Many people believe in the existence of extraterrestrial life. The reasoning is quite simple: Given that we evolved on this planet, and there is such an enormous number of other planets, galaxies and stars out there, surely extraterrestrial life exists on a number of these planets? However, this is based on a number of dubious assumptions that fall apart when subject to examination.
Recently in the New York Times, Adam Frank claimed using the Drake equation (that calculates the number of active, communicative extraterrestrial civilisations in the Milky Way galaxy) that at least a trillion alien civilisations have existed. 
Taking that number at face value for a moment, let’s add another variable to the Drake equation: What is the probability that an active alien civilisation will do or have done something that allows us to verify their existence? Even if the probability is only one in ten million, that means we have verified the existence of 10,000 alien species already! The reason you and I don’t know anything about this? Obviously a CIA plot!
If that reasoning seems dubious to you, congratulations: you have become appropriately sceptical about probability calculations that pull numbers out of nowhere.
One of the biggest assumptions made when using the Drake equation is that all the steps involved are possible. If any of the steps involved are impossible, the number of alien species you get becomes zero, because zero probability multiplied by a billion trillion gazillion is still zero.
This includes the possibility of life evolving from non-life. How do we know that is scientifically possible? “But we got here!”, I hear you say. But that is merely assuming what you are trying to prove. Consider string theory: if you pointed out it has no evidence and someone said “But we have gravity and string theory explains gravity!” what would you say? It is not enough to aim to explain something: you have to show evidence you actually do. Abiogenesis is just another scientific hypothesis like any other: we should not be assuming it’s true unless there is a solid body of evidence to say so.
A statement of the problem
The theory of evolution describes a system of life that is responsible for all the diversity of life that we know. While the life forms that it produces change dramatically over time, the system itself is remarkably stable, barely changing since the beginning. For it to work, this system contains the right amount of flexibility to accommodate change, not so little that it would stagnate and not so much as to collapse.
The DNA system of life is shared by all life as far as we know. No other system of life has ever been observed.
If evolution describes a system of life, that makes abiogenesis the search for a system that produces systems of life – a system that, beginning with whatever was available in Earth’s early days, would produce the system of life that we are all familiar with. Like evolution, this system would have to have a core of stability as it produced the life-system of evolution, else it itself would collapse before it got the job done. Unlike evolution, it would not be able to use natural selection over a large part of the process.
A model of abiogenesis that is not in fact a system, but a large number of ad-hoc steps, stretches credulity and becomes far more difficult to prove.
Comparisons between evolution and abiogenesis inevitably bring out the difficulties of the latter. Evolution started with observations of Galapagos island finches and other animals. Abiogenesis started with nothing but the idea of abiogenesis. Evolution was figured out by a handful of scientists in the 1800s. Abiogenesis has hundreds of scientists investigating it. Evolution had a clear starting point. Abiogenesis could have started with a warm pond, thermal vents, one of Saturn’s moons or who knows where else. Evolution had the fossil record providing a plausible map to guide research. Abiogenesis doesn’t. If evolution research is like navigating a maze with a map, abiogenesis is like an indeterminate number of mazes, with no map, which might be 3D mazes, and you’re blindfolded.
Unlike evolution but like string theory, abiogenesis researchers are focussing on finding a plausible model of abiogenesis first and proving it actually happened later. String theory shows the downside of this approach. Plausibility does not imply actuality and a theory of abiogenesis that has plausibility only will be no better off than string theory.
The discussion so far shows the extreme difficulty involved in proving any theory of abiogenesis; certainly, thus far scientists have not come up with a single plausible model. Most hypotheses of abiogenesis are considered dubious in the extreme; most scientists in the field feel that the ‘RNA World’ model, that proposes an RNA replication system as a halfway point to the DNA system, is by far the most promising.
The ‘RNA World’ model has it’s critics, however. It has been criticised as ‘almost unfalsifiable’ and ‘full of holes’, even compared to a ‘creationist mantra’ (hardly a compliment in mainstream biology!). Certainly the problems involve seem quite daunting. If the ‘RNA World’ model is the best there is, yet ends up collapsing under the weight of it’s problems, perhaps it will be time to give up on abiogenesis altogether.
But wait: “Just because abiogenesis hasn’t been proven yet, doesn’t mean it never will!” you might respond. Of course that is true, but there is still a problem here – this attitude makes abiogenesis unfalsifiable.
Falsification and the limits of science
Scientific theories are supposed to be falsifiable because science can only make progress by discarding theories that produce no evidence. That is how science is supposed to work. In practice, falsification (or verification) of a theory takes a long time. This is because after a failed experiment, rather than abandon the theory, scientists will usually tweak it and try again. It is only after many such tweaks that scientists will draw a line under a theory and move to something more promising.
For a theory of any complexity, if you were so inclined you could simply keep tweaking and tweaking and tweaking forever, never giving up, as there is always a chance you are giving up too soon if you don’t. This makes falsification almost a pragmatic decision. It has been found that scientists are reluctant to consider a theory falsified unless there is an alternative. If no alternative has been presented, they will simply keep tweaking away.
This introduces a danger. Because of the philosophical presuppositions of most scientists in the Western world, they may be inclined to give abiogenesis much more time than it deserves, as there is no obvious scientific alternative to it. Given the huge amount of time and effort already spent investigating abiogenesis, we might be there already or well on the way.
The modern West assumes that given enough time, science can do anything. But there is no basis for assuming this. Science is a human activity; therefore it is limited by human brainpower and human technology. So even if you grant that science can answer all questions in principle, it does not follow that it can do so in practice.
When science was younger, scientists answered the low-hanging fruit questions that were easy for science to answer. As science gets older, the remaining questions will get more and more difficult and eventually only the impossible ones will remain, at which point science will become a bit like philosophy, running around in circles. I think we’re already there in some fields. I expect to see articles about the stagnation of science within 20-30 years.
If abiogenesis is not in fact an impossible question, it sure is acting like one. How you interpret that ultimately depends on the philosophical assumptions you bring to science, rather than the science itself. What is clear though is that making assumptions based on the viability of abiogenesis is unwise, being unsupported by evidence.