Reason Better

Picking the meat off of A.I Engineering
Food for thought!

Picking the meat off the bones of Artificial Intelligence "AI" and Intelligent Systems "AS's" research.
Based on the processes of an artificial intelligent system, we show how humans can be more effective, more intelligent, in reaching their objectives. This is "good reasoning".

If, as in the fairy tales, I could make only one wish for increasing the quality of life for all people on Earth, I would wish that everybody, but particularly young people, would increase their intelligence by learning and applying the methods shown below. A general increase in the intelligence level can only result in a better, more neighborly, and friendly community, in responses that increase the health of all, and a better economic standard of living for all.

The inverse result can be seen easily. Those areas of the world that have famines, several epidemics, and low standards of living are those that have a low average level of intelligence, mainly due to the lack of education.

How to think well is so important that it should be taught at school at the same time as all the other subjects. Knowing these, without being able to think, is of little use.


--------------------------------------------------------------------------------

Better Reasoning
Human beings are the species that dominate this planet because they reason better than all other beings on it. However even between humans there are those that reason better than others. We observe that those that reason better, normally also live better, they have a better standard of living, a better quality of live. Therefore it is of the greatest importance to reason as well as possible. (Here we use the word "reason" to indicate the processes occurring in the brain of the AS's).

This is your opportunity to increase your intelligence by 200% to 300%, believe it, or not. By this we mean that you will reach more of your objectives, and you will do so in less time. Naturally if you are already applying all these methods and in the right order, your intelligence will not change.

The ideas expressed in this chapter are not new; they can be found in other books. What is new is that we give a theoretical basis for the existing empirical ideas. You will find that all the concepts and response rules expressed here will flow naturally from what we now know about intelligence, intelligent systems and their society.

By observing the functioning of Artificial IS's, we can learn a lot about how to reason better, about how to increase our intelligence. Because, if we perform all the steps of reasoning consciously and in their proper order, our reasoning will be more effective and we have a better chance to get the wanted results.


--------------------------------------------------------------------------------


Well, how do these Artificial IS's reason? Their computer program consists of parts, each of which does a function. The main functions are:

Have an objective
Receive sensations
Determine the situation
Make alternate plans
Select best response/plan
Act
Observe results
Store experiences

--------------------------------------------------------------------------------

Objective
The first function that we will analyze is that of having an objective. We can observe very clearly in an artificial IS that when we give it no objective, it cannot act, or only act randomly. The system cannot do what we want, because we have not indicated to it, what we want it to do. This is obvious. On the other hand, how often have we observed that persons start to act, without having clearly defined what their objective is? How then can they possibly expect to reach their objective? Acting, without a clearly defined objective, is possibly the biggest error persons make when thinking. Without an objective there is no basis for choosing between one response and another, and with an ill defined objectives the responses done will surely not be optimum.

What is the relation between the main objective and lesser level objectives? The main objective of all living beings is the survival of the species. Darwin has shown that only species with this main objective have survived and exist today. To reach the main objective we choose lesser level objectives, obtaining them helps in obtaining the main objective. There is a pyramid of objectives where always the lesser level objectives are guidelines for obtaining the superior objective. If we ask ourselves why we have a certain objective, we will come to a next level objective, and finally to the main objective. If we ask ourselves how we obtain an objective we will come to a lower level objective.

When deciding what to do, the artificial IS and also humans should start with the objective: What, exactly, do we want to obtain. The first question should be if that objective really is a good objective and what objective, of a higher level, this objective supports. If you have an important objective that you wish to reach, picture exactly the way you wish to get to it. Imagine the path and the intermediate way stations. These then become your subobjectives.

Well, what can we learn from all this? Knowing well the pyramid of objectives we can analyze if our present objective really is the best to reach our more permanent higher level objectives. Often we invest much time and effort in obtaining temporary objectives that really do not help us appreciably in reaching our important objectives. If we analyze our reasons for having a certain objective we can avoid detours and reach our important objectives in a more direct line.


--------------------------------------------------------------------------------

Obtain Information
Before we can determine what the present situation is, we need information about our environment. This information can come only from our exterior senses and from our body (hunger and pain, for instance). We take this immense amount of information, and concentrating on that which is important for our objectives, we build up the present situation. We should pay special attention to information received through spoken or written words; Is the source reliable? Does the information contradict other, previously obtained, information?
If so, what is the evidence that this new information is true? Can this piece of information be tested? Can it be quantified?
If the new information is true, the older, contradictory, information has to be thrown out.

Information about what another person has said, can be short and complete. Information about occurrences in our environment can never be short and complete. The environment has much to much detail for that, and in any case is only partially knowable. We should always treat with extreme care news that is short and complete and sounds logical. The environment is much too complicated for that to be possible. Somebody has completed and adjusted this piece of news by using his imagination.

If we like the news, we will normally believe it without much checking. But liking a piece of news does not make it any truer. So we have to examine all the information that we receive and try to get as much detail as required for our purposes. Without sufficient and correct information from our environment, we will build up an erroneous situation and reacting to it, our responses will not be adequate to the real situation. We will not reach our objective.


--------------------------------------------------------------------------------

Make Alternate Plans
Intelligence is not something that is general. It is always higher or lower regarding a certain subject, regarding a set of related situations. So one way to increase our intelligence in a certain field is to gain experience in this field; to become involved in many situations in this field. Thus we gain experience, and from it we determine several possible responses or plans that are adequate for the situation. This can be for the composite situation or parts of it. But there have to be several proposals of response. How often have we observed persons, that look at a situation and immediately say: "Now here is what you have to do. . ." Following this the person details out a single plan of response. This is wrong. There is never only one way to deal with a situation. Considering only one course of response is another of the big errors that is usually made when thinking. A good artificial IS and medium and large business organizations always analyze quite a number of plans before they decide what to do. We should do the same in our daily life.

Well, how to we go about this. Possibly we have experienced a similar situation in the past. What did we do then? What were the results? Look for the ingredients that were in the situation and also what the most prominent effects of our response were; what were the undesired side effects.
What would be the other persons point of view. What are his/her objectives. How would the other person react to our action. (See also ethics) Always imagine what the result of our action (probably) is going to be, both immediately and in the long run.
While thinking or talking it over with friends, always use different words for different concepts, do not use the same word for different concepts, that creates confusion. Think only about the situation and your possible actions, do not stray to other subjects.
From our previous experience we select possible responses for the present situation. If we do not seem to be able to find various responses, we make what is called a "brainstorm" (see one of the many books on the subject or on value analysis / functional analysis /value engineering or see weblinks). One or various persons can do this. We write down our objectives in about two words, if possible, a verb and a substantive. For instance "find a job," or "go to Oklahoma." Then we write down whatever comes to mind, without hesitation and without considering if it is reasonable or not. The analysis comes later. After about 15 minutes, we stop. Now we group the proposals into similar responses; we analyze and improve all of them as far as is possible. This leaves us with a few viable alternatives.

Maybe what we want to do is to create something, an object, a method or an organization. This could be something new, or an improvement on something existing. The way to do this is to make a functional analysis, to forget all that exists physically today and to define what exactly the function should be of this new object, method or organization. Knowing the function we can then look for ways for accomplishing this function as cheaply and easily as is possible. Again a good way to do this is by a brainstorming session. We recommend to look up the details in a book on value analysis.

Maybe our objective is to resolve a "problem". A problem is something that previously worked well, but something changed; we are no longer satisfied. The problem can be a person, an organization or a machine. Kepner and Tregoe taught us, in a very good book titled "Problem Solving and Decision Analysis", what to do with a problem (See also Kepner Tregoe worldwide or other web links). We have to look for that which changed, where we can observe the change, when it changed, what part changed. In other words we try to see the situation with as much detail as possible, and contrast it with a previous situation where this change was not noticeable. The contrast can also be with other persons or machines, where we do not observe the problem. Then we list all possible causes that could have produced the observed changes only, but no others (using our experience, our store of response rules). We verify by determining what other changes our presumed cause should have produced, but that we have not noted so far. Then we look for these changes. If we find them, we probably have the cause of the problem. As you can see the procedure is complicated but quite effective. We recommend to read the details in the original book.

So we see that we should always elaborate various alternatives. If we have little time, this analysis will have to be very superficial and short. But if there is time, we should always "think before we act."
In our daily life we come across routine situations, where we have analyzed alternative responses in previous occasions, and we act immediately. Even so, occasionally, it is convenient to check if the reasons for selecting a certain alternative in the past, are still valid today.


--------------------------------------------------------------------------------

Check Your Function
The function of a bottle is to contain liquid. The function of a house is to protect its inhabitants from the weather and intruders. The function of a car is to transport people. The function of an object is that which the inventor intended to accomplish with the object. There are many different ways to realize a function. Both the bike and the car fulfill the function of transporting persons. Also in the reasoning process, the way the human mind and the computer realize the different reasoning functions are quite different, but the functions themselves are the same. Also the results are the same.



--------------------------------------------------------------------------------

Make Best Plan Or Action
We should not select the alternative that at first sight looks best. Often there are undesirable side effects. Often another alternative, even if not as effective immediately, is better in the long run. Naturally, when there is no time, our selection has to be instantaneous, but more often than not, there is time, and we should use it. The best way is to make a selection in writing. Consider the main advantages of each alternative. Write down the same amount of the most important advantages for each alternative (the same amount, so as not to bias one alternative) and give each advantage a value, for instance on a 1 to 10 point scale. Note disadvantages of one alternative as advantages of the other alternatives (In alternative B we do not have the disadvantage of x that exists with high probability in A).

Take into account both short term and long term advantages and disadvantages. Sometimes an alternative, while having important short term advantges, is not so good in the long run.
Also some decisions are less important that others, beside the money or time involved. Some are reversible. If the result of our response is bad, we can undo, with little effort, what we did. Some are irreversible or only reversible with a major effort. With those, our selection of a plan of response has to be very careful.

In a difficult situation it is very important to choose a good response the first time around. If a situation is difficult and we choose a bad response, the situation will get worse. Now it is even more difficult to select a good response. On the other hand, if at first we choose a good response the situation will get better. Now it will be easier to choose the next response.



--------------------------------------------------------------------------------

Act, Observe Results And Store Experiences

Act
Once we have selected a plan, we have to act. The artificial IS has no choice; the response is build in. But that is not the case with persons. Many persons think, but then do not act accordingly. There is no sense in just saying we really should do such and such. We have to "stand by our convictions." Acting, in the best way we know of, is the only way to reach our objectives.

Observe results and store experiences
After each instance the artificial IS stores the experience. This is information in the shape of a response rule, that contains the situation, the response and the result. This storing process fills the memory of the IS rapidly. Therefore, similar to persons, the artificial IS has to forget. It forgets the least important information and the information not used for the longest time. Persons also forget. But in order not to forget what is important, a good way is to review our important experiences once a week and thus keep them fresh in memory.

The IS executes all these functions one after another. There is no way it can change this order. That is not so with human beings. Nearly never do they follow the correct sequence. Normally they act first and then ask themselves if another response would not have been better. They make an incomplete or no definition of the objective and from there they jump to the "only possible" response. Then they look again at the situation, then they analyze the response and finally ask themselves what it was, that they really wanted to achieve. It is obvious that this cannot give good results. To save time and effort and to reach one's objective, it is absolutely necessary to do all the functions in their proper order.

Why don't you try to think in as orderly a manner as a computer? At first it will not be easy, but with time it will become a habit. However it does need practice.



--------------------------------------------------------------------------------

Practice
We have seen how an artificial IS reasons and what we can learn from it. But that is something like learning to ride a bicycle or to drive a car. You cannot learn it by only reading a book. You need practice.

Select first an easy but real problem that you have. Do all the steps in sequence, preferably in writing. What is it that you want? Why do you want it? In other words what is the higher level objective? Is it a good objective? Is it worth your time and effort to reach it? Why? Once you have answered these questions and are satisfied with the objective, go to the next step.

Observe the present situation. Use your senses and sources of information. Get a true and detailed description of the situation. In what way is the situation similar to a past situation and in what way is it different. With a good description of the situation in hand, and based on your experience, select various plans of response. Remember that they have to be various, even though at first sight there is a response that appears to be the best, by far. Try to write down and elaborate several alternatives. Give a numerical evaluation to the important advantages of each. Remember to note down the same amount of important advantages, in order to give all alternatives the same chance. Select the alternative with the highest point rating. Be extra careful with irreversible responses. Once you are convinced of what is the best response, you have to act, otherwise you will not reach your objective. Observe the result and learn from this for the next time.


--------------------------------------------------------------------------------


At the end of the week, review what you have done. At first you will make errors, that is natural. But with time, you will see how your reasoning improves, how much easier it is to do all the steps and in their proper sequence.

Then, when you have learned the method, use it to solve your big problems and use it to achieve your important objectives.

Human beings are the species that dominate this planet because they reason better.