|
|
|
|
|
Naturalism and Reason
The naturalistic approach to reason requires that the existence of reason, and the ability of humans to be rational, must be compatible with scientific knowledge about humans and nature. Supernaturalistic approaches to reason make the rival claim that supernatural explanations are required for explaining the existence of reason and the ability of humans to be rational. Is the supernatural necessary for reason? There are three primary arguments for a necessary connection between religion and reason.
The naturalist replies to each of these arguments that reason and its usage is entirely compatible with, and indeed only explainable by, scientific knowledge about humans and nature. Before examining these arguments, we must be clear about the nature of reason that is under scrutiny in these arguments. Reason is a term that covers the three methods of logical inference. Reasoning is the use of these methods of logical inference to draw conclusions about the truth, or the likelihood of truth, of some propositions from the consideration of reasons offered in support of those propositions. You can read more about the three methods of deduction, induction, and abduction in "Scientific Method". An example of the use of reasoning would be the conclusion that "Today is Monday" from two reasons given in support: "Yesterday was Sunday" and "If yesterday was Sunday then today is Monday". The truth that reason has discerned in this example in "Today is Monday". The principle of reason that is used in this example has this logical form: "If P, and if P then Q, then Q". Because the arguments for supernaturalism considered here depend upon reference to certain allegedly absolute truths and absolutely true principles of reason, the only method of logical inference to be discussed is deduction. Deduction is the only method which is capable of discerning conclusions that must certainly be true so long as the reasons given in support are certainly true. Furthermore, there is a long philosophical tradition which holds that the principles of deductive reason are themselves certainly and absolutely true. Hereafter, when reason is discussed, the method of deduction is meant, unless other methods are specified.
1. What is the Best Explanation for Truths of Reason? Let's look more closely at these three arguments, beginning with "The truth of absolute rational truths requires the existence of a supernatural reality to explain their truth." Consider the following argument, which we can call the "Argument from Non-Natural Truths":
Starting with premise 1, what are these allegedly absolute truths? Examples of absolute rational truths are usually taken from mathematics (the equation 2 + 2 = 4) or logic (the logical axiom If P, and if P then Q, then Q) or geometry (the Pythagorean theorem a² + b² = c²) or ordinary language (bachelors are unmarried adult men). These propositions are absolutely true, it is alleged, because no one who rightly understands these truths can conceive how they could be false for some people (anyone who denied their truth must either be insincere or demonstrating their inability to understand them), or how they could be false at some time (anyone who thinks that one of these truths could become false is not conceiving them correctly). It is very difficult to see how something that can change can be responsible for something universal and eternal. Premise 3 is therefore probably true, because there is nothing permanent about human beings (their bodies and minds keep changing) or human societies (they gradually change their moral standards over time) or human life on earth (survival strategies of the human species gradually change over time) or wider nature (which is always changing). Even if some laws of nature were thought to be absolute (F=MA or E=MC²), they don't seem like appropriate candidates for explaining truths of math or logic (indeed, formulating laws of nature requires math and logic, not the other way around). This argument suggests that we have to look beyond humans, human societies, and nature itself to explain absolute moral truths. What about premises 1? Premise 1 may be false. Naturalists do not agree among themselves about the existence of absolute rational truths. Some naturalists have accepted the truth of premises 1 and 2, and have tried the option of defending absolute rational truths by appealing to pure reason to justify absolute rational truths. Pure reason itself (the source of math, logic, etc.) would explain why there are absolute rational truths. However, the principles of pure reason had better be absolutely true also, and then the naturalist has to explain the absolute truths of pure reason. This option leads to either having to defend a priori true principles quite alien to scientific method (and a naturalist should stick close to scientific method), or it potentially leads right back to a supernatural reality needed to explain those truths. See the second argument about principles of reason considered in the next section. Some naturalists accept premise 1 but reject premise 2. Why should premise 2 be correct? For ordinary propositions such as "the door is open", when this proposition is true, it is because there is an actually existing door that happens to actually be open. It is the relationship between the open door and the proposition (often called a "correspondence" relationship) that makes the proposition true. The open door is responsible for making "the door is open" true; furthermore, the closing of that door is what is responsible for making that proposition false. The implication of premise two is that something other than the truth itself is responsible for making that truth true. What is the truth itself? These absolute truths under consideration here are composed of some meaningful terms set into certain relations with each other. Once a person is capable of correctly understanding the meanings of the terms and the relations into which they are placed, then the truth of the assembled proposition simply shines forth immediately. And maybe that's all that is needed to explain why these truths are absolutely true. Nothing other than the structure of the truth is needed to explain its truth. Indeed, on this theory of absolute truths, that is what makes these truths absolute: we don't try to compare these truths with any other reality to see if they are true: actual reality is entirely irrelevant to their truth (and that is why reality can't make them ever false, as well). If this view of absolute truths is correct, premise 2 would be false. On this theory of absolute truths, they are apprehended by the imagination and can be considered in their rational purity within the imagination, and also applied in empirical investigation (as when a law of nature is given mathematical form). But their truth never depends on their applicability in any empirical inquiry and never depends on science. There are a potentially infinite number of absolute truths (different systems of math, logic, geometry, definitions, etc.), and any number of them may be found to be useful in science at some time or another. If some geometry is found to be irrelevant for measuring the earth (for example, the flat geometry of Euclid is inadequate for mapping large areas of the earth's surface), that fact has no bearing on the truths of that geometry. No geometry (e.g. no mathematics, no logic) is truer than any other. The labels of "analytic" and "a priori" and sometimes "analytic a priori" have been applied to these truths. You can read more about the "Analytic/Synthetic Distinction", "A Priori Justification and Knowledge", and "Logical Truth". Other naturalists reject premise 1, denying that there are any absolute rational truths. They attempt to treat reason, both its methods of inferences and the truths reason discerns, as a type of practical reliable knowledge. These naturalists are radical fallibilists: there is no truth that may someday require revision or abandonment in light of some additional empirical evidence or novel scientific theorizing. In the paragraph above, we gave good reasons why fresh empirical evidence is probably unable to ever supply a counter-example to these seemingly absolute truths. For example, measuring actual triangles will never disprove the Pythagorean theorem, since the theorem is only about perfect right triangles that do not naturally exist. Some radical fallibilists have proposed that, since anyone could make a mathematical or logical error, no one should feel absolutely certain about any seemingly absolute truth. This notion presupposes that every estimate of a proposition's truth is ultimately a matter of mental inference, but for the simpler absolute truths, ordinary minds already familiar with their meanings are quite capable of immediately apprehending their truth without inference. Other radical fallibilists have pointed out that since the meanings of terms gradually drift over time, as languages evolve and populations shift their usage of terms over centuries, then a true proposition can become false. However, this suggestion ignores the point that an absolute truth is composed of certain fixed meaning in certain set relations: to show that meanings shift only succeeds in showing that people can stop believing proposition and start believing another, not that a single proposition has shifted truth-value over time. The most interesting radical fallibilists grasp that no empirical evidence can directly disprove any truth of math, logic, etc., but rather claim that propositions of math, logic, etc. which are used in scientific theorizing become themselves a core part of such theorizing. If a scientific revolution requires the substitution of one set of core truths of math or logic for another set, so that the new scientific theory is a better fit with the empirical evidence, then the evidence has falsified the old core truths and confirmed the new set. Some philosophers believe that this has happened recently, where classical Boolean logic had to be replaced by a modified "Quantum Logic". But it has long been obvious that ever more sophisticated systems of mathematics are required for progress in physics and cosmology, since the days of Newton and calculus. Interestingly, a few naturalists who view math and logic as core components of scientific theories have concluded that a realistic stance must be taken towards such useful mathematical and logical entities, just the same way that scientific method takes a realistic stance towards any other postulated entities of successful theories. Essentially, these naturalists claim, if electrons probably exist, then so do sets and numbers, since they are also indispensible for scientific knowledge. This line of thought can lead towards supernaturalism, since such entities do not have any natural properties (do sets have causal effects? do numbers have mass?). Platonic realism is an exemplary type of this supernaturalism (see "Platonism in Metaphysics"). Alternatively, mathematical and logical naturalists might instead claim that our conception of what is natural simply needs to be broadened to accommodate such entities. You can read more about "Indispensibility Arguments in Philosophy of Mathematics". What sort of naturalism could be so accommodating? Reductive Materialism is probably not; reductivism is more suited to dealing with entities that have only narrowly physical properties and lawfully causal relationships (what causal powers do numbers have?). Perspectival Pluralism and especially Synoptic Pluralism are much more flexible. You can read about these naturalisms at "Naturalism and Science"). To summarize the results from considering the "Argument from Non-Natural Truths," there are various ways for naturalism to counter-argue that nothing supernatural is ever required to account for the truth of alleged absolute propositions. Either there are no such absolute rational truths, or such truths are simply true by themselves without any other reality needed to explain their truth, or nature itself is broad enough to accommodate some mathematical, logical, etc., entities.
2. Can Naturalism Accommodate Rational Principles? The supernaturalist can respond to the naturalist's counter-arguments considered above, by refocusing scrutiny on the foundations of reason and their relationship to naturalism. Consider the following argument, which we can call the "Argument from Non-Natural Rational Principles":
This argument tries to expose the naturalist to a terrible dilemma. According to this argument, knowledge of the rational principles required for any knowledge of nature cannot themselves be part of nature. Therefore, if (a) there is knowledge of rational principles, then we know that something beyond nature exists (naturalism is incomplete), or (b) if there is no knowledge of rational principles, then there is no knowledge of nature (naturalism is false). Either way, naturalism cannot stand. How can the naturalist respond to this "Argument from Non-Natural Rational Principles"? Most naturalists agree that premises 1-4 are true (Synoptic Pluralism alone could create an alternative to premise 4). In particular, premise 3 is correct for a naturalist because most of our knowledge of nature is acquired and justified through the application of rational principles (methods of logical inference, mathematical formulas, geometric axioms, etc.) to past, present, and future experience. A naturalist who denied that any rational principles are involved would regard knowledge of nature simply as the recording of factual data from each passing moment's phenomenal experiences, but humans are obviously capable of far more extensive knowledge of nature, and no modern naturalist would throw away modern science. Naturalists can attempt to deny premise 5, or premise 6, or both. Premise 5 is much harder to deny. Premise 5 would be false if knowledge of nature could use only rational principles which do not have necessary validity. For example, while deductive principles have necessary validity, could knowledge of nature be gained without deduction, using only induction and deduction? Positivists have promised such knowledge of nature, but never delivered, while modern experimental science has reached amazing successes using all three methods of inference and deductive mathematics too (see "Scientific Method"). Naturalism does not get far without deduction. Premise 5 would be false if rational principles could have necessary validity while knowledge of them does not. In other words, the property of necessary validity applies to things like principles, but perhaps not to knowings of them as well. Analogously, if I see a red tomato and thereby know that "There is a red tomato," I see that the tomato has the property of redness, but my knowing "There is a red tomato" does not also have the property of redness. This analogy is not really helpful here, though. In defense of premise 5, we should see that its intent is to guarantee that our deductive inferences have necessary validity -- if we didn't know this, would our use and dependency on them be genuinely deductive? Whether our knowledge is justified using a deductive inference probably depends, in part, on whether we are knowingly using that inference deductively. Some naturalists may disagree, especially reliabilists favoring externalist justification (for reliabilism, your knowledge can originate in an appropriately reliable process, even if you don't know about the process). However, if reliabilists thus locate necessary validity in external nature instead of internal knowledge, then they will have to deal with premise 6 too, like all naturalists. Premise 6 is a specific version of the long-held complaint that nothing in the natural world as described by naturalism could have normativity. Necessary validity is a kind of normativity -- a necessarily valid proposition prescribes what normatively is true under certain conditions. For example, "If P, and if P then Q, then Q" (the 'modus ponens' form of valid inference) is a proposition that prescribes that Q is true when 'P' and 'if P then Q' are true. According to premise 6, this normative relationship is quite unlike, and not reducible to, any natural entity, state, function, or process. And it does seem obvious that none of these natural things, by themselves, can prescribe a normative relationship. The 'is' of their existence entails no 'oughts'. This seems disastrous for naturalism. Yet there is another construal of normativity that is more easily handled by naturalism. Consider again the proposition "If P, and if P then Q, then Q". This proposition can be understood as normative in a somewhat different way, as a normative prescription about belief under certain circumstances: If you believe that 'P', and you believe that 'If P then Q', then you ought to believe 'Q'. Notice that this "belief normativity" interpretation of this proposition (as opposed to the first, "truth normativity" interpretation) is not about how anyone actually forms relationships between these three beliefs -- even if many people fail to believe Q under such circumstances (they don't bother to follow out the inference, they don't want to believe Q, etc.), this is irrelevant to the necessary validity of "If P, and if P then Q, then Q". Modus ponens is not made normatively valid by the actual ways that people draw inferences, nor is modus ponens a description of the statistically average way that people actually draw inferences. This looks bad for naturalism, as it could be immediately concluded that nothing about people, individually or collectively, explains the normativity of "If P, and if P then Q, then Q". However, we should not be deceived by this hasty conclusion. What can the beliefs and behaviors of people have to do with the normativity of propositions? We have already decided that the normativity of "If P, and if P then Q, then Q" cannot reside in the actual way that people make beliefs. Why are we nonetheless so committed to the necessary validity of modus ponens? Is there a naturalistic explanation for this firm commitment? We do believe that everyone should form beliefs in conformity with modus ponens, even if they don't always do so. This fact explains the "truth normativity" interpretation -- we are certain that Q is true when 'P' and 'if P then Q' are true, precisely because we are certain that everyone should form beliefs in conformity with modus ponens. After all, if we were not committed to modus ponens as a rule for belief formation, we would not believe that Q must be true when 'P' and 'if P then Q' are true. A version of Tarski's semantic treatment of truth (see "Tarski's Truth Definitions" and "The Deflationary Theory of Truth") applies here, to commitments about normative validity. Let us stipulate that: "If P, and if P then Q, then Q" is necessarily valid if, and only if, one should believe 'Q' when one believes 'P' and 'If P then Q'. On this "commitment theory of validity", necessary validity is "deflated" to actual human commitment about belief formation. Where in the natural world is this sort of belief commitment? The obvious answer is that modus ponens belief commitment is found in various human communities that sustain such commitment, using the human practices of instruction and enforcement. In such communities (let us call them "logical communities"), humans teach reasoning principles such as the modus ponens rule to their young and thereafter supply regular reinforcement of its proper use, so that new adults are committed to its necessary validity, who in turn pass on these practices to the next generation, etc., etc. The necessary validity of modus ponens, it must be stressed, is not found in the actual processes of belief formation in these communities -- even in a highly logical community, departures from modus ponens will occur. Rather, the necessary validity of modus ponens is found in the commitment of the logical community's members to modus ponens, as each member firmly believes that a person should believe 'Q' when one believes 'P' and 'If P then Q'. This commitment is no abstract unnatural sort of thing -- commitment simply is that community's practices of teaching, enforcing, and applying the modus ponens rule, and these are all entirely natural processes. Premise 6 is therefore false. This "commitment theory of validity" locates the necessary validity of a principle like modus ponens in the practices of a logical community. Various objections to this sort of theory have been raised, too many to cover here. However, discussion of three of the most serious objections can strengthen the plausibility of this commitment theory of validity. Objection One: I can imagine a time when no logical communities exist (a time before intelligent life exists, or a time when intelligent life has lost its sanity, etc.), but I believe that even at such times modus ponens would still have necessary validity. Therefore, validity cannot depend on whether logical communities exist, so the "commitment theory of validity" is disproven. Naturalist Reply: Yes, of course you believe that modus ponens would still have necessary validity -- you are, after all, evidently a member of a logical community which endorses modus ponens and so you are displaying just the sort of commitment to modus ponens that you would have. And you cannot retort that you can go further to imagine your own non-existence, or your own insanity, and still believe that "If P, and if P then Q, then Q" is necessarily valid -- after all, so long as you are using your mind to imagine such things, the logical community is sustained. Objection Two: I can imagine a situation where a variety of logical communities co-exist, each using distinct logical systems so that there is no wide agreement about whether modus ponens is valid -- yet I still believe that modus ponens is universally valid for all people, regardless of whatever other so-called "logical communities" may exist. Therefore, validity cannot be relativized into merely what logical communities happen to be doing, so the "commitment theory of validity" is disproven. Naturalist Reply: Yes, of course you believe that modus ponens would still have universal necessary validity despite the existence of rival logical communities-- you are, after all, evidently a member of a logical community that uses modus ponens, and that's just the sort of commitment to modus ponens that you would have. And you cannot retort that you cannot conceive how you could switch logical communities, as the commitment theory of validity seems to suggest. Your inability to conceive how you could function in a different logical community that doesn't use modus ponens, just because you are so committed to modus ponens yourself, only reveals further what it is like to be a member of a particular logical community. Objection Three: I never was much committed to modus ponens -- I recall struggling with the notion in an undergrad logic class years ago -- and I don't fault others for their occasional departures from modus ponens, either. But I still am a member of a logical community, since I am frequently logical and prefer that others are too. Therefore, some logical communities don't need to use necessary validity, so the "commitment theory of validity" is disproven. Naturalist Reply: The commitment theory of validity embraces the notion that logical commitment can come in degrees -- that is why empirical investigations into peoples' actual inference patterns is so interesting. Logical communities emerge, grow, and change over long periods of time. Human knowledge of math, logic, etc., has been growing and evolving right along with scientific knowledge, and that is the core theme of any naturalistic understanding of reason. This historicist and evolutionary view of reason is especially needed for naturalism in order to answer a serious question about the foundations of reason. How can the naturalist respond to the following dilemma? If humans have judged the cognitive value of various reasoning systems over time, and have improved them in the course of increasing our knowledge of nature, then either (1) these human evaluations and improvements have been made using some independent rational standard, or (2) they have not. If (1), then the naturalist must admit that humans use some rational criteria that have some non-natural origin/foundation, and so naturalism is incomplete. If (2), then naturalism is irrational because its reasoning systems are ultimately grounded on standards lacking cognitive justification. Either way, naturalism is proven inadequate. The naturalist must reply to this dilemma as follows: Reasoning methods and rational systems have been imaginatively created out of earlier and simpler modes of inference, step by step across millennia of human cultural evolution, ultimately tracing back to primitive inference habits shared among the higher mammals. For humans, the "origins" of reasoning were never the ultimate justification for reasoning -- where reason came from is not hardly as significant as where reason is going. Human reasoning habits were developed gradually through the continual test of human practical survival and welfare, resulting in the "common sense" of adults who instill it in children. Modern science is simply an extension and refinement of these common reasoning tools (you can read more about "Scientific Method"), and modern logical and mathematical systems have been tested against our ability to predict and control nature. Our rational principles are tools: now so highly reliable and useful that adults simply take them as certain and necessary (and teach them as such to the next generation). Only scientists working at the cutting edge of novel knowledge ever need question, much less refine further, these systems. This naturalistic account of the cultural evolution of reason is entirely pragmatic, and the non-naturalist might now complain that an ultimate standard has been offered in this account: the pragmatic standard of utility. But there is nothing unnatural about the pragmatic standard of utility -- humans seeking practical accommodation to, or modification of, the environment is about as natural (and reasonable) as it gets. Summarizing, the naturalist can deny premise 6 in order to deal with the "Argument from Non-Natural Rational Principles". The belief commitments, along with the resulting behaviors, of human communities is sufficient to explain why why certain principles possess the normativity of necessary validity. Because "beliefs" are typically understood as "mental" in nature, the non-naturalist may complain at this stage that the naturalist's appeal to belief commitments can't help naturalism, since mental entities like beliefs cannot be natural (but you may now proceed to "Naturalism and the Mind"). This "Argument from Non-Natural Rational Principles," even if successful, does not directly support any type of supernaturalism. How could the naturalist's dilemma help supernaturalism? The supernaturalist is rarely a skeptic about knowledge of nature (supernaturalism need not reject science's knowledge). Rather, the typical supernaturalist agrees with the naturalist that knowledge of nature is possible, and then uses the above dilemma to argue that naturalism is incomplete. This move by itself cannot establish supernaturalism either, but the following supplementary argument can try:
How can supernaturalism offer an explanation of human knowledge of ultimate rational principles? Three main explanations have been developed by theistic theologies: (a) God directly implants such knowledge into humans, (b) God creates humans with the capacity to gain such knowledge using our reflective minds alone, or (c) God created the world with a design that establishes such knowledge for human investigators. Each of these explanations suffers from serious difficulties. Regarding (a), placing responsibility on God only generates a version of the Euthyphro dilemma: Is something necessarily valid because God declares it to be necessarily valid, or does God declare something necessarily valid because it is necessarily valid? (details of this dilemma can be found at "Naturalism and Morality"). The first horn of this dilemma presupposes human knowledge of God (but now we are going in logical circles, since knowledge of God needs rational justification), while the second horn of the dilemma presupposes that we humans can independently know what is necessarily valid apart from God, making God irrelevant. Regarding (b), if we can use our reflective minds to intuitively know ultimate rational principles, why do humans disagree over what counts as such "innate" knowledge, and why do humans have to first learn meanings and concepts in order to recognize and accept such innate knowledge (you can read more about "Intuition and Innate Knowledge"). Regarding (c), the supernaturalist has already concluded by the "Argument from Non-Natural Rational Principles" that knowledge of nature cannot justify knowledge of ultimate rational principles -- who or what created nature is irrelevant. In conclusion, naturalism offers a far superior account of rational principles than any non-naturalism or supernaturalism. There is no "First Philosophy" of epistemological principles that must transcendentally hover above and beyond the natural world.
3. Can Naturalism Explain the Human Ability to Reason? The supernaturalist can pick up where the "Argument from Non-Natural Rational Principles" leaves off: What precisely is the capacity of humans to be rational. According to the "commitment theory of validity," humans acquire commitments to necessarily valid principles by way of membership in a logical community. But consider the following argument:
This "Argument from Rational Perfection" claims that natural processes cannot explain how so many people do frequently succeed in reasoning perfectly in accord with rational principles. The supernaturalist would defend the truth of premise 2 by pointing out that human reasoning cannot be expected to be perfect since it is a highly complex and sophisticated series of processes -- even the naturalist admits this -- which does not look anything like a natural law. Naturalism's appeal to the mathematical perfection of natural laws cannot help here. Natural laws, taken singly, can often be fairly simple and mathematical, but following a rational principle is nothing like following a natural law. People naturally obey the law of gravity perfectly, but they do not naturally obey the law of non-contradiction or the rules of mathematics to any high degree of perfection. If people construct and perpetuate rational systems in logical communities, as the naturalist claims, then reasoning is far more like engineering than falling down. When people do engineering, perfection rarely results, as one would expect. Even after centuries of engineering progress, human constructions only approximate the ideals of engineering perfection. Likewise, from the naturalistic standpoint, actual human reasoning can only approximate the ideals of rational perfection. If the naturalist is right, any people who actually do succeed in reasoning perfectly, and there couldn't be many of them, are simply quite lucky. If premise 3 is correct as well, then naturalism is unable to fully explain rational perfection, and the supernaturalist's conclusion follows: humans must have some supernatural aspect that can account for their rationality. Should the naturalist reject premise 2 or premise 3 (or both)? The sensible tactic is to admit that people do not always reason perfectly (so that premise 2 is partly right), and to claim that people do succeed in reasoning perfectly, but not as often as we would wish (so that premise 3 is partly right. The actual situation of us human reasoners is somewhere is the middle. We do frequently succeed in reasoning perfectly (especially when the reasoning task is not too complicated), and we often fail to reason perfectly (especially when the reasoning task is complicated). Naturalists and supernaturalists can agree on this. Is naturalism unable to explain our mixed reasoning performance? No, premise 4 is false. Our mixed reasoning performance is pretty much what would be expected if the naturalistic account of logical communities is correct. On the other hand, supernaturalism has no easy way to explain our human capacity for reasoning. No supernatural aspect to human mental processes is necessary for explaining how people do actually reason. Furthermore, if people had access to supernatural capacities for perfect reasoning, why don't we reason even better than we actually do? In conclusion, naturalism is not shown to be inadequate by the "Argument from Rational Perfection". Sometimes this argument is designed quite differently by supernaturalists to contain the premise, "Reasoning is a mental process that cannot have any naturalistic explanation". Because this different argument is really about the mental capacities of people, and not so much about reason itself, you may jump ahead to "Naturalism and the Mind".
copyright 2007 by John R. Shook |
