Psychology of Language Bilingual Language Processing

Psychology of Language

Bilingual Language Processing

Bilingualism  •  Definition  of  a  Bilingual  

•  Bloomfield  (1953):  full  fluency  in  two  languages.  •  Grosjean  (1989):  function  in  each  language  according  to  given  

needs.  •  Psycholinguistic:  speaking  two  languages  “fluently”  on  a  regular  

basis.  

•  Types  of  bilinguals  •  Early  (<  5  years)  vs.  late  (>  12  years)  •  Balanced  bilinguals  vs.  unbalanced  •  L2  proficiency  level  (low  vs.  high)  

•  Development  •  Simultaneous  bilingualism  vs.    

 Sequential/Successive  bilingualism  •  Early  Bilingualism/Simultaneous  

Bilingual  Disadvantage  •  Bilinguals  have  a  smaller  vocabulary  in  each  language  than  

monolinguals  (or  at  least  in  one  language)  (e.g.,  Bialystok  &  Feng,  2009;  Oller  &  Eilers,  2002).    

•  However,  overall  they  have  a  larger  total  vocabulary.  

•  Tip-­‐of-­‐the-­‐tongue  (TOT)  state:  bilinguals  reported  more  TOT  states  than  monolinguals  (e.g.,  Gollan  &  Acenas,  2004,  Gollan  &  Silverberg,  2001).  

•  Bilinguals  have  slower  verbal  fluency  (e.g.,  Gollan  et  al.  ,2002).  •  Slower  at  naming  –  saying  words  and  labeling  pictures.  

•  However,  disadvantage  is  limited  to  language    processing  tasks.  

Bilingual  Experiments  

•  Picture  Naming  

•  Semantic  Categorization    Say  “yes”  the  picture  is  a  type  of  clothing.    Say  “no”  if  it  is  anything  else.  

• Weaker  link  Hypothesis:  Bilinguals  are  disadvantaged  relative  to  monolinguals  on  speaking  tasks  because  they  divide  frequency-­‐of-­‐use  between  two  languages.  

  Performance  on  semantic  tasks  is  the  same  in  monolinguals  and  bilinguals  (at  least  in  an  L1).    

  Performance  on  word/picture  naming  is  slower  in  bilinguals  than  monolinguals,  and  slower  in  the  L2  than  in  the  L1.  

 How  can  that  be  explained?    How  are  the  languages  connected,  both  in  terms  of  words  and  

meaning?  

  Do  connection  between  the  language  exist  because  bilinguals  can  translate  between  the  languages  and  learning  an  L2  occurs  within  the  context  of  an  L1?  

Bilingual  Language  Organization  

Word  Association  Model   Concept  Mediation  Model  

Revised  Hierarchical  Model  (Kroll  &  Stewart,  1994)  

Translation  Tasks  •  Translation  Tasks  

•  In  translation  recognition  tasks:  •  Say  “yes”  if  the  words  are  translations.  •  Say  “no”  if  the  words  are  NOT  translations.  

•  Produce  the  translation.  

•  Which  is  faster,  translating  L2L1  or  L1L2?    •  What  might  this  tell  us?  

chien  

dog  

chat    

cat  

avion  tracteur  

dog  

chien  

cat    

chat  vs.  

vs.   plane  tractor  

hamster  mouse  rat  

rabbit  dog  bird  cow  sheep  goat  pig  

fish  dolphin  shark  turtle  

elephant  giraffe  lion  zebra  

monkey  gorilla  

In  your  list,  did  you  include:    cat  horse  tiger  

 Cannot  bypass  the  conceptual  store  on  this  task.  

 Thus  performance  is  similar  in  the  L1  &  L2.    

Semantic  Tasks  •  Semantic  Memory  

•  Give  a  set  amount  of  time  to  study  a  list  of  words.  •  Do  a  recall  test.  •  What  pattern  do  we  see?  And  does  this  differ  

 if  the  task  is  done  in  an  L1/L2?  •  What  does  this  tell  us?    

•  Semantics  &  Translation  Task  •  Translate  the  word  that  you  see.  •  What  pattern  do  we  see?  And  does  this  differ  

 if  the  task  is  done  in  an  L1/L2?  •  What  does  this  tell  us?    

mouton  étoile  voiture  

 Seeing  a  picture  activates  the  concept,  making  performance  similar  in  the  L1  &  L2.    

 This  is  in  contrast  to  a  pure  word  translation  task,  in  which  L1L2  semantics  can  be  bypassed.  

Exhaustive  Activation  •  In  English,  what  semantics  are  activated  when  seeing  the  word  TEAR?  

•  What  does  this  pattern  of  activation  tell  us  about  processing?  

•  Is  this  the  same  for  bilinguals?  Do  they  activate  all  possibilities  that  correspond  to  an  input  even  if  it  is  in  their  other  language?  •  How  would  we  test  this  question?  

  Processing  is  Exhaustive.      All  possibilities  that  correspond  to  the  input  are  activated.  

dog  cat  coin  corner  

  For  English-­‐French  bilinguals,  COIN  activates      ‘money’  &  ‘corner’.  

  Thus,  COIN  influences  response  times  to    CORNER.    Activation  is  Exhaustive  across  all  languages,      or  is  Non-­‐Selective.    

Exhaustive  &  Incremental  Activation    

•  Activation  is  incremental  in  an  L1  and  an  L2.  •  In  other  words  recognition  can  happen  before  the  end  of  the  word  is  

reached.  

•  All  possibilities  that  correspond  to  the  input  are  activated.  •  This  activation  is  Exhaustive/Non-­‐Selective.  

Click  on  the  caterpillar.    

What  happens  in  monolinguals?   What  happens  in  bilinguals?  

Click  on  the  marker.    

stamp  =  marku  

Context  Effects  •  Does  context  influence  the  activation  of  the  two  meanings  of  the  word  TEAR?  

•  What  about  in  a  second  language,  for  a  word  like  COIN?  

•  What  about  for  words  that  are  not  homographs?  •  For  a  sentence  like  the  following  what  word  is  expected?  

 Bill  found  some  ___________  .  •  What  about  for  a  sentence  like  this?  

 Bill  went  to  the  bank  to  deposit  some  ___________  .  

YES  

YES  

What  happens  if  money  vs.  dinero  is  the  next  word?  

  A  rich  context  can  mitigate  Exhaustive/Non-­‐Selective  Activation.    Why  might  this  be  important  for  second  language  teaching?    

What  happens  if  money  vs.  dinero  is  the  next  word?  

Exhaustive  Activation  in  Syntax  

•  Syntax  •  Which  construction  is  more  frequent  in  

English?    Kevin  gave  the  chocolates  to  Jennifer.    Kevin  gave  Jennifer  the  chocolates.    

•  What  happens  in  syntactic  priming?  •  Speakers  use  the  word  order  that  they  

have  heard,  even  if  it  is  the  less  frequent  one.  

•  Does  NOT  depend  on  reusing  words  (lexical  priming),  e.g.  GIVE.  

•  However,  we  do  get  a  lexical  boost  if  the  verb  is  repeated.    • Repeating  the  verb  increase  priming.  

The  girl  gave  the  woman  the  present.  Describe  the  picture  using  the  verb  GIVE.  Describe  the  picture  using  the  verb  PASS.  

Exhaustive  Activation  in  Bilinguals  

•  Syntax  •  Which  construction  is  more  frequent  in  

German?    Richard  gab  Sabine  die  Blumen.    Richard  gab  die  Blumen  Sabine.      

•  Do  bilinguals  experience  cross-­‐linguistic  syntactic  priming?  •  When  syntax  the  same  in  both  

languages,  bilinguals  will  reuse  a  syntactic  structure.  

•  Thus,  syntax  from  one  language  influences  processing  in  the  other.  

•  It  appears  that  activation  of  syntactic  structures  can  also  be  Exhaustive/Non-­‐Selective.  

Das  Mädchen  gab  das  Geschenk  der  Frau.  Describe  the  picture  using  the  verb  GIVE.  Describe  the  picture  using  the  verb  PASS.  

Stronger,  Faster  L1  Picture  Word  Interference  •  Say  the  name  of  the  picture.  

• What  is  the  pattern  of  results?  • Shared  phonology,  even  when  it  is  across  languages,  speeds  picture  naming.  

• Shared  semantics  slows  picture  naming,  but  this  is  more  evident  when  trying  to  name  pictures  in  the  L2.  

• What  does  that  tell  us?  • Language  processing  is  Exhaustive/Non-­‐Selective,  but  the  L2  is  activated  more  slowly  and  may  not  influence  L1  processing.  

Language  Switching  •  RED  =  name  picture  in  your  L1.  •  BLUE  =  name  picture  in  your  L2.  

• What  is  the  pattern  of  results?  •  L1  naming  faster  than  L2.  •  L2L1  switch  is  faster  than    L1L2.  

• What  does  that  tell  us?  •  Asymmetric  switching  cost  

indicates  that  the  L1  is  suppressed  to  produce  the  L2.  

•  Ellen  Bialystok  (2001):  •  Bilingual  children  need  to  be  aware  of  the  language    

 needed  in  a  particular  situation.    •  They  know  that  names  for  things  are  arbitrary.  •  Because  “bilingual  production  requires  the  constant  involvement  of  

executive  control  to  manage  attention  to  the  target  language,  then  it  is  possible  that  this  experience  enhances  that  system  by  making  it  more  robust  for  other  functions”  (Bialystok,  2008).  

•  Bilingual  children  performed  better  than  monolingual  children  on  metalinguistic  tasks  that  required  controlled  attention  and  inhibition,  but  not  on  task  that  involved  grammar  knowledge  (Bialystok,  1988).  

•  Bilingual  children  show  advantages  in  situations  of  conflict  and  distraction.  

Cognitive  Control  

•  The  Stroop  task  is  a  linguistic  task  that  involves  conflict  and  cognitive  control.  •  There  are  two  conflicting  responses:  the  

word  and  the  ink  color.    •  Must  manage  the  conflict  efficiently  to  

respond  correctly  and  quickly.  •  Bilinguals  are  “good”  at  managing  conflict  

as  all  responses  involve  managing  conflict.  

•  Does  the  enhanced  cognitive  control  in  bilinguals  extend  beyond  linguistic  tasks?  

Cognitive  Control  Tasks  

cat  vs.  chat  

Push  the  button  on  the  right  when  a  RED  square  appears.    Push  the  button  on  the  left  when  a  BLUE  square  appears.  

  Bilinguals  perform  better  at  the  Simon  Task,  and  other  non-­‐linguistic  tasks  of  cognitive  control.    

  Thus,  bilingualism  confers  a  general  benefit  for  cognitive  control.  

•  Bialystok  et  al.  (2007)  •  184  dementia  patients,  approximately  half  were  bilingual.  

•  Importantly,  the  age  of  onset  is  significantly  delayed  in  bilinguals.  

•  Effect  of  lifelong  bilingualism:  Maintains  cognitive  function  and  delays  the  onset  neurodegenerative  diseases.  

Neuro-­‐Cognitive  Benefits  of  Bilingualism  

•  Bilingual  children  have  advanced  metalinguistic  awareness  (but  monolingual  children  do  catch  up).    

•  Across  the  lifespan,  bilingualism  slows  verbal  fluency.  

•  Across  the  lifespan,  there  is  no  apparent  influence      of  bilingualism  on  speed  of  syntactic  processing.  

•  Across  the  lifespan,  bilingualism  improves      performance  on  cognitive  control  tasks  (both      linguistic  and  non-­‐linguistic).  

•  Bilingualism  appears  to  have  neurological  benefits  such  that  the  onset  neurodegenerative  diseases  is  delayed.  

•  It  is  important  to  note  that  most  of  these  conclusions  are  based  on  populations  of  bilinguals  who  use  both  languages  from  an  early  age  regularly  and  have  a  high  level  of  proficiency.  

Bilingualisms  Influence  on  Cognition  

Learning  a  Second  Language  

•  Clear  relation  between  age  of  acquisition  and  performance  in  grammar  &  morphology  in  an  L2.  

•  Individuals  differ  significantly  in  the  degree  to  which  they  can  master  a  nonnative  phonetic  code.  •  Eg.  For  Japanese  learners  of  English,  “lock”/”rock”  

distinction  difficult.  

•  Early  “overhearers”  have  more  native-­‐like  accent,  but  are  not  better  at  grammar  &  morphology  (Au  et  al.,  2002).  

•  Intense  perception  &  production  training  can  lead  to  native-­‐like  accent  (Bongaert  et  al.,  1997).  

•  Factors  such  as  working  memory,  age  of  acquisition,  amount  of  exposure,  and  motivational  constraints  play  a  role  in  L2  learning  success.  

 (Johnson  &  Newport,  1989)  

Teaching  a  Second  Language  •   What  can  we  do  to  make  teaching  more  effective?  

•  Providing  a  rich  context  helps  mitigate  the  influence  of  the  stronger,  faster  L1.  •  Slowing  down  the  activation  of  the  L1  can  help.  

•  What  leads  to  higher  L2  proficiency  for  later  learners?  •  immersion  •  enough  input  from  competent  speakers  •  input  adapted  to  the  appropriate  level  •  practice  in  a  comfortable  environment  •  having  time  

•  Immersion  programs  at  Canadian  Schools  have  had  good  success,  leading  to  high  fluency  in  French,  but  no  long-­‐term  cost  to  English  (after  some  initial  cost).  

Conclusions  

•  Bilingual  language  activation  is  Exhaustive  and  Non-­‐Selective.  •  This  means  a  bilingual  cannot  “turn  off”  one  language.  

•  The  L1  is  activated  more  quickly.  •  Thus,  we  see  more  of  an  influence  of  the  L1  on  an  L2  than  of  the  L2  

on  the  L1.    •  To  produce  the  L2,  the  L1  needs  to  be  suppressed,  which  can  make  

switching  back  to  the  L1  costly.  

•  Bilingualism  leads  to  slowed  verbal  fluency,  but      confers  important  neuro-­‐cognitive  advantages      that  extend  beyond  language.