Serendipity
“In my experience, there’s no such thing as luck”
- Obi Wan Kenobi, Star Wars.
- Obi Wan Kenobi, Star Wars.
|
Abstract: It has been claimed that many of the great leaps forward in astronomy have been made serendipitously - someone just happened to be in the right place at the right time. Here I undertake an analysis of some of the most commonly cited examples of "luck" in astronomy and argue they were no such thing - but rather attention directed in the right place; careful observation and problem encountering and solving where required. Luck is either everything, or it is nothing. I argue here for the primacy of the human mind over mere "luck" in pushing science forward.
Part 1: De nova stellar Sundown. November 14, 1572. The sky darkens but another log on the fire to keep them warm would brighten the library. Those who did not head to bed would stay here. It was the warmest place now that the kitchen was closed. Herrevad Abbey was comfortable enough and now that his uncle Sten had paid for a paper laboratory to be built, Tycho had a place he could retreat to when the other workers would stay late drinking in an attempt to keep the winter cold at bay. The laboratory was one of the darkest places of all - for Tycho did not permit any candles to be carried in there and so workers would have to feel along the walls to find the covered lamps. Most could not be bothered and kept themselves in other parts of the abbey that were brighter. But here Tycho could watch the skies and keep his most recent notebooks hidden among the fresh paper. And so here Tycho found himself once more - another November night. Perhaps as cold as most, as dark as most and as clear as most winters’ nights in olden Denmark. Here in a monastery; a monastery without monks praying to the heavens but rather a scientist seeking to understand them. Tycho searched beneath a pile of fresh paper. Much of this was destined for printing presses. But not this pile. This pile was locked in a chest so as to be hidden from his students. Young Johannes was particularly nosey. And in the middle of this pile were Tycho’s most recent observations. Tonight Jupiter and Mars would be high in the sky - he checked the coordinates. The moon would not rise until after midnight - the sky would be dark. He would have to be careful - a step at the rear of the laboratory had come loose - Johannes had again forgotten to have the staff repair it. Last night he almost tripped in the pitch of night. So tonight he thought better of it tonight and unhooked one of the lamps fixed to the wall and guided himself over the threshold stepping over the rickety paving. He looked up - it was a reflex once the Sun was down and he stepped outside for the first time. He looked up and suddenly felt dizzy. He whirled East. There was Jupiter. There was Mars. He turned north west. The Sun’s final rays glimmered and the tail of the scorpion could be glimpsed. South again. He dropped the lamp. Mouth agape, he stood looking at the sky. The seconds ticked by. Then the minutes. He counted the stars in Persus. Then Andromeda. Cepheus. Lacerta. The constellations were as they should be. But Cassiopeia. Cassiopeia was changed. A bright star. A new star. Venus? It must be Venus. No. It’s no where near where Venus is. Or could be. And...it was brighter than Venus. Tycho finally looked down. How much time had passed he could not say, but now all light from the Sun was completely gone. And yet...he could see his notes in his hand and read his handwriting and the lamp had burned out. He ran calling into the monastery “Come, come quickly, something amazing has happened!” u u u u u u Tycho Brahe’s observation of a supernova on November 15th, 1572 from an ex-monastery in what is now Southern Sweden changed human understanding of the heavens forever. Once properly appreciated through careful analysis, this “new star” changed the unchangeable. The heavens were not perfect and immutable for Tycho new that this “new star” was located far beyond the moon in a region that was supposedly perfect and so not capable of change (Krause et al, 2008). But the heavens could change. Aristotle had been falsified. Tycho discovered this. Seemingly, as luck would have it, the great observer of the heavens observed what only very few human beings have seen with the naked eye in recorded history: the violent death of a super massive star. Only around 8 such events have been observed with the naked eye in the last thousand years (PlaitWeb). To be a person fascinated by the heavens and witness such an event is rare indeed. As the question before us suggests - in the case of “discovering” supernovae, someone must be in just the right place at the right time. Tycho it would appear, was lucky. Galileo never witnessed a supernova. Nor did Copernicus or Newton (Keplar did however). Only a serendipitous stroke of fortune seemingly put Tycho in the right place at the right time. He had spent his whole life preparing for this moment. Tracking the motions of the planets and stars as they wheeled overhead. He knew what he was looking at perhaps more than anyone else in history before him. It is likely there had never been a human better placed to notice even the slightest change in the heavens. But it would not have taken a Tycho to notice this change. Indeed no fewer than five other astronomers contemporaneous with Tycho observed this star and indeed even a layperson would have noticed this dramatic new light in the sky. One can only imagine the religious hysteria that would have been stirred up by this event. What distinguished Tycho, however, was the accuracy of his measurements. He was an observer without peer when it came to recording celestial data with the naked eye. It did take work however for Tycho to establish the true significance of the event. After all, the Aristotelean world view would permit only objects closer than the moon to undergo change. This new star was a change. Ergo: This new star must have been “sub lunar”. And yet, painstakingly accurate measurements led Tycho to the inescapable conclusion: this new star was among the heavens. It was far, far more distant than the moon. Although it seems luck brought the discovery to Tycho, it required work for Tycho to bring the discovery into scientific understanding. Part 2: Serendipity Serendipity is defined by the Oxford English Dictionary as “the occurrence and development of events by chance in a happy or beneficial way”. We may see that throughout the history of astronomy such happy or beneficial chance events seem common. Some have even claimed that astronomy itself can be seen as a science of serendipity (see for example Fabian (2009) and Lang (2010)) To remain with just one flavor of serendipitous discovery; the appearance of supernovae might seem to be a rare event in astronomy, but with the advent of modern technology, a professional observer in the business of cataloging these explosions might think it a poor day indeed were less than one per week recorded by computerized measuring devices. In 1998, the High-z Supernova Search team reported observations on 10 very distant supernovae and were expecting to uncover evidence about the extent to which our universe was open, closed or flat (Riess et al, 1998). This team was effectively in a race with the Supernova Cosmology Project to determine the mass density of the universe. What neither team, nor any in the astronomical community at the time expected to find was an accelerating rate of expansion. But that is precisely what both teams did find. However someone has to be first and for the Hi-Z Supernova Search Team the stars aligned and they have been awarded the proverbial blue ribbon. Both teams were equally hard working, and operating with similar equipment, technology and expertise. The difference in who gets the credit for first discovering an accelerating expansion? That falls to simply being in the right place at the right time - surely this too is a sign of serendipity at work in discovery? And yet, as with Tycho’s Supernova, the act of observation alone is not enough for such occurrences to be called scientific. To be scientific the discovery must form part of a theory for which there is evidence. When Tycho or the High-Z team observed their supernovae, it was not until the data had been explained could the observations be considered part of science. If they were, then all observations are equally a part of science if all that matters is observation. Plainly that would be a philosophy of scientific nihilism. Although we do not want to needlessly constrain what we mean by the word science having a definition that is as broad as to include any observation is unhelpful. Much hard work needs to be done on the observations of supernovae to conclude that they lay among the stars and that the expansion of the universe is accelerating. Serendipity then would appear to play a minor role. In terms of Tycho’s observation of De stella nova the appearance of serendipity seems to be little more than blind luck initially. Anyone alive under that supernovae in 1572 was provided with a similar dose of luck, it would appear. So what differentiated them from Tycho? Hard work, a corpus of scientific measurement and (most of all) the ability to explain something that no one else was able to: this new star really was in the distant heavens. Let us turn our attention to an example from the history of astronomy that seems to fit a more textbook, classical notion of the role of serendipity in astronomy. In 1965 Arno Penzias and Robert Wilson were radio astronomers engaged in mapping signals from the Milky Way. They were not cosmologists as such and were certainly not actively engaged in developing or testing cosmological models of the universe. Doroshkevich & Novikov (1964), two Russian cosmologists on the other hand were doing precisely this and they had published a paper suggesting that the CMB was detectable early in 1964. But they published in Russian and when the translation did occur some months later into English, it somehow escaped the attention of people working in the field. Soon after and independently, Dicke, Peebles, Wilkinson and Roll (hereafter Dicke et al) - astrophysicsts from Princeton only 60km away from Penzias and Wilson’s horn antenna had also figured out that relic radiation from the big bang should be detectable and they had even developed instrumentation to do so. Indeed they wrote (Dicke et al, 1965) “Two of us (P.G.R(oll) and D.T.W(ilkinson)) have constructed a radiometer and receiving horn capable of an absolute measure of thermal radiation at a wavelength of 3cm.” So to be clear: Cosmologists Doroshkevich and Novikov predicted the CMB but were theoreticians and did not have the technology to conduct the experiment. Their paper escaped the notice of those who might have had such equipment. Dicke et al likewise understood that the CMB should be detectable and had even begun an experiment to do so. Penzias and Wilson on the other hand were looking at radio signals from the Milky Way but found noise which was homogenous and isotropic to an extraordinary degree (Penzias & Wilson, 1965). One reason they were finally able to determine that it was not “white dielectric material” left by the pigeons in the antenna dish, was that a friend of Penzias -Bernard Burke - also a physicist - said he had read a preprint of a paper by Peebles and finally all the pieces fell into place. As Dicke, et al (1965) wrote, “While we have not yet obtained results with our instrument, we recently learned that Penzias and Wilson (1965) of the Bell Telephone Laboratories have observed background radiation at 7.3-cm wavelength.” Dicke is reported to have taken a phone call about the discovery and called across the room to other laboratory workers, “Well boys, we’ve been scooped”. (APSWeb) Clearly Penzias and Wilson did not know what they should have been looking for or what they were looking at. The other physicists who did, did not yet have the eyes to see (at least their ‘enhanced vision’ was not yet sufficiently fine tuned to have produced results). Penzias and Wilson stumbled across the CMB like a couple of petty thieves in the night tripping over a bag of cash left at the entrance to a house. They weren’t looking for it, didn’t mean to find it - but a windfall with no effort whatsoever was theirs. A serendipitous stroke of fortune. Or was it? This is our third example, after Tycho and the High-Z Supernova Search team to have apparently been given a helping hand by serendipity as though it were like the long lost tenth Greek Muse handing out favors to fortunate researchers. Throughout the history of astronomy there seems to have been degrees of serendipity involved in scientific discovery. Serendipity might range from playing a minor role (such as its contribution to the discovery of an accelerating universe. An argument could surely be made that the major part of this discovery was hard work by a huge team of specialists all toiling towards knowledge creation with highly sophisticated equipment) to being the most significant factor (the chance explosion of a star at just the right time and place such that thousands of years later on the other side of the galaxy it could be witnessed by an astronomer better positioned than any other at the time to record it and discuss its philosophical and cosmological significance). The ways in which serendipity enlightens scientific discovery in astronomy seems to vary from fully illuminating a solution to merely adding to the full spectrum of hard work, expertise and equipment that must all come together to shine a light on the problem. At least this is how the role of serendipity is presented classically. In no less a publication than Science and only recently in 2010 was serendipity credited with varying degrees of importance to the progress of astronomy. Lang (2010) writes of how the moons of Jupiter simply “appeared” to Galileo through his telescope, Hershel’s discovery of Uranus is described as “unanticipated” and Slipher “unexpectedly” discovered the red shift of galaxies which led to Hubble’s explanation that the universe was expanding - to confine myself to just the first 3 paragraphs of the article. Similar sentiments can be found in Cambridge Radio Astronomer A.C Fabian’s (2009) piece “Serendipity in Astronomy” where he writes of the discovery of extra-solar planets. “The problem lay in everyone’s expectations which supposed that other solar systems would resemble our own.” he writes, “In the first extrasolar system found, it was a Jupiter-mass planet in a 4-day orbit!” The exclamation mark of surprise A.C Fabian uses is not uncommon in articles and discussions ranging from Science through to Wikipedia which maintains an article that runs to 75 discrete examples of serendipity in science. Examples in astronomy make up 7 of those as of this writing May, 2011. Examples of the latter include all that I have already mentioned plus: Jansky’s single handed initiation of the entire field of radio astronomy, having discovered radio waves coming from the centre of the Milky Way while investigating static interference on short wave radio and the discovery of Pluto’s moon Charon by James Christy when he took a second look at a photographic plate he was about to throw out. There is one more - and that is the ‘serendipitous’ discovery of Gamma Ray Bursts by military satellites. It is to this discovery I will turn later in some detail. All of the preceding discussion aims to present the consensus that seems to exist among writers on the topic all the way from professional scientists working in the field itself, through to science journalists and all the way down to casual commentators: Serendipity is a real and functioning part of the scientific process and no where is this more evident than in astronomy. So we seem to have two possible tiers of contribution: Serendipity can be an essential component of discovery or it can be but a minor component and and there can be gradations anywhere along the spectrum between these two extremes. That serendipity has these varying roles to play seems obvious and the consensus about its role in science - whether large or small - seems universal. This does not, however prevent this common sense vision of the scientific process from being utterly false. Serendipity does not have a big role in science. Nor does it have a small role. It would even be generous to say that serendipity has no role for this still suggests that serendipity has some reality all it own when in truth serendipity is an illusion - it is a label that is applied to various parts of the scientific process - erroneously - with nothing to unify those parts besides a very subjective feeling. And that feeling is “surprise”. When a scientist, journalist, commentator - essay writer personally feels that a discovery is surprising then to the extent that the discovery feels surprising it is labelled serendipitous. To the degree that a discovery feels peculiarly unexpected it is labelled serendipitous. This feeling is an essential component of what it is for a thing to be labelled serendipitous for we are told by the very definition that serendipity evokes the emotion of happiness. My thesis here, therefore - is that although events are physically real, serendipity is not. But what is my criteria for real? Here I turn to David Deutsch’s definition (Deutsch, 2011), “a particular thing is real if and only if it figures in our best explanation of something.” Now events are real: they figure in our best explanation of space and time. Discoveries are likewise real. They are the events which figure at the beginning of the scientific process. Serendipity I will show is not real in this way as it does not figure in the scientific process. Before we come to my argument in full, consider by way of analogy the concept of a surprise party. Now the party has an objective reality. It is an event. But in what sense does the “surprise” exist except as a function of the psychological state of the “surprisee”? In other words, is a surprise party really surprising? I contend the question makes little sense. Most people in attendance at a so-called surprise party are, of course, unsurprised - they were invited. Indeed, sometimes no one at a so-called surprise party is surprised if the secret is not kept hidden. I hope to show in what follows that serendipity is just like this. Much seems to be made of its role at times in scientific discovery - especially in astronomy but this stems from a basic misunderstanding of the scientific process - even from those engaged in science. It will therefore be my contentious that although there exist discoveries in an objective sense (just as there exist parties), it is questionable as to whether there is, or can be, such a thing as a serendipitous discovery in astronomy unless one wants to label all discoveries in science serendipitous in which case we may as well do away with the word if it is to be tautologically appended to every instance of the word discovery signifying little more than that discoveries are unpredictable. Part 3: The Course of Scientific Discovery Much ink has been spilled and philosophical discourse occupied with the question as to how - and (sometimes with more than a dose of irrationality) if - science creates knowledge. While there has been academic disagreement, it seems that with some minor modifications, the idea that science makes progress should be uncontroversial for any engaged in research. Karl Popper is frequently credited with once and for all finding the demarcation line between science and non-science and so determining what it is about a scientific theory that makes it scientific. Its ability to be falsified. But this caricature of Popper sells his epistemology far short. Popper embraced an inclusive philosophy of science that sort to place on a rational foundation a description of how knowledge is created and grows. It is important for my subsequent argument about the role of serendipity in science that we take seriously the ideas of the most respected philosopher of science ever to have lived. Not because he is the most respected philosopher of science but rather because, with few adjustments, his ideas about the growth of knowledge accurately describe the scientific process. It is beyond the scope of the present essay to go into the many critics of Popper and the obsession that the humanities have with the work of Thomas Kuhn - I have argued about these elsewhere and provide a note for the reader to consult my work on this in a brief appendix at the conclusion of this piece. Hence, I am simply taking here seriously the notion Popper has correctly described how science proceeds and I will see where that takes me. It is also worthy of note that the question writer for our present project sees that there is indeed a distinction to be drawn between “the steady progression of ideas” and “serendipity”. I take “steady progression of ideas” to be what science is about. It is not so much the “steady” which is important but rather progression that sets science apart from any alternative way of trying to create knowledge. Science makes progress. There are objective differences between scientific theories and the direction of change is one towards general improvement (i.e: an increase in verisimilitude to use Popper’s vernacular). Some philosophers needlessly constrain the scope of science and I think it important to clarify what we mean by the word science. Science is just our best attempt to understand reality and so in this way there is no meaningful divide between science and rationality more generally. Science is a continuum of ideas about the nature of reality - but not just any ideas. In order to understand some aspect of reality one needs explanations and this is what scientific theories are. Scientific theories are our best explanations of reality. There are some features of scientific theories that make them superior to other explanations. One feature is testability (i.e: experiments can often be performed - at least in theory - to compare a theory with physical reality). Another is coherence (i.e: it is preferable that scientific theories generally mesh well rather than contradict one another). But how are scientific theories generated in the first place? I take the scheme - represented as a flow chart on the following page - from David Deutsch (1997) who is summarizing the view of scientific progress given to us by Karl Popper (1963). This sees the generation of scientific theories as a problem solving process. In this scheme our ‘discovery’ might be the Problem or it might be the Conjectured Solution. It might even be the Experimental Test. It is important for us to attempt find the location of scientific discovery in this scheme, because it is discoveries that are labelled as serendipitous. And so it is in this scheme that we seek to find serendipity playing a role. |
|
The significance of my diversion into the philosophy of science should be apparent. Where in this scheme do we locate serendipity? If serendipity has a place in astronomy and astronomy is a science and science is adequately described by the picture of Popperian Epistemology given to us above and we are to take all of our best theories seriously - then serendipity should appear here. If not explicitly then at least implicitly, somewhere in the course of scientific discovery. Let us revisit Brahe’s supernova and try to locate ‘serendipity’ in this scheme.
Clearly Brahe’s observation of the supernova is the stage that we consider serendipitous. It was not Tycho’s attempt to explain what he saw (his conjectured solution) rather it was the encounter with the Problem that was serendipitous. A Problem in the scheme given to us by Popper and Deutsch is the best synonym for discovery in this context. Let us take that seriously for a moment. Brahe’s encounter with a Problem if truly serendipitous means that there are two types of scientific problems: Serendipitous Problems and Non-Serendipitous ones. Then what is it that makes a Problem Serendipitous? Well what are the qualities of serendipity that would meaningfully distinguish one kind of problem from others that are not serendipitous? We need to return to our OED definition which was uncontroversial at the outset: “the occurrence and development of events by chance in a happy or beneficial way”. Could “Events by chance” be the feature we need? Well, what does “events by chance” really mean? A chance occurrence is either every occurrence in the literal sense (that is, according to our best scientific theory, reality is governed by the laws of quantum theory which do have a probabilistic or chance nature - but this would get us no where) or rather events which are unpredictable. But, and here is the key - all problems in science are by their very nature unexpected in this way. A problem in science cannot be expected. Because if it was expected (i.e: not due to chance) it would be predictable - and so not a problem.
Once again I turn to Deutsch (2011), “Before a discovery is made, no predictive process could reveal the content or the consequences of that discovery. For if it could, it would be that discovery. So scientific discovery is profoundly unpredictable, despite the fact it is determined by the laws of physics.” The importance of this for our present discussion is profound - there is nothing to distinguish one encounter with a problem from another in any substantive way. Consider again, Brahe’s observation of the supernova. If this encounter with a problem (according to modern philosophy of science) was serendipitous then let us compare it to another encounter with a problem that was not. Clearly we cannot compare it to any other encounter with a supernova because the problem was identified with Brahe. No subsequent observations of supernovae were actually Problems in the same way as Brahe’s. Subsequent observations were simply examples which fit the new theory (i.e: Conjectured Solution) that Brahe put forward. The prevailing cosmology of the time was that the heavens were immutable, perfect and unchangeable. The observation of the supernovae refuted (to use Popper’s terminology) this world view. That is to say the observation of the supernova actually constituted an experimental test of the Aristotelean world view. The new theory (the heavens can change) was thus not challenged by any further sightings of supernovae. However New Problems immediately arose such as “What is a supernova?” And “Why do they occur?” and countless others. Of course Brahe and his contemporaries were more concerned about the impermanency of the heavens - they were not speaking or writing about supernovae as they did not know what they were to pose the question in just this way. The point is, there is nothing we can really compare Brahe’s first observation of a supernova with to establish what role serendipity played. We cannot compare it to other, subsequent observations of supernovae - and we cannot compare it to other chance observations because all chance observations are just that: due to chance and so unexpected and equally deserving of the label serendipitous. It is only by a quirk of our human psychology and perhaps retrospective understanding that we choose to label some discoveries serendipitous. But it is we who do the labeling - we impose on the discovery the label - we do not find serendipity in the discovery like we find a star in the heavens. And (to bring this point into full view), because it is we who do the labeling there can be no role for serendipity in science because it has no reality of its own. Yet to read the writings of some others, serendipity seems to be almost a supernatural force, something akin to the Force from Star Wars, working in mysterious ways to guide the hand of scientific discovery to a greater or lesser extent. It is clearly not this. Or anything like this. It is an illusion. It is actually a by-product of the way we speak - it is, to use Ludwig Wittgenstein’s terminology - a “bewitchment of language” (Wittgenstein, 1953) that allows us to speak about such a thing as though it really existed.
To firm up this picture of science, let us revisit the observation of the CMB by Penzias and Wilson. To Penzias and Wilson, they had encountered a Problem. Namely: What was causing the background noise on their antenna? Dicke et. al (1965) conjectured a solution. They had even devised a way of criticizing their solution with an experimental test (they had built an antenna). But now, as we know, Penzias and Wilson’s Problem was actually Dicke et al’s experimental test. Again, where do we find the serendipity here? Was it in the “events by chance”? No, for as we have already seen - all problems are equally unexpected in the sense that they are unpredictable - and so ‘due to chance’ in that sense. Is it then in the ‘experimental test’? Again, no - the interference Penzias and Wilson detected did not become a test of Dicke et al’s theory until Dicke et al learned of it. And here is where the second part of the OED’s definition of serendipity begins to earn its keep - an event is serendipitous if the event occurs by chance “in a happy or beneficial way”. Happy or beneficial. Again recall the contribution of the word surprise to ‘surprise party’. It depends entirely on an individual perspective. The surprise applies only to the person who the party is being held for and even then only if all have kept the secret. In this sense there is no question that surprise could have an objective reality in the sense that the subjective feeling has an objective corollary in the brain of the person - neurotransmitter concentration increases, synaptic firings and so forth. But serendipity does not have even this much. Serendipity is not supposed to be a property of a person’s brain but rather a property of an event. Consider that Penzias and Wilson’s discovery of the CMB was not at all “happy or beneficial” for Dicke, Peebles et al. The Nobel Prize takes pride of place in the offices of the former, not the latter.
So serendipity is not a feature of a scientific discovery. Or of a scientific theory or its generation. Serendipity might best be described as a psychological attitude one can have towards certain discoveries or sequences of events - and even this is generous. It has nothing to do with the scientific process itself. Either serendipity is about chance (in which case all scientific discoveries are serendipitous in this sense) or it is about feeling happy about all such scientific discoveries (in which case this is a purely subjective feature of one’s individual psychological disposition in which case - again - it is not an objective feature of scientific discovery and is not a part of the process).
Serendipity attempts to qualify with a name the feeling of surprise we have about some discoveries over others. But as we have seen, all discoveries are unexpected - and to the same degree - in the sense that they are unpredictable - and to the same degree. Were they predictable then the prediction would be the discovery. Serendipity is a pretty word, but it is an illusion that should not be taken seriously as a component in the philosophy of science - by which I mean Popperian Epistemology - a characterization of which I have provided here. Serendipity in this way is an example of Wittgenstein’s bewitchment of language and we should exorcise ourselves of it. Scientists are engaged in a problem solving process and there is an infinite stream of them because each new problem solved creates yet more problems as we have seen (Tycho’s Supernovae leads to What is a Supernovae? The discovery of the CMB leads to why is it so homogenous?)
To stay with the bewitchment of language argument for just a moment more.
Consider the sentence: Tycho is happy.
The subject as well as the object have real – that is to say – spatially and temporally located corollaries. Although “Tycho” is trivial to locate, some investigation reveals that “happy” refers to something that really exists. Although we do not know the actual relationship between conscious states of the mind and physical reality we know enough to say it has something to do with a state of the brain and the concentration of neurotransmitters (in particular serotonin). We can even run the experiment: if you deliberately block the uptake of serotonin, a person feels unhappy. Although happiness is a subjective state of the mind, we can speak objectively about its existence. There are objective facts to be known about certain subjective states. This is a distinction that philosopher John Searle has drawn attention to (I take the following from Harris (2010)). There are two very different senses of the subjective/objective distinction. There is the epistemological (how we know) and the ontological (what we know). In this way we can know objectively about the subjective state we call happiness. We are not speaking epistemologically subjectively (that is to say with bias or by deceiving ourselves and so forth).
But consider:
Tycho is lucky.
It would appear that this sentence is very similar to the first. But this is misleading. Because we need only ask: What is the physical corollary of lucky? There isn’t one. Lucky is a purely subjective (that is to say – in this sense biased, the result of opinion, unreliable, etc) term. Does luck exist? No more than magic. There is nothing in physical reality corresponding to lucky. And, given that the project question all but identifies serendipity as simply another term for luck, we can likewise conclude that serendipity has no objective reality either. Rather, mere personal preference for the use of the term (i.e: subjective bias) leads to the use of the term when labeling what can objectively be called scientific discoveries.
Part 4: Evidence and Technology
Consider: Daily the evidence of the existence of supernovae falls all about us. The evidence that the heavens beyond the moon are in constant flux is basically ubiquitous. The only thing that prevented anyone prior to Tycho Brahe from discovering this was the right technology. Certainly “In the fields of observation, chance favors only the prepared mind” as Louis Pasteur quipped in a quote that has become synonymous with the claim that serendipity has a role in science - and in Brahe’s case he knew what measurements to make given this observation. But had Brahe been given a telescope it would not have required a supernovae blast to deliver him - and humanity - from the parochial view that it was only the Earth that underwent change.
Let us turn then to a final text-book case of apparent serendipity in astronomy: the discovery of gamma-ray bursts. Here we shall see that although the discovery seems serendipitous it is merely a direct consequence of the scientific process. A new problem encountered requiring a solution.
At the very height of the cold war, the United States and the USSR signed a test ban treaty. In order to monitor the communists’ compliance with the agreement, the USA launched a number of satellites into orbit which were designed to detect the gamma-ray signatures of nuclear blasts. These Vela spacecraft as they were named (Spanish for “watch”) never detected terrestrial radiation which could be described as unambiguously consistent with nuclear explosions. However on July 2, 1969 gamma radiation was detected but it was unlike anything expected from a nuclear weapon. Between that date and July 1972, sixteen bursts of gamma radiation were recorded by the Vela equipment (Klebesadel, et al. 1973). As this was a military operation, it took some time for the results to be considered either scientifically interesting or fit for publication. Klebesadel et al were the first to detect gamma ray bursts and even conjectured and attempted to test for the presence of supernovae where the bursts had occurred. Their paper spawned an entire sub-field of astrophysics which set out to solve the mysteries of the gamma ray bursts.
The popular view seems to be: because the satellites were not designed to find gamma ray bursts, but did - then this constitutes serendipity. That is, the discovery was an unexpected happy chance to remain within our OED definition. For a final time, let us look carefully at this exemplar of scientific discovery from the point of view both of the contribution of serendipity and in light of our taking seriously Popperian Epistemology as a full account of how knowledge is created. In terms first of serendipity: what role did chance have to play in this? Again, consider that the evidence of the existence of gamma ray bursts (whatever their cause) had been striking the upper atmosphere of the Earth for billions of years and will continue to do so for billions of years to come. So it is hardly serendipitous (i.e: lucky) to have encountered it. Rather all that was needed was the right technology. This is a thread that runs through so many so-called serendipitous discoveries. The augmentation of the human perceptive systems (i.e: equipment to detect radiation we cannot see with the naked eye) leads to unexpected discoveries. But what else could it do? Surely the surprise would be finding nothing after having invented and deployed some new technology specifically designed for the purpose of detecting radiation we have never tried to detect from space before. In one sense this makes a lie of Pasteur’s “chance favors the prepared mind”. The operators of the Vela satellites were entirely unprepared for what they observed. So much so that they discarded the initial measurements. Here it might be argued chance favored unprepared minds.
Now let us look at this same event through the eyes of someone who believes Popper adequately captured exactly how science generates knowledge. The launch of the Vela satellites should be expected to produce problems with our current theories. The reason for this is because such measurements of that radiation had not been made from that place ever before. When Vela detected gamma radiation, this constituted a Problem. The problem is one requiring an explanation. How are we to square these gamma ray bursts with what we already know about astrophysics? Can they be adequately captured by our theories of supernovae? If not, do we need to create new theories (conjecture solutions) about their origin? These new conjectures must mesh (that is to say be coherent with) our current astrophysical theories as far as that seems possible. Hence theories about all gamma ray bursts being spherically symmetric seem to be ruled out leading to solutions like beaming of the gamma rays. Such theories lead to their own predictions and so we get closer to the truth. To a Popperian, the Gamma Ray Bursts were always out there waiting to be discovered. All that was required was the right technology. That is not a serendipity - it is a consequence of scientific progress itself. As science makes progress, technology increases and we will encounter ever more problems that we cannot predict because of the very nature of scientific discovery.
Indeed so it is with Galileo’s discovery of the moons of Jupiter, so it is with Penzias and Wilson’s detection of the CMB and so it is with the discovery of Gamma Ray Bursts. Whoever first gets to use the technology for the first time, gets to see something before anyone else and in general they are not prepared for what they see for that would suggest they have an inkling - that is to say are able to predict what they will discover. And as has become a mantra now in my project: that would be a strict contradiction because no discovery is predictable. In the fields of observation, chance favors the unprepared mind particularly when that mind is connected to reality with technology none before have used.
When Klebesadel et al participated in the Vela Satellite program to detect high energy gamma radiation coming from the Earth, did they expect to find it coming from the heavens? It depends entirely upon what they knew about our best theories of astrophysics. The astrophysicists seemed to already know that supernovae would produce high energy gamma ray bursts (and indeed Klebesadel et al write in their paper about attempting to search for optical traces of high energy photons from supernovae after encouragement from theoreticians). Was this discovery serendipitous for them?
Part 5: Conclusions
The persistence of serendipity featuring in popular accounts of the process of scientific discovery can largely be traced to what might be called “folk-philosophy-of-science”. That is to say “common sense” ideas about how science proceeds include references to constructs such as serendipity. This might seem to be an overly academic ivory-tower position to take: that serendipity is merely a trick of our language and a term referring to an entity entirely devoid of any objective reality. However, although it may be fair to call what I am presenting here as a highly rarefied epistemological position about the abstract reality of a quantity relevant only to the academic study of the philosophy of science it is not merely that. Popular opinion and the way scientists and journalists write about scientific discovery affects research funding and so all parties involved in research should have access to the facts. As it turns out there are facts about epistemology and one such fact is that knowledge is created according to the scheme discovered and explained by Popper. And this scheme does not include serendipity as we have seen. Indeed serendipity does not exist in the way that most writers seem to think that it does. If we were to push the line that serendipity does have a role, then surely those holding the purse strings would rightfully wish to know where in the research grant proposal we can find allocation for serendipity. Or plans to ensure that serendipity is allowed to play a role.
A highly trained scientist with the best equipment does not make discoveries serendipitously. She makes discoveries because she is a highly trained scientist with the best equipment engaged in science. Much goes in to creating the opportunities for scientific discovery: education and hard work, a creative and inquisitive mind, painstaking analysis, access to new technology. At the conjunction of these things begins the creation of new knowledge.
References
American Physical Society Website http://www.aps.org/programs/outreach/history/historicsites/penziaswilson.cfm (accessed 25 April, 2011)
Deutsch, D, 1997, The Fabric of Reality
- 2011, The Beginning of Infinity
Dicke, R., et al. 1965 ApJ, 142: 414
Doroshkevich, A. G.& Novikov, I.D.,1964 Soviet Physics Doklady 9: 4292.
Harris, S. 2010, The Moral Landscape
Fabian, A. 2009, preprint (arXiv:0908.2784v1)
Klebesadel, R., et al 1973 ApJ: 182: L85
Krause, et al. 2008 Nature 456: 617
Lang, K. 2010, Science, 327, 39
Penzias, A. & Wilson, R. 1965 ApJ 142: 419
PlaitWeb: Phil Plait’s Bad Astronomy Website http://www.badastronomy.com/mad/1997/howmanysn.html accessed May 15, 2011.
Popper, K, 1963, Conjectures and Refutations: The Growth of Scientific Knowledge
Riess, A. et al 1998 AJ, 116, 1009
Wittgenstein, L., 1953, Philosophical Investigations.
Appendix
A reader asked me to consider writing my own philosophy of science, having criticized every version of the philosophy of science presented by them as being inadequate. I provide a link here to a blog site where I have posted a more full account of the position I take with respect to how science proceeds. That blog is likewise all my own work. It may be useful to the reader to consult this if they think I have been unfair in my dismissal of figures such as Thomas Kuhn who had much to say on the way in which science proceeds through revolutions and paradigm shifts. Many parallels can be drawn between the errors of the Kuhnian view of science and the misconception I have attempted to identify in this project about the role of serendipity in science.
http://www.orble.com/how-is-knowledge-generated/
Clearly Brahe’s observation of the supernova is the stage that we consider serendipitous. It was not Tycho’s attempt to explain what he saw (his conjectured solution) rather it was the encounter with the Problem that was serendipitous. A Problem in the scheme given to us by Popper and Deutsch is the best synonym for discovery in this context. Let us take that seriously for a moment. Brahe’s encounter with a Problem if truly serendipitous means that there are two types of scientific problems: Serendipitous Problems and Non-Serendipitous ones. Then what is it that makes a Problem Serendipitous? Well what are the qualities of serendipity that would meaningfully distinguish one kind of problem from others that are not serendipitous? We need to return to our OED definition which was uncontroversial at the outset: “the occurrence and development of events by chance in a happy or beneficial way”. Could “Events by chance” be the feature we need? Well, what does “events by chance” really mean? A chance occurrence is either every occurrence in the literal sense (that is, according to our best scientific theory, reality is governed by the laws of quantum theory which do have a probabilistic or chance nature - but this would get us no where) or rather events which are unpredictable. But, and here is the key - all problems in science are by their very nature unexpected in this way. A problem in science cannot be expected. Because if it was expected (i.e: not due to chance) it would be predictable - and so not a problem.
Once again I turn to Deutsch (2011), “Before a discovery is made, no predictive process could reveal the content or the consequences of that discovery. For if it could, it would be that discovery. So scientific discovery is profoundly unpredictable, despite the fact it is determined by the laws of physics.” The importance of this for our present discussion is profound - there is nothing to distinguish one encounter with a problem from another in any substantive way. Consider again, Brahe’s observation of the supernova. If this encounter with a problem (according to modern philosophy of science) was serendipitous then let us compare it to another encounter with a problem that was not. Clearly we cannot compare it to any other encounter with a supernova because the problem was identified with Brahe. No subsequent observations of supernovae were actually Problems in the same way as Brahe’s. Subsequent observations were simply examples which fit the new theory (i.e: Conjectured Solution) that Brahe put forward. The prevailing cosmology of the time was that the heavens were immutable, perfect and unchangeable. The observation of the supernovae refuted (to use Popper’s terminology) this world view. That is to say the observation of the supernova actually constituted an experimental test of the Aristotelean world view. The new theory (the heavens can change) was thus not challenged by any further sightings of supernovae. However New Problems immediately arose such as “What is a supernova?” And “Why do they occur?” and countless others. Of course Brahe and his contemporaries were more concerned about the impermanency of the heavens - they were not speaking or writing about supernovae as they did not know what they were to pose the question in just this way. The point is, there is nothing we can really compare Brahe’s first observation of a supernova with to establish what role serendipity played. We cannot compare it to other, subsequent observations of supernovae - and we cannot compare it to other chance observations because all chance observations are just that: due to chance and so unexpected and equally deserving of the label serendipitous. It is only by a quirk of our human psychology and perhaps retrospective understanding that we choose to label some discoveries serendipitous. But it is we who do the labeling - we impose on the discovery the label - we do not find serendipity in the discovery like we find a star in the heavens. And (to bring this point into full view), because it is we who do the labeling there can be no role for serendipity in science because it has no reality of its own. Yet to read the writings of some others, serendipity seems to be almost a supernatural force, something akin to the Force from Star Wars, working in mysterious ways to guide the hand of scientific discovery to a greater or lesser extent. It is clearly not this. Or anything like this. It is an illusion. It is actually a by-product of the way we speak - it is, to use Ludwig Wittgenstein’s terminology - a “bewitchment of language” (Wittgenstein, 1953) that allows us to speak about such a thing as though it really existed.
To firm up this picture of science, let us revisit the observation of the CMB by Penzias and Wilson. To Penzias and Wilson, they had encountered a Problem. Namely: What was causing the background noise on their antenna? Dicke et. al (1965) conjectured a solution. They had even devised a way of criticizing their solution with an experimental test (they had built an antenna). But now, as we know, Penzias and Wilson’s Problem was actually Dicke et al’s experimental test. Again, where do we find the serendipity here? Was it in the “events by chance”? No, for as we have already seen - all problems are equally unexpected in the sense that they are unpredictable - and so ‘due to chance’ in that sense. Is it then in the ‘experimental test’? Again, no - the interference Penzias and Wilson detected did not become a test of Dicke et al’s theory until Dicke et al learned of it. And here is where the second part of the OED’s definition of serendipity begins to earn its keep - an event is serendipitous if the event occurs by chance “in a happy or beneficial way”. Happy or beneficial. Again recall the contribution of the word surprise to ‘surprise party’. It depends entirely on an individual perspective. The surprise applies only to the person who the party is being held for and even then only if all have kept the secret. In this sense there is no question that surprise could have an objective reality in the sense that the subjective feeling has an objective corollary in the brain of the person - neurotransmitter concentration increases, synaptic firings and so forth. But serendipity does not have even this much. Serendipity is not supposed to be a property of a person’s brain but rather a property of an event. Consider that Penzias and Wilson’s discovery of the CMB was not at all “happy or beneficial” for Dicke, Peebles et al. The Nobel Prize takes pride of place in the offices of the former, not the latter.
So serendipity is not a feature of a scientific discovery. Or of a scientific theory or its generation. Serendipity might best be described as a psychological attitude one can have towards certain discoveries or sequences of events - and even this is generous. It has nothing to do with the scientific process itself. Either serendipity is about chance (in which case all scientific discoveries are serendipitous in this sense) or it is about feeling happy about all such scientific discoveries (in which case this is a purely subjective feature of one’s individual psychological disposition in which case - again - it is not an objective feature of scientific discovery and is not a part of the process).
Serendipity attempts to qualify with a name the feeling of surprise we have about some discoveries over others. But as we have seen, all discoveries are unexpected - and to the same degree - in the sense that they are unpredictable - and to the same degree. Were they predictable then the prediction would be the discovery. Serendipity is a pretty word, but it is an illusion that should not be taken seriously as a component in the philosophy of science - by which I mean Popperian Epistemology - a characterization of which I have provided here. Serendipity in this way is an example of Wittgenstein’s bewitchment of language and we should exorcise ourselves of it. Scientists are engaged in a problem solving process and there is an infinite stream of them because each new problem solved creates yet more problems as we have seen (Tycho’s Supernovae leads to What is a Supernovae? The discovery of the CMB leads to why is it so homogenous?)
To stay with the bewitchment of language argument for just a moment more.
Consider the sentence: Tycho is happy.
The subject as well as the object have real – that is to say – spatially and temporally located corollaries. Although “Tycho” is trivial to locate, some investigation reveals that “happy” refers to something that really exists. Although we do not know the actual relationship between conscious states of the mind and physical reality we know enough to say it has something to do with a state of the brain and the concentration of neurotransmitters (in particular serotonin). We can even run the experiment: if you deliberately block the uptake of serotonin, a person feels unhappy. Although happiness is a subjective state of the mind, we can speak objectively about its existence. There are objective facts to be known about certain subjective states. This is a distinction that philosopher John Searle has drawn attention to (I take the following from Harris (2010)). There are two very different senses of the subjective/objective distinction. There is the epistemological (how we know) and the ontological (what we know). In this way we can know objectively about the subjective state we call happiness. We are not speaking epistemologically subjectively (that is to say with bias or by deceiving ourselves and so forth).
But consider:
Tycho is lucky.
It would appear that this sentence is very similar to the first. But this is misleading. Because we need only ask: What is the physical corollary of lucky? There isn’t one. Lucky is a purely subjective (that is to say – in this sense biased, the result of opinion, unreliable, etc) term. Does luck exist? No more than magic. There is nothing in physical reality corresponding to lucky. And, given that the project question all but identifies serendipity as simply another term for luck, we can likewise conclude that serendipity has no objective reality either. Rather, mere personal preference for the use of the term (i.e: subjective bias) leads to the use of the term when labeling what can objectively be called scientific discoveries.
Part 4: Evidence and Technology
Consider: Daily the evidence of the existence of supernovae falls all about us. The evidence that the heavens beyond the moon are in constant flux is basically ubiquitous. The only thing that prevented anyone prior to Tycho Brahe from discovering this was the right technology. Certainly “In the fields of observation, chance favors only the prepared mind” as Louis Pasteur quipped in a quote that has become synonymous with the claim that serendipity has a role in science - and in Brahe’s case he knew what measurements to make given this observation. But had Brahe been given a telescope it would not have required a supernovae blast to deliver him - and humanity - from the parochial view that it was only the Earth that underwent change.
Let us turn then to a final text-book case of apparent serendipity in astronomy: the discovery of gamma-ray bursts. Here we shall see that although the discovery seems serendipitous it is merely a direct consequence of the scientific process. A new problem encountered requiring a solution.
At the very height of the cold war, the United States and the USSR signed a test ban treaty. In order to monitor the communists’ compliance with the agreement, the USA launched a number of satellites into orbit which were designed to detect the gamma-ray signatures of nuclear blasts. These Vela spacecraft as they were named (Spanish for “watch”) never detected terrestrial radiation which could be described as unambiguously consistent with nuclear explosions. However on July 2, 1969 gamma radiation was detected but it was unlike anything expected from a nuclear weapon. Between that date and July 1972, sixteen bursts of gamma radiation were recorded by the Vela equipment (Klebesadel, et al. 1973). As this was a military operation, it took some time for the results to be considered either scientifically interesting or fit for publication. Klebesadel et al were the first to detect gamma ray bursts and even conjectured and attempted to test for the presence of supernovae where the bursts had occurred. Their paper spawned an entire sub-field of astrophysics which set out to solve the mysteries of the gamma ray bursts.
The popular view seems to be: because the satellites were not designed to find gamma ray bursts, but did - then this constitutes serendipity. That is, the discovery was an unexpected happy chance to remain within our OED definition. For a final time, let us look carefully at this exemplar of scientific discovery from the point of view both of the contribution of serendipity and in light of our taking seriously Popperian Epistemology as a full account of how knowledge is created. In terms first of serendipity: what role did chance have to play in this? Again, consider that the evidence of the existence of gamma ray bursts (whatever their cause) had been striking the upper atmosphere of the Earth for billions of years and will continue to do so for billions of years to come. So it is hardly serendipitous (i.e: lucky) to have encountered it. Rather all that was needed was the right technology. This is a thread that runs through so many so-called serendipitous discoveries. The augmentation of the human perceptive systems (i.e: equipment to detect radiation we cannot see with the naked eye) leads to unexpected discoveries. But what else could it do? Surely the surprise would be finding nothing after having invented and deployed some new technology specifically designed for the purpose of detecting radiation we have never tried to detect from space before. In one sense this makes a lie of Pasteur’s “chance favors the prepared mind”. The operators of the Vela satellites were entirely unprepared for what they observed. So much so that they discarded the initial measurements. Here it might be argued chance favored unprepared minds.
Now let us look at this same event through the eyes of someone who believes Popper adequately captured exactly how science generates knowledge. The launch of the Vela satellites should be expected to produce problems with our current theories. The reason for this is because such measurements of that radiation had not been made from that place ever before. When Vela detected gamma radiation, this constituted a Problem. The problem is one requiring an explanation. How are we to square these gamma ray bursts with what we already know about astrophysics? Can they be adequately captured by our theories of supernovae? If not, do we need to create new theories (conjecture solutions) about their origin? These new conjectures must mesh (that is to say be coherent with) our current astrophysical theories as far as that seems possible. Hence theories about all gamma ray bursts being spherically symmetric seem to be ruled out leading to solutions like beaming of the gamma rays. Such theories lead to their own predictions and so we get closer to the truth. To a Popperian, the Gamma Ray Bursts were always out there waiting to be discovered. All that was required was the right technology. That is not a serendipity - it is a consequence of scientific progress itself. As science makes progress, technology increases and we will encounter ever more problems that we cannot predict because of the very nature of scientific discovery.
Indeed so it is with Galileo’s discovery of the moons of Jupiter, so it is with Penzias and Wilson’s detection of the CMB and so it is with the discovery of Gamma Ray Bursts. Whoever first gets to use the technology for the first time, gets to see something before anyone else and in general they are not prepared for what they see for that would suggest they have an inkling - that is to say are able to predict what they will discover. And as has become a mantra now in my project: that would be a strict contradiction because no discovery is predictable. In the fields of observation, chance favors the unprepared mind particularly when that mind is connected to reality with technology none before have used.
When Klebesadel et al participated in the Vela Satellite program to detect high energy gamma radiation coming from the Earth, did they expect to find it coming from the heavens? It depends entirely upon what they knew about our best theories of astrophysics. The astrophysicists seemed to already know that supernovae would produce high energy gamma ray bursts (and indeed Klebesadel et al write in their paper about attempting to search for optical traces of high energy photons from supernovae after encouragement from theoreticians). Was this discovery serendipitous for them?
Part 5: Conclusions
The persistence of serendipity featuring in popular accounts of the process of scientific discovery can largely be traced to what might be called “folk-philosophy-of-science”. That is to say “common sense” ideas about how science proceeds include references to constructs such as serendipity. This might seem to be an overly academic ivory-tower position to take: that serendipity is merely a trick of our language and a term referring to an entity entirely devoid of any objective reality. However, although it may be fair to call what I am presenting here as a highly rarefied epistemological position about the abstract reality of a quantity relevant only to the academic study of the philosophy of science it is not merely that. Popular opinion and the way scientists and journalists write about scientific discovery affects research funding and so all parties involved in research should have access to the facts. As it turns out there are facts about epistemology and one such fact is that knowledge is created according to the scheme discovered and explained by Popper. And this scheme does not include serendipity as we have seen. Indeed serendipity does not exist in the way that most writers seem to think that it does. If we were to push the line that serendipity does have a role, then surely those holding the purse strings would rightfully wish to know where in the research grant proposal we can find allocation for serendipity. Or plans to ensure that serendipity is allowed to play a role.
A highly trained scientist with the best equipment does not make discoveries serendipitously. She makes discoveries because she is a highly trained scientist with the best equipment engaged in science. Much goes in to creating the opportunities for scientific discovery: education and hard work, a creative and inquisitive mind, painstaking analysis, access to new technology. At the conjunction of these things begins the creation of new knowledge.
References
American Physical Society Website http://www.aps.org/programs/outreach/history/historicsites/penziaswilson.cfm (accessed 25 April, 2011)
Deutsch, D, 1997, The Fabric of Reality
- 2011, The Beginning of Infinity
Dicke, R., et al. 1965 ApJ, 142: 414
Doroshkevich, A. G.& Novikov, I.D.,1964 Soviet Physics Doklady 9: 4292.
Harris, S. 2010, The Moral Landscape
Fabian, A. 2009, preprint (arXiv:0908.2784v1)
Klebesadel, R., et al 1973 ApJ: 182: L85
Krause, et al. 2008 Nature 456: 617
Lang, K. 2010, Science, 327, 39
Penzias, A. & Wilson, R. 1965 ApJ 142: 419
PlaitWeb: Phil Plait’s Bad Astronomy Website http://www.badastronomy.com/mad/1997/howmanysn.html accessed May 15, 2011.
Popper, K, 1963, Conjectures and Refutations: The Growth of Scientific Knowledge
Riess, A. et al 1998 AJ, 116, 1009
Wittgenstein, L., 1953, Philosophical Investigations.
Appendix
A reader asked me to consider writing my own philosophy of science, having criticized every version of the philosophy of science presented by them as being inadequate. I provide a link here to a blog site where I have posted a more full account of the position I take with respect to how science proceeds. That blog is likewise all my own work. It may be useful to the reader to consult this if they think I have been unfair in my dismissal of figures such as Thomas Kuhn who had much to say on the way in which science proceeds through revolutions and paradigm shifts. Many parallels can be drawn between the errors of the Kuhnian view of science and the misconception I have attempted to identify in this project about the role of serendipity in science.
http://www.orble.com/how-is-knowledge-generated/