The First Person Computationalist Indeterminacy

Section: Philosophy Essays

The notion of first person, or subjective, computationalist indeterminacy is a notion which makes possible to explain how, in a context of purely third person (objective) determinacy, experiments can be designed exhibiting, from the points of view of the subjects involved, a necessary lack of apparent determinate outcome (Marchal 1988, 1998, 2004, Barton 2005).



The computationalist hypothesis (comp) in the cognitive science can be shortly described as the hypothesis that there is no privileged part of the (generalized) body with respect to digitally emulable possible prostheses. Specifically it says that for any part of our body there exist a level of description such that those parts can be said functional, and thus substitutable by functionally equivalent prosthesis, so that the subject experimenting that substitution will not experience any changes. It can be shown that if such a level of functionality exists, then that level is unknowable, and that is why we are referring to a generalized body. By definition the generalized body is whatever portion of reality we have to Turing emulate for consciousness to be recovered. Usually it is thought that the physiological brain at the level of neurons or at the biochemical level is enough, but in some application such an hypothesis can easily be eliminated. In case such level does not exist the assumption of computationalism failed, by definition. Computationalist is a modern form of the old mechanist philosophy.



First person and third person view. Let us first define the notion of first person and third person points of view. We don't have to build a perfect definitive definition, we have to build only a definition genuine for the notion of first person indeterminacy we try to convey here. Let us assume the comp hypothesis. It follows that it is in principle possible to survive, in the usual mundane sense of the term, to a substitution of the body/brain for an artificial digital body/brain. It is then possible to use, in principle, classical teleportation, as a means for personal locomotion. The fact that such an experiment is hardly technically feasible today is not relevant for the reasoning. Let us consider some subject using classical teleportation for going from place A to B.

In a classical teleportation experiment, a subject is scanned at some place A, at the right (serendipitously chosen) level of substitution, and then destroyed, and reconstituted, at some place B, from the information which has been scanned at A.



We define the first person discourse by the content of a diary, containing the result of repeated self-localization and local time, in space and time, experiment, with the understanding that the subject takes with him both the diary, and the clock. The third person discourse, in a simple teleportation experiment, is defined by the content of the diary of some external observer, who does not participate to the annihilation/destruction trip. By abuse of language I will talk on the 1-diaries, 3-diaries, 1-clock, 3-clock, 1-discourse, etc.



In the case of a simple use, locomotion motivated say, of (classical) teleportation, the contents of 1 and 3 diaries match, taking into account the permutation of "he/she" into "I" from the 3-diary in the 1-diary, with the usual indexical referents. Without introducing any delay, or better, by making abstraction of any possible delay due to annihilation and reconstitution time, it can be said that the 1 and 3-experiences match. With delay, the 1-experience still matches with the 1-experience without delay. It means the first person is not, and obviously cannot be aware of the delay, without comparing his/her clock with some external clock or without having external similar clues. Indeed, if we introduce some delays in the reconstitution, the 3-clock obviously differentiates, separating the content of the 1 and 3-discourses where the first person discourses does not. This is a key point for some application described below, but here it is used only to illustrate the 1-3 person discourses distinction. Such distinction is fundamental in the experiments (third person describable, and experiences (first person lived and related in the 1-diaries) of self-duplicationa and self-multiplication. Fusion of different individualities through amnesia is plausibly conceivable, but out of topics here.



Duplication. So, after the scanning and annihilation in A, nothing prevents us to duplicate the information retrieved and to reconstitute the 3-person in two different but similar (not distinguishable by the subject) closed rooms. Both rooms contains an envelop with a paper inside on which 0 and 1 are written respectively. With the comp assumption, the candidate survives, by the usual locality assumption (no magical influence at a distance) related to the comp assumption. Suppose that the subject, which already trust the comp assumption and the correctness of the chosen substitution level (and the working of the scanning-annihilation and reconstitution device), is warned in advance that such a duplication will occur. Let us ask the subject where he/she will find him/herself after such an experiment. Put in this way the question is ambiguous. The subject can talk about him/herself in the 3-person way, and answer, correctly (assuming the comp hyp) that he will be in the two rooms at once. So let us ask him more precisely: "where do you expect to feel yourself being after the experiment. Assuming comp, and assuming for the sake of the reasoning some default hypotheses like the absence of any other reconstitutions elsewhere at any time of the scanned state, the subject can expect to feel with certainty to be, after the experiment in some room in front of the envelop. So he/she can expect now, before the experiment, to be uncertain of the outcome of the 0-1 envelop measurement. We take also for granted that if a subject is certain that he/she will be uncertain about some outcome of a future experiment (like looking if 0 or 1 is written on the paper inside the envelop) then he is already uncertainty now about that future outcome verification experience, despite his/her knowing that both its «descendants» will (correctly) attribute its present mind state its its memory past (1-diaries). So the subject is uncertain now of which envelop he will be confronted with, and so the subject, from a first person point of view, is uncertain now which reconstitution, in which room, he will live in his/her future. This is enough to illustrate that a self-duplication experiment, which is entirely conceivable in the computationalist purely 3-person, objective, deterministic frame, entails the existence of experiment where a subject is unable to predict with certainty its personal experience. It is arguable that the subject, in a self-duplication experiment is maximally uncertain, or totally ignorant, about what his personal experience (zero or one) will consist in, and that this complete uncertainty comes from the complete third person determinism inherited from the computationalist assumption, which indeed ensure the numerical identity of the relevant reconstituted body.

No «omniscient» being can predict, to the subject before the experiment, if he/she will feel as if seing zero or feel as if seing one, after opening the envelop. The omniscient being cannot say "you will see both", because this will be refuted by each descendant, which will knowingly put different numbers, zero or one, in their personal observation diary. It cannot say "you will see zero", because this will be refute by the reconstituted subject in the room with the envelop containing the paper with «one» written on it, and this subject would know that the omniscient being is wrong which an omniscient being cannot be.

Betting procedure, and decision theoretic approaches can be introduced through the notion of first person plural, and a locally sharable indeterminacy appears in populations, when population of individuals are copied and multiplied.

First person, or subjective, indeterminacy can be approached in the context of inference inductive computer science, build on the second recursion theorem in computer science. It is enough to substitute the subject by virtual, actually mathematical, robots. Simple combinatorial reasoning ensure most on the sequence of zero and one will be Kolmogorov incompressible, and machines can prove and infer that from the assumption that they are consistent machines. See references below for the use of self-reference logic (the logic of provability and consistency) in that context.

Bibliography:

- The reduction of the computationalist "mind-body problem" to a computer science theoretic «body» problem (Marchal 1988, 1998, 2004, 2005).

- Clarifying Everett formulation of Quantum Mechanics (Marchal 1988, Barton 2005, Wallace 2006)

 

 

Words: 1390