This is an attempt at a style of proof by contradiction, without academic charity to the possibility of other alternative explanations for said contradiction. Briefly, I summarise the argument made before offering my criticisms at the end.
~Summary~
Definitions:
p-zombies := a perfect physical recreation of a given human being except without inner experience ("qualia")
Assume:
A1) It is conceivable that p-zombies exist
A2) A p-zombies would perfectly mimic that human being's observable behavior, provided that the human and the p-zombie interacts with the same external world
A3) Anything that exists (or worth speculating over its existance ["a good theory']) has to be physically observable (create physical differences in the world they exist compared to the world in which they don't)
From (A1) and (A2)
A1, A2) Any observable, and therefore physical, evidence offered a given human being can give to defend the existence of their inner experience, can also be offered by a p-zombie of that human being. There are no physical differences between their worlds.
This is in tension with (A3). To resolve this, Elizier rejects (A1) and accepts (A2) and (A3)
~Criticisms~
Criticisms:
As mentioned by previous commentors, Zombies and Zombies Redacted is longer than it needs to be
The thought experiment suggests that, given p-zombies are conceivable, there is no physical evidence that can be offered to prove the existence of inner experience ("qualia"). This has been obfuscated by polemics against dualism at the expense of clarity.
It is not a given that p-zombies are concievable, merely between the conceivability of p-zombies, physicalism, and the real experience of inner qualia, one of these assumptions have to be mistaken.
It is not a given that qualia exists. A common argument against qualia is that what we think qualia is (the impossibility of capturing experience by simulacra or explanation) is an illusion generated by the limitations of language or thought. This seems very intuitive to me.
Physicalism is not a given either. For example, the framework on which we even talk about these subjects - assumes that such things as "causes" and "consequences", or at least associations between phenomena exists. It is not obvious to me that causality is physically observable, or that, in the style of p-zombie thoguht experiment, that there exists observable evidence that would prove we live in a world that causality exists (or doesn't)!
TLDR;
In mathematics, proof by contradiction works because every other assumption we assume are bedrock reliable. So if we introduce a new assumption, and that results in contradiction, the negation of that new assumption must be true. Even this is not perfect - when there is a contradiction in a theory, and it is not evident where the contradiction emerges, the whole theory may have to be scuttled so that mathematicians can rework something that is dependable, cf- Set Theory and Russel's Paradox.
In philosophy, proof by contradiction doesn't work the same way - as none of the assumptions in philosophy can be assumed to be bedrock. All assumptions eventually boil down to intuition, and so, by framing and reframing, can be more or less intuitive and have to be considered charitably. This analysis by Yudkowsky is fundamentally uncharitable to assumptions held by people with a different view to him, and so more or less begins with the conclusion in mind ("(A1) p-zombies are conceivable, (A3) For a thing to exist, it's existence or non-existence has to result in two physically observably different worlds").
It is perfectly legitimate to argue that (A3) is nonsense, and that a sensible undertanding of the world requires a priori knowledge such as consciousness or causal reasoning, and that in fact, humans use a priori knowledge efficaciously all the time.
This is an attempt at a style of proof by contradiction, without academic charity to the possibility of other alternative explanations for said contradiction. Briefly, I summarise the argument made before offering my criticisms at the end.
~Summary~
Definitions:
p-zombies := a perfect physical recreation of a given human being except without inner experience ("qualia")
Assume:
A1) It is conceivable that p-zombies exist
A2) A p-zombies would perfectly mimic that human being's observable behavior, provided that the human and the p-zombie interacts with the same external world
A3) Anything that exists (or worth speculating over its existance ["a good theory']) has to be physically observable (create physical differences in the world they exist compared to the world in which they don't)
From (A1) and (A2)
A1, A2) Any observable, and therefore physical, evidence offered a given human being can give to defend the existence of their inner experience, can also be offered by a p-zombie of that human being. There are no physical differences between their worlds.
This is in tension with (A3). To resolve this, Elizier rejects (A1) and accepts (A2) and (A3)
~Criticisms~
Criticisms:
TLDR;
In mathematics, proof by contradiction works because every other assumption we assume are bedrock reliable. So if we introduce a new assumption, and that results in contradiction, the negation of that new assumption must be true. Even this is not perfect - when there is a contradiction in a theory, and it is not evident where the contradiction emerges, the whole theory may have to be scuttled so that mathematicians can rework something that is dependable, cf- Set Theory and Russel's Paradox.
In philosophy, proof by contradiction doesn't work the same way - as none of the assumptions in philosophy can be assumed to be bedrock. All assumptions eventually boil down to intuition, and so, by framing and reframing, can be more or less intuitive and have to be considered charitably. This analysis by Yudkowsky is fundamentally uncharitable to assumptions held by people with a different view to him, and so more or less begins with the conclusion in mind ("(A1) p-zombies are conceivable, (A3) For a thing to exist, it's existence or non-existence has to result in two physically observably different worlds").
It is perfectly legitimate to argue that (A3) is nonsense, and that a sensible undertanding of the world requires a priori knowledge such as consciousness or causal reasoning, and that in fact, humans use a priori knowledge efficaciously all the time.