Tuesday, December 13, 2005

Durability of the convenient construct of self

Self may be a convenient construct for intelligence in human packages...how could this change in the future?

In the case of uploading, where a human is copied to a digital substrate (via digital capture of personality, value system, experience history, memory, thoughts, etc. and/or a full copy and simulation of all neurons and other brain functions), would the concept of self be different?

Self has at least two dimensions: corporeal and mental. First the distinction between my body and other objects. With my mind, I can direct the movement of my body. Without employing my body (or more advanced metaphysics) I cannot direct the movement of other objects. This physical agent is "my body" or "my self."

Second, the distinction between my mind and the minds of others. The tableau of my thinking is not overtly visible or perceivable to others. From communication with others, I have the perception that others have thoughts as well, and that they are not explicit to me.

In short, a working definition of self is the perception and control (agency) of the entity that I can direct vs. my surroundings.

In the case of an uploaded human copied to any other physical body with action, self-direction and movement capability, it would still be useful to have the concept of the physical agent self and the mental self, although in a highly functional partial self, a partial set of mental capabilities may be appropriate (an interesting ethical question). If the uploaded human is residing exclusively in digital storage files and does not have physical agency, the concept of the physical agent self would not be useful but the concept of the mental self may still be useful. The mental self concept may evolve in many directions, including the idea of different permissioned access tiers to the mind/processing core and other areas.

What would it be like to be an identity in digital storage without physical agency, particularly knowing that previously "you" had had physical agency? Lots of aspects of physical being can presumably be simulated, the usual pain, pleasure, etc. sensory experiences we are familiar with and probably some new ones. The more important aspect of physical agency relates to survival and ability to act on the surrounding environment for survival. The digital copy would need assurance and control over aspects of survival. This could be a complicated process but at the surface seems to mainly entail the access to a variety of power supplies, computing resources and backups.

Is a saved digital copy of a human alive? Probably not unless its being run. When a digital copy of a human is being "run," how is it known/shown that it is "alive?" As with any new area, definitions will be important. What is alive or not for a running simulation of a human is an interesting topic to be covered later. Here, it is assumed that there are some cases where a simulation is deemed to be alive and others where it is not.

There do not appear to be any ways in which a human simulation could be alive without having self-awareness in the current definition of being able to distinguish between itself and its environment at least mentally and in some cases physically. This analysis does point up the possibility of adapted or more rigorous definitions of being 'alive.'

So the concept of the corporeal and mental self may still be useful in the future but may likely be expanded upon and may exist as one of multiple metaphors for describing the agency of thought and action.