Module 3: Bilingual cognition

topics

 Linguistic relativity

 Cross-linguistic transfer

 Bilingual memory system

Cognition

 Cognition means to know/understand. The collection of mental processes and activities

used in thinking, understanding, perceiving, learning and remembering.
 and the act of using those processes

 the cognitive processes include:

 Learning and Memory

 Thinking and Reasoning (Planning, Decision Making, Problem Solving etc.
 Language
 Vision-Perception
 Social Cognition
 It has already been established that language functions as a part of core cognitive
mechanism among humans.

 This skill works in collaboration with other mental functions, not in an isolated manner, as
was believed earlier.

 These cognitive mechanisms also have their respective neural representations ( which
sometimes co-activate) as findings from neuroscience prove.

 All of these functions, in turn, also connect with the socio-cultural environment of the
language.

 In other words, language is part of embodied cognition.

Language, bilingualism and Cognition

 British Journal of Psychology published two papers in 1920s

 Smith 1923: bilingualism and mental development
 Saer 1923: Effect of bilingualism on intelligence.

 The question: how bilingualism and multilingualism can have far reaching
impact in the cognitive domain.

 This include language and other mental functions.

 The question has far reaching impact for linguists as well as educators and
policy makers.
Two Main questions

 First, linguistic effect:

 This effect is commonly seen in language in terms of cross-linguistic transfer.

 This transfer can be both ways: first language to second language and vice versa.
 Secondly, an equally important question is the impact of a new language on other
cognitive domains, i.e. non-verbal cognition.

 Deals with cross linguistic variation in the way bilinguals perceive the world.

 Do the thought and perception of a bilingual change when they shift from one
language to other?

 Whether bilinguals differ from the monolinguals of the respective languages in

terms of their perception.
Language and thought/cognition

 Languages differ on how they map words onto concepts

 This brings to focus on the relationship of language with thought

 A question that has engaged linguists, philosophers, psychologists and

anthropologists for long.

 One standpoint in this domain is ‘linguistic determinism’ or linguistic

relativity.
Part 1
Linguistic Relativity
 Linguistic relativity
 Point of departure:
 there are innumerable ways to perceive the world around us
 Languages are different in how they ‘cut up’ the same world into concepts
and map these concepts onto words.

 large differences in language lead to large differences in experience and

thought.

 In other words, each language embodies a worldview, with quite different

languages embodying quite different views, so that speakers of different
languages think about the world in different ways.
 This view is also called the Sapir-Whorf hypothesis, named after
Edward Sapir and Benjamin Lee Whorf who made the idea
famous.

 Although linguistic relativism is perhaps the most popular version

of descriptive relativism, the conviction and passion of partisans
on both sides of the issue far outrun the available evidence.
Some examples of the ‘difference’

English Russian references

Staircase Isurin 2000
Ladder Lestnitza

Table
desk stol

Orange one word in Zuni Lenneberg & Roberts

yellow 1956
Pintupi language: Crystal (1987)
Pintupi English Meaning
yarla A hole in an object
pirti A hole in the ground
pirnki hole A hole formed by rock shelf
kartalpa A small hole in the ground
[Needs a lot of
yulpilpa elaboration in A shallow hole in which ants
English to convey live
mutara those meanings] A special hole in a spear
nyarrkalpa A burrow for small animas
pulpa A rabbit burrow
Makarnpa A goanna burrow
Boroditsky, Schmidt &Phillips 2003; p-61

 English: ‘the elephant ate the peanuts’. Mentions time of the

event in past tense.

 Mandarin: indicating time of the event is optional

 Russian: would include tense marking, the elephant’s gender,

whether the eater ate all or some of the peanuts.

 Turkish: needs to indicate whether the event was witnessed or it

was hearsay.
 Given the data:

 Most researchers agree that:

 Differences exist among language

 By implication, speakers of different languages may ‘attend to’ different

aspects of the same world.

 No dispute here
 Why may additional language affect
cognition?

 Reasons for the differences have been understood through :

1. The issue of codability

2. ‘Habitual thought’
 Codability
 Languages code concepts lexically.

 In other words, some concepts are grammaticalized, i.e. expressed

morphologically or syntactically in one language but not in another.

 Having a grammaticalized concept means having a readily available lexical term

for a concept.

 E.g. Alaskan language Dena’ina has different verbs denoting how trees grow on
the mountains, like, ‘growing on the upper mountain side’, ‘growing up the
mountain in strips’, ‘growing up the slope of the mountain’, and ‘growing
through the pass’.
 Similarly,

 Languages differ as to how they segment a continuum into labeled categories.

For example, Italian has a label for light blue (azzurro)and English does not.

 So, speakers of English-learning Italian are exposed to a new concept through a

new label or word.

 Some categories are purely linguistic, for instance animate, inanimate entities
put together in same category based on some culture specific properties on
which the grammaticality depends. For example, women, fire and dangerous
things.

 Lexical and grammatical categories correspond with the conceptual categories

which the L2 learner has to learn.
 Habitual thought

 Habitual uses of language can influence our habit of thought and action.

 For instance, some linguistic practice seems to be associated even with

cultural values and social institution.
 The concept of honorific pronouns in Hindi language.

 The manner of addressing someone differs based on their age, social position, etc.

 For example, “Tum lambi ho” and “Aap lambe hain” (You are tall) have the same
semantic meaning.

 However, while “Aap lambe hain” is used if the listener is someone older than the
speaker or if they have a higher social position, “Tum lambi ho” is used when the
addressee is younger than, or of the same age as the addresser.
 On the other hand, English lacks the explicit presence of honorifics, unlike in Hindi.

 Thus, the sentence “You are tall” can be used irrespective of the age and social
standing of the speaker and listener.

 Thus, Hindi speakers need to take into consideration the extra-linguistic factors
before speaking, which are irrelevant for English speakers.
 English Vs Japanese sentences:

 In an English sentence, pronouns cannot be dropped if they are used as the subject
of a sentence.

 Therefore, “I ate pasta yesterday” is correct, but “Ate pasta yesterday” is not in
standard English.

 However, in languages such as Japanese, pronouns are often dropped from

sentences.
 According to Kashima & Kashima (1998), people living in those countries where
pronoun drop languages are spoken tend to have more collectivistic values than
those who use non–pronoun drop languages such as English.

 The explicit reference to “you” and “I” might remind speakers of the distinction
between the self and other, and the differentiation between individuals.
Where is the dispute?

 Due to the differences, it has been claimed, that language ‘determines’

the way one would ‘see’ the world.

 It would mean that language has an effect on non-linguistic cognition.

 And, that one cannot grasp those concepts that are absent in one’s
language, be it in lexicon or grammar.

 Thus, grammatical/semantic structure of a language has

facilitatory/inhibitory effect on cognition/understanding of concepts.
 Early studies
 (Bloom, 1981)
 English and Chinese in processing counterfactual conditionals
 Counterfactual conditional (a conditional that describes the consequences of events
that did not happen
 E.g.
 If John had seen Mary, he would have known that she was distraught -- i.e., John did not see Mary
Vs.
 If John saw Mary, he knew she was distraught -- i.e., we don’t know whether John saw Mary

 The Chinese language does not distinguish between these two types of conditionals
either lexically or grammatically, and so gives no information as to whether an
event happened or not.
 Bloom created a counterfactual story about what would have happened if a
philosopher named Bier had known Chinese. He asked Chinese speakers to
answer questions about the story.

 Chinese monolingual speakers mostly did not interpret the story counterfactually,
but those who knew L2 English did so more often.

 It provided clear evidence of effects of bilingualism on cognition.

 It was also the first study of bilinguals to look at ‘grammaticalized concepts’.

 Bilingual cognition: areas under investigation

 Perceptual/sensory aspects
 Color cognition
 Pitch
 Taste

 Grammatical aspects

 Grammatical and biological gender

 Grammatical number
 tense marking
 Motion verbs
Sensory aspects
Color: The bilingual perception

 Color perception has been a traditional test set for Whorf’s principle of
relativity, which states that speakers of different languages evaluate
perceptual contrasts differently.

 Early studies showed that the speakers of Zuni, who do not distinguish
between yellow’ and ‘orange’ colors, do not distinguish the colors as
frequently or accurately as English speakers do.
 The domain of color has also served as a prime example of universality.

 Berlin and Kay claimed that despite the way languages name colors, their
underlying representation must be universal. This is because of the physical
property of color itself and arguably the same physiology of vision in humans.

 They noticed that participants from diverse cultural and linguistic backgrounds,
showed English-like color naming and prototype identification after some
training.
 This was taken as a proof of universalism.

 These subjects were immigrants to US with varying degree of

acculturation and proficiency in English.

 Ervin’s study on Navajo-English bilinguals (1961) showed that acquiring

another language may lead to a shift in naming and prototype
identification.
 Athanasopolous carried out a replication of Ervin on Greek English
bilinguals

 Greek has a two way distinction between ble [dark blue] and
ghalazo[light blue].

 The participants were divided into two groups: high proficient L2

speakers living in UK and low proficient L2 speakers living in Greece.
 The task was to point to the best example of ble and ghalazo on the
Munsell chips.

 Low proficient bilinguals showed a tendency to put ble away from blue
focus while the high proficient group tended to put ble closer to the blue
focus.

 This is in line with Ervin’s finding of a shift towards L2 category in terms

of color.
Munsell color chart
 Result of the study (Athanasopolous)
hue 10BG 5B 10B 5PB 10PB
9
8 Advanced
bilingual
ghalazo

7 Low proficient
ghalazo

6
5
4 Monolingual
English blue

Advanced
bilingual ble

3 Low
proficient L2
ble
prototype

2
 However, more interesting was the finding that the advanced bilinguals
put the ghalazo towards the L2 color but towards a lighter hue far away
from the L2 focus.

 Munsell chips shown as a two dimensional projection of hue [horizontal]

and lightness [vertical, value 2 being the darkest]

 This study shows that whereas there is a semantic shift towards L2

protoypes, the speakers also maintain their perceptual distance between
the two categories by further shifting ghalazo to ‘adjust’ this newly
arranged system in their mind.
 Linguistic tone, pitch patterns in languages, musical
tone perception

 With regard to hearing, language has a link to pitch perception.

 Work on language and pitch perception has shown differences of perceiving

the same pitch level differently by speakers of different languages
depending on the pitch range of the first language.

 As a result bilinguals perform differently in these tasks compared to

monolinguals.
 Some researchers have also looked at the ability to perceive and
produce musical tones in bilinguals who speak a tone language.

 E.g. Mang (2006) showed that Cantonese English bilinguals are

more in tune while singing English songs than English speaking
children.

 It is thought to be the result of Cantonese being a tone language.

Taste terms and taste concepts

 Like basic color terms there are basic of taste terms as well in languages.

 In western languages these basic taste terms are sweet, salty, bitter and
sour.

 But in Japanese and in Chinese, there is another taste concept called

‘umami’ which combines attributes of savouriness and meaty taste, found
in parmesan cheese, soy sauce, marmite etc.

 Studies have found that English learners of Japanese language could learn
this concept by exposure to the lexical item and actual food samplings and
this in turn affected their food categorizations.
Grammatical aspects
Grammatical VS biological gender

 This is an interesting domain to check the relation between language

and thought.

 Languages often use masculine, feminine and neuter gender on word

classes like nouns, adjectives and also verb endings.

 In many languages of the world, there is close relation between these

gender marking and the biological gender of the entity
 However, some languages mark gender in inanimate objects too.

 In such cases, the gender assignment is considered arbitrary, because different

languages can have different gender for the same object.

 Like: Sun
 Masculine in Spanish
 Feminine in German
 Neuter in Russian
Guiora1983

 This study investigated the role of gender load of a language on gender identity

 Subjects: Children aged 16 to 42 months

 Environment: monolingual environment of Hebrew, English and Finnish language in

their respective countries.

 Finnish has the least load whereas Hebrew has the highest.

 Task: Michigan Gender identity Test

 Finding: gender loading correlated with gender identity attainment.

 Boroditsky et al 2003
 Question: if a speaker uses a language with grammatical gender, does it
make/mislead him into thinking those objects have real gender?

 Mental representation of gender in Spanish-English and German-English bilinguals

 Task: name first three adjectives that came to their mind for 24 object names.
 Manipulation: the objects had opposite gender marking in Spanish and German.

 Output language for both groups: English

 Result: the groups used adjectives in keeping with the gender marking in their L1.
Grammatical number

 Grammatical number marking (singular Vs plural) depends on two main

features of nouns:
 +/- discrete
 +/- animate

 So,
 +animate + discrete: cat, dog, tree etc
 - animate +discrete: book, table, building etc
 -animate – discrete: water, sand, flour etc
 In English, both the first categories take plural marker.

 Thus, cats, houses etc

 The third category (-discrete –animate) cannot take plurals directly

 There ‘unitizer’ is used: unit of X:
 Thus, glasses of water, sacks of flour etc.
 In languages like Japanese and Yucatec, however,

 +animate + discrete: number marking is optional

 -animate +discrete : cannot take number marking nor can be preceded by a

numeral.

 These are quantified with a classifier

 ‘These classifiers are semantically unspecified with regard to individuation’

 Lucy & Gaskins (2001)
 The authors looked at monolingual Yucatec and English speaker’s mental
representation of objects.

 Subjects were shown three objects in each trial.

 One was target object. The other two were different from the target either in
shape or in material

 Thus
 Plastic comb with handle (target)
 Plastic comb without handle (mismatch in shape)
 Wooden comb with handle (mismatch in material)
 Task: pick one from the two that is most similar to the target

 Yucatec preferred a match on material but the English preferred a match in shape

 Other similar studies (Imai & Gentner 1997) found similar results.

 The final bottom line is that number marking in English makes shape a salient
property of objects whereas the lack of the same draws attention to the material
in case of Japanese or Yucatec.
Consequence: Athanasopoulos
(2006)
 Subjects: intermediate and advanced Japanese learners of
English

 Set of six drawn pictures, each containing nouns in any of the

three categories ( animals, implements, substances).

 Task: picture matching test (introduced by Lucy 1992). The target

picture needed to be matched to the ‘most like’ other from the
set.
 Results: intermediate learners behave like Japanese monolinguals.
Advanced learners behave like English monolinguals.

 Other similar studies showed similar results. (Athanasopoulos and Kasai

2008; Cook 2006; Basetti, Kasai, Sasaki & Takahashi 2006)

 These studies showed that languages does have an impact on cognition,

in terms of mental representations.
 Temporal events
 Temporal relations are encoded differently in different languages.

 In many languages, there are three basic temporal relations, simultaneity, before
and after.

 These have corresponding tense and aspect markers.

 English marks both tense and aspect, modern Hebrew marks tense, but not aspect,
Mandarin Chinese marks neither tense nor aspect.
 Although some morphemes have been identified as aspect marker of some sorts,
they do not carry that function and meaning exclusively.

 For example, ‘guo’, which is understood as an aspect marker, can also mean ‘to pass’
as a verb and so on.

 Even when the aspect marker is present, the time of an event is usually jointly
determined by aspect marker and other factors such as verbal semantics, situation
type of the verb etc.
 In a study conducted on Mandarin Chinese speakers, the participants were presented
with set of pictures depicting three different temporal event [past, present and future]
and they were asked to describe them, individually.

 Chinese participants showed a tendency to describe past and future phases as present.

 However, when told beforehand that each action could assume one of the three
temporal phases, this tendency disappeared.

 This is interpreted as reflecting their ‘habitual way’ of looking at things.

Bilingual Chinese study:

 The study used two groups of Chinese English bilinguals: high and low proficient.

 Material:
 18 action events [blowing up a balloon, crossing a log, erasing something on a
whiteboard etc]. One woman performed all the actions..

 A snapshot was taken at each of the temporal phases of the action event: about to cut
a rope, is cutting a rope, has finished cutting a rope.

 Altogether there were 54 pictures. For each picture a Chinese sentence was
created to describe the event.

 Another 62 pictures and sentences describing people or objects were used a

fillers [this is a teacher, this is a pen…].
 The participants saw a total of 80 sentences [62 non target and 18 target].

 Each sentence was followed by two pictures: one matched the sentence, other
depicting the same action in a different temporal phase [target condition] or a
different object or occupation [in non target condition].

 The participants had to choose which of the pictures depicted the sentence by
pressing a ‘left’ or ‘right’ key. This was RT study.

 The results showed high proficient bilinguals had an advantage in accessing the
temporal phase of the action in past and future phase, though not in present.

 The low proficient bilinguals performed like Chinese monolingual in the previous study.
 Motion verbs

 Languages vary a great deal in how they express motion in space.

 Typically research in this domain focuses on identifying the scope of

variation across different languages and how such variation can affect
behavior.

 Depending on where the path information is presented in the verb phrase,

languages are divided into verb framed or satellite framed languages.

 Le garcon traverse [path] la rue en courant[manner]

 The man runs [manner] across [path] the street
 Hence, French speakers mention manner, when it is an issue, and
are less sensitive to the same.

 But English speakers make widespread communicative and

cognitive use of this dimension.
 An interesting study tried to find out whether speakers of Greek [verb framed
language] and speakers of English [satellite framed language] would attend to
different aspects of a visual scene when watching a motion animation.

 They used an eye tracker to track the participants’ gaze while watching a series of
clip art animations.

 They were told that they would be asked to describe the event after watching it.
 It was found that the Greek participants looked at the path end point first and only
later looked at the instrument depicting manner.

 English speakers showed the opposite pattern.

 However, no such effect was observed when they were told to remember the
event without having to describe them.
 Descriptions of motion events in L2 learners are affected by their first language.

 On the other hand, the L2 also affects conceptualization in L1

 Japanese English bilinguals encode manner more in L1 Japanese than Japanese

monolinguals.

 L1 descriptions of motion verbs in terms of manner and path are affected by L2.

 English learners of French find it difficult to convey the same level of density [path
+ manner] in their second language as expressed in their L1 and as a result often
‘flout’ rules of their L2 to manage the same.
Conceptual Transfer
 Any language, learnt after the first language, does not develop in a vacuum

 Prior knowledge of other language/s plays a role there.

 Influence of the previous knowledge of one language on the knowledge and use of
another language is called ‘conceptual transfer’. (Odlin, 1989; Jarvis & Pavlenko,
2010).

 Various terms used for the same phenomenon: interference, transfer, cross
linguistic influence etc.
 While the term ‘Conceptual Transfer’ was first used by Pavlenko (1998), this idea had
already been displayed in previous work, often under the label of concept-based
transfer or concept-based influence.

 Janse’s (2002) work refers to bilingualism in Greece and that the word ‘barbarian’
referred to someone who either did not speak Greek or spoke ‘bad Greek’; in today’s
terminology, someone who spoke with L1 influence.

 Uriel Weinreich’s Languages in Contact (1953), is considered to be first the work that
initiated the academic research on transfer.
 Weinreich noted several types of transfer (which he calls interference)

 discussed methods for their identification and quantification.

 their relationship with other aspects of bilingualism.

 He developed the concept of interlingual identifications

 This is the situation when an individual identifies the structures of two
languages as same or similar, consciously or unconsciously
 Weinreich’s idea of transfer was not monolithic. (Odlin 1989)

 two types of Interferences (Weinreich) : borrowing transfer and

substratum transfer.

 Borrowing transfer: influence of second language on a previously

learnt language.

 Substratum transfer: influence of source language on the acquisition

of target language
 In 1957, Lado claimed that individuals tend to transfer the forms and
meanings, and the distribution of forms and meanings of their native
language and culture.

 This happens

 Productively, when attempting to speak the language and to act in the

culture.

 Receptively, when attempting to grasp and understand the language

and the culture as practiced by natives (p.2) (Gass and Selinker,
1993:1)
 Similar viewpoints were put forward by others of the same time.

 It was also the time of contrastive analysis wave. The idea of transfer greatly
influenced CA.

 However, with the beginning of empirical research in 70’s, criticisms also started.

 Learning difficulties could not always be predicted by crosslinguistic differences

 Also, many of the difficulties in learning could not be explained by CA.

 Many, like Corder, did not want to use the term ‘transfer’ without proper data.

 Corder (1993:19) believed the use of terms like ‘transfer’ should banned from use
unless carefully redefined

 Corder (1993:25): "If anything which can be appropriately called transfer occurs,
it is from the mental structure which is the implicit knowledge of the mother
tongue to the separate and independently developing knowledge of the target
language. The evidence for such a process is presumably the persistent
occurrence of incorrect mother-tongue like features in the learner’s performance"
 Selinker (1972) took Weinreich’s idea on interlingual identifications to a step
further.

 He brought in the concept of interlanguage.

 Later, Kellerman (1977) formulated the concept of psychotypology (the learner’s

perception of the similarities and differences between languages).

 By 1990’s the focus of transfer shifted from looking for errors in target language to
understanding the nature of cross linguistic influences in its totality.
 Kellerman and Sharwood-Smith (1986) proposed the term ‘crosslinguistic
influence’

 They pointed out that there need not be a transfer always, but the mere presence
of one knowledge system can influence the acquisition of another.

 Pavlenko (1998) used the term conceptual transfer.

 Overall, it tries to account for this influence in terms of “similarities and

differences between the structural properties of the source and the recipient
languages” (Jarvis & Pavlenko, 2010, p. 112)
 Taking into account both similarity and difference between L1 and L2

 errors are just one possibility in case of transfer.

 Other possibilities include:

 the conventional use of L2, sometimes even facilitating or accelerating its acquisition
(Schachter & Rutherford 1979).

 underproduction or overproduction of structures of the recipient language (Ringbom,

1978).

 even in the preference for certain structures over others (Sjöholm 1995)
One example..

 Sjöholm 1995 : acquisition of English phrasal verbs by native speakers of

Swedish and Finnish.

 In case of a choice between single verbs and phrasal verbs, native speakers
of Swedish tend to use more phrasal verbs than native speakers of Finnish.

 That is because Swedish is much closer to English than Finnish, and it also
has phrasal verbs.

 In other words, Swedish speakers transfer their previous linguistic

knowledge, facilitating the acquisition and use of L2-English structures.
 Main points
 Thus, the result of cross-linguistic influence/transfer depend on a number of factors.

 Distance between the languages

 Structural properties of language

 non-structural properties of the language use

 Similarity between the mother tongue and new language:

 positive transfer: facilitates learning of L2

 Differences hinders learning:

 negative transfer: Inhibits learning, avoidance, underproduction, overproduction
 L1 –L2 transfer effects: lexical transfer
 Ringbom (1991; 173) suggests two types of cross linguistic L1 transfer

 Overt: cross linguistic influence may facilitate or inhibit learning of L2. If the
languages are similar, there is facilitation.

 Covert: when learners do not have adequate L2 vocabulary or structures , they

may use L1 structures in its place.
 Lexical transfer:
 "Lexical transfer means that learners assume identity of semantic structure
between the words in Ll and L2, and it is manifested in loan translations and
semantic overextension“
e.g. Malayalam speakers using ‘headbath’ (in English) for shampoo/hair wash. Here
there is no morphological similarity between the words, but speakers assume
semantic similarity.
 Lexical borrowing: this refers to the use of L1 word in L2 discourse, when the
learner fails find one in L2.

 Borrowing can further be of different types

 Content words
 Function words
Some examples

 Over production: Odlin (1989:37) gives an example concerning Japanese

students who in an effort to avoid relative clauses may violate norms of
written prose in English by writing too many simple sentences.

 Under production: "avoid using linguistic structures which they find

difficult because of the differences between their native language and
the target language” (Ellis 1994;304)
Non-structural factors

 These factors include

 Individual : motivation, the age of the learner, the human awareness of the
language, personality influence transfer (Odlin, 1989:129-130).

 Social: context of use, attitude toward mixing.

 Türker (2016) explored effects of L1 conceptual transfer on L2 comprehension of
figurative expressions, like metaphors.

 Participants were native American English speakers who were studying Korean in a
formal classroom setting in an institution.

 Task: provide the L1 equivalents of the 54 Korean metaphorical expressions for

ANGER, HAPPINESS and SADNESS concepts in three different tasks.
 Task 1 was a decontextualized task with isolated metaphoric expressions at the phrasal
level
 Task 2 was a contextualized task that had metaphoric expressions within a dialogue
(limited)
 Task 3 was a contextualized task that included metaphoric expressions within an essay
(elaborate).
 The results revealed that the effects of conceptual knowledge of L1
varied, depending upon the context it was mentioned in.

 Participants performed better in tasks involving figurative language for

which L1 and L2 show similarities at both lexical and conceptual levels,
but only in the conditions of no or limited context.

 In the same conditions, the study also found a significant effect of L1

frequency on L2 processing of figurative language
 L2-L1 effect/influence:
 Majority of studies in this domain investigate the influence of L1 on L2. the reverse effect has
not been studies to the same extent.

 Reasons for this are many

 One, most of these researchers focus on early stages of language acquisition. At this stage, the
influence is almost always from L1 to L2.

 Secondly, often motivated by the question as to how members of immigrant communities can
best master the host country’s language, i.e. the dominant language. As a result, only the L2
was in focus and whatever influenced it. Not vice versa.

 Third, the belief that first language is more stable and robust and hence may not get affected by
other, later languages.
 However it is now clear that cross linguistic influence can work both
ways.

 In a bilingual/multilingual, there exists a dynamic language system, with

much ‘cross-talk’ between various language subsystems.
 Pavlenko & Malt (2011) investigated naming of common household objects in the
first language of Russian-English bilinguals.

 These bilinguals were divided into three groups based on their L2 acquisition:

1. Early Bilinguals, who arrived in the US aged 1-6, grew up speaking Russian at home, and
English outside.

2. Childhood Bilinguals arrived between ages 8-15. They also used Russian at home and
English outside but they did their schooling in Russia for a period of time.

3. Late Bilinguals included those who arrived in the US between the ages of 19 to 27, with
majority of them completing their undergraduate education in Russia. They used Russian
with Russian-speaking people and English with English speaking people around them.
Stimuli :
60 images of drinking containers, varied in terms of shapes, size, materials, and
specific uses.
also included ones made in the U.S. (e.g., a beer stein) and those made in Russia
(e.g., a tea glass in a metal glass holder).

Results:
 The effect of L2 in the naming task was observed to be the strongest among early
bilinguals who were exposed to L2 the longest among the three groups.

 The late bilinguals also showed some influence, albeit limited, of the L2
suggesting that moderate exposure to L2 may influence any group of speakers
over time. Perhaps longer immersion may increase the L2 influence.
To sum up

 A general agreement that transfer, of different types, constitute an

important aspect of L2 leaning.

 This needs to be looked at along with other factors, like natural

principles of L2 acquisition. (Ellis, 1994:341)
Module 3 part 3

Bilingual memory models

Where we have reached so far…
 So far we have seen
 Language acquisition can be of different types
 This leads to different patterns of learning and use of language

 Now, once the learning has taken place, the language information must be stored
somewhere.
 This will include structural, conceptual and pragmatic information.

 From this location, we can access the relevant information and process them
during any linguistic activity

 This brings us to the bilingual memory storage, its access and processing.
Memory

 A familiar smell !!
 A flower
 Blade of grass with a dewdrop !!
Persistence of memory: Dali 1931
 Small child playing in the sun in winter

 How often we have been transported to our childhood by a picture, a

smell, sight, action !!!
 These are triggers for memory to awaken. Long term memory. Or more
precisely, explicit memory.
Memory or memories!!!
 Turns out there are memories, and not just memory !!!
 Theoretically, there are broad distinctions within the memory system, made in the
literature in the following lines:

 Short Term Memory

 Long Term Memory

 Long term memory has further divisions:

 Declarative (explicit) and non-declarative (implicit)
Human memory system

Source: Queensland Brain Institute

 This distinction of human memory system into two types, goes back to James
(1890). He distinguished between primary memory and secondary memory.

 Primary memory: for recent experiences

 Secondary memory: for information stored over a long period

 Later, in 1960’s the new field of Cognitive Psychology emerged, which looked at
human’s information processing capacity (Neisser 1967).

 This new development renewed interest in studying memory systems.

 Atkinson and Shriffin (1968) then came up with STM and LTM, devoted to
maintenance and encoding, respectively.

 Evidence for this difference came form research on amnesia, where

patients displayed specific impairment to one of these memory systems.
(Baddeley & Warrington 1970; Milner 1966)
 Long Term Memory
 Declarative or explicit memory:

 knowing the factual information about events etc, the ‘that’ about information
 Includes episodic and semantic memory
 Scholars’ position on the relationship between episodic and semantic memory have undergone
changes over time.

 Non-declarative or implicit memory:

 procedural memory. This is concerned with information about ‘how ‘things work. The
process involved thereof.
 Also includes what is learned through conditioning
Explicit and implicit memory differences:
 have distinct neural components too ( Eichenbaum & Cohen 2001;Mishkin,
Malamut & Bachevalier 1984; Poldrack & Packard 2003; Voss & Paller 208) .

 play important but distinct roles in language acquisition and processing.

 Explicit memory can be

 explicitly demonstrated by verbally recounting an event
 Answering a query etc
 Is the focus in bilingual cognition research (Pavlenko 2000; Kroll & De Groot 1997)
 Implicit memory:
 can be shown through non-conscious change in performance, due to knowledge gained
over time.
 Acquisition of grammar
Declarative or explicit memory
Episodic Semantic memory

 world knowledge and

 highly specialized memories about
specific time related memory for encyclopedic knowledge,
events and personal experiences. language knowledge, object
This is a person’s unique memory knowledge.
and hence different from another
person.
 This is not personal in nature and
 specific details of the event (time and hence is objective. E.g. Chennai
place)
is the state capital of Tamil Nadu.
 context (what happened next)
 emotions (how you felt).

( Endel Tulving 1972; Howard & Kahana, 2002)

Episodic memory
 contains information stored that has some sort of marker for when the event was
originally encountered.

 It holds the events from personal experiences in the past, exists in subjective time and
space, requires a conscious recollection and a controlled process

 In other words, it contains the records of unique events which occurred at particular
times.

 Specifically, autobiographical memories are stored in the episodic memory.

 Semantic memory

 Semantic memory is thought to store general information, about language and world
knowledge
 It is a kind of mental dictionary containing all the attributes of event-free knowledge,
and has no concern with time or space
 It relates to general facts about the world
 E.g. The sun rises in the east.
 1 + 1 = 2, etc.
 Studies have shown that knowing a second language extends semantic memory and
other cognitive capabilities because they recruit various cognitive operations.
 Short term Memory

 Short Term memory is part of working memory (WM) system

 WM system also includes attentional and control units

 Information from LTM is transferred to STM to process for output.

 temporarily stores and processes information during mental operations.

 The working memory holds information for a short duration during which it retains all
the tools needed at the time to perform an action or make a decision.

 Working memory is the site where conscious thinking takes place.

 In terms of language,

 A major function of the working memory system is the retention and processing of
verbal information.

 Bilingual performance on working memory tasks can be affected by language

dominance, language proficiency, and the nature of the task.
Bilingual memory models:

 In the field of bilingualism research, bilingual memory storage has been a critical
topic.

 The main questions are:

 how is information about the two languages stored?

 Do they go to single storage or two different storages?
 If there are two storages, do they interact?
 What is the nature of interaction?
 There are two main hypotheses dominating the field of bilingual memory:

 the shared or interdependence memory hypothesis

 the separate or independent memory hypothesis.

 These models are concerned mainly with the concept or meaning and not
the other linguistic aspects such as phonology or orthography.
 Shared memory
proposes that a bilingual has
 one memory store for both of his languages.
 This is a language-free sort of storage and is basically a concept store
 a single meaning underlying both words/labels are kept.

 there is a kind of ‘tagging’ mechanism that identifies the words with a

language at the time of retrieval. E.g. for translation equivalents /ghar/
and /house/ for a Hindi-English bilingual, only the meaning is coded or
stored in memory.
(Caramazza & Brones 1980)
Prediction of shared memory

 Because both languages share the same underlying concept, they will
behave the same way. For example, a concept learnt in one language
will be transferred to the other language and one need not learn it again
in the other languages.

 Hence, in experimental set up, both within and between language

conditions would show same/identical results.
Evidence

Task: Spanish English bilinguals [64] performed a memory recall test.

stage 1: familiarization with a list of Spanish or English words.

Stage 2: depending on the test condition [within Vs between language], they were shown another
list in either Spanish or English containing either the same words or their translation equivalents.
For control, there were words that were not part of the first list, entered in the second list.

The result:
when the language of familiarization was English and test language was Spanish, the percentage of
recall was higher in case of familiar words compared to the control condition.

Similarly Spanish –English, Spanish –Spanish and English –English pairs showed facilitation
compared to the control tasks. This shows that language transfer was possible.
(Lopez and Young 1974)
 Another study:

Task: a semantic categorization task taking within language and between language condition as variables.

Stage 1: participants were shown a Spanish or English word representing a category [furniture/muebles].

Stage 2: This was immediately followed by an instance of a member of that category [bed/cama].

The task was to give true/false answer by pressing a key as to whether the second one was a member of
the previous.

RT was measured and results showed lesser RT in case of match in both within and between language
conditions.

(Caramazza and Brones 1980)

 Separate memory
 this model proposes a memory organization where a bilinguals two languages
are stored in two different and independent memory storage systems

 where one language information is not available to the other. Only interaction
is through translation process.

 separate memory hypothesis would predict a different result in within and

between language conditions.

(Heredia & McLaughlin 1992)

 Evidence:
 In a study similar to Lopez and Young , Scarborough et al (1984) did a
primed lexical decision experiment with Spanish English bilinguals.

 Part 1: bilinguals made lexical decision on a series of Spanish and English

words and non-words.

 Part 2: participants were asked for lexical decision on English words [and
non-words]. General findings of the study showed within language priming
effect but not between language effects.

 Other studies on sentence processing also found similar results.

 These mixed results have been explained with task demands by Durgunoglu and
Roediger [1987].
 the free recall tasks were sensitive to the semantic and conceptual processes and these
yielded results consistent with shared memory hypothesis.

 On the other hand, tasks that were sensitive to lexical or perceptual processes, supported
the separate hypothesis.

 This explains why similar studies, using different methods, yielded different results in
Lopez and Young Vs Scarborough’s work.
 To summarize,
 Conceptually driven tasks depend on shared memory as they are required to
access the more general store of the two languages or the general conceptual
system. But the data driven or lexical information related tasks are seen as
accessing the bilingual lexical system.

 these findings point to he fact that both interdependence and independence

hypothesis are correct and they both describe bilingual storage at different levels:
conceptual and lexical level.

 The interaction between these two levels are the subject matter of hierarchical
models
Hierarchical Models

 These models start with the assumption that bilinguals organize their knowledge of
two languages in two levels:

 one general conceptual level

 another lexical level that is particular to each of the languages.

 These models are hierarchical because they distinguish between these two levels.

 Another assumption of these models is that L1 lexicon is larger than the L2 lexicon.

 Now we discuss few of these models

 Word association [WM] model

 this model talks about a bilingual memory structure where the two languages
interact at the lexical level, based on translation equivalents.

 According to this model, words from one’s L2 are directly related to words
from his L1 and access to the general conceptual store from L2 is not possible
directly.

 This connection is always through L1 mediation.

(Potter, So, Eckardt & Feldman

1984)
L1
L2

concept
Concept mediation model [CM]

 as per this model, the L1 and L2 have direct access to the general
conceptual store.

 So words are not related interlingually.

(Potter, So, Eckardt &

Feldman 1984)
L1 L2

concepts
 To test these hypotheses Potter et al [1984] conducted two experiments with Chinese –
English bilinguals: one picture naming task and one translation task.

 Picture naming task: the participants were required to name picture in their L2, i.e.
English. The word association model predicts a long process for this.

 Translation task: participants were required to translate words from their L1 to L2. Word
association model predicts a shorter time for this task.

 Concept mediation models predicts similar time for both tasks.

 The results showed that the participants showed no difference in naming
and translation task, irrespective of proficiency, thus proving the
concept mediation model; i.e. words in L2 are related to L1 via a
common conceptual storage.
 However

 other evidence [Kroll and Stewart 1992] suggested that the word association model
described a bilingual at early stage of L2 acquisition. In contrast CM Model describes
a bilingual structure with high proficiency.

 Another important finding is that bilinguals translate faster from L2 to L1 than the
other way around. Both WA and CM model could not account for this asymmetry.

 To take these into account, Revised Hierarchical Model [RHM] was proposed.

[Kroll and Stewart 1994].

L1 L2

concepts
 RHM
 Lexicons are interconnected bi-directionally.

 The link from L2 to L1 is stronger [marked by solid line] as learning L2 often

happens through L1 and we often learn to map every L2 word with L1 word but
the reverse is not true.
 The link from L1 to L2 is weaker [marked by broken line] because of lack of
translation practice in this direction.

 The conceptual store and lexicon are connected via conceptual link.

 The link of concepts to L1 is stronger as opposed to that between concepts and

L2. This link may get stronger with higher proficiency but will always remain
weaker compared to L1 link.
Distributed conceptual feature model

 This model too distinguishes between lexical and conceptual level

 But goes a little ahead and describes how the conceptual store may be
represented at the word level.

 Thus, feature overlap across languages will signal similarity/difference among the
concepts.

 Thus, for example, cognates have more overlap than non-cognates.

Concrete words are expected to have more overlap than abstract words; abstract
words have language specific information.

Evidence comes from ‘concreteness effect’ experiments where concrete words are
processed faster than abstract ones.

(Distributed feature model of bilingual memory: De Groot 1992)

Bilingual Interactive Activation Model (BIA)
 BIA and BIA+ are models related to bilingual visual word recognition.

 The difference with RHM: integrated storage of bilingual lexicon

 This is a network model, where different levels (features, letter, word etc) interact and are
activated parallel or simultaneously.

 This type of models are called connectionist or parallel distributed models/

 Language input (orthographic and/or phonological) selectively activates those items that
overlap with the input form.

 Thus, overlap is important, not language. Supports non-selective access.

Purpose:
Explain how bilinguals retrieve
orthographical information from
mental lexicon that would
correspond to a word.

e.g. ‘A’ = /, -,

(Dijkstra & Van Heuven 2002)

 Bilingual Language Interaction Network
for Comprehension Model (BLINCS)

 This is s model of bilingual spoken (auditory) language comprehension. This too is a connectionist
model.
 In this model ‘interconnected levels of processing are created using dynamic, self-organizing maps’.

 In this model,

 The received acoustic signal travels to a feature level, then to phonemic, to lexical, and then finally to
semantic level.

 The levels interact bi-directionally

 Within levels, both language specific and shared representations are allowed.
(Shook & Marian 2013)
 summary
 Studies on the organization of bilingual memory have come a long way.

 We discussed a few of the models and studies that have shaped this domain.
Many other models exist too. E.g. Bilingual Dual Code theory (Paivio and Desrochers
1980) ; SOMBIP (Li & Farkas 2002), Inhibitory Control Model (Green 1998), Adaptive
Control Hypothesis (Green & Abutalebi 2013) etc .
 Initial models were quite general or let’s say underspecified( shared Vs separate).

 Gradually more nuances were brought into each aspect of that representation and
retrieval.
 At present, hierarchical models (including de Groot’s) are considered the clearest
ones to describe the bilingual memory storage.

More work is still

underway.
 References

 https://www.researchgate.net/publication/
231869051_Conceptual_transfer_Crosslinguistic_effects_in_categorization_and_construal
 Friedenberg, J., & Silverman, G. (2006). Cognitive science: An introduction to the study of the mind. Sage
Publications Ltd.
 Galotti Kathleen M. (2008). Cognitive Psychology: Perception, Attention, and Memory. Cengage Learning.
 Jarvis, Scott. (2011). Conceptual transfer: Crosslinguistic effects in categorization and construal. Bilingualism:
Language and Cognition. 14. 1 - 8. 10.1017/S1366728910000155.
 Jarvis, S. (2000). Semantic and conceptual transfer. Bilingualism: Language and Cognition, 3(1), 19-21.
doi:10.1017/S1366728900260112
 Kashima, E. S., & Kashima, Y. (1998). Culture and language: The case of cultural dimensions and personal
pronoun use. Journal of Cross-Cultural Psychology, 29(3), 461–486
 Mang, E. (2006). The effects of age, gender and language on children's singing competency. British Journal of
Music Education, 23(2), 161–174.
 SMITH, F. (1923), BILINGUALISM AND MENTAL DEVELOPMENT. British Journal of Psychology. General Section,
13: 271-282. https://doi.org/10.1111/j.2044-8295.1923.tb00101.x
 Athanasopoulos, P. (2009). Cognitive representation of colour in bilinguals: The case of Greek
blues. Bilingualism: Language and Cognition, 12(1), 83–95. https://doi.org/10.1017/S136672890800388X
 Athanasopoulos, P. (2006). Effects of the grammatical representation of number on cognition in bilinguals. Bilingualism:
Language and Cognition, 9(1), 89–96. https://doi.org/10.1017/S1366728905002397
 Weinreich, U. (1953). Languages in Contact. The Hague: Mouton.
 Pavlenko, A. (1998). Second Language Learning by Adults: Testimonies of Bilingual Writers. Issues in Applied Linguistics,
9(1). http://dx.doi.org/10.5070/L491005268 Retrieved from https://escholarship.org/uc/item/6496g7v9
 Guiora, A. Z. (1983). Language and Concept Formation: A Cross- Lingual Analysis. Behavior Science Research, 18(3), 228–
256. https://doi.org/10.1177/106939718301800304
 Lucy, J. A., & Gaskins, S. (2003). Interaction of language type and referent type in the development of nonverbal
classification preferences. In D. Gentner & S. Goldin-Meadow (Eds.), Language in mind: Advances in the study of
language and thought (pp. 465–492). MIT Press.
 Janse, Mark. (2002). Aspects of Bilingualism in the History of the Greek Language.
10.1093/acprof:oso/9780199245062.003.0013.
 Odlin, T. (1989). Language Transfer: Cross-Linguistic Influence in Language Learning. Cambridge: Cambridge University
Press. http://dx.doi.org/10.1017/CBO9781139524537
 Jarvis, Scott & Pavlenko, Aneta. (2007). Crosslinguistic Influence in Language and Cognition. Crosslinguistic Influence in
Language and Cognition. 1-287. 10.4324/9780203935927.
 Selinker, L. (1972). Interlanguage. Product Information International Review of Applied Linguistics in Language Teaching,
10, 209-241. http://dx.doi.org/10.1515/iral.1972.10.1-4.209
 Kellerman, Eric. “TRANSFER AND NON-TRANSFER: WHERE WE ARE NOW.” Studies in Second Language Acquisition 2, no. 1
(1979): 37–57. http://www.jstor.org/stable/44488152.
 Schachter, Jacquelyn and William E. Rutherford. “Discourse Function and Language Transfer. Working Papers on
Bilingualism, No. 19.” (1979).
 Pavlenko, Aneta & Malt, Barbara. (2011). Kitchen Russian: Cross-linguistic differences and first-language object naming
by Russian–English bilinguals. Bilingualism: Language and Cognition. 14. 19 - 45. 10.1017/S136672891000026X.
 SAER, D.J. (1923), THE EFFECT OF BILINGUALISM ON INTELLIGENCE. British Journal of Psychology. General Section, 14:
25-38. https://doi.org/10.1111/j.2044-8295.1923.tb00110.x
 Isurin, L. (2000). Deserted island or a child's first language forgetting. Bilingualism: Language and Cognition, 3(2), 151-
166. doi:10.1017/S1366728900000237
 Crystal, D. (1987) The Cambridge Encyclopedia of Language. Cambridge University, Cambridge.
 Boroditsky, Lera et al. “Sex, syntax and semantics.” (2003).
 Kashima, E. S., & Kashima, Y. (1998). Culture and language: The case of cultural dimensions and personal pronoun use.
Journal of Cross-Cultural Psychology, 29(3), 461–486. https://doi.org/10.1177/0022022198293005
 Bloom, B. S., Madaus, G. F., & Hastings, J. T. (1981). Evaluation to Improve Learning. New York: McGraw-Hill.
 Ervin, S. (1964). Language and TAT content in bilinguals. The Journal of Abnormal and Social Psychology, 68(5), 500–
507. https://doi.org/10.1037/h0044803
 Imai, M., & Gentner, D. (1997). A crosslinguistic study of early word meaning: Universal ontology and linguistic
influence. Cognition, 62, 169-200.
 Ringbom, Håkan. “THE INFLUENCE OF THE MOTHER TONGUE ON THE TRANSLATION OF LEXICAL ITEMS.” Interlanguage
Studies Bulletin 3, no. 1 (1978): 80–101. http://www.jstor.org/stable/43135191.
 Lantolf, J. P. (1996). Book review: Sharwood Smith, M. 1994: Second language learning: theoretical foundations. London:
Longman. 235 pp. Second Language Research, 12(1), 107–109. https://doi.org/10.1177/026765839601200105
 Türker, E. (2016). The role of L1 conceptual and linguistic knowledge and frequency in the acquisition of L2 metaphorical
expressions. Second Language Research, 32(1), 25–48. https://www.jstor.org/stable/26375836