Chapter 4: Some Points from Epistemology and Philosophy of Science To Help Us Think
4.6 A deeper dive into philosophy of science
Science itself isn’t pre-determined to be what it is – it could be done in other ways, and there have been plenty of debates about the trajectory of science over time. These varying trajectories suggest that there are many competing worldviews that have shaped science, and indeed religious worldviews have accompanied science (and scientists) all along. After all, some of the most important scientists have been religious people (such as Isaac Newton or Gregor Mendel), and many prominent scientists today are also people of faith (such as Francis Collins, director of the Human Genome Project, who is an evangelical Christian). At the same time, religious authority has been roundly opposed by science at times, and of course many important and contemporary scientists are not religious (and some are firmly anti-religious). We have already noted the “scientific imperialism” that has framed our era, and the influence of scientistic thinking continues to convince some that science is the only legitimate path to knowledge. If we pull back the veil of some of these views, it’s easier to reduce some of the confusion between science and religion that tends to fan the flames of conflict and obstructs complementarity. If so, one implication is that “conflict” is an unfortunate and unnecessary category for the relation between science and religion (it really only makes sense when people either try to use science like a religion, as in scientism, or when people try to use religion like a science, as in some varieties of creation science); I think some amount of independence, dialogue, and integration all seem more reasonable.
For instance, we see examples of independence and dialogue and integration all happening at universities – different fields delve into different realms of ideas and understanding, and often don’t talk much with each other, and yet more and more there is interest in multi-, inter- and trans-disciplinary dialogue.  In addition to these possible ways of seeing how religion and science relate to each other, Barbour also comments on the more specific question of how science is related to reality.  Barbour lays out four options that have held sway in the past 100 years: positivism, instrumentalism, idealism, and realism. Barbour asserts that most scientists hold to some form of realism; Barbour himself favors critical realism. Another way to understand these categories of understanding science is as different understandings of what “theories” are:
Theories as summaries of data: Positivism
Positivism is a philosophy of science that rejects metaphysics and theism; it holds that every rationally justifiable assertion can be scientifically verified or is capable of logical or mathematical proof. This view includes logical positivism and the verification principle (see further discussion below). However, positivism doesn’t adequately represent the crucial role that concepts and theories play in the history of science. Since “the real” appears to be more than just “the observable,” this philosophy seems lacking. Despite the fact that positivism doesn’t seem very likely to be correct and has been largely discredited, the attitudes, traditions, and habits of positivist assumptions still hold more sway that we’d think they should for a notion that doesn’t match reality. (The idea that “objectivity” is an end-all/be-all criteria for all science and knowledge is one such tradition based on positivism.)
Theories as useful tools: Instrumentalism
In contrast to positivism, instrumentalism recognizes that the knower does more than record and organize data. She abstracts, idealizes, constructs, invents, etc. Here, theories are tools to make accurate predictions, organize information to guide further research and achieve technical control (whether results are true or false is secondary). We recognize that laws and theories are invented, not discovered. However, Barbour reminds us that it seems that valid concepts are true as well as useful, so instrumentalism doesn’t seem a perfect match with reality either.
Theories as mental structures: Idealism
This notion posits that the structures of theory are entirely imposed by the mind on the chaos of sense data. As such, theories are subjective; the activity of the mind imposes structure on uninterpreted data – sense-data have no structure otherwise, according to this view. Thus, reality changes as discovery proceeds! However, realists say: our ideas change, but the world/physical reality doesn’t, and concepts represent the structure of events in the world.
Theories as representations of the world: Realism
In realism, being is prior to knowing. For many, intelligibility rather than observability is the hallmark of the real. This view is critical of both positivism and idealism, picking up on the critiques noted above. Barbour claims that scientists usually assume realism in their work. Naive realism overlooks the role of our ideas (it looks a bit like positivism: if you assume realism is “true” and then assume that humans can discern a 1 to 1 relationship between the real and our theories about it, then such a naive view will hold some of the same assumptions as positivism). Critical realism – Barbour’s preferred view – acknowledges both creativity of human thinking and the existence of patterns and events not created by human minds. It posits that no theory is an exact description of the world, but yet the world is such as to bear interpretation in some ways and not in others.
If good science aims to understand nature, we can unify the concern for empirical testing of positivism with the concern for intellectual coherence of idealism, while avoiding the excessive preoccupations of either. Science requires both logical processes and a creative imagination transcending logic. For Barbour, this describes critical realism.
A few additional points emerge from the concepts discussed above. The first regards the verification principle: this was part of the suite of beliefs of the positivists, and it states that only what can be verified (repeated and predicted) is real or true or knowable.  But if only the verifiable is real or has meaning, this is a contradiction, because we can’t verify the verification principle itself. That it is true is not empirically verifiable – it appears to be a tautology.  The problem can be phrased like this, too: a statement is only legitimate if there is some way to determine whether the statement is true or false (or what the answer to the question is), but the standards of discerning true and false are not verifiable. Indeed, if we look at how much non-verifiable information we base our lives on, the verification principle seems less compelling already – it’s as though our empirical experience in this world, ironically, points towards even us empiricists relying on non-empirical knowledge to know what we know about life and empiricism itself! This is like Jodie Foster’s character in the movie Cosmos;  she is a rationalistic scientist, who realizes that the thing she knows with the most certainty in her life is something that she can’t prove scientifically, namely that her deceased father loved her. A related point is that all knowledge, if you question it’s foundations far enough, is revealed to rely on unverifiable presuppositions. 
Another point that needs attention, or is likely to have arisen from the preceding discussions, has to do with “objectivity.” One hangover from positivism, somewhat akin to the verification principle, is a bias toward thinking that only “objectivity” is trustworthy or can provide knowledge or a basis for science, but the above categories make clear that some of the more “subjective” functions of knowing and thinking do indeed play a role in good science, always have, and always will.  Barbour addresses this by redefining “objectivity” as “intersubjective testability with commitment to universality.” By this he means that if we look at how science actually proceeds to create and confirm knowledge, we see that any particular field will seek the expert judgement (through peer review) of its own practitioners to test any new ideas or conclusions against what is already known and trusted as reliable information. Here “universality” suggests a focus on aspects of experience that are universal and applicable to all. The result is a perspective on scientific knowledge that isn’t merely “subjective” and thus somehow unreliable, but rather an acknowledgement of the elements of science that draw on subjective capacities in an intersubjective way (such that any idea will have to pass the scrutiny of a consensus of experts in order to be seen as reasonable).
Laws and Theories
Barbour further comments on laws and theories in science to describe the differences between the two. Essentially, we think of laws as descriptions of observed phenomena, and of theories as explanations of phenomena (laws don’t explain why a phenomenon exists or what causes it). It is tempting to think of theories and laws as existing on a spectrum, where the hunches and ideas of scientists (based on careful study) eventually start to gain coherence as a theory that finally is verified in terms of a law of nature, thus completing the quest for certain knowledge about some part of the world. In some cases, this sequence of development might hold true – we might craft a theory based on limited observations to explain what we’re observing in the behavior of gasses, and then, with time and more and more accurate observations, derive a universal law of physics. But it is important to remember that not only do theories not necessarily lead to laws — we might instead first discern a law through observation before we can figure out a theory to explain why that law is true — but even if some theories lead to laws (as sometimes happens in physics, chemistry, or other physical sciences), it is another thing to think that all of reality works this way – thinking that it does would assume a positivist (or naive realist) view of the world that doesn’t seem to merit such trust. It would assume that all of reality is governed by a sort of physical/material Newtonian/causal set of formulas. 
Rather, it seems more the case that laws and theories work at different but complementary levels, and though intuitions and hunches may play a role in the formation of theories, by the time something is declared a law, it’s been thoroughly tested for repeatability and “always-trueness.” With things that are purely physical/Newtonian in character, formulation of a “law” may be the goal all along. But there are also many types of things where we probably shouldn’t ever expect a “law” to emerge or for “law-like” relationships to be the norm.  Many aspects of life are messy, chaotic, chancy. For instance, imagine that we think we could reduce human relationships, say marriages, to a set of laws. That would suggest that we could predict the outcomes of marriages and relationships, and also that we could expect to achieve happiness in relationships by simply following the formulas allowed by whatever laws of marriage are true. This seems either highly unlikely, or like an achievement that would require either extreme restrictions of human freedom or a great deal of brainwashing. And yet we’re not about to therefore say that relationships and love are not real or are insignificant or irrelevant, just because we can’t derive “always true and predictable” laws about these things. So it behooves us to attend to what is a theory, what is a law, and why they might matter.
That said, laws are usually observed or deduced – they have to do with directly observable relationships. Laws can be deduced from theories (at which point we test empirically to see if that law holds true in reality). Laws are about correlations between two or more concepts that are closely related to observables. These are less complex mental constructs – once we figure out the correlation/pattern/relation, it seems pretty straightforward to see and describe what is going on. Meanwhile, Barbour says that theories are usually deduced, intuited, or interpreted. They are mental constructs. They are unified and generalizable conceptual schemes from which laws can be derived; as such, theories are things that can’t so much be observed (at least not yet).
For example, you can derive laws from theories: from Newton’s theory of gravitation, you can deduce Kepler’s laws about the elliptical orbit of planets around the sun. However, it can go in the other direction as well – to account for Boyle’s Law, which had been observed and figured out in the 1600’s, Kinetic Theory was later developed (and Kinetic Theory accounted for other laws, too, and led to new discoveries).
So we’d say that (ala Bacon, Hume, Mill) the inductive ideal is the example of coming up with a law – “generalizing from particular experiences/data to universal patterns.” We take observables, figure out the general pattern, and viola, we are headed towards the discernment of a law. The deductive ideal, on the other hand, is when we derive verifiable observation statements from general theories – it’s like taking Newton’s theory of gravitation to deduce Kepler’s laws (if those theories are true, our calculations show that there must be another planet out there, and voila! There is!). We can arrive at laws by deduction from a theory, but we can also discover laws by inductive processes working with the data/observations that turn out to be governed/predicted by a law. Note, however, that creative imagination plays a key role in inspiring theories as well as in deducing laws from observables.
In a sense, we might think of it this way: you figure out a theory based on what you can measure/observe and on the ideas you have about what might explain those observations, but theories can be about things that are still too complex, or maybe not repeatable enough, to deduce a law about them. But, you can also figure out a law based on what you can measure/observe and on the ideas you have about how those things are related, which will be influenced by your understanding of what explains those observables. So, good theories make it easier to sort out complex data to see repeatable patterns that might be reduced to laws (of motion, say, or behavior). 
Understanding the relations between laws and theories in science is challenging, and it invokes both historical and philosophical points. And yet all of this conceptual work still falls within the limited bounds of what and how we think we know things in science. For those who believe (as in scientism) that scientific laws and theories are all that is, it will be hard to see these concepts as complementary or in dialogue with religion, except as a means to exclude religion as a legitimate mode of knowing. But putting all of these concepts within the framework of Barbour’s critical realism, and with modes of dialogue and integration in mind as alternatives to the conflict model required by the faith of scientism, we will be able to open up a much deeper and far reaching conversation.
- Note that these three correlate to some extent with the independence, dialogue, and integration categories from Barbour. ↵
- Ian Barbour, Issues in Science and Religion, Harper and Row, 1971 ↵
- Barbour notes that Hume and the empiricists said that observation is the only way we can acquire knowledge. ↵
- A tautology involves the needless repetition of an idea, especially in words other than those of the immediate context, without imparting additional force or clearness, as in “widow woman”; it is a compound propositional form all of whose instances are true, as “A or not A” (http://www.dictionary.com/browse/tautology), thus not actually meaningful. For a verificationist, non-tautologous statements are meaningful only if it is – in principle – possible to establish empirically whether they are true or false. But if your standard is the observable, then we would need to be able to step back and watch a universe where people act only on what they observe versus a universe where people may know other things that aren’t verifiable, and I guess then we’d have to judge that the empirical-only universe is somehow “better” or more “correct.” However, the standard we’d have to use to make that judgement itself would then be an opinion, not verifiable as such, so we would need to appeal to something that can’t be verified, which would seem to argue against the verification principle. ↵
- Based on Carl Sagan’s novel of the same title. ↵
- For one who believes that there can be no reality that isn’t “verifiable”, this claim could be disorienting or might seem impossible. It is akin to the story of the foundations of the earth resting upon the shell of a giant turtle: one then asks, “Well, what is the turtle resting on?” The answer is “another turtle.” The obvious next question becomes: “Well, what is that turtle resting on?” The answer again is “another turtle” and so on, until the questioner finally says “Don’t you see? This is absurd! That makes no sense! You just keep invoking another turtle!” which occasions the final reply: “Don’t you see? It’s turtles all the way.” This is the philosophical problem of infinite regress, here exploited to question “certainty.” ↵
- The intuitive leaps, eureka moments, and hunches of scientists might be examples. The mode of knowing that Baer described as intellectus fits in this space. ↵
- Prior to the elucidation/clarification of quantum theory, philosophers were left to debate this, but physicists themselves have provided contrast to such a Newtonian worldview with quantum physics. ↵
- The relative lack of “laws” in the social sciences compared to the natural sciences might be an indication of the limits of assuming that physical/causal “laws” govern all of reality. ↵
- Likewise, bad or inaccurate theories based on incomplete data can delay our elucidation of a law. ↵