Lecture 24: Brain & Language II

As usual, these notes are provided as a supplement to the class lecture, and not as a replacement.

More Brain & Language

• Updated understanding of aphasias
• Steps towards a more detailed understanding (briefly begun)

Revised Understanding of Aphasia

Wernicke's patients clearly don't have just a comprehension problem

a. speech is typically somewhat incoherent

b. patients perform very poorly on semantic judgments

"which two out of these three words are related?"

dog, cat, turnip

mother, father, trout

--> deficit may be in semantic component of language

Broca's patients turn out to have a comprehension problem, when carefully tested

a. passive sentences

i. The dog chased the cat (active: BA patients understand)

ii. The girl threw the ball (active: BA patients understand)

 

iii. The cat was chased by the dog (passive, 'reversible': BA patients misinterpret 50% of time)

iv. The ball was thrown by the girl (passive, 'non-reversible': BA patients interpret more reliably)

--> they may be using word-order 'strategies' rather than syntax proper

b. attention to determiners

i. He showed her baby pictures (ambiguous for normals and BA patients)

 

ii. He showed her the baby pictures (unambiguous for normals; ambiguous for BA patients)

iii. He showed her baby the pictures (unambiguous for normals; ambiguous for BA patients)

--> deficit may be in syntactic component of language

Some Remaining Problems

Broca's Aphasia

Although patients show problems in syntactic comprehension, they are not entirely blind to syntax or function words

Ability to make grammaticality judgments (Linebarger, Schwartz & Saffran 1983); e.g. passives

a. ok John has finally kissed Louise.

b. * John was finally kissed Louise.

c. ok John was finally kissed.

d. ok John was finally kissed by Louise.

Given the sensitivity of these patients to the details of what does and what doesn't make a well-formed passive construction, it seems like we can't say that they have missing syntactic knowledge -- the problem seems to be that they can't use this knowledge in comprehension, though they can use it to make grammaticality judgments!

Wernicke's Aphasia

Although patients show evidence for lack of semantic knowledge on some tasks, more sensitive tests show retained knowledge (Milberg & Blumstein 1981)

e.g. Lexical Decision Task

Task: "Is this letter string a word?"

 

GLUB

TABLE

DOG

FORL

FORK

 

CAT ... GLUB

CAT ... TABLE

CAT ... DOG (faster to say "yes" to DOG)

KNIFE ... FORK (faster to say "yes" to FORK)

"priming" effects preserved in Wernicke's patients

Since priming effects seem to be contingent on having knowlege of semantic relations, we again can't really say that these patients are missing semantic knowledge, because then we couldn't account for the presence of priming effects. It seems that Wernicke's patients are unable to use this knowledge for making conscious semantic judgments, but it is available when it is used unconsciously!

How Brain Damage can Affect American Sign Language Speakers

Signers show same basic aphasic syndromes as hearing people, but sometimes effects are more striking

Brenda I. (reported by Poizner, Klima & Bellugi 1986), Right-hemisphere damage

• Visual problem: left visual field often ignored or moved
• Using space to describe bedroom: objects squeezed into right visual field
• Using space as part of ASL syntax: entire space used

Disorders of Aging...and the English Past Tense (Ullman et al. 1997)

Irregular verbs: past tense must be memorized

go/went, buy/bought, sing/sang

Regular verbs: past tense computed by rule ("add 'ed'")

want/wanted, play/played, fix/fixed (existing regular verbs)

blick/blicked, plag/plagged (novel verbs)

Alzheimer's Disease:

• typically associated with severe memory deficits; pat little trouble with skill learning
• at micro-level, seems to be due to neural 'tangles', especially in posterior temporal and parietal lobes

Parkinson's Disease:

• best known as motor problem (classic symptom of trembling); few problems with memory, great difficulty with skilled motor sequences
• known to be caused by problem in production of dopamine, a neurotransmitter important to the functioning of the frontal part of brain

When Ullman and colleagues tested Alzheimer's and Parkinson's disease patients on their ability to generate the past tense forms of verbs, they found that Alzheimer's patients showed problems with irregulars but not with regulars; Parkinson's patients, on the other hand, showed problems with regulars but not with irregulars. Even more strikingly, they found that the degree of difficulty that Parkinson's disease patients had with regular past tenses was closely correlated with the degree of difficulty that they showed in moving the right-hand side of their body. This is because the left-hemisphere areas of the brain that may be required for processing linguistic rules (i.e. Broca's area) are adjacent to the left-hemisphere motor areas, which control movement on the right hand side of the body.

Micro-level understanding of how the brain supports language

• Very little known to date
• Largely because we're not even sure what to look for!
• Situation is strongest in the case of speech perception - we at least understand a good deal about how the sounds are organized.

"Cortical Stimulation" (Boatman et al. 1995)

• Patients being treated for intractable epilepsy
• Mapping of brain functions prior to surgery
• Grid of electrodes placed on surface of cortex, allows recording and stimulation
• At one pair of electrodes over auditory cortex
  • selective inability to identify or discriminate stop consonants
  • (relatively) spared ability to perceive vowels
  • less difficulty with tones (non-linguistic)

Electrophysiological Recordings

Measurements can be made of voltages or magnetic fields outside the head ('brainwaves'), based on which it is possible to draw conclusions about thought processes inside the head. Electric recordings are known as 'event related potentials' or ERPs; magnetic recordings are known as 'event related fields' or ERFs.

Semantically anomalous words give rise to a characteristic brain wave, occurring about 400ms after presentation of the semantically anomalous word. This wave is known as the N400 (the 'N' stands for 'negative', because this is the normal polarity of this waveform). [cf. Kutas & Hillyard, 1980]

Example sentences:

I drink my coffee with cream and dog.

The man spread the warm bread with socks.

Before we go to bed, we make sure to turn off the living room giraffe.

Syntactically anomalous words give rise to a different kind of brain wave, generally occurring about 600ms after the presentation of the syntactically anomalous word. This wave is generally known as the P600 (the 'P' stands for 'positive'). [cf. Hagoort et al. 1993; Osterhout & Holcomb 1992; Neville et al. 1991.]

Example sentences:

The broker hoped to sell the stock was sent to jail.

The old man want to eat a hamburger.

Detecting Phonemes: role of Auditory Cortex (Phillips et al. 1997, 1998)

Within 200ms of the beginning of a speech sound, brain activity (observed based on magnetic fields recorded outside the head) shows that phoneme information has been extracted in left-hemisphere auditory cortex.

Conclusion

• Many properties of human language look like properties of a human 'instinct', which our brains are specially equipped for
• Our linguistic knowledge is highly sophisticated, and largely unconscious: includes phonology, syntax etc., etc.
• Our conscious linguistic knowledge (of prescriptive rules, speaker variation etc.) covers only a very small part of our linguistic abilities
• Languages are learned effortlessly by children under appropriate circumstances, with no conscious linguistic knowledge
• Acquiring literacy is an effortful task, which requires some conscious linguistic understanding (e.g. phonology)

In sum...

There's much more to human language than we're normally aware of.