Science and Explanation
The Magical Flower of Winter is an essay series exploring reality and our relationship to it. It deals with philosophy, science and our views of the world, with an eye on the metacrisis and our future. Sign up to receive new essays here:
The order and regularity in the appearances, which we entitle nature, we ourselves introduce.
Immanuel Kant - The Critique of Pure Reason
Science, as we should now have started to become familiar with, is many things, depending on perspective and context. It is a profession of vast variety, a practice by millions of people worldwide, an intricate family of knowledge domains, a label of a great many frameworks, and an element in a great many world views. This essay will begin with an exposition of a particular dimension of what science is: an extension to our experiential concepts. As always the hermeneutic circle requires us to enter this explication in medias res, to step into a motion already underway, and to take certain things as given, which we later revisit and validate according to the resulting view. I take as a starting point that an essential part of experience is expectation, that an inseparable component to normal experience is the effort to anticipate and predict what is to come (This component will be further investigated, and this particular hermeneutic circle closed, in World, Model and Mind). In our experience we face phenomena and events that do not conform to the expectations we generate. These we try to find explanations for in order to minimize future discrepancy between expectations and experience. Part of this construction is undoubtedly automatic, an innate ability to update our explanation model without active effort, but another part of it requires a conscious effort, to the extent that we are aware of and weigh alternatives and imagine different outcomes. We thus have explanation as a tool for prediction, we extrapolate into the future those explanatory constructs we create for structuring our past. Science is the inter-subjective extension of this cognitive ability, a practice among participants that allows an increase in the scope of our explanations and predictions. As a practice among several practitioners, science introduces elements not present at the individual level of explanation, for instance standards of evaluation and theory choice like simplicity and empirical completeness and adequacy. The important point is that science is an extension to our cognitive abilities, it projects the explanatory and predictive constructs into territories unknown to first-hand experience.
That science is an augmentation means that there is no conceptual difference between the terms we use in describing our direct experience (“observational terms”), and the terms we use in science (“theoretical terms”), due to the holistic working of language (See Language and Meaning and Wittgenstein and the Private Language Argument). Scientific terms and concepts have no different status than experiential ones, they are equally part of and subject to the evolving web of language and meaning. If we analyze a scientific term (theoretical, synthetic), we will find only a vague border with non-scientific terms (observational, analytic), that there is no separation between these two kinds, and that the distinction dissolves1. This is consequential, because due to this the authority and justification of the claims of science should be treated no different than claims we make about our direct experience. The claims might be subject to different standards of justification, critique or evaluation, but conceptually they are equal. The reason I stress this aspect is to make clear that science can have no special status except as what we make it to be, except as how we use it. Science is just as anti-foundational as language and meaning, equally self-upholding. As van Fraassen stated it: “We must accept that, like Neurath’s mariner at sea, we are historically situated. We rely and must rely on our pre-understanding, our own language, and our prior opinion as they are now and go on from there. Rationality will consist not in having a specially good starting point but in how well we criticize, amend, and update our given condition.”2
Explanation
To say a bit more about explanation, let us view it along two related axes: reduction-emergence and causal-teleological3. Along the first axis we have reduction at one end and emergence on the other. With reference to Figure 1, unprimed letters represent macro-phenomena, while primed letters represent micro-phenomena. To say that A reduces to A’ is to say that A is nothing “over and above” A’, that there is some (sufficiently) perfect mapping, law or bridge in terms of which an understanding of A can be transformed into an understanding of A’. This does not exclude that one may not understand things more simply or differently at one level. The classical example of reduction is thermodynamic: heat as a macro-phenomenon reduces to molecular kinetic energy at the micro-level. At the other end of the scale we have emergence. To say that A is emergent on A’ is to say that there is something to A “over and above” A’, i.e. that there is something more to the macro-phenomenon we are unable to represent at the micro-level, indicated in Figure 1 by CA. An example of an emergent property would be the wetness of water. For completeness, another term used in this discussion is supervenience, which more generally denotes a relationship between properties or facts at two levels where A supervenes on A’ if and only if some difference in A’ is required for any difference in A. We see that all reduction is supervenience, while not all supervenience is reduction. This area is lively discussed, but there are those who argue that emergence in the strong sense is incompatible with supervenience: that there can exist properties in A completely autonomous from properties in A’. The classical example of strong emergence is consciousness, in particular qualia, that there is something it is like to have consciousness or be conscious. The question of whether consciousness is in fact reducible to the micro-level (to neural correlates) is the mind-body problem we have encountered before, the central issue in the physicalism/non-physicalism debate in the philosophy of mind.
The holistic approach makes much of the reduction/emergence debate moot: Reduction is the case when the context of phenomena, C, is either sufficiently irrelevant, i.e. the elements of reduction are adequately separable from their context, or context is already part of our understanding of the phenomena. The latter can be due to meaning plurality, like our understanding of heat as both experiential and thermodynamic. Before a conceptual connection was made between experienced heat and thermodynamics, i.e. prior to the thermodynamic paradigm, the former was irreducible. Emergence is the case when the context of phenomena is relevant and not accounted for. In the emergent case, A is something over and above A’, because an understanding of A is dependent on a context CA that is untranslatable to the micro-level. The wetness of water is an entirely contextual property, a property that requires an experiencing subject for its existence and not just water molecules. In a sense, emergence is what the reductionist finds when faced with the contextual and holistic aspects of reality, or when the physicalist runs up against the limits of meaning. As explicated in Wittgenstein and the Private Language Argument, we will not satisfyingly close the gap between mind and body, because this separation is built into the very premises of the problem, there is no concept of mind at the micro-level as this is external to the contexts that consciousness acquires its meaning in.
Let us next look at another, related axis. The causal-teleological axis becomes relevant when we consider processes in time, and not only at different levels. At one end we have causal explanation, that an understanding of B is complete in terms of its causes A. The typical example would be idealized billiard ball collisions: the motion of a billiard ball that has been struck is completely determined by the mechanical details of the collision. This case overlaps with reduction to the extent that we cannot have a causal account of emergence insofar as B manifests properties irreducible to properties in A. From the holistic perspective, the causal account is one where we are only interested in a limited domain of aspects, for only in a limited sense can causal explanation work, being irreducibly reductive. At the other end of the axis we have teleological explanation4: that an understanding of A is incomplete without an understanding of B, that “the end” B figures into the explanation of A. Teleological explanations are usually considered to be controversial, standard examples being evolution by design, the belief that some agency is guiding evolution. But there is another, less irrational conception of teleology that enters the stage when we see explanation from a holistic stance that bears a connection to emergence. In order for an explanation to be intelligible, we require a connection from our macro-level concepts to micro-level theory. It is often the case that our knowledge of the limitations of a particular process, or of the space of possible configurations, will impact how we frame the process or how we specify the configuration. This knowledge is posterior and external to the process under consideration, thus macro-level concepts (“final causes”) will influence what micro-level considerations we care about based on this connection. Explanations may seem teleological for epistemic reasons of simplicity and intelligibility. From the holistic view, this is another contextual phenomenon: phenomena are inseparable from their context, in this case a context of intelligibility, CB. In order to understand a process we cast it in the terms of a framework known to us, and this “downwards” process of construction can be easily perceived as teleological. An example is the anthropic principle: that the universe could not have been different from what it is in certain respects, for then we would not be having the experience we are having. In its cosmological variant, this principle explains that cosmological parameters for our universe will appear fine-tuned because if they were not we would not be here to discuss them at all. This is of course not a satisfying or insightful explanation, because from the context of particularist science, we expect that a satisfying explanation would be reductive and causal, not teleological. But there are cases where teleology is utilized successfully, and its use is explanatorily satisfactory: approaches where one is required to consider the whole. Such an approach is for instance the stationary action principle in classical mechanics (and in Feynman’s path integral extension in quantum theory): To find the equations of motion for a system we need to optimize the action of the system, and this optimization requires a consideration of all possible trajectories. The knowledge of these trajectories is both over and above the micro-level of the system and posterior to the process that is considered. Thus, a consideration of the whole influences the parts, and this prior dependence on the final whole makes the approach appear teleological5.
Something should also be said about the temporal aspect of phenomena. We usually consider phenomena in part as snapshots in time, thinking we are justified in separating a system and its environment at one particular time. Relativity teaches us that this can only be achieved relative to a frame of reference, i.e. that “one particular time” can only be specified relative to a perspective. The breakdowns into past and future in the explanatory diagrams are as such entirely relative. Quantum theory further teaches us that the specification cannot be complete, i.e. that quantum effects makes specification itself an irreducibly incomplete process, in the sense that completeness means perfect specifiability. This means that the sharp separation into system and context, or even into past and present, is a conceptual tool with limited applicability. These aspects should further motivate a holistic understanding of science and explanation, as opposed to a purely causal-reductive one. Crucially, what this account showcases are the limits of explanation. In future articles on quantum theory and general relativity we will come to know further examples where we come up against the very boundary of explanatory reach, where our notions of explanation and the epistemic itself seemingly come to an end. The contextual considerations of explanation should amply put weight to the participatory relationship of us as explainers to our explanations. As I have cautioned before, we must keep in mind that all these aspects pertain to explanation and our framework for representing reality. Reduction, emergence, causal or teleological are classifications of the epistemic, of our framework. If my ordering “Fetch me that hammer!” is followed by you giving me the hammer, we naively think that the order, the higher-level sign, is a cause of the lower-level execution of the action, when the sign is just a piece of our framework, purposive to us in explaining and justifying, a consequence of language and epistemics. Both reductionism and emergentism is confusing the map for the territory, as reality itself happens, regardless of our framework. But do I not then contradict my own critique of “independence”? Reality and experience is co-dependent, that reality is, is irreducible and thus independent of the epistemic. But reality is dependent on experience through requiring a subject to be for.
Scientific Realism and Constructivism
A common position in philosophy of science is scientific realism. This view can be distinguished by three dimensions6: ontological independence, semantic correspondence and epistemic optimism. Ontological independence is the component whereby there is a reality independent of us, our existence, thought and language. Semantic correspondence indicates the relation between this independent world and language: Truth is understood as correspondence between world and language. According to the scientific realist, the goal of science is truth. The last component, epistemic optimism, indicates the belief that science is the best tool for uncovering “truth”, and that one chooses a scientific theory if it is closer to “truth” than the alternatives. In order to make explicit the claims and beliefs entailed in scientific realism I will narrate a story related to the experience of heat and cold and the concept of temperature. The story is not intended as representing any actual turn of events, but it aims at being a plausible and coherent narrative, though it may seem “fantastical” to the particularist part in us. You will find that the story I tell is colored by a constructivist7 variant of scientific anti-realism, but this should not matter for the points I am trying to bring out.
It all starts with experience. One more or less repeating aspect of experience is gradually identified, there is similarity to what we find in our memory, we recognize a certain dimension of our experience as something resembling what we have felt before. A set of terms, again gradually, are created and iterated on by the community in order to describe this aspect (for a range of purposes relating to communication, cooperation, survival etc.), like “hot” and “cold”, “sweating” and “freezing”. The terms form a cluster that mutually define each other: the meaning of one term is in relation to the others. We start correlating certain other aspects of experience to these concepts, like water freezing and boiling and the pain of fire. Thus the conceptual cluster grows. At some point, someone, by chance or by observation, by trial and error, by hypothesis and experimentation, by a creative spark, identifies a set of behaviors that covary with “hot”, “cold” and the rest of the conceptual cluster, and reproducibly so through a systematization of these behaviours. Explainability, predictability, simplicity, adequacy etc. are all likely criteria that play a part in this process to some extent. This systematization is an apparatus, an instrument, that in some sufficient range of conditions predictably varies constantly with the experience of “hot”, “cold” and so on. This instrument gets the name “thermometer” and it is useful, because it allows one to read off from a scale a number that represents how “hot” or how “cold” it is, a scale one learns to associate with both one’s experience of “hot” and “cold” and other experienced phenomena that covary with one’s experience of these. We can call the scale “temperature”, abbreviated T. The number we read off we can call a T-measurement. T, T-measurement, and other terms belonging to the process that brought about the instrument constitute a conceptual cluster that, from the start, is co-dependent on the pre-existing cluster of concepts relating to “hot”, “cold” etc. The latter in fact receive a reinterpreted meaning in terms of the instrument. “Hot”, in addition to its old thermometer-independent meaning, now also has a thermometer-dependent meaning, i.e. a more or less vague correspondence to a range on the scale. Thus, we could talk about “instrument-independent hotness” to imply the aspect in experience, or “instrument-dependent hotness” to imply the added aspects related to the instrument.
What the realist now does constitutes the truly realist turn. The experiential narrative so far is burnt, and a new one is set up in its place. We take the instrument-dependent and experience-dependent aspect as reflecting something that is “really” independent, an objective world outside our experience: «In the beginning there was nothing, then a great bang…» What was primary, experience or mind or consciousness, is now secondary to a physical reality of objects or fields. Experience now reduces to, emerges from or supervenes on a material substrate that science is able to describe. Consequently, we believe in ontological independence. Furthermore, since we now have ontological independence and our conceptual structure is in seeming correspondence to the ontological aspects, we believe in semantic correspondence: The correspondence between the independent world and our language must be as it is because of this relation being one of truth. Furthermore, because of the success of the scientific part of our language and of our practice of science we believe in epistemic optimism: that science must be the best tool for uncovering this truth-relation to the independent world. In this way the scientific realist sees his practice entirely outside its historical and social context, and is led to a narrative for the success of his/her view divorced from the evolution of the very framework, science, he/she claims dominance of. The scientific hope is founded on the idea that in the statistical aggregate, individual variance is eliminated, and so this, in addition to good design and methodology, removes the subjective effects due to all experiments being interaction. But what if subjective effects are systematic? Also, what manner of quantity have we arrived at in the statistical aggregate? For it corresponds to no instantiation in reality. The realist scientist has placed its creation outside itself, and forgotten how the creation came about in the first place.
An alternative narrative goes differently in its commitments and beliefs. Picking up the story, instead of the realist turn, we instead observe that we normally use only the one term, “hot”, to indicate both “experience-dependent hotness” and “instrument-dependent hotness”. Meaning is plural. There are conditions, contexts, in which the T-measurement does not align with our experience of “hot” and “cold” (e.g. very high air pressure or contamination). So the thermometer is not perfect, it displays contextual variability, but part of the process of making the apparatus was to create it in such a way that this contextual variability is minimized, though never zero. Thus, instruments are made so as to maximize the observation of certain aspects, and minimize others. The T-measurement is only meaningful, is only a T-measurement, in a very specific context. To talk about temperature or T-measurement outside of this context is to project theory outside the context in which it gets its meaning. We could not give a thermometer to a child without previous experience and have it understand how it works or what its purpose is. This the child would have to learn from others, or by extended observation of the behavior of the thermometer in a range of experiential situations. We are not licensed to say that T or T-measurement exists independently of this context, in terms of the “normal” meanings of “exist” and “independent”. If we do say that T or T-measurement exists, this is a different usage of “exist” than in normal experience, thus if we do use the word “exist” in this setting, the meaning of “exist” is expanded (varied) to include this projective, contextual and theoretical existence, just like the meaning of “hot” was extended to include the contextual concept cluster introduced by the thermometer. Meaning, now of “exist”, is once again plural. The instrument and the conceptual structure attaching T and T-measurement to the experiential cluster of “hot” and “cold” and so on is as such a construction.
Scientists, being equipped with a complex organism and embedded in constantly changing physical and social surroundings, used ideas and actions (and, much later, equipment up to and including industrial complexes such as CERN) to manufacture, first, metaphysical atoms, then, crude physical atoms, and, finally, complex systems of elementary particles out of a material that did not contain these elements but could be shaped into them. Scientists, according to this account, are sculptors of reality but sculptors in a special sense.
Paul K. Feyerabend - Conquest of Abundance
We could track a similar narrative for concepts like “electron” or “quark”8, though now the construction is far more complex, involving layers of instruments and a far larger cluster of conceptual structure and range of parameters and criteria that go into what constitutes scales and measurements. The interface with our experience that we conceive as the “measurement” (e.g. a data point or visual on a computer screen) stands on top of a complicated and dependent series of other interfaces. Is there an “ultimate” interface with reality? Not as independent, for each of the components is either covarying with some aspect of experience or covarying with some other component or concept in the whole. The measured is inseparable from the contextual process of its measurement, we can neither isolate it in time nor in space, for whatever is isolated stands without the relations and context by which it is what it is. Reality as something in itself, disconnected from our experience, drops out of the construction. And none of this is to say that reality exists solely in our minds, or that “we create the world” like magicians. The world happens, experience happens, we do not control this happening, but we do act in the world towards impacting the future, and what we control is how we structure this happening in our conceptual framework, though this is heavily constrained by our immersive context. Reality, experience, as a whole is far too vast for our complete description of it, so there will always be room for the novel, for deviation from prediction and explanation. Following the argument traced in Wittgenstein and the Private Language Argument, the criteria we use in evaluating reference and meaning of the ontology of science is still public and pragmatic, in the open, and thus “ultimate reality” too cannot be “hidden” at all. «The image and the reality are in one space.»9 What does this say about science? What does this say about the historical progression of “hidden reality” from what was “underlying”, “grounding” or “beneath” atoms, then electrons and nuclei, then quarks, fields and now possibly strings or loops? Do we not have incontrovertible proof of the existence of at least some of these fundamental elements? There is no hidden, no beneath, this presupposes the separation. It is all of one kind, and quarks and fields are equally constructed out of the same contextual frameworks as everything else. Their reality cannot be transposed outside the conceptual web in which they are constructed, for their very “existence” cannot survive the translation for lack of this concept of “existence” outside the context in which it gets its meaning. This does not in any way say that physics or science is unimportant, but it does constrain what expectations we can coherently have as to what physics and science can and should answer. The atom, electron, and quark all exist, but not outside the context and background of a conceptual web and an understanding of “exists” relative to this web. Evidence for this dependent relationship has become increasingly clear with results from quantum theory, which we will return to in more detail in another article. The dependence of our world on the epistemic is so deep and vast that it has become transparent to us, the depth of our immersion in the conceptual has convinced us that the conceptual really is an independent world, and we have near lost the ability to resurface from the depths.
Given this constructivist stance, how can we explain the success of science, or understand a claim such as “Science uncovers the hidden in reality”? We can only understand this with “uncovering the hidden” as meaning: the interdependent construction of new entities through concept and instrument creation. Instead of construction we could equally well say manufacture. By doing science, we make what was previously inaccessible, due to the limits of perception and instrumental extension, accessible. This is different from “hidden”, we must understand the task of science as “making available” rather than “uncovering”. We talk as if these constructions had independent existence, but in a similar fashion do we talk about subjective things that by definition are dependent. It is simple to now assume the ontic, but the only path to “it” is constructive and epistemic. And none of this is to say that what we are doing through science isn’t measuring or observing reality. It is to say that it would be at least partly wrong to call this a process of “discovering nature and her secrets”, for this indicates a hiddenness and detachment that isn’t there, and says nothing about the irreducibly participatory relationship, creativity and constructive effort that is integral to all discovery. It is to say that what is measured and observed is dependent on how we measure and observe, and what framework, context, the measurements and observations take place in. I think Feyerabend didn’t go far enough: it is not only scientists, but every single being that are sculptors of reality. In a special sense, yes, but exactly in that sense that we should realize brings some participatory magic back to the cold and barren universe of the particularist.
Both the realist and constructivist narrative “fit the data”, and they, as well as a multitude of both intermediate and completely different accounts, co-exist in our reality. This plurality of world views naturally leads to a discussion of incommensurability, the idea that parts of different world views or theories might be untranslatable into each other. Aspects of the constructivist account should further remind us about epistemisation, the process by which reality is only ever accessible to us as the epistemic. We will follow up on incommensurability and epistemisation in the next article.
Thank you for reading! If you enjoyed this or any of my other essays, consider subscribing, sharing, leaving a like or a comment. This support is an essential and motivating factor for the continuance of the project.
References
Barrow, J. D., & Tipler, F. J. (1996). The Anthropic Cosmological Principle. Oxford University Press.
Braver, L. (2007). A Thing of This World: A History of Continental Anti-Realism. Northwestern University Press.
Button, T. (2013). The Limits of Realism. OUP Oxford.
Chalmers, D. J. (1997). The Conscious Mind: In Search of a Fundamental Theory. OUP USA.
Davidson, D. (1980). Essays on Actions and Events. Clarendon Press.
Feyerabend, P. K. (1981). Explanation, Reduction and Empiricism in Philosophical Papers Vol. 1: Realism, Rationalism and Scientific Method. Cambridge University Press.
Feyerabend, P. (1993). Against Method. Verso. [1975]
Feyerabend, P. K. (2001). Conquest of Abundance: A Tale of Abstraction Versus the Richness of Being (Ed. B. Terpstra). University of Chicago Press.
Hanson, N. R. (1958). Patterns of Discovery. Cambridge University Press.
Kant, I. (2008). Critique of Pure Reason (Penguin Modern Classics) (M. Weigelt, Ed.; M. Muller & M. Weigelt, Trans.). Penguin Publishing Group. [1781]
Kim, J. (2010). Essays in the Metaphysics of Mind. OUP Oxford.
Kuhn, T. S. (1996). The structure of scientific revolutions. University of Chicago Press. [1962]
Kuhn, T. S. (2022). The Last Writings of Thomas S. Kuhn: Incommensurability in Science (Ed. Bojana Mladenovic). University of Chicago Press.
Pickering, A. (1984). Constructing Quarks: A Sociological History of Particle Physics. Edinburgh University Press.
Polanyi, M. (1998). Personal Knowledge: Towards a Post-critical Philosophy. Routledge. [1958]
Putnam, H. (1981). Reason, Truth, and History. Cambridge University Press.
Quine, W. V. O. (1980). Two Dogmas of Empiricism. In From a Logical Point of View: Nine Logico-Philosophical Essays. Harvard University Press. [1953]
van Fraassen, B. C. (1980). The Scientific Image. Clarendon Press.
van Fraassen, B. C. (2002). The Empirical Stance. Yale University Press.
Wittgenstein, L. (1980). Philosophical Remarks (R. Rhees, Ed.). University of Chicago Press. [1964]
See e.g. Feyerabend (1981), Kuhn (1996), Hanson (1958), Quine (1980).
van Fraassen (2002).
Some inspirations for the discussion in this section are Kim (2010), Chalmers (1997), Davidson (1980), Polanyi (1998).
See e.g. Chapter 3 of Barrow (1996).
Feyerabend (1993) p. 17: «Creation of a thing, and creation plus full understanding of a correct idea of the thing, are very often parts of one and the same indivisible process and cannot be separated without bringing the process to a stop.»
See e.g. Button (2013), Braver (2007), Putnam (1981) and the editor’s introduction to Kuhn (2022).
Inspired in part by the constructive empiricism of van Fraassen (1980, 2002).
See Pickering (1984).
Wittgenstein (1980) §IV.38.