This number four in our series. I will put the links for the first three in the series at the end this time. I am now going to focus on some of the key ideas developed in Robert Rosen's book Anticipatory Systems. (See the links below). The groundwork I have laid in the first three parts should make you aware that Rosen did not approach these topics from the standard viewpoints but challenged them severely. This will be the case in all of what is to follow as well for this is indeed a different way of looking at the world. We will come back to why this has to be so when we review Rosen's concept of "complexity". Right now there is much to be done to prepare for that. Let us start with this question from Rosen's book:
How is it, then, that the ubiquity of anticipatory behavior in biology could have been overlooked for so long? Should it not have been evident that the "reactive paradigm", as we may call it, was grossly deficient in dealing with systems of this kind? To this question there are two answers. The first is that many scientists and philosophers have indeed repeatedly suggested that something fundamental may be missing if we adopt a purely reactive paradigm for consideration of biological phenomena. Unfortunately, these authors have generally been able only imperfectly to articulate their perception, couching it in terms as "will", "Geist", "elan", "entelechy" and others. This has made it easy to dismiss them as mystical, vitalistic, anthropomorphic, idealistic, or with similar unsavory epithets, and to confound them with teleology.He then compares the Copernican and Ptolemaic representations of our solar system and explains that the Ptolemaic epicycles are not "real" but a simulation device to make the model work.He goes on to say something that will echo again and again throughout this series in many other ways:
The other lies in the fact that the reactive paradigm is universal, in the following important sense. Given any mode of system behavior which can be described sufficiently accurately, regardless of the manner in which it is generated, there is a purely reactive system which exhibits precisely this behavior. In other words, any system behavior can be simulated by a purely reactive system.
Nevertheless, I believe that it is precisely the universality of the reactive paradigm which has played the crucial role in concealing the inadequacy of the the paradigm for dealing with anticipatory systems.These are very loaded words and they need some discussion. Read on below and see what they entail.
First of all, the idea that the reactive paradigm can be seen as an impediment to understanding even if we can create acceptable and seemingly accurate explanations for the way systems behave is a bit shocking is it not? If this is shocking hold on to your seat for it is part of a larger challenge to traditional thinking but that will come later.
Also, the distinction between "anticipation" and "teleology" is so fundamental and so poorly understood that recent books trying to break this subject open do exactly what Rosen was complaining about back then.
We will come back to these issues but we need to go forward. Rosen says this:
Indeed , it is clear that if we are confronted with a system which contains a predictive model, and which uses the predictions of that model to generate its behavior, we cannot claim to understand the behavior unless the model itself is taken into account.I hope the logic of that statement is clear. It is the reason that we need to look at what we do when we make and use predictive models as we do all the time. I'll skip the many arguments that other systems besides us are anticipatory, and therefore make and use models, for the time being since we certainly all can identify with the idea that we plan for future events. The fact that we do makes us model makers. Now comes the punch line. If we operate this way we certainly should want to do it in the most effective manner possible. Maybe you think you already do. I would wonder if that is true if you are doing it unconsciously and without caring about whether or not it can be done better. I suggest that in the following careful look at what this is all about you will be given tools that will help in the future. That is because the way we all behave in this respect has been in very large part a product of the reactive paradigm and the larger body of well accepted reductionist thinking from which it comes.
So what are these models we are so concerned about? I will introduce the basic ideas here, knowing I am skipping over a lot of the preparation Rosen lays down in the book, and fast forward to the sound bite version.
First of all the model is based on a process of turning our perceptions of our world into some sort of "picture" (for lack of a better sound bite) of how that world is organized and operates. The level of thinking this entails is a level of systems thinking. It is not static and deals with events and processes. One does not need a model to perceive a large boulder at the edge of the driveway nor does one expect to encounter it unless one sets into motion a whole host of other things. For example, cutting grass around the boulder entails actions and changes and the use of tools and so on.
I will be very surprised if the above example did not create some sort of picture in your mind and also evoke strategies for avoiding mishaps when mowing around the boulder. So I trust you do "get the picture" so far.
Lets look more carefully at some of what we dealt with in this example. First of all we dealt with your perception of the situation out there. Your senses told you that a number of things were true. The grass was in need of cutting. The configuration of your yard entails objects like the boulder that need to be taken into account as you act out your desire to get the grass cut.
Let's focus on the boulder for now. You have a visual perception of it because you have seen it there before. You may not have its details in your consciousness but thinking about it has brought some attributes up from your unconscious. For example, it is made of some pretty hard materials. It is large enough to create problems if ignored.
Let's take a closer look at this act of perception. let's change the scenario to one where you forgot the boulder was there and you came upon it while mowing. Your visual system gave you that information. Your visual system did not tell you the boulder was made of hard materials and that it could cause you a problem if ignored.
STOP! THINK! Here is where the myth of passive, objective observation breaks down. We now have the advantage of knowing that we could record brain activity during these events and not only was information carried by nerve impulses from your eyes to various parts of your brain, but also information was simultaneously mobilized to be sent to those input sites from your memory and from various association centers. In other words there is no such thing as pure, clean sensory input unclouded by your past experiences. Nor, do I think, would you like it if it did work that way.
What went on was the use of information from outside to encode into a system of thinking already there in your brain. That system of thinking took the natural system outside you and encoded it into a "formal" (again for lack of a better sound bite at the moment) system that predicted that if you ignored the boulder you would have problems. If that prediction could be verified by action you would now be in possession of a working "model" of your world.
If per chance the "boulder" was actually a big puff ball mushroom ready to seed the prediction would have been far less accurate and there would not have been a useful model created by the process.
I have struggled to take a very difficult idea and give a simple illustration. The entire process, the perception of a natural system in the world, its encoding into a formal system in your brain, its decoding into an action on the real world and the observation of the result of the action on the natural world, when it works (Or as we say when it commutes) is a model and this model of modeling is called the modeling relation. This has taken some words and they need to be digested at this point. Rosen uses the modeling relation to discuss the way anticipatory systems work. His use of it to talk about science generally involves a mathematical formal system, but not always. Its usefulness will become apparent more and more as we go on. For now I need to know how our model of modeling commutes for you.
Number five in the series will involve a jump ahead to Rosen's "definition" of complexity and how the modeling relation shows what went wrong over the past few hundred years.
Here are the first three in the series: