Robin Hanson’s book The Age of Em is a fascinating and important book applying economic theory to analyzing a dramatic technological possibility in the future: the possibility that before we truly understand the human mind and before artificial intelligence built up from first principles can match human intelligence, we might be able to make functional, faithful, software emulations of the workings of particular human brains. Robin gives these emulations the affectionate nickname “ems.”
There is much too much in The Age of Em to talk about to fit into a single blog post. Today I will address only one issue: if a copy of your own mind is made into an em, what is the experience like once that em is started up?
At one level the answer is easy. The em that is a copy of you will act just like you and say and do the same kinds of things you would say and do if you suddenly found yourself a bit of software in a virtual world but otherwise were yourself.
As I say in “On the Effability of the Ineffable,” the mysteries we talk about are not really ineffable, because we are talking about them! (Ineffable: Beyond expression; indescribable or unspeakable.) An em that is a copy of your mind will talk about things in exactly the way you do. And what the em says and does, or readings of its brain emulating activity will be the only access anyone else has to what it is like to be that em, just as what you say and so and readings of your brain activity are the only access anyone else has to what it is like to be you.
There are two obviously important cases. (I’d be glad to hear about others.) One is if dualism is true. If there is a spirit or soul inside of you that does the experiencing, then what the experience of an em is like depends on whether ems get to have a spirit or soul or not. That then depends on facts about what the gods or natural processes that grant or produce spirits or souls do.
The other obviously important case is if our experience comes from the interactions of particles and fields that are either now or someday will be known to physics, none of which individually has any more of spirit or soul than any other particle or field. In that case, it is hard to see why ems would not experience things. Since—if they are truly faithful emulations—they would speak and act as if they are experience things at exactly the same depth as human beings, it is also hard to see why they wouldn’t be experiencing things in the same way as human beings.
But is an em that is an emulation of your brain more like another human being who is eerily like you, or more like a you? On the assumption that experience comes from particles and fields known to physics (or of the same sort as those known to physics now), and that the emulation is truly faithful, there is nothing hidden. An em that is a copy of you will feel that it is a you. Of course, if you consented to the copying process, an em that is a copy of you will have that memory, which is likely to make it aware that there is now more than one of you. But that does NOT make it not-you.
You might object that the lack of physical continuity makes the em copy of you not-you. But our sense of physical continuity with our past selves is largely an illusion. There is substantial turnover in the particular particles in us. Similarity of memory—memory now being a superset of memory earlier, minus some forgetting—is the main thing that makes me think I am the same person as a particular human being earlier in time.
Noah Smith’s religion guest post “You Are Already in the Afterlife” makes this point nicely: we continue to become different than we were before, yet consider ourselves the same person. ]
Why should I consider myself the same person as the Miles Kimball twenty years ago, who in many ways was very different in characteristics, but think an em copied from me right now and started up a minute from now, who is much more similar to me, is a different person?
Just as important, if I seem to have a continuity of conscious experience despite the fact that the particles making me up keep changing, there is no reason to deny that there is a continuity of conscious experience from me to the em that is a copy of me. The weird thing is that with such copying, there would be several different continuities of conscious experience. One line of conscious experience that ended up outside any computer and another line of conscious experience that ended up inside a computer. (Note that distance in time is not a big issue: we are used to what counts as a “continuity of consciousness” having a sleep state intervening. Being in “suspended animation” as a recorded computer state is less of an interruption than a sleep period, because a sleep period changes the state more.)
In the technological environment Robin and I are considering, after the copying event these two lines of conscious experience are isolated from one another as any two human beings are mentally isolated from one another. But these two consciousnesses that don’t have the same experience after the split are both me, with a full experience of continuity of consciousness from the past me. If one of these consciousnesses ends permanently, then one me ends but the other me continues. It is possible to both die and not die.
The fact that there can be many lines of subjectively continuous consciousness that are all me may seem strange, but it may be happening all the time anyway given the implication of quantum equations taken at face value that all kinds of quantum possibilities all happen. (This is the “Many-Worlds Interpretation of Quantum Mechanics.”) Indeed, although one type of splitting originates in macroscopic events and the other type of splitting originates in picoscopic events or smaller, they have the same qualitative description as things that have all the same observable consequences of real splitting with all paths continuing, even if one wants to deny that it is real splitting with all paths continuing. There is complete observational equivalence.
Note that, along the lines of what I said in “On the Effability of the Ineffable,” if there is any way I can even think to myself something ineffable, that is enough brain activity that it could, in principle, be revealed to someone else. So—short of hardcore dualism with a spirit or soul—it is hard to see what distinctive magic there could be to the “true me” to distinguish it as me and the copy of me as not-me. Is being composed largely of water really the magic that makes it the real me?
And my problem with hardcore dualism is this:
If a spirit or soul influences any of my decisions, then it has enough effect on particles in the brain that it should be detectable by physics with the sensitivity of instruments we have now.
If a spirit or soul is affected by the body but does not itself have any effect on the body (Epiphenomenalism), then it is not through any causality from that spirit or soul the spirit or soul that we talk about because it has no causal pathway to move our mouths. God might make our bodies so they talk about our epiphenominal spirits or souls. But our spirits or souls in this case are not talking about themselves on their own behalf.
Conclusion. The bottom line is that I think an emulation of my brain would have a genuine continuity of consciousness with me as me. There would be a weirdness of there being more than one of me, or one me that ends and another me that continues, but that would be “just the way it is.”
Don’t miss these related posts: