Computer science, a science beyond computers?
A lot of people, including myself, have an intuition that computer science is delivering new insights into some bigger part of reality. I believe that, while it is mainly about computers, it is at the same time also about something very fundamental.
However, it has proven difficult to formulate what this fundamental ‘new’ insight is, exactly. Two weeks ago, I read an interesting paper written by a philosophically-minded computer scientist,Amnon Eden. I partially disagree with this paper, but it helped me to formulate my thoughts about this ‘big idea’ from computer science.
‘Information’ as a property of nature
In the paper, an argument is made that computer science is a natural, fundamental science, instead of mathematics or engineering. Computer science is partially the study of informational processes, like coding and algorithms. These processes do not only take place in man-made computers. They occur in nature also. Think of DNA that carries the instructions for building organisms.
Therefore, informational processes are natural phenomena, just like electromagnetism or the human circulatory system. A science studying these natural phenomena is therefore a natural science. This got me thinking. Is ‘information’ really a property of nature, like radiation or magnetism? And what makes a process informational?
I agree with Eden that information processes can be found everywhere in nature. In that sense, information is indeed a natural property. This would make computer science a natural science. But I also believe that there is a subtle difference between ‘ordinary’ natural properties like radioactivity or magnetism, and the property of information. This difference is ignored in Eden’s paper.
I do not think that ‘information’ is a natural property in the same way radioactivity or magnetism is.
Radioactivity and magnetism are physical properties of a process. We can measure whether a natural process has these properties, by using a Geiger counter or magnetism detector. The property of information-ness is not a physical property. Its presence cannot be measured in the same way. Its presence can be established, however. When we encounter a physical process that results in the creation of the complex from the relatively simple, we have good reasons to believe that the process is somehow ‘informational’.
But this is an interpretation, made by us. And I believe that a natural property that can only be established by interpretation of what happens is of another nature than a property that can be established by measurement.
The fact that it cannot be measured does not mean that information-ness does not exist in nature. It does! But information is not a physical property tied to a process itself. Information is a way for us to interpret and understand the causal processes that generate the complex from the simple.
A process like DNA coding may be very intricate. But essentially, there is nothing that makes it physically distinct from other physical processes. It is just an ordinary chain of cause and effect, like thunderstorms or forest fires. In principle, we could trace the causal links between a DNA molecule and the characteristics of the organism. But because these causal chains are so very intricate, that would be very difficult for us.
So we make a genius move.
Information as a way of understanding
In DNA coding, a relatively simple molecule results in the creation of complex structures. In this way, it is akin to the mental activity of mathematics, in which we build complex arguments from simple axioms. Therefore, we link it to that activity. We interpret what is happening in terms of information, instructions and codes. We can understand these intricate causal chains better if we interpret them in terms of a conscious mind processing meaningful information.
But unobserved, DNA codes do not mean anything. They are not a message for anybody to read. They work because they cause certain things to happen. So there is no ‘information’ in there. We are the ones who interpret it as information, as a tool for us to understand what is going on.
I believe that information cannot exist in itself. It only exists in relation to an observer, for who the information is meaningful. Information is always information for (something or somebody). And I state that, since DNA coding is a blind, causal process, not involving any consciousness observer, there is no meaning.
Information and meaning
At this point, some readers might raise a criticism. Isn’t my definition a bit too narrow? Does the observer need to be a conscious human, in order to give a process meaning? Can’t the information in DNA be meaningful to the cells and biological processes of the organism, which need this information to function?
This is a valid point. DNA does contain meaningful information for the cells of the organism, even if there are no humans to observe it. But the fact that these processes are meaningful to the cells does not change the fact that at the same time, these processes are made of blind, meaningless causality, no different from a domino chain. Cell and DNA form a dumb, physical machine, from which smart behavior emerges.
Meaningful information processes are everywhere. They are meaningful to the organisms, cells and processes that ‘employ’ them. They allow their ‘users’ to display clever and complex behavior, indispensable to their preservation. But at the same time, they are nowhere. You cannot grasp information, because information is not a physical thing. Information is a way of understanding the world.
Computer science: a new way of seeing the world
I believe that this this new way of understanding causal processes, in terms of conscious, mental processes of information and calculation is computer science’s big idea. I believe that computer science is relevant beyond hard drives and PC’s, because it offers us a whole new way of understanding reality.
This is not to say that computer scientists were the only ones, or even the first to make the connection between physical, causal processes and meaningful information processes. But computer science has specialized in developing an informational look at the causal world.
In order to do that, computer scientists need to make one step above usual mathematics. Instead of focusing on knowing the right mathematical principles, they need to develop an understanding of mathematics as an activity itself. It is the process of calculation one needs to understand in order to interpret and understand other, physical calculation processes (luckily, the activity of mathematics lends itself to mathematical analysis pretty well).
That makes computer science different from other natural sciences. Other natural sciences use also use mathematics in their study of the physical world. But for them, mathematics has a more descriptive role. Mathematical formulas are used to create descriptions of causal, physical processes, but the process itself is interpreted and understood in causal terms. Computer science interprets the causal process as a logical, informational process. Computer scientists use their understanding of the activity of mathematical calculation to interpret what is going on.
Life, the universe and everything…
Understanding physical processes as information processes helps us to reason about all kinds of processes, like the aforementioned DNA coding, animal behavior, the limits to cognition and communication, and even the process of natural selection. It may also help us look at the creation of complexity in a more abstract sense. It may specify the conditions under which blind causal processes may form ‘positive feedback loops’, not only preserving complexity that was accidentally formed, but actually reinforcing it.
The creation of life and the process of natural selection is a nice example of such a positive feedback loop for complexity. Chemical processes may spontaneously create all kinds of molecular structures, most simple, some complex. Complex structures may be shaped in such a way that they display some ‘behavior’. This behavior may turn out to be self-preserving, but most likely, it not. Those that happen to show self-preserving behavior will continue to exist for some time. Eventually, these structures are likely to get destroyed again. But among the most complex of them, some may happen to be shaped in such a way that they self-replicate. Most of these self-replicating ‘families’ will probably also get wiped out, soon enough.
But of those families, the ones that are so complex that they do not only self-replicate but also have the ability to adapt themselves to their surroundings are more likely to stick around. And among those, the ones that have adapted themselves to their surroundings in the most sophistic way, will not only stick around, but will get multiplied. And among those, who multiply the, most sophisticated, adaptable ones will win out. And on and on it goes.
The chance that such a complex molecular structure is spontaneously created is very small . But if nature runs the dice long enough, even the more unlikely events will even happen. And when that happens, complexity can get over a certain threshold form where it keeps reinforcing itself.
Insights from computer science may help us understand these processes, explaining how and why the simple, physical laws of nature could have created the dazzling complexity of living nature, human beings and society.
 The chance may be small, but it might not be as small as previously suspected. Jeremy England (2013) has shown that the laws of thermodynamics predict that molecules driven by an external energy source (such as the sun) and surrounded by a heat bath, (like the ocean) will rearrange themselves in organized structures, providing a first step towards the complexity threshold.