We think of science as producing theories, the final result of scientist’s work. But when we try to identify what kind of thing a theory is, we run into a big problem: None of the accounts available seem to accurately describe what we know about how scientists work. Perhaps we should simply abandon the idea that science produces theories altogether and shift our focus on the actual practices of scientists. If we do that, we’ll not only gain a better understanding of the history and philosophy of science, but we’ll get closer to answering the biggest question of all: “How does science actually work?”, writes Steven French.
What is a scientific theory? We can point to any number of books and papers, seminars and YouTube talks in which all kinds of theories are presented and discussed but it would be odd, to say the least, to take any one of those books or papers and identify the theory concerned with that. Let’s take the General Theory of Relativity as an example. We can hold up and read Einstein’s original journal publication but, again, it would be bizarre to take that particular paper as the theory, not least because it was written in German, which would mean that the English translations and the re-presentations of it in countless books and other journal articles were something else, something other than that theory.
Considered as an entity of some sort, the General Theory of Relativity seems to transcend these physical manifestations. Indeed, many have suggested that theories in general are abstract things, like numbers or, more contentiously perhaps, musical works. Karl Popper, for example, argued that as well as the world of physical things and the world of mental entities, there existed a ‘World Three’, inhabited not just by scientific theories but by pieces of music, such as Beethoven’s Fifth Symphony and works of literature, such as Shakespeare’s Hamlet, waiting to be discovered. More recently Amy Thomasson has presented a similar view, arguing that such artworks, and scientific theories, can be regarded as ‘abstract artefacts’ that are brought into being by their creators’ intentions. But such views, sophisticated as they are, nevertheless jar with what we know about the process of coming up with a theory.
Consider again Einstein’s General Theory of Relativity. This did not just pop into Einstein’s head in some sort of ‘lightbulb’ or flash of inspiration moment, as it is so often portrayed in popular accounts of scientific work.
Where is the theory? Steps, missteps and proto-theories
Consider again Einstein’s General Theory of Relativity. This did not just pop into Einstein’s head in some sort of ‘lightbulb’ or flash of inspiration moment, as it is so often portrayed in popular accounts of scientific work. The steps and missteps, false starts and dramatic leaps forward that Einstein took have been extensively charted and we now have a pretty good understanding of the heuristic moves that led him from his initial thoughts in 1907 to the presentation of the core set of equations to the Prussian Academy of Sciences in 1917. So when, exactly, in that process was the theory discovered or come into being? Picking the time at which he made the final full stop in his 1917 paper, say, seems absurdly arbitrary. Alternatively, you could say that all of his initial efforts, his proto-theories and half-way attempts are also ‘there’ as abstract entities in World Three, or wherever. However, if that’s the case, and not only for Einstein’s work of course but for all theories, then Popper’s world is going to be very densely populated! In either case, then, it seems hard to reconcile this view of theories as ‘out there’ with what we know about the actual practice of theorising.
My own suggestion is to cut through the confusion that arises when trying to identify what exactly a theory is and simply deny there any such things as theories. There is an immediate objection that can be made, of course: we talk about them, we describe them, we ascribe various attributes to them, and so on. So, we might say that ‘General Relativity is elegant’ or more prosaically perhaps, ‘General Relativity has been confirmed by the LIGO observations of gravitational waves’. What makes those statements true if not the possession of certain qualities by the theory itself?
However, there are lots of cases in which the feature that actually makes the statement true, isn’t directly referred to by the statement itself. Consider another example: ‘The desk that I’m typing this on is solid’. Fundamentally, what makes that statement true is the arrangement of the relevant collection of elementary particles. In this case, what makes the statement true doesn’t appear in the statement at all. Likewise, we can take ‘General Relativity has been confirmed by the LIGO observations of gravitational waves’ to be made true by the various practices involved in making those observations, together with the theoretical practices involved in deriving the relevant prediction (originally made by Einstein himself in 1916). What makes ‘General Relativity is elegant’ true is less straightforward but again will involve certain practices, having to do, perhaps, with the way certain sets of symbols are set down in a certain order, or, alternatively, with the way a certain range of phenomena can be explained in a parsimonious manner. So, we can eliminate theories as things that exist in Popper’s World Three, say, but still regard statements that mention them as true. (We can also make a similar move when it comes to desks and eliminate them as well, but that is much more contentious!)
Steve Fuller, Kenneth Cukier Lawrence Krauss ask whether science along can uncover the truth
Why Thomas Kuhn was wrong
There is more to be said, of course, but what I want to explore here are the implications of this shift from a focus on theories as things of some kind, to the practices that make true those statements that are purportedly about them – in particular, the implications for our understanding of science, and its history.
So, let’s consider one of the most well-known books on the history of science that famously opened up an entirely new perspective on our understanding of how science works: The Structure of Scientific Revolutions by Thomas Kuhn. According to Kuhn, science has proceeded in the following manner: within a given discipline there is a dominant ‘paradigm’, organised around some theory construed as a kind of exemplar which shapes what Kuhn termed ‘normal’ science. Eventually, however, anomalies begin to accumulate, in the form of phenomena that the theory cannot explain, and a sense of crisis takes hold until there is a shift to a new paradigm, with a new theory at its heart in terms of which normal science can once again be pursued, albeit in a different form than before. An example that might appear to fit this mould is the so-called Quantum Revolution of the early twentieth century in which the ‘classical’ theory of mechanics originally developed by Newton struggled to account for a range of phenomena, from atomic spectra to the photoelectric effect. As the number of anomalous phenomena grew, Newton’s theory was replaced by the new quantum mechanics of Bohr, Heisenberg and Schrödinger and a new ‘paradigm’ was established.
However, Kuhn’s focus on theories being replaced in abrupt shifts is misleading. Kuhn himself says very little about quantum theory in his famous book, perhaps because when he conducted interviews with those quantum physicists for the American Institute of Physics’ Oral History project in the early 1960s, the interviewees steadfastly declined to adhere to the above narrative! If, instead, we look at the various practices, both theoretical and experimental, that were engaged in at the time, a very different picture emerges in which different streams come together and coalesce or separate and fade altogether, different approaches are suggested and developed, and shown to be inter-related and different stances are taken on how to understand these developments. Just a quick glance at the various textbooks produced at the time will reveal the different practices in play and focussing on those and the different views they embody reveals a much richer and more nuanced picture. Indeed, with those practices brought to the fore, it is hard to escape the conclusion that the claim that there was a point at which the old, ‘classical’ view of the world was dramatically overthrown is one that has been read into the history retrospectively and from a theory-oriented standpoint. Moving away from that and dropping the idea that theories are ‘out there’ somewhere, as abstract attractors to which scientists will naturally gravitate, frees us to envisage a subtler and more sophisticated history.
Kuhn’s focus on theories being replaced in abrupt shifts is misleading.
Beyond the art analogy of representation
Changing the way we think about theories also allows for new perspectives on how we understand issues such as the nature of scientific representation. Typically, discussion of this has oriented itself around a comparison with representation in art, with detailed and complex accounts that accommodate both paintings, on the one hand, and scientific theories, on the other. However, such discipline spanning frameworks encourage the view that theories, like paintings, are essentially static representations and just as these frameworks typically treat a painting as a thing, sitting at one end of the representational relationship, with some scene or other at the opposite end (think of Constable’s Haywain, for example), so they implicitly regard a theory likewise, perhaps as an abstract entity, representing some phenomenon or aspect of the world.
Moving away from that way of thinking opens up space to consider representation in science as much more fluid and transitory as, again, both theoretical and experimental practices change and evolve. So, rather than thinking of the development of science in terms of scientists coming up with fixed representations of phenomena, as artists do with their paintings, we should think of it in terms of complex, messy practices that happen to make true - at certain points, in certain contexts - claims about representations.
Furthermore, making that move encourages those of us engaged in constructing such accounts to reflect on what it is we are doing in our efforts to understand how science works. Attempts to understand science in terms of how theories represent phenomena should themselves be understood in terms of certain practices that conceive of that relationship of representation in different ways but – and this is the crucial point – given the different aims involved, it may not be the case that any one such way can be deemed unequivocally better than another.
According to one, long-held philosophical position, theories are sets of sentences, organised logically, whereas according to another, currently dominant, they should be conceived of as families of models. Each offers a different account of scientific representation, as well as the relationship between theory and evidence, and, indeed, how science works more generally. There is a long history of vigorous philosophical dispute between the advocates of these two positions but what I am suggesting is that rather than regarding them as alternative descriptions of theories, as things ‘out there’, we should consider them as offering different tools that we can deploy to better understand scientific practice.
Reconceptualising both our historical reconstructions and our philosophical ‘understandings’ by shifting our focus to the relevant practices, rather than on theories as things that are fixed in some abstract realm or other, opens the door to novel and exciting ways of thinking about how our understanding of science can be enhanced. And more generally, if we accept that there are no such things as theories, this can free us up to entertain and explore new responses to that old question, ‘how does science work?’.