1- The modularity debate Flashcards
What is modularity about?
Different modules have different functions e.g. different areas if the brain do different things.
“Specialist systems” reporting back to “Command Centre”
They can function independently of each other (COGNITIVE IMPENETRABILITY/ INFORMATIONAL ENCAPSULATION).
Cognitive impenetrability (does not process any information from outside that module)
What is Fodor’s vision of modularity?
3 levels-
Transducers e.g. eyes, ears for perception of shape, colour, face recognition, prosodic and voice recognition. Turns physical signal into neural signal
Modules- integrate neural signals (not under voluntary control)
The central system (voluntary/executive control, thinks/feels, makes decisions, abstract thought, problem solving, forming beliefs, unencappsulated)
What is the Muller-Lyer illusion
There is a gap between what we know about the lines, and what we perceive.
Visual perception is a module, even though we KNOW the two lines are the same length, visual perception module does its own thing (tells us one is shorter) and we cannot change this.
How does this Muller-Lyer illusion arise?
Result of a depth perception heuristic.
“fast but dumb” procedure, a shortcut that give the right results MOST of the time.
Empirical argument for modularity
Data from performance on Muller-Lyer illusion, modularity of syntax
Deficits affecting very specific functions e.g. - Prosopagnosia = difficulty recognising faces
- Semantic category deficits (natural kinds versus artefacts). Will find it hard to label artifacts like spoon, cup but can name natural kinds like trees. This suggests they are stored in different areas of the brain.
Double dissociations (looking at 2 abilities in different populations). In first population, A is affected but B is intact, in 2nd population B is affected but A is intact.
- at behavioural level, e.g. “SLI” versus Williams Syndrome - at behavioural and neurological level, e.g. Wernicke’s versus Broca’s aphasia
Brain scanning
- for most cognitive tasks certain regions of the brain are more strongly activated than others
Evolutional argument for modularity
We need to rapidly make sense of the world in order to make informed decisions which will affect our survival–> division of labour between modules and central system.
Central system must be rapidly fed information from modules which will enable making right choices
Muller-Lyer illusion created by 3D perception module which cannot be “switched off”
“Heuristic” = special-purpose procedure
A ‘strategy’ or ‘shortcut’ procedure- It is fast but dumb.
It gives us the right results most of the time
Efficiency argument for modularity
David Marr - vision researcher
“Any large computation should be split up and implemented as a collection of small sub-parts that are as nearly independent of each other as the overall task allows” - otherwise “a small change in one place will have consequences in many other places”
Compare and contrast modules and reflexes
Modules and reflexes both involve information encapsulation, cognitive impenetrability. Not under voluntary control.
But modules are computationally complex, involving higher structures. Whereas, reflexes are computationally simple involving the brain stem.
Give characteristics of modules
Prespecified
Domain-specific
Fast
Mandatory
Shallow outputs
Informationally encapsulated
Fixed Ontogeny
Localised
What is domain specific?
Process a certain type of info e.g. visual, language (subdomains of semantics, syntax, phonology etc)
What is mandatory?
They operate automatically.
Will automatically tune in to conversations in your own language but not another.
The stroop test- asked to say the colours but graphemic representation is mandatory and difficult to suppress so interferes with sensory representation.
E.g. if a sentence obeys syntactic rules, we cant help but construct an interpretation for it ie syntactic processing is mandatory. Syntax comes before semantics in the processing chain.
What is meant by informational encapsulation
Modules are sealed from outside information e.g. Muller Leyer illusion.
McGurk effect contrast this. Integration of information across domains.
What is meant by shallow outputs
The information that modules pass onto other systems is fairly superficial.
What is meant by localised?
Modules are spatially localised in your brain. Evidence from brain scanning.
Recruitment of different regions for different tasks
Double dissociations at behavioural level, corresponding with deficits at neural level. E.g. production and comprehension in Brocas vs Wernickes area
What is fixed ontogeny?
Ontogeny=development.
Modules are pre-specified genetically so their development follows a fixed path.
What does pre-specified mean?
It is in your genes.
Infant brain may not be strongly modular, but there is a genetic program resulting in modularity.
Genetic instructions determine “activity-independent” phenomena
(a) the type of neuron a cell will become
(b) where they end up in the brain
(c) roughly which direction the axons point in.
However, a lot of brain connections emerge from experience (activity-dependent)
Modularity and language- SLI and Williams Syndrome both demonstrate modularity and double dissociation
Syntax in “SLI”
Severe difficulties producing and perceiving regular morphemes, e.g. He laugh (He laughed)
Moderate difficulties producing and perceiving irregular morphemes, e.g. It was broke (broken)
Difficulties understanding complex sentences, e.g. passives; the cat was chased by the dog.
What is Williams syndrome?
A rare genetic disorder, physical symptoms like smaller brain and cardiac problems as well as moderate-severe phycological symptoms.
Have good vocab, pragmatic ability, regular morphology but POOR IRREGULAR MORPHOLOGY.
Why is there regular vs irregular morphology in WS?
- spared computational system= good regular morphology
- impaired associative memory system= poor irregular morphology / unusual word choices
Compare and contrast WS and SLI
Regular morphology = good in WS, poor in SLI
Irregular morphology = poor in both but poorer in WS
Lexical abilities= good in WS, poor in SLI
Pragmatics= good in WS, sometimes impaired in SLI.
Double dissociation between WS AND SLI
In WS verbal language is good, but non-verbal abilities are impaired. Children with SLI show opposite pattern, whereby lang is affected but non-verbal abilities are not.
This provides evidence for a language module
Other dissociations in WS
Poor spatial abilities (spatial awareness, drawing) and poor numeracy
BUT
Good facial recognition.
Counter arguments for modularity
Neuroconstructivism- development results from constant 2-way interaction between genetics and environment.
Much of our neural wiring is activity dependent
e.g. in individuals with congenital blindness reading braille activates parts of primary visual cortex.
Hippocampi of cab drivers
Plasticity= ability of the brain to reorganise itself because of external experiences.
What is the nativist view of modularity?
Modularity is innate, present in the infant brain. Would expect to find a lot of double dissociations
What is the neuroconstructivist view of modularity?
The modular structure is not present in the infant brain but rather emerges through development.
Double dissociations emerge during development should not expect to find clear double dissociations in developmental disorders.
“Dethroning the Myth: Cognitive Dissociations and Innate Modularity in Williams Syndrome”.. what is the “myth”
The “myth” is the idea that people with Williams Syndrome have intact language and face processing (used to argue for independent modules). The authors argue that although these are relatively good compared to other areas of cognition, they are not “intact” and still develop atypically compared to typically developing individuals.
Challenges to Intactness Claims and Arguments for Atypical development in William’s Syndrome
Karmiloff-Smith and colleagues challenge the notion of intact language in WS. They argue that no aspect of language, be it syntax, semantics, phonology, or pragmatics, is truly intact. They say Nativist literature misinterpret findings, treating relative strengths as absolute strengths.
They highlight the limitations of small sample sizes and ceiling effects in studies that claim intact language abilities. For instance, they conducted a larger study on past tense formation and found no selective deficit in irregular past tense formation in WS when controlling for verbal mental age.
Instead of “intactness,” these researchers suggest that WS language is delayed and develops along a different trajectory compared to typically developing individuals.
They present evidence suggesting that individuals with WS may rely more on phonological information and less on semantic information during language development. This is supported by findings such as the naming spurt preceding fast-mapping ability and the lack of sensitivity to subcategory violations during sentence monitoring tasks in WS.
The authors argue that the good verbal memory often observed in WS can mask underlying difficulties with semantics and syntax, making language appear more intact than it actually is
It is not the case that people with WS have an intact face processing module and an impaired space processing module. Both follow atypical developmental trajectories.
What claims regarding morphological profiles in Williams Syndrome do nativists make?
Nativists believe individuals with Williams Syndrome exhibit intact linguistic functioning ‘hyperlinguistic’ and morphosyntax, whilst cognition is impaired (IQ measured around 50), suggesting the presence of an innate, independently functioning language module.
For instance, Pinker (1991, 1999) has claimed that syntax is spared in WS, citing observations of relatively good grammatical abilities in controlled testing. He contrasts WS with Specific Language Impairment (SLI), suggesting that in WS, genes impair intelligence while sparing grammar, whereas in SLI, genes impair grammar while sparing intelligence.
Clahsen and Almazan (1998) also support this view, proposing a double dissociation of innate mechanisms. They argue that WS involves impaired lexical memory but intact syntax, while SLI presents the opposite pattern. They base this on studies of tasks like past tense formation and sentence comprehension. They claimed WS individuals have a specific deficit in forming irregular past tenses but intact performace on regular past tense.
How do neuroconstructivists respond to these claims?
The authors of this article, Karmiloff-Smith and colleagues, critique these nativist claims, pointing out methodological limitations such as small sample sizes and ceiling effects in the studies cited by nativists. Ceiling effects can simply suggest task is not sensitive enough.
Given that WS language is seriously delayed initially, Thomas et al. (2001) argued that it is not sufficient to show that irregular past tense formation is poorer than regular past tense formation, because this is also true of some stages of typical development. the WS group displayed no selective deficit in irregular past tense formation.
The best way to characterize WS language is that it is delayed. If the WS infant brain presented with an intact morphosyntactic module, as Nativists suggest, this severe delay would surely be surprising.
What role does sample size play in research Williams Syndrome?
What does Karmiloff-Smith claim about the relationship between phonological and semantic abilities in Williams Syndrome?
WS language system develops differently, placing more weight on phonological information and less on semantic information. For example, the naming spurt in WS precedes fast mapping ability. In typical development, these two coincide (Mervis & Bertrand, 1997).
The naming spurt in WS does not coincide with exhaustive category sorting (indicative of semantic representation) Suggests that vocabulary growth relies less on semantics than in the typically developing case.
WS group displayed equal levels of reading for both concrete and abstract words (Laing, Hulme, Grant, & Karmiloff-Smith, 2001). Controls found concrete, imageable words much easier to read. Imageability effects are weaker in people with WS
What is the evidence against the claim that grammar is “intact” in children with Williams Syndrome
Vocabulary levels are usually better than syntactic levels in WS but both are significantly below conrtol (Karmiloff-Smith et al., 1997).
In imitation tasks, WS show impairment with complex syntactic structures like embedded relative clauses.
A recent study by Grant, Valian, and Karmiloff-Smith (2002) showed that despite having a mean vocabulary test age of 9 years, the participants with WS performed significantly worse on relative clauses than the 6- and 7-year-old controls.
Length of sentence did not explain the results because the shortest of the sentence types was the most difficult for the WS group who performed at ceiling on non
embedded filler sentences of varying length. These findings are inconsistent with the view that WS syntax is intact.
What is the “Modularity Continuity Hypothesis”? Is this supported or undermined by studies investigating dissociations in linguistic and numerical abilities across age in children with Williams and Downs syndrome?
“Modularity Continuity Hypothesis”= brain is organised into innate, genetically determined modules that are capable of being dissociated from one another across the lifespan.
Studies investigating dissociations in linguistic and numerical abilities in children with Williams syndrome (WS) and Down syndrome (DS) challenge this hypothesis…
Paterson (2000) compared the performance of toddlers and adults with WS and DS on vocabulary comprehension and numerosity.
Adults with WS performed significantly better on vocabulary tests than adults with DS.
However, toddlers with WS and DS showed equal impairment on vocabulary comprehension tasks, both performing significantly worse than typically developing children.
This suggests that either the developmental trajectories of language acquisition differ between the two syndromes, or children with DS experience a progressive decline in linguistic skills relative to their WS counterparts.
Adults with WS were significantly more impaired than adults with DS on numerosity judgement tasks.
In contrast, toddlers with WS demonstrated unimpaired performance on numerosity judgement tasks, while toddlers with DS were significantly impaired.
FINDINGS DEMONSTRATE- the relative strengths/weaknesses observed in adulthood do not reflect the cognitive profiles present in early development. The differing patterns observed across age suggest that the development of these abilities follows distinct trajectories in WS and DS, and that inferences about innate modularity based solely on adult phenotypic outcomes are misleading.
