In Defense of Rationalist Epistemology

Below is a compilation of material I have written up in various contexts in my theology and philosophy classes defending a rationalist (or broad empiricist) over against an empiricist (or narrow empiricist) epistemology.  (For those more familiar with the theological meanings of these terms than the philosophical ones, I am using the words in a more philosophical sense.  Basically, the difference has to do with whether or not the logical analysis of concepts is a source of knowledge in addition to more direct sense experience.)  See here and here for more on this subject.

"How can the analysis of concepts give us categorical/universal knowledge of reality?"

Our concepts are derived from our observations and constitute the very foundation of all our ideas.  Reality itself is a concept derived from our observations.  So when we examine concepts to see what belongs to their essences, we can make universal/categorical claims regarding them.  For example, we can know that all cats are mammals (even cats we will never have particular experiential contact with), because the essential definition of a “cat” involves being a mammal.  If it isn’t a mammal, it isn’t a cat by definition.  Or a “triangle,” by definition, has three sides.  If a shape doesn’t have three sides, it’s not a triangle.  So we can know that all triangles everywhere will always have three sides.  Or we can look at the concept of “being” and see that it excludes “non-being” as part of its essential definition.  Therefore, we can know that all beings will exist and not not-exist, and they will be what they are and not what they are not.  This is the law of non-contradiction, which is the basic essence of logic, and so we can know that all reality is logical in nature.

"How would you respond to the objection that, when we analyze concepts, we are really only learning about things in our minds and not things in the real world?"

The distinction between our concepts and “the real world” is an illusory one, because reality itself is a concept derived from our observations/experience.  So when we talk about “reality,” or “real things,” or “the real world,” or whatever other synonyms we want to use, we are bound to the concept of reality.  It is that concept that we are talking about.  To wonder if “the real world” matches our fundamental concept of reality is like wondering if there are red things out there somewhere which aren’t red.  It’s meaningless gibberish.  There are lots of things about reality that we don’t know, but we do know that reality will be like reality.

"According to you, what was Kant’s fundamental error?"

Kant’s fundamental error was that he made an illusory distinction between “the real world” and the concept of reality, thus binding himself to the erroneous impression that all our knowledge relating to our concepts has nothing to do with “the real world” outside of our concepts.  This mistake bound him to the very skepticism he was hoping to escape.

"Rationalistic arguments are nothing more than an analysis of concepts in our minds, without checking that analysis against the real world as encountered in the empirical data, and so they are nothing but word games or mind games and tell us nothing about reality.”

There is a fundamental difference between “playing word games” and making a logical analysis of concepts.  Here’s a good example of a word game:

A. No cat has eight tails.
B. One cat has one more tail than no cat.
C. Therefore, one cat has nine tails.

This argument is obviously fallacious.  It does indeed play games with words.  It uses equivocation—that is, it changes the meaning of words in the middle of the argument so as to reach a fallacious conclusion.

But what I am doing in my arguments for God is something fundamentally different.  I am examining concepts derived from experience and exploring their logical implications.  It might be objected that concepts are not the same as objective facts in the world, and so the examination of concepts is nothing other than examining things that aren’t real.  But this objection fails to grasp what concepts really are.  They are categorizations based in our experience, and so long as we are using words that relate to concepts and have definite meaning, we are bound to the meaning of the concepts and words we make use of.

For example, if someone asks whether or not an infinite past is possible, well, this question refers to words and concepts, like “infinite” and “past."  These words have definite meanings.  Without such definite meanings, language is impossible.  It turns out that the essential idea of “infinite” includes within it the idea of “untraversability."  That is, an infinite cannot be “got through."  That’s just part of its meaning, like “having three sides” is part of the definition of “triangle."  And “past” refers to “that part of time that has already happened, that has already been got through.”  So the concepts of “infinite” and “past” are inherently contradictory.  These two concepts inherently exclude each other.  Does that mean that an infinite past cannot exist?  Yes, because “existence” is also a concept.  “Reality,” “being,” existing thing,” etc., are all synonyms of the same basic concept, and it has definite meaning.  For one thing, inherent in the idea of “being” is that it excludes “non-being."  Every particular claim of being inherently excludes its opposite.  So “red," for example, excludes “non-red," “up” excludes “down," “can fly” excludes “can’t fly," etc.  To use language with meaning, our propositions must exclude their opposites.  So logic—A is not non-A and therefore excludes non-A—is an essential component of “being," and so contradictions cannot “exist."

This is not a word game.  This is simply to use words meaningfully and to not equivocate in our ideas and arguments.  Sure, we can redefine words if we want to.  You are free to redefine “infinite” to mean “pickle” if you want to.  But you’d better make that clear to others if you want to have a productive conversation, because that’s not what it usually means.  It is not a word game to hold people to the meanings inherent in their own words and concepts.  If someone wants to consider the possibility of an infinite past, they have to deal with the fact that “infinite past” is a contradiction.  If they want to assert that contradictions can exist, they have to deal with the fact that “being” excludes “non-being” and that therefore “contradiction” is incompatible with “existence."  If someone wants to assert that there may be small, red, reptilian, egg-laying kangaroos in the Andromeda Galaxy, they have to face the fact that “small, red, reptilian, egg-laying” excludes “kangaroo," and therefore their proposition is meaningless and absurd.  If someone wants to assert the existence of four-sided triangles at the heart of the planet Mars, they will have to deal with the reality that “four-sided” and “triangle” exclude each other and that therefore a “four-sided triangle” is meaningless gibberish and can’t exist.  If they wish to keep on using these concepts, they are necessarily bound to their essential meanings.

Ironically, it is the empiricist who feels himself unbound to the meanings of his own words and concepts whose argumentation most resembles the “word game” example I gave above.  It is not I, who insist that we remember what we mean by “infinite” and “past” when we discuss the possibility of an infinite past, but the empiricist who feels free to ignore the inherent meanings of his own words and ideas and to draw conclusions based on ignoring these meanings, who is committing the “word game” fallacy of equivocation.

Therefore, logical argumentation rooted in analysis of concepts is a perfectly valid way of learning about “reality”—which is also a concept.  So arguments of this type cannot simply be dismissed on the grounds that they are not sufficiently “empirical” in nature.  The argument, “It is inherent in the nature of a triangle that it has three sides, and therefore all triangles have three sides wherever they may be found” is a different sort of argument from “sea turtles lay eggs because I’ve seen them doing it," but both are valid, and the former argument cannot be rejected merely on the ground that it is not the latter sort of argument.  I have just as much right—that is, no right at all—to dismiss empirical arguments merely on the grounds that they aren’t “abstract” or “rationalistic” enough.

In fact, the validity of empirical argumentation depends on the validity of rationalistic argumentation.  If we cannot make universal, categorical claims—even something so basic as “A always excludes non-A”—based on logical analysis of concepts, then no empirical claim could ever have any foundation or even any meaning.  Every proposition is only valid to the extent that it excludes its opposite.  If “There is an apple on the table” does not exclude “There is not an apple on the table” (assuming we’re not equivocating on the meaning of the words), then it has no meaning at all and therefore can establish no knowledge.  But you can’t prove the universal validity of logic empirically (that is, by simply observing it with your senses).  Rather, logic is a precondition for even being able to derive any knowledge from one’s senses or to form any propositions rooted in the evidence of one’s senses.  So if you reject the rationalistic method of reasoning as invalid, you must, to be consistent, reject all claims whatsoever as ungrounded and meaningless.  To put it bluntly, the empiricist's epistemology, to the extent that there is any validity in it at all, is parasitical upon rationalism.  And when the empiricist rejects rationalist epistemology, he rejects any possibility of any validity in his empirical claims.

"I understand where you are coming from when you say that it seems obvious that, by definition, you can’t get something from nothing.  But, still, I don’t see how we can know, without empirical confirmation, whether our idea of nothing corresponds with reality.  We might think that 'nothing' implies certain things, but reality often surprises us, and we have to correct our concepts and definitions in response to new things we learn through empirical investigation.  For example, perhaps you have seen many cats, and you reasonably come to the conclusion that ‘cat’ involves the idea of ‘having fur.’  But then, after many years, you encounter your first hairless cat, and you have to adjust your definition to accommodate your new empirical experience.  It would be unreasonable to say that it is impossible that there could ever be a hairless cat simply because you’ve decided to define the word 'cat' in a certain way based on your limited experience.  Our ideas are always provisional and could be corrected by further evidence.  We can never claim anything as an absolute.  And therefore, when it comes to matters of ultimate reality outside the realm of our experience, an appropriate recognition of our epistemic limitations demands that we be agnostic."

I certainly agree that we can learn new things about reality.  And I agree that we might sometimes adjust the use of our terminology to better accommodate reality when we have learned something new about it.  The example of the hairless cat is a good one, potentially (though I’m not sure if there are any cats that are literally completely hairless, strictly speaking).  If you didn’t know about hairless cats before, you might have included in your definition of a cat the idea of “having hair.”  Using this definition, a hairless cat is not a cat.  But in official taxonomy, hairless cats are classified as real cats because they are the same species, the same branch of the evolutionary tree as it were, even though they have been bred to be unusual in the matter of hair.  It makes more taxonomic sense to classify hairless cats as cats rather than exclude them simply because of this unusual trait.  Of course, words and labels are a human invention, and we can categorize things however we like, so long as we don’t misrepresent what we are categorizing.  If we wanted to exclude hairless cats from the cat family and call them by some other name, we could.  It wouldn’t be objectively wrong, simply less taxonomically convenient and efficient.

In some cases, such as when labeling empirical creatures that we don’t necessarily know everything about, we might find it useful to shift our definitions from time to time in light of new information that causes us to revise our taxonomies to make them more efficient.  This is not likely to happen with things we know through-and-through - such as, for example, a triangle.  We’re not going to learn something new about triangles in the future that will cause us to want to reclassify them in some other way.  A triangle, unlike a cat, is a pure, abstract concept which we know thoroughly, so we will not have occasion to find some new kind of unknown triangle, as can be the case with animals we are continually learning more about.

But in either case, none of this has any effect on the validity of the rationalist epistemology.  Rationalism does not deny the fact that we can learn more about the world through empirical means, nor does it deny that sometimes it makes sense to revise our categories and definitions in order to better accommodate what we know.  We can do these things, but what we can’t do is ignore or forget the meaning of a word or a concept at a particular given time while we are using it.  Although we can change the meanings of our words, at any given time our words have particular meanings, and we are bound to the meanings of our words when we are using them unless we will be speaking utter gibberish.  Sometimes a definition can be open-ended to some degree.  (“Reptiles are cold-blooded, at least all the ones we’ve come across so far.”)  Sometimes exceptions can be built into a definition.  (“Mammals give birth to live young, except for duck-billed platypuses.”)  But still, while we are using a word, we are bound to its current meaning.  The fact is that the word and concept of reality (and all its synonyms), as it is used by people in general, includes in it essentially the idea of the exclusion of non-being.  If something exists, it doesn’t not exist.  If something has certain characteristics, it does not not have those characteristics, and it does not have other characteristics that exclude those characteristics.  (“Thngs are what they are and they aren’t what they aren’t,” as I like to put it.)  All of that is clearly included in the concept of reality as the word is ordinarily used.  And that is the law of non-contradiction, or logic.  So reality, by definition, is logical.  And that’s all we need to be able to make metaphysical arguments such as those commonly used in arguments for the existence of God.  So rationalism, and its usefulness in learning important facts about reality, is established.

Perhaps someone will say, “I want to give a different meaning to the concept of reality.”  Well, OK, but just make it clear how you are defining the term.  And don’t expect to win an argument simply by redefining words.  That’s the fallacy of equivocation.  For example, if you and I are arguing about whether or not there is an apple on the table, and, in the middle of the argument, you suddenly change the meaning of “apple” to have the meaning usually assigned to the word “mug,” and then you declare that there is an apple on the table because there is a mug on the table, you have not really proved your case.  You’ve cheated by redefining the word in the middle of the argument.  Similarly, if you want to redefine reality to mean something different from how pretty much everyone else uses it, fine, but make it clear what your new definition is, and realize that your new definition is probably going to be irrelevant to the conversation at hand which is about reality in the ordinary sense.  I claim that all reality is logical, using reality in the ordinary sense.  If you say, “Well, I’m defining ‘reality’ to mean ‘something that both exists and doesn’t exist at the same time and in the same way,’ and in that definition, reality is not logical,” you’re right, reality in your definition is not logical.  But reality in your definition also doesn’t exist in the ordinary sense of that word, because, in the ordinary meaning of exist, to exist includes not not existing.  So everything that we see around us that exists in the ordinary sense of that word isn’t reality in your new definition.  So your new definition is irrelevant to my claim that all reality (in the ordinary sense) is logical, just like redefining “apple” to mean “mug” doesn’t help at all in the argument about whether or not there is an apple (in the ordinary sense) on the table.

As I’ve mentioned in class, to talk about the difference between empiricism and rationalism, I like to use the analogy of receiving a package in the mail.  The empiricists are like a person who receives a package but doesn’t open it, content with just receiving it.  The rationalists are like a person who opens the package and explores its content.  We gain our concepts from our observations of reality, and then we learn more about reality by unpacking those concepts and their logical implications.  That’s how rationalism works.  Empiricism’s failure to unpack concepts to learn more is why empiricism is an impoverished epistemology that misses the most important things about the real world.

Published on the feast of St. Mary Magdalene.