فلسفه علم

The term “holism” refers to a variety of positions which have in common a resistance to understanding larger unities as merely the sum of their parts, and an insistence that we cannot explain or understand the parts without treating them as belonging to such larger wholes. Some of these issues concern explanation (see explanation). It is argued, for example, that facts about social classes are not reducible to facts about the beliefs and actions of the agents who belong to them; or it is claimed that we only understand the actions of individuals by locating them in social roles or systems of social meanings.

The most discussed forms of holism in recent philosophy of science have concerned epistemological and semantic issues which have sometimes appeared to threaten reasonable assumptions about rationality. When we make predictions in order to test hypotheses, we rely upon extensive background knowledge: from the reliability of our senses to the other theories presupposed by our experimental techniques, from information about the context of observation to techniques drawn from logic and mathematics, and so on. If our prediction is disappointed, this shows only that something is wrong somewhere: but judgment is needed to conclude that the best overall response is to reject the hypothesis we were testing. In principle, we could reject the embarrassing observations, place some of our background knowledge in doubt, or even question parts of logic or mathematics. Holistic theories of confirmation hold that the rules we follow in making such adjustments are concerned with the overall virtues of whole systems of beliefs, rather than with the support of particular claims. For example, they enjoin us conservatively to minimize change or to seek overall simplicity.

Holistic approaches were defended by Duhem. Poincaré, and others. A classic but sketchy version is in Quine's “Two dogmas of empiricism” (1953. ch. 2): “our statements about the external world face the tribunal of sense experience not individually but only as a corporate body” (p. 41): “total science is like a field of force whose boundary conditions are experience” (p. 42). And one consequence Quine draws is that “the total field is so underdetermined by its boundary conditions, experience, that there is much latitude of choice as to what statements to reevaluate in the light of any singular contrary experience” (p. 43). His more recent “moderate holism” is less extreme, holding that “more modest chunks” of theory are at stake in each observation (1981, p. 71). But he still denies that confirmation focuses on particular claims (see (quine).

Two apparent consequences of such positions seem particularly disturbing. First, once attention is drawn to the virtues of overall systems of beliefs, or the comparative merits of different revisions in our beliefs, some common assumptions about rationality seem questionable. Objectivity seems to require that the bearing of evidence on our beliefs and theories should be a matter of consensus among all rational inquirers, and it is supposed that relatively specific rules guide us in evaluating our opinions. The holistic picture suggests that different revisions in our beliefs may be equally good in the light of the available evidence, or that different “total theories” may be equally satisfactory in the face of all possible evidence. There need be no fact of the matter concerning what scientific rationality requires of us. An instrumentalist view of theories beckons.

The second consequence concerns meaning. Logical empiricists held that our understanding of a proposition was fixed by a set of analytic truths specifying which experiential results were compatible with its truth (see logical empiricism). These analytic truths provided standards of rationality: we should reject the hypothesis if the predictions they license are disappointed. Holism undermines this picture of meaning and analyticity. No distinction is to be drawn between meaning-constitutive analytic propositions and synthetic ones, and any opinion could (in principle) be abandoned in order to make sense of surprising experiences. Our account of what is involved in understanding a proposition must instead talk of the varying degrees to which beliefs are embedded in our overall system of opinions: some could be abandoned with little disruption to the rest of the corpus, while others could not. Understanding, like confirmation, is thus holistic. Our understanding of theoretical propositions is a function of their relations to all of the rest of our knowledge. In principle, any of our opinions could influence how we evaluate the proposition in question. If we endorse the empiricist project of defining meaning in terms of confirmation, the unit of meaning becomes “the whole of science.” Either we need a holistic theory of meaning (cf. Dummett 1975), or we conclude, skeptically, that the concept of meaning has little philosophical role.

These corollaries of holism depend upon an empiricist understanding of confirmation and meaning. They may be avoidable if this understanding is abandoned. Suppose that we emphasize the role of activities such as experimentation in scientific testing, and note that this involves a causal interaction with theoretical entities. We may then be able to explain how the reference of such terms is not fixed wholly by our current views about how to test statements involving them. As our knowledge progresses, we try to preserve reference, although our concepts and conceptions evolve (cf. Putnam 1990; Hacking 1983). This may enable us to defend realism, and to make better sense of how we decide which revisions to our beliefs are appropriate.

Analogous problems emerge in the philosophy of mind. Granted that our actions and predictions reflect the total pattern of our beliefs and desires, observations of how people act and what they expect may be compatible with very different ascriptions of beliefs and desires to them. And the contents of beliefs and desires will depend upon their place in this entire network of attitudes. Holism about meaning and content can lead to skepticism about whether propositional attitudes can have any role in psychological explanation (Fodor and Lepore 1992). This can be used to support the elimin-ativist view that the proper vocabulary for psychological research is physiological, or to argue that explanations in terms of propositional attitudes do not form part of an inchoate or informal science of mind.

References