From: Rick Minto <rminto@julian.uwo.ca>

Date: Thu, 01 Feb 1996 14:31:47 -0500

To: ronmerrill@bix.com

Dear Mr. Merrill,

I was intrigued by a comment that you made in your interview in the November issue of _Full Context._ Responding to a question Karen asked about additions to _The Ideas of Ayn Rand_ you'd like to see in a hypothetical second edition, you responded: "I would dig into Rand's metaphysics much more deeply, as I now understand some of the deficiencies in her ideas on causality much better than I did five years ago."

I am currently writing a doctoral dissertation on the subject of causation, which in a nutshell is an attempt to synthesize the various clues about causation that Objectivists, Aristotelians, contemporary scientific realists, and ecological psychologists have been tossing around for years, and to embed the clues (or at least the good ones) within a broader metaphysical framework, in virtue of which the pieces of the puzzle might acquire some theoretical unity.

So I'd be interested in "sharing notes" with you on this issue, if you are willing...

I'll look forward to hearding from you.

Regards,

-------------------------------------------------------------------------- William R. Minto, M. A. rminto@julian.uwo.ca Dept. of Philosophy, University of Western Ontario London, Ontario CA

February 2, 1996

Dear Mr. Minto:

I have been working for some time on a paper on causality, and this project led to the thoughts on the subject in my interview, about which you inquired.

Briefly, the background is this: As you know, with the development of modern science, Aristotle's "four causes" were truncated, so that only the efficient cause is subsumed under the modern concept of causality. This reaction against the rococo development of final causes by the medieval Scholastics came to its culmination in the Behaviorism of Watson and Skinner, in which even human action was considered to have no final cause. Moreover, the Aristotelean view of causality leads to the problem of first causes; which he himself, of course, solved by invoking theistic solutions. Of course efficient causality also raises the first-cause question, but this can be resolved by infinite regression, an expedient not so repugnant to the modern mind as it was to Aristotle.

Ayn Rand also truncated the four causes, but in a different way. Like most other moderns, she rejected any approach that seemed to demand theistic conclusions. But she seems to have associated the efficient cause with determinism (as indeed do many scientists). She therefore took the position that only the formal cause should be considered (or perhaps a blend of the formal and material causes; Peikoff's helium balloon example might be relevant here).

Now, I come to this issue as a scientist. The standard approach (using efficient causality only) has shown disquieting signs in recent years and scientists seem to be on the verge of reconsidering the validity of other types of causality. At the same time, I don't think a coherent account of phenomena can be developed from Rand's approach.

I therefore think we need to go back to Aristotle and start over. Specifically, I want to reconsider the doctrine of the four causes in the light of modern science; unify Aristotle's concept of causality with Rand's more sophisticated axiomatic method; and confront, rather than evade, the problem of first causes.

First causes might be identified, in the modern way of thinking, with the "laws of nature." These are commonly taken to be "brute facts." But I ask the question--and this is by no means abstruse, it is a very practical operational question for scientists: By what means do we know when a fact is a brute fact?

As you can see, this is an ambitious project. I should be glad to exchange notes with you in the hope that we both may benefit.

Best wishes,

Ron Merrill

ronmerrill@bix.com

From: Rick Minto <rminto@julian.uwo.ca>

Date: Sat, 03 Feb 1996 02:44:15 -0500

To: ronmerrill@bix.com

Dear Mr. Merrill,

I don't mind being informal, so you can address me as Rick if you like. May I call you Ron?

ronmerrill@BIX.com wrote:

Ayn Rand also truncated the four causes, but in a different way. She therefore took the position that only the formal cause should be considered (or perhaps a blend of the formal and material causes;Peikoff's helium balloon example might be relevant here).

I think this is sort of right; I would quibble with the use of "truncation," the idea that the only relevant cause is the formal-material cause. As I see it, Rand did subscribe to the causal *primacy* of substance, which gives pride of place to a sort of hybrid formal/material mode of causation. Presumably Rand rejected treating formal and material causation as distinct modes because she rejected Aristotle's formal realism, and its hylomorphic consequences. If this is right, then other modes of causation are admissible and real, but can be explained in terms of formal-material or *substantial* causation. Harry Binswanger and Jim Lennox are both convinced that final causation is a valid concept, so at least some have thought that Objectivism countenances final as well as substantial causation. I think efficient causation is a valid concept as well, as long is it is explained in terms of substantial causation and recognizes its metaphysical priority.

At the same time, I don't think a coherent account of phenomena can be developed from Rand's approach.

Do you mean the approach that equates efficient causation with determinism and thus rejects efficient causation since it is incompatible with free agency?

On this issue, we seem to have widely diverging views. I don't see where first causes come into our world-picture. What explanatory work to they do? There's nothing _a priori_ untenable about an infinite succession of efficient causes, is there?

My formal definition of law of nature is: a capacity ascription to a substance. Examples: matter has the capacity to gravitate, acetaminophen has the capacity to relieve headaches, cells have the capacity to replicate. Formally, capacities are relational properties of substances defined in terms of characteristic kinds of interactions, properties that in general supervene on intrinsic structural and configurational features of the substances to which the capacities are ascribed. Since laws of nature are amenable to analysis along these lines, I don't think they make good candidates for brute facts. I suggest this: in order to be "brute" a fact would have to be a "simple" unanalyzable fact--one that reports a metaphysically irreducible state of affairs.

Regards,

-------------------------------------------------------------------------- William R. Minto, M. A. rminto@julian.uwo.ca Dept. of Philosophy, University of Western Ontario London, Ontario CA

Feb. 5, 1996

Dear Rick:

I have no objection at all to informality.

If I understand Binswanger rightly, he explicitly rejects final causation as a valid concept (see pp. 36-39 of his book). He seems to take the position that final causality is reducible to the efficient cause. I am not familiar with Lennox's position. It is not quite clear to me where Rand herself stood on the issue of efficient causality. Clearly she rejected the modern presumption that all causality is efficient causality, with its implied determinism. For she is explicit that human beings can cause their volitional actions by their nature--that is, that volitional actions at least have formal causes. Instead of, or in addition to efficient causes? That's not obvious from her extant writings, but it's easy to get the impression that Rand completely rejected the validity of efficient causality. Thus Peikoff (*OPAR* p. 15), in defining the law of causality, specifies it as formal causality (though he does not use that term). However, he qualifies it by referring to "circumstances", which might seem to leave the door open to efficient causality also. But then, on page 16, he makes it crystal clear that he rejects efficient causality. How the disagreement with Binswanger is reconciled (if at all), I do not know.

Turning to the question of first causes, the difficulty arises thus. You are quite right that (as I myself said in my first message to you) there is nothing automatically unacceptable in assuming an infinite regression of efficient causes. And in fact modern physics accepts this possibility. Even under the Big Bang hypothesis, such an infinity can be accommodated by General Relativity, for instance using the elegant model of Hartle and Hawking. The same is not true, however, of the formal and material causes. Though we might be willing to accept an infinite regression here also, as a matter of observation this does not occur. Each level of causal explanation is enormously more powerful than the next level lower, and such a progression hardly seems likely to go on forever.

The difference arises because efficient causation explains actions in terms of earlier actions--which are qualitatively the same things. Formal causality explains actions or qualities (ultimately it is not clear there is any crucial difference; or rather, "actions" are merely a somewhat arbitrarily delimited subset of qualities or attributes) in terms of the structure or organization of entities. These structures are themselves explainable (caused by) more fundamental structures, and

so on. But the scope of causality expands at each level.

Now, as to laws of nature. "A capacity ascription to a substance" has some difficulties as a definition. For one thing, one would not normally call biological cells "a substance." Acetaminophen is certainly a "substance" by the definition used in chemistry, but it's not so clear that it is "substance" as the term is used in philosophy. (In fact, modern scientists don't really believe in "substance" as the term is used by philosophers. To a scientist, an entity doesn't have anything except its qualities or attributes, for that is all that can be observed. Once you've specified the organization or structure or "form" of the entity, there's no "substance" or "material" independent of it. For if "substance" as such existed, it would have to be observable, ie, have attributes. But by definition substance is what is left over when you have, so to speak, removed all the attributes. Rand might agree with this; did she not say, "existence *is* identity"?) Set that aside, though. A "capacity" is an option; I have the capacity to drink whiskey, but that does not mean that I do so. On the other hand, matter has no option whatever but to gravitate; indeed, it always gravitates in a particular, very quantitative way. And indeed this is why we refer to a "law" of nature; that which is a law of nature does not leave any option, it specifies exactly how entities of certain identities *shall* behave.

Let me point out where the problem lies. Aristotle's concept of causality is that of explanation; when we ask "why?" something is the way it is (or "why?" something happened) we are asking for a *cause*. Now let us suppose that I as a scientist observe that sodium reacts violently with water. Why? Because sodium undergoes the highly exothermic reaction 2Na + 2H2O ---> 2Na+ + 2OH- + H2. Why? Well, because sodium is a very electropositive metal. Why? Because the sodium atom readily loses an electron to give the stable neon configuration. Why? Because of the shell structure of atoms. Why? Because under the laws of quantum mechanics certain electronic configurations of atoms have unusual stability. Why? Well--why *does* matter organize itself according to the laws of quantum mechanics? Is there a deeper explanation? Or are the laws of quantum mechanics "brute facts"? How do I know whether they are, or are not? They *seem* to be--but, not so very long ago, the fact that sodium is electropositive seemed to be a brute fact.

The scientific process consists of developing just this kind of chain of formal causes, in which each cause is a deeper or more general quality of the substance in question. (The answers given in the above example might be summarized: because of the nature of sodium--because of the nature of metals--because of the nature of atomic substances--because of the nature of matter.) At some point we must have a most general or first cause. The problem is, how would we know when we found it?

Now I am well aware that the whole difficulty can be (and is) removed simply by defining it out of existence. We can say, as is the scientific tradition, that Aristotle's notion of causality as explanation is too general, and that we will limit our concept of causality to the efficient cause. It turns out that this leads to difficulties, not only philosophically but scientifically. We can adopt a Randian approach and take up the formal cause--but then we're right back in the soup, as discussed above. Well then, simply *define* it so that entities "cause" actions, but nothing "causes" entities or their qualities (as Peikoff does). But this doesn't really get you out of it, because the questions still remain.

Where Aristotle was fundamentally inspired--and I think Rand is completely with him on this--is in his conviction that *reality makes sense*--that "why?" is a meaningful question that has an answer. For everything we observe, we are entitled to ask, "why?" Why did that happen? Why does this exist? Why does this entity have the attributes that it has? Why do entities of a certain type exhibit certain qualities or behavior? The Aristotelean approach, which Objectivism follows, takes it that questions of this sort have answers. We do not say, as do almost all other kinds of philosophers, that such questions either are meaningless or that the answers are unknowable.

Such answers have been variously called "causes" or "reasons" or "explanations". Whatever terminology we adopt, and whatever distinctions we make, we must address the issue of the limits of the question "why?" Are there questions which have no answers? If so, how shall we recognize them? We may refer to "brute facts" or to "laws of nature" or to "first causes"; I suspect that ultimately all these terms refer to the same thing.

Well. As you can see, these are random jottings, not an organized account of the matter. Though I have not attempted to discuss the potential solutions on which I am working, I hope this gives you some idea of what I see as the problem, and why it is important.

Best Wishes,

Ron Merrill

ronmerrill@bix.com

ronmerrill@BIX.com wrote:

Binswanger indeed does seem to think so. Notice, thought that is Nagel's position, which Binswanger accepts uncritically. Lennox thinks that justifying reduction claims depends on having at least one actual reduction on the books. There are no such reductions. So its an empty claim. I think Lennox believes in efficient causes.

determinism. But then, on page 16, [Peikoff] makes it crystal clear that he rejects efficient causality.

I don't know about the "crystal" part. Some modes of causation are those in which an alteration in a passive subject is induced by the activity of another object. This is how efficient causation is always characterised, and I think you have to grant that it does happen that way, even if you attach the proviso that the response of the subject to the external activity is partly determined by the nature of the subject.

Now, as to laws of nature. "A capacity ascription to a substance" has some difficulties as a definition. For one thing, one would not normally call biological cells "a substance."

I was using the term substance in an anachronistic Peripatetic sense, "entity" will do fine, and alleviate the difficulties you see with the definition.

Set that aside, though. A "capacity" is an option; I have the capacity to drink whiskey, but that does not mean that I do so. On the other hand, matter has no option whatever but to gravitate; indeed, it always gravitates in a particular, very quantitative way. And indeed this is why we refer to a "law" of nature; that which is a law of nature does not leave any option, it specifies exactly how entities of certain identities *shall* behave.

I see the distinction you are making. Let me try to flesh out my theory of causal laws in order to clarify my own reasons for saying that laws of nature are capacity ascriptions. (Besides causal laws, there are laws of nature that express quantitative relationships between physical magnitudes, laws of quantitative covariance, that are not themselves causal laws.)

First, laws are ceteris parabus statements. The law of universal causation, for example, states that the gravitational force exerted by a body upon others--one that will always accelerate objects towards it, *given the absence of other forces* that oppose the gravitational force, and attentuate, modify, negate, or overcome the force (eg. electromagnetic fields, human lifting, tractor beams).

Second, laws of nature are relations between universals. When we are applying a law of nature to explain, we do so by introducing the particulars subsumed by the law--which necessarily find themselves embedded in a messy causal nexus, connected to many things in their environment.

So when you say that you have the capacity to drink whiskey, but it is not a law that you do so because you have an option, you are not really stating a law. Not all capacity ascriptions are (causal) laws. What I'm saying is that all (causal) laws are capacity ascriptions.

Another example of a causal law would be: human beings have the capacity for sexual reproduction.

Now of course that does not say that you must reproduce, or that you are always engaged in the act of reproduction, or that you have no choice over your reproductive activities. Nor does it say that every actual human being has the capacity--some people are sterile. What is does say is this: here is a capacity ascription. Now, given the presence of the conditions for the exercise of its power, and the absence of inhibiting factors, the object will exemplify a certain characteristic mode of activity: sexual reproduction.

There are some objects which exercise their powers

unconditionally--such as gravitating planets. Yet the absence of inhibiting factors is necessary for the planet to do what planets do.

There are some objects whose power is exercised uninhibitedly--neutrinos apparently can pass through anything.

There are two questions here: The metaphysical question is "is there a level of brute-factuality of formal causation--an irreducible level?" The epistemological problem is "given that there is an irreducible level," how do we know when we've hit it?"

Ron, what makes you think that the series of formal causes *must* terminate? I must have missed something in your earlier post, because I'm still unclear why you think this is *any* different than the case of the temporal series of efficient causes--an infinite series cannot be ruled out _a priori_. I don't think its a necessary condition for scientific inquiry, nor does the absence of brute facts make science fruitless. Grover Maxwell, in his article, "The Ontological Status of Theoretical Entities" argued that the observable-theoretical distinction does not reflect any intrinsic structure of the world, but reflects the state of technology--how powerful our instruments are at detecting microstructural features and interactions of things. Likewise, maybe what counts as a "brute fact" just is a function of our level of scientific devices technology.

One last point: it seems to me that explanation is an intransitive relation. If X explains Y, and Y explains Z, it does not follow that X explains Z. Thus if someone wants to have an answer to "Why Z?" You can say "Y," but not "X." To use your example, "Why does sodium react violently with water?" If I answered, "Because of the shell structure of the atoms," that is not an explanation--it does not explain why sodium specifically reacts violently with water specifically, it (partly) explains what any chemical reactions happen *at all.*

Let's take your apartment-building view of things, where lower levels are causally more general. If explanation is just giving causes, then causation is intransitive too. So while a formal cause on the "ground floor" may give rise to some higher-order phenomena on the "second floor," which in turn might be the formal cause of a "third floor," the third floor phenomenon does not necessarily have the first floor formal causes *as its cause*.

It looks like the intransitivity of the formal causal relation ought to be accepted as a constraint on the possible models. Do you agree?

Regards,

Rick

-------------------------------------------------------------------------- William R. Minto, M. A. rminto@julian.uwo.ca Dept. of Philosophy, University of Western Ontario London, Ontario CA

I'm afraid I still have a number of problems with the explanations you offer.

Peikoff on the efficient cause: ". . . the causal link does not relate two actions." And again, "The law of causality states that entities are the cause of actions . . . and not that the cause of action is action . . ." (*OPAR*, p. 16). Peikoff is absolutely explicit on this issue; he simply does not accept anything resembling efficient causality as either Aristotle or the moderns conceive of it.

Laws of nature: If I understand you rightly, you are taking, eg, the law of gravitation to be a "law of quantitative covariance" that is not a causal law. Certainly this is in violent disagreement with the way scientists think; for we consider that the tides, for instance, are *caused* by the gravitation of the moon and sun.

The idea of laws of nature as capacity ascriptions completely contradicts the modern conception of the term. Indeed, I do not think that Aristotle would go along with it either; he stresses that natural law refers to that which occurs "by necessity". Certainly in Rand's idea of causality, "capacity ascription" is crucial; for an entity is simply the sum of its attributes, ie, its capacities for actions. But the term "law of nature" already has a well-established meaning, which is utterly incompatible with the definition you wish to attach to it. It is always difficult to get people to abandon an established and useful meaning and use the same term to mean something else entirely. You might be wiser to come up with some neologism for "capacity ascription."

That causality operates on a ceteris parabus basis is a truism. A law of nature identifies, however, something that must happen if nothing interferes to prevent it; and must not happen if something *does* interfere to prevent it!

The chain of formal causes: One can easily imagine an infinite regression of efficient causes. The eight-ball went in the corner pocket because it was struck by the five-ball, which in turn was struck by the cue ball, and so on. This is possible because (in general) each efficient cause has no greater explanatory power than its successor. Each impact between billiard balls has just about as much power to cause things to happen as any other such impact, earlier impacts no more than later impacts. But this is not true of the formal cause. Thus in the example I previously gave, when I explain that the reaction of sodium with water is caused by the electropositivity of sodium, I not only explain the immediate fact (action), but many other facts (actions) of sodium--its reaction with alcohol, its dissolution in liquid ammonia to give a blue solution, its formation of a white coating on exposure to air, and so on. At the next level of formal causation, the shell structure of atoms, I am enabled to explain the chemical behavior not just of sodium but of all the elements. At the next level, the laws of quantum mechanics, I can explain far, far more. Clearly this process cannot continue indefinitely; indeed, it is widely believed by scientists that the next level of causality above quantum mechanics will account for (that is, be established as the ultimate cause of) all physical phenomena in the universe--it will be, as the saying is, a "theory of everything."

Of course, one could conceive of an infinite regression of such theories (causes), each as powerful as its successor, starting at some point. (It would be rather like the old idea that the earth rests on a giant turtle, which in turn stands on a still larger turtle, and "it's turtles all the way down.") This is equivalent to the assumption that all general attributes (universals) of entities are reducible to still deeper universals. For instance, we would say that the weight of a stone is due to (caused--formal cause--by) its mass, which in turn is caused by a super-mass, which in turn is caused by a super-duper-mass, and so on to infinity. I admit that I can't refute this possibility offhand, but I think an analysis of the metaphysical implications would rule it out.

Theory and causality: There is a great deal of confusion on this, even among scientists. One sees this in the widely accepted assertion that "evolution is a theory". Scientists and creationists agree; they disagree only on how strong the theory is. But both are wrong. Evolution is a fact; natural selection is a theory.

Anything which we can observe, directly or indirectly, is in the realm of fact; that is, it involves statements about the attributes or actions of entities. That living things evolve is a fact; we can observe it, in the fossil record as well as in the laboratory. That there is a black hole at the center of the Andromeda Galaxy is a fact (if it turns out to be true). That electrons have magnetic moments is a fact. When we say that we have a "theory" about something, what we mean is that we have an *explanation*. *A theory is an assertion of a **causal** connection between observed facts.* That smokers are more likely to get lung cancer is a fact. That lung cancer is *caused* by smoking is a theory. Where Grover Maxwell, and many other thinkers on science, make their mistake is in failing to understand this distinction.

The intransitivity of formal causality: I'm not clear on why you would consider this a desirable property. In any case, it seems a bit problematic. The difficulty arises from composite causality. That is, for instance, both A and B might be necessary to cause C; but B might be caused by A. Example: The flavor of an apple is caused by (among other things) the presence of both fructose and glucose. But the presence of the glucose is caused by the presence of the fructose.

Summing up, I would like to point out how crucially important this issue is. We must understand causality thoroughly in order to handle such problems as the epistemology of science, the meaning of fact and theory, the so-called "problem of induction," and many other questions of philosophy. But causality is an issue of metaphysics. How absurd it is to say that metaphysics is "otiose" (as the Logical Positivists did, and most "conventional" philosophers still do), or to dismiss it as "a highly delimited subject" (as Peikoff does). To paraphrase an author you may recognize, if Objectivism is to survive, what it now needs is not to reform metaphysics, but to discover it.

Best Wishes--

Ron Merrill

ronmerrill@bix.com

From: Rick Minto <rminto@julian.uwo.ca>

Date: Thu, 15 Feb 1996 03:14:42 -0500

To: ronmerrill@bix.com

Peikoff is absolutely explicit on this issue; he simply does not accept anything resembling efficient causality as either Aristotle or the moderns conceive of it.

Unfortunately for Peikoff, Aristotle did countenance analyzing agent causation as efficient causation in some respects. Peikoff should certainly have said more about this; he should have at least agreed with Binswanger on a common "Objectivist" usage, if for nothing more than coherence.

Look at what A says in Nicomachean Ethics 1039a32-33:

"The origin of action-its efficient, not its final cause-is choice, and that of choice is desire and reasoning with a view to an end."

Sure, but the statement of the law says nothing about causation.

You might be wiser to come up with some neologism for "capacity ascription."

I'll think about rewording this idea. I don't mean to deny that laws of nature codify necessities. I am saying that when capacities act, they necessitate or causally determine their effects...*but* that the capacities of an entity need not always be in a state of actualization. Capacities may (and often are) latent--awaiting the conditions for their actualization. Laws of nature describe *how* capacities necessitate their effects, *if and when* they act. Laws of nature do not presume to insist that they only address those entities which are always in the active state.

> The chain of formal causes:

Ok, I see the argument now. Each deeper level of microstructure reveals a level of causation that is more pervasive, and thus broader in explanatory scope. The limit of explanatory scope is "everything in existence," thus the number of levels of formal (or to use the non-hylomorphic preferred term, "substantial") causation is finite.

But I still don't see how you avoid the intransitivity objection: telling me about quantum mechanics does *not* explain to me the action of the rabbit's circulatory system. The explanation of Z needs to refer to the proximate formal cause Y, not to the formal cause X of the formal cause Y.

This is just a case where the antecedent is not satisfied. All that intransitivity says is that if A causally explains B, and B causally explains C, it does not follow that A causally explains C. Or 'CE(A,B) & CE(B,C) |-/-> CE(A,C)'.

Summing up, I would like to point out how crucially important this issue is.

You're preaching to the converted, man. But as an Objectivist, you're excused. :-)

-------------------------------------------------------------------------- William R. Minto, M. A.

Dept. of Philosophy, University of Western Ontario, London, Ontario CA

February 18, 1996

Dear Rick:

Scientific laws: Your "law of quantitative covariance" (which seems to resemble Hume's "constant conjunctions" model), I must maintain, does not represent scientific laws as they are currently understood. When we observe certain quantitative values for the tide, we do not presume that the tides cause the existence of the moon, nor that these two phenomenon just happen to be connected. I think the challenge for you, if you want to defend this notion, is to show that we have any way of justifying such "laws"; for if the "quantitative covariance" is not due to a causal connection, by what means do we know that it will continue to hold in the future?

Intransitivity of causality: Well, I think I now see the issue here, at least. To say that causality is intransitive is just an indirect way of denying necessary causation.

If we take causation to correspond to explanation, we may ask: What would we consider to be an explanation of a phenomenon? Suppose we observe C; what does it mean to say that "B caused C"? We would not ordinarily understand a requirement of inverse necessity: "B caused C" implies that "if C, then B was necessary". This may sometimes apply, but not always. If the eight-ball rolls into the corner pocket, it might be because it was struck by the cue ball; but it might have been struck by some other ball, or moved by a different force. But in ordinary discourse, we do assume causation implies forward necessity: "B caused C" means "B happened (or existed), and therefore C necessarily had to happen". For if this necessity is not accepted, how can we say that we have explained C at all? (Of course composite causality complicates things, but this can be dealt with, eg by the Mill-Mackie "INUS conditions" method.) But if we accept causation as necessity in this sense, then intransitivity cannot hold. For if A necessarily results in B, and B necessarily results in C, then A necessarily results in C. So the two formulations are equivalent. But is it easier to prove intransitivity than to prove non-necessity? If not, what's the point?

Necessary causation and its consequences: Of course, this whole issue of necessary causation was ground through by Aristotle, so I won't try to go into this in depth. But the key issue, obviously, is this: Another way of expressing necessary causation is "uniformity of nature." And without uniformity of nature there can be no science, nor indeed any firm knowledge of reality.

Since necessary causation (if universal) implies determinism, this leads to a dilemma. Many modern scientists have chosen to embrace determinism rather than accept the alternative. Clearly you (rightly, I think) are not willing to go this route. But the alternative of totally abandoning necessary causation strikes me as too desperate. Im my view we should make the minimum concession possible. While affirming necessary causation in general, we must admit two exceptions: volitional action, and metaphysical chance (as in quantum phenomena). Such phenomena must be recognized as first causes in themselves. We must further require that there must be some way to decide whether a phenomenon falls into one of these classes--that is, we must be able to know whether causal necessity applies, or does not, in every case. (I addressed this problem, though I do not claim to have effectively solved it, in my paper "On the Physical Meaning of Volition.")

This general approach I think would be congenial to Aristotle, who advises us that "The educated man looks for as much precision in each subject as the nature of the subject allows." It's not quite so congruent with Rand's approach, as I understand it. And yet, even she would seem to seek to minimize exceptions to necessary causality. Mote that Rand limits volition to the choice to think. That is, after this choice, even human action is necessitated!

Best Wishes,

Ron Merrill

ronmerrill@bix.com

From: Rick Minto <rminto@julian.uwo.ca>

Date: Tue, 20 Feb 1996 00:48:13 -0500

To: ronmerrill@bix.com

for if the "quantitative covariance" is not due to a causal connection, by what means do we know that it will continue to hold in the future?

I think you mistake my meaning here. I _do_ say that for some laws of quantitative covariance, they are what they are in virtue of some natural causal relation between the entities whose magnitudes are so related. That does not deny, however, that what the statements of the laws actually say, is one thing while the causal connections that make them true is another. Equations are not statements of causal necessitation relations.

Intransitivity of causality: Well, I think I now see the issue here, at least. To say that causality is intransitive is just an indirect way of denying necessary causation.

If we take causation to correspond to explanation, we may ask: What would we consider to be an explanation of a phenomenon?

I think we need to apply the method of objectivity to get the order of conceptual priority right here. I think you correctly point out that causation is transitive. My point was that substantial causal explanation is intransitive. Talking about electrons does not explain rabbit mating practices, but at the same time, rabbits mating practices are realized, physically, in matter whose nature requires electron talk.

This suggests that causation and causal explanation (at least substantial causal explanation) are not the same concept.

Another way of expressing necessary causation is "uniformity of nature." And without uniformity of nature there can be no science, nor indeed any firm knowledge of reality.

I agree with this.

Well, isn't it?

Regards,

-------------------------------------------------------------------------- William R. Minto, M. A.

Dept. of Philosophy, University of Western Ontario, London, Ontario CA

March 3, 1996

Dear Rick:

On quantitative covariance: I take you, then, to be saying that we may observe regularities in nature for which we propose a mathematical relationship even though we have not established a causal connection. However, there are other such regularities which we do know to be causal. This is uncontroversial. But does it tell us anything very new?

On causation and causal explanation: The key issue here is reductionism. We must distinguish between metaphysical reductionism and epistemological reductionism.

Metaphysically, all causation relationships are interconnected. (Before the era of modern science, there were doctrines that asserted the contrary--eg, vitalism, which said that biological causation could never be related to chemical phenomena.) There are no isolated areas of causation. This means that all phenomena, no matter how complicated, can ultimately be reduced to basic "physics"; where "reduced" in this sense means that the causal chains are real and can be identified.

This must be separated from epistemological reductionism. As a matter of conceptual practicality, explaining the reproductive behavior of rabbits directly in terms of atomic physics (to use your example) is of course totally impractical and absurd. We thus construct concepts for separate use in each field of study.

However, I do not think this distinction really has implications for our ideas of causality. Anti-[epistemological]reductionism is merely a matter of operational convenience. Thus, when we do tensor calculus, we use the summation convention to simplify the notation and make it possible to handle extremely complex formulae in a manner that would not be practical if we wrote out all the components. That does not in any way change the underlying mathematics. Similarly, when we use biological concepts such as fitness, fertility, genetic recessives, and so on, to explain reproduction in rabbits, that does not change the fact that we have not fully or ultimately explained these phenomena until we have carried the causal chains to their ultimate conclusions, which will lead us into chemistry and ultimately into physics.

Ron Merrill

ronmerrill@bix.com

March 22, 1996

Dear Rick:

If, as I understand you now to be saying, you are denying the principle that scientific (causal) laws can only be established by reproducible experiments, then indeed it is a "radical proposal". For it is by means of reproducibility that we establish that a "regularity" is ceteris paribus.

I never thought the "law of supply and demand" expressed a causal relationship between supply and demand; does it not express a causal relationship between these variables and price? And is it not qualitative, rather than quantitative?

Now, returning to reductionism: I don't know what a "mereological" reduction is. However, suppose we observe a biological phenomenon. For instance, we see that a T-cell "recognizes" a host cell as being infected with a virus. At the biological level, we can say that this process caused the T-cell to take certain actions (such as killing the infected cell). But it is meaningful (and fruitful) to ask: *How* did the T-cell perform this act of "recognition"? At a biochemical level, we may answer that the host cell processed viral peptides and presented them on its surface, where the T-cell, on contacting the host cell, found non-host peptides and responded accordingly. But how did the T-cell "identify" the peptides? By means of certain complex protein-binding interactions, which can only be explained in chemical terms such as van-der-Waals bonds. And of course we must then ask how such bonds exist, which involves explanations in terms of physics (viz, quantum mechanics).

Now, we can discuss causality purely at the biological level; but if we want to *completely* explain the phenomenon (in accordance with the Aristotelean notion of causality as explanation), we must invoke these deeper levels of explanation from chemistry and physics. Metaphysically, the biological phenomenon actually occurs *because* of the chemical and physical causes--the connection is there in reality.

I take epistemological anti-reductionism to be simply a matter of convenience. When we are talking about biological phenomena as such, explanations in terms of van-der-Waals bonds are hopelessly inefficient. We construct biological concepts for conceptual efficacy in thinking about biology. But this does not mean we cut off the connections between conceptual levels. Biological "recognition" is explained by (and caused by) biochemical "binding," which in turn is explained by (and caused by) chemical "bonding," which in turn is explained by (and caused by) physical "forces."

Ron Merrill

ronmerrill@bix.com