The following is a web adaptation of a paper I wrote for a seminar on epistemology. I think this idea has the potential to revolutionize epistemology in the 21st century - maybe I’ll write a better paper about it later.

The discipline of cybernetics, created in the 1940’s by Norbert Wiener, is the science of information processing systems. Recent developments in cybernetic technology have given us reason to re-evaluate the way we think about knowledge. This is because knowledge is certainly related in some way to information. However, I do not think it has been made clear in exactly what way they are related. When I sit down at a computer with access to the internet, I feel smarter; I am able to answer questions that I would never be able to otherwise, with very little effort on my part. One likely explanation of this phenomenon is that I am somehow more knowledgeable when I am sitting at a computer. But this does not fit into the usual conception of knowledge; I doubt that sitting down at a computer automatically changes any beliefs of mine, nor does it justify them. This leads me to think that there should be a theory of knowledge that allows for this kind of thing to be called knowledge. Generally, I believe that a claim to knowledge is a claim to a certain competence; it is commonly said that knowledge is power, and “power” is often used analogously with other types of competence. I intend to show that a claim to knowledge (in the sense of knowledge about the world) is a claim to have access to information.

Suppose someone asks me, “Do you know what time the movie is playing?” I respond, “Yes, it’s playing at 7:00.” In epistemology, many people would say that my claim to knowledge is justified in the event that “The movie is playing at 7:00” is a justified true belief. I will later show that this formulation is flawed, but for now the question I would like to ask is, just what steps am I allowed to take before answering? Let us examine some cases to find the answer to this question.

Now suppose I could not remember right away but had to think for a second or two before responding. I think we would still say that I know what time the movie is playing, even though I did not necessarily remember at the exact instant that the question was uttered. But did I know before or after I remembered? It is not obvious how the justified true belief model will answer this question, but more on this later.

Suppose I had the movie time written on a notepad in my pocket. Someone asks me, “Do you know what time the movie is playing?” and I answer “Yes,” look at my notebook, and say, “it’s playing at 7:00.” Was I still correct in claiming knowledge in this case? I believe I was, and this is certainly an ordinary manner of speaking.

Now suppose it was a PDA I looked at, where I had written the movie time. Does this change the case at all? Certainly when it comes to epistemology an electronic notepad should not be viewed any differently than one made of paper. And so I think it is at least obvious that the claim to knowledge in the PDA case is reasonable in the event that it is also reasonable in the notepad case.

Now suppose instead of my notepad, it was a newspaper with the movie times in my pocket. I hadn’t actually read the times yet, but I knew it had accurate times for the movie and that’s why I picked up the newspaper. In this case, it seems as though answering “Yes”, looking at the newspaper, and saying “it’s playing at 7:00” is at least as justified as the notepad case.

Now let’s combine the previous two examples. Suppose I took out my PDA, clicked on a bookmark to the theater’s website, and then read the time. I could conceivably do this even faster than remembering, and in terms of observable action it seems to be the same as using my PDA as a notepad, though I’m accessing someone else’s information, like I did in the newspaper case. Then when I answered, “yes” and took out my PDA and answered, “It’s playing at 7:00”, I was making a true statement about my knowledge, even though my knowledge preceded my accessing the information via the Internet. This suggests that the knowledge is actually my ability to access the information, not any particular propositional attitude I may have had about the movie theater.

So if knowledge is merely access to information, don’t we know everything knowable? In principle, we have access to every knowable piece of information, even though the process of access may take a great deal of time and effort on our part, and may even include the possibility of faulty results. However, this does not seem to match our usual use of the term, “to know.” If I had answered “Yes,” to the question and then started asking passersby until I learned when the movie was playing, I think we would ordinarily say that I was being untruthful in claiming knowledge. But our usual assertion that we do not in fact know everything can be addressed by appeal to contextualism in epistemology, as described by Keith DeRose.

According to DeRose, the context of utterance affects the truth conditions of knowledge attributions (DeRose, 111). Such factors as what has been said in the conversation so far and the importance of being right will affect whether “S knows that p” is true. This should not be too hard to accept – it is common to say that a proposition’s context must be fixed in order to determine its truth value. “It is raining” is true in some contexts, false in others, but once the context has been fixed it is either true or false. And so DeRose asserts that contextual factors about a claim to knowledge will affect its truth value.

If knowledge is defined as the ability to access information, surely some methods of access are better than others. In fact, some methods might be so bad that in a relevant sense they do not count as knowledge. Thus by appeal to contextualism we can see how depending on context we can set the bar for what counts as a good method of access to information. It matters how quick, easy, and reliable the method of access is (though this is not intended as an exhaustive list of the virtues of such methods). Thus, “I don’t know” is shorthand for “I do not have access to the information within acceptable parameters of ease, quickness, and reliability.”

This can be seen in DeRose’s bank cases (DeRose, 109). My wife asks me, “Do you know whether the bank will be open tomorrow?” and I answer, “Yes; it will be.” She then says, “Are you sure? A lot of banks are changing their hours, and it is really important that we get to the bank by tomorrow.” And I admit, “No, I guess I don’t know – let’s call the bank.”

This sequence of events can be explained in terms of standards for speed and accuracy (amongst other concerns) of information access. When I am first presented with the question, I of course assume that my memory of the bank’s hours should be a sufficient way of determining what the bank’s hours will be tomorrow. However, my wife points out that banks have been changing their hours, which suggests that my method is not as reliable as I thought. She also points out that it is very important that we are correct about the bank being open, which suggests that our standards for reliability in this case should be relatively high. Thus I admit that I do not know, which is simply a way of saying that my method of information access does not meet acceptable parameters for this particular job. I thus suggest that we call the bank, which is a different method which sacrifices ease and quickness in favor of reliability.

What, then, of justified true belief? I believe that this model of knowledge is lacking in many ways, but one of its main problems is the assumption that knowledge is propositional in nature. This is simply not the way that we normally talk about knowledge (at least when we’re not doing epistemology). Someone might ask me, “Do you know the phone number for the pizza restaurant?” and I would respond, “Yes, it’s 203-555-1234”. However, I would not likely find myself in a situation where I would report, “I know that the phone number for the pizza restaurant is 203-555-1234.”, at least I would only say such a thing in the context of talking about knowledge. Thus, one who claims to know does not have a proposition in mind, but is rather talking about information (a fact).

According to Wittgenstein, “’I know’ is supposed to express a relation, not between me and the sense of a proposition (like ‘I believe’) but between me and a fact” (Wittgenstein, 16). Thus, knowledge is not a type of belief. He goes on to say, “I know that a sick man is lying here” is “nonsense … And ‘I know that a sick man is lying here’ used in an unsuitable situation, seems not to be nonsense but rather seems matter-of-course … the words ‘I know that…’ are always in place where there is no doubt” (Wittgenstein, 15). This tells us that “I know that” statements, or to put it another way “S knows that p” statements are actually not the usual way of talking about knowledge. I contend that knowledge is not propositional, and statements of propositional knowledge are only constructed in the act of talking about knowledge. I will further examine the implication of this when I discuss Heidegger.

In the above case about the pizza restaurant, my knowledge entailed an ability to give the phone number when asked. This is a pragmatic way of looking at knowledge; the meaning of my claim to knowledge is composed entirely of its material consequences. (According to Peirce, the meaning of any term is the sum of all practical effects it might have (Peirce, 48).) In this way, any time I claim to know anything, I am promising that I can perform some action which would follow from the meaning of my knowledge claim.

So working under the pragmatic view, if it is true that I know that the color of the ball is red, how would I know that it is true that I know that the ball is red? Someone asks me, “What color is that ball?” and I answer, “Red”. Or to put it another way, someone asks me, “Do you know the color of that ball?” and I answer, “Yes; it is red.” The only consequence of any knowledge is that the knower can perform some action; thus, this is the entirety of what we mean by “knowledge”.

The idea that knowledge is performative and not propositional is not without precedent. Peirce associated all meaning with action, but there was a Neo-Confucian philosopher in China who specifically made the connection between knowledge and action. Ch’eng I explained that when one claims to know evil and still does evil things, this is not truly knowledge (Koller, 533). This is because evil is against the ends of the agent, and thus knowledge of evil entails acting in a way that is not evil. (This analysis should be familiar to those who have read Plato). Ch’eng I spoke of “true” knowledge, where true knowledge has a special relationship to action. True knowledge entails being able to act appropriately in a test of knowledge. Clearly, Ch’eng I was touching upon the same idea of knowledge that I put forward here.

Some critics of this view, including Armen Marsoobian, have raised the following objection: I am describing knowledge completely in terms of degree, but in effect I am denying any difference in kind (between the one who knows and the one who does not know). I believe this is not really a problem. It has become commonplace to assert that all differences in kind are merely the way in which we (humans; oneself) carve up the universe. The chair and the table are part of the same continuum of matter and space; at the particle level it is impossible to tell where one ends and the other begins. However, we make distinctions between them because it is useful. There are phenomena which emerge from the scattered motions of particles, and those phenomena which are relevant to us become distinctions. In this way, the continuum of the universe is seen as the multitude of distinctions with which we are familiar. This is what the Buddhists would call “interdependent arising”, or “dependent origination” (Koller, 233). It also may reflect the pragmatist concept of theory-laden observation; we need background theory to carve up the world and we might end up carving it up different ways with different background theory.

Let me give an example of how this applies to a “difference in kind”. We say that a woman is either pregnant or not pregnant, and one cannot be a little pregnant. However, when does this transition take place? There is a sperm which starts interacting with an egg, and some time later we have a fetus. But between there, the sperm enters the egg, the tail falls off, the membrane around the genetic material of the sperm breaks down, the genetic material of the sperm attaches to the genetic material of the egg, the genetic material starts multiplying, and the egg starts dividing. And in each of those steps, we can see many smaller steps which can be further divided. What we see is that pregnancy is a process which does not involve a sudden change from pregnant to not-pregnant. The distinction is merely one way of specifying the overall biological state of the woman, in a way that is convenient for us.

And so it is with knowledge. For there to be some sudden change from one-who-does-not-know to one-who-knows would mean that knowledge does not work like any other process in the world. Why should we expect knowledge to be anything other than a useful distinction amidst a continuum of possible distinctions?

Another objection that has been raised to my theory of knowledge is as follows: If I can know something simply by having access to a machine, then wouldn’t that make the machine an extension of the knower? I have no problem with this characterization, but this is an important issue in the field of cyberphilosophy (Okay, nobody uses that term anymore). But I do not think it is out of the ordinary to consider objects a part of ourselves. Objects like my clothes, my eyeglasses, the fillings in my teeth, or an artificial limb may seem just as much a part of myself as any living tissue. (Is non-living hair an object or a part of the body?) Once again, this seems to be a distinction that is made on the basis of its usefulness.

In The Origin of the Work of Art, Heidegger expresses this relationship between “equipment” and the person who works with it (Heidegger, 158). To Heidegger, equipment is a thing that is used for work and thus becomes a part of the self. He describes a pair of peasant shoes, which are even moreso a part of the peasant because she is not aware of them while she is working. As she works in them, one can look at the shoes and see the worker. And so it is not strange to think of a thing as an extension of the self. For more on this connection as it relates to computers, see Michael Heim’s “The Computer as Component: Heidegger and McLuhan”.

I believe there is something else that can be taken from Heidegger’s examination of the peasant shoes. Heidegger claims that when the shoes are not being used to work in the fields, an examination of the shoes will not show what the shoes are in truth. This is because what the shoes are is actually a thing that is used for work. This analysis relates to what I said earlier about propositional knowledge (by way of Wittgenstein): The times when we talk about knowledge in terms of propositions are not the times in which we are using knowledge; in these situations, it is not even properly knowledge. Only by examining the way knowledge is actually used (for example by answering a question) can we get at what knowledge actually is, and this supports my claim that knowledge is non-propositional by nature.

This theory of knowledge has some interesting implications. This explains why in philosophy of education we would like to say that it is more important to teach children how to think than to fill their heads with facts. According to my theory, helping a child to grow into a knowledgeable person is helping him learn how to find those facts. Making a child memorize the capitals of every African country is useless if the child will not be able to use this information when he is called upon to do so.

This also explains why people from the humanities are sought-after in Information Technology. In Information Technology and other computer-related fields, employers have found that they would prefer a college-educated person with a good background in the humanities to someone who has extensive job training and certification. This is because a good education helps one learn how to be knowledgeable – how to access and use information – unlike computer certification training which aims to teach specific information about computers (which will doubtless become obsolete in a matter of months).

This theory of knowledge also has implications for the Fitch-Church knowability paradox. If knowledge is not categorical, then it makes sense to say that everything is known to some extent; thus one of the assumptions of the Fitch-Church knowability paradox is false (that some things are currently not known). However, it also has the same effect on many logical methods in epistemology; such arguments rely on the modal knowability principle Kp, where p is a proposition. If what we mean by knowledge is not about propositions, then Kp does not refer to knowledge, and so such logical arguments have nothing to do with knowledge.

This also provides answers for questions that come up in the study of the cybernetic integration of humans with machines. If I have a computer chip in my brain which provides massive amounts of data, traditional views of knowledge would leave us asking whether or not any of the information there is known. This theory of knowledge does away with such questions, since it does not matter where the information is stored when making a knowledge claim, only how accessible it is.

At this point I feel that I have written on a very diverse set of topics and so I must summarize what I have uncovered about the nature of knowledge. My pragmatic leanings tell me that knowledge, like all concepts, should be understood entirely on the basis of its practical effects. The practical effect of knowledge of a fact is the ability to perform some task, such as answer a question, that requires one to successfully access some information. Furthermore, whether a putative claim to knowledge should be considered valid is based on the context in which it was uttered, and this has to do with the importance of speed, ease, and reliability of the method of information access. Therefore in most cases, knowledge does not merely imply the ability to access information from memory; in many cases, one can be knowledgeable by use of some equipment, such as a computer, which allows fast, easy, and reliable access to information. In this way, computers and other tools are constantly reshaping our identity and becoming a part of who we are.

Works Cited
DeRose, Keith. “Contextualism and Knowledge Attributions.” Epistemology: The Big Questions. Ed. Linda Martin Alcoff. Malden: Blackwell, 1998.
Heidegger, Martin. Basic Writings. Ed. David Farrell Krell. San Francisco: HarperCollins, 1993.
Heim, Michael. “The Computer as Component: Heidegger and McLuhan.” Philosophy and Literature. 16.2 (1992): 304-319.
Koller, John M. and Patricia Koller, eds. A Sourcebook in Asian Philosophy. Upper Saddle River, NJ: Prentice Hall, 1991.
Peirce, Charles Sanders. “Definition and Description of Pragmatism.” Pragmatism: The Classic Writings. H. S. Thayer, ed. Indianapolis: Hackett, 1982.
Wittgenstein, Ludwig. “On Certainty.” Epistemology: The Big Questions. Ed. Linda Martin Alcoff. Malden: Blackwell, 1998.