Tom L. Beauchamp is Assistant Professor of Philosophy at Georgetown University, Washington, D.C. 20007.
The following article appeared in Process Studies, pp. 296-300, Vol. 2, Number 4, Winter, 1972. Process Studies is published quarterly by the Center for Process Studies, 1325 N. College Ave., Claremont, CA 91711. Used by permission. This material was prepared for Religion Online by Ted and Winnie Brock.
According the author, there is a lack of discussion about Whitehead’s view that scientific laws state principles which are immanent in nature but which evolve concurrently with novel changes in the entities actually constituting the universe.
The author examines Whitehead’s view that scientific laws state principles which are immanent in nature but which evolve concurrently with novel changes in the entities actually constituting the universe..
According to Whitehead, scientific laws state principles which are immanent in nature but which evolve concurrently with novel changes in the entities actually constituting the universe.1 This suggestive and almost entirely neglected thesis provides a lever for criticizing contemporary views concerning the scope and character of laws of nature. The received view among philosophers of science, whether they be of a regularity or necessity persuasion, is that a statement s is a law statement or nomological generalization if and only if it satisfies the following logically necessary specifications:
(1) s is universally quantified,
(2) s is omnitemporally and omnispatially unrestricted in scope,
(3) s is omnitemporally and omnispatially true, and, depending upon one’s regularity or necessity predilections, either (4a) or (4b):
(4a) s is contingent,
(4b) s is necessary.
Whether any philosopher has ever held precisely this theory, I do not know. But certainly philosophers as diverse as Popper, Pap, Kneale, and Carnap have endorsed some form of crucial premises (1), (2), and (3).2
I contend that (2) and (3) cannot be defended against an epistemic construal of Whitehead’s position and therefore should be abandoned as necessary conditions of nomological generalizations. Some philosophers -- viz., those who accept either a weak (natural necessity) or a strong (logical necessity) version of lawful necessity -- have argued that (4b) directly supports (1) - (3) in a way (4a) logically cannot. But I intend my arguments to have an equal impact on the supporters of both (4a) and (4b). These arguments also show that Whitehead’s account raises afresh, somewhat surprisingly for those who regard him as Hume’s opposite on induction, some of the latter’s doubts concerning the justifiability of sweeping conclusions reached through scientific inference.
Whitehead’s theory directly challenges (2), a specification traditionally introduced in order to escape the problem of accidental universals. If nomological generalizations were restricted to specific finite space-time regions, the class denoted would be closed and the law statement could be expressed as a finite conjunction of singular statements. Since it is thought that counterfactual conditional statements cannot be adequately grounded by summative inductions and that laws must sustain counterfactuals, it is argued that no restricted universal qualifies as nomic. Regularity theorists generally argue that. the distinction between unrestricted universals of law and mere universals of fact is to be accounted for in terms of the epistemic and contextual support which unrestricted universals receive within a system of scientific theory. Necessity theorists maintain that a distinction must be drawn between necessary universals and mere universals of fact. But both parties agree that unrestricted universality is a logically necessary condition of laws.
Whitehead’s dissenting opinion is best understood as an extension of one dimension of Hume’s polemic against induction. Hume maintains that there is no adequate justification for inference from statements of the form
(a) All examined cases of X are to statements of the form
(b) All cases of X are Y.
Still less was Hume satisfied with inferences from (b) to
(c) All cases of X are necessarily Y.
Whitehead takes the rather different view that while we are conditionally justified, for metaphysical reasons, in making these inferences (when confined to a particular space-time region where there is traceable continuity), we are not justified in construing the empirical generalization unrestrictedly. It may be true but not unrestricted in scope:
In every inductive judgment, there is therefore contained a presupposition of the maintenance of the general order of the immediate environment, so far as concerns actual entities within the scope of the induction. . . . The anticipations are devoid of meaning apart from the definite cosmic order which they presuppose. . . . Thus the completely unknown environment never enters into an inductive judgment. (PR 311f; cf. AI 143f)
Whitehead is arguing, against (2), that laws should be construed as spatially or temporarily restricted and that they only support counterfactual conditionals whose conditions are within the scope of the law.3 Rational warrant, in the sense of epistemic worth, will thus diminishingly vary in accordance with the space-time region within which the law is assumed to hold. This is the epistemological side of Whitehead’s metaphysical speculation. It indicates that he is far more skeptical of scientists’ capacities to make accurate predictions beyond the relatively immediate future than is Hume, who had a strong faith in induction beyond the immediate environment but could not rationally justify it.
William Kneale is the only philosopher to my knowledge who has explicitly disagreed with Whitehead’s general thesis and has met it with an argument. Kneale objects that when special caution is employed by scientists concerning remote inferences beyond local conditions,
their caution can be explained and justified without reference to Professor Whitehead’s theory. . . . When we reason in this way, we are not abandoning a belief in laws of unrestricted universality. We are merely admitting that in our attempts to formulate laws we sometimes overlook factors which should be mentioned and say ‘All a things are b’ when we should say ‘All ag things are b.’ Indeed, the considerations which make the scientist cautious in such a case cannot be stated without an assumption that precisely formulated laws of nature have unrestricted universality. It seems unnecessary, therefore, to discuss [Whitehead’s] alternative further.4
Kneale is contending that the concepts ordinarily used to describe universally connected features in nomological generalizations may differ radically from those which always could be mentioned and which would be contained in a complete formulation of the law. This is the received interpretation of causal laws. Donald Davidson has more rigorously developed the point -- though not specifically as an objection to Whitehead -- by drawing a distinction between a strong and a weak way of interpreting causal law statements which satisfy the Principle of Uniform Causation.5 On a strong interpretation, says Davidson. a causal law statement consists of those predicates X’ and ‘Y’ which are specifically used in singular statements to describe particular objects or events x and y which are instances of the general law. The singular statement "x caused y," in other words, entails a particular law incorporating the predicates actually used in describing x and y. On a weak interpretation, there might be some true descriptions of x and y such that the sentence derived by substituting these descriptions for ‘x’ and ‘y’ in the singular statement "x caused y" follows logically from a true nomological generalization. The second interpretation is weaker because no particular law using the predicates ‘X’ and ‘Y’ is directly entailed by ordinary singular causal statements and the latter can be defended without having to defend any particular law. Yet there is an unrestrictedly true law. This weak interpretation precisely expresses Kneale’s objection to Whitehead.
The argument, however, is question begging as an argument against Whitehead. Kneale presumes that there must be a specifiable, though overlooked or hidden, condition g. In the aforementioned passage, Whitehead refers to this as a "presupposition of the maintenance of the general order." This demand assumes that a full characterization of the antecedent will invariably reveal perfect uniformity, i.e., that if the law statement is true and fully described the same consequent will always result from the same antecedent. This assumption rests on the a priori same cause -- same effect" Uniformity Principle which Hume thought to have no rational warrant, but which he thought we believe nonetheless. Whitehead, in his distinctive way of extending Hume, is agreeing but is also adding a principle of novelty. We would expect this of a process and time-oriented philosophy such as Whitehead’s; and we would expect the reverse of a philosophy such as Kneale’s where laws are necessary principles and are stated in the timeless form "All a things are b."
The following schema is a simplification and reconstruction of Whitehead’s view that it is logically possible, and not known to be empirically impossible, that relevantly similar and fully described causal items may produce dissimilar effects in different cosmic epochs.6
(i) "All ag things are b&-p" is true in Cosmic Epoch 1
(ii) "All ag things are b&p" is true in Cosmic Epoch 2
(iii) "All ag things are p&-b" is true in Cosmic Epoch 3
Obviously it cannot be objected that (i), (ii), and (iii) are different, unrestrictedly true laws. If held to be omnitemporally unrestricted, they could not be omnitemporally true, for each falsifies the other. It also cannot be maintained that the conjunction of (i), (ii), and (iii) is logically impossible. This would not only render the presumably factual principle of uniformity logically necessary, it would also beg the question by simply denying the tenability of the claim that laws evolve and insisting on the notion of uniform causation. Whitehead is, of course, directly challenging this principle of uniformity. This will come as no surprise to those familiar with his theories of concrescence and novelty. But, again, it is somewhat surprising to find Whitehead in the position of being at once more skeptical than Hume about induction, while also being one of Hume’s most outspoken and non-skeptical critics.
Whitehead’s thesis that laws may be restricted in scope must be accepted, I think, unless it be held a conceptual truth that nomological generalizations express unrestrictedly exceptionless regularities. I cannot see that this is a conceptual truth, but even if it were so treated, Whitehead’s argument provides plausible grounds for revising the concept of a law of nature. The main import of this argument for contemporary philosophy of science is simply to show that the "omnitemporal and omnispatial" requirements of (2) and (3) are gratuitous and unjustified. If one follows Whitehead here, the most one can say about laws is that (A) no known data indicate that examined instances of contemporary laws constitute the complete class of instances, and that (B) we have reasonable grounds for holding laws to be unrestrictedly universal within a certain spatiotemporal scope (whereas we can, by experimentation, know that accidental universals are closed or subject to exceptions). The evidence for the knowledge claim in (A) coincides with the reasonable grounds of (B). Once again this epistemic approach to laws is largely, though not fully, shared by Whitehead and Hume. Whitehead thought, of course, that more could be said metaphysically. But it has not been my concern to explore his metaphysical views.
1Whitehead usually states this view categorically rather than hypothetically. It is doubtful, however, whether his speculative system can sustain a categorical claim. Even if one grants the full complement of premises in his metaphysics of novelty, it follows only that laws may evolve -- not that they do evolve. Cf. AI 143-46; PR 139; FR 27-29; MT 211f; SMW 156. Some of Whitehead’s clearest, but least technical, statements are found in AI, chapters 7-8.
2Cf. R. Carnap, Philosophical Foundations of Physics (Basic Books, 1966). pp. 209-14; W. Kneale, Probability and Introduction (Oxford, 1949), sections 13-19, and "Universality and Necessity," The British Journal for the Philosophy of Science, 12, 46 (1961), 89-102; A. Pap, An Introduction to the Philosophy of Science (Free Press, 1962), pp. 292f; K. Popper, The Logic of Scientific Discovery (Harper and Row, 1959), Appendix X.
3Whitehead also maintains that scientists do not find, even within the environment of our own cosmic epoch, laws which are exactly obeyed (AI 143-46; PR 140).
4Kneale, Probability and Induction, p. 73.
5Causal Relations," The Journal of Philosophy, 64 (1967), 691-703.
6This schema indicates my disagreement with Gary Gutting’s recent interpretation of Whitehead in "Metaphysics and Induction," Process Studies, 1, 3 (Fall, 1971), 171-78, esp. 175-77.