Science is a derivative of the more general method of induction, characterized by its specific approach and methodology. As such it is subject to all the rules and considerations pertaining to inductive reasoning in general.
As inductive reasoning results in certainty, so too does science.
Of particular note is the contextual certainty achievable by induction. As inductive reasoning results in certainty, so too does science. For example, the existence of atoms is not some kind of probabilistic estimate that might be overturned by new findings tomorrow, but a proven, certain, objective fact.
The definition and methodology of science derive from and reveal its philosophical basis: that reality (the subject of science) exists, that we can know it (the purpose of science), and we know it by objective means (the methodology of science). The success of science is (and can only be) a consequence of its ability to achieve truth, which itself is a consequence and demonstration of the truth of these philosophical underpinnings.
Yet modern philosophy is dominated by a pretense that we can know nothing
Yet modern philosophy is dominated by a pretense that we can know nothing, reveling in doubt as its alternative to understanding and explanation, and one symptom is the state of the philosophy of science. While science has been progressing at a staggering rate—from the Wright brothers to supersonic jets, from rockets to men on the moon, from the discovery of DNA to sequencing the human genome, each in a matter of decades—mainstream philosophies of science see only doubt. Rather than explain the success of science, they attempt to undermine it with claims that nothing is certain and that all science is tentative, if not imaginary. We now examine such claims.
The most popularized and widely known philosophies of science are those of Karl Popper and Thomas Kuhn. We will briefly analyze these in the forms in which they are generally understood and promulgated.
The essence of Popper’s philosophy of science is that a valid scientific hypothesis or theory is necessarily falsifiable but inherently unverifiable. This is because if theory A makes a prediction B (If A then B), then the rules of deductive reasoning mean that if B is false, A must be false; while in contrast B being true does not prove A (a true conclusion does not imply true assumptions). Popperianism concludes that science can never be certain of anything, and a valid scientific theory is one that is open to disproof—not one that has been proven, because that is impossible. The potential for disproof is the best we can do to link our theories to reality, leading to the strangely inverted notion that what defines a good scientific theory is the capacity to be false.
The essence of Kuhn’s philosophy of science is that science consists of a series of paradigm shifts. That is, there is a progression of accepted theories (paradigms), each of which has its heyday before facing increasing data to the contrary, finally crumbling—generally when its adherents leave the field by death or retirement. It is then replaced by a new paradigm fitting the new data—until it too is superseded in the same manner. For example, the geocentric model of the solar system was replaced by the heliocentric model, which was refined to a heliocentric system with elliptical orbits; and the ‘static’ paradigm of the cosmos was replaced by the ‘steady state’ paradigm of an eternally expanding universe, which was replaced by the ‘big bang’ paradigm, which will be replaced by something else.
Both the Popperian and Kuhnian philosophies of science suffer from internal contradictions.
Thus Popperianism recognizes that science must be tied to reality, but starting from the primacy of deduction, and finding no certainty of proof there, it consoles itself with the certainty of disproof: exalting falsifiability above knowledge and sacrificing truth on the altar of doubt. Meanwhile, Kuhnianism sets science adrift in a sea of ignorance, with ‘truth’ an elusive and illusory construct that fits the facts of today but not the facts of tomorrow: just a fashion of the day , depending more on the subjective bias of scientists than on objective facts.
In contrast, a philosophy of science based on an objective epistemology—itself a consequence of the manifest hardness of reality (see Philosophical Reflections 1–3)—starts from the primacy of induction and ends in the achievement of certainty.
Induction and contextual certainty were detailed in Philosophical Reflections 24-27, and we don’t need to repeat the proofs in detail here. But basically, because things in reality have a nature that affects the things around them, and different things in reality share all or some aspects of their natures: all we know and can know comes ultimately from inductive reasoning, and within the context of our observations inductive reasoning gives us certain knowledge of the nature of things. That is, while we don’t know everything, what we know, we know—and the rules of inductive reasoning enable us to know the limits of our knowledge, so that we can know that we know it and why we know it.
Robotic probes do not escape Earth and investigate Jupiter by luck, and smallpox was not eradicated by chance or wishful thinking. In the face of such eloquent proofs of the efficacy of science, how do philosophies of doubt remain standing? Like all good lies, both the Popperian and Kuhnian philosophies of science are partly true. But what is true about them is true not because their premises or arguments are true, but as consequences of the wider principles of induction.
A theory that makes no predictions about what will happen is cut off from the realm of evidence and therefore is essentially arbitrary, hence without value. Conversely, any theory that does make predictions is by definition testable. If such a theory is correct, it will pass the tests; if not, eventually it will fail. Thus the primary feature of valid scientific theories is testability, of which falsifiability of incorrect ideas is a corollary not the primary. And for the reasons explained when discussing inductive reasoning, the result is not perpetual uncertainty but increasing certainty.
For example, ordinary matter is composed of atoms, which do consist of small, dense positively charged nuclei surrounded by electrons in orbital ‘clouds’: clouds whose structure determines the chemical properties of atoms, which in turn determine the properties of matter. As science progresses, we will learn more about this and why it is so: but the knowledge we have now is certain. We will find out more about the fine details of what atoms are, and why atomic nuclei are what they are and electrons do what they do: but we will never find that ordinary atoms don’t after all have that structure.
Similarly, the history of atomic theory could indeed be described as a series of ‘paradigm shifts’—from pre-atomic theories, to atoms as little balls, to atoms as little puddings of positive and negative charges, to atoms as tiny central nuclei with orbiting electrons, to the modern understanding of electron ‘shells’ and quantum mechanics. But these are not random jumps through a sea of ignorance, but the proof of the existence of atoms at the expense of other ideas, then proof that they contained positive and negative components, then further refinements in the light of observed reality toward a theory of increasing detail, accuracy, and precision. Such a process, in basic form, is echoed in all branches of science, and is a consequence of the fact that we learn about reality by observing it, generating hypotheses, and testing them.
As a result, we have progressed from guesses and speculations about the fundamental nature of matter, to the ability to see and move individual atoms (e.g. with atomic force microscopes), and even to create designer atoms to improve our lives (such as certain medical isotopes).
Regarding the former, as all knowledge of the world ultimately must come from induction (reason applied to sensory evidence), the very data that is used to ‘falsify’ a theory is itself derived from induction. So if one claims that induction is never certain, no theory can be disproved either! Falsifiability is only possible as the flip side of inductive certainty: thus Popperianism in fact rests on an assumption of inductive certainty, which it then evades.
The crucial question on paradigm shifts is why, in fact, do they shift? The answer is: because while earlier observations were consistent with the theory, further ones were not. But then the next paradigm must account for both the former set of observations and the latter. And so on. Thus, each paradigm is a better explanation of more facts of reality than the previous—which is a finite progression. Eventually we must settle on a complete and confirmed paradigm, or at worst, one in which the areas of ignorance are clearly defined.
The less we know about something, the wilder our guesses to explain it and the bigger the ‘paradigm shifts’ that result. Examples of that are what give Kuhnianism a veneer of plausibility. But by the nature of facts and truth, no fact can ever falsify a true theory: which can therefore never be caused to ‘shift’. Eternally shifting paradigms are only possible if there are no true theories, which is only possible if there are no facts of reality to describe—but there are. Which is why science works, and works so well.
Scientific knowledge is dynamic, and the changing landscape of facts and theories mirrors the valid aspects of falsification and ‘paradigm shifting’. It is worth a closer look at the elements that comprise scientific knowledge, to illustrate the boundary between knowledge and doubt, and to dispel misunderstandings that have been used to attack science. For example, creationists often say that “evolution is just a theory,” as if theories are some kind of inferior, dubious alternative to facts. Such an attitude to theories is common, but it is a critical misunderstanding of science, of which theories are a crucial component (see Philosophical Reflections 30, on the nature and process of science).
Scientific statements can be divided into facts, laws, theories, and hypotheses.
A fact is some proven item of knowledge, such as “water is a compound of hydrogen and oxygen” or “Earth is a planet orbiting the Sun.”
A law is a general descriptive principle that applies to all relevant existents. Thus a law describes a fundamental quality of a broad range of things. For example, the law of universal gravitation applies to all mass, and the laws of thermodynamics apply to all systems of matter and energy.
A theory is a causal description that explains facts. For example, the atomic theory explains the chemical qualities of elements and compounds, and the theory of evolution explains the origin of the multitude of kinds of living things. A theory is not some weaker second cousin of a fact, but an explanation of facts.
A hypothesis is a proposed explanation of facts, essentially a provisional theory. A ‘good’ hypothesis is one that explains some facts, does not absolutely contradict any other facts (though it might not be able to explain everything), and makes testable predictions. Hypotheses abound at the interface between what we know and what we don’t know, like expendable scouts probing a wilderness.
It is plain from the above that whereas hypotheses are never facts (though they might become facts), both laws and theories can also be facts, indeed wider-ranging and more fundamental facts than the specific facts they encompass. For example, the atomic theory is also a fact, as it is a fact that elements and compounds are made of atoms. In contrast, hypotheses by their nature never have enough evidence to be called ‘true’: but the generation of hypotheses is how theoretical science progresses. Many hypotheses fall by the wayside, but by spotlighting where further observations should be made, even the wrong ones can contribute to the growth of knowledge. It is out of the melting pot of facts, hypotheses, and experiment that validated laws and theories are refined.
Thus we can see how science grows by objectively looking at reality, and as it grows, we see ignorance give way to knowledge by the primary process of induction—which entails the secondary effects of falsification and paradigm shifts of wrong or incomplete ideas. In a way, science is like a city growing in a wilderness. The edges are in great flux as knowledge extends in fits and starts into the unknown, while the established center grows in height and maturity of understanding, filling in details and rebuilding where required.
Einstein once said in a similar vein:
“Creating a new theory is not like destroying an old barn and erecting a skyscraper in its place. It is rather like climbing a mountain, gaining new and wider views, discovering unexpected connections between our starting point and its rich environment. But the point from which we started out still exists and can be seen, although it appears smaller and forms a tiny part of our broad view gained by the mastery of the obstacles on our adventurous way up.”
We have looked at philosophies of science that focus on uncertainty and seen how both certainty and uncertainty fit into the scientific process. Now we turn to more severe criticisms of science, which claim the scientific process is fundamentally flawed.
‘Holists’ argue that science is too ‘reductionist’, while true understanding requires a holistic approach that looks at the whole instead of the parts. Consider this excerpt from Technological or Media Determinism:
“Reductionism contrasts with ‘holism’, which is broadly concerned with the whole phenomenon and with complex interactions within it rather than with the study of isolated parts. In holistic interpretations there are no single, independent causes. Holistic interpretation proceeds from the whole and relationships are presented as non-directional or non-linear. It is holistic to assert that the whole is more than the sum of its parts, a proposition with which it is difficult to disagree when you think of a working motor compared with the stacked parts.”
Such views represent both a misrepresentation of science and a flawed view of cause and effect.
In terms of what science actually is, firstly, many scientific disciplines are explicitly interested in whole systems (e.g. ecology and even physiology). Secondly, there is no logical equivalence of ‘reductionism’ with ‘linear or one-way’ effects. Attempts to understand individual causes do not imply that those causes have linear effects, and most definitely do not imply an absence of multiple interacting effects or feedback (an effect directly or indirectly affecting its own cause). Indeed, both are well known in science, as in the regulation of cell growth and even such commonplace phenomena as temperature regulation in our own bodies.
While it might be ‘holistic’ to assert that the whole is more than the sum of the parts, that is not unique to holism: it is where holism and science overlap. The question is not whether the whole is greater than the sum of the parts, but whether the whole can be understood by studying the parts, or without studying the parts. Ironically, the holist’s own example, a working motor, proves the former: for the working of a motor can be understood completely by reductionist science—motors having been invented using the principles of such science in the first place. One wonders exactly how a holist could achieve the same end.
Even when the whole is more than the sum of its parts, it remains caused by simpler things. Nothing is causeless, and no whole can cause itself nor can it be more than the sum of its parts and their interactions. Unless you tease out specific causes, you can in fact learn nothing about causality, hence, you cannot explain what you are studying. Hence the attraction ‘holism’ holds for mystics, who replace objective understanding with subjective, arbitrary assertions.
It is only through reductionism that the whole can be understood. Isolating simple causes, then finding out where they are insufficient to understand the whole, is how one ultimately discovers how the various parts interact individually and as systems in order to generate the whole. One finds the easiest causes first, then the more subtle or complex ones, until complete understanding is achieved.
Science bases its pursuit of and claim to truth on objective enquiry. Denials of the possibility of objectivity therefore attack science at its root.
Such denial lies at the base of diverse but related criticisms of science, from post-modernism’s “everything, including science, is just a text,” to multiculturalist claims that science is a subjective social construct no more valid than the beliefs of non-scientific cultures, to radical feminist theories of science as specifically ‘male’.
The basic assumption behind such arguments is that science cannot be objective because it is a human activity. That is, objectivity is impossible to human beings, therefore science’s claim to it is false: all human discourse is subjective and culturally (or racially, or sexually) defined.
Such collective–subjectivist claims are in essence racist or sexist, ignoring the many excellent scientists of both sexes from all over the world. But that is not their worst feature. There is no honest attempt to avoid objectivity (as any honest attempt can result only in silence): only an attempt to disparage evidence in order to enthrone arbitrary claims in its place. A person who has evidence does not seek to deride objectivity, but to appeal to it. Only someone who knows their views cannot stand objective scrutiny attempts to drag the objective down to their own level. Thus the spectacle of such claims being made by people who wish to avoid the rigors of science being applied to their own beliefs, while insisting that you should believe what they are saying, despite their avowed subjectivity and lack of proof!
Arguments for the cultural subjectivity of science that are not outright attacks on objectivity as such amount to the obvious but irrelevant—“Scientists study what interests them, and what interests them depends on where they live.” What else should scientists study, than what interests them or those who pay them, and what should interest people, besides what is relevant to their lives? But the whole meaning and validation of objectivity is not that you don’t care or choose what you study, but that the answers you get depend on the nature of reality, not on your prior beliefs, prejudices, or desires. The many serendipitous discoveries in science (from nylon to conductive polymers, from penicillin to recombinant DNA), where an important new fact or principle has been discovered while studying something else, are a consequence and validation of this.
Indeed, whatever plausibility such claims might have is similar in nature to that of the ‘paradigm shift’ theory of science. That is, in the absence of good evidence, all sorts of theories are possible. A few poorly understood facts can support any number of wild ideas and, as a consequence, they can also be shoehorned to fit baseless prejudices. But in the long run, false theories will fail. The presence of good evidence (as determined by the rules of inductive reasoning) severely limits the possible theories one can validly propose to explain the facts.
Thus, while one could argue that anthropological ‘proofs’ of European superiority in the 19th Century were ‘culturally determined’ by the prejudices of Europeans—the evidence adduced was never sufficient for such claims, even at the time. On the other side, one cannot argue that modern aeroplanes do not fly, whatever prevailing prejudices existed before powered flight was made a reality.
The history of science is filled with debates, some vicious. Yet scientists are meant to be objective. Thus people often have a picture of the ‘ideal scientist’ as a person who is disinterested in the results of their experiments, has no emotional commitment to a particular theory, and will change theories to fit the best current evidence: and any deviation from this ideal is proof that science isn’t objective.
This is a misunderstanding of objectivity—a misunderstanding whose ultimate root lies in the ancient reason-emotion dichotomy. But there is no necessary dichotomy between reason and emotion or between facts and values. Rather, values should follow from facts, and emotions rooted in rational values serve reason. Yes, a passion that overrides or replaces reason is irrational: but passion flowing from rationally chosen values is not.
A scientist should care—about what is true, and what can be achieved by knowing the truth. For the purpose of science is to learn the truth, and the proper purpose of knowledge is the improvement of human life. It is literally, if long-range, a life or death activity. As witness the improvements to human life from modern agriculture, genetics, electronics, and medicine.
Scientists should be passionate about the truth, and such passion is more rather than less likely to make them strong defenders of their favored theories. This is a good thing. It is an important part of scientific progress that theories have partisans, who don’t accept contrary evidence at face value and are motivated to find weaknesses and alternative explanations.
However a passion for the truth must be just that: a passion for the truth. Thus any such passion has to be subservient to rationality and the rules of inductive reasoning, as our only means to know the truth. Science derives from an objective epistemology, and has no place for subjectivism or any other irrationality. Scientists must place nothing above the facts, not even their own or favorite theory. Thomas Huxley called it:
But it is not a tragedy, it is the virtue and strength of science.
Whatever the basis of fundamental criticisms of science, they come up against one stubborn fact of reality: science works. Indeed, science is so successful that its opponents all use its results daily even while attacking it. Whether you look at nutrition, medicine, agriculture, metallurgy, transport, computers, clothing, engineering, water supply or waste disposal, science is pervasive in its benefits. Science has mapped the moving of continents and the motion of galaxies, measured events lasting femtoseconds to billions of years, sequenced the genomes of many organisms including man, enabled machines of metal to fly hundreds of people around the world in a matter of hours, and sent spacecraft to other planets and beyond the solar system. And all that in the few hundred years that science has existed, out of the hundreds of millennia humans have walked the earth eking out a living by spear and plough.
The questions to ask any critic of science as a tool for discovering truth are simple: “How do you explain what it has achieved?”, and “Why do you use so many of its results in your own life?” There is no explanation, except that science does what it claims.
Individual scientists are individuals, and are not always rational or objective. But while such irrationality is bad for them and can slow the progress of science, fundamentally it doesn’t matter for the integrity of science itself, just as the fact that some scientists commit fraud doesn’t matter in the long run. As discussed in Philosophical Reflections 30, science is a process designed for objective enquiry, designed to catch and correct errors and, whatever the errors that might happen along the way, the truth will eventually out. This is another example of where the principles underlying paradigm shifts apply: conflicting opinions, whether justified or unjustified, will eventually triumph or fail according to, and only according to, the evidence found in reality.
Because science is based on objectivity, it should never be treated as dogma: and thus a scientist should never expect to be taken on faith, but rather to prove what they say by reference to facts. Conversely, their audience needs to be able to evaluate that evidence. That topic is discussed in another essay, When Science Meets Philosophy.
This article is modified from an essay first published in TableAus, the journal of Australian Mensa (2005).