In this blog post I shall be analysing the notion of the self from a naturalistic point of view. The issue is a complex one and I shall end by highlighting the limitations (of which I am aware) that my approach suffers from.
I shall start with some ground rules:
1) Any reasonable definition must not reference any special quality of our universe. A definition of the self should not be something that only applies to universes like our own.
2) The definition should give the "correct" answers when applied to people in everyday situations. Without this restriction it would be questionable if the use of the word "self" was justified.
So now let us look at some simple possible definitions and my criticisms of them:
A) The self is simply a collection of atoms moving according to the laws of physics.
This definition is often quoted as the materialist position. There may be people who do advocate it. However, it certainly is not implied by materialism and it is unsatisfactory. This self hood would then apply to all collections of atoms. Without a method to decide which collections to take seriously we aren't much better off.
More seriously this definition would not meet (2) as the atoms in people are recycled quite quickly. The you of today contains very few of the same atoms as the you of a year ago.
B) The self is a lump of matter that has some coherence to it. New atoms may join it and old atoms leave it but like a football team it can be defined never the less by relabelling atoms as they become part of the greater lump.
This definition is more refined. It would manage to identify the movements of many people through their lives. However, consider case of Siamese twins attached at the pelvis. If the two are genetically identical who is to say when one twin ends and the other begins. If what we are is a coherent lump of matter then we must treat Siamese twins as one individual. Most people would acknowledge though that Siamese twins can and do have independent selves.
Note that we shouldn't rule out the possibility that there are (or could be) cases where selves do not come in neat packages but can blur into each other, split apart or fuse together.
C) The self is the pattern of electrical firing of neurons within the body.
This definition is often considered the 'scientific' notion of selfhood. It is a notion that makes sense of almost every everyday scenario. However, it is still flawed (as are A & B) in that it depends fundamentally on the physics of our universe and is totally inapplicable to many other possible universes (hence failing criterion (1)).
D) The self is a pattern of interactions between entities where the entities engage in greater exchange of information between each other than they do with the outside world.
This notion keeps the advantages of C whilst being able to deal with almost any possible universe. It does, however, allow for the existence of non-living selves and vegetative selves. This isn't necessarily a problem if we weigh the different selves by their computational capacity. That way would get us a more familiar emphasis on those individuals with more significant selves.
Problems with my approach:
P1) How is the term entity to be understood? Neurons are certainly entities. Is a hurricane an entity? Is a current in the ocean an entity?
P2) How do we measure the computational complexity of an assemblage of entities in a reasonable manner?
P3) How do we use such a definition to track the motivations, thought processes and decisions of a 'self'?
I do actually think that these problems can be solved and a general and simple notion of the self defined.
Q1: What problems do you see with this approach?
Q2: Are the criteria I use correct? (and have I applied them correctly?)
Q3: What criteria should be added to them?