Does "identical" mean "one"?

Eugene Leitl (eugene@liposome.genebee.msu.su)
Sat, 18 Jul 1998 19:15:25 +0400 (MSD)

Harvey Newstrom writes:
> During the upload discussion, many people claimed that if two objects
> were identical except for location, that they were the same object. Not
> just two of the same type of object, but literally the same object.
> They would be counted as one object. The claim applied to an original
> and an exact duplicate upload in the same room. Some claim that there
> is only one person, one body, one brain, one physical entity.

Two systems in the same quantum state are indistinguishable. You can find the outline of a proof for that in the appendix of Tipler's 'The Physics of Immortality'. I don't understand which 'counting' you refer to, but that proof says there cannot be any measurement procedure to distinguish one object from the other. Finis.

Of course this is pretty useless, for no macroscopic objects occupy the same quantum state. But we are talking about the user front end, right? If _I_, the user, cannot distinguish between an object, and a copy, then it is meaningless to make the distinction between them, obviously?

Addressing your position re discontinous (destructive) uploading. Current physics seems to forbid realtime in vivo scans with enough resolution without destroying the object. It would be ideal to have working incremental/gradual uploading before we croak, but it's availability is highly uncertain, because it is pretty difficult to do even with mechanosynthetically produced devices. So a destructive scan of a vitrified cerebrum ("digitizing neuroanatomy") and a computational model of it with sufficient precision is currently the best we can hope for. Lucky thing this buys you time, lots of it, because basically no deterioration occurs to tissue in glass matrix (let's not discuss the involved topic of suspension damage right now). So there is no original to compare for harvey-nature, so we have to devise some other metric for harveyness. So you can check some operation signatures: high-level ones (behavioural), fMRI/MEG/EEG/firing pattern fingerprint, etc. etc. Obviously the metric for harveyinity is not boolean, but fuzzy (reality is always messy, right?). You have a region, not a point in persona space, as you obviously lose information due to the suspension damage inflicted, have uncertainties due the limited emulation resolution, etc. If you could have recorded lots of additional information, as somehow tracked your cellular growth, and recorded anything you saw and did during your life from your perspective, you could use that as a source of constraints to narrow down the region in persona space labeled 'Harvey'. But I very much doubt we need _that_ much detail, because the changes you have to cope with during your daily life, especially towards the end of it, are (hopefully) so very much more drastic than the artefacts due to the uploading construction and operation process. You seem to value personal continuity so very much, but you dispense with it when you go to nonREM sleep, suffer a head concussion, total anaesthesia or a stopped bloodflow for several minutes. So this definition of personality is very fragile it gets violated habitually. The question of persona trajectories being divergent has been handled with, already, so I believe we can skip that.

Summa summarum I think your attitude towards preservation (or failure thereof) of persona in an discontinuous upload is not rational. I, too, have felt emotional discomfort with discontinuous uploads in the past. Being a question of faith (does it harvey-nature, or not?), it is best addressed from a point of faith; at least I have found it so.

ciao,
'gene