Hempel, Carl Gustav (1905–1997)
HEMPEL, CARL GUSTAV
Carl Gustav Hempel was born in Germany, immigrated to the United States, and became a naturalized citizen. He taught at Yale, Princeton, and Pittsburgh. Along with Sir Karl Popper and Thomas S. Kuhn, a former colleague, he would become one of the most important philosophers of science of the twentieth century. Popper exerted more influence upon natural scientists and Kuhn upon social scientists and the public alike, but Hempel's impact upon professional philosophers of science was unparalleled. His work, including the problems he addressed and the methods he employed, virtually defined the philosophy of science, not just for a few years, but for several decades.
Hempel sought solutions to philosophical problems that were not only well-supported by suitable arguments but which were also precisely formulated by means of symbolic logic. He proposed subtle and nuanced formulations of scientific philosophy and promoted the transition from logical positivism to what would become known as logical empiricism. Hempel was committed to extremely high standards of philosophical clarity and rigor, which enabled his explications to be subject to the most demanding inspection and critical examination. He cared more about finding the right solutions than whether his own solutions were right.
Thus, "logical positivism," the leading movement of the 1930s and 1940s, was based on three principles: the analytic/synthetic distinction; the observational/theoretical distinction; and the verifiability criterion of meaningfulness. Logical positivism thus affirmed that all a priori knowledge is analytic and that all synthetic knowledge is a posteriori, denying the existence of knowledge that is both synthetic and a priori. Sentences that are nonanalytic but also nonverifiable, including various theological and metaphysical assertions concerning the divine or the absolute, thereby qualify as cognitively meaningless.
The precise manner in which scientific theories are to be related to experience therefore became a crucial issue. Observation language is assumed to consist of names and predicates whose applicability or non-applicability, under suitable conditions, could be ascertained by means of direct observation or relatively simple measurement. Theoretical language, which makes reference to nonobservables, such as malleability and conductivity as well as electrical fields and gravitational forces, must therefore either be reducible to observables or is empirically meaningless.
Hempel (1950, 1951) demonstrated that empirical knowledge was thereby restricted to observation sentences and their deductive consequences, which reduces scientific theories to mere logical constructions from observables. In articles on cognitive significance and empirical testability, he persuasively demonstrated that the verifiability criterion implies that existential generalizations are meaningful, but that universal generalizations are not, even though they include general laws, the principal objectives of scientific discovery.
Moreover, on the assumption that a sentence S is meaningful if and only if its negation is meaningful, Hempel demonstrated that implementing the verifiability criterion generates inconsistent consequences. The sentence, "At least one stork is red-legged," for example, is meaningful because it can be verified by observing one red-legged stork; yet its negation, "Not even one stork is red-legged," cannot be shown to be true by observing any finite number of red-legged storks and is therefore meaningless. Assertions about relative frequencies in finite classes and their negations are meaningful, but those about limits in infinite sequences are not.
These realizations suggested that the logical relationship between scientific theories and empirical evidence cannot be exhausted by means of observation sentences and their deductive consequences alone, but needs to be expanded to include observation sentences and their inductive consequences (1958). The concepts of confirmation and disconfirmation (as forms of partial verification and partial falsification) warrant renewed attention, where the crucial feature of scientific hypotheses is their empirical testability rather than their verifiability.
Hempel (1960) argued further that the application of inductive logic supports certain logically impeccable, but psychologically surprising, consequences, such as that the observation of a white shoe confirms the hypothesis that all ravens are black because it is an instance of the hypothesis that everything is either not a raven or black, which, using extensional language, is logically equivalent to all ravens are black. And he proposed that cognitive significance should best be envisioned as a matter of degree that may only be evaluated relative to multiple criteria (1965a).
Dispositions and Definitions
In Fundamentals of Concept Formation in Empirical Science (1952) he addresses the problem of definability in relation to dispositional predicates, such as "malleable," "soluble," and "magnetic," which designate, not directly observable properties, but rather tendencies on the part of some things to display specific reactions (say, attracting small iron objects) under specific circumstances (the presence of small iron objects in the vicinity). On first consideration, it might seem appropriate to define this predicate by means of a formulation employing a conditional: "x is magnetic at t" is taken to mean, "if, at t, a small iron object is close to x, then it moves toward x."
Interpreted as a material conditional, whose meaning is synonymous with "either not … or ____," however, the proposed definition would be satisfied by things not subject to the test condition at all—such as brown cows—when there are no small iron objects in their vicinity. This result threatened the integrity of the project of developing an adequate philosophical framework for understanding the language of science. Both Hempel and Rudolf Carnap displayed great ingenuity in employing the resources of formal logic to cope with it. Ultimately, Carnap would embrace intensional logic as the solution, but Hempel preferred extensional logic, which imposed methodological boundaries upon explications he found acceptable.
Explications of Explanation
Hempel's most important contribution to the philosophy of science, no doubt, was his masterful explication of the structure of scientific explanations as a refinement of the theory of explanation by means of subsumption by general laws, an approach whose precursors date from Aristotle. Hempel developed this conception by means of his "covering law" model, which he elaborated in several versions, understood as arguments whose premises ("the
explanans") include at least one general law, Li, which explain why the event that is described by the conclusion ("the explanandum") occurred by showing it was to be expected relative to its initial conditions, C1-Cm (Hempel and Oppenheim 1948).
Thus Hempel presented a schema that has become familiar to generations of graduate students of the philosophy of science, which incorporated those conditions as follows in Figure 1. A simple example might explain why a small coin expanded when heated by invoking the law that copper expands when heated and noting it was copper. Hempel considered a vast variety of modes of explanation, contending that those which—implicitly or explicitly—conform to this conception are scientific.
Hempel included explanations of empirical generalizations by laws and of laws by theories within the scope of his approach, but devoted most of his attention to elaborating several precise and detailed accounts of the scientific explanation of singular events. And he advanced deductive-nomological and inductive-probabilistic versions to account for differences between subsumption by universal and by statistical covering laws. The differences between them, especially the peculiar difficulties generated by probabilistic explanations, would preoccupy much of his efforts for more than two decades, including Hempel (1948, 1965b, 1968).
The crucial problem turned out to be that of the rationale for the logical link between explanans and explanandum when the covering laws were not universal but statistical. Suppose, for example, that a statistical law of the form, P(B/A) = r, assigned probability of value r to the occurrence of an outcome of kind B, given conditions of kind A. Then an explanation of the form (see Figure 2), invites the presumption that the bracketed variable [ r ] should be understood as a measure of evidential support. Hempel initially adopted such an approach, which reflects an epistemic interpretation of [ r ], but he would subsequently reject it on the grounds that the truth of the explanandum is already known: what we want to explain is why it occurred (Hempel 1968).
While the covering law approach dominated the philosophy of science in the 1950s and the 1960s, such difficulties, which were rooted in deep problems about the nature of explanatory relevance and of probabilistic laws, stimulated other investigations, the most important being the statistical relevance model of Wesley C. Salmon, which denied explanations were arguments and captivated the discipline in the 1970s. Salmon would later abandon the interpretation of nomic probabilities as relative frequencies for the Popperian alternative of propensities as probabilistic dispositions in the context of probabilistic explanation. During the 1980s and the 1990s, no approach would exert its grip upon the discipline as had Hempel's covering-law model, which made explanation a central function of science.
The Problem of Provisoes
One of the most remarkable features of Hempel's career is that he continued to publish original and innovative papers well into the eighth decade of his life. He authored a series of studies that moved away from the standard conception of scientific theories as formal calculi combined with empirical interpretations and, in Philosophy of Natural Science (1966), a widely used introduction to the philosophy of science that would be translated into ten other languages, he even advanced the novel explication of scientific theories as consisting of internal principles and bridge principles, where the general hypotheses that distinguish a theory are connected to observation and experiment by principles expressed in mixtures of ordinary and of technical language, where antecedent understanding replaces explicit definability.
More strikingly, Hempel (1988) noted that the application of scientific theories presupposes the absence of factors that might affect the internal principles of the theory, which goes beyond the content of the theory itself. Deriving predictions and explanations from classical mechanics, for example, presupposes that bodies are being acted upon exclusively by gravitational forces, where the presence of electromagnetic forces would invalidate those derivations. The function of these provisoes means that instrumentalist constructions of scientific theories as mere calculating devices and programs for the elimination of theoretical language by reduction to observational language alone are misguided and cannot be sustained.
See also Carnap, Rudolf; Confirmation: Qualitative Aspects; Confirmation Theory; Explanation in Science; Explanation, Theories of; Kuhn, Thomas; Logical Positivism; Popper, Karl Raimund; Salmon, Wesley; Scientific Theories; Verifiability Principle.
works by hempel
"The Function of General Laws in History." Journal of Philosophy 39 (1942): 35–48.
"Studies in the Logic of Explanation." With P. Oppenheim. Philosophy of Science 15 (1948): 135–175.
"Problems and Changes in the Empiricist Criterion of Meaning." Revue Internationale de Philosophie 41 (11) (1950): 41–63.
"The Concept of Cognitive Significance: A Reconsideration." Proceedings of the American Academy of Arts and Sciences 80 (1) (1951): 61–77.
"The Theoretician's Dilemma." In Minnesota Studies in the Philosophy of Science. Vol. 2, edited by H. Feigl, M. Scriven, and G. Maxwell, 37–98. Minneapolis: University of Minnesota Press, 1958.
"Inductive Inconsistencies." Synthese 12 (1960): 439–469.
"Empiricist Criteria of Cognitive Significance: Problems and Changes." 1965a. In Hempel (1965c), 101–119.
"Aspects of Scientific Explanation." 1965b. In Hempel (1965c), 331–496.
Aspects of Scientific Explanation. New York: Free Press, 1965c.
Philosophy of Natural Science. Englewood Cliffs, NJ: Prentice-Hall, 1966.
"Maximal Specificity and Lawlikeness in Probabilistic Explanation." Philosophy of Science 35 (June 1968): 116–133.
"Valuation and Objectivity in Science." In Physics, Philosophy, and Psychoanalysis: Essays in Honor of Adolf Grunbaum, edited by Robert S. Cohen and Larry Laudan, 73–100. Dordrecht, Netherlands: Kluwer, 1983.
"Provisoes: A Problem concerning the Inferential Function of Scientific Theories." In The Limitations of Deductivism, edited by Adolf Grunbaum and Wesley C. Salmon, 19–36. Berkeley: University of California Press, 1988.
Selected Philosophical Essays: Early and Late, edited by Richard C. Jeffrey. Cambridge, U.K.: Cambridge University Press, 2000.
The Philosophy of Carl G. Hempel: Studies in Science, Explanation, and Rationality, edited by James H. Fetzer. Oxford: Oxford University Press, 2001.
works on hempel
Davidson, Donald, et al. Essays in Honor of Carl G. Hempel. Edited by Nicholas Rescher. Dordrecht, Netherlands: D. Reidel, 1969.
Esler, Wilhelm, et al., eds. Epistemology, Methodology, and Philosophy of Science. Dordrecht, Netherlands: Kluwer, 1985.
Fetzer, James H., ed. Science, Explanation, and Rationality: Aspects of the Philosophy of Carl G. Hempel. Oxford: Oxford University Press, 2000.
Salmon, Wesley C. Four Decades of Scientific Explanation. Minneapolis: University of Minnesota Press, 1990.
James H. Fetzer (2005)