Stuart A.Kauffman
Oxford University Press 



Investigations is, without a doubt, the strangest and most surprising pilgrimage of my scientific life. At the outset, in December of 1994, I began an ongoing notebook. 

Thinking of Wittgenstein and his shattering break with logical atomism in his Philosophical Investigations and sensing that something vast and inarticulate waited to be uncovered, I entitled that notebook "Investigations." For the entire winter, spring, and summer of 1995 and again in the fall of that year, after a month trekking in the Everest region, I struggled sideways, backward, and sometimes for-ward. A rough version was published as a Santa Fe Institute preprint in September of 1996. I returned to the topic a year later and have worked at it since then. Whatever Investigations is—useful, as I hope, or foolish—it is not normal science.

Perhaps the most astonishing aspect of Investigations, both as process and the resulting book, was my puzzled realization that the way Newton, Einstein, and Bohr taught us to do science may be incomplete. 

You see, in following their cornerstone examples of physics, we are taught to prestate the particles, forces, laws and initial and boundary conditions, then compute the consequences. 

In this enterprise, we are able to state ahead of time what the full space of possibilities is, that is, we can finitely prestate the configuration space of possibilities of the system in question. 

This capacity to prestate the configuration space, for example, is the central conceptual presupposition of the classical statistical mechanics of a liter of gas in its 6N-dimensional phase space of all possible positions and momenta of the N particles of gas.

But I was, to my deep surprise, led to doubt that we can ever prestate the configuration space of a biosphere. 

The core issue arises with what Darwin called "preadaptations," namely, causal consequences of parts of organisms that were not of adaptive significance in the normal environment of the organism, but might come to be of adaptive significance in some new environment and end up being se-lected by natural selection. Thus arose hearing, lungs, flight—virtually all major and probably most or all minor adaptations. But could we say ahead of time all the odd, context-dependent causal consequences of bits and pieces of organisms that might be of selective significance in some odd environment, hence, to come into actual physical existence in the biosphere? I think not. And if not, then we cannot finitely prestate the configuration space of a biosphere.

So the biosphere, it seems, in its persistent evolution, is doing something liter-ally incalculable, nonalgorithmic, and outside our capacity to predict, not due to quantum uncertainty alone, nor deterministic chaos alone, but for a different, equally, or more profound reason: Emergence and persistent creativity in the physical universe is real.

I think I am correct in the above statement. But then what are the implications for science and society? We have thought, in part, that the unfolding of society and culture could be brought under the sway of science. On the other hand, Sun Tzu, four centuries prior to Christ, and Clausewitz, in the early nineteenth century, in, respectively, The Art of War and On War, both stressed the totally unexpected ways of battle and the need for intuition and command genius, whatever the science lying behind strategy and tactics. 

Science and art—the practical getting on with it, wissen versus können in German, "know that» versus "know-how" in English, min-gle in our daily lives. Yet können, "know-how," has no place in our science. Why?

But that takes me to the starting questions of Investigations. Consider a bac-terium swimming upstream in a glucose gradient. We readily say that the bacterium is going to get food, that is, the bacterium is acting on its own behalf in an environment. 

Call a system able to act on its own behalf in an environment an "au-tonomous agent.» All free-living cells and organisms are autonomous agents. But the bacterium is "just" a physical system. In its Kantian form, my core question be-came, What must a physical system be such that it can act on its own behalf?

The stunning fact is that autonomous agents do, every day, reach out and ma-nipulate the universe on their own behalf. Yet that truth is nowhere in contemporary physics, chemistry, or even biology. So, what must a physical system be to be an autonomous agent?

I think I have found a plausible answer. I had not expected even the outlines of my answer. It may be that I have stumbled upon the proper definition of life itself. 

In the process, I have been led down strange pathways to a critique of the physi-cist's concept of work and work cycles—Maxwell's demon—toward the heart of a proper concept of organization that is not matter alone, energy alone, entropy alone, or information alone. In fact, a bacterial cell or colony is doing something that we cannot yet state clearly: 

The cell or colony is a 'propagating organization", that is, that it literally constructs more of itself. What the cell and colony are doing has no statement in current physics or biology but constitutes that which con-structs a biosphere.

Moreover, we are on the verge of the capacity to create novel molecular au-tonomous agents. When we do, or if we discover life on other planets and solar sys-tems, science will enter a vast new phase in which we will create a "general biology," freed from the limitations of terrestrial biology. 

Such a general biology will necessarily lift physics and chemistry up to new levels as these disciplines and biology struggle to understand what general laws may govern biospheres anywhere in the cosmos. And the experimental creation of molecular autonomous agents will un-leash a technological revolution equal to the computer revolution. We will create physical systems able to evolve by selection to carry out tasks we desire.

Investigations broaches four candidate laws for any biosphere. The deepest is re-lated to the following question: Might there be a fourth law of thermodynamics for self-constructing open thermodynamic systems such as biospheres? 

There are good reasons to think that no such law is possible. Yet I believe there may well be such a law: Biospheres, as a secular trend, that is, over the long term, become as di-verse as possible, literally expanding the diversity of what can happen next. 

In other words, biospheres expand their own dimensionality as rapidly, on average, as they can.

And the coconstructing behaviors of autonomous agents spill over to the econ-omy, with surprising implications for the foundations of economics, for economic growth, and for the development of adaptive firms that coevolve in corporate ecosystems whose dynamics almost certainly express the same laws as do biological ecosystems, with small and large gales of Schumpeterian creative destruction, weeding out old species and technologies, ushering in the ever new species and technologies whose nonprestatable features are expressions of the very creativity of the universe.

And if we cannot prestate the configuration space of a biosphere, how can we prestate the configuration space of the universe? But this bears on the problem of time in general relativity and has led me, in collaboration with Lee Smolin, a quan-tum gravity cosmologist, to quite unexpected ideas about how the universe might select its own laws and, somewhat like a biosphere, coconstruct itself.

No, Investigations is not normal science. I do know what normal science is. My previous two books, Origins of Order and At Home in the Universe, are examples. The first sought to exemplify processes of self-organization in biology and to lay out the need to rebuild evolutionary theory as a marriage of two sources of order in biology—self-organization and selection. Fitness landscapes, coevolutionary av-alanches, the unexpected ordered behavior of model genetic regulatory networks, the edge of chaos, models of the origin of life and morphogenesis, and bits of eco-nomics are to be found in Origins and in a latter-day form in At Home.

Nothing, however, had led me to expect even the outlines of Investigations. Nothing had led me to expect the answers I would struggle toward. And having completed Investigahons, I remain profoundly puzzled by what I have said, despite the fact that I think I am correct.

I suspect that Investigations poises us for a new understanding of what it means and is to know and make our world together. I suspect that Investigations broadens the scope of science, yet will demand of us a novel synthesis of wissen and können, science and art, and ultimately, science and civilization.