|
KNOWLEDGE, SCIENCE, AND RELATED
CONCEPTS
It would be desirable to try to define knowledge, science, and some other concepts closely related to them, such as truth, facts, laws, explanation, theory, hypothesis, and evidence. I. Truth and Knowledge First consider the concept of truth. Here is one possible definition: a proposition is true if and only if it corresponds to reality or "tells it like it is." That is, there timelessly obtains some actual state of affairs out there in the real world to which the proposition corresponds. It should be noted that propositions might be true even if no one believes them or has any way to test them. For example, consider the following two propositions: (a) Abraham Lincoln ate at least one egg on the morning of July 22, 1825. (b) Abraham Lincoln did not eat any eggs on the morning of July 22, 1825. It must be the case that one of these, (a) or (b), is true, yet neither of them is testable and probably neither of them is believed by anyone in the world today. Note also that neither proposition coheres with any systematic set of propositions and neither of them is useful for the solution of any problem. So any definition of truth that appeals to such features (such as the coherence theory or the pragmatist theory) could be refuted by these "Abe Lincoln" counter-examples. Let us now turn to the concept of knowledge. We could say that knowledge consists of all propositions known to be true. And a person X knows a proposition P to be true if and only if: (1) P is indeed true (as defined above), (2) X strongly believes P, and (3) X believes P on an adequate and non-erroneous basis. It is hard to say exactly what is needed for X's belief to have "an adequate and non-erroneous basis," but I think one requirement is that the basis of X's belief must be such that anyone who believes anything on that basis is widely recognized (at least among educated people) as being reasonable in his/her belief. However, it is not required that the belief itself be widely held or enjoy "a consensus of rational opinion." Scientists can come to know things long before there comes to be any consensus on the matter. One question that may be raised here is whether there can be such a thing as religious knowledge. The main issue is whether the basis of religious belief is "adequate and non-erroneous." Is it a basis such that whoever believes on that basis is widely recognized (at least among educated people) as being reasonable in his/her belief? My own view on this issue is a negative one, though the matter is complicated by the fact that different people base their religious beliefs on different sorts of consideration. Some appeal to "faith." Others appeal to revelation or religious authority. Others appeal to religious (or mystical) experience. And still others appeal to reason (which includes scientific method), though the application of reason to the beliefs in question may turn out to be inadequate and erroneous. Because of all this diversity, it is hard to deal with the "religious knowledge" issue in a general way. II. Science Let us consider now the concept of science. Here are five different definitions of the term: D1: Science is a branch of knowledge or study dealing with a body of facts or truths systematically arranged and showing the operation of general laws. [Random House Dictionary] D2: Science is a body of knowledge which consists of the following, coherently organized in a systematic way: (a) Statements which record and classify observations which are relevant for the solution of a problem in as accurate and definite a way as possible. (b) General statements - laws or hypotheses - which assert regularities among certain classes of observed or observable phenomena. (c) Theoretical statements which connect and account for the largest possible number of laws. (d) Other statements which are deducible from the initial descriptions and from laws and theories and which are confirmed by further observation and testing.[1] D3: Science is public knowledge the goal of which is a consensus of rational opinion over the widest possible field.[2] D4: Science is mankind's attempt to observe, understand, and explain the operation of the universe and its inhabitants.[3] D5: Science is any attempt by members of a community to establish a framework that makes their observations of nature intelligible.[4] Although D1 leaves open the possibility that science is not a branch of knowledge but is, rather, "a study," the next two definitions definitely make science into a type of knowledge. One reason for connecting science with knowledge is that the word is derived from the Latin word "scienta" (which means "knowledge"). Another reason is that science is a subject matter to be learned by students, and people look to science for information and understanding. On the other hand, there are also good reasons for denying that science is a type of knowledge. First, it contains theories and hypotheses, including ones just recently put forward to explain phenomena, and they are certainly not anything that can be said to be known, or classified as "knowledge." And second, almost all people concerned with the nature of science grant that one of its essential characteristics is tentativeness. Scientific conclusions are accepted only tentatively until something better comes along. This is regarded to be an essential part of the scientific outlook or attitude. But to be knowledge, something must be known to be true: it must be certain, and cannot be merely tentative. Hence, science cannot properly be classified as a type of knowledge. It may be that science used to be thought of as a type of knowledge, as suggested by its etymology, but the meaning of the term has changed, perhaps within the past two centuries. Tentativeness has come to replace certainty. So, although science may contain knowledge, as indicated by the fact that students learn it and people look to it for information, nevertheless it would be wrong to define it, as D2 and D3 do, as a type of knowledge. D1 is somewhat vague what a "study" is supposed to be. It is also not clear that D1 excludes theology and metaphysics as science, whereas it ought to. And it may erroneously exclude geography and archeology, since they are forms of science that do not clearly show the operation of general laws. The main defect of Klemke's D2 is that it does not allow for the fact that science is an activity. People can "do science" or engage in science, so it is clearly not just a set of statements. Even when science is taught as a subject matter, D2 is still too narrow, for there are sciences that do not contain all the items on the D2 list. For example, geography and archeology, which were mentioned above, do not contain any "theoretical statements which connect and account for the largest possible number of laws." Another criticism of D2 has to do with its use of the word "statements." When students learn science, some of what they learn is merely inferred and not explicitly stated, and what they eventually do with what they learn may not be to state it, but only to accept it or just entertain it in thought. So a more general term should be used in that context. I recommend the term "proposition." It includes things that are stated or asserted, but it also refers to things that are merely believed, known, supposed, entertained in thought, verified, or confirmed, and having the properties of truth or falsity, testability, deducibility, consistency, and so on. Thus, when we speak of students as "learning science" (or of science as being "taught"), science should in that context be construed to be a set of propositions. (Of course, scientific method can also be taught, and I do not mean to exclude that by what I say here.) Ziman's D3 is a rather unusual definition. It is both too broad and too narrow. It is too broad because it includes things as science which clearly do not belong, for example, theology and metaphysics. Surely the attainment of a consensus is as desirable there as it is in science. D3 is also too narrow because seeking a consensus is not the only goal of science, and is not even essential to it. Although a consensus is desirable, scientists seek other things which do not depend on the attainment of a consensus, such as truth and knowledge. They also seek precision in their observations and simplicity and comprehensiveness in their theories. They also seek applications of their results in the realm of applied science. [It may be that the discovery of applications is an extra-scientific goal (like the securing of research funds), but in that case, so is the attainment of a consensus.] Since definition D3 would erroneously exclude from science all knowledge gained by scientists who focused on goals other than a consensus, it is thereby too narrow. Although I reject Ziman's definition of "science," I agree with his point that science is essentially a social or community activity. He says in his essay that science cannot be done by Robinson Crusoe, alone on his island. With some qualification, I would agree with that. Crusoe could make observations that have the potential for later becoming a part of science, or at least relevant to it, and he could to a certain extent perform "scientific theorizing." However, he could not "do science" in any full-blown sense for the reason that his observations would need to be personal rather than inter-personal. He could never be assured that he wasn't merely hallucinating, for there would be no one around to check his observations. Surely that would be a source of genuine concern for him, especially after several years had passed. Ziman is right that science needs to be a social activity, since the observations on which it is based need to be interpersonal ones, but he goes too far in maintaining that science's only goal is the attainment of a consensus. Like D3, D4 and D5 are both too broad, including not only theology but religion itself. Many centuries ago, religion was all that humans had for explaining nature and the universe, but that is not enough to make primitive religion a science. D4 is also too broad in that it would include as science all attempts at truth-seeking, even the unsystematic type often called "common sense." The distinction between common sense and science will be commented on below. D5 would also include the religions of present-day primitive tribes around the world, which makes it particularly broad. More is needed for a system of thought to be a "science" than merely that it aim at understanding and explanation. What, then, is science? I agree with D1-D3 in various ways. I agree that science is closely connected with knowledge, though is not itself a type of knowledge. I agree with the appeal to "a body of truths systematically arranged" contained in both D1 and D2. Science is both something that can be presented or taught and something that can be done or engaged in. It is in part a certain sort of pursuit of knowledge: knowledge is indeed its aim insofar as it has any aim. But the methodology by which science pursues knowledge is of a special sort. For one thing, it does not employ any appeal to authority initially, though authority may be appealed to in the later teaching of (the results of) science. Science employs what I shall call "empirical method," which may be defined as follows: (1) the making and recording of precise interpersonal observations, (2) the classification of those observations according to subject matter, (3) the formulation (possibly through the use of induction) of general laws and theories which serve to explain the observations, and (4) the testing of those laws and theories by appeal to further precise interpersonal observations that are derivable from them as predictions. My own rough working definition of "science", which I shall label "D6", makes use of the above definition of "empirical method." It is the following: D6: Science is the pursuit of precise, systematic, comprehensive knowledge by empirical method, and the propositions which are involved in and which result from that pursuit. I see science as bifurcated into two parts: an activity and a set of propositions produced by the activity. When we speak of "doing" or "engaging in" science, it is the activity to which we refer. But when we speak of teaching or learning science, it is the propositions to which we refer (except in the special case of teaching or learning scientific method, which is a way of referring back to the activity of science). The set of propositions produced by science need not actually be precise, systematic, and comprehensive, but that sort of proposition or set of propositions must be the sort that is pursued or aimed at. That is, to be doing science, one needs to aim at precision, systematicness, and comprehensiveness (the greatest generality possible in one's explanations). It is these features which distinguish science from what might be called "commonsense (or everyday) knowledge." An example of "commonsense knowledge" is the proposition that a block of ice will melt if it is left in the sun on a hot day. An example of scientific knowledge is the proposition that the melting point of H2O in the solid state under standard pressure is 0 degrees Centigrade. One objection to D6 is that it excludes formal sciences like logic and mathematics. I would say that logic and mathematics play an important role in science, but they are not themselves sciences. Although they are used in "empirical method" as defined above and also in the process by which the set of propositions produced by empirical method is made increasingly systematic, logic and mathematics have no subject matter of their own. It is only through application to something else that they are ever used to "express knowledge." If I am wrong about that and logic and mathematics really do have a subject matter of their own (perhaps some sort of Platonic entities), then I would want to distinguish "a priori science" from "empirical science." In that case, D6 could be taken as a definition of "empirical science" and some other (quite different) definition would be given for "a priori science." The concept of "science in general" could then be defined as "anything that is either empirical science or a priori science." Another objection to D6 is that it excludes applied sciences such as engineering, agriculture, and medicine. In reply to this, it might be said, first of all, that D6 is not a definition of "a science" but rather "science" (in general). Second, insofar as science is done within the applied sciences, I should think that it does conform to D6. What science is there within engineering, say, that does not pursue precise, systematic, comprehensive knowledge by empirical method? I am not aware of any. The objection seems misguided. No doubt there are other important objections to D6 as a definition of "science" (or "empirical science") and these will emerge as the discussion of the topic progresses. Another concept that I would like to consider in this essay is that of "scientificness," especially as applied to theories. Theories are sets of propositions put forward to explain facts or observations. They could come to be widely known to be true and thereby become facts (or sets of facts). One example of that is the heliocentric theory of the solar system. Though not initially known to be true, several centuries ago it became a set of facts. What, then, is the distinction between a scientific theory and a non-scientific one? I would say that a scientific theory is one that is capable of being produced by, or used in, science. Hence, it must be capable of employing empirical method, which (as described above) relies on interpersonal observations within a framework of natural law. The main criterion for such capability is whether or not the theory is: (1) testable (i.e., test procedures that appeal to interpersonal observations can be clearly described for the theory) and (2) compatible with natural law (i.e., conforming to the known laws of nature). The theory need not be true and need not be used by current scientists, but it does need to be the kind of theory that might fit in. Scientific theories satisfy this "empirical method" criterion, as it might be called, whereas unscientific ones do not. If a theory is unscientific, it is because it is not part of an empirical pursuit of knowledge but something else, perhaps a system of thought based on revelation or authority, or something derived only from personal experiences or imagination rather than interpersonal observations. As a paradigm example of an unscientific theory, I offer the "gremlin theory of car failure." It maintains that the reason why some cars won't start is that they have gremlins in them, where gremlins are by definition undetectable beings who dwell inside engines and cause them on occasion to malfunction. They do this by means of a supernatural power which also enables them to forever elude observation, no matter how sophisticated the investigator's detection devices may be. Consider the prediction that gremlins will prevent a certain car from starting. Is this testable? I would say, "Only partly, not fully." The prediction can be analyzed as the conjunction of these two statements: (a) The car's engine won't start. (b) The cause of that is gremlins. Only statement (a) is testable, not (b). Since statement (a) can be falsified, the entire prediction can be falsified. But falsifiability of a prediction is not enough to make the theory on which the prediction is based testable. Since part of every prediction made by the theory is untestable, the theory itself is untestable and should not be called "scientific." It also fails to conform to natural law, since the gremlins do things by means of forces that lie outside the known laws of nature. That would be another reason for regarding gremlin theory to be unscientific. It has been objected that the attempt to draw a line of demarcation between scientific and unscientific theories is misguided. Two reasons are commonly given. One is that all the proposed definitions of "scientific theory" are failures because of some obscurity within them. And the other is that, even if the boundary could be clearly defined, the distinction is useless, because, whatever purpose one might try to fulfill by means of it, it would be better to try to achieve that purpose by drawing a different boundary, in particular, that between "good" theories and "bad" theories. Let us look more closely at these objections. The first objection maintains that all the various attempts to demarcate between scientific and unscientific theories have been failures. But what is wrong with the criterion that I propose above: that a theory is scientific if and only if (1) it can be empirically tested and (2) it is compatible with natural law? It might be objected that some theories in physics do not meet requirement (1). For example, quantum theory cannot be empirically tested because it deals with entities that are in principle unobservable. My reply is that this is an overly narrow conception of testability. Quantum theory does generate empirically testable predictions and that should suffice to satisfy requirement (1). The details of this issue lie beyond the scope of the present essay. It should be noted, though, that the testability requirement is a fairly standard one and is used by scientists themselves, as well as philosophers of science. The second objection maintains that even if a clear line of demarcation can be drawn between scientific and unscientific theories, it is not required for any useful purpose. Why draw the distinction at all? One reason is pedagogical. Schools need to decide what to teach under the heading of "science" and what to exclude. A related reason pertains to publications and libraries. Editors and librarians need to decide which manuscripts and essays are "scientific" and which books belong in the "science" section. Still another reason has to do with funding. Agencies like the National Science Foundation need to decide whether or not a proposed research project is "scientific" or not. There are many such purposes served by the given line of demarcation. The objection being considered is that instead of distinguishing "scientific" from "unscientific" theories, it would be more appropriate to distinguish "good" theories from "bad" theories. Thus, the reason why gremlin theory should not be taught in schools, or funded, or published, or given library space, is not because it is "unscientific" but because it is a "bad" theory. My reply is, first, that the line of demarcation between "good" and "bad" theories is not any clearer or easier to draw than that between "scientific" and "unscientific." And second, there is still a need for the scientific-vs.-unscientific distinction within our present institutions. Libraries and schools, for example, will always have "science" sections and "science" classes, and not "good theory" sections or "good theory" classes. So, the suggested shift in focus is simply impractical, given the organizational structures within our present institutions. For these reasons I find the objection to be a failure. It seems to me, at least at present, that a line of demarcation between scientific and unscientific theories can be drawn, somewhat along the lines suggested above, and that the standard objections to such an approach can all be satisfactorily rebuffed. III. Other Concepts Two other concepts related to science are facts and (empirical) laws. There are several meanings for the word "fact." According to one of them, a fact is any proposition known to be true. By that definition, a fact has to be a truth, but a truth may not be a fact. One of the "Abe Lincoln" propositions, formulated in section I, above, as (a) and (b), is a truth, but neither of them would be a fact. A law (i.e., an empirical law, as opposed to a normative one) is a proposition that corresponds to some regularity or uniformity of nature (or uniformity of behavior in the case of laws in the social sciences). Laws become facts when they come to be known to be true. One very important concept is that of explanation. An explanation for an event or state of affairs is any relevant answer to the question "Why did the event occur?" or "Why does the state of affairs obtain?" The explanandum (or explicandum) is that which is to be explained and the explanans (or explicans) is the explanation given. If the explanans logically entails the explanandum (or a description of the explanandum), then the explanation may be said to be "complete." A scientific explanation is one that might be put forward in the process of doing science. Two main criteria that would need to be met are: (1) that each proposition in the explanans be (directly or indirectly) intersubjectively testable by appeal to sensory observations, and (2) that the explanans be compatible with the known laws of nature. There is a certain model for "complete scientific explanation" which is sometimes called the "D-N (or deductive-nomological) model" (or the "covering-law model"). It requires, in addition to satisfaction of the two criteria mentioned above, that the explanans contain at least one law of nature that is required for the derivation of the explanandum. For example, suppose the fact to be explained is "this wire conducts electricity." An explanation that meets the D-N model would be: "the wire is made of copper and all copper conducts electricity." The explanation is "deductive" in that the explanandum does indeed follow logically from the explanans, and it is "nomological" in that it makes use of the law of nature that all copper conducts electricity, which is a law required for the derivation of the explanandum. There can also be complete scientific explanations of individual events. For those, the D-N model requires that the explanans specify the cause of the event and make use of a causal law. For example, why did the outside of the glass become wet? Answer: the glass was cold and it cooled the surrounding air sufficiently for water vapor in the air to condense upon the outside of the glass. Such condensation occurs whenever the cooling of air causes its water vapor to have a density greater than the saturation density of water vapor in air. The relevant causal law here is that the cooling of air beyond a certain specifiable level causes the water vapor in the air to condense. And since the explanandum is derivable from the explanans, the requirements of the D-N model are satisfied. Laws themselves can be explained by showing how they follow from more general laws. Consider, for example, the law that ice always floats on water. That can be explained by appeal to the more general laws that the density of ice is always less than the density of water (the reason for this being that water always expands when it freezes), together with the law that objects less dense than fluids always float on them. Pushing the explanation one step further, why is it that objects less dense than fluids always float on them? Answer: because (by Archimedes' Law) a fluid always buoys up an object immersed in it with a force equal to the weight of the fluid displaced by the object, and that force is always greater than the weight of the object if the object is less dense than the fluid. (Of course an object floats if it is buoyed up by a force greater than its own weight.) This process of explaining laws by appeal to "higher" laws cannot go on indefinitely. There must be some "highest" (or "basic") laws which cannot themselves be explained in accord with the D-N model of explanation. Thus, the D-N model has the implication that not everything can be scientifically explained. There must be some laws, at least, which must simply be accepted without explanation. All that can be said is that the universe contains those uniformities, and science at present can provide no answer as to why it contains them. Some have suggested that there might eventually be discovered a "theory of everything" that would explain, among other things, why the universe contains the uniformities that it does. It is not impossible that there should be such an explanation, but it is at best a goal of science in the future. Theories are sets of propositions that are put forward to explain facts or observations. They could come to be known to be true and thereby become themselves facts (or sets of facts). One example of that is the heliocentric theory of the solar system. Though not initially known to be true, several centuries ago, it became a set of facts. A hypothesis is any explanatory proposition, usually one that has not as yet been strongly supported. It may or may not be part of a theory. One final concept to be considered is that of evidence. Here is a possible definition: D1: E is evidence for hypothesis H if and only if H explains E better than any alternate hypothesis does. This may be too narrow because we speak of particular instances being evidence for a proposition even if the proposition does not explain the instances. For example, suppose ten marbles are drawn at random from an urn containing 100 marbles and all ten are found to be red. That is some evidence that the next marble to be drawn will also be red. However, the proposition that the next marble to be drawn will be red does not explain why the previous ten were red. Perhaps the definition could be expanded as follows: D2: E is evidence for hypothesis H if and only if: (1) H explains E better than any alternate hypothesis does, or (2) there is a hypothesis H' such that: (a) H' explains E better than any alternate hypothesis does, and (b) H' logically entails H. In the case of the marbles, there is another hypothesis, namely, the generalization that all the marbles in the urn are red, which explains the given facts better than does any alternate hypothesis and which logically entails the proposition that the next marble to be drawn will be red. Thus, the marble case would be covered by the second (expanded) definition of "evidence." This issue of the nature of evidence is troublesome and there may be other counter-examples to be considered. These definitions of the various concepts are intended only as suggestions designed to stimulate thought. It would not be surprising for all of them to be found wanting in one way or another. NOTES [1] Editors' Introduction to Part I of E.D. Klemke, Robert Hollinger, A. David Kline (eds.) Introductory Readings in the Philosophy of Science, Revised Edition (New York: Prometheus Books, 1988), pp. 16-17. [2] John Ziman, "What Is Science?" Ibid., pp. 30-31. [3] The Institute for Creation Research, "Impact," no. 161. [4] Formulated by Stephen McCluskey of the WVU History Department. |