locally appropriate
those who learned L2 through didactic instruction, typically in
7ua9es, he married a Kunwinjku-speaking woman from
school These factors, among others, interact to yield a large
Ae^!L/Trf- cla,n.and WOUM regularly speak Kunwinjku with

I
Recommended video muttidimensionat space of bilingual types. It is clear, however.
k,.a he.rrelatives-- In th's system your clan language
that some combinations of traits are more common than others.
» title deed, establishing your claims to your own
Karen Emmorey: The signing brain: What sign languages tell us about human Language For instance, the most proficient bitmguals usually acquired
^Z'J,wr,sp1ntual 5afetya^ luck in the hunt there.
https://www.youtube.com/watch?v=9fp9kGd7TCO L2 early and implidtly, whereas the least proficient ones usually
hne the knowledge of other languages gives you the
acquired L2 late and explicitly.
Other Topics
The Bilingual Brain

Table 17.1 Clans and languages, many of which are mutually unintelligible, along the 200-kilometer track from Wilyi to Jabiru in findings that are currently available do not yet provide a clear
each language both before and after the onset of injury. This is
Arnhem Land, Northern Australia. Residents of this territory are highly multilingual, often speaking half a dozen languages answer to the question of how the human brain represents and
often a formidable challenge, however.
by the time they are adults processes multiple languages.
It is generally assumed that a patient's pre-injury LI
dan Language Language family Reflecting on this situation, Costa (2019, p. 94) recently
proficiency was quite high, since Ll is the native tongue.
remarked that "the evidence we currently have about how
Murran Iwaidja Iwaidjan; Iwaidjic In contrast, estimating a patient's pre-injury L2 proficiency
bilingualism sculpts the brain is somewhat inconclusive
ManangkaLi Amurdak Iwaidjan; Iwaidjic is sometimes hard, since it depends on a host of variables.
and confusing." Similarly, Green and Krotl (2019, p. 261)
The most reliable judgments can usually be made by
Minaka Manangkardi Iwaidjan; Iwaidjic acknowledged that "we have no theory yet of how the use of
combining informal measures—e.g., self-ratings, interviews,
Born/Kardbam Bminj Gun-wok Gunwmyguan (Central) more than one language changes the brain...." And in the same
and observations from friends and relatives—with formal
Mandjurtngun Bininj Gun-wok Gunwinyguan (Central) vein, Sulpizio et al. (2020b, p. 834) observed that "a coherent
measures—e.g., published questionnaires that probe the
Bunidj Gaagudju Gaagudjuan (Isolate) picture of the brain's capacity to acquire, store, and use more
individual's L2 history, exposure, and use (Paradis & Libben,
than one language efficiently is currently lacking...."
Mandjurlngun Mengerr Mengerrdji Giimbiyu 1987; Kohnert, 2013; Marian et aL, 2007; Li et at., 2014).
Nevertheless, some definite patterns can be discerned in
Manilakarr Urningangk Gnmbiyu Regarding post-injury LI and L2 assessment, the main
the empirical literature. As described below, L2 often recruits
obstacle is that most standardized aphasia tests have been
Bunidj Gun-djeihmi Bininj Gun-wok Gunwinyguan (Central) many of the same brain regions as LI, but the degree of
developed and normed for monolinguat English speakers and
overlap is strongly moduLated by several factors, including
have only been carefully adapted to a few other languages.
In the literature on bilingualism, the subtitle of a classic studies have led to many fascinating and valuable insights (for the three that we have already encountered: proficiency, age
Of course, researchers and clinicians are free to adapt these
1989 paper by Fran^ois Grosjean called "Neurolinguists, beware! reviews see Lorenzen & Murray, 2008; Abutalebi & Detla Rosa, of acquisition, and manner of acquisition. Moreover, even in
test batteries to additional languages whenever they need to,
The bilingual is not two monolinguals in one person" has been 2012; Marini et aL, 2012; Goral, 2012; GoraL & Conner, 2013; the most skilled bilinguals, the extent of neural convergence
but doing so is not only laborious but fraught with problems,
quoted so often that it has effectively become a slogan. But Higby et at., 2013; Li et aL, 20Ua; Wong et al., 2016; Green & between LI and L2 is rarely if ever complete—a discovery
since items that reflect unique linguistic properties or cultural
what does it really mean? In his recent memoir^ journey in Krotl, 2019; Schwieter, 2019; UlLman, 2020; for popular science that is not really all that surprising because, after alt, they're
associations of English cannot easily be converted, and the level
books see Hernandez, 2013, and Costa, 2019). different languages!
languages and cultures, Grosjean (2019, p. 114) explains: of difficulty cannot always be matched.
The chapter is organized as follows. The first section, whkh
More positively, the Bilingual Aphasia Test (BAT) was
The coexistence and constant interaction of two or more is also the longest, deals with a central topic involving the
Evidence from aphasia deliberately designed to enable the systematic evaluation and
languages in biiinguals has produced a different but bilingual brain—namely, the degree of overlap in the neural
comparison of a patient's Ll and L2, taking into account the
complete language system. The analogy I used comes from representation and processing of each language. The next For well over a century, researchers have been investigating how
structural and cultural characteristics of each language (Paradis
the domain of track and field. Hurdlers blend two types of section then considers the cortical and subcortical mechanisms aphasia is manifested in bilinguats. However, opinions differ as
& Libben, 1987; see also Paradis, 2011). The test has three
competencies—that of high jumping and that of sprinting— that collectively allow bilmguals to manage and switch between to how much has been teamed about the neural substrates of
sections: Part A focuses on the patient's language history;
into an integrated whole. When compared individually their two languages. Finally, the Last section discusses some of bilingualism by exploring the various ways in which Ll and L2 Part B assesses the patient's knowledge of many aspects of
with sprinters or high jumpers, hurdlers meet neither level the cognitive consequences of bilingualism, highlighting recent can be impaired by brain damage. On the one hand, Kuzfnina
each language through the modalities of hearing, speaking,
of competence, and yet when taken as a whole, they are evidence that the regular use of two languages, as opposed et at. (2019, p. 2) assert that "reports of individuals with
reading, and writing; and Part C deals with the patient's
athletes in their own right. No expert in track and field would to just one, enhances executive functions and delays the onset bilingual aphasia have emerged as an important constraint on
ability to translate and switch between the two Languages. The
ever compare hurdlers to sprinters or high jumpers, even of dementia. theories of the neurobiotogy of Language." On the other hand,
BAT is available in over 70 Languages, and it includes online
though the former blend certain characteristics of the latter Before proceeding, a brief terminological note is in Del Maschio and Abutalebi (2019, p. 200) claim that "the Instructions for creating versions in other languages as well
two. In many ways, bilinguals are like hurdlers: unique and order: Some of the studies described in this chapter included aphasiological literature has not been successful in defining the
(https://www.mcgitLca/tinguistics/research/bat).
specific communicators. individuals who spoke more than two languages, but for the "eural tocalization of multiple languages, nor in illuminating
To be sure, the BAT is a valuable tool that has been widely
sake of simplicity we will not distinguish between bilinguals the mechanisms underlying their simultaneous management."
used, but caution is still warranted when considering the
Keeping Grosjean's caveat in mind, we are now ready to tackle and multih'nguals unless it is necessary to do so. As described below, although these two positions seem
literature on bilingual aphasia. This is because the relationship
the core question of this chapter: How are two different contradictory, they are both, to some extent, true.
between a patient's pre- and post-injury L2 abilities always
languages handled by the brain of a single person? This question ^e will first consider the unique challenge of evaluating
reflects a complex mixture of factors, including the patient's
has inspired an entire branch of neurolinguistics for several The degree of overlap in the igual aphasia. Then we will discuss the most common
history and use of L2, whether L2 was the dominant language
decades, but researchers have struggled to develop a compelling Patterns of impairment in bilingual stroke patients, together
neural representation and processing W1th the factors that influence them. Next, we will examine
in the patient's environment pre- and post-injury, and of course
answer that applies to all bilinguals. This is not just because the etiology and distribution of the patient's brain damage
of the daunting challenges that neurolinguistic investigations of each language -ne relationships between behavioral data and lesion data in
(Gorat & Conner, 2013; Lerman et aL, 2020).
normally confront, but also because, as mentioned above, •'"gual stroke patients. And finally, we will look at the sorts

bilinguals vary greatly with regard to factors like proficiency, To what extent are a bilingual's different languages subserved 1"ts that have been found in bilingual patients with
Patterns of impairment and the factors
age of acquisition, and manner of acquisition. On the bright by the same macroscopic brain regions? This is a fundamt Pnniary progressive aphasia.
that influence them
side, however, during the past few years an increasing number of issue in the neurolinguistics of bilingualism. To address it we
In the field of bitingualism, Michel Paradis, who is now a
studies have directly explored how these factors, among others, will first survey several forms of evidence from diverse bram challenge of assessment
professor emeritus at McGi'tt University, is famous not only for
influence the neural implementation of multiple languages in mapping methods, and then we will consider some general Ll ^\to determ'ne the status ofa brain-damaged bitinguat's
spearheading the development of the BAT (Paradis & Libben,
both healthy individuals and brain-damaged patients, and these theoretical perspectives. As we will see, however, the various a"d L2, it is necessary to evaluate his or her competence in
1987), but also for being among the first researchers to
Other Topics
The Bilingual Brain

document many different patterns of impairment in bilingual As described in previous chapters, brain damage can cause in at least two ways that are consistent with the assumption
In one of the most sophisticated and informative studies
aphasia (Paradis, 1977, 1989, 1998, 2001): naming difficulties that are significantly worse for certain classes that both languages have largely overlapping neural substrates
to date. Hope et at. (2015) sought to determine whether the
of words than others, such as nouns vs. verbs or proper nouns (Green & Abutalebi, 2008). First, if a stroke directly disrupted
relationships between behavioral data and lesion data that are
• Parallel: Fairly comparable degrees and types of impairment vs. common nouns (see especially the end of the discussion some of those shared neural substrates, L2 might be more
found in monotinguat stroke patients can be generalized to
and recovery for both languages. of lemma retrieval in Chapter 6; see also Kemmerer, 2019c). impaired than Ll because, before the injury, it was less
bilingual stroke patients. To this end, they examined two groups
• Differential: One language is impaired significantly more Category-related deficits like these have been observed mainly ip entrenched in those mechanisms and hence recruited them
of chronic left-hemisphere stroke patients. The monoh-nguat group
than the other. monolingual stroke patients, but they have also been found to more heavily. Second, if instead a stroke affected some of the
consisted of 174 native English speakers, and the bilingual group
• Selective: One language is partially recovered, but the other affect both LI and L2 in a few bilingual stroke patients (ALmagro regions that mediate language control, L2 might again be more
consisted of 33 non-native English speakers who had a wide
is never recovered. et al., 2003; Kambanaros & van Steenbrugge, 2006; Poncelet disrupted than Ll because, before the injury, it was more reliant
range of first languages (Dutch, Farsi, Finnish, French, German,
• Successive: At first one language is partially recovered; then et al., 2007; Faroqi-Shah & Waked, 2010; for similar patterns on effortful control operations to enable sentence formulation.
Gujarati, Hakka, Ibo, Italian, Jamaican Patois, Lithuanian, Lunda.
later, sometimes many months Later, the other is partially in patients with neurodegenerative diseases see Hernandez word retrieval and selection, speech-motor programming, and
Mauritian, Polish, Portuguese, Punjabi, Russian, Serbian, Spanish,
recovered. et at., 2007, 2008; and for reviews see Kambanaros, 2009, and the inhibition of interference from the dominant LI at multiple
Swahili, Tamil, and Yoruba). Of the 33 bitinguals, 24 indicated the
• Antagonistic: Recovery of one language progresses while Miozzo et aL, 2010a). For instance, an aphasic Catalan-Spanish levels of processing. These ideas clearly cannot be tested
age at which they learned English: mean = 10.7; minimum = 2;
recovery of the other regresses. speaker exhibited significantly worse production of nouns than without lesion data, but the first account is supported by a
maximum = 24. Thus, there was a mix of early and late bilinguals.
• Alternating Antagonism: Availability shifts back and forth verbs in both LI and L2, despite still knowing the meanings study described in the next subsection (Hope et al., 2015), and
In addition, 26 of the 33 bitinguals rated their pre-stroke
between one language and the other. of all the target words (ALmagro et aL, 2003). This patient the second one fits with research on the link between bilingual
proficiency in English on a scale from 1 (lowest) to 9 (highest):
• Blending or Mixed: Properties of the two languages are acquired L2 early in life; however, some other patients who aphasia and Language control (Green, 2008).
mean = 8.1, minimum = 5.8; maximum = 9. Thus, the majority of
combined—e.g., the patient speaks one language with the didn't learn L2 until adulthood, yet still achieved conversational Finally, let's return to the topic of category-retated naming
patients had been fairly fluent before their brain injury.
accent of the other, or applies the inflectional affixes of fluency in it, have also been found to display category-related deficits and their implications for lexical organization. We
The experimental procedure was as follows. The researchers
one language to the stems of the other. anomia in both LI and L2 (Kambanaros & van Steenbrugge, noted above, in the context of parallel impairment, that some
first gave each patient the standardized Comprehensive Aphasia
2006). These remarkably fine-grained manifestations of parallel bilingual stroke patients exhibit significantly worse production
Test, which evaluates knowledge of English with 22 tasks (e.g.,
The first two patterns—i.e., parallel and differential—are by far impairment constitute powerful evidence that, at Least regarding of nouns than verbs, or vice versa, in both Ll and L2—patterns
understanding spoken words, repeating spoken sentences, writing
the most frequent, and for this reason we'll concentrate on them fundamental word class distinctions, LI and L2 have common which suggest that the corticat basis of the lexicat distinction
to dictation), yielding a score for each task (Swinburn et aL,
in the following discussion. neural substrates (for detailed theoretical discussion see Miozzo between nouns and verbs is somehow shared by the two
2004). The researchers also obtained a high-resolution MRI scan
To begin with parallel impairment, it is especially common et at., 2010; and for qualifications reflecting cross-h'nguistic languages. Here, in the context of differential impairment, it's
of each patient's brain and measured the amount of damage in
among patients who acquired L2 in early childhood and differences in word classes see Kemmerer, 2014). also worth noting that other bilingual stroke patients have
each of 199 predetermined left-hemisphere regions that included
continued to use it on a regular basis up to the time of brain Shifting now to differential impairment, the most been found to have tanguage-specific dissociations between
both gray and white matter. Then, using multivariate statistical
injury. For instance, Fabbro (2001) employed the BAT to comprehensive study to date is the meta-analysis by Kuzmina word classes. For instance, Dai et at. (2012) reported an aphasic
analyses, they identified the 22 "feature sets" that were most
evaluate 20 Friulian-Itatian patients who learned L2 between et al.(2019)that, as mentioned above, included 119 bilingual Cantonese-Mandarin speaker who, when given a naming task,
capable of predicting all of the monolingual patients' task
the ages of five and seven and spoke both LI and L2 routinely aphasic patients (see also Lerman et aL, 2020, for a series of manifested a noun-verb dissodation in L2 (Mandarin) but not
performances on the basis of all of their Lesion data, together
before suffering a stroke between 1 and 96 months prior to the interesting cases). We already saw that those patients who in Ll (Cantonese). Furthermore, Ansaldo et at. (2010) reported
with data regarding their total lesion volume, time post-stroke,
assessment. He found that 13 (65%) of the 20 patients were learned L2 before the age of seven had, on average, equivalent an aphasic Spanish-English speaker who displayed a double
age, and sex. In other words, for each of the 22 tasks in the
comparabty impaired in both languages. In addition, although deficits in both languages—i.e., parallel impairment. What's dissociation that was both category-related and language-related,
aphasia assessment, and for just the monotingual group, the
one of the other seven patients performed worse in LI than L2, notable here, however, is that the patients who learned L2 after such that he was able to produce nouns better than verbs in LI
researchers calculated which combinations of features, based on
the p-value was only 0.07, which indicates that the difference that age had, on average, significantiy worse deficits in L2 than (Spanish) but verbs better than nouns in L2 (English).
both tesion data and demographic data, could best distinguish
only approached statistical significance. More recently, Kuzmina Ll—i.e., differential impairment, specifically of the language These intriguing findings do not necessarily threaten the
between higher and tower scores. Finally, to achieve their main
et al. (2019) conducted a meta-analysis of 119 published cases acquired late. There were also some relatively small effects of v'ew that, at some level of organization, the neural substrates
goal, they investigated how well the feature sets derived from
of stroke-induced bilingual aphasia. When they focused on the proficiency. In the entire sample only four patients, all of whom °f the noun-verb distinction are shared by a bilingualperson's
the monolingual group could be applied to the bilingual group.
entire sample, they found an overall tendency for L2 to be more were early bih'nguals, reported higher pre-injury competence in two languages. They do, however, suggest that the relevant
They did this by comparing the predicted scores with the actual
impaired than LI; however, when they focused on the subgroup L2 than LI (due to attrition of the Latter), and they continued to Processing system must simultaneously be configured in such a
scores for the tatter group.
that acquired L2 before the age of seven and appeared to have perform better in L2 than LI post-injury. In addition, among the waV as to allow these different word classes to be separabLe by
Several important results emerged, as shown in Figure 17.1,
high pre-injury proficiency in it, they discovered that the most late bilinguals whose speech production was significantly more aln damage in just one language or the other. Clearly, a great
First, the feature sets that were most predictive of the
prevalent pattern was parallel impairment. These studies suggest impaired in LZ than LI, this dissodation was greatest for those of further research will be needed to elucidate this complex
monolingual patients' scores turned out to be strongly predictive
that if a bilingual person's two Languages are learned early and patients who also had lower pre-injury competence in L2 than ne"rocognitive architecture. (For a comparison involving spoken
of the bilingual patients' scores too. However, for 13 of the
used frequently, they are likely to be disrupted in similar ways According to Kuzmina et al. (2019), these results challenge written production of nouns and verbs, see Figures 2.2-2.3
22 tasks, the expected scores for the bilingual patients were
by a stroke, which in turn suggests that they are likely to be the view that L2 has, for the most part, the same neural and the associated text in Chapter 2.)
significantly higher than their actual scores. Interestingly,
subserved by largely overlapping neural networks. substrates as LI, but one could argue otherwise. As just
/?e/oh-, further analyses revealed that the bilingual patients' worse
Interestingly, although the studies described above were described, the most common kind of differential imp; 'onships between behavioral data and lesion data
Turn performances on these 13 tasks could not be attributed to
primarily concerned with language processing in general, other namely, worse deficits in L2 than LI—appears to occur most [j^1ng to the tesion correlates of bilingual aphasia, the current
lite different tesion sites than the monolinguat patients. On the
research has shown that bilingual stroke patients sometimes often in bilinguals who learned L2 late and/or lacked L2 arature is unfortunately rather limited. Nonetheless, several
contrary, for the 13 critical tasks, increasing damage to the
display parallel impairment of very selective aspects of both before suffering brain damage. In Light of these points, one es have helped to illuminate the degree of overlap in the
very same brain regions was associated with decreasing scores
LI and L2. A striking example involves lexical organization. could potentially explain the greater impairment of L2 than representation and processing of each language.
in both groups. These deficit-tesion correlations, however.
Other Topics
The Bilingual Brain

Evidence from functional neuroimaging Patterns of activation and the factors

During the past few decades, PET and fMRI have been used
that influence them

with increasing frequency to investigate the extent to which We have already seen that the major patterns of impairment

Ll and L2 have common neural substrates. So far, however, the in bilingual aphasia are, at least in cases of stroke, typically

findings that have come from alt these functional neuroimaging moduLated by an individual's pre-injury competence in L2,

studies are, as SuLpizio et al. (2020b, pp. 834-835) put it, by how early in life an individual learned L2, and by whether

"patchy and largely inconsistent." This is mainly because an individual learned L2 implicitly or explicitly. As one would

the studies have varied in many ways, such as the level(s) of expect, these same factors also influence the normal patterns

language that they focused on, the task(s) that they used, of cortical and subcortical activity in healthy bHinguats when

and, perhaps most importantly, the type(s) of bih'nguals that they process their different Languages. The following discussion

they enrolled as participants. Nevertheless, some order can be stresses the first two factors— i.e., proficiency and age of

detected amid the chaos. acquisition—because their separate and interactive impacts

We will begin by discussing evidence that the three factors have been explored in a variety of functional neuroimaging

described above—namely, proficiency, age of acquisition, and studies of the bilingual processing of natural languages. The

manner of acquisition—strongly influence the similarities and third factor—i.e., manner of acquisition—has received Less

differences between the general activation patterns associated attention in this literature, but its effects on activation patterns

with LI and L2 processing. Then we will turn to a number have been closely monitored in the wetl-controlled context of

of more specific results involving the bilingual processing of model language" learning, as described in Box 17.1.

particular aspects of language, including speech perception and The importance of proficiency has been demonstrated by

production, lexical semantics, and morphotogy and syntax. several reviews and meta-analyses of functional neuroimaging
studies of bitinguatism (Abutatebi et al., 2005; Indefrey,

Figure 17.1 Results from Hope et al:s (2015) comparison of brain-behavior relationships in monolingual and bilingual stroke patients. Multivariate
statistical analyses revealed the left-hemisphere lesion patterns that were most predirtive of the monolinguat patients' performances on
22 tasks assessing knowledge of English. Crucially, the same patterns were strongly predictive of the bilingual patients' performances too. BOX 17.1 The neural substrates of learning "model languages'
The color bar in (A) indicates the number of tasks (out of 22 total) that were significantly associated with damage to the corresponding
brain regions in both groups. For 13 tasks, however, the bilingual patients' actual scores were Lower than their expected scores. These worse It is common for adults to learn foreign languages, but Little
analysis focused on 12 studies that involved "non-Dedarative
than anticipated performances could not be attributed to different areas of damage in the bilingual patients than the monolingual patients. is known about how their brains allow them to do so. This is
Grammar Learning (nonDGL)"—i.e., acquiring the relevant
Instead, they were due to the bilingual patients being even more sensitive than the monoLingual patients to damage in the same set of largely because acquiring a full, natural language after the
grammatical rules in an implicit, ski'tt-like manner.
specific regions. (From Hope et al., 2015.) ©Thomas M. H. Hope, et at. Licensed under a Creative Commons Attribution 4.0 International
age of about 12 is a very complex, time-consuming task that
License. The color bar in (B) indicates the number of tasks (out of 13 total) that were significantly associated with damage to the Uttman's theory predicts that DGL should engage the left
corresponding regions in both groups; however, these associations were greater for the bilingual than the monolingual patients which is to 1s not only extremely hard to accomplish, but also extremely
hippocampus, and perhaps also some adjacent regions in the
say that increasing damage in a given region caused even more severe task impairment for the former patients than the latter. (From Hope hard to study, especially at the neurobiotogicat level. As an
medial temporal [obe (MTL), butTagarelLi etal. (2019) did not
et aL, 2015, pp. 1077 & 1080.) © Thomas M. H. Hope, et aL Licensed under a Creative Commons Attribution 4.0 International license. alternative, it has become increasingly popular to examine
find such activations in their initial, whole-brain meta-analysis
how adults Learn so-called "model languages," which
bilingual patients representing all three variants of PPA—i.e., (Figure 17B1.1). They argue, however, that this failure may
turned out to be significantly greater for the bilingual group
include artificial languages that consist of made-up words
than the monoLingual group, thereby accounting for the former the nonfluent variant (nfvPPA), the semantic variant (svPPA),
and grammatical rules, as well as miniature versions of real
group's lower than expected scores. As the researchers put it, and the logopemc variant (IvPPA) (for a refresher on these
languages that consist of subsets of words and grammatical
these findings "reflect enhanced sensitivity to lesion damage syndromes see Chapter 4). Not surprisingly, which aspects of
rules. From an experimental perspective, such model
in the bilingual group in regions that play a similar role in the language were most compromised in these cases differed as
languages are ideal not only because subjects can team them
monolingual group" (Hope et aL, 2015, p. 1077; see also the a function of which type of PPA they had. However, Malcomb
quickly, but also because researchers can control the amount,
following YouTube video, which contains interviews with the et al. (2019) were still able to identify some similarities.
type, and timing of exposure that subjects receive during the
authors: https://www.youtube.com/watch?v=Qk4ZqcWOVik). For one thing, none of the patients were significantly more
acquisition process (Culbertson, in press).
More generally, this study clearly supports the view that impaired in LI than L2; instead, all of them either had parallel
Many PET and fMRI studies have investigated the neural
monolinguaLs and bih'nguals share essentially the same neural impairment of both languages over the entire course of the
substrates of model language learning in adulthood, and
circuitry for language, with bilinguals tending to recruit parts of disease, or were initially more impaired in L2 than LI and later
Tagarelti et at. (2019) recently conducted two meta-analyses of
it more heavily than monoLinguals. became equally deficient in both. It was unclear, though, which
th)"s literature to test the predictions of ULlman's (2001, 2004,
factors determined whether a given patient first displayed
2016, 2020) Dedarative/Procedurat Model (this framework is
Primary progressive aphasia parallel or differential impairment. Indeed, as MaLcomb et at.
described m the section called "Theoretical perspectives").
So far, our discussion has been restricted to research involving (2019) emphasize, perhaps their most startling discovery was
°ne analysis focused on 10 studies that involved "Dedarative Figure 17B1.1 Activation maps for Dedarative Grammar Learning
bilingual, patients who suffered strokes. As a final topic, then, that the patterns of impairment that the 13 patients manifest
Grammar Learning (DGL)"—i.e., acquiring the relevant (DGL) non-Dedarative Grammar Learning (nonDGL),
let's take a quick look at what has been learned about bilingual were not reliably influenced by those factors that have been and their conjunction (pink). ALE - Activation
Srammatical rules in an explicit, fart-tike manner. The other Likelihood Estimation.
patients with primary progressive aphasia (PPA). In a recent found to affect biLinguaL stroke patients—most notably,

review of this Literature, Malcomb et aL (2019) analyzed 13 proficiency, age of acquisition, and manner of acquisition.
(continued)
 The Bilingual Brain
Other Topics

• Simultaneous bilinguals (/V = 16): Grew up in a bilingual

(continued) family and regularly interacted with those relatives, and

that is predicted by the [theory] to rely particularly with other people, in both languages since birth.
have arisen from technical limitations of their meta-analytic • Covert Simultaneous bilinguals (/V = 8): Grew up in a
on procedural memory showed basal ganglia (anterior
method and from the notoriously poor sensitivity of the monoh'nguat family that used LI, and acquired L2 implicitly
caudate/putamen) activation....
anterior MTL to fMRI. They therefore carried out a more spedfic because it was present in the surrounding environment.
region-of-interest analysis that zoomed in on the left MTL. This • Sequential bih'nguals (/V = 8): Grew up in a monoh'ngual
For additional support from electrophysiology, see Morgan-
revealed that 27% of the studies involving DGL reported MTL family that used LI, and acquired L2 implicitly before the
Short et al.(2012).
activity, especially in the hippocampus, but none of the studies age of six when the family emigrated to another country.
invoMng nonDGL did so-a statistically significant difference • Late bilinguals (A/ = 12): Grew up in a monolinguat family
that supports the theory (Figure 17B1.2). that used LI, and acquired L2 explicitly in school after the
UUman's theory also predicts that nonDGL should engage DGL nonDGL
age of nine.
the Left basal ganglia and inferior frontal cortex, and both p=.03—|

regions were in fact robustly activated (Figure 17B1.1). Interestingly, all 44 subjects had also studied a third Language
Indeed, the left basal ganglia responded exclusively to in school and developed considerable competence in it. The
Figure 17.2 Results from Sebastian et al.'s (2011) meta-analysis of
nonDGL, and even though a small part of the left inferior functional neuroimaging studies of bilingualism. Activation most commonly spoken languages were English, French, and
frontal cortex responded to both nonDGL and DGL, most of 27% 0% likelihood estimates from eight studies involving subjects Swiss German, with the others being Bulgarian, Catalan,
this frontal territory responded only to nonDGL. with high L2 proficiency are shown in (A), and activation Finnish, Greek, Hebrew, Hungarian, Indonesian, Italian,
likelihood estimates from six studies involving subjects
Overall, then, the functional neuroimaging literature Percentage of Studies Showing Activation in MTL Japanese, Portuguese, Russian, Serbo-Croatian, SLovenian,
with tow/moderate L2 profidency are shown in (B).
on model language learning appears to bolster the basic Spanish, and Turkish. During the fMRI sessions, each subject
i)^«^^. 100% Representative sagittal (left), coronal (middle), and axiat
assumptions of ULlman's theory. As TagareLli et aL (2019) (right) slices are shown in both (A) and (B). Red = LI; silently narrated what they did at certain times the previous
blue = L2; and purple = overlap. In (B) two of the regions day (morning, mid-day, and evening), with separate blocks
point out,
Figure 17B1.2 Percentage of studies in DedarativeGrammaI_ uniquely engaged by L2 were the dorsolaterat prefrontal
for LI, L2, and L3. Each subject also performed a baseline
Learning (DGL) and in non-Dedarative Grammar cortex (DLPFC) and anterior cingulate cortex (ACC), both of
only grammar learning that is predicted by the [theory] task that was intended to maintain his or her attention and
Learning (non'DGL) that show medial temporallobe which are associated with effortful controlled processing.
to rely particularly on deciarative memory showed
(MTL) activation.
(Adapted from Sebastian et at., 2011, p. 807.) involved joining together the thumb and forefinger of the right
hippocampai activation..., while only grammar learning
hand in response to a cue.
When the researchers analyzed the imaging data, they

regions, including the dorsolateral prefrontat cortex (DLPFC) and Turning now to age of acquisition, its importance has likewise focused on the regions traditionally referred to as Broca's
2006; Sebastian et al., 2011; Abutalebi & DelLa Rosa, 2012; been demonstrated by several reviews and meta-analyses of
the anterior cingulate cortex (ACC), both of which have been area (defined as BAs 44 and 45) and Wernicke's area (defined
Blackburn, 2019a, 2019b; Cargnelutti et aL, 2019; Del Maschio functional neuroimaging studies of bilingualism (Abutalebi
implicated in effortful cognitive processing (Figure 17.2B). These as posterior BA22). At the level of the whole group, the
& Abutalebi, 2019). For instance, Sebastian et al. (2011) et al., 2005; Stowe & Sabourin, 2005; Indefrey, 2006;
regions probably don't contribute directly to the representation activation patterns elicited by LI, L2, and L3 in these two
conducted a meta-analysis of U studies-eight involving Abutalebi & Detla Rosa, 2012; Liu & Cao, 2016; Berken et al., classic language regions did not differ significantly. However,
of L2; instead, they most likely mediate the greater processing
subjects with high L2 proficiency, and six involving subjects 2017; Blackburn, 2019a, 2019b; Cargnelutti et aL, 2019; Del
demands of L2, spedficaLly by carrying out executive functions at the Level of individual subjects, they varied a lot. For this
with Low/moderate L2 proficiency. Although the two sets of Maschio & Abutalebi, 2019). As with proficiency, this factor
like helping to select the right words in L2 and suppress reason, the researchers decided to compare the four subgroups
studies differed systematically in terms of the subjects' L2 influences the extent to which areas associated with cognitive
intrusions from LI (see the later section on language control in ^ with regard to the amount of variance across LI, L2, and
proficiency, they both included a mix of early and late bilinguals control, such as the DLPFC and ACC, are recruited to facilitate
bilinguals). Support for this interpretation comes from L3 in the spatial extent of activation (i.e., the number of
as well as a wide a range of experimental and baseline tasks. L2 processing. In particular, compared to early bilinguals activated voxels) in each region. The results are portrayed in
and Della Rosa (2012, p. 527):
For each set of studies, Sebastian et al. (2011) concentrated on O'.e., those who learned L2 before the age of about seven), Figure 17.3. These graphs indicate that in both regions the
how the contrasts between experimental and baseline tasks for late ones tend to engage these areas more often and more
Consider the following very simple question: Why should the L2 variability of activation across languages increased with the age
LI related to those for L2. 'ntensety during L2 processing, though the degree of activation of L2 acquisition. The Simultaneous and Covert Simultaneous
learner have her LZ more extensively represented at the brain ^
The results are shown in Figure 17.2. Regarding the eight decreases as proficiency increases (see Grant et aL, 2015, for a
level... ? Consider that a low-profident L2 speaker such as our^ bilinguals displayed the least amount of variability, or, in other
studies that involved high-profidency subjects, it is apparent longitudinal study that reveals such effects in the ACC).
learner may know only, for example, no more than 1,500 i words, the greatest amount of homogeneity; the Sequential
that the activation patterns for LI and L2 were almost completely It's also notable that compared to early bilinguals, late ones
as compared to the 15,000 or so words in her native language- bilinguats displayed more variability; and the Late bilinguats
overlapping and fell within regions of the left hemisphere that e><hibit more variable activation patterns in so-called Broca's
Following any principle of neural efficiency, it would be a ^ displayed the most. According to Bloch et al. (2009, p. 631),
are well-established as subserong basic Linguistic functions a"d Wernicke's areas—i.e., the left posterior inferior frontal these findings suggest that
paradox that these 1,500 words are represented in larger' ?'"•"
(Figure 17.2A). This suggests that once an individual has Syrus (IFG) and left posterior superior temporal gyrus (STG)—
areas. As a consequence, the prefrontal effect [i.e; the;
achieved substantial competence in a second language,rtis when processing their different languages. This was shown in early exposure to two languages results in the formation
recruitment of the DLPFC and ACC] cannot be a question o]
processed primarily by brain regions that also implement the an fMRI study by Bloch et al. (2009) that's worth describing of a language processing system that is able to handle
language-spedfic neuroanatomical representations.^ but
native language, and in a fairly automatic fashion. Shifting to -re' This investigation included 44 subjects who alt had all languages acquired, including late teamed ones...,
necessarifyan issue of differences in processing demc"1d5' '^.
the six studies that involved [ow/moderate-profidency subjects, n1edium to high L2 proficiency but differed with respect to the [whereas] monolingual upbringing results in a language
neural effort to process a weak system, such as the L2 ofa^\,
some core language areas in the Left hemisphere were engaged a9e and context of L2 acquisition. They were therefore divided network which is not able to accommodate late learned
proficiency speaker, is higher than the neural effort toproLK-
again by both LI and L2, suggesting some shared neural nto four subgroups: languages to the same degree....
strong system, such as the L2 of a high-profidency sp
substrates; however, L2 also uniquely engaged several other
 The Bilingual Brain
Other Topics

than after the age of about seven (for further information about mentioned, but also engage other areas associated with
Broca's area
the so-called critical or sensitive period for L2 acquisition, see language control, like the DLPFC and ACC. These control-
(A) 70000 Berken et at., 2017). related regions may be needed to inhibit intrusions from Ll

60000 With these points in mind, let's now look at some of what and facilitate word selection in L2 (Abutalebi & Green, 2007;
has been revealed by functional neuroimaging research about see also the later section on language control in bilinguals).
50000 the extent to which LI and L2 have common vs. distinct neural It has also been found, though, that even when bitinguats
substrates for particular domains of language. First, though, a who have high L2 proficiency generate spoken words, they
g 40000
c3 caveat: Because the studies in each domain differ in numerous still activate core speech production areas more strongly than
'.§ 30000 ways (e.g., stimulus characteristics, task demands, and types monoh'nguals, regardless of whether production is in the native
of participants), they cannot easily be compared. For this or non-native language. This was demonstrated in a frequently
20000
reason, we will limit our discussion to some salient trends and cited fMRI study by Parker Jones et at.(2012) that is worth
10000 influential studies (see also the muttiple meta-analyses by describing in detail.
Sulpizio et aL, 2020b). The experiment included 67 subjects: 36 monotmgual
0 Late
Covert Sequential English speakers, and 31 bilinguals whose LI was German,
Simultaneous
SwndtamoitS (n=8) (n=12) Speech perception and production Italian, Dutch, or Czech, and whose L2 was English. Although
(n = 16)
(n=8) So far only a few functional neuroimaging studies have directly the bih'nguats learned English at very different ages (range =
investigated bilingual speech perception (for a review see 1-15), they all had very high proficiency in English, according
Wernicke's area Golestani, 2016). Overall, these studies suggest that both to multiple types of assessment. In order to distinguish between
(B) 30000 native speech sounds in LI and non-native speech sounds activation patterns associated with different stages of word
in L2 are analyzed and recognized by essentially the same generation, the researchers created eight conditions that
25000-1 corticat areas—most notably, the superior temporal regions involved two kinds of response—overt speech production and
that subserve increasingly complex auditory processing, and button press—for each of the four kinds of "triadic" stimuli
20000-1
8 the temporoparietal and frontal regions that map auditory shown in Figure 17.4:
§ 15000 representations of speech stimuli onto the corresponding motor

i 10000
representations (for a refresher on these regions see Chapter 5).
In addition to these neural commonalities, however, many Meaningful Meaningless
differences have been found between the perception of Ll
5000 and L2 speech sounds regarding the intensity and distribution
Piano 5555
of activation. Unfortunately, these differences are not very Words
Late consistent across studies due to variability in experimental
Oven Harp eeee 5555
Simultaneous Covert Sequential
(n=16) Simultaneous (n = 8) (n=12) procedures; hence, they do not allow us to make confident
(n=8) generalizations. Furthermore, virtually all of the studies in the
existing literature have either used subjects with low/moderate
*
H,«».3 ^^^^m^^^^^;,^^^^^^^^^;r
Pictures
LZ proficiency or trained subjects in L2 over several sessions. As
F"ml" '~"(

4 *
./'
a consequence, almost nothing has been discovered yet about
the corticat underpinnings of bilingual speech perception in

0.05 or ***p-< 0.005. (From Bloch et al., 2009, p. 630.)

people who have achieved a high level of L2 competence as Tasks Naming/readmg Say "1,2,3"
the result of early implicit acquisition or many years of explicit Semantic matching Identity matching

We are left, then, with a question: Which factor has 'nstruction. Hopefully, though, this topic will receive greater
It's worth noting that Bloch et al.'s (2009) study has several attention in the near future. Figure 17.4 Experimental paradigm in Parker Jones et al.'s (2012)
more influence on the cortical representation of a second
limitations. For one thing, because the narrative production task fMRI study of bilingual speech production in subjects
In contrast, considerably more functional neuroimaging
language-how well one can use it, or how early one with high L2 proficiency. On each trial three stimuli were
was "both silent and open-ended, the researchers had no way of research has addressed bilingual speech production (for a review
it? The answer is by no means clear, but there does appear simultaneously shown as a triad, with one on top and
knowing how complex the subjects' covert sentences were in each see Klaus & Schriefers, 2019). The available data suggest that two on bottom. Word retrieval was tapped by naming
to be a general consensus that these two factors affect
language. In addition, the baseline task was not well-matched Ll and L2 production have mostly shared neural substrates, meaningful pictures and reading aloud meaningful words.
dWerent aspects of L2 to different degrees (and as continuous
with the experimental one. Nonetheless, the study does suggest Articulation without word retrieval was tapped by saying
Particularly in several left temporal and frontal regions that are
dimensions; see Sulpizio et al., 2020a). Proficiency is | "1,2,3" in response to meaningless pictures and symbol
that even when a bilingual person's proficiency in L2 is just as 'own to contribute to lexical retrieval, form encoding, and
manifested most strongly in the realm of Lexical semantics.^ sequences. Semantic processing was tapped by pressing a
high as in LI, the degree to which the two languages evoke articulation (for a refresher on these regions see Chapter 6). button to match the meaningful stimulus on top with the
After all, regardless of when one first began to learn L2, one
similar activation patterns in core language regions depends on ^ lr"ficant effects of both proficiency and age of acquisition related one on bottom (e.g., piano with harp rather than
can always keep increasing one's vocabulary. In contrast,agej
the age at which L2 was acquired. And this is certainly a valuable oven). And perceptual processing was tapped by pressing a
acquisition seems to have the greatest impact on artic /s also been documented, however.
button to match the meaningless stimulus on top with the
discovery. (See also Kim et aL, 2007, for a closely related fMRI Regarding the impact of proficiency, when subjects who
and grammatical processes. For instance, it is much eas identical one on bottom. Further details about the specific
study that obtained partly convergent results when the same not yet very skilled at using L2 perform picture naming conditions are provided in the main text. (Adapted from
speak L2 without an accent and correctly use its most :
narrative production task was given to high-profidency bilinguals <s- they not only hyper-activate many of the regions just Parker Jones et at., 2012, p. 893.)
syntactic constructions if one started learning it before
who acquired L2 either together with LI or in adolescence.)
Other Topics The Bilingual Brain

involving two groups of French-Engtish bilinguals who had that a single person spoke both languages ftuentty, how would
• Meaningful pictures:
• Speech production: name each picture—first the top
equally high L2 proficiency but differed in L2 age of acquisition, these hugely different consonant repertoires be implemented
with one group being called "simultaneous" because they learned in his or her brain? (For a pertinent study which suggests that
one. then the bottom Left one, and finally the bottom
both languages from birth, and the other group being called learning a second language increases the size of one's auditory
right one
"sequential" because they teamed L2 after the age of five. The cortex, see Resset et at., 2012.)
• Button press: make a semantic judgment by indicating
experimental condition required the subjects to read aloud simple
which of the two bottom pictures is more closely
sentences in either French or English, whereas the baseline Lexj'cat semantics
related to the top one
condition required them to view strings ofXs as if they were In Chapters 8 and 9 we encountered considerable evidence
• Meaningful words:
sentences. For each group the researchers investigated whether, for the Grounded Cognition Model, which maintains that
• Speech production: read each word aloud—first the
Figure 17.5 Results from Parker Jones et al.'s (2012) fMRI study of after subtracting out the baseline condition, there was greater
top one, then the bottom Left one, and finally the the meanings of object nouns (e.g., dog, apple, spatula)
bilingual speech production in subjects with high L2 brain activity when reading aloud in L2 (i.e., English) than in and action verbs (e.g., kick, grab, bite} tend to be anchored
bottom right one
proficiency. Relative to the button press conditions, the LI (i.e., French). The simultaneous group did not display any
• Button press: make a semantic judgment by indicating m anatomically segregated modality-specific systems for
speech production conditions engaged many Left temporal
significant differences at all; however, as shown in Figure 17.6, perception and action, such that a great deal of semantic
which of the two bottom words is more closely related and frontal regions that have previously been associated
with talking. However, in the five regions shown here, the the sequential group evinced greater activity not only in the left processing involves the partial reconstruction of sensory and
to the top one
magnitude of activity was significantly greater in bilinguals posterior IFG and ventral premotor cortex, both of which have motor states in different parts of the brain, usually in a rapid,
• Meaningless pictures:
than monolinguals during picture naming and reading been linked with phonetic encoding (see Chapter 6 for details),
• Speech production: say "1,2,3" while viewing first the automatic, and unconscious manner, but with sensitivity to
aloud than during semantic judgments for pictures and
but also in the left posterior fusiform gyrus near the Visual task, context, and individual experience. All of the empirical
top picture, then the bottom left one, and finally the words. PTr, pars triangularis of inferior frontal gyrus; POp,
pars opercularis of inferior frontat gyrus; PrC, precentral Word Form Area, which has been linked with orthographic-to- studies that we discussed in those chapters involved
bottom right one
gyrus; STG, superior temporal gyrus; PT, planum temporale. phonological transformations (see Chapter 15 for details). These monolinguat subjects, but several fMRI experiments with
• Button press: make a visual judgment by indicating
(From Parker Jones et al., 2012, p. 898.) Adapted and results indicate that even when a person's general proficiency in
which of the two bottom pictures is identical to the bilingual subjects have shown that the Grounded Cognition
reproduced permission of Oxford University Press. OUP is
L2 is quite high, if that person acquired L2 several years after LI, Model also applies fairly well to words in L2 (for reviews see
top one not responsible or in any way Liable for the accuracy of
the adaptation. The Licensee is solely responsible for the he or she tends to recruit the neural substrates of speech-motor Kuhne & Gianetli, 2017; Monaco et aL,2019).
• Meaningless symbol sequences:
adaptation in this publication. planning more heavily for L2 than LI. Such observations help In one fMRI study, for instance. De Grauwe et al. (2014)
• Speech production: say "1,2,3" while viewing first the
to explain why it is so hard for Late learners of L2 to speak that administered a lexicat decision task to native German speakers
top sequence, then the bottom left one, and finally
some of the same regions that are sensitive to processing demands language without at Least a mitd foreign accent. who achieved high proficiency in Dutch as adults. As indicated in
the bottom right one
during monolingual speech production—an idea that fits well with Lastly, it's worth noting that further fMRI research is needed previous chapters, a lexical decision task simply requires subjects
• Button press: make a visual judgment by indicating
the lesion study by Hope et al. (2015) that we discussed earlier. to explore the degree to which Ll and L2 speech sounds to distinguish between real words and pseudowords, so whatever
which of the two bottom sequences is identical to
In further analyses, Parker Jones et al. (2012) compared the diverge in the brains of bitinguaLs whose two languages have semantic processing is evoked by real words is mostly reflexive
the top one
response properties of the five regions across conditions in order substantially different phoneme inventories. For instance, as rather than deliberate. The researchers found that, compared to
to get a better sense of which stage(s) of speech production mentioned in Box 5.1 in Chapter 5, Rotokas, which is spoken the recognition of semanticalty similar abstract verbs in both
All of the subjects performed the verbal tasks in English, and a
they Likely subserve. In two regions—namely, POp and PTr—the 'n Papua New Guinea, has only six consonants, whereas !X6o, languages (e.g., German zogern 'hesitate' and Dutch aarzelen
subgroup of the bilinguals also performed the verbal tasks in LI,
effect of bilingualism was greater for both naming pictures and which is spoken in Botswana, has 122. In the unlikely event hesitate'), the recognition of semanticatly similar action verbs
but always on a different day.
Relative to the button press conditions, the speech production reading aloud than for saying "1,2,3," which supports the view in both languages (e.g., German werfen 'throw' and Dutch gooien
conditions elicited remarkably similar distributions of activity that these regions contribute more to accessing, selecting, and throw') engaged some of the same precentral motor areas and

across many Left temporal and frontal regions that have preparing word forms than to overtly articulating them (see postcentrat somatosensory areas. These results suggest that for

previously been associated with talking. However, in five different Chapter 6 for details). In the other three regions—namely, PrC, highly competent bih'nguats, the meanings of closely related

regions the magnitude of activity was significantly greater in STG, and PT—group differences were present not only for naming concrete words in LI and L2 have shared neural substrates in

bilinguals than monolinguaLs during picture naming and reading pictures and reading aloud but also for saying "1,2,3," which modality-specific cortical regions. As one would expect, however,
aloud than during semantic judgments for pictures and words. supports the view that these regions underlie articulation perse other work suggests that if L2 proficiency is relatively low, the

These regions are shown in Figure 17.5 and Listed below: (PrC) and the subsequent processing of auditory feedback (STG sensory/motor grounding of L2 words is Likewise relatively weak
and PT; again, see Chapter 6 for details). Parker Jones et al. (Birba et aL, 2020; see also Zhang et aL, 2020b).
(2012, p. 901) state their main conclusion as follows: "In sum' In Chapters 8 and 9 we also saw that the multifarious
• Pars trianguLaris (PTr) of IFG
the effect of bilingualism on regional activations m naming and Fi3"re 17.6 Results from Berken et al:s (2015) fMRI study of bilingual semantic features of concrete words appear to be integrated by
• Pars opercuLaris (POp) of IFG
reading is best explained by greater demands on processes of speech production in subjects that had high L2 proficiency transmodal hubs such as the anterior temporal tobes (ATLs). Once
• Precentral gyrus (PrC)
word retrieval, articulation, and postarticulatory monitoring but differed in L2 age of acquisition. When reading aloud
again, all of that literature was based on data from monolingual
• Superior temporal gyrus (STG) simple sentences in LI and L2, simultaneous bilinguals
are in common with word processing in monolinguaLs. subjects, but there is some evidence that the left ATL also
• Planum temporale (PT) (i.e., those who learned L2 together with LI) did not
What about age of acquisition? Although not many functior display any activation differences; however, as shown here, represents unified lexical concepts in L2. For instance, Crinion
neuroimaging studies have attempted to isolate the potential. sequential bitinguals (i.e., those who teamed L2 after age et at. (2006a) found that in high-proficiency German-English
As Parker 3ones et al. (2012) point out, other fMRI research
five) displayed greater activity for L2 than LI, specifically in
effects of this factor on speech production, there is some and Japanese-Engtish bilinguals, priming effects were elicited
has shown that when monoh'nguals generate words, the signal the left premotor cortex, inferior frontal gyrus, and fusiform
evidence that it does modulate activation patterns in cortK in the left ATL by pairs of semantically related words such as
strength in almost all of these areas increases as the frequency gyrus. These effects were found after subtracting out a
areas associated with phonetic encoding. This was shown, ^ control condition that involved viewing strings ofXXXs as if trout-salmon (compared to pairs ofsemanticalty unrelated words
of the target words decreases (Chee et aL, 2002; Graves et aL,
instance, by Berken et al. (2015), who conducted an fMRI st they were sentences. (From Berken et aL,2015,p.214.) such as trout-horse), regardless of whether the words in each pair
2007). This suggests that bilinguah'sm amplifies the workload for