Language and the brain ch 1

3 levels of description in Cognitive Science

1-Computational description

2- Algorithmic description

3-Implementational description

Each level captres diff question about how brain works

• answers to consider when you have questions: hypothesis

• Each level, different kind of answer to "How does the brain represent nouns?"

1-Computational description

• defines the problem the brain is trying to solve

• inputs and outputs

Ex: converting  2D patterns of light → 3D map. - converting sound waves to discrete speech sounds

2- Algorithmic description

• steps a system goes through to solve a problem in a computational order.

Ex: brain recognizes speech sounds by keeping a table with sound waves matched to appropriate sounds. Can match up/link in when it hears a sound

3-Implementational description

• how neurons/physical system carry out algorithm that solves a problem

Ex: how do neurons represent info physically in brain -how they interact  to look up" items in table

Scientist struggle to get a deep understanding of language in the brain by focusing only on its implementation

shows how we can’t learn how the brain represents focus by simply measuring brain activity and seeing the result. I- develop idea of what a noun is (computational level desc) 2- add hypothesis of how brain recognizes each objects (algorithmic) 3 - understand how neurons might implement those ideas (implementation)

Focus only on implementation - can’t understand language deeply 

• Ex: neuroscience couldn’t really understand video game by only paying attention to electric circuits 

• Diff strategies couldn’t make sense of microprocessor 

• Similar to brain being studied through neurologic disorder lens 

Barn owl example - diff levels can be combined

• computational desc of problem

• - sound waves reaching each ear need to be convented in representation that indicates where sound is coming from

  
  

goal: calculate difference in timing and loudness of sound detected at each ear

Barn owl example -

Algorithmic: Owls need to convert sound → location 

Delay lines: Parallel Pathways - one starting at right ear be one at left each path has forks, sound reaches or and goes down path. way of each ear, eventually meet

⬇

Coincidence detectors: circuits light up iff both inputs receiving signal at same time place where delay lines intersect a directly proportional to where in space they cave from

Ex: sound from right side of Owl: reaches right ear Ist and travels down right delay line

• Bc came from right ear travels further down the right delay line → coincidence detector lights up and senses this → spatiality mapped in the brain.

Barn owl example - implentation

Implementation : Parts in brain used to convert sound → location 

combines pair of delay lines w coincidence detector right ear Ist

- Activation passes @ fixed rate

- Coincidence detectors identify when signals cross

- Location encoded where cd activated

all parts need to be combined for entire thing to make sense

Paul Broca

found Brocas area

• 2 case studies with evidence of language disruption in specific area - language somewhat localized 

• area concerning speech comprehension - fluent aphasia - left tomporal lose

Carl Wernicke

found Wernicke's anen

• survey of language disorders

• found area concerning speech production difficulties - non tluent aphasia - left frontal loke led to "classical model" of language in the brain

Key pieces 

1- division of labor betwen language centers in brain  temporal area - comprehension frontal lobe area - production 2. regions are necessarily comected in lang - we produce language that we  understand 3. Centers are connected to other brain regions that language functions in

Cognitive view - language reflects a kind of knewledge

• ability to use lang → knowlege abt what u know

• knowledge alot mental entries, how to articulate sounds, maning, etc.

Outputs for 2 diff strategies are the same

Linguistics tree - word tree  represent diff algorithens/ steps      ⬇ differences have consequences for implementation in the brain Implementation hypothesis depends on syntactic structure & whether it was buit earlier or later on in sentence

Neurolinguistic research requires deep thinking abt

all levels:

→ answers at one level don't recessarily depend on answers at other levels

Marr - neurolinguistc

says too

→ Answers at separate levels must & can be brought together into a more comprehensive ace of cognitive system 

Linking hypotheses

• Neuro linguistics very concerned with these

• hypotheses about how these levels connect to each other