Monday, 23 May 2016

Revisiting Scientific Revolutions  Incandescent Collection
Posted by Thomas Scarborough 
Thomas Kuhn was wrong.  He failed to understand the dynamics of scientific revolutions. Far from such revolutions occurring through an accumulation of evidence – until, so to speak, the dam bursts – they fail to occur until such time as scientific constraints have been weakened – namely, the scientific method.  I shall explain.
In recent generations, we have witnessed a rising awareness of an inter-connected world, and cosmos. One of the results of modern science in particular is the perception that 'everything is related to everything else'.  Yet paradoxically, even at the same time, we find that science requires the very opposite of openness to the totality of things, to survive and to thrive. For science to advance, there is the need for scientific control on the one hand, and a strictly normed language on the other. In the words of Wilhelm Kamlah and Paul Lorenzen, science must 'screen things out'. This applies to all four phases of the scientific method: characterisations, hypotheses, predictions, and experiments.

There is something equally true about science which we typically do not much pay nearly as much attention to. If the scientific method should exert any influence on those potential influences which it excludes, then scientific control is compromised. For instance, if in seeking to establish how much energy is required to convert a kilogram of ice into steam, I find that I am warming the laboratory at the same time, then the procedure is fundamentally flawed. Energy is being lost. We therefore require what I shall call a 'double isolation' in science. Not only does science screen things out, but it needs to screen itself out from its environment.

This 'double isolation' has led historically to two major problems:

Firstly science, having screened itself out from the world, ultimately needs to 're-enter' the world. After the final, experimental stage of the scientific method, with the artificial conditions of the laboratory removed, science begins again to have an effect on the world. Yet little thought is given to what happens at this point. Science, when it re-enters the world, typically goes beyond anything that was formally taken into account in the scientific method. The disasters which have here occurred have led various thinkers to suppose that science is responsible for the ruination of our world. Stephen Hawking puts it simply: science may score an own goal.

Secondly, the isolation of science from the world has resulted in confusion as to how science really advances. The orthodox view is that science advances by and large through an inductive process: by making broad generalisations from specific observations. Yet consider that those specific observations have already followed the procedure of 'screening things out'. That is, such science has already minimised the effects of variables. It has excluded a great many possible relations in order to trace the relations which it does. There is a limit, therefore, to what can be achieved with previous scientific observations, as far as the tracing of relations is concerned.

Not only this. Experience tells us that scientific conjectures are not adequately explained by an inductive method. Here is an example from my own experience, dating from 2004. In that year I came up with a new principle for metal detecting, called coil coupled operation (CCO).  I was already familiar with the transformer coupled oscillator. This is governed by theory which, even in its full complexity, has little or no interest in outside influences on the oscillator. Now consider that such outside influences could include coins beneath the soil, which may change the frequency of the oscillator. In order to turn this into a metal detector, my mind needed to leap outside of the theory, to discover a principle which rested precisely on those influences which the present theory excluded.

Science, therefore, would seem to require not only the inductive method, but something far larger – namely intuition. Albert Einstein wrote, 'Knowledge is limited to all we now know and understand, while imagination embraces the entire world, and all there ever will be to know and understand.' This has important implications for the scientific method. The inductive method should be taught only as one possible means of doing science – and probably not the best way. Rather, the emphasis should be on a more frenetic and imaginative thinking. This is borne out, among other things, by the fact that many scientists of note were inter-disciplinary or multi-disciplinary in their pursuits – among them Archimedes, Leonardo da Vinci, and Albert Einstein.

On the other hand, science should take account not only of the individual mechanisms which are isolated in controlled experiments. It should deliberately keep track of those mechanisms which are excluded from such control. These may potentially be infinite – yet it is crucial that there be an attempt to list them. No experiment is truly complete until this has been done, and no experimenter has been truly responsible without it. Inconsistently, today, some of our scientific pursuits are systematically regulated and supervised after the final, experimental stage of the scientific method – most notably in the areas of food and drugs – while vast areas remain ill-considered. The scientific method, far from being closed after four stages, should be an open-ended process.

This is intimately connected with the philosophy of scientific revolutions. In the process of 'screening things out', scientists' thinking is constrained. Yet a paradigm shift requires an eye for the wider canvas of relations. Therefore science, through the very scientific method, works to prevent paradigm shifts. However, as a science advances, the need for scientists to 'screen things out' becomes weaker. The work of scientific control has been done, and the ability to think creatively becomes stronger. Rather than paradigm shifts occurring through an accumulation of evidence, they occur where the scientific method is weakened – like a housewife, perhaps, who after kneading her bread, looks up to see the sun rise.


  1. It is indeed true, Thomas, that much science advances through orthodoxy and induction. In many scenarios, that’s all that’s called for and is therefore fitting: What’s being performed may be of an everyday (even routine) nature, not requiring soaring flights of imagination. That’s typically not the most exciting, inspiring science, but it often suffices to fit the bill—it gets the job done. And its outcomes may still matter in this everyday scheme of things. Other science, though, requires the shattering of orthodoxy and evokes creative ways to alternatively model and describe how things work in the world (and the universe), and what might happen differently. I would suggest this entails not just tweaking the variables (inputs to hypotheses), but the grander ability to envision alternative realities from a bird’s-eye, big-picture perspective—along with ‘intuition’. Accordingly, one way I’d differ with the essay is in this statement: “Yet paradoxically . . . we find that science requires the very opposite of openness to the totality of things, to survive and to thrive.” Rather, in the many scenarios where the best of science envisions alternative realities, with sweeping perspectives, both the method and the result necessarily steer abruptly away from orthodoxy (including strict induction), leading to the kinds of historical ‘paradigm shifts’ you refer to.
    This brand of science, characterised by the model-shattering mentioned in the preceding paragraph, requires ‘imaginative thinking’. And, I would argue, it requires even more: the gift to think in terms of quickly moving what-ifs that come in and out of the pantheon of ideas swirling around in the scientist’s head; to mentally mould and re-mould—and tirelessly re-mould yet again—alternative realities, with unbounded mental agility; the scientist’s resistance to prematurely vetting his or her ideas; and not to be intimidated by the possible succession of models (paradigms) that the scientist must course through, perhaps many times unsuccessfully, in order to discover what works best in shaping an entirely fresh vision of reality. All this still qualifies as the ‘scientific method’—especially theoretical, though also practical—but, clearly, a scientific method of a vastly more bold, mind-boggling, and ambitious kind. That is, how science does, at its best, work for most imaginative effect—essential to achieving the ‘scientific revolutions’ you refer to! I enjoyed your take, Thomas, on how science does and does not play out.

  2. Thanks for the thoughtful comment, Keith. All this reminded me of Paul Feyerabend. ONe of his many great points again:

    ‘When Copernicus introduced a new view of the universe, he did not consult scientific predecessors, he consulted a crazy Pythagorean such as Philolaos. He adopted his ideas and he maintained them in the face of all sound rules of scientific method. Mechanics and optics owe a lot to artisans, medicine to midwives and witches. And in our own day we have seen how the interference of the state can advance science: when the Chinese communists refused to be intimidated by the judgement of experts and ordered traditional medicine back into universities and hospitals there was an outcry all over the world that science would now be ruined in China. The very opposite occurred: Chinese science advanced and Western science learned from it. Wherever we look we see that great scientific advances are due to outside interference which is made to prevail in the face of the most basic and most ‘rational’ methodological rules. The lesson is plain: there does not exist a single argument that could be used to support the exceptional role which science today plays in society.’

    - from a talk called 'How To Defend Society Against Science'

  3. Thank you Martin. The trouble with Feyerabend is that he gets brushed off as the lunatic fringe. Perhaps his thoughts lack the requisite system.

    Laurent Schwarz wrote, with regard to mathematical drudges, that "once their task is accomplished, the ideas of the scientists with a penchant for generality come into play". He did not apply this to paradigm shifts, though, as I do in my post.

    With regard to Keith's comments, I personally obtained this: that I probably underplay the significance of the drudges, and I shall revisit that (my post is a precursor of a chapter in my second edition metaphysic).

    1. If we're talking philosophy of science, Feyerabend is taken very seriously, as is Kuhn (who he detested).

      If we're talking 'mainstream science', then he is brushed off, yes, but such people don't have anything to say about how science itself operates.

  4. Your very stimulating essay, Thomas, prompted a few thoughts regarding the scientific method more generally …

    As for induction, Karl Popper offered this bold, unequivocal statement: “Theories can never be inferred from observation statements or rationally justified by them. I found [David] Hume’s refutation of inductive inference clear and conclusive.” He continued by observing that one can never conclude reasonably that scientific generalizations are true—no extrapolation from a track record. But science does succeed, and eminently so, through induction, of which there are myriad examples, without having to mine deep into history. Here’s just one …

    A recipe for future abundant, safe, clean energy: Fusion. Combine deuterium and tritium. Heat the mix until the isotopes fuse, releasing energy. Confine the process in a magnetic field. Scale it up to market levels, with energy out far exceeding energy in. The result is a miniature ‘sun,’ ready to service the world. Okay, any physicist worth her or his weight in plasma would point out that I’ve vastly oversimplified. But amidst the usual chorus line of energy sources—oil, coal, natural gas, solar, wind, water, geothermal, biomass, and, sure, nuclear (fission)—sits often-overlooked fusion.

    Much more hard science and engineering are needed to nail this, which the United States, the EU countries, Russia, China, Japan, Canada, India, and Korea are all doing, in spades. It’ll continue to be tough sledding. The timeline is undoubtedly measured in decades, not just years. But scientists and policymakers alike have long since thought that we should also think long term, while meanwhile proceeding with short- to medium-term fixes to the human-made carbon emissions driving climate extremes and their nasty effects.

    Fusion power plants have tremendous promise as a source of global energy. And there are selling points: Deuterium, available from the oceans, and tritium, from lithium, are inexhaustible. Zero carbon emissions (it’s clean). No long-lived radioactive waste. A safe, nontoxic by-product in the form of helium. No runaway risk (unlike the Chernobyls and Fukishimas of the world). And the production of mammoth amounts of energy to fuel the world.

    The pathway to fusion energy represents a classic success story of the best of induction. Including what’s left to do before this source eventually enters the global marketplace. It’s induction that has helped direct scientists and engineers in this process—round and rounds of theorising, testing, analysing (experimental data), abandoning the bad, and revising the model yet again. Underscoring Richard Feynman’s point: “If [a new principle] disagrees with experiment, it’s wrong.” Deduction can’t litigate where the mistakes are, or whether theory is wrong. Popper notwithstanding, induction does that.

    Anyway, just thought I’d toss a few random ideas into the hopper related to this interesting topic of the scientific method.