Linguistic term paper

Ling C140. Fall 2019 G. Silvestri, Oct 15

1

Two tongues in the brain:

neurolinguistic aspects of bilingualism 1. Key concepts of neurolinguistics (cf. Ahlsén 2006)

§ Neurolinguistics studies the relation of language (and communication) to different aspects of brain function

§ it tries to explore how the brain understands and produces language and communication 1.1 Different views on the relation between brain and language § Localism tries to find locations or centers in the brain for different language functions § Associationism situates language functions in the connections between

different areas of the brain (associations of perceptions of different senses with words and/or “concepts”)

§ Dynamic localization of function: functional systems of localized subfunctions perform language functions - dynamic systems

§ Holistic theories: many language functions to be handled by widespread areas of the brain § Evolution-based theories: relationship between how the brain and language evolved in

different species, how they develop in children, and how adults perform language functions 1.2 The central questions of neurolinguistics Some of the central questions for neurolinguistics are: i. What happens to language and communication after brain damage of different types? ii. How did the ability to communicate and the ability to use language develop as the species evolved? iii. How can we relate this development to the evolution of the brain? iv. How do children learn to communicate and use language? How can we relate their acquisition of language to the development of their brains? v. How can we measure and visualize processes in the brain that are involved in language and communication? vi. How can we make good models of language and communication processes that will help us to explain the linguistic phenomena that we study? vii. How can we make computer simulations of language processing, language development, and language loss? viii. How can we design experiments that will allow us to test our models and hypotheses about language processing?

Ling C140. Fall 2019 G. Silvestri, Oct 15

2

1.3 Questions concerning bilingualism in the brain Two central questions: (1) How are the two (or more) languages represented/encoded in the brain? - in the same area or separately? - equally lateralized or unequally lateralized? (2) What factors affect brain organization and processing of multiple, e.g. age of acquisition, language proficiency, similarity of L1 and L2, etc.? 2. Main studies on language in the brain § Localism: different “higher functions” localized in different centers of the brain, mainly the

cortex - these centers are “sisters,” being equally important OR - one center, such as the prefrontal area (in front of the frontal lobes), is superordinate to the others

• Broca

- neurolinguistics is said to have been born in 1861 with P. Broca - a patient’s symptoms and dissection of his brain: patient had major difficulties producing speech (He was called “Tan,” since this was his only speech output) - At the autopsy, the oldest brain damage localized in a specific area in the brain => Broca’s area, the third frontal convolution (gyrus) in the left hemisphere

1. that it was possible to localize psychological functions in brain convolutions 2. that linguistic symptoms were caused by lesions in the left hemisphere and that consequently language was lateralized, which was totally unexpected

Figure 1. Broca’s area

• Broca called the linguistic disorder aphémie = ‘inability to speak’ • Broca’s theory led to new conflicts between unitarists and localists • since then, researchers worked to localize symptoms such as “agraphia,” “word

deafness,” and “word blindness”

Ling C140. Fall 2019 G. Silvestri, Oct 15

3

• Wernicke

- theory based on Broca’s findings - patients with problems understanding language

Figure 2. Wernicke’s area 2.1 Main subsequent neurolinguistic studies 2.1.1 Studies on APHASIA • These studies have told us a lot about the location/representation of different language

abilities in monolinguals - Left hemisphere dominance - Selective impairment of language independent of other cognitive abilities

• Defining aphasia: “language loss due to brain damage”

§ due to an infarction (blockage of blood vessels in the brain), a hemorrhage (bursting of a blood vessel), or a head trauma

§ effects: changed comprehension and/or production of language § studying aphasia helps understand how functional systems related to language work and

which brain structures are involved § data acquired from studies of aphasia have roughly the same degree of reliability as the

data obtained with other neurolinguistic methods § re-defining aphasia: acquired language disorder, often defined as a focal lesion (i.e., a

lesion of one or more specific areas) • Acquired disorders: caused by progressive neurological diseases, such as dementias

(language and memory closely connected) • Aphasiology or linguistic aphasiology is the dominant branch of neurolinguistics

§ assuming aphasia caused by damage to the connection between Wernicke’s and Broca’s areas, Wernicke predicted “conduction aphasia” (most likely caused by a lesion bundle of fibers connecting the two areas)

§ Wernicke’s aphasia:

Ling C140. Fall 2019 G. Silvestri, Oct 15

4

- association fibers connect different cortical areas and create associations between visual, auditory, and tactile representations/conceptions/ ideas of the same object. - each item is born out of associations that are repeated and stabilized, so that for example the visual image alone is enough to evoke the other representations: from “primary images” in one sensory modality to “secondary images,” which combine and unite all the images of an object, and then on to “tertiary images,” which are the foundations of abstract concepts and words

Figure 3. Language areas in the brain 2.1.2 Brain mapping and bilingualism/multilingualism Important (and compatible) hypotheses about brain representation/storage of different languages: A. Dual system hypothesis (Paradis 1981) The two languages are represented in “overlapping” but distinct areas: - whether the pattern of aphasia and recovery is the same or different in 2 languages or not depends on whether damage is to area of overlap or area where only one language is encoded B. Lateralization hypothesis - L2 is less strongly lateralized (more right hemisphere involvement) and that is why it recovers first in some cases

• age of acquisition crucial: late bilinguals are more likely to show increased Right Hemisphere involvement (Vaid 1983).

• (sequential?) bilinguals are more likely to suffer aphasia than monolinguals (Galloway 1980)

o After brain damage, do two languages show the same aphasic syndrome/same recovery? Important for ascertain: - whether the two languages are localized in the same areas of the brain

Ling C140. Fall 2019 G. Silvestri, Oct 15

5

- whether the same LH regions support bilingual’s two languages - if there is more RH involvement in bilinguals - if age of acquisition matters to brain organization of 2 languages, or other factors related to manner of presentation. o Hypotheses about recovery (1st half of 19th century)

- Ribot’s Law (1882): The first learned language should be less impaired and will recover first - Pitres’ (1895) rule: The most familiar or most recently used lg returns first

o Current hypotheses (e.g. Dual system hypothesis; Lateralization hypothesis) about language recovery in aphasics relate it to specific localization of the two languages

Some results: • Obler and Albert (1977): initial recovery of L1 is at chance, while

initial recovery of most recently used language is more than chance - supporting Pitres’ rule and potentially the Lateralization Hypothesis (the later the L2 is acquired, the less LH lateralization, i.e., the more spread out, the less chances of complete loss) • M. Paradis (1987) did a meta-analysis of 138 cases reported in the literature:

- About 50% recovered both languages at the same rate (parallel) - In other cases L1 returned before L2

- In other cases L2 recovered before L1 § Currently, there are about 200 documented cases of recovery of the 3 categories above § Some studies: faster recovery of both languages when they are more related (Sasanuma &

Park 1995) o Problems with aphasia studies:

§ depend on correspondences between area of damage and language abilities – which are both difficult to assess or are often not assessed carefully

§ Assessment of damage is difficult since it is usually diffused (spread out across and through various areas)

§ Status of language loss in different languages is difficult to determine because patients are not examined in health (assessment of loss is based on reports by relatives and doctors)

§ Assessments given in global terms (e.g. more or less fluent) but no attention is given to linguistic details.

2.2 Methods of language mapping in the brain 2.2.1 Cortical stimulation studies

Ling C140. Fall 2019 G. Silvestri, Oct 15

6

- Mild stimulation to specific areas of cortex while patients are undergoing treatment for epilepsy

Figure 4. Cortical stimulation • Patients asked to name pictures of objects in a specified language • Electrical stimulation inhibits activity where it is applied leading to inability to name objects. • Cortical stimulation studies generally support dual system and lateralization

hypotheses (e.g. Rapport & Tan 1983; Lucas II, McKhann II, Ojemann 2004) • But the studies have been criticized: subjects were not neurologically intact

2.2.2 Divided visual field paradigm studies • anatomy of the visual system, visual information is directed to the hemisphere that is opposite to the stimulated half-field (Bryden 1982) • Lateralized stimuli presented to eliminate the participant’s ability to make an eye-movement • Language lateralized in the Left Hemisphere faster/easier for a person to read something that is presented in the right visual field (RVF) (Dornbush & Winnick 1965; Mishkin & Forgays 1952, a.o. For a review, see Hellige, Laeng, & Michimata 2010)

Figure 5. Divided visual field

Ling C140. Fall 2019 G. Silvestri, Oct 15

7

• Soares & Grossjean (1981) - Subjects: adult right-handed males (Portuguese-English bilinguals and English-speaking monolinguals) - Procedure - Word-reading task using the divided visual field paradigm - In Experiment 1, the words from the bilinguals’ two languages were presented in mixed blocks - In Experiment 2, they were presented in separate blocks - Results - similar level of left hemisphere advantage for language in the bilingual and the monolingual groups - No evidence of greater heterogeneity of asymmetry patterns in bilinguals - A significant correlation (r=.61) for lateralization levels of the bilinguals’ two languages => language is processed primarily in the left hemisphere for both (simultaneous?) bilinguals and monolinguals • Willemen & Richardson (1994) - Papua New Guinea university students: reading of words presented to Right and Left visual field both for English (L2) and Tok Pisin (L1) - Response times greater for stimuli presented to Right visual field than to Left visual field (suggesting Left Hemisphere dominance) for people who learned their L2 before age 4 - L2 learners after 9 the pattern was reversed suggesting more Right Hemisphere participation 2.3 Recent studies 2.3.1 Event Related Potentials (ERPs) § ERPs show electrical activity in the brain measured at the scalp § different patterns of electrical activity depending on the kind of linguistic stimulus § crucial finding: syntactic and semantic anomalies elicit qualitatively distinct ERP effects

- Semantic anomalies (e.g., The cat will bake cookies) elicit a negative wave that peaks at about 400ms after the anomalous word appears => N400 effect (Kutas & Hillyard 1980; Osterhout & Nicol 1999); see Figure 6a - Syntactic anomalies (e.g., The cat will eating the food) elicit a large positive wave that onsets at about 600ms after presentation of the anomalous word and persists for at least half a second => P600 effect (Osterhout & Nicol 1999); See Figure 6b

Ling C140. Fall 2019 G. Silvestri, Oct 15

8

Figure 6. ERP responses to semantically (a) and syntactically (b) abnormal sentences. • The great thing about ERPs studies is they are non-invasive and you can test healthy brains! § Weber-Fox and Neville (1999) • Procedure: - monolinguals and bilinguals on their ERP response while reading sentences with semantic and syntactic violations: a) Semantic violation, e.g. The boys heard Joe’s orange about Africa b) Syntactic violations of 3 types: i) Phrase structure violations: Word order errors, e.g. *The boys heard Joe’s about stories Africa. ii) Specificity constraint: Don’t wh-move out of a definite DP, e.g. *What did the boys hear [DP Joe’s stories about __ ]? (cf. What did the boys hear [DP more stories about __ ]? ) iii) Subjacency constraint: Don’t move a DP out of a DP, e.g. *What do the boys believe [DP the story that Joe sleeps with __ ]? (cf. Do the boys believe [DP the story that Joe sleeps with a teddy bear]?) - They were also given a grammaticality judgment task (behavioral) • Subjects: - 61 Chinese-English bilinguals (18-44 years) - Age of Acquisition groups: 1-3 years, 4-6, 7-10, 11-13, >16. - All had lived in the US for at least 5 years • Results (simplified): - The L2 learners had ERP responses to the semantic errors comparable to the native speakers - The ERPs of L2ers to syntactic violations differed as a function of age of acquisition: - Proficiency (as determined by judgment data) and age of acquisition were correlated - Semantic processing seems less vulnerable to age effects than syntactic processing § Summary of other ERP studies

Ling C140. Fall 2019 G. Silvestri, Oct 15

9

- Studies of morphosyntax (e.g. subject verb agreement, plural marking): highly proficient L2ers whose L1 is similar to the L2 can show native-like ERP responses (difference between Russian and Japanese leaners of German; see Rossi et al. 2006; Ojima et al 2005; Hahne et al 2006; Osterhout et al 2006) - ERP responses: affected by the particular syntactic structure being tested - Violations in syntactically obligatory constituents stronger (i.e. more native-like) response

§ Does this mean that balanced bilinguals represent the syntactic rules of their languages in the same area, but L2ers don’t?

2.3.2 Functional Magnetic Resonance Imaging (fMRI) • fMRIs show brain activity by detecting changes associated with blood/oxygen flow • cerebral blood flow and neuronal activation are coupled § undisputed finding: in normal monolingual right-handed individuals, cortical activation

associated with language processing is strongly lateralized to the Left Hemisphere and involves: - Broca’s area and motor projection areas related to speech - left temporal lobe (most of Wernicke’s area) - parietal lobe (some of Wernicke’s area on the LH)

§ bilingual studies: languages of the bilinguals activate different or partially different brain regions (Dehaene et al., 1997; Ding et al., 2003; Kim et al., 2002; Kovelman et al., 2008, Marian et al., 2007, Perani et al., 2003, Tan et al., 2003, Tham et al., 2005)

• The majority finds overlapping brain activation for the two languages (e.g. Briellmann et al., 2004; Frenck-Mestre et al., 2005; Gandour et al.; 2007, Klein et al., 2006; Yokoyama et al., 2006) - this could explain why the two languages sometimes interfere • brain activation in overlapping areas for the bilingual’s two languages does not imply there cannot be a functional/neuronal separation - the two languages could be separated at the functional/neuronal level, but still be overlapping at higher levels of brain organization § How can we use fMRI to investigate whether the bilingual brain has differentiated neural

patterns of activation for different grammars? • Kovelman, Baker & Pettito (2008):

Ling C140. Fall 2019 G. Silvestri, Oct 15

10

- English monolinguals have higher Left Inferior Frontal Cortex (LIFC) (i.e. Broca’s area) activation for the more complex object relative clauses in English (1a) rather than for the less complex subject relative clause sentences in English (1b) (Caplan 2001; Caplan et al. 1998; Stromswold, Caplan, Alpert, & Rauch 1996) 1) a. [The juice that the child spilled _ ] stained the rug “SO” (object relative clause attached to the subject) b. The child spilled [the juice _ that stained the rug] “OS” (subject relative clause attached to the object) - this same pattern in Spanish-English bilinguals when processing in English o Remarkably, this was not observed with the Spanish-English bilinguals processing the Spanish counterparts to those sentences

Figure 7. Inferior Frontal Cortex activation o In Spanish (Bates, Devescovi, & D’Amico 1999), one would expect no (significant) differences in brain activity between the two sentence types with varying word order o This is exactly what the authors observed in the bilinguals in Spanish o The authors conclude that early bilinguals with extensive dual language exposure develop predominantly differentiated representations for each of their languages in one brain. 3. Summary of neurological studies

o Simultaneous bilinguals show brain activation for the two languages in (broadly) the same areas as monolinguals

o When language is acquired early, it is lateralized in a specific area of the brain where it will be more efficiently learned/processed

Ling C140. Fall 2019 G. Silvestri, Oct 15

11

o On the other hand, sequential bilinguals and L2 learners may show more spread brain activation for the non-dominant language

Assigned (**) and suggested (*) readings * Iacoboni, M. 2008. Mirroring people. The science of empathy and how we connect with others. New York: Picador. * Knightly, L-M, Jun, S-A, Oh, J-S, and T. Aux. 2003. Production benefits of childhood overhearing. Journal of the Acoustical Society of America 114 (1), 465-474. ** Polinsky, Maria, and Olga Kagan. 2007. Heritage languages: In the 'wild' and in the classroom. Language and Linguistics Compass 1(5): 368-395.

** Polinsky, Maria. 2018. Bilingual children and adult heritage speakers: The range of comparison. International Journal of Bilingualism 22(5), 547-563.

=========================================================

Appendix A. On ‘written’ code-switching: Eng-Ita email “Vediamoci nel mio ufficio, if you can. XXXX Vedrai il mio nome sulla directory, you buzz me and I’ll let you in. For last minute changes of plan, il mio campus number è XXXX” B. The case of bilingual Romance-Greek speakers in southern Italy.

§ negative on social stigma on bilingual competence § called ‘double-tongued’ and ‘double-minded’, i.e. ambiguous and untrustworthy