The Mythology of Science
August 29, 2007
What is the nature of change in scientific theories? Is it reasonable to consider science as a journey from ignorance to truth? Or is there a mythology of science that distorts the nature of the scientific endeavour?
When I first went to university, it was to study astrophysics. I was enthusiastic about the subject, and eager to learn. My expectation to some degree was that the workings of the universe were going to be explicated to me, and that I was going to be shown experiments which underpinned the theories of nature that had been developed, and from which those theories could be derived. But in fact, the theories were taught as strictly factual, and only then were the experiments enacted. Students, having already learned the prevailing models of reality, approached each experiment with the certain knowledge of the expected results.
When we were asked, for instance, to take measurements from which to calculate the gravitational constant of the universe, there was no doubt as to what the correct answer was expected to be. In practice, few if any students produced results that would yield an answer sufficiently close to that dictated by prior theory. In the face of conflicting experimental evidence, most students would attempt the experiment again, usually once again yielding results which were not of the kind expected. After a few failed attempts, many students then adjusted their data in order to more closely resemble the expected results of the experiment.
This experience greatly challenged my expectations of science. It was not that the theory in question was fundamentally in error – if we had, for instance, averaged all the data gathered in the lab by all the students, the mean result would almost certainly have resembled the expected results. But what I observed in the physics laboratory was students of science faced with unexpected data and then, instead of reporting this honestly as my preconceptions of the scientific method demanded, changing their measurements to conform with the expected results. That this was the way to get the highest marks from the laboratory experiments is not in doubt, but if the central value of science is truth, the students were not learning this value – they were learning how to toe the line with existing theory.
Scientists, in general, are not taught philosophy of science, and as a result rarely question their abstract beliefs about the nature of the scientific process. As a result, science has developed a persistent mythology, central to which is the idea that science uncovers the truth, and as the theories and techniques of science develop, science moves closer and closer to absolute truth.
The celebrated historian and philosopher of science Thomas Kuhn, in his seminal work The Structure of Scientific Revolutions (first published in 1962) put forward one of the decisive criticisms of this view. He observed, from study of the development of European science, that scientists’ commitment to particular theoretical frameworks acted as a barrier to seeing data in a different light – even when that data was entirely contradictory to the theory. Like the students in the physics lab I encountered, theory was considered to trump observation to some degree.
Kuhn’s historical analysis resulted in a model of science that denied the conventional belief in science as a process of accretion. The idea that science gradually assimilated building blocks and advanced in steady and discrete steps did not match up to the historical record. Instead, Kuhn saw periods of what he termed normal science, during which scientists worked to adapt their current theories to a range of experimental observations, and periods of scientific revolution – when the old theories came into crisis, and were eventually supplanted by new theories.
The view of scientific fields proposed by Kuhn was that a particular field does not emerge within science until scientists working in this area begin to develop a common framework. Kuhn terms this framework a paradigm, a term he uses ambiguously in a number of subtly different contexts. On the one hand, a paradigm describes the collection of symbolic generalisations, experimental methods and common assumptions shared by practitioners of a given scientific field. On the other, paradigms represent specific exemplars of scientific puzzle solving, including both experimental and theoretical results to problems. We can see a paradigm as a set of common beliefs that a group of scientists share. Despite the intense faith that is placed in science, these beliefs do not necessarily correspond to reality – rather, they provide a framework that enables the scientists to investigate reality in a particular way.
Kuhn suggests that during periods of normal science, scientists are mostly attempting to “force nature into the preformed and relatively inflexible box that the paradigm supplies.” No paradigm explains all the facts with which it can be confronted, and the general process of science therefore works on the solutions to puzzles – problems in adapting the theoretical models of that paradigm to the description of reality. This commitment to a particular paradigm is essential to the scientific process in Kuhn’s opinion. Normal science is only able to proceed on the predicate that scientists know “what the world is like.”
Anomalous data – that which cannot be made to fit into the box provided by the paradigm – is usually ignored, or interpreted in a manner consistent with the paradigm’s assumptions, or else the theoretical framework of the paradigm is adjusted to compensate for the discrepancy. Experienced scientists, having had great success with their paradigm, are unwilling to abandon it in the light of contradictory evidence. This can be seen as a necessary commitment, because to give up a particular paradigm without adopting a new one would be to abandon science entirely: there can be no science without a particular model by which the investigatory process can proceed.
Kuhn observes that new paradigms gradually replace old ones as a result of a growing view (usually among newcomers to the field, since these are least committed to the old beliefs) that the old methods are no longer able to guide the exploration of a particular area in which the old paradigm used to lead the way. But the transition between paradigms does not often proceed without difficulties – the commitment to the old paradigm remains strong, and it is hard for rational discussion to take place between two individuals who are using different paradigms.
Often, the elder scientists are unwilling or unable to abandon the paradigm which has lead to such success and progress during their time, and this is only natural. As Max Planck observed: “a new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.”
There is no objective means of resolving these disputes between competing explanations. Scientists must necessarily premise their own paradigm in order to argue in its favour, a circular process that can nonetheless be quite persuasive. Although arguments are usually couched in terms of the capacity of the competing theories to explain experimental facts, it is not generally possible to determine which is the better theory by comparison with facts. All significant scientific theories fit the facts to some extent. But when two theories are competing with each other, the issue is not how much each theory fits the facts, per se, but rather whether or not one theory fits the facts better than another.
Even this may not be the means in which the commitment to a paradigm is changed. What is being asked for is a fundamental alteration to the way a particular scientific field is practiced, and for this argument to be persuasive what must be offered is something more than correspondence to facts – as already mentioned, most scientific theories can be made to adequately explain the facts they are faced with. For example, chemists abandoned phlogiston theory, but at the time that the theory of combustion in the presence of oxygen was forming, phlogiston adequately explained the majority of the observed facts. Many of these facts presumed phlogiston in their formulation. But the new idea carried with it the promise of future progress, and it was faith in this possibility that helped drive this paradigm change forward.
The model of reality suggested by one paradigm generally
involves quite different entities to its rival. Consider the view of the world
Part of the mythology of science is to see changes like these (from an old successful theory to a new one) both as progressing towards truth, and also to view the old theory not as having been disproved but as a special case of the new theory. This tendency is particularly prevalent in relation to the paradigm shift from Newtonian to Einsteinian physics, and scientists can indeed demonstrate how the Newtonian equations can be derived from the later theories. But this process is misleading. An abandoned theory can always be viewed as a special case of its successor, but to do so one must transform the original theory in the terms of the new one – and this is something that can only be done in hindsight.
This rewriting of history is endemic to science. The scientists of earlier ages are represented in textbooks and the like as having worked on the same set of problems, and in accordance with the same guidelines, as modern science. This makes science appear to be cumulative, but it is a fiction created by ignoring the crises that accompany scientific revolution. Earlier scientists worked on very different problems, because their models of the world lead them to different puzzles to solve. Because the results of scientific research do not seem to depend upon the history of their development (that is, scientific theories do not depend upon their historical origins for their veracity) it seems acceptable to abstract over the details of their development. In doing so, the nature of the scientific endeavour becomes distorted to appear both linear and cumulative.
Furthermore, this view distorts the nature of the abandoned paradigms. There is a tendency to look at certain discarded theories as myths that have been disproved – but this perspective will not stand up to scrutiny. The old paradigms were derived through the same essential scientific process as the new paradigms, and from a historical perspective both the old and the new must be seen as scientific – else all scientific knowledge can be accused of being a myth. It is not that the old beliefs were false and the new ones true – for a future paradigm shift may well render the current ideas false by this reasoning. Rather, the science of the past contains views and beliefs wholly incompatible with the models of present scientists, yet all such paradigms are still by their very nature scientific. They proceeded from fact and measurement, with the goal of providing explanations of these facts.
Kuhn challenges our preconceptions of science as evolving towards objective truth, and instead suggests we should understand scientific progress as evolution from the community’s prior state of knowledge. There is no perfect conception of reality that science is travelling towards; there is no ideal goal state that science will ultimately evolve into. Such ideas are anachronistic fallacies that do not match up with the history of science.
The idea of goal-directed progress was abandoned in the
Ultimately, scientific knowledge is, like language, an intrinsic property of a group of people. To understand that knowledge, it is necessary to understand the nature and characteristics of the groups that create and use this knowledge. Science is the name we give to the practices of scientists, who by dedication to an empirical view of the world gradually refine their ideas, and produce exceptional instruments which in turn allow for the creation of new technologies. Later scientific theories show progress – they are better at solving puzzles in often very different environments to those of their predecessors – but it is a mythological view of science that sees science as truth, or evolving towards truth. Science evolves as scientists refine their perspectives, but this refinement is an adaptation to new conditions of knowledge, and not an inevitable march towards perfection.
The opening image is Dancing Light, a refraction caustic by Alan Jaras which I found here. As ever, no copyright infringement is intended and I will take the image down if asked.
Whats funny is when Scientists see something that contradicts their preffered models, they often disparrage the champions of new ideas as quacks, even when their process is highly scientific. Richard Dawkins dances on the verge of being such a prig, but look at Aubrey de Grey and how he's been recieved by the scientific community, and you see this in action.
I think this probably applies to the game design community as well.
Posted by: Patrick | August 30, 2007 at 04:44 PM
It's a very human thing, premature certainty. Kuhn argues that the scientific community could not operate without this commitment to "knowing how things are", and he may have a point.
As for the games industry, yes, the parallel is rather too apposite. :)
Posted by: Chris | August 31, 2007 at 03:49 PM
Science evolves as scientists refine their perspectives, but this refinement is an adaptation to new conditions of knowledge, and not an inevitable march towards perfection.
Hear hear. And less fit theories - eventually - become extinct. But that can take thousands of years, especially as the fitness function includes some aspect of "appealing to humans" in it. Belief in a flat earth is not extinct because the idea appeals to some.
Posted by: Peter Crowther | September 03, 2007 at 09:17 PM
There's an element of circularity in this statement of position, since by implication you are saying that there is no objective truth (and I know this to be your opinion), and so how can you dismiss anything as not being true if you have no grounds for comparative judgement?
Posted by: zenBen | September 04, 2007 at 02:11 PM
zenBen: I have never said there is no objective truth - only that we are incapable of establishing it. We can dismiss things as being 'not true' relative to a particular set of propositions, but only in such a relative state of affairs. This, however, is more than sufficient for science to function.
Posted by: Chris | September 05, 2007 at 02:26 PM
Chris, in reference to our earlier exchange I will just add that it is often also the practicing "technologist" who starts to find "gaps and discrepancies" in the predictions provided by theoretical scientist - the practioners may be a lot less enthusiastic about the "state of truth" than the "theorists" (a.k.a "professors").
with regard to "textbooks" and "scientific education" I couldn't agree more with your view - I find "textbook teachings" (or is "preachings" more adequate here ;) very often failing to stress that the experiment (even if it were only a thought experiment) must have priority over the stating or derivation of "laws".
on the scientific community's belief in the "progress towards truth" I'd add another important point you seem to be neglecting (what does Kuhn ahve to say about it?):
ever since at least the invention of calculus (if not since the time of Euclid) the "progress" in development, refinement and extension of mathematics has played a crucial role in the development or creation of "world-views" that seem acceptable to the incubent gate keepers of scientifc process and interpretation (the "authorities" of the field) - and this seems to have become even more important if you consider the leading position of "mathematical physicist" for the last 150 years starting with Maxwell, Gibbs, Boltzmann and Poincare, the obvious "Einstein-Schrödinger-Dirac" group over to von Neumann, Feynmann and so forth - it seems to me that it is here in the supposedly "universal language" of mathematics that the "post-modern" scientist seeks his last ressort when talking of "objective truth"
Posted by: translucy | September 07, 2007 at 12:13 PM
It seems that by placing "inverted commas" around key words, we can express our "disagreement" with some position, without actually elaborating any reason why this position may be "wrong".
Posted by: zenBen | September 07, 2007 at 01:56 PM
zenBen: snarky, but not necessarily helpful. ;) translucy seems to have picked out more terms than perhaps was strictly necessary, but I didn't find the meaning obfuscated, personally. I don't think the intent was disagree without stating objections: in fact, if you read translucy's comment and ignore the quote marks, it seems perfectly complete.
translucy: I would love to see a collection of interviews with technologists, and their attitude to scientific theory - I feel that could be quite illuminating! Increasingly, I feel that while the experiment may be the exemplar of the scientific tradition, it is technology which is the final validation of any theoretical framework. In fact, those scientific fields (evolutionary biology in particular) farthest from producing technology seem to be the areas most prone to devolve into metaphysics.
As for the universal language of mathematics as a means to objective truth - there is some validity to this claim, in so much as the formalism of mathematics and logic is in fact the only sure way of expressing true and false. Of course, being able to express such matters in mathematics doesn't give us purchase on application to the world around us. In fact, it is effectively a kind of retreat from that world.
As for the validity of appointing mathematics as "a universal language" - there is at least one sense in which this claim is valid, namely that mathematics - being a formal system - has the potential to be understood by any species. If we were to meet an alien race, it would presumably be easier to communicate mathematically than linguistically, at least at first.
Coupled with this, however, is the acknowledgement that mathematics is not a universal language in the sense of being able to use it to communicate with anyone - since only the minority are fluent in its formalisms. Similarly, as a language mathematics is severely limited: it is, strictly speaking, meaningless - without meaning - since meaning as a conscious phenomena is entirely inexpressible in mathematical terms, something no conventional language might claim or want to claim for itself.
Posted by: Chris | September 11, 2007 at 12:10 PM
Well I just thought translucy's choices of which words to "inverted comma", and which not, were quite arbitrary. E.g. - (the "authorities" of the field) - why not "field"? Are these "authorities" not really authorities, but they do work in a real field? Why? If the field exists, in a very real sense it is only because there are authorities in it. It's a confusing comment, and seems to be going further than just saying some of the terms have ambiguous interpretations, but without real definition as to what it is saying.
As to language...our personal experience of the world, I see as a quite fuzzy but mutually clear and comprehensible model of the world. Scientific experiment, is a precise snapshot of the world, a facsimile model but only for the case of the measurement. Scientific theory, (ideally) a generalisation of (several of) these experimental measurements, is a close to precise model. And mathematics is an overly precise model, since the world itself seems to become a little fuzzy when you get down far enough.
So we're always dealing in error margins. There is no objective truth...well, I would say there isn't even any 'objective', so why not assume that the model that allows the clearest communication is the truth?
Posted by: zenBen | September 11, 2007 at 03:32 PM
I don't know whether the fact that translucy's first language is German figured into this at all. We'll have to wait for comment from the source. :)
"I would say there isn't even any 'objective', so why not assume that the model that allows the clearest communication is the truth?"
The clearest communication between who and whom, exactly? And why should communication and truth be comparable categories?
Posted by: Chris | September 11, 2007 at 04:43 PM
Chris: points well taken, thank you for adding your perspective on mathematics as language to the above post. I do find it quite important though for understanding the original post. (And I do happen to agree with what you say, for all it is worth :)
zenBen: in my view your emphasis on modeling the inference process itself based on multiple and multi-contextual measurements goes to the heart of the matter - you do seem to offer a view that I find to be common among the scientists and technologists I work with. On reading my above comment a second time I can see why the amount of "inverted commas" must seem excessive to you - let's see if I can do "better" next time ;) For this time simply delete all >"< while reading ;)
Posted by: translucy | September 11, 2007 at 06:30 PM
"The clearest communication between who and whom, exactly? And why should communication and truth be comparable categories?"
Whoever and whomever - two individuals, groups, what you will. The comparability comes if you strictly disallow assumption of shared frames of reference between sentiences. You and I both live within our skulls, looking out on the world, and believe that we see (almost!) the same world because our models have some level of precision and have been communicated between us (mostly by proxy through similar basic education). So here, as I use it, truth relates to communication in an inverse manner to noise. I'm composing these definitions on the fly, so that ironically the 'truth' of what I'm saying (my point of view) may be obfuscated by careless language.
Posted by: zenBen | September 12, 2007 at 12:09 PM
It's funny... now I understand what you're saying, but have lost how it connects to the wider conversation. :) I have a feeling we finished our discussion a while back, and now we are just mutually scribbling in the margins. :)
Posted by: Chris | September 12, 2007 at 08:25 PM
Thanks for the great article, what an interesting read!
Posted by: Law of Attraction | May 08, 2008 at 04:02 PM
Just discovered your blog today and I find it to be quite good. I don't know Kuhn, but I've spent a fair bit of time with Cassirer. Have you read him? His Philosophy of Symbolic Forms and Essay on Man (and Myth of the State...but now I'm giving too many book suggestions!) have something to say about this. Particularly in that there are always elements of myth making their way into the other symbolic forms (art, language, science, etc). A tension is always existing between these forms, so that, to take them as wholly separate and unaffected, is itself already supposing a myth. (perhaps it might be said that I am saying that from a mythical presupposition-exactly. You can't get out of it, only, maybe, become aware of it)
Anyway, if you haven't, check Cassirer out!
Posted by: Jb00m | January 05, 2011 at 03:24 AM
Hi Jb00m, and welcome to the Game!
I haven't read any Cassirer but he does sound right up my street. My reading list is absolutely booked solid right now, but I'll certainly start looking for his books as I scour the shelves of second hand book stores for philosophical oddities. :)
I'll have a look over his entry in my Encyclopaedia of Philosophy if I get a chance this afternoon.
All the best!
Posted by: Chris | January 05, 2011 at 02:25 PM
Jb00m: I picked up a copy of Cassirer's "Language and Myth", which has been absolutely fantastic! Right up my street, and incredible to think it was written in 1925. Many many thanks for the tip in this regard!
Posted by: Chris | February 23, 2011 at 08:44 AM
Awesome! I find myself being blown away by what he says, when he said it, quite a bit too. Good stuff! Also, congratulations on the birth of your son. Pretty momentous.
Posted by: Jb00m | March 07, 2011 at 03:35 AM