Will we ever find ETs out in the Universe? No – but we might find them inside a computer!
Where is everybody?
Enrico Fermi, 1950
The fact that we have not yet found the slightest evidence for life – much less
intelligence – beyond this Earth does not surprise or disappoint me in the
least. Our technology must still be laughably primitive, we may be like jungle
savages listening for the throbbing of tom-toms while the ether around them
carries more words per second than they could utter in a lifetime.
Odyssey: the Unauthorised Biography of Arthur C. Clarke, Neil McAleer, 1992
Extraterrestrial intelligences are out there. But the people involved in the search for them are going about it entirely the wrong way. Even if they were going about it the right way, the task would be nigh on impossible. And, even if it was not impossible, and ‘contact’ was miraculously made, there would be nothing much we could learn from extraterrestrials that we could not discover ourselves simply by playing about on a desktop computer!
This is the controversial opinion of Stephen Wolfram, evangelist of the idea discussed earlier that nature has created all the bewildering variety of the Universe simply by repeatedly applying a simple complexity– generating computer program.* Wolfram is not saying this just to be controversial. Once upon a time, he was an enthusiast for SETI, the search for extraterrestrial intelligence. ‘It’s just that the work I’ve done over the past twenty years has taught me that SETI as currently practised is not the most sensible thing to do,’ he says.
Of course, it is not at all obvious that the sensible thing to do is to abandon any systematic scanning of the heavens for signs of extraterrestrials and instead look inside a computer. In fact, the suggestion seems utterly mad. But Wolfram has his reasons. And the first step in appreciating those reasons, he says, is to understand what is wrong with the current strategy for finding extraterrestrial intelligence.
Take the search for alien signals. Electromagnetic waves are continually raining down on the Earth from the Universe. These span a vast range, or ‘band’, of frequencies.† For instance, there are radio waves – the most sluggishly oscillating waves – and gamma rays – the most rapidly oscillating waves. The visible, or ‘optical’, light picked up by the human eye is pretty much in the middle of this electromagnetic ‘spectrum’.
SETI scientists focus their attention on the radio and optical bands for the simple reason that both types of wave can travel across large swathes of the Galaxy without being absorbed and that both can easily penetrate the Earth’s atmosphere. In effect, SETI researchers look for the interstellar equivalent of a radio station. This is an electromagnetic wave which spans a narrow range of frequencies and which carries information – the information is impressed on the ‘carrier wave’ by continuously varying, or ‘modulating’, its frequency or amplitude.*
In short, what SETI scientists do is sift through the chaotic babble of radio ‘static’ coming from space in the hope of finding an interstellar radio station – a signal with some kind of regularity, some kind of pattern. ‘But a patterned signal is an inefficient signal,’ says Wolfram.
Take, for instance, a patterned signal which consists of a 1 followed by a 0 a billion times over – in other words, 1010101010 … It would take ages to send. It would be terribly inefficient. However, the pattern can be exploited to compress the signal down to something far more efficient that can be transmitted far more quickly: ‘Repeat 10 a billion times’.
The trend towards more efficient communications is evident today in terrestrial communications. The need to cram as much information as possible into a signal has led to communication technologies in which information is sent using a vast number of frequencies simultaneously. Take CDMA – Code Division Multiple Access – the cell phone standard used in North America. Here, the carrier, rather than being a periodically undulating wave, is a pseudo-random sequence of binary digits – something a lot more complicated. ‘The key point is that the signal being transmitted has very little pattern – it looks almost random,’ says Wolfram.
There is a lesson here for SETI, says Wolfram. Because a patterned signal is an inefficient signal, it is very unlikely to be used by an advanced technology. Instead, Wolfram maintains, the transmissions of an advanced civilisation will be as far from the patterned signals that people are searching for as it is possible to imagine. In fact, the transmissions will look pretty much like the random radio ‘static’ which comes from astronomical objects such as stars and interstellar clouds of gas. ‘Distinguishing a signal from an extraterrestrial intelligence from the natural background of cosmic signals is going to be extremely difficult,’ says Wolfram.
This may explain why SETI, after more than four decades of effort, has drawn a disappointing blank. As the Italian-American physicist, Enrico Fermi, asked: ‘Where is everybody?’ If Wolfram is right, extraterrestrials are out there but they are a lot more difficult to spot than anyone ever imagined.
In fact, Wolfram can see only one way to distinguish a signal from an extraterrestrial intelligence from the natural background – and that is to use a tedious process of elimination. Consider the radio emissions our radio telescopes pick up from our Galaxy as a whole. This consists of the sum total of the radio waves emitted by all the objects in the Milky Way – for instance, the radio static from the nearest star, Alpha Centauri; the radio static from the Crab Nebula; the radio static from the clutch of newborn stars in the Pleiades; and so on. According to Wolfram, we have no choice but to compare the sum of this static with the static we pick up from the Galaxy as a whole. ‘Only if there is some radio static left over can we say we may have picked up an intelligent signal,’ says Wolfram.
The ‘may’ here is the operative word. The leftover radio emission might actually have a natural cause but one which we have not yet identified. ‘Even if we were to pick up a cosmic radio signal containing the digits of pi, it is always possible that a natural process could be generating it,’ says Wolfram.* According to Wolfram, we may never be able to exclude the possibility that a signal we have detected is natural. ‘It is very unlikely that the question “Is this an intelligent signal?” will have a simple yes/no answer,’ he says.
Of course, it is always possible that an advanced extraterrestrial intelligence decides to talk down to backward civilisations like ours. So, despite using efficient, pretty much random, signalling for its own communications, it chooses to broadcast just the kind of inefficient, patterned signal we would easily recognise. ‘They might use a beacon that, by design, is energy inefficient, directional and easily detected by our current methods,’ says Paul Shuch, Executive Director of the SETI League in New Jersey.* Max Tegmark of the Massachusetts Institute of Technology can imagine such an inefficient signal being broadcast for a deeply sinister reason. ‘Say an extraterrestrial civilisation wanted to infect the computer network of a target planet with a devastating computer virus,’ he says.
Wolfram thinks Tegmark and Shuch’s scenarios are very unlikely, however. And for a simple reason. Any extraterrestrial intelligence that specifically tailors a signal for a civilisation like ours will be tailoring it for a civilisation that has developed radio communication but has not yet developed efficient radio communication. ‘That’s an awfully brief window in time,’ says Wolfram.
What Wolfram means is that technological civilisations are not likely to be at such a stage for very long. So, at any particular time, there are likely to be very few civilisations in the Galaxy scanning the heavens for such a signal and, consequently, very little chance of the signal attracting the attention of its intended recipient.
Shuch, however, sees another possibility even if extraterrestrials do not deliberately broadcast a signal of the type that SETI scientists are looking for. ‘They might broadcast such a signal inadvertently,’ he says. There may be emerging civilisations, pretty much like our own, he says, which waste energy by using inefficient, ‘narrow-band’, telecommunications. Inevitably, some of this electromagnetic chatter will leak away into space. ‘Our leakage is currently detectable out to a distance of about 100 light years,’ says Shuch. ‘Theirs will be too.’
He concedes that most of our cosmic neighbours are likely to be far more advanced than us and so will not be using the kind of primitive communications we would recognise. ‘Nevertheless, there are about 100 billion stars in our Milky Way,’ he says. ‘It stands to reason that there must be a few civilisations at our level, plus or minus a century or two.’ Wolfram, however, sees the same problem as before. There are likely to be far too few civilisations out there at roughly our stage of development, making the chance of our detecting their electromagnetic leakage entirely negligible.*
However, deliberate broadcasts and inadvertent broadcasts may not be the only way that a technological civilisation gives away its presence. According to Seth Shostak of the SETI Institute in Mountain View, California, even a highly advanced society will want early warning of comets and asteroids that might be involved in a catastrophic collision with their home planet. ‘And the best way to do this is to sweep interplanetary space with powerful radars,’ says Shostak. Shuch points out that the radar signal necessary to locate city-sized chunks of rocks and ice flying through space is quite different from a telecommunications signal. ‘By its very nature, it is inefficient and detectable from a long way off,’ he says.
The Achilles’ heel of this idea is once again that an extraterrestrial civilisation might use such a radar warning system for only a brief period of its history. Perhaps they will find a more effective means than radar to alert them to rogue comets and asteroids. Or perhaps they will vacate planetary surfaces for interstellar space. If the use of planetary defence radar is brief, it follows that the chance of SETI searches finding them is also slim.
Wolfram believes that, if extraterrestrials really want to communicate with us, there is a simple and effective way to get our attention. ‘Why not simply use a big flash of light?’ he says. ‘It could be seen directly simply by looking up at the sky. There would be no need to rely on the recipients of the signal having constructed complicated radio telescopes.’
So much for picking up an ET signal. What about locating an extraterrestrial artefact? In 2005, Luc Arnold of the Observatoire de Haute-Provence in France suggested searching for extraterrestrial intelligence by looking for peculiar structures such as giant triangles orbiting nearby stars.* In the past few years, astronomers have detected planets which happen to pass in front of their parent star by the way such ‘transits’ temporarily dim the star. Arnold points out that, if an unusual-shaped structure were orbiting a nearby star, it would dim the star’s light in a way distinctly different from a planet. Such a structure might be part of the normal infrastructure of an alien civilisation – it might, for instance, be a giant collector of solar energy. Or it might be an artefact fabricated specifically to signal to civilisations around other stars in the Milky Way.
Wolfram, however, takes issue with Arnold. For pretty much the same reason he believes aliens would not use simple, patterned alien communication signals, he does not believe they would build simple, patterned artefacts. ‘They are inefficient,’ he says. ‘They are very unlikely to be used by an advanced technology.’
To illustrate his point, Wolfram cites the example of trains arriving at a train station, say, every half an hour. According to Wolfram, even if we were to look down on the station from a great height – so high that we could see none of the details of the trains – we would still recognise their behaviour as artificial. The giveaway would be the regularity. ‘However, this kind of regularity turns out not to be a necessary feature of a mass transit system,’ says Wolfram. ‘It is merely a reflection of the current primitive state of our transport technology.’
A more efficient transport system, according to Wolfram, would consist of small taxicabs shuttling back and forth on demand. Controlling them would admittedly require more computing power. But more computing power, according to Wolfram, is synonymous with greater efficiency. ‘In the future, when our technological artefacts are continually communicating with each other and are able to carry out computations, the world will look a very different place,’ he says. ‘It will be far more complicated, far less patterned than it is today.’
At the present time, it is easy for us to distinguish between artificial objects and natural objects – for instance, between buses and trees. Nature’s artefacts are without fail more complex. ‘But this will not always be true,’ says Wolfram. ‘In the future, all our technological artefacts will be as complicated as nature’s – if not even more complicated.’*
It follows that the artefacts of an advanced civilisation will look as different from the giant space triangles envisioned by Arnold as it is possible to imagine. In fact, according to Wolfram, they will look pretty much like natural objects. This prompts an extraordinary question – to which Wolfram supplies an extraordinary answer. Could the stars be artificial? ‘Yes, in the abstract,’ says Wolfram. ‘It is extremely difficult to rule out the possibility of them being built for a purpose.’
Extremely difficult but perhaps not impossible. Wolfram points out that, compared with biological entities which have evolved by natural selection, artefacts created by intelligence for some purpose tend to do fewer things which are not related to their principal purpose. A plane, for instance, does fewer spurious things than a bird. A bulldozer does fewer spurious things than the muscles of a human being. ‘Artefacts created by intelligence are optimised for their purpose,’ says Wolfram. ‘On the other hand, it is inconceivable that a tree is optimised for its purpose.’
If Wolfram is right, in order to figure out whether or not an extraterrestrial artefact or signal is of intelligent origin we must first guess its ‘purpose’, then see whether it does anything spurious to that purpose. But how do we guess the purpose of a signal or artefact? ‘With great difficulty,’ says Wolfram. ‘In fact, I am not sure that it is possible at all.’
In fact, things are even worse than this. ‘What do we even mean by “purpose”?’ says Wolfram. ‘This is a deep philosophical question.’ And, according to Wolfram, the whole field of SETI is fraught with such philosophical difficulties. ‘We are looking for “intelligence” but how do we even define “intelligence”?’ he says. ‘For that matter, how do we define “life”?’
The central qualities of life appear to be the ability to move about, compete for resources, reproduce, pass information from generation to generation, and so on. Living things also use characteristic ‘biological molecules’, including DNA, the genetic material; proteins, the scaffolding and chemical engines of cells; lipids, the molecules of cell walls; and carbohydrates, the fuel of life. ‘But all of these things really are just characteristics of “life as we know it”,’ says Wolfram. ‘They are specific properties of life on Earth – we have no idea of the general properties of life.’
So, what characteristic might fit the bill as a bona-fide general property of all life? Racking his brains, Wolfram can come up with only one: complexity. Unfortunately, if there is one thing Wolfram’s exhaustive research over the years has revealed, it is that complexity of the degree seen in living things is not unique to living things. On the contrary, it appears to be exhibited in a vast range of other physical phenomena, from subatomic particle collisions to turbulence in a fluid to the circulation of the Earth’s atmosphere. This discovery Wolfram has elevated to the status of a universal principle. Put crudely, the ‘Principle of computational equivalence’ says that systems of similar complexity are equivalent. In other words, a system which is as complex as a living thing – for instance, the circulation of the Earth’s atmosphere – has exactly the same right to be classed as a living thing as you or me.
If defining life is difficult or impossible, defining intelligence is every bit as challenging. Take birdsong, says Wolfram. Zoologists have been able to say that a particular cluster of brain cells, or neurones, is responsible for a particular birdsong. They have then taken this as evidence that the birdsong, though undoubtedly complex, is not a manifestation of intelligence. ‘The trouble with this argument is that each and everything we do – from reasoning to exercising creativity – is in principle traceable to discrete sets of neurones,’ says Wolfram, ‘What quality do we have that a bird does not have that makes us intelligent? The answer is not at all clear to me.’
With all the tremendous difficulties in recognising intelligent signals and artefacts – and even in recognising ‘intelligence’ at all – should the people doing SETI give up and go and work in petrol stations? Wolfram is not actually saying this. He is simply saying that SETI is going to be a lot harder than anyone imagines.
At present, SETI scientists look for patterned radio and optical signals not because this is the best thing they can do but because it is the only thing they can do. Two pioneers of SETI, Guiseppe Cocconi and Philip Morrison, said it all: ‘The probability of success is difficult to estimate; but if we never search, the chance of success is zero.’* Wolfram understands this and that SETI scientists have little choice but to search in the way they do. Nevertheless, he thinks that, if anything worthwhile is to be achieved, the search strategy will have to be made a lot more effective.
At present, SETI scientists analyse the morass of data they collect from the sky with a technique known as ‘Fourier analysis’. This picks out any patterns in the form of periodic signals. Wolfram, however, thinks people should also look for more subtle patterns in the data. He gives the example of testing whether or not a binary number – a sequence of 0s and 1s – is random. The first step is to see whether the number has as many 0s as 1s. A random number will have 50 per cent of each. The next step is to see whether there are any unusual blocks of 0s or 1s – say, nine 0s in a row. A random number will not have any such blocks. ‘The point is that something like twenty tests are commonly used to check for randomness,’ says Wolfram. ‘In the same way, I think SETI signals should be analysed by a whole battery of tests. Relying on just one – the Fourier test – is really weak.’
Those in the SETI community have no problem with this. ‘I agree wholeheartedly,’ says Shuch. Shostak points out that there is already a more sophisticated approach now being used by SETI researchers in Bologna, Italy: ‘It’s called the Karhunen-Loeve Transform.’ KLT promises to dig far deeper into the ‘noise’ than Fourier analysis. ‘It stands a chance of detecting not only periodic patterns but also “aperiodic” patterns – ones which are not quite periodic,’ says Schuch.
Nobody denies that SETI faces an enormous challenge. It is just that Wolfram thinks the challenge is even more enormous than anyone else does. But what if his pessimism is misplaced? What if, against all the odds, SETI scientists succeed in recognising an intelligent signal, an intelligent artefact, intelligence itself? Without doubt this will be a momentous day for the human race. However, Wolfram sees a problem even here. ‘If we make contact with extraterrestrials, what could we possibly trade with them?’ he says.
The only thing Wolfram can think of is computer programs that can do useful or interesting things. He imagines the existence of an abstract ‘computational universe’ that contains all possible computer programs. When we find computer programs that do useful things – for instance, encrypt data or run spreadsheets – according to Wolfram, we are simply plucking these programs from the computational universe.
Even technological artefacts can be thought of as computer programs, according to Wolfram. In fact, he says that in the future there will be absolutely no distinction between computer programs that do things and physical artefacts that do things. This is because the electrons that shuttle about computers carrying out computations are the very same entities that glue together the atoms of matter. One day, Wolfram says, we will devise ‘universal constructors’ whose computational output will not simply be sequences of binary digits but actual physical artefacts – objects constructed from the Lego bricks of atoms. Arguably, nature already possesses such universal constructors.*
Wolfram points out that extraterrestrials will have access to the computational universe just like us. And this fact may severely limit what we can trade – what they can learn from us and what we can learn from them. ‘After all, what could they possibly tell us except “We’ve done more computations than you and here are some good computer programs we’ve found”?’ says Wolfram.
Of course, we may discover extraterrestrials who are far in advance of us and have explored much more of the computational universe. Leaving aside the issue of why they might want to give us the benefit of their experience when there is so little in it for them – how benevolent do we feel towards pigs, or bacteria? – might we not gain immeasurably from them? Yes, we might indeed. However, this is beside the point, according to Wolfram. The question is: How long is it going to take to find such an extraterrestrial civilisation compared with how long is it going to take to search the computational universe for their knowledge? According to Wolfram, there is no contest. Searching the computational universe is going to be quicker.
It has often been said that, if a monkey sat at a keyboard and hit the keys randomly, eventually it would write the complete works of Shakespeare. Of course, to stumble on the correct sequence of letters, spaces and punctuation, the monkey would have to sit at the keyboard for many times the present age of the Universe. Wolfram’s key discovery, however, is that simple computer programs can create fantastically complex outputs. In other words, there exists a program which is enormously simpler than the complete works of Shakespeare that can generate the complete works of Shakespeare or, at the very least, generate an output in the style of Shakespeare.
The program for writing the complete works of Shakespeare will not be as difficult to find in the computational universe as the complete works of Shakespeare in the universe of all possible letter sequences typed by monkeys. Similarly, other interesting programs will be easier to find than might be expected. Since this is the case, why go to the trouble of searching the physical Universe for aliens to tell us about the computer programs they have found, asks Wolfram? Not only will we have to search 100 billion star systems in our Galaxy but it will be a darn sight harder than sitting at a computer on Earth and searching for the programs ourselves. ‘It’s more efficient for us to use a computer to simply search the computational universe,’ he says.
Wolfram has already made a start at exploring the computational universe. In his monumental book, A New Kind of Science, he carries out a systematic search for all simple computer programs that have complicated and interesting outputs. According to Wolfram, the search for computer programs that do useful things in the computational universe is simply the latter-day equivalent of the search for useful commodities such as spices in the real world. Except, of course, it might take a lot longer.
Another logical consequence of Wolfram’s view is that extraterrestrial intelligences themselves, just like human beings, will merely be particularly complicated computer programs. In other words, they will exist in the computational universe, as we will. So, instead of scanning the heavens for alien broadcasts, we could look closer to home. Much, much closer: your office PC or laptop. You could literally have an alien at your table! Of course, if we find ET in a computer, it will not be a flesh and blood alien – it will be a cyber version. (Unless of course we have acquired universal constructors and can build one from the digital blueprint.) But this does not mean that we cannot communicate with it. We could still converse with a virtual version of an alien civilisation, and learn plenty from the conversation.
Of course, all the knowledge we might gain from extraterrestrials – surely the principal reason for trying to contact them – is already ours for the taking. It is in the computational universe. We do not have to find extraterrestrials and ask them for it – unless, of course, we are feeling particularly lonely and would like someone to talk with. If we can find computer programs for aliens, we can find computer programs for what they know. We can cut out the middleman.
Wolfram is not saying do not search the physical universe for somebody to talk to. ‘I’m merely saying we can look for extraterrestrial intelligence in the physical universe or we can look for extraterrestrial intelligence in the computational universe,’ he says. ‘But the task is a whole lot easier in the computational universe and there’s a lot more there.’
* See Chapter 2, ‘Cosmic Computer’.
† The frequency of a wave is simply how fast it oscillates up and down. Normally, this is measured in Hertz (Hz), where 1 Hz is 1 oscillation per second.
* The amplitude of a wave is its maximum excursion from its average level. Think of it as the wave’s height.
* This alludes to the fact that the number pi, though it appears complex, its digits never repeating, can in fact be generated by a very simple program. Since Wolfram believes that the secret of nature’s complexity is the repeated application of simple programs with complex outputs, it is conceivable that there is a natural process somewhere in the Universe that is currently employing the pi-generating program.
* The SETI League is an international band of radio and radio astronomy enthusiasts, dreaming of the day when one of them will catch ET phoning Earth. To this end, they are using satellite dishes and signal-processing software to listen out for extraterrestrial signals from their back gardens. (The SETI League, 433 Liberty Street, PO Box 555, Little Ferry, NJ 07643, USA; http://www.setileague.org.)
* Interestingly, a giant array of telescopes, being planned for operation in 2020, will be capable of picking up TV and radar broadcasts from planets round nearby stars. The Square Kilometre Array, which could be built in Argentina, Australia, South Africa or China, will consist of a cluster of dishes within five kilometres of each other, with outriders up to 3,000 kilometres away. With 100 times the collecting area of any existing telescope, just keeping the dishes working together as one will require data rates 100 times greater than the total UK Internet traffic in 2005!
* See ‘Transit Lightcurve Signatures of Artificial Objects’ by Luc Arnold (Astrophysical Journal, vol. 627, p. 534). Also: http://xxx.lanl.gov/abs/astro-ph/ 0503580.
* The reason our artefacts will eventually be more complex than nature’s is that natural selection is more restricted than human inventiveness. Rather than making radical changes to a creature, natural selection tends to make incremental changes – lengthening a neck or increasing the size of a brain. Human technological inventions are subject to no such constraint.
* ‘Searching for Interstellar Communications’ by Giuseppe Cocconi and Philip Morrison (Nature, vol. 184, no. 4690, pp. 844‒6, September 1959).
* Cells take molecules such as amino acids as their input and, after carrying out the equivalent of a computation, produce as the output proteins, biomolecules that can perform a vast array of tasks from providing the scaffolding of cells to speeding up chemical reactions crucial to life. Nature’s universal constructors, however, are constrained by natural selection and so cannot explore every nook and cranny of the computational universe. The universal constructors we may one day build will suffer no such constraint.