Transcript Surface structure

3/29/2016
• Language
» Articulatory Phonetics
◊ Top down processes
» Syntax
» Lexical and semantic factors
Oh freddled guntbuggly, thy micturations are to me
As plurdled gabbleblothchits on a lugid bee
Groop, I implore thee, my foonting turlingdromes
And booptiously drangle me with crinkly bindlewurdles
Or I will rend thee in the gobberwarts
With my blurglecruncheon, see if I don’t
Prostetnic Vogon Jeltz
Hitchhiker’s Guide to the Galaxy
◊ Case Grammar
» Neuropsychology of language
◊ Classical Localization Model
» Neurophysiology of language
Study Questions.
• How has sentence ambiguity been used to study the psycholinguistics
of grammar.
• Describe the classical localization model of brain and language. How
do different aphasias relate to the model. Give examples.
Language
• Articulatory Phonetics
» Voicing
Language
• Is speech special?
» Do we possess specialized neural mechanisms for
perceiving speech?
» Categorical perception
Percentage Identified
◊ Voice onset-time and distinguishing /d/ from /t/
100
80
60
40
20
0
10
20
30
40
50
60
70
Voice-onset time (ms)
Language
• Articulatory Phonetics
» Vowels
◊ Positioning and part of tongue
– Height
High (/i/ beet)
Med (/e/ bait)
Low (/a/ pot)
– Part
Front (/I/ bit)
Central (but)
Back (/o/ boat)
Language
• The search for invariants
» Distinctive features
» Problems with a simple bottom-up approach
◊ There are no periods of silence between phonemes
– The speech spectrograph
Language
• The search for invariants
Language
• The search for invariants
» Phonemic information is presented in parallel
◊ Coarticulation
◊ E.g. Cf. /M/ in “Tim” vs. “/M/ in “mad”
» We perceive them as the same, but they are different
» We perceive the same sound differently according to the
context
◊ E.g.: Writer vs. Rider
◊ E.g.: Insert a silence between /s/ and /i/ --> “ski”
Insert a silence between /s/ and /u/ --> “spew”
Language
• Top down processes
» Phonemic restoration effect (Warren, 1970)
◊ Their respective legi*latures
◊ Found a *eel on the axle
◊ Found a *eel on the shoe
Language
• Perceiving conversational speech
» Two main problems:
◊ There are no physical boundaries between words
– Anna Mary candy lights since imp pulp lay things
– ( An American delights in simple play things)
◊ Speech is sloppy
– He wants to kiss this Guy?
– Misheard Lyrics (www.kissthisguy.com)
– This was the best buy
vs. She is a bad girl
Language
• Perceiving conversational speech
Language
• Perceiving conversational speech
W h a d’ a
ya
D oo w i n
Language
• Top-down processes and speech perception
» Phonemic perception
◊ The McGurk Effect
» Sentence comprenension
◊ Miller & Isard (1963)
– Participants shadow sentences:
Grammatic: Bears steal honey from the hive.
Semantically incorrect: Bears shoot honey on the highways.
Ungrammatic: Across bears eyes honey the bill.
Language
• Top-down processes and speech perception
◊ Miller & Isard (1963)
– Results
Gram.
Nonsem.
No noise
89%
79%
Noise
63%
22%
Nongram.
56%
3%
Language
• Syntax
» Finite state grammar
◊ E.g, Miller (1958).
1
N
Start
0
G
0´
X
N
S
G
3
X
S
2
Finish
Language
• Syntax
» Finite state grammar
◊ E.g, Miller (1958).
Structured
L1
L2
SSXG
NNSG
NNXSG
NNSXG
SXSXG
SXXSG
Etc.
Etc.
Random
R1
GNSX
NSGXN
XGSSN
Etc.
R2
NXGS
GNXSG
SXNGG
Etc.
Language
• Syntax
» Finite state grammar
◊ E.g, Miller (1958).
Language
• Syntax
» Finite state grammar
◊ Problems with finite state grammar
– Linguistic competence
– Judgements of grammaticity (Chomsky)
e.g., Colourless green ideas sleep furiously
(For words never paired together)
– Judgements of agrammaticity (Miller and Selfridge)
e.g., Was he went to the newspaper is in deep end.
(For words often paired together)
– Resolving/explaining ambiguity
e.g., They are cooking apples.
Language
• Syntax
» Phrase structure grammar
» Constituent analysis
The boy hit the ball
Noun phrase (NP)
Det.
The
Noun
boy
Verb phrase (VP)
NP
Verb
hit
Det.
Noun
the
ball
Language
• Syntax
» Phrase structure grammar
◊ Three types of sentences
– Grammatical / meaningful: maps onto only one phrase structure.
– Nongrammatical: cannot be mapped onto a phrase structure
– Grammatical / ambiguous: maps onto more than one phrase structure.
e.g., They are flying planes
Sentence
NP
Sentence
VP
Pronoun
Cop.
They are
Verb
NP
NP
Pronoun
Part. Noun
flying planes
VP
Verb
Adj
They
NP
Noun
are flying planes
Language
• Syntax
» Transformational grammar
◊ Some ambiguous sentences are not explained by phrase
structure
E.g, Visiting relatives can be boring
◊ Both interpretations map onto the same phrase structure...but,
they map onto different meanings
» Surface structure: Superficial appearance (i.e., phrase
structure).
» Deep structure: The meaning of the sentence.
» Transformational rules: convert the deep structure into a
surface structure (a sentence ready to be spoke)
Language
• Syntax
» Resolving ambiguity (McKay, 1966)
◊ Lexical ambiguity
– E.g., Although he was continually bothered by the cold …
– Control: headache
◊ Surface ambiguity
– E.g., Although Hannibal sent troops over a week ago …
– Control: almost
◊ Underlying ambiguity
– E.g., Knowing that visiting relatives could be bothersome ….
– Control: visiting some
Language
• Syntax
» Resolving ambiguity (McKay, 1966)
◊ Results
Language
• Syntax
» Case grammar
◊ Semantic analysis involves determining the semantic role of
each word or concept and computing sentence meaning based
on that analysis.
◊ E.g,
A] The key will open the door
B] The janitor will open the door with the key
– The ‘key’ is the subject of [A] and an object in [B] but serves the same
role in each sentence.
Language
• Syntax
» Interaction between syntax and semantics
◊ Fillenbaum (1974)
– Had subjects read and then paraphrase several sentences
– Normal sentences:
Threat: Don’t print that or I will sue you.
Control: John got off the bus and went to the store
– Perverse/disordered
Threat: Don’t print that or I won’t sue you
Control: John went in the store and got on the bus.
– Results:
Perverse: 50% normalized in their paraphrases
Disordered: 60 % normalized.
When subjects checked their work, they missed half of the errors…
Nice we’re having weather isn’t it?
Brain & Language
• Neuropsychology of language
» Aphasia: Language deficits resulting from brain-related
disorders and injury.
◊ Very common
– 40 % of all strokes produce some aphasia
» Broca’s Aphasia
◊ Paul Broca - studied patient Leborgne (A.K.A.’Tan’)
–
–
–
–
–
Treated for leg injury
Died a few days later
Autopsied brain
Discovered ‘Broca’s area’
Left Hemisphere dominance for language
Brain & Language
• Neuropsychology of language
» Broca’s Aphasia
Paul Broca
Brain & Language
• Neuropsychology of language
Paul Broca
» Broca’s Aphasia
◊ Production Deficits
– Problems in producing fluent language
– Range from ‘Tan,tan,tan,…’ to short phrases
– Lack function words and grammar
May retain idioms (‘fit as a fiddle’) or songs
– Proximity to motor cortex
Dysarthria: loss of control over articulatory muscles
Speech Apraxia: Unable to program voluntary articulatory
movements.
Brain & Language
• Neuropsychology of language
» Broca’s Aphasia
◊ Comprehension deficits
– Unable to analyze precise grammatical information
E.g.
“The Boy ate the cookie”
Who ate the Cookie?
“Boy ate cookie”
Implied grammar (cookies don’t eat boys)
“The Boy was kicked by the girl”
Who kicked whom?
“Boy kick girl”
Paul Broca
Brain & Language
• Neuropsychology of language
» Wernicke’s Aphasia
Carl Wernicke
◊ Carl Wernicke, 1870s
– Examined two patients
Problems understanding language following strokes
Fluent but nonsensical speech
Poor language comprehension
Proximity to auditory sensory areas
(Wernicke proposed word memory area)
Brain & Language
• Neuropsychology of language
» Wernicke’s Aphasia
Carl Wernicke
◊Production deficits
– Sounds fluent (e.g., foreign language)
– Neologistic (invented words)
– Semantic substitutions
E.g.
I called my mother on the television and did not
understand the romers by the door.
Brain & Language
• Neuropsychology of language
» Wernicke’s Aphasia
Carl Wernicke
◊Comprehension deficits
– Do not recognize the incomprehensibility of their own
sentences
– Do not comprehend written or spoken language
◊“Here and gone again”
– Aphasia improves over time
– Anomia: Losing the ability to retrieve words (nouns)
Brain & Language
• Neuropsychology of language
» Classical localization model (Lichtheim, 1885; Geschwand,
1967)
Conceptual
Information
Broca’s
Wernicke’s
Brain & Language
• Neuropsychology of language
» Classical localization model (Lichtheim, 1885; Geschwand,
1967)
◊ Damage to main areas
– Broca’s Aphasia
– Wernicke’s Aphasia
◊ Damage to connections
– Conduction aphasia
– Transcortical sensory aphasia
Brain & Language
• Neuropsychology of language
» Conduction aphasia
◊ Damage to the connection between Wernicke’s and Broca’s area
– Arcuate Fasciculus
Brain & Language
• Neuropsychology of language
» Conduction aphasia
◊ Production deficits
– Problems producing spontaneous speech
– Problem repeating speech
– Sometimes use words incorrectly
◊ Comprehension
– Can understand spoken/written words
– Can hear their own speech errors, but cannot correct them
Brain & Language
• Neuropsychology of language
» Conduction aphasia
Conceptual
Information
Broca’s
Wernicke’s
Brain & Language
• Neuropsychology of language
» A prediction
Conceptual
Information
Broca’s
Wernicke’s
Brain & Language
• Neuropsychology of language
» A prediction
◊ Disconnecting Wernicke’s from the conceptual area
should lead to repetition without comprehension.
◊ Transcortical Sensory Aphasi
– Damage to the angular gyri
Brain & Language
• Neurophysiology of language in the intact brain
» ERP studies
◊ The N400: Semantic violations
Brain & Language
• Neurophysiology of language in the intact brain
» ERP studies
◊ The P600: Syntactic Positive Shift (syntactic violation)
Brain & Language
• Neurophysiology of language in the intact brain
◊ Dispreferred continuation of ambiguous sentences
– E.g.,
The spy saw the cop with the binoculars
Who has the binoculars?
“The spy has the binoculars” -> preferred continuation
◊ The N400 and aphasia
– Swaab et al.
Patients listened to sentences that had an anomalous word
at the end.
Brain & Language
• Neurophysiology of language in the intact brain
◊ Swaab et al. (1997)
Brain & Language
• Neurophysiology of language in the intact brain
◊ A caveat: Individual differences
– Stimulation mapping of the brain
Neurosurgery around left hemisphere language areas
– A couple hundred of partients
– Correlation with effects in Wernicke and Broca’s area are
week
Some patients have naming problems in the area, not all.
– Anatomical localizations vary considerably.
Brain & Language
• Neurophysiology of
language in the intact brain
◊ A caveat: Individual
differences