Theories of Language

1
Theories of Language

• Mental lexicon store of language information
• Semantic (word meaning)
• Syntactic (word combination)
• Word form (spelling and sound pattern)
• Phonological (sound-based)
• Phoneme smallest unit of sound in a language
  that has meaning
• Older, evolutionary-based
• Orthographic (vision-based)
• Newer, object-based

2
Theories of Language

• Lexicon organization (Levelt's, 1994, model)
• Hierarchical, information-specific networks
• Conceptual semantic knowledge of word/meaning
  relations
• Lemma syntactic, grammatical
• Lexeme sound form

3
Theories of Language

• Conceptual organization (e.g. Collins Loftus,
  1975, semantic network model)
• Categorical networks of conceptual nodes
• Semantic priming studies
• TruckCar vs TruckTulip

4
Theories of Language

• Conceptual organization
• How many conceptual systems?
• Warrington's physical (living objects) vs
  functional (man-made objects) properties
• Caramazza's animate-inanimate categories

5
Theories of Language

• Conceptual organization
• How many conceptual systems?
• How many levels?
• Damsio's category-specific deficits in object
  naming
• Failures at lexical level

6
Theories of Language

• Components of a language analysis system

7
Theories of Language

• Components of a language analysis system
• Spoken input
• Phoneme analysis
• Voicing
• Vocal chords for and some consonants (e.g., b, m)
• Point of articulation
• Lips (b)
• Lips and mouth (f)
• Manner of articulation
• Blocked airstream (p)
• Unblocked airstream (l)
• Physical boundaries
• Prosody rhythm
• Stress

8
Theories of Language

• Components of a language analysis system
• Spoken input
• Neural substrate
• Superior temporal gyrus
• Wernike's area in LH
• Ventrolateral superior temporal sulcus

9
Theories of Language

• Components of a language analysis system
• Spoken input
• Written input
• Phoneme (alphabetic) systems letters represent
  sounds
• Shallow orthography close correspondence between
  letters and sounds (e.g., Spanish)
• Deep orthography lack of correspondence between
  letters and sounds (e.g., English)
• Syllabic systems characters represent sounds
  (e.g., Kana)
• Logographic systems characters represent words
  (e.g., Chinese, sort of)

10
Theories of Language

• Components of a language analysis system
• Spoken input
• Written input
• Feature analysis "what" pathway
• Modular models bottom-up processing
• Selfridge (1959) pandemonium model
• Interactive models top-down can influence
  bottom-up
• McClelland and Rummelhart (1981) connectionist
  model

11
Theories of Language

• Components of a language analysis system
• Spoken input
• Written input
• Feature analysis "what" pathway
• Modular models bottom-up processing
• Selfridge (1959) pandemonium model
• Interactive models top-down can influence
  bottom-up
• McClelland and Rummelhart (1981) connectionist
  model
• Word superiority effect supports interactivity
• Respond "t" or "k"
• Trip
• t
• Pirt

12
Theories of Language

• Components of a language analysis system
• Spoken input
• Written input
• Neural substrate
• Feature-level analysis same as object recognition
• Letters activate occipitotemporal sulcus compared
  to faces (Puce et al., 1996)
• Pure alexia damage to occipitotemporal areas can
  result in reading deficit with otherwise normal
  language

13
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Lexical access
• Lexical selection
• Lexical integration

14
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Lexical access differs for spoken and written
  input
• Dual-route for written
• Assembled grapheme-to-phoneme
• Direct whole-word orthographic
• Alexia without agraphia deficit in reading, but
  not in comprehending or writing
• Deep or phonological dyslexia cannot read
  pseudowords that are not in lexicon, make
  semantic errors (direct route)
• Surface dyslexia rule-based speech (assembled
  route) translates letters to sounds

15
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Lexical access differs for spoken and written
  input
• Dual-route for written
• Temporal for spoken
• Cohort model (Marslen-Wilson, 1980) phoneme
  analysis activates cohort of matching words until
  best match is selected

16
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Neural substrate
• Hierarchical model (Binder at al., 2000)
• Primary auditory cortex (Heschl's gyri) to STG
  no differential of speech sounds
• STG to superior temporal sulcus speech
  distinguished from nonspeech, but no
  lexical/semantic distinction (activated by
  speech, backward speech, nonwords)
• STS to MTG, inferior temporal gyrus, angular
  gyrus, temporal pole (more LH than RH) words
  distinguished from nonwords

17
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration higher-order processes
• Semantic and syntactic integration
• Grammatical structure

18
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration higher-order processes
• Semantic and syntactic integration
• Grammatical structure
• Parser process by which sentences are decomposed
  into syntactic units
• Garden-path model (Frazier, 1987) syntactic
  nodes have preferred structure based upon economy
  (minimal attachment and late closure), modular
  model doesn't allow for context

19
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration higher-order processes
• Semantic and syntactic integration
• Grammatical structure
• Parser
• Garden-path model (Frazier, 1987)
• Parsing is an interactive process (Münte et al.,
  1998) temporal organization of sentence affects
  processing

20
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration
• Neural substrate Broca's area
• Agrammatic aphasia deficit in producing and
  understanding complex grammatical structure after
  lesion in Broca's area

21
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration
• Neural substrate Broca's area
• Agrammatic aphasia
• PET studies (e.g., Caplan et al., 2000) find
  activation in Broca's area for normals when
  processing complex ("The juice that the child
  enjoyed stained the rug.") vs simple ("The child
  enjoyed the juice that stained the rug.")
  grammers.

22
Theories of Language

• Components of a language analysis system
• Spoken/Written input
• Word Recognition
• Sentence integration
• Neural substrate other areas
• Wernicke's (and Broca's) area and homologous RH
  locations (Just et al., 1996)
• Processing meaningful stories activates rCBF in
  the left middle temporal gyrus, the left and
  right temporal poles, and a superior prefrontal
  area in the left frontal lobe, but only the
  temporal poles remain activated whenever
  sentences with acceptable syntax and prosody are
  presented. (Mazoyer et al., 1993)

23
Theories of Language

• Speech production system (Levelt's model)
• Conceptual level
• Macroplanning communicative intent, planning
  goals of communication
• Microplanning planning word choice and grammar
  to convey intent
• Lemma level
• Formulator process of getting message into
  grammatical and phonological form
• Lexeme level
• Mapping phonemes to motor program

24
Theories of Language

• Speech production system
• Neural substrate
• Dronkers et al. (1996)
• Compared brain lesions of 25 stroke patients
  (aged 32-79 yrs) with apraxia of speech with
  lesions of 19 stroke patients (aged 52-80 yrs)
  without such deficits to determine brain regions
  potentially involved in coordinating speech. A
  robust double dissociation was found between the
  2 groups. All Ss with articulatory planning
  deficits had lesions that included a discrete
  region of the left precentral gyrus of the
  insula, a cortical area beneath the frontal and
  temporal lobes. This area was completely spared
  in all Ss without these articulation deficits.
  Thus, this area seems to be specialized for the
  motor planning of speech.

25
Neuropsychology of Language

• Aphasia disorders of language comprehension and
  production following neurological insult
• Primary aphasia problems with language processes
• Secondary aphasia problems with attention,
  perception, or memory that affect language (e.g.,
  alexia)
• No animal models for language, so most of what is
  known about the localization and structure of
  language systems comes from patient studies

26
Neuropsychology of Language

• Classical localization model of language system
• Broca's area LH inferior frontal lobe (areas 44
  45)
• Damage produces deficits in language production
  (Broca's or expressive aphasia)

27
Neuropsychology of Language

• Classical localization model of language system
• Broca's area
• Wernicke's area LH posterior STG
• Damage produces deficits in language
  comprehension (Wernicke's or receptive aphasia)

28
Neuropsychology of Language

• Classical localization model of language system
• Broca's area (M) word production - speech
  planning and motor programming
• Wernicke's area (A) word comprehension - storage
  of phonological lexicon
• Conceptual area (B) word representation - widely
  distributed (e.g., supramarginal and angular
  gyri)

29
Neuropsychology of Language

• Evidence for the classical localization model
• Damage to arcuate fasciclus between M and A
  produces deficits in repeating and producing
  speech but not in comprehension (conduction
  aphasia)
• Damage to fiber tract between B and A produces
  deficits in comprehension but not in repeating
  and producing speech (transcortical sensory
  aphasia)

30
Neuropsychology of Language

• Limitations of the classical localization model
• Not all patients that have damage to Broca's or
  Wernicke's areas show classic deficit
• Classic model fails to fully account for
  contributions of subcortical structures and
  damage to fiber tracts
• Comprehension often improves with time after
  damage to Wernicke's area
• Damage also produces anterograde and retrograde
  degeneration in other areas
• Anatomical location of focal lesion may not
  capture function loss in brain metabolism

31
Neurophsyiology of Language

• Semantic integration N400
• Large N400 with difficult semantic integration

32
Neurophsyiology of Language

• Semantic integration N400
• Syntactic processing P600/SPS
• Syntactic positive shift with violations of
  grammer

33
Neurophsyiology of Language

• Semantic integration N400
• Syntactic processing P600/SPS
• Syntactic positive shift with violations of
  grammar
• Syntactic processing LAN
• Left anterior negativity with violations of
  noun-verb relations