‘A Picture Held us Captive’: Science as the triumph of Truth over ‘bias’ in all its forms
This is the third instalment in the Philosophy of science series. A modified version appeared here in Convivium, a publication of the Cardus think-tank, where I am currently a fellow in the NextGEN program.
In the links below, you can find the Series overview, parts 1 and 2 of the first instalment on Methodology and Subject matter, and the second instalment on Correlation and Causation.
Many scientists and lay people alike look to the fundamental fields of science as the paramount example of knowledge, and the sole reliable means of arriving at truth. Whether that belief is explicit or implicit, many would shutter at the idea that there are values at work in science.
As noted earlier, a regnant view of the very practice of science is that its success depends on the fact that its methods and assumptions effectively purge the observational realm, processes of inference, and testing of certain kinds of values.
By doing so, it is claimed, scientists can arrive at an unclouded understanding of the natural world.
Values in Science? Why yes, all the way down!
There are two broad types of values operative in scientific inquiry – the epistemic and the normative. Though, as we’ll see, they run into one another due to their common status as values. Both are essential components of scientific practice and theory.
Epistemic values are principles that orient scientific inquiry. They assist the inquirer in the goal of scientific endeavour – explanation of some sort. There are different kinds of explanations – rendering a data point or set more consistent with other findings, making a connection between related phenomena more logically rigorous, or providing better explanations by filling in gaps in the causal story, for example. All of the above involve an explanans (that which explains) and an explanandum (that which is to be explained). A scientific theory explains (the explanans) a set of phenomena (the explanandum).
This small list is already very revealing – scientific inquiry is an activity with a purpose. As such, the standards of success and the methods used to achieve them are determined, in no small part by the goal of the endeavour.
Here are a few examples to illustrate the point.
First, ontological simplicity – the idea that we should not multiply the entities that populate the object domain of study, or terms in a theoretical lexicon in order to explain some phenomenon. It is also known as Ockham’s Razor after the 14th century Franciscan Friar, William of Ockham. By adhering to this principle, the temptation to make up a concept in order to solve some problem is easier to avoid. It is also closely linked to testability – the more entities and assumptions that are made in theorizing, the less testable is the theory, for there are simply too many moving parts.
However, there is no clear, empirical reason why there should be fewer entities in one’s theoretical lexicon, or that a simpler explanation is the better one. In some sense, the principle is hygienic – it forces the theoretician or practitioner to be more rigorous about their practice than they might otherwise be. Very well then, but it is hardly a necessary feature of scientific inquiry.
Second, a preference for the falsifiability of a claim or theory. It is argued that if a claim is falsifiable, it marks it out as something with tangible implications. Without the ability to falsify a claim, or body of knowledge, it is said to be unworthy of scientific inquiry.
Though there are naturally limits to such a view, for not every piece of knowledge is falsifiable, even if it nonetheless knowledge. Furthermore, what standard of falsifiability must we adhere to in order for something to count as scientific? Such simple questions demonstrate the extent to which the preference for falsifiability is indeed a value choice, determined by the nature of the question we are attempting to answer.
Third, replicability. As we saw earlier, if we cannot replicate a finding, we do not often take it to give grounds for a general truth claim. If we do, it is only in a qualified sense – subject to further verification.
There is no reason to think that a finding must be replicable in order to be an accurate explanation. The idea that explanations should have this feature reveals another common assumption of the regnant philosophical understanding of science – that most basically, the universe is a set of fundamental laws that determine the particular manifestation of particles, and larger composite bodies. This is a philosophical assumption, not a scientific one.
Along with many other contemporary philosophers of science, Nancy Cartwright has argued that laws of nature are only true under artificial circumstances that often don’t obtain. Laws are derived through careful experimentation in which the environment is modified and controlled so as to isolate relevant factors, and derive a general conclusion. For example, Coulomb’s law of electrostatic attraction and repulsion describes with strict accuracy only when they aren’t being subject to gravitational effects. The same applies to Newton’s law of inertia and Kepler’s law of ellipses. Her view is that what’s really going on is that objects have powers and capacities that are active or dormant, and that this view better captures the imperfect regularity we see in the world.
The problem is that through this process of hypothesizing, measuring, correlating measurements, and then making a generalized statement about the relation between terms, much of what is causally relevant in the broader sense is bracketed out. The picture you get is one of mathematically described relations between entities, but only of a highly idealized ‘slice’ of reality.
Another value that we operate with is the preference for elegance. We prefer elegant theories that we can easily understand, or whose conclusions can be presented in such a way that they have a kind of appealing coherence. We wouldn’t very much like a theory that had so many unreconciled, extremely difficult pieces to grasp. We naturally try to synthesize those findings, connect and present them in a way that is aesthetically and cognitively appealing.
Lastly, predictive power. It is hard to imagine science being done to such an extent without the promise that it yields predictive power. If a finding did not yield any added powers of prediction, or the ability to make tools that do, we would not take the findings to be true. This is most especially the case today, when people routinely mistake something’s usefulness for a particular end, with its truth, goodness, or beauty. Indeed, the uncritical belief in the goodness of novel technological creations is very much one of the dominant spirits of the ‘progressive,’ technological age in which we live.
What then of the Value-Purge?
Evidently, values play a role in shaping the standards of inquiry, methods, and evaluation of the logic of a scientific conclusion, but must that be the case? Is there not a set of principles that others are reducible to?
If, for example, we take correspondence (to empirical observation), or consistency (with other known empirical facts) as our standard of truth, would it not follow that questions about method, explanation, and ontology have a clear, and determinate answer?
To return to the discussion in an earlier piece, it would seem that there are foundational claims that are evidently true, and so form the bedrock and the litmus test for theories to pass. These claims would also serve as the starting point for hypotheses, for it is from these most basic statements that more complex ones can be deduced and inferred.
New scientific findings would have to correspond to repeated empirical observation. They would also have to be consistent with all of our other established knowledge, most importantly the bedrock claims.
It is here where we can see that the problem is much deeper than this sketch would have it seem. Independently, a philosopher, Willard Van Orman Quine and a mathematician and physicist, Pierre Duhem, arrived at a thesis about the underdetermination of theory by evidence.
In its various forms, it states that the evidence for a given claim (a proposition, or hypothesis, for example), or a larger body of claims (theory), is always underdetermined; the evidence can never provide sufficient grounds to conclusively deduce a particular conclusion from it.
It could be that there are other theories that are equally consistent with the facts, but the evidence does not itself necessitate the choice of one explanation over another. Other, extraneous considerations enter into the equation in the process of selecting the explanation for the piece of evidence in question. This is a logical consequence of the relationship between evidence and a conclusion, as well as something that is empirically demonstrable.
First, the logical – there is nothing contained in an isolated piece of evidence that strictly entails a conclusion of some sort. We can see this by examining claims about the genetic vs. environmental determinants of behaviour, and the effects of isolated practices like education or media consumption habits on specific behavioural outcomes, such as aggressiveness or violent behaviour.
If it is demonstrated that a collection of genes is correlated with a pattern of behaviour, such as a personality trait, there is in principle no amount of data (by itself)that entails the conclusion that the genes cause the behaviour. This requires additional assumptions about the presence of other behaviours that affect the outcome, the path between individual genes and phenotypic expression in space and over periods of time, as well as non-empirical arguments about the understanding of what a cause is, and the contemporary taxonomy used in biology that cuts up the world in such a way that assigns privileged ontological and causal power to strands of DNA, for example.
To determine whether watching certain kinds of TV shows will make aggressive behaviour more likely, the same problems arise. Is it that more aggressive people watch certain kinds of shows more frequently, or that the shows cause more violent behaviour because of their content? How many other factors enter into the equation, etc. Though there are many ways to disentangle the correlations from causation, there is never enough for the evidence in a given case to strictly entail one type of explanation over another.
This is not to say that all explanations are equal – that is nonsense. It is about showing that additional considerations extraneous to empirical observation are needed to fill in the logical gaps from the evidence to local causal explanation, and larger explanatory frameworks.
Quine’s epistemological holism is informed by this view – the idea that no knowledge claim can be tested in isolation. We must always have recourse to a set of background assumptions – part of the web of understanding in his case – in order to test even the simplest claim.
I would argue that this does not go nearly far enough. As many others have already pointed out, most notably Kant, our perceptual faculties are deeply informed by our own biological nature, habits, and personal and cultural history. Before we are faced with a fork in the road in choosing between competing theoretical explanations for a set of data, we bring our values to bear on our world in perceiving and interacting with it.
For our purposes then, the important conclusion is that values play a necessary role in the evaluation of the simplest claims to knowledge, as well as more complex choices between theories, or explanatory frameworks.
All of these considerations go to show that much of scientific practice is non-empirical. There is no reason why any of the most commonly held assumptions about the standards and methods used in the sciences are required in order for something to be scientific, or true. These values are principles used to help us in the pursuit of an explanation of some sort.
Far from this being a licence to say things such things as atoms or particles are not real, genes do not causally influence behaviour, or everyone’s favourite, ‘reality is socially constructed’ – such humdrum observations show what any scientist worth there salt would readily admit – scientific findings are empirically-based descriptions of the patterns and regularities that we find in the world around us. They are not the be all and end all of explanation. They are local explanations of aspects of the world around us, that are provisional in nature.
However, the most common conclusion to be drawn from this sort of thinking in the 20th century Western tradition – and now culture – is some form of skepticism about knowledge of a variety of types – scientific, philosophical, ethical, aesthetic, etc.
Yet, this is in no way entailed by the above considerations. Rather, we ought to realize that scientific inquiry consists of a set of methods that we use to rigorously observe a narrow range of reality around us, and systematically explain the causal relations between entities in that object domain – an understanding of causation, albeit one that must remain necessarily opaque at some level.
Skepticism in all its many colours, stems from a misunderstanding of the activity of scientific inquiry. Science does not say anything about the ‘meaning of life’, the nature of causation, the origins of the universe, whether there is ‘free will’, etc., until its findings are combined with additional premises in an argument. Arguments, made as they are in human language, are strictly speaking, philosophical in nature.
This may sound absurd to some, or to be expected coming from someone with a background in philosophy, but it is really rather prosaic. Philosophy is a practice that everyone does, every day of one’s life. It is the practice of thinking about the meaning of things. Simply put, it is reflection.
Like science, reflection – systematic or freewheeling, rigorous or intuitive – does not have much to say in isolation about the chemical structure of water, the causes of poverty in 16th century Italy, or the physics of black holes. Why? Quite simply, because it is a different kind of activity, with a different area of inquiry, and a different set of questions that animate it.
Ends determine Means, and the Partiality of Perspective
Rather than science providing an unmediated access to reality, the scientific outlook affords a perspective. Not a trivial one in the slightest, not the only one, not a self-sufficient one, but a perspective nonetheless.
Some would argue that there is nothing but perspective, from such a view. But this is hardly warranted. Only the person who denies that they are making sense when they speak, or acting as the agent of their own thoughts when they think can utter that kind of nonsense, and actually take themselves seriously. Yet, this is done frequently, and is even the default position among many advocates of scientific naturalism.
Scientific reasoning can never prove the truth or falsity of its own assumptions (which are values), nor can it have much to say at all about normative questions, only indirectly. A scientific argument can be used to support a premise used in a philosophical argument about some conclusion, but it cannot constitute the argument.
Finally, I think the best science is rooted in the best values. Insofar as they claim their views and practices are straightforwardly scientific, Social Darwinists, eugenicists and abortionists, materialist psychiatrists and physicians are mistaken. Not because they are fraudulently engaged in the process of searching for causal explanations, explanatory frameworks that make sense of diverse sets of local phenomena, or appropriate medical treatments. Those theories and the practices that flow from them represent bad science and medicine because of the assumptions that underlie them, the ontology they assume, and the ends that they are in service of.
Addendum: Bridging the Gap between Facts and Values
Values are indeed objective, but they are known in a different manner than facts about mind-independent reality.
We can have objective knowledge of the properties of abstract, general concepts that go beyond that which we can experience through our senses – things that are universal, and refer to all possible objects of a certain kind. We do not ‘sense’ the abstract similarities in objects, but understand them with our intellect.
We can also judge whether a statement is true or false, and whether that truth is necessary – like whether an effect is preceded by a cause, whether everything that exists has a reason for doing so, and whether 3+7=10. Each are true in the present, past, and into the future – necessarily true.
All of these ways of knowing pertain to values and their application, as much as they do about questions of mathematics and empirical fact.
A value judgment involves the use of reason to assess the truth or falsity of claims about questions of goodness. A good is something that is proper to a thing, based on its nature – it is that which a thing strives for, and that which renders it complete. For example, hunger desires food and is satiated by it; a seed is completed by becoming a plant or tree; sperm fertilizes an egg, to become a living being; and inanimate objects have a constrained range of potential ways in which they can interact with other objects, better described as tendencies arising from their particular nature, rather than the general laws and forces that impact them as well.
Everything tends toward an end that completes it. Humans, in particular, have a range of ends, or goods that we seek. That which we desire through our senses, emotions, appetites, actions, will and intellect all find their fulfilment in properly aligning the means that we use in pursuing them with the the achievement of the end, which in turn affects the degree of quality in which the end is achieved.
As humans, we have many ends that flow from all of our different capacities, but our most important one is the capacity for reasoning. Rationality is our distinguishing feature – the ability to identify a goal, and select the means to achieve it by reasoning. Rationality organizes and circumscribes the proper use of all our other capacities through its ability to understand by way of the intellect, and to control and direct intentions, desires, and action by way of the will.
Because we have the ability to understand and act in accordance with it, we can use our other capacities in ways that are commensurate with the knowledge of what is good and right in a particular domain, and we can fail to do so. Through experience over time, we come to learn what things are good for us, and which aren’t, and to recognize which things are of higher value than others. By using our capacities well, we instil virtues that direct us reliably toward the goods we seek. The final natural end that we seek is happiness; achieving it reflects the harmonious ordering of all our lower capacities and ends that we seek out.
Key to this picture is the recognition that there are precepts that establish the proper use of one’s capacities. These precepts are built right into the use of our practical reason, when we think about how to act on, and use any of our capacities. They are revealed to us by their self-evidence, and are undeniable on pain of contradiction. Furthermore, experience confirms their correctness.
The first precept is that the good is always sought, and evil avoided.
To see why this is the case, take any action. Even if we do something we know to be bad for us upon reflection, such as insult someone in anger, make fun of another, procrastinate, overeat, or jump to a hasty conclusion, we do so under the appearance that it is good when we do it. Even if we reflect and say to ourselves one moment that this is not a good thing to do and the next moment we do it, in doing so, we are acting as if we are seeking it as a good that will fulfil some desire that we have. It is simply definitional of any kind of practical rational action that it is directed toward some good, either true or only apparent.
Other derivative precepts can be drawn from this more basic fact that we always seek out goods, based on our nature, and the particular natures of our various capacities. We have a natural inclination to pursue the various goods afforded by our sensory and rational faculties – the eye seeks sight, faculties of taste good tasting food, the intellect truth, and so on. In acting in accordance with any of these capacities and the goals that they seek, we recognize that it is good to seek them, and bad not to.
For example, the good of truth is found in the intellect, insofar as it apprehends a thing that is known; beauty is found in the cognitive faculty, when it apprehends the form of an object, which has proportion and integrity. By pain of contradiction, we recognize that the ends we seek are determined by their nature, our own, and the faculties we have, in relation to one another.
The proper use of any and all of our capacities is circumscribed by the relations between our various faculties and how they contribute to our happiness. Certain things are ruled out as directly contrary to our good, because in doing them, we cut ourselves off from that which we seek, our happiness. However, the specifics of how and when and why are not prescribed by precepts or rules in fine-grain detail. For, one of the things that makes any action good is taking into account all of the particulars of a situation, and using one’s intellect to understand, and will to act freely on that basis.
In summary, the objectivity of values and the norms that emerge as rules governing their pursuit are derived from a recognition of the fact that we are directed toward various ends. These ends are goods that we cannot help but seek because of the very nature of the faculty that pursues them and in turn the nature of the good sought, and the conformity between the two. In realizing that an end is sought, we recognize that that end also has a nature, which dictates the means through which it can be pursued.
However, the most common view today is that values are subjective, and facts objective. To the contrary, values are very much factual and objective, but not in the same way as facts about events, physiological properties, and natural processes.
Far from there being a divorce between scientific truth, and moral and aesthetic truth (goodness and beauty), an explanatory scientific theory that describes the nature of reality must be consistent with truth as it pertains to all three aspects of being.
I will explore the relation of these topics in greater detail in the final instalment on the unity of explanation.
This is the third instalment in the Philosophy of science series. In the links below, you can find the Series overview, parts 1 and 2 of the first Instalment on Methodology and Subject matter, and the second instalment on Correlation and Causation.
 Nancy Cartwright, Nature, the Artful Modeler: Lectures on Laws, Science, How Nature Arranges the World and How We Can Arrange It Better, Paul Carus Lecture Series 23 (Chicago: Open Court, 2019); Nancy Cartwright, The Dappled World: A Study of the Boundaries of Science (Cambridge, UK : New York, NY: Cambridge University Press, 1999).
2] See Christopher Baglow’s writing on faith, science and reason for an introduction to these topics. Examples taken from: Christopher T Baglow, Faith, Science, & Reason: Theology on the Cutting Edge, 2019, 242.
 First part of the Second part, Question 49 in St. Thomas Aquinas, Summa Theologica, Complete English ed (Westminster, Md: Christian Classics, 1981).