Eliezer has posed some fascinating puzzles, and I hope that people will take the time to read them and try to give answers.
> Let's suppose that we record the tape, states, position, and actions
> of a conscious Turing machine, and play it back. Is the playback
> conscious? Probably not, because there's no causality, right? And if
> the playback was generated by quantum randomness, at some extreme level
> of improbability, then the conscious being "recorded" doesn't exist
> yet, right?
One aspect of most of these puzzles is that they focus on the issue of playback. I've seen several other paradoxes which also are based on this idea.
Are playbacks of conscious calculations themselves conscious?
My approach to this problem is to look at it from the point of view of the (simulated?) consciousness itself. Suppose, right now, I am aware that I am running in a computer of some sort, and that the computer's state information will be recorded and then played back using the various scenarios that Eliezer suggests.
To ask whether the playback is conscious is to ask, from my perspective, might I be experiencing a playback right now? Is this "instance" of my consciousness the result of a dynamic, causally-determined calculation, or the result of a passive playback?
The problem is, I can't give a meaningful answer to the question. Any answer I could give about what larger external environment I might be a part of cannot be correct in most cases. Even if my consciousness only exists during a causally based run of a computer, it could still be run multiple times. Each time, I will behave in exactly the same way. (I am assuming that there are not uncomputable elements necessary for consciousness.) Each time, I will think exactly the same thing. If I think to myself, "*this* instance of my consciousness is happening on the third run of the experiment", I will think that every time. It has to be wrong most of the time.
What, then, can I reasonably believe about the circumstances which surround this instantiation of myself? As I see it, I can't conclude anything specific about them. I might be run one time, I might be run 1000 times, and it will feel exactly the same. My "inner" time and sense of conciousness has no relation whatsoever to outer time and place. I might be run one million times slower than real time, or one million times faster. My brain could be simulated by a galaxy-wide computer with one processing element per star system, with each thought taking eons. None of this will be perceptible to me.
>From my own perceptions and senses, I can't conclude anything about the
external world. I can't tell how many times I am instantiated; I can't tell which particular systems instantiate me.
In particular, I can't tell whether a replay is conscious. It might be, or it might not be. It would make no perceptible difference to me.
If even I, the inner conscious entity, can't tell whether a replay is conscious, is it really a meaningful question?
Throughout the philosophy of consciousness, we continually face a dichotomy between the objective and the subjective. We try to link them; computationalism and, more generlly, materialism, try to say that there are objective facts corresponding to the subjective facts of consciousness.
But in this case, there is no subjective fact to begin with. There is no subjective meaning to the question of whether a replay is conscious. Therefore, I'd suggest, we can say that there is no objective meaning to the matter, either.
Simply put, it is meaningless to ask whether a replay is conscious; it is meaningless even to ask whether a second run of a conscious program further instantiates the consciousness. There is no subjective fact of the matter to resolve, and so there is no objective problem which can be solved, either.