Friday, January 21, 2011
Does Natural Selection support Veridical Perceptions?
One claim that is used to rebut some forms of the AFR, including Plantinga's EAAN, is that natural selection selects for truth. According to this paper by Mark, Marion, and Hoffman, maybe not.
Subscribe to:
Post Comments (Atom)
19 comments:
A semi-interesting approach. Given the variables they play off one another:
"We find that the costs in time and energy charged to truth
can exceed the benefits it receives from perfect knowledge,
so that truth ends up less fit than simple."
Is it a surprise that if they charge enough for truth, they can find situations in which truth is traded for utility? That is, if we increase the cost of truth enough, of course we will end up with cases in which it is cheaper to get it wrong.
That said, I do find their overall approach conceptually fun, if a bit biologically untethered so very hard to interpret in a biological context.
Also, as they point out the strategy still perserves the correct ordering of quantities in the world, and to change this they have to introduce cost functions that have less biological plausability (they are not monotonically increasing functiongs of resource). Frankly, that section smells of fudging to get the desired answer.
However, I think this is an interesting line of research. Better to do simulations than to not do them, everything else being equal. I am certainly not wedded to the claim that evolution must produce veridical perceptions.
On the other hand, we know evolution has produced extremely accurate machinery for sensory representation. This is what we observe empirically in real (not simulated) sensory systems. They might in the future use more biological constraints on their evolutionary algorithms with some real biology.
Dr Reppert, I'm interested in your labelling the EAAN a form of the AFR. whilst the arguMent comes to a similar conclusion I believe the route taken to get there is markedly different. Personally I'm not convinced by Plantinga's claims about seeking for truth but I'm more persuaded by your arguments from casual closure. Hence it seems like a real possibility is agreement with your central premises, which don't rely on evolutionary arguments, but statement with Plantingas reasoning.
Selecting*. Not seeking, sorry
It is a type of the AFR, in that it says that if naturalism is true it undermines epistemic confidence. I think his argument actually does include an argument from mental causation as well as an argument that natural selection would select for truth. But it isn't the argument that I have spent the most time defending.
What's the deal here?
If I've said it once, I've said it a million times. We are not better adapted at living in African rivers than crocodiles, but that completely misses the point.
Our strength is in adapting to new environments (by finding niches within them) within a lifetime. We can evolve at the speed of learning instead of at the speed of genetic replication. Our strength plays out in our ability to migrate, expand our territory, adapt to changes in climate, and even change our environment. The power of this mechanism is in its generality, not its specificity to any particular environment.
You hand us a paper in which the environments are fixed, and none of the advantages of rational thought are examined at all. And you suggest that it's relevant to the EAAN?
It's stuff like this that makes me think you're not truly interested in the answer to the question.
DL the dangers of being a modeller rather than experimentalist!
DL: note you focused on humans, their model could be a useful model for less flexible species such as tree frogs. They are focused on perception over theoretical cognition, after all.
BDK,
I know. I think there's nothing wrong with their research.
The problem is in thinking this research is relevant to the EAAN. I'm not sure if the authors of the paper thought it was relevant, but Victor seemed to think so.
Doctor Logic says:
If I've said it once, I've said it a million times. We are not better adapted at living in African rivers than crocodiles, but that completely misses the point.
Our strength is in adapting to new environments (by finding niches within them) within a lifetime. We can evolve at the speed of learning instead of at the speed of genetic replication. Our strength plays out in our ability to migrate, expand our territory, adapt to changes in climate, and even change our environment. The power of this mechanism is in its generality, not its specificity to any particular environment.
You hand us a paper in which the environments are fixed, and none of the advantages of rational thought are examined at all. And you suggest that it's relevant to the EAAN?
No, it does not miss the point, and of course the environment is fixed. Evolution has no foresight, and therefore cannot select for potential usefulness in changing environments (we are not talking here about the usual variations within a particular, fixed environment in which evolution takes place).
Al,
You say:
Evolution has no foresight, and therefore cannot select for potential usefulness in changing environments
You're arguing that some as-yet-non-existent environment cannot affect the evolution of a population. By this you mean that a population cannot be selected for the specific parameters of an environment which does not yet exist. This is true, but, again, you're missing the point.
Variability is an aspect of almost all environments, and you hint at this in your response. For example, temperate zones have seasons, i.e., variability in temperature, cover, food, etc. If you like, you can say that such a temperate environment really consists of four environments: Spring, Summer, Winter and Autumn. Species that are adapted to this climate have a strategy for dealing with a specific set of seasonal changes in temperature, food availability, etc. That is, they are adapted to the seasonal shifts from one environment to the other. Migratory animals can move from one environment to another similar environment with the seasons.
But the world consists of a very large number of geographically-connected, seasonally-changing environments. It can take millions of years for a species in one environment to move into a different adjacent environment, e.g., for weasels to become river otters, or river otters to become sea otters.
Yet humans moved into all of these environments in just a few tens of thousands of years. This is because our mental faculties allow us to learn the rules of each new environment within a human lifetime.
The meta-environment (the ensemble of variable environments) already exists, so it has a selecting efficacy now. If I can adapt better to any unspecified environment, I have an advantage today (as long as new environments are being discovered or invented).
Godfrey-Smith wrote a book about this general topic, not sure if you have seen it, Complexity and the function of mind in nature. I haven't read it, but know he is a good philosopher, he worked with Philip Kitcher at UCSD.
But the world consists of a very large number of geographically-connected, seasonally-changing environments. It can take millions of years for a species in one environment to move into a different adjacent environment, e.g., for weasels to become river otters, or river otters to become sea otters.
Yet humans moved into all of these environments in just a few tens of thousands of years. This is because our mental faculties allow us to learn the rules of each new environment within a human lifetime.
Again you argue, when it comes to humans, as if evolution has foresight. Yes, humans can adapt to a new environment within a human lifetime, which gives them an advantage, but how did they achieve that capability in the first place?
Al: plasticity can be selected for too. If a population of animals enters a different niche, then those that are able to adaptively change their behavior (i.e., learn) will have a selective advantage. For animals that move around a lot, there would be selection for plasticity of the brain, for the ability of the brain to be shaped by its context, to learn (as opposed to fixed action patterns).
The cortex is partly an organ of plasticity. Humans are the sine qua non of adaptability (well, within the vertebrate world anyway).
That's the story anyway. No foresight needed, just selection for creatures with brains that learn better.
We could apply similar logic at the phylogenetic level, such as bacteria selected for faster evolution (as they presumably don't learn within a lifetime).
"No foresight needed" is the claim, but the problem is that when you start getting into these preadaptations, foresight starts to become strongly implied. This remains even if preadaptations are called "exadaptations" in order to try and shave off the teleological implications.
The history of evolution is loaded with deep homologies, preadaptations, convergences, and more across the board. In other words, it's starting to look tremendously teleological, even foresighted.
anon i think you don't get it. sorry. it's standard natural selection.
ok done with blog comments for a while need to write.
anon i think you don't get it. sorry. it's standard natural selection.
Waving the "natural selection" magic wand does diddly here, because selection on its own is entirely compatible with what I said. Teleology, direction, foresight and and purpose still is strongly implied specifically by what we see in evolutionary history and development.
Pre-adaptation, deep homology, convergence and more are making the evolutionary history which includes "standard natural selection" look teleological. Sorry, I know evolution is supposed to be atheist property. It's not anymore. ;)
Fight the Status Quo and its Natural Selection exclusivity! See why here: http://nature-sucks.blogspot.com/
Post a Comment