Let us look at these issues in a little more detail. Physicist and philosopher of science Pierre Duhem (1861–1916) taught us that there were no critical experiments in physics that could establish that a hypothesis was true. Theories face experimental refutation collectively. An experiment that refutes a hypothesis refutes a network of interconnected ideas—rather than a single idea—pointing to a problem within that network rather than pinpointing the problem. This is because predictions that are tested are deduced from theoretical hypotheses, auxiliary hypotheses and other knowledge. Consequently, it is always possible for the scientist to save a hypothesis by adjusting the auxiliary hypotheses. Duhem pointed out further that the choice of hypotheses to test is governed by considerations of order, symmetry and elegance rather than by their ability to describe the world accurately. Quine extended Duhem’s idea to take in the whole of science, suggesting that it was an entire web of belief—even a world-view—that faced refutation as a whole. He also points out that it is possible to save any belief if we are prepared to make the necessary adjustments elsewhere. This is a phenomenon that is widespread and has been well documented by anthropologists in the case of beliefs in such things as magic. Norwood Hanson (1924–67) taught us that what we took to be facts depended on our conceptual system.
Consistent with the above body of criticism, Popper rejected the nineteenth-century attempt to prescribe a method of discovery or of verification. He tells us:
Science is not a system of certain, or well-established, statements; nor is it a system which steadily advances towards a state of finality…The old scientific idea of epistēmē—of absolute certain, demonstrable knowledge—has proved to be an idol…It may indeed be corroborated, but every corroboration is relative to other statements which, again, are tentative. Only in our subjective experience of conviction, in our subjective faith, can we be ‘absolutely certain’.
Popper does not provide us with a logic of science, nor does he believe that such logic is possible. He also rejected the positivist attempts to distinguish the meaningful from the meaningless along the lines proposed by the positivists. Instead, he sought to divide all human knowledge into two categories: science and non-science. In his view, science is distinguished from non-science by its method of formulating and testing propositions, not by its subject matter and not by a claim to certainty of knowledge. Nevertheless, Popper draws no absolute line between science and non-science, as falsifiability and testability are matters of degrees. All ‘true’ theories are merely provisionally true—having so far defied falsification. Because no individual scientific hypothesis was ever falsified conclusively, Popper suggested certain normative limits on the methods that could be used to safeguard theories against falsification based on what he believed to be sound practice. Let me emphasise the point: there exists no formal method to rule out ad hoc assumptions to save a hypothesis, and Popper has to employ normative rules to save his conjecture–falsification approach from such tinkering.
The empirical basis of objective science has thus nothing ‘absolute’ about it. Science does not rest upon solid bedrock. The bold structure of its theories rises, as it were, above a swamp. It is like a building erected on piles. The piles are driven down from above into the swamp, but not down to any natural or ‘given’ base; and if we stop driving the piles deeper, it is not because we have reached firm ground. We simply stop, when we are satisfied that the piles are firm enough to carry the structure, at least for the time being.
This does not mean, however, that falsification thereby ceases to be a valuable practical scientific tool—though there will be occasions when a hypothesis fails a test because of the inadequacy of auxiliary assumptions. Importantly, for Popper, a theory is scientific only if it gives rise to a known set of conditions that are testable and which will falsify that theory if they do not occur. This was the basis of his critique of Marxism: that it could not be subjected to empirical test and therefore was not scientific. A similar criticism is often made about the core of the neoclassical economic program. Blaug complains—with considerable justification—that mainstream economics preaches falsification but does not practise it. He sees this as a problem all through the social disciplines and even in the natural sciences.
Hungarian Imr Lakatos (1922–74) followed in Popper’s footsteps but talked about progressive and degenerating research programs, suggesting that it was a research program as a whole as it developed over time that should be the focus of attention, rather than its state at a particular point in time. He sees research programs as comprising a hard core, which is essentially untestable, and auxiliary hypotheses, which are testable. He suggests that a research program is theoretically progressive if it predicts some novel, hitherto unexpected fact, and empirically progressive if each new theory leads to the discovery of some new fact. He also cautions about being too hasty in assessing a program, while acknowledging that it is possible to persist with a degenerating research program for too long. This approach, however, while apparently reflecting real practice, weakens the normative significance of Popper’s message. It also gives little practical guidance to a researcher or observer evaluating such a program at any particular point in time.
Importantly, Thomas Kuhn (1922–96), the most influential modern philosopher of science, argues that the appeal to falsification is misleading, because in practice scientists seem to be trying to verify rather than to falsify theories, and because theories that are falsified by particular experiments are rarely abandoned. His seminal work, The Structure of Scientific Revolutions, looked at the history of scientific practice and concluded that all science was based on an agreed framework of unprovable assumptions about the nature of the universe, rather than simply on empirical facts. These assumptions—a paradigm—comprise a constellation of beliefs, values and techniques that are shared by a given scientific community, which legitimise their practices and set the boundaries of their research. Importantly, this view undermines directly the claimed objectivity and value-free neutrality of scientific investigations. Kuhn argues that what he calls ‘normal science’ ‘aims to elucidate the scientific tradition in which [the scientist] was raised rather than to change it’. It uses the same methods that the rest of us use in everyday life. He suggests, therefore, that examples are checked against criteria, data are fudged to avoid the need for new models and guesses—formulated within the current jargon—are tried out in the search for something that covers the cases that cannot be fudged. He goes on to argue that radically new theories arise not as a result of falsification but by the replacement of a hitherto explanatory model—or paradigm—with a new one. Such revolutionary science—the overthrow of a paradigm as a result of repeated refutations and anomalies—is the exception in the history of science. Implicit in this view is the idea that science does not advance in a steady, linear process.
Normal science is a thoroughly social process in which the problems to be examined and the general form the solutions should take are the result of agreement among a scientific community. It is a self-sustaining, cumulative process of problem solving within the context of a common analytical framework. The breakdown of normal science is marked by a proliferation of theories and by methodological controversy. In this climate, a new framework can appear offering a decisive solution to hitherto neglected problems. Conversion to the new approach takes on the nature of an identity crisis or a religious experience. Importantly, Kuhn tells us that there is no neutral algorithm or systematic decision procedure that will determine choice between competing paradigms. He claims that new paradigms are not only incompatible with their predecessors, they are incommensurable. This is because there is no third, neutral language within which rival paradigms can be expressed in full.
Importantly, Bernstein likens this decision process to Aristotle’s practical reasoning—the type of reasoning in which there is a mediation between general principles and a concrete situation that requires wit, imagination, interpretation and the judicious weighing of alternatives—reasoning that is shaped by the social practices of the relevant community. Resolution does not take place by an appeal to the canon of deductive logic or by any straightforward appeal to observation, verification or falsification. Rather, ‘the cumulative weight of the complex arguments advanced in favour of a given paradigm theory, together with its successes, persuade the community of scientists’.
Kuhn subsequently listed five criteria for choice—accuracy, consistency, scope, simplicity and fullness—stressing that these criteria, which functioned as values, were imprecise and were frequently in conflict. He explains that this does not involve a total abandonment of rationality in science, but rather a shift to a more realistic understanding—to a different model of rationality. Indeed, this shift from a model of rationality that searches for determinate rules to one that emphasises the role of exemplars and judgemental interpretation is a theme that pervades all of Kuhn’s thinking. It is a view that picks up on Michael Polanyi’s strong emphasis on the tacit knowledge of the scientist—knowledge acquired in the practice of science, which cannot be formulated explicitly in propositions and rules.
Importantly, for Bernstein, the real character of rationality in the sciences in general—but especially in theory choice—is closer to the tradition of practical reason than to the image of epistēmē . MacIntyre puts it this way:
Objective rationality is therefore to be found not in rule-following, but in rule-transcending, in knowing how and when to put rules and principles to work and when not to. Consider how practical reasoning of this kind is taught, whether it is the practical reasoning of generals, of judges in a common law tradition, or surgeons or of natural scientists. Because there is no set of rules specifying necessary and sufficient conditions for large areas of such practices, the skills of practical reasoning are communicated only partly by precepts but much more by case-histories and precedents. Moreover the precepts cannot be understood except in terms of their application in the case histories; and the development of the precepts cannot be understood in terms of the history of both precepts and case histories.
As the new framework achieves dominance, it becomes the normal science of the next generation.
Kuhn subsequently acknowledged that his earlier description of scientific revolutions involved some rhetorical exaggeration. Paradigmatic changes during scientific revolutions do not imply total discontinuities in scientific debate. In this later account, scientific development is characterised by overlapping and interpenetrating paradigms, some of which can be incommensurable. Paradigms do not replace each other suddenly; rather they achieve dominance in a long process of intellectual competition. Nevertheless, Kuhn’s stress on the role of normative judgements in scientific controversies—and sociological factors such as authority, hierarchy and reference groups—remains intact, along with a mistrust of the role of cognitive factors as determinants of scientific behaviour. Because it is a social process, scientific research is heavily dependent on the norms and ideals underpinning society in general and the norms embedded in inter-subjective communication in particular.
Reflecting on the above literature, Kincaid has suggested that good science requires at least the following evidential virtues—though he acknowledges that they are abstract and simplistic and admit multiple interpretations:
falsifiability as the first line of empirical adequacy
empirical adequacy—the more predictive success the better
wide scope—predicting a wide variety of different kinds of phenomena
coherence with the best information from other sciences
fruitfulness in terms of a past track record and a future promise
Kincaid goes on to stress the importance of fair tests, independent tests and cross tests. Given what has already been said, we might baulk at the possibility of achieving objectivity, but at a practical level this seems a useful suggestion—particularly in respect of normal science—as long as we do not get carried away with its status as ‘the method’ to the exclusion of other lists or with the knowledge status of the results.
Austrian-born Paul Feyerabend (1924–94) goes further than Kuhn and Kincaid. He points out that the physical sciences—the usual exemplars of scientific practice—have not advanced in a manner consistent with the canon of the strict methodologists, including those of Popper. Rather, he believes that progress has depended on a willingness to breach those canons. In part, this is because he sees normal science as a process of indoctrination; and, because science cannot be grounded philosophically in any convincing way, he warns us expressly that scientific findings are no more than beliefs that should not be privileged over other beliefs. Indeed, there is a substantial anthropological literature arguing that the religious stories and beliefs of other cultures are no less rational than our own scientific beliefs—making good sense of experience to the members of those cultures. It is simply that the frames of reference, the paradigms and the tools available differ significantly. It is the height of ethnocentric intellectual arrogance to suggest otherwise. In this regard, Shweder tells us:
A remarkable feature of the entities of religious thought is that they are thought to be external, objective, and real. But it seems to me, it is precisely that feature that marks a point of strong resemblance with scientific concepts, for one of the features of scientific thinking is that ‘representations’ of reality are typically treated as though they were real, and unseen ideas and constructs are not only used to help interpret what is seen but are presumed to exist externally, behind or within that small piece of reality that can be seen.
Nevertheless, Feyerabend believes that scientists should test their perceptions—seeing this willingness as the difference between science and non-science, though these beliefs are no less culturally, socio-politically and historically conditioned. He also draws our attention to the ways in which scientific communities can become closed, rigid and intolerant of new ideas, even though science is often seen as the very model of openness. It is an important part of the argument that will be advanced in Chapter 8 that the community of neoclassical economists has become such a closed group.
The fact that the creation of scientific knowledge is a social process has an important corollary. There are power relationships within any scientific community, as within any other community. Those power relationships, associated with such things as prestigious professorships, the editorship of journals, the referring of papers, participation in funding and appointment committees and positions within the broader community, can have a big influence on the acceptance or rejection of particular theories. And, of course, scientists are no more virtuous than the rest of us. American social scientist and methodologist Donald Campbell (1916–96) described this social construction of knowledge as a quasi-conspiratorial social negotiation involving ambiguity, equivocality and discretionary judgement. This group identification can suppress intra-group disagreement, while exacerbating disagreements between groups and restricting the flow of information and people between them.
Let me reiterate that there is significant agreement on some essential points among these critics of the nineteenth-century image of science. They are all anti-positivist. Strict justification cannot be achieved. In particular, we cannot stand outside our current language and structure of thought. Ultimate justification is not achievable; neither is inquiry free of presuppositions. Consequently, the belief that scientific knowledge is an accurate representation of reality has had to be abandoned. As Rorty put it: ‘We understand knowledge when we understand the social justification of belief, and thus have no need to view it as accuracy of representation.’
It is also clear that any explanation is an explanation from a particular partial point of view—an attempt to reduce the unfamiliar to the more familiar—so that there can be multiple, even inconsistent explanations. All such explanations are tentative. All theories involve abstraction from the complexity of the world and any particular theory highlights particular attributes only. It is an extraordinary leap of faith to believe that any particular point of view can capture successfully the essence of any phenomenon.
In particular, Rorty tells us that the attempt to isolate science from non-science through the use of words such as ‘objectivity’, ‘rigour’ and ‘method’ assumes that scientific success can be explained in terms of discovering the language of nature. Galileo—in claiming that the book of nature was written in the language of mathematics—meant that mathematics worked because that was the way things really were. For Rorty, this was simply a bad metaphor; rather, Galileo’s reductionist mathematical vocabulary just happened to work—something that lacked a metaphysical, epistemological or transcendental explanation. Consequently, for Rorty, the moral that seventeenth-century philosophers should have drawn from Galileo’s success was that scientific breakthroughs were not so much a matter of deciding which of various alternative hypotheses were true but of finding the right jargon in which to frame hypotheses in the first place. What is clear is that the extent to which our mathematical vocabulary matches that of nature—whether nature can reasonably be described as having a vocabulary—will always remain problematic.
It follows that empirical sciences cannot claim an essential grasp of reality and, as a result, a privileged status in the human conversation. It also follows that economics, sociology, political science or even philosophy cannot claim to be objective and rational in a way that moral philosophy, aesthetics and poetry are not. In all cases, justification is a search for persuasive arguments—a fully social phenomenon—not a transaction between the inquirer and reality. In this connection, Peirce referred to the ‘indefinite Community of Investigators’, while Mead spoke of the ‘Community of Universal Discourse’. As we will see in Chapter 7, these concepts have much in common with Habermas’s ideal speech conditions. When it comes to matters of the basic structures of society and major issues of public policy, this community is to be found in the ordinary citizens of the society—not in some intellectual elite who would be the equivalent of Plato’s guardians.
The real problem lies with us and the excessive faith we want to place in scientific knowledge—including knowledge about economics—and the faith we want to place in its practitioners, and in their capacity to free us from anxiety. As Gadamer tells us, ‘[T]he problem of our society is that the longing of the citizenry for orientation and normative patterns invests the expert with an exaggerated authority. Modern society expects him to provide a substitute for past moral and political orientations.’
Further, he says that philosophical hermeneutics ‘corrects the peculiar falsehood of modern consciousness: the idolatry of scientific method and of the anonymous authority of the sciences and it vindicates again the noblest task of the citizen—decision-making according to one’s own responsibility—instead of conceding that task to the expert’.
The above difficulties in grounding rationality and science undermine any sharp distinction between science, philosophy and any other critical inquiry—undermining the special status that we have hitherto attached to science. They also point to our inability to insulate scientific inquiry from the need for practical reason, for judgement and even wisdom. As German critical rationalist philosopher Hans Albert tells us, ‘[T]he problem of adequate criteria is a very general problem. It is to be found in every field of social activity—in every kind of problem-solving activity; in law, morals, politics, literature, the arts, etc—and not merely in the enterprise of acquiring knowledge in science.’