CHAPTER IV-2 THE CONCEPT OF CAUSALITY: AN ADVANCE FROM HUME* INTRODUCTION Especially since David Hume addressed himself to the subject first in 1739 and again in 1748, the concept of causality as used in scientific work has tantalized philosophers and has occasionally occasioned furious controversies. All that could be agreed upon has been lack of agreement. Dissatisfaction with prevailing views of causality and the need for a new and better concept has recently reappeared in the field of artificial intelligence, as researchers have attempted to show "common sense" and imitate aspects of our "intuitive" understanding of the world (Waldrop, 1987, p. 1297). This essay offers a method of handling the causality concept in social-scientific work that is intended to be practical and useful as well as intellectually satisfactory. The key is to not ask what a causal relationship "is", because in principle one cannot arrive at a useful statement of properties of the relationship of cause to effect. Rather, we should ask which characteristics a relationship must have for us to usefully choose to call it "causal". The cue for such a shift away from thinking about properties is Einstein's break with the conventional notion of time, and his adoption of the new viewpoint that underlay his discovery of the idea of relativity. Reorienting one's mind in this fashion turns out to be very difficult, however, even after Einstein has shown us the path. Bits and pieces of the approach presented here may be foreshadowed in Hume and in other writers, depending upon how you read them. But even if this approach is not completely novel, stating it explicitly and systematically must be new. (Perhaps the reader will more quickly forgive this ado about credit when it is remembered that Hume -- whom friends such as Adam Smith thought to be as admirable a person as has lived -- confessed a yearning for scholarly distinction. In his extraordinary short deathbed autobiography, he spoke of "my love of literary fame, my ruling passion..." If Hume could be so up-front about his yearning, who am I to be bashful?) The idea suggested here must be distinguished from the useful line of work, starting with Paul Lazarsfeld and continuing through Herbert Simon and Herman Wold, that addresses the problem of which among a set of variables should be considered to be causal. These writers have developed a body of statistical techniques for unraveling the skein of causality in available statistical data series. They study how to disentangle the relationships within a complex set of variables, and how to understand the relative causal status of a closed set of relationships. Their works assume causal relationships within the system and consider which variable causes which, i.e., the direction of causality. This view is clearly stated in the advertisement for Glymour et. al. (1988) about "a computer program which uses artificial intelligence techniques to help investigators discover causal models -- that is, systems of linear equations -- that are consistent with a body of correlational data being probed." In contrast, my aim -- as was Hume's -- is to classify _w_h_e_t_h_e_r or not one variable causes another. That is, my purpose is to better understand the nature of the relationship between two given variables, though with reference to other relevant variables, of course. The Lazarsfeld-H. Simon-Wold type of analysis resembles the advances made by experimenters in biology and agriculture who have developed methods for effectively arranging experiments so as to determine which among a set of fellow-traveling variables are the operative ones, e. g. is it the paper or the tobacco that causes cancer? Implicit in such experiments, without explicit justification, is the notion that if you manipulate an independent variable and the dependent variable changes in response, the independent variable may be called "causal". As we shall see later, however, this implicit definition does not cover some crucial situations, which require a more explicit and more extended treatment of the concept of causality. While Hume's discussion is not perfectly clear in all respects, and may even be self-contradictory in some secondary issues, this much is crystal clear: All that can ever be known about events is what can be observed. And the most that can be observed is that there is a "constant conjunction" between events, a statistical correlation. Hume specifically denied that a causal relationship can be established by a priori logical analysis of the features of events or objects. That is, the Lazarsfeld-H. Simon-Wold line of thought cannot provide the needed concept of causality, though it may assist our thinking in other ways. Cleansing though Hume's analysis may have been, it is not fully satisfactory, because both in scientific work and in everyday life we are prepared to label "causal" some correlations but not others. A statement that the flight of birds overhead precedes rain seems to be a different sort of statement than is the statement that actuating the starter of the automobile precedes the starting of the engine; indeed, we behave very differently toward these two statements. And there seems to be a difference in meaning between the empirical relationship of prices on the Dutch stock exchange to the number of houses built in the U. S., and the empirical relationship of mortgage interest rates in the U. S. to the number of houses built in the U.S. This difference between statements that are only predictions and those that seem to have additional meaning causes people to continue wrestling with the concept of causality. And it is that nut that I sought to crack. When puzzled or stumped in one's scientific work, it is often sound strategy to examine a set of concrete examples of the phenomenon in which you are interested, rather than simply reflecting on the discussions of the topic that are found in the scientific literature; confining oneself to the scientific literature rather than looking out the window or in the street for problems and for examples of interesting phenomena is one of the main perverse tendencies in (at least) the social sciences, I believe. Therefore, I doodled up a variety of situations in business and economics, some of which we would ordinarily refer to as causal and some of which not, to look for the elements which discriminate between the two classes. Here I was influenced by that strain of philosophy sometimes called "natural language analysis" which, as I understood it, asks that we clarify a concept by finding out what it generally means in everyday parlance; this is not to say that we cannot or should not use terms in a restricted technical sense, but we should not fool ourselves by thinking that our scientific or philosophical terms refer to phenomena to which they do not in fact commonly refer. More precisely, I examined a set of situations where there seems to be scientific consensus about whether the label "causal" should or should not be used. My aim was to find criteria for classifying these instances into those that are conventionally called "causal" and those that are not. If chosen well, these criteria should accurately identify situations in which the term is and is not used. The next step after the development of criteria is to put these criteria into such form that they can be used as a guide for social scientific work -- a sort of check-off list. The aim is to improve scientific communication. If a social scientist uses the term in such a manner that there will be consensus among his/her colleagues about what the scientist claims to say, her/his meaning will be clearer. Such a set of criteria constitute what might reasonably be called a _w_o_r_k_i_n_g _d_e_f_i_n_i_t_i_o_n (a less loaded term for the key idea in the operational definition) of the term "causality". But even the term "working definition" is not acceptable to some contemporary philosophers, and therefore I shall stick to the term "criteria set". The key element that this criteria set has in common with an operational definition is that the definition consists of a set of reasonably unambiguous _i_n_s_t_r_u_c_t_i_o_n_s for the scientist to follow. The output of the instructions is a judgment about whether or not a particular item should be placed into the category in question -- that of "causal relationship," in this case. It is absolutely crucial to distinguish this criteria set from the long-time object of philosophical search -- definition of causality by reference to its material or physical or "existential" or "ontological" properties. That is, we must not ask what causality "is". In a great many complex and ambiguous issues, the term "is" and also related terms such as "are" and "be" and "exist", is (sic) one of the greatest sources of confusion in English. One reason is that "is" sometimes indicates equivalence ("Two and two is four" and "A car is an automobile") and sometimes a connector ("Jack is skinny"). "Is" is related to some of the grave paradoxes in logic that required unraveling by the pathbreaking discoveries of Bertrand Russell and Alfred North Whitehead. It is (indeed it is) a difficult and fascinating exercise to try to speak or write without using any version of "is." The result is a purified English which its inventor, David Bourland, Jr., calls "E-prime".1 Hume discovered the impossibility of a definition of causality in terms of physical properties. But Hume did not replace the material-property definition with a concept that fits the needs of the working scientist. Recognizing that a material-properties definition could not work, some philosophers tried to define "causality" by inspection of its _l_o_g_i_c_a_l properties as, for example, the formidable "counter-factual conditional". We can skip the details because there is now rather general agreement that this device also has not worked, a main reason being that "causal" is not mainly a term in formal theories but rather belongs primarily to the context of empirical research methods. Later I think it will be clear why such attempts _m_u_s_t fail. A _d_e_n_o_t_a_t_i_v_e definition of causality is implicit in scientific training. The scientist observes examples of relationships which other scientists do or do not label "causal" relationships, and learns to follow these examples. But denotation of "causality" apparently does not suffice, or else there would never be arguments about whether a given relationship should be called a causal relationship. Definition by giving synonyms is a sort of indirect denotation, and like denotation it may help, but it surely is not enough. Yet mere synonymization has often been offered as the solution, resulting in even more befuddlement. For example, Hubert Blalock says that "causality is conceived [in his discussion] as involving the notion of production, i.e., causes produce effects..." [l964, p. l73]. I find this worse than unhelpful because it purports to explain the matter even though it really does not. Einstein's study of Special Relativity pointed in a radically different direction. Percy Bridgman explained how the development of pre-Einsteinian physics had been held back because of the old habit of defining concepts in terms of their properties. It was only after Einstein had forsworn property definitions, especially of simultaneity, that he could make his crucial discovery. As Bridgman put it, Before Einstein, the concept of simultaneity was defined in terms of properties. It was a property of two events, when described with respect to their relation in time, that one event was either before the other, or after it, or simultaneous with it. Einstein now subjected the concept of simultaneity to a critique, which consisted essentially in showing that the operations which enable two events to be described as simultaneous involve measurements on the two events made by an observer, so that 'simultaneity' is, therefore, not an absolute property of the two events and nothing else, but must also involve the relation of events to the observer (l927, pp. 7-8). Philosophically-inclined readers may find it useful to connect the concept of the operational definition with Wittgenstein's ideas. "Wittgenstein warned against seeking the essence of concepts. Instead, he argued that a term's meaning could be found in its use" (Judis, 1996). In this vein, the appropriate concept of causality is a concept framed in terms of the operating criteria for applying the label "causal" to a relationship, rather than in terms of the properties of a relationship and/or some underlying reality. Just as with the case of simultaneity, one needs a great wrench of the mind to accept this point of view. The trouble is that all of us have continually used the concept of causality in our everyday life without great confusion, able to get by with a vague intuitive understanding of when it is appropriate and when it is not, because most everyday situations are relatively straightforward in this regard. The concept of physical causality causes no difficulty when a small boy throws a baseball through a window (though it may be difficult to adjudge causality if we are trying to fix moral or legal responsibility). Therefore we are not accustomed to the difficulty one can face in applying the term usefully in scientific situations that are not similarly straightforward. It was Einstein's great achievement to have made the first such recorded wrench of mind when he weaned himself away from property definitions of time and simultaneity, and thereby taught us how to accomplish similar liberations with respect to other troublesome concepts. The difficulty of the task for physics and for Einstein -- it took him a decade to think it through -- should put us on our guard about how difficult it is for each of us to rearrange our thinking in similar manner, either about causality or about time. And indeed, it is likely that my lack of success in getting this view of the causality concept accepted, or even noticed (Simon, 1970), stems from my underestimation of the mind-wrenching difficulty even for clear thinkers, even after an extended explanation of the matter. Einstein seems to think about the concept of causality on exactly the same lines. He agrees that closed-system deductive logic cannot supply the appropriate concept. "Hume saw clearly that certain concepts, as for example that of causality, cannot be deduced from the material of experience by logical methods" (1949, p. 13). Or more generally, "Hume saw that concepts which we must regard as essential, such as, for example, causal connection, cannot be gained from material given to us by the senses" (1954, p. 21)3. And he then goes on to assert that the appropriate causal concept should depend upon our scientific needs. "All concepts, even those which are closest to experience, are from the point of view of logic freely chosen conventions, just as is the case with the concept of causality" (1949, p.13). Max Planck, too, asserted that the choice of causality concept should be on pragmatic grounds. "In his attempt to build up his hypothetical picture of the external universe the physicist may or may not, just as he likes, base his synthesis on the principle of a strict dynamic causality or he may adopt only a statistical causality. The important question is how far he gets with the one or the other." (1981, p. 99). As in the first flush of development of many intellectual advances, vastly overblown claims were made for operational definitions. There arose a body of thought called Operationism which claimed that application of operational definitions to all terms in science would resolve all problems of scientific conceptualization. Counterattacking, Fritz Machlup (1978, Part 3) vigorously argued that many theoretical terms in economics, and in the rest of social sciences, cannot possibly be operationalized -- that is, reduced to measurement. Machlup even seemed to doubt whether any terms could properly be operationalized, though when pushed he admitted the necessity for at least some of the terms in any theory and discipline to be related to observed reality with some device akin to an operational definition. Let us not be sidetracked into controversy about that matter, however. Our aim is a useful concept of causality in social science. CAUSALITY AND EXPERIMENTATION Natural scientists often say that an experiment defines causality. Indeed, a positive experimental result is quite a good test of causality when experimentation is possible. If the stimulus is followed by response, and non-stimulus is followed by non-response, as in John Stuart Mill's canons, the stimulus- response relationship is commonly said to be causal. Experiments are replicable, and hence the definition has high reliability. This explains why there is relatively little dispute in the experimental sciences about which relationships to call causal. A single experimental relationship is _n_o_t, however, a _p_e_r_f_e_c_t_l_y valid indicator of causality. In the famous Hawthorne experiments, for example, variation in light intensity in the work room was followed by variation in the work performance. But it became obvious that it was other factors -- perhaps the attention of the experimenters, or perhaps associated changes in the rate of pay, an issue still the subject of controversy -- which _c_a_u_s_e_d the increase in work output. It is intuitively clear to experimental scientists that causality is shown better by a _s_e_r_i_e_s of experiments that vary the parameters of the original experiment, than by a single experiment. One can therefore state the criteria of causality in experimental situations as follows: (l) Keeping all other conditions the same, vary the stimulus and observe the response. (2) If the variation in stimulus is followed by variation in the response, yielding a statistically significant relationship that is also strong enough to be of some importance, vary the conditions and repeat the experiment. (3) If the original relationship continues to appear even under different parametric conditions, call the relationship 'causal.' Two important points should be noted about these criteria of causality when experimentation is possible: (l) It is the actual operation of experimenting that defines the term; _t_h_e_ _e_x_p_e_r_i_m_e_n_t_ _m_u_s_t_ _a_c_t_u_a_l_l_y_ _b_e_ _c_a_r_r_i_e_d_ _o_u_t; the experiment is important as an act, and not as a model, in this context. This definition specifies that an actual experiment cannot be replaced by a hypothetical experiment. (2) Whether or not a relationship will be called "causal" is not an automatic and perfectly objective process; rather, it requires _j_u_d_g_m_e_n_t based on unspecifiable contextual knowledge, e.g., judgment about whether the _a_p_p_r_o_p_r_i_a_t_e conditions have been varied, whether _e_n_o_u_g_h conditions have been changed, and whether the observed relationship is sufficiently important or strong. CAUSALITY IN NON-EXPERIMENTAL CONTEXTS Now we move on to the much harder task, a set of criteria of causality for observed relationships that are not subject to experiment. One suggestion is what Wold called "the fictitious experiment," equivalent to this test: Judge whether the observed situation has the properties of a controlled experiment. If you so judge, call the observed relationship 'causal.' But this definition differs from the experimental working definition in that it does _n_o_t include the crucial operative phrase, i.e., "carry out the experiment...." Furthermore, it is clear that this definition has low reliability; that is, there is much room for disagreement among scientists about whether or not an observed situation does indeed have the properties of an experiment. One frequent suggestion has been to deny the label "causal" to any non-experimental observed relationship, to say "correlation does not prove causation." But there are several drawbacks to this suggestion: (l) The term "causal" is frequently used in scientific descriptions of non-experimental relationships, and therefore we need to discern its meaning when it is used. (2) The "pure" scientist may be able to withhold the appellation of causality (though it may well be a useful word in his vocabulary), but the decision-maker (or the scientist _q_u_a adviser for "decision-makers") certainly cannot duck the issue. The 1964 Surgeon General's Committee on Smoking and Health knew that many people would not decide to quit smoking who otherwise would if the Committee did not use the word "cause," and the progress of legislation might also depend upon whether they wrote "causal." Therefore they chose to use the word. (3) Our intuition tells us that there is an important difference among various observational relationships, a difference that corresponds to our usual sense of the word "causal." For example, there is a difference between a) the statement that when one clock's hour hand reaches l2, another clock strikes the hour, and b) the statement that when you remove the plug from the socket the electric clock ceases to run. For an economic example, we sense a difference between the observed association relating prices on the Dutch stock exchange to the number of houses built in the U.S., and the observed association relating mortgage interest rates in the U.S. to the number of houses built in the U.S. Similarly, in sociology there seems to be a difference between a statement that certain phases in the moon precede or accompany a rise in the murder rate, and the statement that a rise in the temperature precedes or accompanies a rise in the murder rate. Here is the working definition that I propose for the term "cause and effect relationship" in non-experimental situations: l. _S_t_r_e_n_g_t_h_ _o_f_ _C_o_r_r_e_l_a_t_i_o_n. The relationship must be strong enough to be interesting and/or useful. For example, one is not likely to say that wearing glasses "causes" (or "is a cause of") auto accidents if the observed correlation is .07, even if the sample is large enough to make the correlation statistically significant. (A correlation is measured by a number between -1.0 and +1.0, with zero indicating no correlation. In almost every discipline except perhaps educational research, a correlation of .07 is usually considered to be of no importance at all.) In other words, unimportant relationships are not likely to be labeled "causal." Of course this criterion by itself is not enough; that is the grain of truth in the expression "correlation does not prove causation." But nothing else "proves" causation, either; that is the larger truth. 2. _F_e_w_n_e_s_s_ _o_f_ _S_i_d_e_ _C_o_n_d_i_t_i_o_n_s. The relationship in question must not require too many "if's," "and's," and "but's." That is, the "side conditions" must be sufficiently few, and sufficiently observable, so that the relationship will apply under a wide enough range of conditions to be considered useful or interesting. For example, one might say that an increase in income "causes" an increase in the birth rate if this relationship were observed everywhere. But if the relationship is only found to hold true in developed countries, among educated persons, among the higher-income groups, among those who can be assumed to know about contraception, then one is less likely to say the relationship is causal -- even if the correlation is extremely high once the specified conditions have been met. 3. _N_o_n_-_S_p_u_r_i_o_u_s_n_e_s_s. For a relationship to be called "causal" there should be good reason to believe that even if the "control" variable is not the "real" cause (and it never is), some "more real" variables will change consistently with changes in the control variables. (Between two variables, v may be said to be the "more real" cause, and w a "spurious" cause, if v and w require the same side conditions except that v does not require a side condition on w.) This third criterion (non-spuriousness) is of particular importance to policymakers. The difference between it and the previous criterion concerning side-conditions is that a plenitude of very restrictive side-conditions may take the relationship out of the class of causal relationships even though the effects of the side-conditions are known. But the criterion of non-spuriousness concerns variables that are as yet _u_n_k_n_o_w_n and unevaluated, but which have a _p_o_s_s_i_b_l_e ability to upset the observed transformation. Examples of "spurious" relationships and hidden-third-factor causation are commonplace. For a single illustration, toy sales rise in December. One runs no danger in saying that December "causes" an increase in toy sales even though it is "really" Christmas that causes the increase, because Christmas and December (almost) always accompany each other. One's belief that the relationship is not spurious is increased if _m_a_n_y likely third-factor variables already have been investigated and none reduces the original relationship. This is a further demonstration that the test of whether an association should be called "causal" cannot be a logical test; there is no way that one can express in symbolic logic the fact that "many" other variables have been tried and have not changed the relationship in question. The more tightly a relationship is bound up with (that is, deduced from, compatible with, and logically connected into) a general framework of theory, the stronger is the relationship's claim to being called causal. For an economic example, the positive relationship of the interest rate to business investment, and the relationship of profits to investment, are more likely to be called "causal" than is the relationship of liquid assets to investment. This is because the first two statements can be deduced from neo-classical price theory whereas the third statement cannot. This element of theory in scientific thinking and in the explication of the concept of causality is perhaps the biggest difference between Hume's thinking and contemporary thinking. Hume focused on each relationship all by itself, and all by itself there is no more that one can say except that there is "constant conjunction". But the presence or absence of other statements of relationship that are either connected to the relationship in question by commonsensical logic, or even more strongly by an integrated body of theory, is very important in deciding whether it is sensible to think that the statement in question should be considered as only a predictive relationship, or whether one should go further and call it "causal". One likely reason for absence of this consideration in Hume's thinking is that in his time there was no branch of science, except perhaps physics, that possessed such an integrated body of theory. Economics lacked an integrated body of theory until Adam Smith came along to weld together the various fragmentary observations that already existed; William Letwin (1975) has persuasively argued that this was Smith's greatest achievement. Certainly there was at that time no philosophy of science that analyzed the importance of an integrated theoretical framework, as has been done for us by recent writers. Indeed, none of the social sciences other than economics yet has a well-developed body of deductive theory, and hence this criterion of causality is not weighed as heavily in those social sciences. Rather, the other social sciences seem to substitute a weaker and more general criterion -- whether the statement of the relationship is accompanied by other statements that seem to "explain" the "mechanism" by which the relationship operates. Consider, for example, the relationship between the phases of the moon and the suicide rate. The reason sociologists do not call it "causal" is because there are no auxiliary propositions that sensibly "explain" the relationship and describe an operative mechanism. On the other hand, the relationship between broken homes and juvenile delinquency is often referred to as causal, because a large body of psychological theory serves to explain why a child raised without one or another parent, or in the presence of parental strife, is not likely to "adjust" readily. Depending on who is reading him, it might be argued that Kant sensibly brought theory back into the discussion of causality. But if he did so, it was in such a confusing fashion that progress was not made, but rather the opposite. This does not exhaust the list of possible criteria of causality. For example, on pragmatic grounds Guy Orcutt demands that a relationship be controllable for policy purposes. And Travis Hirschi and Hanan Selvin studied various _f_a_l_s_e criteria. But my survey of the use of the term "causal" in the social- scientific literature suggests that the definition proposed above captures more of the essence of the concept than any other definition. Others may disagree, and I hope that they will construct better definitions--but heuristic definitions. THE ROLE OF THEORY In addition to the analysis above, as stated in the 1960's, additional attention should be given to the place of theory in the decision to label a relationship causal. Hume begins and ends his analysis with the idea observing the frequency with which two events are conjoined. But I have asserted that the existence of a theoretical connection between the events is one of the predisposing conditions for us to call a connection "causal". Therefore it behooves us to consider what is meant by a theoretical connection, and why its existence in a particular situation makes it more appropriate to label a relationship "causal". As noted above, Letwin makes much of the idea that Adam Smith's great contribution was the welding together of a great many existing economic propositions into an interacting system relating those propositions to each other logically. It is this logical structure that made economics into a mature discipline rather than a grab-bag of observations. And it is this structure as a whole that constitutes a theory. A single proposition does not constitute a theory, Letwin argues; only an integrated structure of such propositions does. The practical importance of such a structure is that one properly has much more confidence in the validity of a proposition that is part of such a larger interacting structure than in the validity of a similar proposition that stands alone. A proposition that is part of such a structure calls forth more confidence because it rests not only upon the evidence available to support that proposition alone, but also upon all the evidence that supports the other propositions in the theoretical system. In the same way, we feel more secure in calling a connection "causal" if, in addition to the empirical evidence bearing directly upon that connection, we can also call forth a theoretical explanation deduced from a set of propositions that interact with it logically. Hume uses as an example one billiard ball striking another. He says that it is only because we have experience that a moving ball colliding with a ball at rest sets the latter in motion that we predict that the same will happen next time there is such a collision. But if we possess an integrated body of physical propositions about moving objects, force, mass, and the like, we feel more confident in predicting such a phenomenon than if we did not have such theory available. Furthermore, we often find ourselves using logical deduction in bringing theory to bear in such a case, though Hume leaves the impression that logical deduction from observations has absolutely no role in affixing the label "causal". But it is important to note that, at the bottom of the theoretical structure from which we make our deductions, there necessarily lie empirical observations. The theory does not exist purely in the mental realm; rather, it necessarily has some observed links to observation, or else it has no validity as a theory. The theory serves as a device for taking advantage of a wider set of observations in making predictions about a particular connection, such as that between a pair of billiard balls. Perhaps it will help to explicate the matter if we notice that some judgments are necessary before we can make any prediction. For example, a prediction about the behavior of a pair of billiard balls requires that we classify these billiard balls, and the force that starts one of them in motion, as similar to previous billiard ball events. And a prediction about another pair of more-or-less round objects based on experience with billiard balls requires more delicate judgment as to whether the similarity is sufficiently great so that our experience with billiard balls is relevant. A body of theory is a systematic way of taking advantage of experience with not-so-similar phenomena so as to increase the body of our experience that is relevant to a particular phenomenon. Hume never seems to bring to bear upon a particular phenomenon such as a billiard-ball collision any experience with similar-but-different phenomena, that is, with a theoretical structure. This is one reason why his discussion of causality stops only with constant conjunction and prediction. And this is why he cannot distinguish between the sorts of connections that we call "causal" and those that we consider only predictive. Einstein may be interpreted as saying much the same thing: Our present rough way of applying the causal principle is quite superficial. We are like a child who judges a poem by the rhyme and knows nothing of the rhythmic pattern. Or we are like a juvenile learner at the piano, just relating one note to that which immediately precedes or follows. To an extent this may be very well when one is dealing with very simple and primitive compositions; but it will not do for the interpretation of a Bach Fugue. Quantum physics has presented us with very complex processes and to meet them we must further enlarge and refine our concept of causality. (in Planck, 1981, pp. 203-4). DISCUSSION My heuristic definition is a checklist test against which one can compare a given relationship. If the relationship seems to meet most of the checklist criteria reasonably well, then you probably will (and ought to) call the relationship "causal"; if not, not. This test is not automatic or perfectly objective, of course3. Whether the relationship meets any one of the criteria, or enough of the criteria, is a subjective judgment and demands knowledge of the substantive context of the problem. This is why no _l_o_g_i_c_a_l definition of the causal concept can ever capture its essence. Furthermore, the judgment about whether or not the relationship passes the test of the checklist also depends upon the discipline and area of knowledge of the judge and of the materials he is working with. For example, as noted above, the requirement that the relationship be compatible with deductive theory is much more important in economics than in sociology. The reader may remark on the absence from the definition of a time-direction concept. The reasons are twofold: First and most important, time-dating cannot help determine _w_h_e_t_h_e_r there is _a_n_y causal relationship between the two variables. Additionally, time-dating is itself a difficult and uncertain operation; often one cannot say which event preceded which. This is especially the case when human intentions and expectations of future events influence a person to take an action to affect another event; the effect event then precedes the causal events by way of the expectations. Of course one could argue a forward- moving chain of events, but this easily becomes ambiguous. Whether or not this is a _g_o_o_d definition must be considered from several points of view. First of all, it ought to fit common scientific _u_s_a_g_e. The definition given above actually evolved from an inductive study of statements of economic relationships, and it can best be understood in that concrete context. A second test of this definition is that it fit the reader's _i_n_t_u_i_t_i_o_n. One's intuition is closely related to one's experience with usage, of course, but the intuition has some life of its own, too. In other words, I hope that the reader agrees that the definition offered here really stands for what the concept means to the reader. A third test concerns the reliability of the definition. Clearly it is much less reliable than almost any other working definition in science; the need for contextual knowledge in judging causality assures that there will be many more cases about which judges disagree than for most other definitions. But the better question is whether this definition is _m_o_r_e reliable than other methods of classifying situations into "causal" and "non-causal." If this definition is a helpful improvement, then it may lead to others that are even better. One might argue that all of the criteria proposed above lack good definitions themselves, and hence listing them does not improve the situation. Perhaps so. But often one can make a better judgment when one breaks up an overall judgment into parts, e.g., one can usually make a better judgment about the height of a skyscraper if one estimates the number of floors, and then multiplies by a guessed-at height of floor, than if one has to guess the height of the building directly. Similarly if one at least examines a correlation coefficient, self-consciously thinks about the relationship to the body of theory, etc., one may arrive at a better judgment of causality than if one makes a judgment directly. The concept of causality suggested by Hirschi and Selvin for use in juvenile delinquency research, where the issue is particularly troubling, is quite similar to that suggested above. I have also come across a similarly-spirited discussion of the concept in the context of cigarette smoking by (I think) Richard Doll (citation lost), and a very similar definition in an epidemiology text, stated without fanfare in a practical fashion (Mausner and Bahn, 1974). Epidemiology shares with many social- science situations the characteristic that experiments cannot be undertaken with human subjects. But these discussions apparently have never entered the literature of philosophy, perhaps because these writers did not realize that their analyses had something fundamental to contribute to philosophy, but rather assumed that they were working out a tool for the benefit of working scientists in their own field. Another reason may be that they did not frame their analyses in terms of a general concept such as the operational definition. DIRECTIONS FOR FUTURE WORK One possible cause of the original paper falling deadborn from the press was that readers did not see directions for further work building upon it. Yet this work does strongly imply additional work. The criterion set proposed here has been tested only against my own casual review of research in economics. It needs to be tested against work in other fields of social science, and perhaps adapted for the needs of particular fields. And perhaps even more important, a systematic testing procedure should be devised to empirically compare the efficacy of this and other definitions of the term "causality", because such systematic testing is the only valid way of deciding which concept should prevail. Unfortunately such empirical work is not part of the philosopher's common practice. Such work might also have the benefit of promoting such empirical work in philosophy. SUMMARY Property definitions of causality are a dead end. Definitions referring to logical properties have failed, and must always fail. What is needed is a set of criteria of causality. This is the set of criteria I propose: A statement shall be called causal (a) if the relationship correlates highly enough to be useful and/or interesting; (b) if it does not require so many side-condition statements as to gut its generality and importance; (c) enough possible "third factor" variables must have been tried to give some assurance that the relationship is not spurious; (d) the relationship is deductively connected into a larger body of theory, or (less satisfactorily) is supported by a set of auxiliary propositions that "explain" the "mechanism" by which the relationship works. This definition is a checklist of criteria. Whether a given relationship meets the criteria sufficiently to be called "causal" is not automatic or perfectly objective, but rather requires judgment and substantive knowledge of the entire context. ENDNOTES *I am grateful to Steven Goldman, David Kelley and Joseph Agassi for useful suggestions about this chapter. 1 Albert Ellis and Robert A. Harper changed their language to E-prime when they revised their excellent self-help book, A New Guide to Rational Living (1961; 1975), and they assure us that it clarified their thinking greatly. My knowledge of E- prime comes from the introduction to their book. 2 Einstein's admiration for Hume is impressive and charming. "If one reads Hume's books, one is amazed that many and sometimes even highly esteemed phillsophers after him have been able to write so much obscure stuff and even find grateful readers for it. Hume has permanently influenced the development of the best of philosophers who came after him." (1954, p.21) 3 I do not suggest that mechanical systems of deciding whether or not a relationship should be called "causal" are impossible or impractical. With a sufficient number of specifications, carefully made, there is no reason why a computer program could not effectively sort into "causal" and "non- causal". For some discussion of this issue, see Pearl (1988).