Re: Forward vs. Reverse Engineering

From: HARNAD Stevan (harnad@cogsci.soton.ac.uk)
Date: Mon May 13 1996 - 16:32:44 BST


> From: "Chatwin Judy" <JAC295@psy.soton.ac.uk>
> Date: Mon, 13 May 1996 14:03:59 GMT
>
> Secondly, ' What is reverse engineering?'
>
> Again I understand that the best way to prove how something works is
> to break it down and create one but I do not feel able to go beyond this.
> Any help would be greatly appreciated. Thanks Judy

Have a look at the skywriting thread of discussion on this topic,
http://cogsci.soton.ac.uk/~harnad/Hypermail/Explaining.Mind/0046.html
and also at Dan Dennett's paper of Reverse Engineering:
http://cogsci.soton.ac.uk/~harnad/Papers/Py104/dennett.eng.html

> From: Rowe Anna <ajr395@soton.ac.uk>
> Date: Mon, 4 Mar 1996 16:28:36 GMT
>
> Please could you tell me again what reverse and
> forward engineering are, and also what the difference is
> between behaviour and cognition.

Behaviour: What a person (or animal or machine) DOES and CAN DO:
its observable actions

Cognition: The processes going on in the head, both conscious and
unconscious that cause behaviour. Cognitive psychology is the field
that tries to EXPLAIN behaviour, to explain HOW we can do what we can
do.

Forward engineering is normal engineering. It builds devices that can
do certain useful things for us: bridges, furnaces, cars, planes. It is
forward, because WE build the devices (by applying the principles of
physics and previous engineering). We know HOW forward-engineered
devices work, because we designed and built them.

It's reverse engineering when the "devices" were already "designed" and
"built" by nature (by evolution) and we have to figure out how they
qork, how they can do what they can do. (A "device" is just a causal
system that can do something; in this sense, we too are devices.)

To reverse-engineer our behavioural capacity, we need a theory of
cognition: a theory of what structures and processes are capable of
generating our behaviour. To do this, we study behaviour and brain
function, and we build models that have some (eventually all) of our
behavioural capacity; models that can do what we can do. For if the
model can do it, than we know how it can be done.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:41 GMT