How do we know anything at all?
This is perhaps the single foundational question of philosophy, for after this question is answered, we should have a method to follow to answer any other question, and the rest becomes a debate over the facts. But coming up with such a method of knowledge, called an epistemology, is somewhat paradoxical, because we have to assert facts about facts — to know that we know anything, we kind of have to beg the question and assume we can know. For how could we ever come to know that we can never come to know?
Where does truth and knowledge come from? Every fact ever known can be responded to with a “how do you know this?” which requires appeal to yet another fact that is known, and then we must inquire as to where this new fact came from. Eventually, we keep justifying facts with other facts until we either loop in a circle (and thus have circular logic) or base ourselves in an uninterpreted personal experience, such as a sensory perception, thought, or emotion prior to being described by words. Therefore, every fact we have is either justified by another fact or by an uninterpreted personal experience.
This is where the real story starts. Many say that the story starts at the Big Bang, but for everyone, the story starts individually at personal experience until knowledge of the world is built up, and then used to know of what happened back in time, which includes the Big Bang. It is only once we have experience and how to use it to have knowledge, that we can know of the Big Bang and continue our story. The first experience is the real beginning of everything; from where all else flows.
Today I will outline the theory of knowledge that I currently hold, that is inspired and heavily based off of the works of Richard Carrier, Eliezer Yudkowsky, and David Hume.
Uninterpreted Personal Experiences and the Origin of the Model
Uninterpreted personal experiences have the unique property of being literally undeniable. While we may be confused and wrong about what the string of letters “uninterpreted personal experiences” refers to, we cannot be wrong about the fact that we do experience things — one cannot doubt the existence of sensory perceptions without having sensory perceptions, ending up in a contradiction. This is basically the concept of “I think, therefore I am” — the idea of the existence of these perceptions being false is unintelligible.
However, the interpretations of these experiences can be false — you know you are receiving a complex combination of colors, but the fact they correspond to a computer could be false. Perhaps you are being misled or seeing things, and they correspond to something else, or nothing at all. Uninterpreted personal experiences are just that — experiences that are not yet interpreted, and exist prior to language.
The Need for Knowledge
Another one of these experiences we have are goals (also known as desires, wants, needs, or urges). Again, this occurs entirely prior to the use of language, and we could have misremembered what the word “goals” referred to and thus be wrong about having goals because the word “goals” actually refers to something else, we cannot be wrong that some of our perceptions compel us to want things. Therefore, we must act in certain ways to get these things and accomplish our goals.
In order to accomplish goals, we must have an accurate model of the world — we must possess true knowledge and understand to some degree how the world works. We must know what food is poisonous and what food is edible, we must know which actions result in death and which actions result in survival. Therefore, we want to adopt the system of acquiring knowledge that is least likely to give us false beliefs, while still allowing us to acquire a usable model of the world. While a system of knowledge that held no beliefs at all would indeed contain no false beliefs, it would be completely impractical.
This answers the questions (1) of what good is knowledge? and (2) why do we care about true beliefs? The answer to both is that we need true beliefs, and therefore knowledge, to get anything else we care about — an accurate model of the world is the ultimate means to any end we could realistically have. We need a model to survive and accomplish things.
Having a Model
Therefore it is in our interest to have a model and assume as undeniable only what truly is undeniable (uninterpreted personal experiences) and hold everything else to be deniable. For everything else, it is possible for us to be wrong and mistaken, and therefore it is in our interest to doubt these things.
However, we still need to come to some beliefs — we need to actually hold a model of how the world works. A model with no beliefs would indeed have no false beliefs, but this model would be entirely useless; we would just sit around doing nothing until we die; and we would be surprised by our death.
Therefore, in order to have a model, we should use our uninterpreted personal experiences and use them as reasons to form certain beliefs, and use these beliefs to accomplish our goals. However, before we figure out what is true and what is false, we must first invent a system of language to allow us to interpret our experiences, store them for later recall, and communicate them to others. Very little interpretation exists prior to language:
Constructing a Language
Languages are, more or less, arbitrary. The word “oompla” could refer to anything — we could start calling walls, cats, or dogs “oompla” and very little would change. This means that there could be hundreds of different languages and they could all be successful — and indeed this is the case. Whether a house is referred to with the symbols “casa”, “huis”, “家”, “dům”, “maja”, “bahay”, “talo”, “maison”, “haus”, “σπίτι”, “社内”, “dom”, “дом”, “hus”, “வீடு”, “منزل”, “گھر”, “nhà”, “הויז”, or “oompla” is irrelevant — all sounds are, more or less, equally functional.
The best language is whichever language best accomplishes the goals of allowing us to interpret, store, and communicate experiences, and ease of learning and adopting. However, as the failure of others to adopt artificial languages specifically constructed for optimizing these goals, there isn’t much demand for a fully optimized language. Perhaps here, one could apply the old adage if it ain’t broke, don’t fix it.
Instead, what matters is if the language accomplishes these goals enough for people to keep using it. But how exactly do these languages work? It all starts with a personal experience, such as seeing the color red — a visual sensory perception. Everyone who has seen something red (and who speaks English) has agreed to call this perception “red”, so that they can talk about this perception to others. They can then talk about other perceptions they have that correspond with this “red” perception, such as the perception of apple, blood, stop signs, etc.
The fact that you occasionally experience the sensory perception of red is undeniable — what is deniable is whether this experience really is of red or whether you mislabeled it, or whether this red object you perceive really exists in the external world or is just a product of your imagination. For example, a whole host of experiences might combine to give you a belief in the existence of an apple — the perception of weight of the object in your hand, the perception of taste when you go to eat the apple, the visual experience of the apple, all the experiences of others that indicate they are also seeing this apple, etc. You then use language to describe the combination of all of these experiences as “apple”, and anyone else who has had all of these experiences can understand you when you say apple.
Imagining Through Language
However our language allows us to do a lot more than communicate experiences that we’ve personally had — it allows us to conceive of experiences we have not yet had, or may never have. For instance, I could tell you what my college, Denison University, is like, and you’ll begin to have a model of what Denison University is like, even if you have never been to Denison University.
This is because we can imagine experiences — we can have the same sensory perceptions we would have had, had we actually experienced what is being communicated. I can tell you about trees and you can imagine the sensory perception of seeing trees and be able to relate to what I’m talking about. You can know what it would be like to see red, even if you’re not currently seeing red.
But what is even more amazing about the power of imagination is the ability to consider experiences that you will likely never have in real life, such as imagining a blue apple or imagining a space unicorn in orbit around the moon. Once you have a model of how things work, you can further understand the implications of what are not real — you can know how things would be different if other things were different.
Imagination is not only useful for a system of language, but required. It makes no sense to talk about experiences if you can’t communicate those experiences meaningfully to someone who has never had them, and it makes no sense to talk about how the world is if you have no idea of how the world could be different. Without imagination, you wouldn’t be able to understand the difference between true and false, and wouldn’t be able to hold any beliefs at all.
The Shades of “Not Real”
With imagination, our language also becomes very complex. For example, a dog could be alive or dead, reality or fiction, existant or hallucinated, brown or yellow, televised or not, or associated with any other concept. Our default perception of a dog is that of a dog we encounter in reality to be alive and right in front of us, but we could encounter a fictional story about a dead green dog that is televised, and it would still be a dog.
So what is the essence of dog, the sine qua non dog, that which makes something a dog and not something else, like a cat? It seems like the basis for all of these words are experiences of a group of four-legged furry creatures with tails that share a common physiology and characteristics, such as barking, running, and wagging. We can remove some of these components, such as imagining a dog without a tail, but we could never imagine a dog with no tail, only two legs, that never barks but instead meows, and is capable of flight by flapping of wings. Such a creature would no longer fit our definition of dog.
Live vs. Dead
But how do we tell between various kinds of dog? How do we tell between species of dogs? How do we know whether a dog is alive, dead, fictional, or hallucinated? Definitions are therefore abstractions of experience — we categorize a certain group of similar experiences into a word and use that word to communicate these experiences to others. Qualifiers like “live”, “without a tail”, or “televized” allow us to clarify the experience with more specificity.
An alive dog has a completely different set of actions than a dead dog, and we can experience these actions and tell the difference. If we experience the dog running around, barking, wagging a tail, and eating food, we say that we are experiencing the dog being alive. If we experience the dog just lying there, not breathing, having no pulse, and not responding when whacked with a stick, we say that we are experiencing a dog that is dead.
Real vs. Fictional vs. Nonexistent
We can also extrapolate this into more abstract concepts like “fictional dog”. A fictional dog is an experience not of a dog, but of writings of a dog or images of a dog that look like dogs, but do not correspond into furry creatures that we can personally pet. Lassie is a fictional dog, because no dog that we can personally experience outside of books and television correspond with all the characteristics of Lassie.
Some concepts we have only exist in fiction, such as dragons and unicorns. No such creatures exist outside these books and films; we can’t experience petting a dragon, but we can imagine it. Some concepts don’t exist at all — on March 20, 1997 the concept of “Blugga Blugga, the red hippo capable of interstellar travel” didn’t exist, because I just invented him for this essay, today on July 2, 2011. Of course, any time we imagine a concept, we create it in our mind, thus it is impossible to ever imagine something that doesn’t exist at all, since imagination makes things exist in your imagination.
It’s true that Blugga Blugga isn’t real in one sense, since nowhere can we pet Blugga Blugga or actually experience a red hippo travel through space. But Blugga Blugga is real in another sense, because we can have experiences of Blugga Blugga, they just are experiences of a different, admittedly much less important, kind — experiences of text that cause experiences in our imagination. This is what I mean by the different shades of “not being real” — it depends on what real is relative to. Usually real is used in the first sense, not existing outside of fiction and imagination.
Another complex topic is the idea of something being potentially real. This is something that isn’t real (outside of fiction and imagination), but could become real — it has the potential of being real. For example, if you own a penny, you also own a potential pool of melted copper, because you could potentially melt the penny back into copper. If you own enough pennies, you also own a potential copper ring, because you could shape the copper into a ring.
Different arrangements also get different properties — you could easily roll a copper ring and easily flip a copper penny, but it is difficult to do the vice versa. Additionally, you could use the copper penny to buy goods, but you could not as easily use the copper ring. The copper penny also has a picture of Lincoln on it, whereas the copper ring does not. However, some properties are excluded — no matter how you arrange the copper penny, you could never safely eat it or use it to water a plant.
Some dragons are potentially real and some dragons are not. For example, we could eventually use science to create a lizard with wings, but it would be far more difficult to create one that is capable of sustained high-speed flight or one capable of biologically breathing fire, since those latter properties seem to conflict with physics. Likewise, some unicorns are potentially real and some are not, because we could create a horse with a horn, but we could never create a horse capable of crying cures to disease or capable of flight without wings, helicopter blades, or jet engines.
Once we have a consistent system of language, we will automatically have a system of logic. This is because any system that is consistent must never produce a contradiction and anything that never produces a contradiction must follow the laws of logic. Therefore, language must not only require imagination but also automatically entail the laws of logic.
For example, consider the law of non-contradiction — that something can’t be A and also be not-A. This means that if we define something to be “lemonade”, the same object at the same time can’t simultaneously be defined as “not lemonade”, even though it may become or have been “not lemonade” at a different point in time.
But why must the law of non-contradiction hold? Well what if I were to tell you that I had an object that was simultaneously lemonade and not lemonade? In order to make sense of this statement of mine, you would look up in our mutual language codebook (English) the meaning of all of my words, the sentence structure, and the grammatical context. You would then try to imagine what I was communicating to you; you would attempt to imagine an object that was simultaneously lemonade and not lemonade.
However, you would fail, because something that is simultaneously lemonade and not lemonade could never be experienced, even in principle. Therefore I have failed to communicate anything to you, and you know I possess no such object. You don’t even need to investigate any further and you already have close to absolute knowledge that my claim is false. In fact, the only way you could be wrong about the impossibility of my object is if you misapplied the codebook and I was really communicating something else.
Consider an alternate formulation of the law of non-contradiction: the old joke of someone “walking to school, uphill both ways”. Clearly this is something we could never experience, even in principle, because we know that in our world, when you start at a location and go up, the only way to get back to that location is to go down, and going up again will only result in being even further away. Furthermore, a hill can’t be simultaneously sloped upward and sloped downward, since an upward slope is, by definition, not a downward slope. Saying “uphill both ways” does not communicate an experience that can exist, even in principle — we can’t even imagine it. This means we’ve encountered a contradiction.
The Law of the Excluded Middle
Now consider the other foundational law of logic, the law of the excluded middle. This law states that something is either true or false. While we know from the law of non-contradiction that something cannot be simultaneously true and false, we cannot derive from this whether or not there could be additional states, for example something that is not true and not false, but something else entirely; a third value, or even a fourth value.
But again, any values of truth besides “true” or “false” can never be experienced or imagined, even in principle. We can only imagine something that exists or does not exist, something that is or is not. We can’t imagine something in some alternate state.
Keep in mind not to confuse additional values of truth with probability (probably true), with potentiality (potentially true), or with an evaluation of multiple facts (partially true). These are all different qualifications of our state of knowledge, but not of the fact itself.
No matter how probable or improbable, the fact itself is either true or false. Potentiality also speaks about something changing from true to false (or vice versa), but not being anything different than those two. Furthermore, partiality speaks about multiple facts being true and false, but not about any individual fact being something other than true or false. It is impossible to conceive of a fact that is something other than true or false.
The Law of Identity and Modus Ponens
And as has been established, all other laws of logic are derived from these two laws. Consider the law of identity that A is A — that if something is a bananna, it is a bananna. This isn’t very stunning or surprising, but it is also a consequence that if something is A, there are only two options for it to be: A or not-A (law of the excluded middle) and it can’t be not-A (law of non-contradiction), so it must be A. Yes, this is a circle of justification — that’s just what makes these laws so sound.
As an additional example, consider modus ponens, the idea that if you believe A and believe A implies B, then you must believe B. For instance, if you believe that “I go to work every Tuesday, without exception” and you also believe that today is Tuesday, you must believe that you will go to work (or that you went to work earlier in the day).
Believing this not to be true would mean that you do not really believe that “A implies B” (“Being Tuesday implies going to work”), because the nature of believing “A implies B” entails that you will believe “B” when you believe “A”; this is the law of non-contradiction again, and we can’t experience a contradiction.
Again, logic is a property of language. Any consistent language worth having implies having these logical properties, and if we didn’t have logic, we wouldn’t be able to communicate or imagine anything. And if logic were false in reality (as opposed to being something that we were unaware of), nothing would be consistent and life as we know it (or even as we can conceive of it) could not exist. If we exist, logic must exist, and if we communicate, logic must be known, as any language worth having is consistent, and any consistent language requires logic.
But having logic allows us to do a lot more than recall and communicate — it also lets us interpret and give names to our experiences, and then infer and deduce from these experiences to how the world works, and from these grounds we can bridge the gap from our experiences to reality and create a model of the world. And once we have a model of the world, we can work to make it more accurate, because we’ll know what “more accurate” means. And then we’ll have an epistemology.
I now blog at EverydayUtilitarian.com. I hope you'll join me at my new blog! This page has been left as an archive.