Implications of the Copernican Principle for Our Future Prospects J Richard Gott 11

HYPOTHESIS Implications of the Copernican principle for our future prospects J Richard Gott 11/

Making only the assumption that you are a random intelligent observer, limits for the total longevity of our species of 0.2 million to 8 million years can be derived at the 95% confidence level. Further consideration indicates that we are unlikely to colonize the Galaxy, and that we are likely to have a higher population than the median for intelligent species.

THE Copernican revolution taught us that but also against whatever calamities may these species becoming extinct as a func

it was a mistake to assume, without affect its observability in the future, tion of time tf in the future is sufficient reason, that we occupy a because all that is required for equation privileged position in the Universe. Dar (1) to work is that in the end your position N(tf) dtf = NtotalAo exp (-Aotf ) dtf (4) win showed that, in terms of origin, we as an observer turns out not to have been Let rl and r2 be independent random num are not privileged above other species. Our special. Just to illustrate, in 1969 I saw for bers each distributed uniformly over the position around an ordinary star in an the first time Stonehenge (tpast = 3,868 interval [0, 1]. For a given species alive ordinary galaxy in an ordinary super years) and the Berlin Wall (tpast = 8 years). today cluster continues to look less and less Assuming that I am a random observer of tp = -(Ao)-Iln r special. The idea that we are not located the Wall, I expect to be located randomly l

in a special spatial location has been in time between tbegin and tend (tend occurs tf = _(AO)-I In r2 (5) crucial in cosmology, leading directly when the Wall is destroyed or there are 'p) = r2 to the homogeneous and isotropic no longer any visitors left to observe it, ri'/ Friedmann cosmological models in gen whichever comes first). The Wall fell 20 Let Y> 0 be a constant eral relativity theory which have been years later giving tfuture = 2.5 tpast , within 1 2 remarkably successful in predicting • the the 95% confidence limits predicted by P([trft ] > Y) = dr! = 1/(Y+1) (6) p Lri existence and spectrum of the cosmic equation (1). Applying the (P = 0.95) delta 3 microwave background radiation .4. In t argument also correctly predicts that The above probability distribution for the astronomy, the Copernican principle Stonehenge should still be observable ratio of trf tp is exactly that found earlier works because, of all the places for intel today, 24 years later (as tfuture> 99 years, through what I have called the delta t ligent observers to be, there are by from equation (1)). In 1977 I visited the argument: 50% of the time tf > tp (P = 0.5 definition only a few special places and U.S.S.R. (tpast = 55 years). Although at that for Y= 1),50% of the time ttp< tf <3tp many nonspecial places, so you are likely time its existence into the indefinite future (P=0.25 for Y=3 and P=0.75 for Y= to be in a nonspecial place. This idea can was generally assumed, it ended only 14 t), and 95% of the time~tp < tf < 39tp (P = be used to estimate the likely future years later (tfuture = 0.25 t , consistent past 0.025 for Y = 39 and P = 0.975 for Y = ~); longevities of various observables, includ with the limits in equation (1)). Equation see Fig. 1. ing that of our own species. I will discuss (1) was satisfied not because my visit its implications for the far future, for the somehow caused the demise of the Search for Extraterrestrial Intelligence U.S.S.R. but simply because in hindsight (SET!) and for space travel. we can now see that the timing of my visit 0.6 was unremarkable. Nature has been pub Delta t argument lished for 123 years; the delta t argument a.. Assuming that whatever we are measuring predicts (P = 0.95) its future publication .£ can be observed only in the interval will last more than 3.15 years but less than :0 jg 0.3 between times tbegin and tend, if there is 4,800 years. e nothing special about tnow we expect t now 0... to be located randomly in this interval. Our future longevity The estimate tfuture = (tend - tnow ) = tpast = Suppose, consistent with the findings of (tnow - tbegin) will overestimate tfuture half CarterS, intelligent species are currently the time and will underestimate it half the being formed in the Universe at a uniform 1 1 1 10 100 1,000 time. If r1 = (tnow- tbegin)/(tend - tbegin) is rate. By 'intelligent' species we mean one a random number uniformly distributed that is self-conscious and has the cognitive 1.000 100 10 between 0 and 1, there is a probability abilities to reason abstractly, think about tFu!ure l t Past

P = 0.95 that 0.025 < r1 < 0.975, or the future, create art and so forth. So far, equivalently on Earth, we, homo sapiens, are the only 6 FIG. 1 Probability per unit logarithmic interval species to fit this description . Assume that (base 10) of finding a given ratio ttuturel tpast for ~tpast < tfuture < 39 tpast (1) these intelligent species are subject to (95% confidence level) an observable from the delta t argument. The some unknown extinction rate Ao. The dis probability that ttuturel tpast is greater than Y is Similarly, tribution of ages tp of the Ntotal intelligent 1/(Y +1) [see equation (6)]. There is a 95% proba• species alive today is bility that ttuturel tpast lies between (1/39) and 39. ttpast < tfuture < 3 tpast (2) (50% confidence level) N(tp) dtp = NtotalAo exp (-Aotp) dtp (3) Our species, homo sapiens, is roughly Equation (1) tells us that the length of as only a fraction f = exp (-Aotp) of those tp = 200,000 years old. (This is conserva time something has been observable in the species born at an epoch tp ago survive tive, as recent estimates include> 100,000 7 8 past is a rough measure of its robustness today. Species alive today are subject to years , >150,000 years , 200,000 years9 not only against the calamities of the past, the extinction rate Ao so the number of and 250,000 years6; if improved estimates

NATURE· VOL 363 . 27 MAY 1993 315 © 1993 Nature Publishing Group HYPOTHESIS become available they can be substituted associated with the fact that you are an Consider the following toy model. Sup in an obvious way). Clearly, tp is much intelligent observer. For example, the pose the birth rate b is a constant and that less than the age of the Universe, to = places and times that intelligent observers the death rate has the constant value d l < b 13 X 109 years. Using equations (1) or (6), are born may be limited by the laws of before time tmax and the constant value l4 the estimate for the future lifetime of our physics . This is the weak anthropic prin dz > b after time tmax . The population p (t) l5 species is ciple as formulated by Carter and is basi will thus have a rate of increase = b - d l = cally just a self-consistency argument. It t;-I before time t and a rate of 5,100 years < t < 7.8 X 106 years max f (7) has had notable success in cosmologyI6,17, decrease = d - b = (;1 after t ' The (95% confidence limits) z max giving us our best explanation of the Dirac number of people born as a function of This gives our species a total longevity large-numbers coincidencel4. Carter5 has time is thus between 0.205 million and 8 million years. pointed out that there is a remarkable NB(t) dt = bp(t) dt The average longevity for most species order-of-magnitude coincidence between is between 1 and 11 million years 10-12, and the time required to develop intelligent life = bNmax exp ([ t - tmax ]/ tl ) dt for mammals it is 2 million years13. Our on the Earth, tl = 4.5 X 109 yr measured for t < tmax direct ancestor homo erectus lasted only from the formation of the Solar System, (11) about 1.4 million years and our near rela and the future main-sequence lifetime of NB(t)dt=bp(t)dt 9 tives, the Neanderthals, flourished for the Sun tfms = 6 x 10 yr. Carter uses this roughly 200,000 years. There is thus an fact and an ingenious argument to deduce = bNmax exp ([ tmax - t]/ tJ dt order-of-magnitude coincidence between that there must be of the order of 1 improb for t> tmax the range of total legevity of our species able event required for the formation of as predicted by the delta t argument and intelligent life. I agree; further, this coin Regardless of the values of tl and t2 , 50% the observed longevity of other species, cidence is just what one would expect from of observers are born at times when the particularly species like ours. Thus we the delta t argument if the intelligent life population p( t) is within a factor of 2 of should not assume that our intelligence is is simply formed at some random time its maximum value of N max (and 95% of likely to increase our longevity vastly during the main-sequence lifetime of the observers are born when the population above that of other species. star. Interestingly, Carter's argument is within a factor of 20 of Nmaxl. It is thus If Ao« (200,000years)-I, only a tiny depends implicitly on the idea presented interesting that the current population of fraction f < 200,000 years x Ao« 1 of all formally here: that according to the Cop the Earth is over 5 billion and many pro intelligent observers alive at present would ernican principle, among all intelligent jections of future growth show the popula zl 23 observe tp ~ 200,000 years, and you would observers (including those yet to be born) tion topping out at -8 to 12 billion - • be unlikely to be among them. In the limit you should not be special. Estimates of the theoretical maximum where we expect our species to live for Let us formalize this as the 'Copernican population for the Earth include 20 bil 21 24 ever, Ao goes to 0 and P( tp ~ anthropic principle': that the location of lion and 40 billion . Exponential growth 200,000 years) goes to O. What we would your birth in space and time in the Uni and decline is common in biological sys expect to observe in the limit as Ao goes verse is privileged (or special) only to the tems, so you should not be surprised to to 0 is that the value of to we observe goes extent implied by the fact that you are an be born in an epoch with problems of to infinity, tp = to and (to - tp)« to. These intelligent observer, that your location overpopulation. The total number of would be the telltale signs that we might among intelligent observers is not special people born before tmax is NI = bNmaxtl survive into the indefinite future. In fact but rather picked at random. Knowing and the number of people born after tmax we see exactly the opposite: tp« to, and only that you are an intelligent observer, is N2=bNmaxt2, so NdN2= tdt2' In the (to - tp) = to. These are the telltale signs you should consider yourself picked at limit where t2 goes to infinity, this model that we can expect to last only a short time random from the set of all intelligent crudely follows a logistic curve, a period into the future tf = tp(X39±1). observers (past, present and future) any of exponential growth followed by a long Replace the words 'intelligent species' one of whom you could have been. period of equilibrium where the birth and with intelligent genus, intelligent family, Assume therefore that you are located death rates are equal. But this is not intelligent order or intelligent class, and randomly on the chronological list of allowed because if you find yourself on the argument remains the same. So the human beings. If the total number of intel the rising exponential you are a member estimates of tf that I have derived apply ligent individuals in the species is a posi of the set N I, and we require that N2 < not only to us but to any (by an argument tive integer N tot = Npast + 1 + Nfuture where 39NI in order that your good luck should analogous to the delta t argument there N past is the number of intelligent not be more spectacular than 2.5%; this are not likely to be many) intelligent individuals in the species born before a means that tz < 39 tl , and as you are likely species (including intelligent machine particular intelligent observer and Nfuture to be born near tmax , that tf ~ 39 tp Uust as species) descended from us. is the number born after, then we expect in the delta t argument). But the set N z Gould6 called tp« to "geology's most Npas, to be the integer part of the number to which you do not belong can be as small frightening fact", because "If humanity r l N tot where r l is a random number uni as 0, which means that the only lower limit arose just yesterday as a small branch on formly distributed between 0 and 1. Thus: on tz is O. In the limit tz = 0 (sudden extinc a flourishing tree, then life may not in a tion) you are likely (P = 0.95) to be born (1/39) Npast -1 < Nfuture < 39(N 1) (8) genuine sense, exist for us or because of past + between 3.7 and 0.025 e-folding times (tl ) us. Perhaps we are only an afterthought, (95% confidence level) of the end, giving a lower limit of tf> a kind of cosmic accident, just one bauble and if N past » 1 0.025tl = tp/40 In (Nmax ). on the Christmas tree of evolution." I find We might reasonably expect, as a first it frightening too, but for a different rea (1/39) Npast ~ Nfuture~ 39 N past (9) guess, that an eventual population crash son: if tp« to then the delta t argument (95% confidence level) would roughly mirror its rise but perhaps suggests tf = tp(X39±1) and thus tf « to, The number of human beings born so far be multiplied by a scale factor in time. In meaning that on astronomical timescales, is of the order of 70 billion (refs 18-20). this case we would still expect the 95% we and our intelligent descendants prob By equation (9) the number of human confidence level upper limit on tf to be ably will not be around very long. beings yet to be born is expected to be - 39 tp just as in the delta t argument Population effects. When you evaluate (because if the fall mirrored the rise but your observations of the Universe you 1.8 billion < Nfutme < 2.7 trillion (10) was stretched out by a factor of 39 in time, must take account of the selection effects (95% confidence level) then 39/40 of the people would be born

on the declining portion of the curve). The 95%-confidence-Ievellower limit on t can f Ehrlich & Ehrlich be calculated directly by noting that the Type (2) 95%-confidence-Ievel lower limit on the number of future human births is 1.8 bil P(Q lion (equation (10». At current rates25 it will take only about 12 years for another o 1.8 billion people to be born (assuming 0 4.2 x births continue without abatement before Ehrlich & Ehrlich a sudden extinction). This extremely Type (3) pessimistic lower limit would need a double dose of bad luck, bad luck at the P(Q 2.5% level coupled with an instantaneous I % , . < • extinction. The number of years is so short o '" 6 because a reasonable fraction, 7.7%, of all or------. • 4.2 X 10 yr the people who have ever lived are alive Galactic colonization todayI8-20,25. This is a dangerous situation. 5 X 1018 Thus, including population effects, P(Q 12 years < tf < 7.8 million years (12) (95% confidence level) 0 Disturbingly, even extraordinarily low 0 values of tf cannot be confidently excluded (P = 0.95) but high values of tr, such as 5 x 1033 many billion years, which we might hope for, can be. Ehrlich and Ehrlich24 propose three possible future population scenarios for P(Q the human race: (1) topping out at 10 billion in the next century and then dying out; (2) topping out at 10 billion in the 0 next century and then collapsing to a few •0 hundred thousand people eking out a livelihood on an impoverished planet for the next 4 million years; and (3) a sustain ble stable population of 1 billion for the FIG. 2 Population against time for various scenarios discussed in the text shown on linear scales. In next 4 million years. The number of future the pessimistic Ehrlich & Ehrlich Type (2) scenario, the population reaches 10 billion in the next century births implied by these scenarios are and then collapses to a few hundred thousand for the next 4 million years; on this scale, only the roughly (1) 10 billion, (2) 30 billion and population spike near the present is visible. In all but the Type (2) scenario your position (black arrow) 12 turns out to be extraordinarily lucky to be in the first tiny fraction of the intelligent observers in our lineage. (3) 40 x 10 . Unfortunately, only the two pessimistic cases (1) and (2) are within the limits in equation (10) (see Fig. 2). 12 Combining Nrutuce < 2.7 X 10 (equation 12 years old, more than 1.8 billion people where Copernicus's theory is known, but (10» with the current rate of 145 million have already been born after you, so you you were not lucky enough to have been 25 births per year we find tr < 19,000 years already know that you are not in the last around when it was discovered, and in unless the rate of births drops. If we wish 2.5% of the chronological list. In another such an optimistic scheme it would be to stretch our survival out to the upper 12 years, with the current rate of births, likely for you to be in the same position limit of 7.8 million years (from equation everyone alive today will be off the last with respect to equation (9). The only hope (12» we require the average rate of births 2.5% of the chronological list. So bad luck to force you into the first 2.5% of the to drop by a factor of more than 400. is required for you to be on the last 2.5% chronological list is to live in a type (2) Although an optimist might hope for this of the list. If you are an optimist and scenario where civilization collapses in the to be produced by extraordinary means believe that civilization will go onward near future and we return just as before such as genetic engineering producing a and upward from here, then any future to a hunter-gatherer phase (which forgets greatly increased human lifetime, a humans should live in a civilization the equation and is not sophisticated pessimist would note that any civilization advanced enough to remember or rederive enough to rederive it) which one might potent enough to produce a large linear equation (9); so of all those observers who hope would last a long time. But this will increase in its life expectancy would be live in epochs sophisticated enough to not work either, because without civiliz potent enough to produce an exponential apply equation (9), less than 2.5% of them ation we would lose our possibility of increase in its numbers as well, and argue can be in the first 2.5% of the entire being especially protected against extinc that such a drop could be more easily chronological list (because they occupy all tion and we should at that point have an produced simply by a population crash. of the chronological list except for a seg expected future longevity of about 2 mil Equation (9) claims that it is likely that ment at the beginning). lion years or less, like other hominid you are not in either the first 2.5% or the In such an optimistic scenario if species. The popUlation in the final hunter 12 last 2.5% of the chronological list of Nfutuce> 2.7 X 10 , you would have to ask gatherer phase should be low so that future human beings. Can you avoid this con yourself why, of all the more than 2.7 x births should be of order 30 billion (con clusion by arguing that you occupy a 10 12 humans who will live in epochs where sistent with equation (9». (In such a situ special position on the list because you equation (9) or its variants are known, you ation, the civilization, by Carter's5 argu are born into an epoch where the level of are so lucky (P < 0.002) as to be among ment, should form at some random time sophistication is great enough to know the 5.4 billion people alive when it is dis during the lifetime of the species and the equation (9)? Well, if you are now over covered. For example, you live in an epoch delta t argument should apply). Being in

the first 2.5% of the chronological li st would only increase life-expectancies to assuming a civilization's rate of births is requires at least good luck at the 2.5% level 104 years and even extraordinary increases not correlated with L; if it were positively in all these cases. in lifespans (of up to say 10 100 conscious correlated, which might seem natural, then Any biases due to the fact that we live years) would still require the Dyson super the limits would be even stronger). As you in a technological civilization we expect civilization to have over time of order are randomly located on the list of obser 125x 6 to be small in any case, because popula 10 10 individuals and our probability vers in the catalogue tions are so high during technological of becoming a Dyson supercivilization ; N X 6 civilizations that we expect a substantial would be less than of order 10- 1.25 I0 ; I L, = r2 I L, = '2(L)N fraction of all intelligent observers to be still there could be as many as _10375xl06 i = 1 ; = 1 Dyson supercivilizations in the Universe. born into them. ""' jLj ~NLj =NLp /rl (13) So some intelligent life may last into the Prospects for the far future far future; it is just not likely to be us or where r2 is a random number uniformly The probability that we will colonize the our descendants. distributed between 0 and 1. Thus Galaxy (increasing our population by a In the limit where (cosmology permit 9 9 ( L)~ L,'/('lr2) (14) factor of at least 10 ) is of order p "", 10- , ting) a supercivilization is able to accu as in such a situation we expect N future> mulate an infinite amount of elapsed The probability P = 0.95 that (rl ' 2) > 9 9 9 10 N past and P(N!uture> 10 N p",,) "'" 10- • conscious time and an infinite number of 0.0087 so (Seeding other worlds with microorgan intelligent observers (N ,), the fraction of to (L) < 12,100 years, N < 121 (15) isms26 might be possible, but by Carter's5 ordinary civilizations such as ours that will o (95% confidence level) argument only a small fraction would ever develop into such a supercivilization must develop intelligent life. Even if one enter go to zero so that the set of observers born independent of the form of the distribution tained the notion that life on Earth was on the original home pl anet is not an of the Ls, and considerably lower than 6 28 6 seeded in this wa/ - , the typical delay infinitesimal minority of all intelligent Cameron's37 estimate of (L ) = 10 years. of several billion years to evolve intelligent observers. Thus there is some possibility that a life when it does occur guarantees that the radio search for other civilizations in our number of direct 'ancestor' civilizations in Implications for SETI Galaxy or others could be successful, but our lineage must be at most a few, and, Colonization is not important in the sense a targeted radio search of 1,000 nearby using an argument similar to the delta t that galactic colonists and their descen stars is not likely to succeed. The upper argument, the number of our 'descendant' dants must not dominate the numbers of limit on No is generous because Carter's5 intelligent species is also not likely to be intelligent observers in the Universe argument implies that the probability of large). The probability that we will build (otherwise you would be likely to be one). forming an intelligent species on a habit a Dyson29 sphere of O'Nei1l3o solar Significantll\ this explains why we able planet around a given star during its powered space colonies surrounding the should not be surprised that we have not main-sequence lifetime may be many Sun (using all the Sun's radiant energy and been colonized by extraterrestrials. orders of magnitude below unity rather mUltiplying our population by a factor of Assume 109 habitable planets in the than the optimistic one order of magnitude 8 8 at least 10 ) is of order p"", 10- . The prob Galaxy (surely optimistic). Carter's5 argu below unity that we adopted. ability of our establishing a Kardashev ment shows that the fraction of these that As a human being today the probability type III civilization, defined as one that will develop intelligent life is at least an is large (P = 0.97) that you were born in uses the energy output of its entire galaxy order of magnitude and perhaps many a country with a population of more than by placing a Dyson sphere around each orders of magnitude less than 1. As the the median value of 6.3 million25 . For the star (multiplying our population by a fac main-sequence lifetime of their star tms is same reason, if there is a large spread in 17 17 10 tor of at least 10 ) is of order p "", 10- • of the order of 10 years, the rate at which the populations of intelligent species, Dyson31 showed that operating at lower intelligent civilizations are forming in the which we would expect for noninteracting and lower temperatures as the Universe Galaxy is 7]<0.01 yr-I. If the average lon opportunistic species, then it is likely that expands (thinking ever more slowly, and gevity of these civilizations for radio you, as a random intelligent observer, interspersing periods of hibernation), transmission is (L) then (as colonization would find yourself in an intelligent intelligent life could in principle continue is not important) by the Drake species with a population larger than 35 forever, having an infinite number of equation ,36 we expect to be able to the median. (In the canonical a = thoughts on only a finite amount of energy. observe within our Galaxy No = 7] (L) < 0.2 lognormal distribution which best fit s As an example, consider Linde's32 chaotic 0.01 yr-I (L ) civilizations transmlttmg a wide variety of observed species-abund inflationarl 3 cosmology where our Uni now. You are born into a radio-transmit ance curves, 98.6% of all individuals come verse may continue its normal expansion ting civilization, so you should be picked from species having populations above the 5x 6 for _10 10 years. The amount of effective at random from the set of observers in median value)38-42 . Civilizations sig conscious time elapsed would be 10 1.25X lo· radio-transmitting civilizations. Assume nificantly larger than our own must be years because of the slow thinking and our radio transmission longevity is Lj sufficiently rare that their individuals do hibernation. Because the mass within the (radio transmission may end because we not dominate the total. Thus, we do not horizon grows linearly with time in the become extinct, technological civilization expect to see a Dyson sphere civilization Linde model up to 105xlo' years, even with comes to an end, or we simply move on within our Galaxy, or a Karadashev type an extremely pessimistic probability for to a different method of communication). III civilization within the current observ the formation of intelligent life on a habit Then by the delta t argument we know able horizon. 4 able planet (say 10- ,00°) there would still that Lp = r Lj where Lp = 105 years is the 6 l be of the order of 105 x 10 intelligent civiliz past longevity of radio transmission on Implications for space travel ations in our entire inflationary domain Earth, and 'I is a random number uni Why is our probability of colonizing the that would eventually become visible. formly distributed between 0 and 1. Galaxy so low given that it would be very Intelligent observers with life-expectan Arrange all N radio-transmitting civiliz good for our survival prospects to expand cies longer than 102 conscious years must ations in a catalogue in order of their radio our habitat vastly and that we already not be sufficiently common to dominate longevity so that for all i, L, "'" L,+I' know how to travel in space? Although the total, otherwise you would be likely to Assume that the number of intelligent we have been around for 200,000 years or be one. Elimination of all current causes observers in a radio-transmitting civiliz so, civilization (with cities and writing) of death other than murder and suicide ation is proportional to L (equivalent to has only been around for 5,500 years and

technological civilization sophisticated ably the best we can make. At present we of equally inappreciable length" [italics enough to engage in space travel for only have no data sufficient to reject this mine]. This is essentially the delta t argu 32 years. But the delta t argument tells us hypothesis and it does explain a number ment applied to our position among all that the capacity and motivation to engage of facts: that you were born during a living things on earth. If our species does in space travel may be with us for only of period of high population and that the become extinct sometime within the next the order of another 32 years (x39±1). The Earth has not been previously colonized few million years, and we do not colonize Cold War which was responsible for the by extraterrestrials. Like any good scien space and are not potent enough to destroy space race in the first place is now over. tific hypothesis, this hypothesis is fal all life on Earth either, then we will indeed Human flights to the Moon lasted for only sifiable, either immediately by discovery be like other species (albeit especially four years, and human activity in space of intelligent radio signals from a civiliz interesting) and we might indeed be has unfortunately retreated at present to ation around a nearby star; or eventually, expected to occupy a random position in low Earth orbit. The delta t argument sug if more than 2.7 x 1012 more human beings the history of life on earth. Life on Earth gests that there may be only a brief window are born. If you believe that our intelligent began over 3.6 billion years ago, and we of opportunity for space travel during descendants will last 10 billion years and might expect it to last another 6 billion which we will in principle have the capa colonize the Galaxy, you must believe that years until the Sun becomes a red giant, bility to establish colonies (which could you will, in the end, turn out to have been in agreement with equation (1) and Dar in turn establish further colonies). If we very lucky to have been in the first tiny win's prediction. let that opportunity pass without taking fraction of the members of our intelligent The odds are against our colonizing the advantage of it we will be doomed to lineage (see Fig. 2). If you were not lucky Galaxy and surviving to the far future, not remain on the Earth where we will event enough to find yourself on the first page because these things are intrinsically bey ually go extinct. So far, not a single human of the phone book or were not even born ond our capabilities, but because living being has been born outside the gravita on January 1, can you feel comfortable things usually do not live up to their tional potential well of the Earth. assuming that you will turn out to be even maximum potential. Intelligence is a capa The methods that I have used here are luckier in the ultimate chronological list? bility which gives us in principle a vast very conservative; if the results are dra You should be suspicious of any claim potential if we could only use it to its matic .it is only because the facts are dra that future events will conspire to make maximum capacity, but so does the ability matic (tp« to). This paper only points out you in the end turn out to be exceptionally to lay 30 million eggs as the ocean sunfish 44 and defends the hypothesis that you are a lucky, like the claim that you will win the does • We should know that to succeed random intelligent observer. Sometimes lottery tomorrow or get rich by participat the way we would like, we will have to do we say that the future is unpredictable. ing in a chain letter you have received. something truly remarkable (such as Another way to look at this is to say that What about the future of all life on colonizing space), something which most at birth you have no information on where Earth? Darwin43 said" And of the species intelligent species do not do. fractionally in the chronological list of now living very few will transmit progeny After completing this paper, I have lear human beings you will in the end turn out of any kind to a far distant futurity .... ned that ideas on future human population to be. You are no more likely eventually As all the living forms of life are the lineal similar to some of those presented here to turn out to be in the first 2.5% of the descendants of those which lived long were discussed by Brandon Carter in a chronological list of human beings than before the Silurian epoch, we may feel 1983 talk, although never in print (see 45 47 you are to be in the first 2.5% of the alpha certain that the ordinary succession of treatment by Leslie - ), as well as in 48 beticallist of human beings in your home generations has never once been broken, dependently proposed by Nielson • 0 town's telephone book. Short of having and that no cataclysm has desolated the J. Richard Gott 11/ is at the Department of Astrophy• actual data on the longevities of other whole world. Hence we may look forward sical Sciences, Princeton University, Princeton. intelligent species, this hypothesis is argu- with some confidence to a secure future New Jersey 08544, USA

1. Gamow, G. Phys. Rev. 74, 505-506 (19481. 18. Desmond, A. Population Studies: Selected Essays and (Academic, Boston, 1990). 2. Alpher. R. A. & Herman, R. C. Nature 162, 774-775 (1948). Research (ed. Kammeyer, K. C. W. 436-450) (Rand McNally. 33. Guth. A. H. Phys. Rev. D23, 347-356 (1981). 3. Penzias, A. A. & Wilson, R. W. Astrophys. 1 142, 419-421 Chicago, 1969). 34. Hart, M. H. & Zuckerman, B. Extraterrestrials- Where Are They? (1965). 19. Brraben, J·N. Population 34e Annee, no. 1.13-25 (1979). (Pergamon, New York. 1982). 4. Mather, J. C. et al. Astrophys. J. 354, L37 -L40 (1990). 20. Thatcher. A. R. Guiness Book of World Records 1991 (ed. 35. Drake, F. & Sobel, D. Is Anyone Out There? 52-68 (De lacorte, 5. Carter. B. Phil. Trans. R. Soc. A310, 347-363 (1983) McFarlan, D.) 202 (Facts On File, New York. 1990). New York. 1992). 6. GOUld. S. J. Wonderful Ufe, 44-320 (Norton, New York. 1989). 21. McEvedy, C. & Jones. R. Atlas of World Population History. 351 36. Shklovskii, I. S. & Sagan. C. Intelligent Life in the Universe. 7. Caroll, R. Vertebrate Paleontology and Evolution, 475-476 (Penguin, Middlesex, 1978). 409-418 (Dell. New York, 1966). (Freeman. New York, 19881. 22. Demeny, P. World Population and Us. Policy (ed. Menken, J.) 37. Cameron, A. G. W. Sky and Telescope 26, 258-261 (1963). 8. Stringer. C. B. Scient Am. 263, 98-104 (1990). 65 (Norton. New York, 1986). 38. May, R. M. Ecology and Evolution of Communities (eds Cody, 9. Cann, R. L., Stoneking. M. & Wilson. A. C. Nature 325, 31-36 23. United Nations Population Projections: Problems and Solutions, M. L. & Diamond, J. M.) 81-120 (Harvard Univ. Press, 'Cambridge, (1987). Workshop on Population Projections. Budapest, 1980, 118 MA,1963). 10. Wilson, E. 0., Biodiversity (ed. Wilson, E. 0.) 8 (National (United Nations. New York. 1981). 39. Preston. F. W. Ecology 29, 254-283 (1948). Academy Press. Washington DC. 1988). 24. Ehrlich, P. R. & Ehrlich, A. H. The Population Explosion 16-237 40. Preston. F. W. Ecology 43, 185-215 (1962). 11. Raup, D. M. Biodiversity (ed. Wilson, E. 0.) 54 (National Academy (Simon and Schuster. New York. 1990). 41. Preston. F. W. Ecology 43 410-432 (1962). Press, Washington DC, 1988). 25. Haub. C., Kent. M. M. & Yanagishita. M. 1991 World Population 42. Whittaker, R. H. Taxon 21, 213-251 (1972). 12. Raup, D. M. Science 231, 1528-1533 (1986). Data Sheet (Population Reference Bureau, Washington DC, 43. Darwin, C. On the Origin of Species, 1859 (reprinted by Harvard 13. Stanley, S. M. Proc. natn. Acad. Sci. 72, 646-650 (1975). 1991). Univ. Press. Cambridge, MA. 1964). 14. Dicke, R. H. Nature 192, 440-441 (1961). 26. Crick, F. H. C. & Orgel, L. E. Icarus 19, 341 (1973). 44. McFarland. D. Guinness Book of World Records 1991, 42 (Facts 15. Carter, B. Confrontation of Cosmological Theories with 27. Arrhenius, S. Die Umschau 7, 481 (1903). On File, New York, 1990). Observation (ed. Longair. M.S.) 291-294 (Reidel, Dordrecht, 28. GOld. T. Air Force and Space Digest, May. 65 (1960). 45. Leslie, J. Bull. Canad. nucl. Society, 10 (3), 10-15 (1989). 1974). 29. Dyson, F. J. Science 131, 1667-1668 (1960). 46. Leslie. J. The Philosophical Quarterly, 40, 65-72 (1990). 16. Carr, B. J. & Rees, M. J. Nature 278, 605-612 (1979). 30. O·Neili. G. K. Physics Today 27, 32-40 (Sept. 1974). 47. Leslie, J. Mind 101.403 521-540 (1992). 17. Barrow, J. D. & Tipler, F. J. The Anthropic Cosmological Principle, 31. Dyson. F. J. Rev. mod. Phys. 51, 447-460 (1979). 48. Nielsen, H. B. Acta phys. Polon. B20, 427-468 (1989). 231-248 (Clarendon. Oxford, 1986). 32. Linde, A. D. Inflation and Quantum Cosmology. 139-141