The Null Device

How to identify zombies

Consciousness researcher Christof Koch claims that we are almost zombies, or rather, that the vast majority of our lives are spent unconsciously, on autopilot: (via bOING bOING)
You drive to work on autopilot, move your eyes, brush your teeth, tie your shoelaces, talk, and all the other myriad chores that constitute daily life. Indeed, he says, Any sufficiently well-rehearsed activity is best performed without conscious, deliberate thought. Reflecting too much about any one action is likely to interfere with its seamless execution.
Given the range and effectiveness of these zombie agents, Koch believes the great mystery is why we are not complete zombies. Or to put it another way: What purpose does consciousness serve? Why does it exist at all?

Consciousness, Koch argues, is a local phenomenon, residing in a specific part of the brain, and serving a specific purpose, and not an emergent property of a complex system, as some have claimed:

In principle, Koch says, there is no reason why consciousness is necessary to life. With enough input sensors and output effectors, it is conceivable that A zombie could pretty much do anything. But since every zombie behavior must be hard-wired, the more situations it must respond to, the more complex its internal mechanism must become. Instead, Evolution has chosen a different path, synthesizing a much more powerful and flexible system that we call consciousness. The main function of this innovation, he and Crick propose, is to enable organisms to deal rapidly with unexpected events and to plan for the future. As Koch likes to say, consciousness puts us online, allowing us to override our instinctual offline programming.

Which lends itself to a possible test for detecting consciousness, and thus differentiating between humans and zombies and other unconscious life:

Since zombie agents operate purely according to preprogrammed rules, a zombie would have no need for short-term memory, and hence Koch believes the absence of this feature would serve as an indication that consciousness was also missing. Consider the following situation: You see an outstretched hand, but instead of shaking it immediately, which instinct would dictate, you are required to close your eyes and wait several seconds before doing so. Koch and Crick suspect that without a short-term memory, a zombie could not do this task, or any other in which an artificial delay was imposed between an input and the associated motor output.
For the moment, he is concentrating not on humans but on biologys most common test subject, the mouse. He and his colleagues are trying to develop a mouse model of consciousness, a rigorous way of determining if and when a mouse is aware. Over the past decade, biologists have learned how to turn individual genes on and off in the developing rodent fetus. With a mouse model of consciousness, Koch could begin to explore what genes are essential for this phenomenon. One question he would like to pursue is whether it is possible to genetically engineer an animal without conscious awareness -- a zombie mouse.

There are 3 comments on "How to identify zombies":

Posted by: Matt http://www.godhatesfags.com Tue Nov 25 01:26:26 2003

If you include subconscious and automatic things like breathing, not falling over, sleeping, blinking etc., I think the figure is conservative for most people.

Posted by: kenny http:// Wed Nov 26 00:49:31 2003

cosma shalizi had something to say about that:

http://cscs.umich.edu/~crshalizi/weblog/archives/000145.html

Posted by: mitch http:// Wed Nov 26 09:42:24 2003

As Cosma points out, any standard cognitive operation can be performed by some finite-state machine. If the possession of consciousness is supposed to be nothing but the possession of a particular cognitive ability, then presumably possession of consciousness is nothing but possession of the appropriate finite-state machine (or even, nothing but *being* the FSM in question). I find it odd that Cosma agrees with Dennett that "qualia are a red herring", but then has problems with this sort of functionalist theory of consciousness. (I have a problem with it too, but then I'm a quantum-mind zealot who thinks consciousness and intelligence are quite different things.) Couldn't Koch just bite the bullet and say, yes, I think that anything with the appropriate computational architecture is conscious? If one rejects "qualia", what reason would one have for objecting to such a theory?