...I looked to these very skies,
And probing their immensities____
Robert Browning, Christmas Eve
Interstellar travel is certainly difficult, perhaps impractical, but not impossible. Even with our present level of technology mankind has succeeded in launching a craft that will some day drift out to the stars. Imagine an ETC with a technology only slightly in advance of our own; suppose its craft travel at the sedate speed of, say, c/40. Then, if the ETC makes one more technological advance — the development of Bracewell-von Neumann probes — it possesses a strategy to colonize the Galaxy. And quickly. * * *
of the many contributions to science made by von Neumann (a partial list is on page 28), the most important may have been in the theory of computing. He became interested in computing at Los Alamos, where he was in charge of the calculations needed for the design of the bomb. Crude calculating machines had been developed to help von Neumann's team in its tasks; after the War, von Neumann turned his mind to what was required of more general-purpose computing machines. His considerations led to many of the important principles of computing, and most of today's
computers — which are based on the general logical design and mode of operation he championed — are known as von Neumann machines.98
The questions involved in the design of a general-purpose computing machine led von Neumann to ask an even bigger question: What is life? As a step toward answering this, he developed the idea of a self-reproducing automaton, a device that could (a) function in the world and (b) make copies of itself. (Such a device is also sometimes called a "von Neumann machine," but this leads to confusion with the von Neumann machine — the architecture that is at the heart of present-day computers. I will use the term "self-reproducing automaton" when I refer to this hypothetical device.) In von Neumann's scheme, the automaton has two logically distinct parts. First, it has a constructor, which manipulates matter in its environment to carry out tasks, including the construction of units it can then use to assemble a copy of itself. A universal constructor has the capacity to make anything — as long as it has suitable instructions. Second, it has a program, stored in some sort of memory bank, which contains the instructions needed by the constructor.
An automaton can reproduce itself as follows: The program first tells the constructor to make a copy of the program's instructions and place the copy in a holder. It then tells the constructor to make a copy of itself with a clear memory bank. Finally, it tells the constructor to move the copy of the program from the holder to the memory bank. The result is a reproduction of the original device; the reproduction can function in the same environment as the original and is itself capable of self-reproduction.
Of course, von Neumann did not give explicit details of how to build a self-reproducing automaton. Even today, we are far from being able to build such a device (although the seeming convergence of several technologies suggests that we may be able to do so in a few decades). What von Neumann was interested in was the logical underpinnings of self-reproducing systems, rather than any particular mechanism for achieving reproduction. In a lecture first given in 1948, he discussed the relevance of self-reproducing automata to the question of life. He argued that a living cell, when it reproduces, must follow the same basic operations as a self-reproducing automaton. Within living cells, there must be a constructor, and there must be a program. He was right. We now know that nucleic acids play the role of the program, and proteins play the role of the constructor. All of us are self-reproducing automata. (We discuss the function of nucleic acids and proteins later; see page 189.) What concerns us here is not what von Neumann's self-reproducing automata might tell us about life. Rather, it is how to use such automata to colonize the Galaxy. Frank
Tipler outlined a possible scenario.
First, we must remember that the transport of intelligent beings to investigate planetary systems would be expensive: food, water, life support — all these items are necessary, but require energy to transport. Probes do not have this problem. Indeed, this is why Crick's motto for directed pansper-mia was "bacteria go further"; a small probe filled with a payload of bacteria would be cheaper to build and propel, and would enable an ETC to seed the Galaxy. With probes we are on the right track; but a bacteria-filled probe is of little use to an ETC wanting to explore and learn about the Galaxy. For an inquisitive ETC, it makes more sense to launch Bracewell-von Neumann probes. (These devices are usually called simply von Neumann probes in the literature. However, to the best of my knowledge, von Neumann never considered the possible uses of probes in interstellar exploration. The first person to suggest that probes would be useful for interstellar exploration and communication was Ronald Bracewell.99 It seems reasonable, therefore, to refer to these devices as Bracewell-von Neumann probes.)
In Tipler's scenario, a Bracewell-von Neumann probe can be small: the payload need be nothing more than a self-reproducing automaton — one with a universal constructor and an intelligent program — and a basic propulsion system for use within the target system. After arriving at the target star, the program instructs the probe to find suitable material with which it can reproduce itself and make copies of the propulsion system. (If the planetary system resembled our own, then there would be plenty of raw material available for the constructor; asteroids, comets, planets and dust could all be broken down and utilized.) If necessary, radio signals from the home planet could send revisions to the program, so that the probe would never be out of date. Soon after arrival there would be a host of probes, each undertaking some pre-programmed task. Some might explore the planetary system, sending back scientific data to the home world. Some might construct a suitable habitat for later colonization by the home species. Some might even raise members of the original species from frozen embryos stored as part of the payload. And some would travel to another star, where the process would be repeated until every star in the Galaxy had been visited.
If probes traveled between stars at the rather stately speed of c/40, and if the propagation of the probes was directed rather than random, then a colonization wave could surge through the Galaxy in roughly 4 million years — a period that equates to just 2 hours 46 minutes of the Universal Year. figure 29 Ronald Bracewell has long been an advocate of seti. This time is shorter He was also the first to suggest the use of probe technology as a than the colonization means of exploring the Galaxy.
times in the models of Newman and Sagan, and Fogg, but this is to be expected. Probes need not stay in a planetary system and wait for colonists to give them instructions on how to proceed: they already have their instructions. The Galactic colonization time is short because the process is planned to be efficient. Not only is colonization by probe quick, it is cheap. An ETC simply has to send the first few probes; after that, the Galaxy picks up the tab in terms of providing raw material for the continuing process.
Can such probes be built? Well, intelligent self-reproducing automata are certainly possible: Nature has already built them in the form of human beings. (As John Watson points out, humans are a good example of what
we expect of a certain type of Bracewell-von Neumann probe! Perhaps we are not a "natural" species, but the probe technology of some advanced ETC?) Whether mankind can match Nature's accomplishment, or perhaps improve upon it and build better self-reproducing automata, is unknown. Certainly there are significant technical and engineering hurdles to overcome before we can build Bracewell-von Neumann probes. But even if mankind is not bright enough to develop probe technology, surely a technological civilization thousands or millions of years in advance of us could build probes. There seems to be no theoretical reason why they could not.
Colonization of the Galaxy by probe is technologically possible; it is quick; and it is cheap. Even if the aim is contact rather than colonization, Bracewell showed there are circumstances in which probes are more effective than radio signals. So as Fermi would ask: where are the probes?
We touched on this question in Chapter 3, when we discussed the possible use of probes in directed panspermia and when we considered places where a monitoring probe could hide. But such probes are not the Bracewell-von Neumann probes that can dismantle planets, undertake as-troengineering projects, and colonize the Galaxy in the cosmological blink of an eye. There is no evidence of such probes ever having visited the Solar System, nor is there any evidence for their activity elsewhere in the Galaxy.
Even if an ETC has the ability to construct Bracewell-von Neumann probes, perhaps it would choose not to deploy the technology. There are risks, after all. The probes reproduce (like living beings) rather than replicate (like crystals), so inevitably there will be reproductive errors. There will be mutations. Probes would evolve, just as biological creatures evolve. The Galaxy could soon be home to different probe "species," each with its own interpretation of its goals. There would be a risk, for example, of a probe returning to the home system and failing to recognize it; not good news for the ETC if the probe's orders are to dismantle planets and use the material to construct something else. But is it a risk every ETC declines to take, and a problem every ETC fails to solve?
Since colonization of the Galaxy by probe seems straightforward, some authors argue there is a strong motivation for an ETC to engage in colonization: if species A does not do it, species B will. Stake your claim early, in other words. This sort of argument might have appealed to von Neumann, who was a strong proponent of the nuclear first strike. (In an interview with a Time magazine reporter, von Neumann said: "If you say why not bomb them tomorrow, I say, why not today? If you say five o'clock, I say at one o'clock.") We must be grateful that, in the 1950s and 1960s, wiser counsel than von Neumann's prevailed. Perhaps we can hope intelligent species develop to the stage where they have no urge to own every star, inhabit every planet, and populate the Galaxy with beings just like them selves. Nevertheless it takes only one ETC to reason that it should not take the risk of losing out on all that real estate.
A discussion of Bracewell-von Neumann probes is relevant to any discussion of the Fermi paradox, but you may ask why I present it in a part of the book devoted to solutions of the paradox. A surprising number of people seem to believe that probe technology does resolve the paradox. They argue that we do not see aliens because aliens would send probes rather than travel interstellar distances themselves. Of course, this entirely misses the point. Fermi's question refers either to aliens or the product of alien technology. After all, if we detected an object in space that was clearly artificial yet not man-made, then presumably we could deduce the existence of an extraterrestrial civilization that constructed the object. We see no evidence of aliens nor of their probes. Far from resolving the paradox, the possibility of Bracewell-von Neumann probes provides the paradox with real bite.
Was this article helpful?