User Tools

Site Tools


a_philosophy_for_complexity

From http://groups.haas.berkeley.edu/gem/essays/complex.html

A Philosophy for Complexity, from Managing Complexity

C. West Churchman


I should warn you in the beginning that there are two strange characteristics of philosophers. One is that they dearly love to ask questions, and if they sniff out that there is going to be an answer somewhere they are going to be very unhappy. So if you sense that I seem to be approaching a conclusion and get as far away from it as possible, you sense correctly. The other is that they equally like to point out that nothing is new about complexity. It has been going on for a very long time, and other era s in history seem to reflect the same concerns that we have in the 1970's about the idea of complexity. There are three key words that I will make the theme of my remarks, and they are the only technical words I will use. One is ontology , which is the philosophy of reality. It is basic because we really want to know if complexity exists. Is the re complexity in the real world? That is the ontological question. The second term is episistemology. For the philosopher this means the theory of knowledge: the question of how we know certain things to be true and other things false. Our question is: If the world is complex, then what is our theory of knowledge which will enable us to face that complexity? And the third term is ethics. The ethical question is: Is complexity a valuable thing in our society, or an evil thing, or neither?

I am interested in how the world is going to manage its affairs, and these three branches of philosophy are intended to give me the base for management thinking about this crucial problem.

I will begin with epistemology, because we have now quite a literature on the epistemological ideas of complexity. The first point is differentiation between complexity and simplicity. This differentiation is based on the notion that to know something we have to know it in a simple form. And the primary knowledge task that the manager faces is to go from complexity to simplicity and from simplicity to knowledge. A word of caution: I am not saying that is the answer – there are no answers here. But th at is a classical approach to complexity. It consists essentially in trying to take the complex and get it into a simple form where we can understand it. Now Holling (in a subsequent paper) presents a perfect example of that, although he hides from us the enormous complexity that lay behind his research. With a few charts he excites us about budworms and other things in the forest. His is a classical approach to getting into a state of knowledge by taking the complex and putting it into the simple. “It ta kes a genius to create simplicity out of complexity.” One of the great geniuses of all time is Spinoza, and his Ethics is essentially based on this epistemological theme. He considers all the complexity that we face in life, filters it down to th e elementary postulates that make up the beginning of his book, and comes out with a very simple theme for everyone of us – namely, that the ethical mode of life is understanding. That is the only message you need – if you are a Spinozist. So out of the total complexity comes this genius' creation of simplicity.

But this is not the epistemology that we are trying to use today in Systems Science. We are not essentially struggling to cull the simple truths from the complexities. Rather, the spirit of our age is to hit complexity straight on, taking it for what it is. So complexity comes to be a characteristic of another mode of attack on knowledge – namely, letting everything that is there be examined and trying to put complexity together into a system which we call a model. For example, complexity is partiall y measured by the number of variables in the model. Consider the growth of this measure. When we first started doing operations research in the 1950's, the biggest linear program we could handle contained about 15 variables and a few constraint equations. I understand that in 1975 there was a linear program that had 2 million variables and 35,000 constraint equations. Such an approach to understanding complexity makes little attempt to sift out simplicity. Rather, it takes the world to be the way all thos e variables are, and does not try to deduce some simple truths.

Another meaning of complexity says that it consists in the complex interaction between the variables. We have come to be aware that, when you change a variable in a system, the impact on many other variables throughout the system occurs either immedia tely or with a time lag. For example, Forrester's World Dynamics model essentially had only five basic variables in it – not very “complex” in terms of the number of variables; but when you think of the interactions in the model, it appears quite complic ated. Thus, DYNAMO, which is the simulation acronym for the model, could be thought of as a fairly complicated model in this second sense.

Another meaning of complexity recognizes that we live in a world of uncertainty. In making historical references in the first definition of complexity, I referred to Spinoza. For the second and third, which relate to large-scale model building, I refe rred to no historical figure. I think this is one thing which is truly modern: very large-scale model building. In the case of uncertainty, however, the historical roots are many. I will select one figure, Carneades, who lived in the post Aristotelian per iod. It was Carneades who emphasized the point that no assertion is ever known with certainty. We live as human beings in a universal uncertainty, and uncertainty about everything that is happening to ourselves and in the outside world. Butt said Carneada s, that does not stop us from making assertions. In fact, the only certainty we have is that we will go on making statements even in a world of uncertainty. But Carneades argued that some statements are more appropriate than others. So he suggested a meas ure of confidence in the assertions we make, which is based on the word “appropriate.” The word in English had exactly the same root as probability. In fact, all probability theory is simply an extension of Carneades' idea that we need to be able somehow to measure the appropriateness of the statements we make. Accordingly, our world becomes complex to the extent that it is uncertain for us. The obvious suggestion here is that we need to develop a calculus of uncertainty. We can measure unexpectedness, sa y, on a scale ranging from zero for the completely unexpected, or the impossible, to one, for the completely certain.

Thus we have developed, in history, a theory of probability, and, more generally, of one type of uncertainty. But there is much more fundamental uncertainty about our understanding of the world that is not reflected in probability theory. This has to deal with data - the information we use to build our models, to describe the parameters that drive the models. We need that kind of information if our models are going to have any kind of content. But the character of information in our understanding of the world is totally different from the rather simplistic notion of information that empiricism gave us. Empiricis m says that if you want an answer to the question, “Are all swans white?” then look at some swans. And if all of them should turn out to be white, then you are on the right track. But if you suddenly go to Kyoto, japan, and happen to see the black swans, that's done it. There is a black swan (unless the Japanese are painting them). So, not all swans are white. That is simplistic.

But systems scientists cannot proceed in that manner. To illustrate, suppose we would like to know the cost of the new Systems Science Program at Portland State University. President Blumel (1) has praised the program, but if I were on the Board, I wo uld like to know, “What is the program costing?” If you were an accountant you might add up all the salaries, equipment, computer time, student pay, and all the rest of it, and say that that figure with an overhead is the cost of the program. But, form a systems point of view, this procedure would not be right. Epistemologically, that is not the cost of the program. The real cost is the lost opportunity of all the individuals who are in this program at Portland State. What could they be doing now if they were not doing this? We need to estimate the value of the activity they could be doing, because they are losing the opportunity of doing it by being in the program. It is the lost opportunity that makes the cost. That is why we are so unpopular as systems people., We go into a room, and they are arguing about which programs to have at Portland State. And we keep saying, what other programs could we have? Those are the ones we need to talk about.

No you can see why empiricism does not work for us. Where do you go to look for cost? What do you look at? You have nothing to look at ; you must think about lost opportunities. In other words, you need a large value-loaded model to estimate those lost-opportunity costs.

For example, if you wonder whether Forrester's model reported in The Limits to Growth used good data or not, the answer is no. There are no models around today that use good data, in the sense of being epistemologically sound. Now what do we do? We use past data. We use the period which, say, The Limits to Growth does – i.e., from 1900 to 1970. Many things happened in that period, but surely there was a lot of mismanagement. If we use historical data, they are partially generated ou t of bad judgment. And if we use those historical data to make forecasts into the future, they are bad data.

But how do we get the necessary data? I do not want to discourage budding systems scientists. We can have a procedure for estimating opportunity costs in Operations Research. If we want to know the costs of holding inventory, we cannot get them out of the accounting books. what we do is to talk to the smart controller, who seems to have some grasp of the financial system, and we talk and talk and talk, or rather we listen and listen and listen to him, because 95% of what we get is irrelevant. But out of all the talk we cull an estimate. That is the number we use, relying on our own judgment. It is a guess based on the best judgment we can make. But, of course, it may be terribly wrong.

Now you may be a systems expert and ask: Why do you not talk about sensitivity analysis? the answer is: I have not because some of the data have to be critical. If they are not, then we are really not doing anything. If none of the data matter at all, we might as well forget the whole story. So, I do not think the answer lies in sensitivity analysis.

Thus, the real epistemological complexity lies in the essential properties of your information needs. We are not laboratory technicians with good measuring devices in front of us.

Now is the time for an ontological interlude in this symphony. If the previous movement depresses you, this music is a little sweeter. So I will bring out the violin section accompanies by a few cellos, and point out that in classical rationalism the question of complexity had a beautiful solution. Descartes realized he was dealing with a very complex question. His question was: Is there any proposition I know with absolute certainty? He realized he lived in a complex world, and somehow, while he was conducting the search for the unqualifiedly true proposition, he needed a mode of living, and some solid support. The answer is the Guarantor. His task was to prove the existence of a benign God. If we are the children of a benign God, then the problem of complexity is not even threatening. It depends upon how you define “benign,” to be sure, but I am not going to go into all that detail. The point is that God will take care of us poor children in a messy world no matter how messy it may be. Because He is bening, He is not evil. So you now have the opportunity fo being a reactionary (and there is nothing wrong with that) and of returning to the classical rationalism of the seventeenth century, and to your own satisfaction prove the existence of a benign G od. In such a case, the complexity is there; but it resides in the infinite wisdeom of the deity and not in your rather limited capabilities.

That is the ontological interlude: it says you have an optimist option. Any time somebody is in a gloomsday mood, you can reflect on that question. If you do and succeed, then you have the epistemological, ontological, and ethical answer. The key hist orical figure in the discussion is Descartes, because in the Meditations and Discourse on Methods he was living through that same experience of existing in a complicated world and finding some way in which he could feel the guarantee.

Now the interlude is over, and we are back to present-day society which has not opted to go that route – not, I might argue, on any rational grounds that I have been able to find. It is still an option, not by any means closed, even though some of th e uglier people in the scientific community may say, “I will not entertain that hypothesis because there is no test of it.” The problem of the Guarantor is still with us! Is there any guarantee that human development and progress is going to take place?. If there is – that is a God.

But now I will come back to today, and leave the world Godless for the time being and talk about the ethics of complexity. Is it good, or is it bad, that we live in a complex world? I am assuming that ontology has given us the answer that we live in a complex world, and so has epistemology. First we have the simple-minded, gloomy side that says complexity is bad. Why? Because, it will kill us all! If you look at a future population chart, pollution increases drastically and kills much of the populatio n. Complexity is bad because it frightens me, or complexity is bad because I want it to frighten me. Whatever your answer is at this level, you must become part of the general gloomsday philosophy. I am not going to talk about that, because it is not very interesting although highly publicized.

I will talk instead about the bright side of complexity. Complexity is good! I will play this tune with a little bit of tongue-in-cheek. The theme here might be called the Club of Rome syndrome. Being very concerned about the future of the world, the group asked Haas Ozbekhan to write a position paper. In that paper Ozbekhan first introduced into Systems Science literature the French word “problematique,” standing for the complexity that I mentioned earlier, the interaction of the world's problems. I might say, parenthetically, at this point, that many systems scientists and a lot of people in the world do not recognize this to be true. The Battelle Memorial Institute has a project called Dematel, in which they asked the world's experts to identify th e world's problems, and the experts identified forty-seven of them. (It is a mystery why these things always come out to be prime numbers (2)) They then asked experts in the world of transportation education, etc., to fill in that forty-seven by forty-sev en matrix with numbers from 1 to 7 indicating the interaction of the pairs of problems. I said to them, “My God what are you doing that for! The only way to fill in that matrix is to put nothing but sevens in it.” They said, “Yes, we know that, but look a t how the experts of the world do it.” True enough, many of them quite frankly saw no connection at all between energy and education. I do not know how they think the school buildings are lit, or how they think in general.

But back to the Club of Rome's attitude. It says, in effect, that the world is in a complicated situation, that complexity is there, and, therefore, we need, in effect the best minds to try to tell us how to get out of it, how to progress, how to get ahead, and how to make things easier. Now when you get into that syndrome complexity is pretty good because planners become very important. For a while we did not know whether we were important: in the 1950's and 1960's everyone was muddling around with q ueuing theory and inventory theory, and trying to solve little problems of little companies or little problems of little governmental agencies, and so on.

But now, by God, we are hired by the Club of Rome and we can really take the whole world on. And as we do it, the titles of our books reflect our modesty: The Limits to Growth; Mesarovic's Mankind at the Turning Point; Stafford Beer's < I>Platform for Change; and finally, Russ Ackoff's Redesigning the Future!

I really think the world is getting great for us planners. But that's just a facetious side, and I do not think that you have to take that point very seriously. What these “holistic” people are doing in the world today is making us pay attention to th e future generation, as a moral obligation on our part. Now I am dead serious. There was a time in my young life when I was struggling with the question, “What is morality?” I came to the conclusion that morality is what a future generation would ask us t o do if they were here to ask us. I believe that the voice of future generations is a morally critical voice today, because a lot of things we are thinking about today have their implications for the future generation. Consider nuclear energy production a nd its waste. We do not know what to do with the waste now. The salt mine is no longer a necessarily good solution. Some people think it should be taken to the Antarctica and sunk in the snow. One suggestion is that, since we do not know what to do with i t, what about orbiting it, and let some smart technologist of the future take care of it. They will figure out what to do with our waste! That is a serious space application. I feel that it is clearly immoral. It is immoral because future generations woul d very likely ask us not to do this if they were here to ask us. We may have to do it, incidentally, but at least we have to admit we are being immoral in doing it.

Linstone talked about discounting the future (Part 1). Any positive discount, I think, is immoral. My children are a lot more important than I am in my life and their children's children are still more important, and so on. The value of future generat ions keeps increasing, and becomes an amplifier rather than a diminisher. The bright side is that we are being made aware, in the 200-year projection, of the things we are doing today and their implications for future generations. We are facing the proble m of how we are going to assess values that are out there a hundred years from now. Given our incapabilities of assessing our own values, that is a tremendously complex problem. It still does not diminish its importance, and, therefore, I think that compl exity is really on the bright side of making us face up to it. There is a gloom side which says we cannot cope with the problem. So you can present the issue dialectically in your mind between the bright side, in that we are deeply concerned, and the gloo my, because we do not know how to handle the problem.

Now I come to another bright side. A vast portion of economic literature involves aggregation. Economic models have to aggregate a number of things, and one of the things they aggregate is you! In great globs you are aggregated into statistical classes. There is nothing more frustrating than the damn statistician! You can get up on the Golden Gate Bridge. You have been thinking for months that you will take your life. You finally go and stand on the bridge, and as you are falling, some statistic ian says, “Yes, that fits right into the probability distribution.” That is the moral of your glorious suicide, ruined by these aggregators! They ruin our lives in very, very deep ways because they aggregate. Aristotle is probably at fault, or at least I will blame him. He receives so much credit – it cannot hurt to give him a little criticism. He is trying to figure out how logic works, and he knows pretty well in his own mind that if all men are animals, and all animals die, then all men will die. Then he thinks about Socrates and how he would do in a syllogism. So it turns out that Socrates is man, all men will die, therefore, Socrates will die. At that point in history, Socrates was made into a class. That was deadly, but people have been doing it ev er since. They have been taking us unique individuals, classifying us, and making decisions about classes. Not about us as individuals – but on us as classes! They forget that there is another side, another kind of reality. We are back to the ontological problem. There is something just as complex as classes, just as real, and that is our own inner self. The self, the individual, the psyche, the soul - not the brain, not the mind, but self.

Is this the first time it has been said in history? No! It goes back to Indian philosophy, to the compassionate Buddha, to the whole mythology of self in Hindu philosophy. It is said over and over again in poetry and drama. One of my favorite characte rs is Immanuel Kant. Kant, after writing his Critique of Pure Reason, which deals with empirical science, decided that something had been left out. His own language is different, but he finds that there is another aspect to reality which he calls “the will.” He could have just as well called it the self, for that matter. He then develops in his second critique, the Foundation of Metaphysics and Morals, a story about that world – The Kingdom of Ends. In The Kingdom of Ends, we are all wills – nobody is a class. Nobody is a man, woman, adult, black, white etc.; we are all wills, not distinguishable by any class categories. Here is a man who spent his life in what we would today call hard science. He worked in physics and astro nomy, wrote a basic text on the philosophy of science, etc. In the 1780's he comes to the realization that something in his life has been left out. For all the complexity lies on the phenomenal side, there still is another world. At the end of one of his books, there is this statement: “Two things fill my heart with never-ending awe: the starry heavens above, and the moral law within.” If I could paraphrase that, without ruining Kant's original saying, I would do it as follows: Two things fill my heart wi th never-ending awe: the complexity of the total social system forever, and the self within. No one should stand up and say, “Look, you have to forget that self within; we have other problems –pollution, poverty, etc. – on our hands.” The inner world is just as important, but is it as immense as the starry heavens? Yes it is! I think it is just as immense, just as awesome, and just as compelling. In modem phenomenology the call for attention to the inner self is to be found in Husserl and Heidegger, and in psychology, especially in Carl Gustav Jung and the Jungians, but there are many other people writing on the same theme, the development of the self.

In Two Essays on Analytical Psychology, Jung says that, until you have gone through the process of individuation, which is his label for the understanding of the self, you will not be able to face the social problems. You will not be able to b uild your models and tell the world what to do.

For the Christians among the readers, I will cite Matthew, chapter 25, as well as tell a little story. Some of my students were at Laguna Beach (California) trying to help the community in its desperate plight. It really is a community that has a desperate plight, since it is torn apart at its very foundation. Some people in a lovely l ittle town on the Pacific beach south of Los Angeles want to put up high-rise developments. Some people see development as the end of old Laguna: the artists will be gone, and shops and hotels will take their place. Then there are the hippies, some living off drugs and painting pictures, and the three rich-man ghettos, where you have to have a card in hand in order to enter. So the community is torn apart.

The students and others were trying to help the community pull together. When I visited them, they had a little place in the center of town they called the Volunteer Post, where people could drop in and talk about their problems. Even the police chief occasionally dropped by. That day they had a schizophrenic on their hands, a very disturbed young man. Eventually he went out to get us some coffee, and they said to me, “We've been talking to this guy all day long. It seems to us that this big city of L aguna Beach is being neglected because we are spending all our time with one disturbed man.” I then got out Matthew, chapter 25, and read it. It says that the King on Judgment Day turns to those on his right hand saying, “you fed me when I was hungry, you clothed me when I was naked . . .,” and so on, and they say to him, “Lord, when did we do any of these things?” The answer is, “Even as you did it onto the least of these, my brethren, you did it to me.” One case.

Thus, from that perspective of the unique individual, it is not counting up how many people on this side and how many on that side. All the global systems things go out: there are no trade-offs in this world, in this immense world of the inner self. Y ou do not trade off your evil by going out and doing good. Forgiveness happens, but that is not a trade off. All our concepts that work so well in the global world do not work in the inner world, which is essentially a world that is ineffable for the Engl ish language. We have great trouble describing it very well in scientific language, but it is there, exists, and is important. I know what the global planner's answer will be: How am I going to put the two together? How am I going to put an individual per son into the model? Of course, the answer is, You are not going to do it! “Well, what am I going to do? Are you telling me that I have to pay attention to the unique individual, or aren't you telling me that? Anyway, I don't know how to do it bec ause I have to handle 300 years and billions of people and you are telling me that I have to handle the unique individual. Tell me what to do.” The philosopher is not going to tell you what to do. I warned you at the beginning, I am not going to answer an y questions, so I am not going to answer that one.

Hegel was one figure in history who had a suggestion. He suggested in his writing that the mature individual is the individual who can hold conflicting world views (Weltanschauungen) together at the same time, and act, and live, and that his or her life is enriched by that capability - not weakened by it. If you can think about that - that will be the end of my message here. That is complexity, that is really complexity. To be able to see the world globally, which you are going to have to be ab le to do, and to see it as a world of unique individuals - a Kingdom of Ends, each individual infinitely valuable in itself, not to be “compared” excepting perhaps through some key words. The key words for that world are: faith, hope, and love. With those in hand, and Paul's message, complexity can be handled if these exist. If they don't, it can't be. But these come out of the inner self; do not try to define them, please. If you do, realize that you are ruining them. You cannot define hope. But if it is not present in today's society, then according to Paul, a systems man, the whole of human destiny will fail. The title of this paper is “A Philosophy for Complexity,” and what I have been trying to do is to develop for you a kind of mode of questioning, essentially using historical literature as a basis. I have simply been generating many questions so that you c an reflect on the issue better. Wherein is its reality and our way of perceiving it? How do we understsnd complexity? And, what is the value of complexity?


ENDNOTES (1) Of Portland State University (2) Editors' Note: True to form, the recent Stanford Research Institute's “Assessment of Future National and International Problem Areas” for the National Science Foundation identified forty one problems! (Cf. Teige, Harman, and Schwartz, Part v.)


a_philosophy_for_complexity.txt · Last modified: 2015/01/31 23:55 (external edit)