Theory of Knowledge

I should put these old essays on my blog. Some of them are worth a read, I think. They did not get very good marks, because that is the problem with doing university courses. They get marked by post graduate students and mediocre academics. The former are hoping to become mediocre academics and do not want to jeopardize their chances.

In other words, a bunch of partial autistics; people who are incapable of following a multi dimensional argument and so do not think I am making one. Since no matter how well researched and worked out my stuff is, I cannot get a fair mark, I stop caring much and just look into some topic that interested me. I have already put some on my blog.

Below is a typical example. I did this last summer when I did the course in the history and philosophy of science. It is not exactly as I handed it in. I actually got not too bad a mark on it, even though I pretty much ignored the question.

I chose the topic "experiment and theory in the science of mind". It gave me a chance to look at Artificial Intelligence (AI) and all the trouble cognitive scientists have had playing doctor Frankenstein. I did an introductory course to cognitive science awhile back. From this I have had the idea that science of mind is mainly about cogsci which is mainly about AI.

I was disturbed to realize that the framer of the question had the idea that science of mind is psychology, especially Freudian psychology. Freudianism is philosophy, not science, as others in the class had pointed out. So I do not think that is a viable question and it offends me. In my original conception I at least had a question I could do something with, and I was not starting over after the work I had already done. So this essay went in.

I was also concerned that the marker would think I am recycling old researches, but I did some new research and thinking for this. I had not yet had the opportunity to put my ideas about this in writing. It fit well with what has been taught so far in this class. Especially, the four theories about the basis of knowledge; rationalism, empiricism, authority, and history. I say knowledge is based on history, which means context, which means intuition. That is my thesis.

That theory can be separate from experimentation is a false dichotomy in the science of mind or any other science. It is not exactly, as Stanford Encyclopedia of Philosophy has it; "Although theory without experiment is empty, experiment without theory is blind." It is more like, thinking is theory and experiment working closely together in a constant cycle. The rationalist thinks theory can happen independent of experiment, the empiricist thinks experiment has meaning without theory.

Those pursuing AI are generally thought to be rationalists. They think intelligence is about the ability to do complicated reasonings and calculations and think that this is all they need. This idea has had a destructive effect in western civilization starting all the way back with Plato. According to Kenaw; "For Plato, the world of particular things is merely a world of our opinions. Real Knowledge is...directed at the world of forms. Understanding....deals with ideas, or universals. These ideas or universals are timeless, ahistorical, or context free." I could not have put it better myself, so I did not.

Rationalism has been disputed down the ages, and Kenaw noted Heidegger and Wittgenstein as having "rigorously argued and critiqued" rationalism. The problem these people would have had is in trying to argue abstractly against abstraction. What AI has done is finally give a way of concretely falsifying rationalism. That it has been impossible to build a computer that works like the human mind shows that the mind does not work like a computer; does not work logically or mathematically.

Here is the critique that has developed of cognitive science, which seems to mean almost the same thing as AI, at least older cognitive science, and at least to Stanford Encyclopedia of Philosophy. A computer cannot duplicate the emotions that are a large part of human thinking. A computer cannot duplicate the experience of being inside a human body and living in the actual world and within a society. There is a debate about exactly what human "consciousness" is, and it is pretty sure that if you cannot identify human consciousness, you cannot identify it in AI. The human brain simply does not have the computational capacity to deal with life in a computational way, for example our working memory only handles about seven bits of information. The human mind is a dynamic, not computational, system.

Yet AI scientists still think all they need to do is "expand and supplement" their approach. Their arrogance about their own intelligence and ability to perform faster and more operations than most people gives them the idea that this is all intelligence is about. A scan of a fairly recent book intended as a compendium of ideas about the failure of AI, and asking if the attacks on its critics were justified, shows that, while some AI people have felt a need to adopt a stance of humility, but they are all still arrogant rationalists. By example, one thought that the problem has been that the size of the task was underestimated. He wants massive government funding for his "Massively Parallel Knowledge Representation" project. He still does not acknowledge emotion, experience, or consciousness.

Some recent cognitive scientists are more or less on the right track. They have discovered "combinatorial explosion", which means that solving any real life problem by calculation would require an impossible number of calculations. So, nature designed the brain to do "relevance realization"; to sort the relevant information from the irrelevant, to define problems so as to reduce the number of elements needed to be processed in order to solve the problem or carry out the task. Intelligence is about making the complex simple.

They have also discovered context, which means that instead of starting from scratch with each new problem, the mind starts from what it has already learned. It uses that context to sort the relevant from the irrelevant. Therefore, the way that AI falsifies Rationalism also falsifies Empiricism; which is the idea that knowledge is from observation and experience, or experiment. What we observe depends on what we expect to see, which comes from the context of previous experience.

Reason and experience, theory and experiment, they are both governed by context, which means intuition. Humans and other higher organisms build knowledge by building a model of the world in their heads, with which to predict events and decide actions in order to ensure survival and well being. The model gives a context to what we see, and is built on past events and objects encountered, and decisions and conclusions made about them. We are not consciously aware of most of this process, meaning that there is no way to really reconstruct how we know most of what we know. So, contrary to Plato, the world is the world of my opinions. My opinions are the total of my experiences of living in the world.

The other idea of the source of knowledge is Authority. This seems to me to be a kind of infinite regression fallacy. Where does the authority get his, her, its authority from, to decide truth for everyone? And were does that entity get it?

If knowledge is historical then scientific knowledge is a historical process. It is somewhat like the way an individual mind grows knowledge. Everything is built on previous experiences. Elements are reduced to simplicity. Intuition points from the last experiment to a theory which is confirmed or falsified by the next set of experiments. There is nothing rational or "formal" about it.

Illustrative of this, from the class notes, there is the more rapid development of knowledge in the eighteenth century by continental scientists using an analytical, algebraic approach, over those using a synthesizing, geometric approach. Analysis is about working from unknown to known, meaning breaking apart and reorganizing the premises to find the simplest explanation of them. Or, to find which ones do not fit with the rest.

Synthesis is working from known to known, meaning, to create a backward procession of premises to show how you know something. This can be futile because there is no first cause, every premise must be proved by other premises, ultimately some of them must be taken as given, and some of them are likely wrong. But it is more "rigorous", meaning, it takes more work, so those who use it must be more "rational", therefore smarter.

However, no evidence against their assumptions will deter rationalists or empiricists, who are much the same thing and might as well be called rational-empiricists. They need to build their complicated theories to show how smart they are. They want a way to explain everything, even if it is wrong. This is why, when their basic assumptions are proven wrong by experiments such as AI, they have a tendency to brush it off and say that they just need a greater effort to make it work.

Looked at in this context, a review of the content of the course shows that the best and most productive scientists are the ones who have good analytical skills, but who also work intuitively. They learn from their experiments and base their theories on what they learned, and use that as the context for more experiments. They project the past into the future by means of intuition. They use what they know. They do not try to understand why they know, which is futile.

Thinking is mostly applied intuition, therefore knowledge is historical, and that was my thesis.

This was fun!