mutants are so cool

this probably doesn't relate all that directly to the project, since it has nothing to do with brains or minds, but it does have to do with development and modularity, and i think it's cool.

http://www.ucl.ac.uk/~ucbzwdr/teaching/b250-99/homeotic.htm

basically when fly genes are messed with, they get legs on their heads. fully formed legs. I'm not totally clear on what homeotic genes are, but i think it has something to do with regulating development. like they tell the body part: grow in this location at this time and at this speed. and other genes (not homeotic ones) are responsible for deciding what the body part will be like, basically. so there are (at least) 2 different kinds of genes, some for the nitty gritty details, and some for the overall organization.

if someone who knew anything about biology were to read this, they's probably think i'm a total idiot, but i think this is the basic genetics for dummies sort of idea.

one way in which this might sort of relate to the project... it reminds me of something I read about kids with Williams Syndrome. they're these pixie-like kids who have strange shaped heads and who talk a lot, have really big vocabularies, and are hyper-social. they just talk a lot of bullshit and don't actually seem to understand what they or anyone else are saying. And what I read about them is that what seems to be wrong with their brains, is that they're missing a part. like it's just not there, and the rest of their brain ended up forming in a different way to fill in the gap, which is why their heads are a funny shape and they look like pixies. so that seems like it might be the same sort of mutation, where something went wrong with a homeotic gene.

Fodor book, part 1

for mental faculties, as for bodily organs, ontogenetic development is to be viewed as the unfolding of an "intrinsically determined process"

distinction between neo-cartesian and epiricist theories of mind
- innate knowledge, organs, modules
- blank slate where all knowledge is from observation

Chomsky:
what's innately represented must be/have propositional content
because it must interact with linguistic data
those interactions are assumed to be computational

but memory couldn't be innate in that way. it's more like a mechanism than a set of rules.
Plato's faculty psychology was spacially differentiated - birdcage for memories
Fodor's is functionally differentiated, leaving open whether there are locations in the brain for each

mental processes are computational insofar as they're cognitive
what is instinctive is genetically determined, but the reverse doesn't have to hold

vertical faculties:
- domain specific
- genetically determined
- distinct neural structures
- computationally autonomous

traditional faculty psychologists and associationists would agree on the virtual architecture of cognitive capacities, perhaps, but would disagree on the nature of the virtual architecture: whether it's direct faculty, or a construct (computation?) that makes the association between inputs and outputs.

Fodor asks: why would anything other than the virtual architecture be of interest to psychologists? why should anyone care whether faculties are assembled? if they're assembled, we could get to the primitive elements of cognition, and thereby have a way at an empiricist theory of cognitive development. (associationists) (environment to mental states by learning)

neurological speculation that the physical architecture closely parallels the virtual architecture.
do environmental conditions alter neural connectivity (specificity) on a large scale, or does the brain innately have large structures that aren't alterable? (presumably both sides would accept that small changes happen)
- reason for stressing the distinction between virtual and physical architecture is the exhibit thr actual organization of the mind a sjust one of the possibilities that could have been realized had the environment dictated an alternative arrangement of the computational elements. 36

taxonomy of systems:
- vertical vs. horizontal cog. org. -> domain specific or cross-domain operations?
- computation system innately specified, or formed by learning?
- assembled of stock elementary processes, or does virtual arch. map directly onto neural implementation?
- hardwired localized elaborately structured or implemented by relatively equipotential mechanisms?
- computationally autonomous, or share resources (memory, attention...) w/ other cog systems?

Fodor's view:
will all be fuzzy answers, but roughly,
modular systems are domain specific, innately specified, hardwired, autonomous, and not assembled.
(vertical follows)

- the set of processors for which the modularity view currently seems most convincing is coextensive with a functionally definable set of the cognitive systems
- there is some (more or less a priori) reson to believe that cognitive systems which do not belong to that functionally defined subset may be, in important respects, nonmodular
- our cognitive science has made approximately no progress in studying [putatively nonmodular processes like thought and fixation of belief] and this may well be because of their nonmodularity.

Modularity book intro

notes I took from the introductory sections of Modularity - Understanding the Development and Evolution of Natural Complex Systems (which I refer to as "Sandy's book on Modularity" in other posts)

complex ststems:
- hierarchical structures
- consist of subsystems through many levels
nearly decomposable = subsystems interact, but less frequently by an order of magnitude or two than the interactions within each subsytstem

why does complexity in the universe generally take this form?
- complex systems evolve from simpler systems by evolution
- independence of organs makes natural selection of beneficial designs more likely to occur, since they can improve independently w/o mutual dependence of design

Bonner (1988) and Lewontin (1974) argue that modularity is necessary for the very existence of adaptation.
Brandon (1999) transcendental argument says this is logically valid but not explanatory

is genetic modularity necessary for evolvability?
Gerhart + Kirschner (1998)
Raff + Raff (2000)
Wagner + Laubichler (2004)
- many other conceivable principles of evolvability: Hansen (2003)
- genetic evolvability might not appear logical and simple to us

further criteria for modularity:
- own structural and/or functional identity
- persists as identifiable units
- repetitive, reusable building blocks

ontologically, 2 varieties of modularity:
- primary property of how organisms are built due to organizational principles
- or an evolved property
can result from either parcellation, or from integration

Bechtel + Richardson (1993)
structures often do not map neatly one to one onto functions

top-down (Fodor, Chomsky) - modules as explanans
bottom-up (genetic modularity) - modules as explanandum
(huh? need to look up Latin, assuming that is Latin ;)

Churchland + Sejnowski (1992) brain module
neuropsychology -> link behavioural data with regions of the brain using brain-damaged patients and brain-imagery

Kosslyn + Koenig (1992) weak modularity
???

Fodor - information encapsulation

Evolutionary Psychologists
massive modularity hypothesis
adaptationism, nativism, computationalism
(my notes were pretty vague at this point, just lists of words without context, but i think this was a list of different ways that modularity could be understood, or different ways of splitting it up)

modularity of knowledge representation

computational - specific computational device
vs
chomskyan - domain-specific device for representation of knowledge to account for a cognitive capacity

preliminary bibliography

papers from the course that look relevant, at least at first glance:
Simon - The Architecture of Complexity
(for the argument about why evolution might have to be hierarchical in a particular way and so modular)
Schlosser - The Role of Modules in Development and Evolution
Schank and Wimsatt - Evolvability: Adaptation and Modularity
Raff - Modularity, Dissociation, and Co-option

papers from Edouard's class that might be relevant:
(will fill in later)

reference book on neuroscience:
Kandel, Schwartz, Jessell - Principles of Neural Science

chapters from Sandy's book on modularity:
(will fill in later)

other books:
Fodor - The Modular Mind
Minsky - Society of Mind
Karmiloff-Smith - Beyond Modularity

some possible leads to chase down:
Bonner (1988)
Lewontin (1974)
Brandon (1999)
Gerhart + Kirschner (1998)
Raff + Raff (2000)
Wagner + Laubichler (2004)
Hansen (2003)
Churchland + Sejnowski (1992)
Kosslyn + Koenig (1992)

notes from preliminary meeting with Sandy

Some of the questions that I might want to explore for the project:
- what do cognitive science and neuroscience mean by modularity?
- is there a mismatch in granularity between them?
- is there a conceptual misunderstanding between them?
(the idea is that they mean different things by modularity and the different ideas might not match up)

one hunch about this is that many cognitive scientists, like Fodor, talk about functional modularity, while neuroscientists might talk more about structural or physiological modularity. I don't know if either part of that is true yet, but it's a starting hunch.

some more possible questions:
- why do we need modularity?
(arguments for how evolution has to happen in hierarchical steps for it to feasible at all)
- what does modularity do for us?
(not sure what Sandy meant by this exactly)
- how is this idea of modularity (from cognitive science and psychology) mapped onto the brain? or how could it be?

a few new vocabulary terms:
epigenetics = how the expression of genes is affected by environmental factors
evo-devo = (evolutionary-developmental biology) how evolution and development interact

this evo-devo thing sounds really interesting and like one of the things i'm interested in but didn't know the name for, but for some reason Sandy was directing me away from the papers about it. not because it's crap, but because it wasn't what I was interested in, she thought. Maybe she's right about that, with respect to this project. Maybe she doesn't realize that that's what I want to look at because my articulation of the idea consisted mainly of hand-waving. Or maybe I still don't really know what it means. I scanned some chapters from the evo-devo section of the book we were looking through anyway, so that i can see what they're about and maybe figure this out.

to start

I'm going to keep a record of my research project here.

it's about how developmental work informs ideas about brain complexity.
or something along those lines.

so far i've read some intro material from a book on modularity Sandy lent me, scanned 5 chapters from it, and picked up a few books from the library: Fodor's "Modularity of Mind", Minsky's "Society of Mind", and Karmiloff-Smith's "Beyond Modularity".

if i weren't too tired at the moment, i'd write up some of the notes i took after my meeting with Sandy.