Probability Theory and Multiple Conscious Lifetimes W. J. Cocke Steward Observatory University of Arizona Tucson AZ 85721 email:
[email protected] 6 July 2009 Abstract— Having a conscious, first-person awareness carries information. I use simple probability arguments to show that the chances are large that I have experienced more than one conscious lifetime. An elementary lottery model is worked out in the context of conditional probability methods and the Copernican principle. I argue that the methods of physics are in principle not able to deal with a first-person subjective awareness. I. Introduction It seems unlikely that probability theory would have anything useful or plausible to say about a subject like consciousness and its persistence over multiple lifetimes. There is, however, information content in the fact that I have a first-person conscious awareness at this point in time. Some of the probability arguments that use this information involve conditional probabilities, and I introduce these later in the paper (Section V). The core of my argument has two parts. Part (a) — My being alive now is improbable, in the sense that my life could have taken place at any time over the past several million years. Or, of course, it could take place at some time in the future, perhaps the distant future. This is like being a big winner in a lottery. I am very lucky to be here now, as a winner of this lottery. The critical point is the current now-ness of my win. I am considering the probabilities from my own first-person standpoint, but in the Appendix I discuss a real break-in incident that contrasts the first-person standpoint with the standpoint of an outside observer or third person. These points of view must be equivalent, if both “persons” have the same data. Part (b) — The second step is to ask why I have won the current biological lottery. Was there a bias or some kind of arrangement that made a win more likely at this time? For example, I might have acquired a large number of tickets. 1
The thesis of this paper depends on the assumption that it makes sense to ask whether or not my conscious awareness appears only once in all of time, from the Big Bang to the end of the universe. This is certainly what people mean when they say: “You only live once.” It follows that the converse statement would also make sense, so that we can inquire about my having multiple lifetimes, as in the title of this paper. My life — my conscious awareness — is certainly happening now. I have won the lottery, but there are lots of current winners, and indeed this has been going on for millions of years. The reader may well ask, “What is so special about you at this particular time?” The really special thing is the coincidence between my awareness and geological (physical) time now. It is the point of this lottery that winning the current drawing means existing now. People who are already dead have won in past drawings. Certainly, the actual number of current winners is irrelevant — the probabilities are the same, whether I am merely one winner out of 10 billion or the only winner. But as a current lottery winner, I have to deal with the fact that there might have been some kind of bias or cheating involved. Using probability theory to deal with these questions in entirely appropriate. It would be disingenuous to say that there is nothing improbable happening, since the laws of physics have actually caused my awareness to appear at this time. After all, we know much less about the physics of subjective consciousness than about the physics of hands in a game of cards, and yet we are perfectly content to make statements like: “I was dealt a royal flush in a poker game last night. This was very unlikely!” The analysis in this paper depends on knowing simply that my lottery win coincides with time now. I know nothing about the results of the many past drawings, only the results of this particular drawing. In other words, a particular drawing was chosen for scrutiny — we are at time now — and the result was a win for me. This procedure contrasts with picking out an improbable win from a long sequence of drawings. That is a different problem and is irrelevant to the multi-lifetime issue. It is occasionally said that statistics and probability should be a science of last resort, and that one should use it only if the results are reasonable and not counter-intuitive. In considering the question of multiple lifetimes, there seems to be no way to make a definitive judgment. Many people consider the idea of reincarnation to be quite reasonable, and certainly any number of people — including many Westerners — believe in the survival of consciousness after death. In Section III, I continue the lottery model in more detail and consider the idea that 2
bias has occurred in the lottery. We will see in Section V that a non-zero probability that bias has occurred can enhance the probability that bias occurred in my own lottery win. II. The Remarkable Fact of First-Person, Subjective Awareness The hypothesis that the lottery might be biased parallels the idea that my own conscious awareness has made more than one appearance throughout the ages — hence “multiple lifetimes.” And it may be that we must consider more species than just homo sapiens. My consciousness could just as well have been centered in the brain of any mammal, or perhaps any vertebrate. Can anyone doubt that dogs and cats, for example, share with us certain key elements of this conscious awareness? In what follows, I sometimes refer to the Copernican principle, which is the idea that there should be nothing special or remarkable about one’s particular location in space or time. Gott (1993, 2001) states that this principle has been very successful in puzzling out our relation to the rest of the universe. Some of the reasoning in the present paper is an appeal to this principle. As far as my own history is concerned, there should be nothing special about this particular point in time. This paper grew out of my feeling that there is something very significant in the fact that I happen to be alive and conscious at this particular moment. If I am to have only one lifetime, then the coincidence between physical or geological time now and my being alive and aware would seem really improbable; I should have been dead for many millennia or simply not yet born. This is a restatement of Part (a) in the Introduction. In terms of the Copernican principle, my being alive at time tnow is very remarkable, and it is an interesting point that physical science has nothing to say about what “now” might mean. How does it happen that geological time is now, say, tnow = 4.6 billion years since the formation of the solar system? Certainly, one can determine this fact from studying the concentrations of various isotopes, but the point is this: there seems to be no basis in physical science for concluding that there should actually be a time tnow , let alone what that time might be. See Nagel (1986, pg. 57n) and Dummett (1978). Rudolf Carnap (1963) reported a conversation with Albert Einstein, who told him that physics had nothing credible to say about the distinction between past and future, a distinction that we humans find very important. There is a strong sense that the existence of our individual first-person awarenesses is extremely peculiar. This sense runs through some philosophical discussions of the “hard 3
problem” of consciousness. Nagel (1986) and Himma (2005) discuss this question clearly. It is hard to improve on this statement by Nagel (1986): It isn’t easy to absorb the fact that I am contained in the world at all. It seems outlandish that the centerless universe, in all its spatiotemporal immensity, should have produced me, of all people and produced me by producing TN [i.e., Thomas Nagel]. There was no such thing as me for ages, but with the formation of a particular physical organism at a particular place and time, suddenly there is me, for as long as the organism survives. In the objective flow of the cosmos this subjectively (to me!) stupendous event produces hardly a ripple. How can the existence of one member of the species have this remarkable consequence? I have observed that nearly everyone is very accustomed to his or her own first-person, subjective awareness — so much so that there is no sense of how strange and extraordinary this individual awareness is. It is easy to understand our lack of wonder at this strange phenomenon, for it has been the most prominent feature of our mental life since birth. As a conscious entity, I am aware right now and am looking out into the world. I am aware not only of the particular time-interval that my life inhabits, but of the slow incrementing of time tnow inside this interval. This incrementing is the mapping of my consciousness into geological time. My sense of the passing of time may or may not continue after my death, but right now it is concurrent with geological time. To me, this seems quite odd, since my conscious lifetime could have happened at any time during the past several million years. Nagel (1986, pg. 55) would separate this issue into two questions. (1) Why should my own particular awareness have settled on this particular physical network, with its unique set of circumstances, active from the year 1937 until perhaps 2020? If consciousness is a purely biophysical phenomenon, physics itself would seem to rule out such a choice. There are, after all, currently about 10 billion possible choices here. (2) More generally, why is it that my awareness has any kind of specific locus at all? The above two questions constitute one way of stating a well-known difficulty concerning the third-person laws of physics: there seems to be no way to reconcile these laws with the existence of my own first-person, subjective awareness (Goodman 2005, Himma 2005, Nagel 1986). However, this difficulty might not stand in the way of using biophysics to model human thoughts and sensations. We discuss this further in Section VI. Bryce DeWitt (2005) wrote an interesting article about theoretical physicists as am4
ateur theologians. Although this paper has nothing per se to do with theology, different religions do have different points of view about the episodic nature of human consciousness. In Section III, I discuss the two data-analysis scenarios that come up in this context, and I briefly categorize the types of evidence for and against the possibility of multiple lifetimes. In Section IV, I use the lottery model — aided by the Copernican principle — to make a crude estimate of the probability that I should be alive now, if indeed I live only once. This probability amounts very roughly to 1 part in 40,000 and embodies Part (a) of the Introduction. Section V continues the lottery model and develops the conditional analysis, as required by Part (b). This analysis shows how very skeptical one must be to conclude that I have only one lifetime. That is, I conclude that it is very improbable that no bias is involved. In Section VI, I summarize my conclusions. The Appendix is devoted to the break-in problem mentioned in the Introduction as an illustration of the similarities and dissimilarities between the first-person and third-person points of view. There is nothing unusual or profound about this distinction, however; and one should note that, if all parties have the same data, they should all reach the same logical conclusions about the probabilities. III. Data-Analysis Scenarios and Possible Evidence I wish to present my arguments as clearly as possible. Applying the tools of probability and statistics has many pitfalls, but here the data are quite simple. The data set is this: I am alive now and have a first-person, subjective awareness centered in my own body. I refer to this data set as A. I have, in effect, won the current lottery. There are two possible ways of using this data set. Scenario I: I confront the Lottery Commission, thus showing that I am alive and aware; i. e., that I am a current winner of the biological lottery. This is data set A. It is highly improbable that I have won the current lottery. Lots of people are (current) winners, but from my standpoint, my own win is a very significant coincidence. This is like a break-in occurring on the only night that a door was left unlocked. (See the discussion in the Appendix.)
5
Scenario II: The Lottery Commission is again confronted with me as a winner, but they decide to consider the possibility of bias or cheating, as discussed in the Introduction. Have I won as a one-time occurrence? Or was I able to submit a large number of tickets, in violation of the lottery rules? (Was the break-in a random, one-time occurrence, or did it involve some kind of systematic surveillance? Again, see the Appendix.) Dealing with the question of bias turns out to be very productive. This is Part (b) in the Introduction, as reflected in Scenario II. As I have said, Scenario I does not pose the interesting question about bias. The point made by Part (a) is that, if this is my only lifetime, its present coincidence with geological time is a serious violation of the Copernican principle and implies a special coincidence in time for my own particular consciousness. This idea is reinforced by the conditional analysis in Section V required by Part (b). We are now addressing issues of the sort discussed in the cautionary remarks following Parts (a) and (b). That is, I assume that it makes sense to ask whether or not I live only once. If the reader agrees that this is a sensible question, then it also makes sense to introduce the Copernican principle into the argument. The main point is going to be that we can resolve these difficulties only if this lifetime is one of many for my awareness. In the Appendix I look at the related problem about a car break-in. This problem illustrates my first-person point of view vs. that of an outside observer, or third person. The actual evidence for and against this multiple-lifetime hypothesis is, to say the least, controversial. There seem to be two types: (1) Revelation by higher authority. Here I include statements of religious doctrine. For example, in many Eastern religions, reincarnation is a major feature. But in other typically Western religions, the multiple-lifetime idea is either not mentioned or is declared heretical. Also in this category is testimony by individuals who say that they have, in deep meditation, connected with past lifetimes. The reader might or might not consider these as evidence, depending on his/her cultural and educational background. (2) Anecdotal evidence by people who have shown knowledge or connections that they could not reasonably have acquired during their present lifetimes. This evidence has sometimes 6
surfaced while under hypnosis. Other kinds of evidence appear, as well. For example, Schwartz (2002) discusses some striking experiments involving spirit media. His book refers to other documented experiences that seem to confirm the survival of conscious awareness after death. Some physical or biological scientists consider the multiple-lifetime idea to be unphysical. But such an opinion is hardly evidence; and I argue in Section VI that physical science is in principle not capable of dealing with the fact of my own first-person awareness, though it might be able to model emotions and thought processes in general. See also the remarks in Section II, especially regarding the problem of how there can actually be a “time like the present” tnow . In this regard, I can paraphrase Nagel (1986, pg. 29) by suggesting that there is no way to construct my own subjective awareness out of 80 kg of subatomic particles. This impass contrasts with the possibility of modeling awareness in general by using biophysics. Again, the crucial distinction is the my. Thus it seems that the evidence for and against multiple lifetimes and/or the survival of consciousness is inconclusive. And on the whole, it seems hard to justify a firm stand for or against the multiple-lifetime hypothesis — at least, on the basis of evidence alone. Schwartz (2002) reports that many scientists and academics have a deeply unreasonable prejudice against psychic and spirit phenomena, a condition he calls “skeptimania.” IV. The Lottery Ticket Model We now apply probability methods to the data set A: the fact that I am alive and aware now. Since we have only the vaguest understanding of how consciousness might work (Ellis, 2005), the methods of probability and statistics are germane here. In fact, they are perhaps the only tools usefully available. To fix our ideas, consider a sequence of time intervals, each of which is perhaps 50 years long, and each of which represents a possible lifetime for me. The exact length of these intervals is unimportant; we need only suppose that the number N of possible intervals is large. Thus the geological time from which we can draw the intervals would not be limited to only the near past or the near future. On the face of it, if I am allocated only one lifetime over the past and future N intervals, the chances are small that I would be consciously connected with geological time tnow , and not simply be dead for something like 1,000,000 years, or not yet born. This is the argument of Part (a) of the Introduction. 7
In the lottery model, consider that each person’s awareness has its own box containing N lottery tickets. The tickets are identical except for their color: L are red, indicating a possible conscious lifetime for that person; and the rest N − L are white, symbolizing an interval when he or she is not aware of anything. In my own particular lottery, suppose that whatever is in charge of my intervals (“the lifetime gods” perhaps) draws a ticket from my box every 50 years. If the ticket is red, I become aware for that interval. If white, I do not. In either case, the ticket is not returned to the box, but discarded. A red ticket drawn in the past or in the future does not constitute a current win. For the moment, I am interested only in the current drawing. The hypothesis that I would enjoy only a single lifetime corresponds to L = 1; that is, there is only one red ticket out of the N in my box. Thus I would be guaranteed one and only one lifetime. And we see that the chances that I would have won the current lottery — and thus be alive at geological time now — would indeed be only 1/N , presumably a small number. Because the gods are drawing tickets without replacing them, there are actually fewer than N tickets in my box for the current drawing. But we don’t know the results of the past drawings, so the result 1/N is still correct. See Jaynes (2004, Chapter 3). What might a possible value of N be after all? Here I am asking about the number of 50-year intervals available for consideration. This number is simply how many intervals the lifetime gods are willing to deal with as my potential lives. For example, I could ask for the number of such intervals over the current age of the genus homo, about 2 million years. This implies that N = 2,000,000/50 = 40,000. And therefore the probability of my current win would be 1/N = 0.000025, a small chance indeed. At the risk of being repetitive, I’ll state again that my chances of winning the current lottery would be small (1/N ) only because of my contemporary first-person, subjective awareness. There are billions of people alive right now, and if there were no individual awarenesses, but only a kind of hive mentality, this argument would not be valid. There would be no way to distinguish among the many current awarenesses. In effect, my having an individual subjective awareness allows me to apply the Copernican principle: there is nothing special about my being alive at some time; but the fact that I am alive at geological time now puts me in a very special position, unless this interval is only one of many — unless, that is, my lottery box holds many red tickets. Again, this is the point of Part (a) of the Introduction. 8
Of course, using 1/N as the probability of winning the current lottery treats all intervals as having equal probability. If we were to consider only homo sapiens, we would have to weight intervals near the year 2000 more highly than those in the past. This weighting would not be a very large effect, however, since less than 7% of all homo sapiens past and present are alive at this particular time (Curtin, 2007). And there is no reason to believe that our particular species has a monopoly on first-person consciousness. Thus my simple 1/N assumption seems appropriate. In the next section, I use the lottery model to complete the conditional analysis required by Part (b) of the Introduction. This analysis is conclusive, but the reader who has little experience in probability theory might want to proceed directly to Section VI. V. The Bias Hypothesis and Conditional Probability in the Lottery Model As stated in the preceding section, the probability model for my lifetime(s) consists of a box containing initially N tickets. These tickets are identical except for their color: there are L red tickets and N − L white ones. Every 50 years, the lifetime gods draw a ticket from the box. If it is red, I am aware or “have a life” for that time interval; if white, I remain unaware. In both cases, the ticket is not returned to the box. This is called sampling without replacement and is thoroughly discussed by Jaynes (2004, section 3.1), among others. See also Feller (1968). We assume that the total number of tickets N (the number of possible intervals) is large; perhaps N ≈ 40,000, as estimated in the previous section. So the question I wish to answer is “What can we say about the value of L, the number of red tickets?” That is, how many lifetimes would I actually experience? For the current lottery, we know that the gods have drawn me a red ticket (data set A). Certainly, 1 ≤ L ≤ N . The skeptic would say that I should live only once, so L = 1. Let H1 stand for the hypothesis that I experience only one lifetime. In this case, we would say that the lottery is unbiased. But some people would claim that L is large. Then let HL stand for the corresponding hypothesis, so that H2 stands for the hypothesis that I actually experience two conscious lifetimes, and so on. Note that I am starting to set up the conditional analysis foreshadowed in Scenario II. In terms of the lottery analogy, H1 is winning by a lucky draw for which I had only one red ticket in my box. Let the bias hypothesis Hb correspond to the collective HL , where 2 ≤ L ≤ N . The bias hypothesis would thus involve having entered more than one 9
red ticket. For example, H3 stands for the hypothesis that I somehow entered three red tickets, so that my chances of winning in the current draw would be increased to 3/N . I use the usual | notation, where P (A|HL ) is the probability of coming up with the event A, if I assume that the hypothesis HL is true regardless of any misgivings and prejudices. Thus I ask the following question under the auspices of Scenario II. Out of those N possibilities, what is the probability P (A|HL ) of drawing data set A, given the hypothesis that there are L lifetimes out of N possible intervals? (Or the probability of winning the current lottery, given that there were L red tickets in my box.) It is, of course, P (A|HL ) = L/N . I now compute the probabilities of the hypotheses P (HL |A), given the data set A. This is a conditional (Bayesian) analysis, so I need to work with a set of prior probabilities P (HL ) that summarize our expectations prior to the realization that I am conscious at this time. In Section III, I discussed the evidence for and against the hypothesis that L > 1, and I concluded that the available evidence does not support a strong belief one way or the other. I abbreviate p1 ≡ P (H1 ). And for clarity, I use the prior multi-lifetime bias probability pb ≡ 1 − P (H1 ) = 1 − p1 as a parameter indicating a prior state of mind about there being more than one lifetime. That is, more than one red lottery ticket in the box. For simplicity, I assume that all the other probabilities are equal: P (HL ) =
pb N −1
for L = 2, . . . , N.
(1)
I refer to the collective multi-lifetime hypothesis as Hb , as in the hypothesis of bias in the lottery discussed above. If, in spite of the evidence, I think that this multi-lifetime hypothesis Hb is complete nonsense, then pb = 0. Thus p1 = 1, and P (HL ) = 0 for 2 ≤ L ≤ N . If I don’t have any particular prejudice about multiple lifetimes — maybe there are, maybe there aren’t — I might choose P (Hb ) ≡ pb = 1/2; whereupon p1 = 1/2, and equation (1) gives P (HL ) = 1/[2(N − 1)] for L = 2, . . . , N . This choice would be in better agreement with the evidentiary discussion in Section III. The information A that I am alive now is equivalent to a red ticket’s having just been drawn. I wish to compute the conditional probability P (H1 |A, pb ). This is the posterior probability for H1 (only one lifetime, or only one red ticket in the box), given the data set A and the prior value of pb . Note that 1 − pb ≡ P (H1 |pb ) ≡ P (H1 ). The rule for conditional probabilities for P (H1 |A, pb ) simply restricts the consideration of H1 to the 10
subset (or subpopulation) A (Feller, 1968). Renormalization then gives P (H1 |A, pb ) = P (H1 A|pb )/P (A|pb ). Here, H1 A is the intersection of H1 and A. Since H1 A = AH1 , we clearly have P (A|H1 , pb ) = P (H1 A|pb )/P (H1 |pb ). Therefore one easily gets the product rule (Jaynes, 2004) P (H1 |A, pb ) =
P (A|H1 )P (H1 |pb ) . P (A|pb )
(2)
The first factor on the right is just the probability for drawing a red ticket for this interval, given that there is only one red ticket in the box and that I do not know the results of any earlier drawings. It is P (A|H1 ) = 1/N . The denominator is the total probability of drawing a red ticket for this interval, given a value for pb . Because P (A|HL ) = L/N , this total probability is, from equation (1), P (A|pb ) =
N X
P (A|HL )P (HL ) = pb
L=1
=
1 1 + N 2
+
1 (1 − pb ) N
1 (N pb + 2), 2N
where I have used the identity Equation (2) then yields
PN
L=2
P (H1 |A, pb ) =
L ≡ (N + 2)(N − 1)/2.
N pb N1
1 − pb 2(1 − pb ) . = 1 N pb + 2 + 2 + 1 − pb
(3)
Accordingly, this is the probability that I have only one lifetime, given the fact A that I am alive now and given the prior value of the multi-lifetime bias parameter pb . Note that, if the multi-lifetime hypothesis Hb is deemed to be a priori nonsense (P (H1 ) ≡ p1 = 1 − pb = 1), equation (3) implies that it is still nonsense: P (H1 |A, 0) = 1, as of course must be the case. But if I am unprejudiced about whether or not there can be more than one lifetime, (p1 = pb = 1/2), then P (H1 |A, 1/2) =
2 . N +4
(4)
For large N this is very small; i. e., the chances that my box had only one red one ticket are very small. Thus it is very improbable that I would have only one lifetime. Indeed, for N = 40,000 we get P (H1 |A, 1/2) ≈ 0.00005. 11
But we see that, no matter how large N is — no matter how many possible intervals there are for me to have been alive — there exists a value of pb so small that the posterior probability for more than one lifetime is still less than, say, 1/2. So I can ask, given a value for N , what value of pb gives P (H1 |A, pb ) = 1/2; that is, how skeptical do I have to be about multiple lifetimes to raise the posterior probability for this hypothesis to 1/2? Let us call this prior pb (skeptical). Setting the left-hand side of equation (3) equal to 1/2, I find pb (skeptical) =
2 . N +4
This expression is the same function of N as the above expression for P (H1 |A, 1/2) because equation (3) is actually symmetric under an exchange of P (H1 |A, pb ) and pb . If we use the rough estimate N ≈ 4×104 , we must conclude pb (skeptical) ≈ 5×10−5 = 0.00005. This amounts to a rather high degree of skepticism about multiple lifetimes, especially in the face of the discussion at the end of Section III. An interesting argument arises in this connection: if I hadn’t won the current lottery, I wouldn’t be here to talk about it. This is certainly true, but it seems irrelevant. Consider that, when I won, I received a plane ticket to join the other winners in a celebratory meeting at a deluxe resort. At this meeting I strike up a conversation with a third party — Mr. Smith — and I explain why I think the lottery is biased. After some thought, Smith says that there are lots of winners here at the resort and that we have no idea how many losers there were in this particular draw. Thus he is reluctant to conclude that there is any kind of bias. However, we both know nothing about the number of losers; and the data we both have are the same: a coincidence between my being alive and physical time now.
The
fact that there are perhaps 6 billion human winners at the resort celebration is irrelevant. Since Smith and I both have the same data, we must both reach the same conclusion about bias — the conclusions embodied in equations (3) and (4). So again, I have to argue that the first-person and third-person points of view are consistent and logically the same. It is certainly possible that there is some other explanation for my astonishingly contemporary existence. Given that we do not understand what consciousness is, we could enlarge our set of hypotheses to include an “ignorance hypothesis” Hi in the spirit of the deception hypothesis HD of Jaynes (2004). He introduced this hypothesis in connection with data alleged to confirm the existence of powers of ESP. However, it seems pointless to do this formally, because all we can say is that if the unknown mechanism can produce the 12
data set A with some finite probability P (A|Hi ), and if the reader prefers Hi to Hb , then P (Hi |A, p) can easily be made greater than P (Hb |A, p) by adjusting the prior assigned to Hi . VI. Conclusions Many readers would say that there is something unphysical about the possibility of multiple lifetimes for a single consciousness. If consciousness is strictly biophysical, this multiplicity may raise valid concerns. See Searle (1992, 2004) for discussions of the doctrines of materialism and dualism and how they relate to the mind-body problem. On the other hand, it is difficult to reconcile the existence of a first-person, subjective consciousness with the third-person laws of physics, which are viewpoint-invariant (Goodman 2005, Himma 2005, Nagel 1986). Ellis (2005) has argued that the bottom-up methods of physics can in principle never explain the high-level complexities of life, let alone consciousness. Physics may indeed be able to explain or model emotions and sensory experiences as biophysical events. These conscious experiences are often called “qualia.” The thought or feeling “I am hungry and want pizza” may correspond to a certain identifiable series of biophysical processes. Physics excels at making accurate models of physical systems, perhaps including qualia, but one can safely say that physics is in principle not set up to deal with the problem of my individual first-person, subjective awareness. Consider, for example, the problem of modeling the spectrum of radiation emitted by a system of hydrogen atoms. What might be the rationale for singling out a particular one of these atoms from all the others? A particular atom may be said to have emitted an Hα photon at approximately time t0 ; that is no different from Mr. Jenkins’ becoming angry at having to wait through three cycles of a traffic light at 5:30 PM. But it is simply not equivalent to my being angry at the traffic light, even if my name is Jenkins. In effect, there is no way for the “my” to enter the discussion. Searle (1992) discusses why it seems impossible to successfully attribute mental processes and first-person awareness to arrangements of physical particles and fields. The reader can find an excellent pr´ecis of these arguments in Nagel (1995). Thus it seems unlikely that a strictly materialist position could account for the properties of a first-person conscious awareness. Hofstadter (2007) argues that consciousness is a sort of neurological dance of symbols 13
in the brain. He might say that neurological patterns, or feed-back loops, are the bearers of consciousness. This may in fact be true, but what makes a particular one of these loops me is a question that remains unanswered. I wish again to make the point that most people are so inured to being who and where they are that the very unusual and really remarkable nature of this phenomenon remains unappreciated. Indeed, we understand very little about the biochemistry or biophysics of the higher levels of such phenomena. And in any case we must deal with the remarkable coincidence between geological time now and my present state of consciousness. The Copernican principle, applied to the lottery model, shows that this coincidence is highly unlikely. I have shown that applying conditional probability methods to this model leads to the conclusion that it is unlikely that I would enjoy only one lifetime (see equation (4)). My one lifetime could have occurred at a vast array of geological times. But in truth I happen to be alive right now. And my conscious, subjective egocentric awareness, centered in my body, heightens the significance of this fact. Acknowledgments I would like to thank James O. Allsup, B. Roy Frieden, J. L. Denny, Claire I. Cocke, Mark Gettings, Allan Goodman, Noah Goodman, J. R. Gott, D. E. Orne, and Yervant Terzian for their helpful comments. References Carnap, Rudolf (1963). “Intellectual Biography,” in P. A. Schilpp, ed., The Philosophy of Rudolf Carnap, La Salle, IL: Open Court Press. Curtin, Ciara (2007). Scientific American, Sept. 2007. See also Carl Haub’s article with the Population Reference Bureau on www.prb.org/Articles/2002/HowManyPeopleHaveEverLivedonEarth.aspx DeWitt, B. (2005). Physics Today, January 2005, pg. 32. Dummett, M. (1978). Truth and Other Enigmas, Oxford: Oxford University Press. Ellis, G. F. R. (2005). Physics Today, July 2005, pg. 49. Feller, W. (1968). An Introduction to Probability Theory and Its Applications, Vol. 1, 3rd Ed. New York: John Wiley & Sons. Goodman, N. D. (2005). private communication. Gott, J. R. (1993). Nature, 363, pg. 315. Gott, J. R. (2001). Time Travel in Einstein’s Universe. Boston: Houghton Mifflin Co. 14
Himma, K. E. (2005). Psyche, 11(5) (June 2005). (psyche.cs.monash.edu.au/ symposia/metzinger/Himma.pdf) Hofstadter, Douglas (2007). I am a Strange Loop. New York: Basic Books. Jaynes, E. T. (2004). Probability Theory: The Logic of Science. Cambridge: Cambridge University Press. Nagel, T. (1986). The View from Nowhere. Oxford: Oxford University Press. Nagel, T. (1995). Other Minds. Oxford: Oxford University Press. Chapter 10. Schwartz, G. E. (2002). The Afterlife Experiments. New York: Atria Books. Searle, John R. (1992). The Rediscovery of Mind. Cambridge, MA: M. I. T. Press. Searle, John R. (2004). Mind: A Brief Introduction. Oxford: Oxford University Press. Appendix: The Car Break-in Incident This appendix describes an unusual coincidence that happened to my wife and me. Analyzing this coincidence is useful and illuminating because it parallels what happens in Scenario II. Our interpretation of this incident, as described below, was not popular with the sheriff’s deputy who responded to our phone call. In effect, the deputy denied that the coincidence was significant. He said that car break-ins occur often and that there are certainly coincidences that occur by chance. But as you will see, this point of view is not defensible. The deputy had the same data about the break-in that we did. And logically, he should have reached the same conclusion: the coincidence was significant, and increased surveillance of our neighborhood would be justified. I show below that the deputy’s point of view is analogous to that of my fellow lottery winner Mr. Smith in Section V. Here’s what happened: My wife and I live in a quiet neighborhood outside the city. There are the usual burglaries and similar crimes, but nothing out of the ordinary. We have no garage, and so our car is parked outside in our driveway. Usually the car is locked overnight, but one night we carelessly left it unlocked. And we claim that this was the only night in recent memory that the car was left unlocked overnight. When we looked at the car the next morning, we found that someone had entered the car that night and peeled back the plastic shroud over the steering column. We called the Sheriff’s Department, and soon a deputy arrived to investigate. He said that the would-be thief had tried to unlock the steering column and start the car, but that 15
our car is resistant to that sort of tampering and that the thief had given up. He said that this sort of molestation occurs occasionally. Hardly unusual. But the fact that this had occurred on the only night that we could remember leaving the car unlocked seemed significant. So I told the deputy that this could only mean that someone was systematically and carefully checking the neighborhood for unlocked cars. Otherwise why would this break-in occur on the very night that we had left the car unlocked? The deputy was not impressed with this argument and said that this sort of coincidence would happen now and then, given the fact that many cars are stolen every night. So he was unwilling to initiate any sort of special surveillance of our neighborhood. What are we to make of this? Was the deputy right to claim that nothing unusual was happening in our neighborhood? Note that we all agree on the data, which consist of two facts: (1) The break-in occurred on the only night that the car was left unlocked. (2) In general, there are lots of car break-ins and thefts. This latter fact is actually irrelevant; to include it in the analysis, we would need to know the general frequencies and types of break-ins and possible coincidences. So we are forced to analyse the problem like this: let HI be the hypothesis that the break-in was just a coincidence; and let HII be the contrary hypothesis, that indeed the neighborhood was being systematically checked every night. We agreed with the deputy that HII is improbable, and assigned a prior P (HII |ps ) ≡ ps
