[Comp-neuro] Re: (3) A Matter of Religion

Harry Erwin harry.erwin at sunderland.ac.uk
Fri Aug 22 10:30:56 CEST 2008

On 21 Aug 2008, at 22:43, James Schwaber wrote:

> (3) A Matter of Religion
> I am really struck by the forceful and repeated confessions of faith  
> in this thread: we are materialists with every fiber of our body, we  
> are certain the material world is well understood and the brain is  
> part of it, reductionistic accounts are in principle complete,  
> determinism makes freewill an illusion, and so on. My experience is  
> that this faith is universal among us neuroscientists, and to a  
> somewhat lesser extent biologists in general. We want to believe we  
> are real scientists, just like physicists, and will prove it by  
> mechanistic reduction of behavior and brain to physical-chemical  
> levels.
> It is my interesting experience that this faith is often not shared  
> by physical scientists and philosophers of science - that this kind  
> of hard reductionism/determinism may be passé, not considered a  
> workable approach.  That is, a bottom-up, seamlessly mechanistic- 
> deterministic account of all of nature is typically not considered a  
> reasonable goal of science, e.g. given the limits we have hit in  
> 20th century physics. In this context, a local account of some  
> mental or behavioral or neuronal function in its own terms would be  
> a success to the extent it was predictive-useful, but never taken as  
> complete or eternally true.
> Obviously this is a matter of faith, of 'religion'. I wonder if the  
> faith of neuroscientists is inhibitory to how we consider how  
> brain's work?
> _______________________________________________

My PhD research was in bat behaviour—bat echolocation and target  
capture—where I studied how bats use sound to search for and capture  
their food. Echolocation is a low data rate sense that requires a lot  
of processing to make sense of stimuli, so a bat has to efficiently  
ask quite specific questions of its acoustic environment. The process  
of sound perception in bats and other animals is complex, using  
internal neural representations only indirectly related to the sounds  
the bat originally heard.

Study of how bats do this has led to some interesting conclusions. For  
instance, bats seem to live in an internal model of their environment.  
This was first noticed by Möhres and Öttingen-Spielberg when working  
with a bat that was accustomed to roosting in a cage in a room. The  
researchers rotated the cage and eventually removed it, and noted that  
the bat continued to behave in detail as if the cage were in its  
original position.

The bat lives in its world model and uses it to choose its behaviour.  
Until forced to reorient, the bat continues to perceive the cage in  
its original position. These points led me to emphasize cognitive over  
behavioural models and have channelled my philosophical approach to  

I am not a theoretical reductionist, which might be assumed from my  
being a materialist. Reductionism claims that higher-level systems can  
always be explained—that is ‘modelled’—in terms of lower-level  
systems. For example, that the mind is reducible to neural network  
activity; neural network activity can be reduced to biology; biology  
to chemistry, and chemistry to quantum physics. The problem is that  
you can’t do the modelling. The number of steps from each lower-level  
model to the next higher-level is essentially infinite—for example, to  
calculate the folding of a medium-sized protein from quantum mechanics  
requires more computational power than the universe can possibly  
contain. This is what Ian Stewart and Jack Cohen call ‘ant country’.  
So to understand a higher-level system like the mind, you must develop  
a theory—anchored in careful experiments—at the same level as the  
mind. Anything restricted to reductionism is certainly infeasible and  
likely to be impossible. Certainly, the protein folds, but you can’t  
model it. I do the modelling, but it is 'cum grano salis.'

I am also an enactive constructivist—I study embodied cognition (minds  
in bodies)—not a behaviourist. Again, we need to unpack the  
terminology. According to Wikipedia, enactive (or embodied) cognitive  
scientists emphasize embodiment and action over passive internal  
representations. The individual grows into the world—he has no a- 
priori representation of the world, and has no core self separate from  
the world. I reject behaviourism—the proposition that behaviour can be  
explained without assuming internal mental states—and I also reject  
Cartesian dualism—the proposition that mental phenomena are in some  
respects non-physical. Again, my evidence is based on that  
experimental work with bats.

So to summarise, I am an experimental scientist, a materialist but not  
a reductionist, and an enactive cognitive scientist, but not a  
behavioural scientist or a dualist.

> --

"an academic who listens to pleas of convenience before publishing his  
research risks calling into doubt the whole of his determination to  
find the truth." (Russell 1993)
Harry Erwin

More information about the Comp-neuro mailing list