Work in Progress: Observational and Theoretical Components of Scientific Discovery
Friday 29 July 2022
Despite the efforts of philosophers over thousands of years, there is no science of science, as I have often pointed out. Science can formulate an organized body of knowledge of some empirical phenomena, but it cannot give an account of itself as an organized body of knowledge. There are many philosophical attempts to put science on a firm footing, but none that command the respect of all who are competent to judge.
Even in terms of large-scale structures of scientific discovery, it is difficult to converge on any generalization. For example, in the case of big bang cosmology and the expanding universe, it was Hubble’s observational approach that ultimately turned the tide on recognizing a universe with a natural history, and not the work of theoreticians, although some models of general relativity (specifically, Lemaître’s) did predict an expanding universe. On the other hand, it was the theoretical work of Darwin that provided the framework for the natural world with a natural history, and not the observational work of paleontologists, geologists, or biologists.
There is a good passage from Darwin’s autobiography that makes this point nicely:
“It has sometimes been said that the success of the ‘Origin’ proved ‘that the subject was in the air,’ or ‘that men’s minds were prepared for it.’ I do not think that this is strictly true, for I occasionally sounded not a few naturalists, and never happened to come across a single one who seemed to doubt about the permanence of species. Even Lyell and Hooker, though they would listen with interest to me, never seemed to agree. I tried once or twice to explain to able men what I meant by Natural Selection, but signally failed. What I believe was strictly true is that innumerable well-observed facts were stored in the minds of naturalists ready to take their proper places as soon as any theory which would receive them was sufficiently explained.”
Sometimes it is an observational breakthrough that drives scientific change (as with Penzias and Wilson detecting the CMBR without knowing what it was), and sometimes it is a theoretical breakthrough that drives scientific change. Mostly, it is theory and observation working hand-in-hand, sometimes in series and sometimes in parallel.
Einstein’s cosmology was a theoretical breakthrough, but Einstein was slow to acknowledge the expanding universe or quantum theory (Einstein wasn’t so much “quantum hesitant” as he was “quantum refusing”), so the developments of contemporary cosmology might be said to have occurred in series. With Watson and Crick’s initial paper on DNA, theory and observation was more in parallel, though perhaps it could be said that the fundamental breakthrough here was the insight into the structure of DNA, more theoretical than observational, but it would not have been possible without the observational component. Maybe if I thought through enough examples I might come up with a better example (or examples) to illustrate my divergent point: there seems to be no single component of scientific method that drives scientific progress. In a sense, this is as much as saying that the whole of scientific method is necessary to scientific discovery, which is true, but one might also spin this after the manner of Feyerabend, pointing to the essential anarchy of science, such that anything can count as an example of the scientific method under the right circumstances.
Now that I have mentioned it, it would be an interesting project from a philosophical point of view to specifically study scientific discoveries, from Archimedes to the present, with an eye toward identifying the crucial component, if any, in each instance of scientific discovery. To do this right it would be necessary to adopt certain conventions in laying out the scientific method (already done in various ways) and to identify certain practices with certain components of the scientific method. The result would be rough, given the compromises that would be forced upon one by the conventions adopted, but it might also offer some insights.
In any case, I have been thinking about the theoretical in relation to the observational for several months now, and my thoughts on this haven been sharpened (although not always clarified to the degree I would like) by my work on my upcoming NoRCEL presentation, “Toward Universal Biology: An Observational Scientific Research Program in Origins of Life.” I have previously included the abstract for my presentation, but here is it again:
Technology for exoplanet search has come into use only in the past thirty years, and we are now on the cusp of exoplanet atmospheric spectroscopy, which will reveal the atmospheric composition of some exoplanets. If life is not a cosmological imperative, if its appearance is not inevitable once the necessary prerequisites are in place, then there will be geologically complex planets that also develop chemical complexity without the particular form of chemical complexity that we identify as life. Differentiation of non-biological chemospheres from biospheres based on atmospheric spectroscopy may be possible through following the model of reconstructing stellar evolution. A taxonomy of planetary types surveyed at different periods of time in their development in different planetary systems may allow for the reconstruction of chemospheric evolution, in turn revealing a taxonomy of typical forms of chemospheric development for given planetary types. The more comprehensive our observational research program in reconstructing the history of chemospheres across diverse types of planets, the more likely we are to be able to reconstruct the typical developmental pathway of a living biosphere, allowing us to distinguish between non-living chemospheres and living biospheres, and to focus on the turning point at which the transition is made from chemosphere to biosphere.
Clearly, the emphasis here falls on the observational component of scientific research programs. However, even a primarily observational research program — one that, as Martin Dominik says, prioritizes exploration over speculation — still has a considerable theoretical component. This isn’t (or shouldn’t be) news to anyone.
The idea that all observation is theory-laden is usually traced to Karl Popper. Though many philosophers since Popper have elaborated this idea in considerable detail, Popper put it compellingly in his Conjectures and Refutations:
“…I tried to bring home the same point to a group of physics students in Vienna by beginning a lecture with the following instructions : ‘Take pencil and paper; carefully observe, and write down what you have observed!’ They asked, of course, what I wanted them to observe. Clearly the instruction, ‘Observe!’ is absurd. (It is not even idiomatic, unless the object of the transitive verb can be taken as understood.) Observation is always selective. It needs a chosen object, a definite task, an interest, a point of view, a problem. And its description presupposes a descriptive language, with property words; it presupposes similarity and classification, which in their turn presuppose interests, points of view, and problems.”
An observational scientific research program faces exactly this dilemma: what exactly are we to observe? In order to make our observation productive, and for it to be the basis of constructing scientific knowledge, we need to have already, if only in a vague and uncertain way, envisioned the end in view. That is to say, we need an hypothesis that we can attempt to confirm or disconfirm with our observations (what Popper called “a definite task”). And between our observations and our hypothesis, we require a relatively sophisticated structure of concepts to direct and control our observations, to establish the parameters of controlled observation (what Popper called “a descriptive language, with property words… similarity and classification”). If this conceptual framework does not exist, we have to construct it. This is just another way of coming at the same problem of observational and theoretical language that I discussed in newsletter no. 182 (inter alia).
Working on my presentation I have come to understand that the conceptual framework requisite for my hypothesis (as well as a number of subsidiary hypotheses that cluster around the main hypothesis) does not conform to the familiar division of scientific concepts into the classificatory, the comparative, and the quantitative. This is doubly interesting insofar as the concept of habitability does conform to this model of concept formation, and there is a very close relationship between the concept of habitability and the concept of chemospheric succession outlined in the above abstract.
Habitability is a comparative concept (as I discussed in newsletter no. 192), and thus well-suited for exactly the kind of concept formation delineated by Carl Hempel in his Concept Formation in Empirical Science. However, the kind of chemospheric taxonomies I am postulating as part of the observational scientific research program in origins of life is not a comparative concept, nor does it seem to conform to the familiar categories of a classificatory concept or a quantitative concept, although it could be assimilated to either or both of these. But it really stands out that a typical developmental succession of a chemosphere is not a comparative concept, and this is interesting in and of itself.
The idea of a typical planetary development (like the typical development of stars familiar from stellar evolution) is more like a biological concept, or an ecological concept, than it is like a concept from geology, physics, or chemistry, which latter sciences usually lie closer to the base of all schemes of scientific unification. And logical empiricists (like Hempel) never had much truck with biology and ecology, notwithstanding the best efforts of J. H. Woodger. It is interesting both from a theoretical standpoint, as well as from a sociological standpoint, that logical empiricists usually liked to cite physics as the paradigmatic example of science, and they focused on the kind of concepts that are employed in physics. Philosophers of biology and ecology, who would cite biology as a paradigmatic example of a science would also likely to focus on the kind of concepts that are employed in biology. This is only natural; one should not realistically expect otherwise.
I have many times mentioned Carl Hempel’s Fundamentals of Concept Formation in Empirical Science, as one of my go-to works for clarification, but now I must go where classificatory, comparative, and quantitative concepts have not gone before. My reliable guide to concept formation in science, Hempel, has left me bereft. I might compare myself to Dante, who is conducted through Hell and Purgatory by Virgil, only to be left by the poet having traversed Purgatory, since the pagan poet cannot enter the Christian heaven. But whereas Dante is then taken up by Beatrice at the gates of Heaven, I have not yet found my Beatrice to guide me through the realms of heavenly concepts.
The analogy, while fun, quickly breaks down, because the distinctive thing about the ecological concepts I need to formulate is that time is essential to them (the analogy breaks down if we assume that Heaven is essentially atemporal, i.e., eternal). Classificatory, comparative, and quantitative concepts might be temporal, non-temporal, or atemporal; they are indifferent to time. Indeed, it has often been pointed out that the equations of physics work with time running either forward or backward, and this has led to a lot of nonsense speculation about the possibility of time going backward. Rather than conclude that the currently understood mathematical expression of physics is incomplete, some will say that the equations are just fine, and we are simply only seeing one side of physics, which could just as well run backward from the familiar directionality of time.
What is needed is a careful philosophical analysis of the formation of essentially temporal concepts, that is to say, concepts that would be incoherent without their temporal component. Perhaps I need to go back to Heinrich Rickert’s The Limits of Concept Formation in the Natural Science, which is primarily concerned with history, and see if I can find what I am looking for there, after a fashion, as one could assume that the historical sciences make use of essentially temporal concepts. But perhaps the definitive Fundamentals of Concept Formation in the Temporal Sciences has yet to be written.