Editor's Note: This is a supplement to "The Evolution of Future Wealth"

At first, the insight that predicting future innovations in goods and services is fundamentally impossible might sound deceptively obvious. It is after all the common experience of mankind that we do not know exactly what tomorrow holds. But what our work reveals is that the depth of our ignorance about certain types of future developments is far more impenetrable even in principle than one commonly supposes, and the larger significance of that finding for sciences such as biology and economics, if true, is profound.

Going back at least as far as Galileo and Descartes, if not Pythagoras, science has generally viewed the universe as understandable, at least in theory, in terms of a finitely statable number of natural laws or principles. In that earlier reductionistic view, the universe was thus like a vast deterministic machine, and if one knew the initial states and positions of all its atoms and all the applicable laws and boundary conditions governing their interactions, one could predict all future states of the universe. While deterministic chaos has limited predictability, and quantum mechanics has eliminated the determinism of this earlier reductionistic view, the belief that the universe is fully describable by natural laws remains our scientific worldview.

For deterministic systems, even if one could not know exactly what would happen in practice (because of inexact or otherwise incomplete knowledge of earlier states), well-formed probabilistic estimates of the most likely outcomes might still be possible. Thus, even if we cannot know whether a particular flipped coin will end up heads or tails, we know that its end state will surely be either heads or tails and even what the likely distribution of heads and tails will be over any repeated set of tosses.

Implicit in that probabilistic analysis, however, is an assumption that all the possible outcomes can be stated in advance. Without foreknowledge of this sample space, one lacks the basis for compiling the appropriate statistics. Systems for which the sample space cannot be prestated seem to preclude the formation of meaningful probability statements.

As described in the main text, we believe that economic systems fall into this category. One cannot anticipate all possible future innovations by simply listing all possible combinations of prior ones: it is impossible to know what features of those earlier goods might be useful in the future because new goods and services can invent completely novel uses of old ones (and sometimes eliminate old ones).

Ample reasons further suggest that the ceaseless creativity of the universe similarly manifests itself throughout the biosphere. (Indeed, it was the Darwinian notion of preadaptations and their significance in evolution that inspired much of our thinking about economics.) Living things and their parts--whether at the cellular, anatomic or organismal level--can have causal consequences of no selective significance in their current environment but which might prove highly advantageous under new circumstances.

For example, some early species of bony fish had lungs that assisted their gill systems in providing oxygen to their tissues. Because these lungs were filled with air, they incidentally affected the fishes’ buoyancy. Consequently, those primitive lungs were preadapted to evolve later into swim bladders, which help fish control their neutral buoyancy in the water column (and indeed, lungs have vanished from most fish species because they are no longer necessary, whereas swim bladders remain widespread). It seems impossible to finitely prestate all possible Darwinian preadaptations for all living species, or even humans. We cannot prestate the adjacent possible of the biosphere.

2 Comments

1. 1. paolomagrassi 08:38 AM 5/10/08

   I would like to inform the Authors that the "post-cartesian view of science beyond pure reductionism" has now been in effect for about a century.
   
   Non-determinism, chaos, complexity and emergent properties have been identified in "hard" science as early as at the end of the XIX century by (among others) Poincaré, Weierstrass and Boltzmann, then studied by, among others, Hadamard, Cantor, Schroedinger, Prigogine, Lorenz, Mandelbrot (in addition to Maturana, Varela, and many other "soft" science champions).
   
   This is the classical dualism between hard science on the one side and soft science on the other. Soft science is often more complex (see, e.g., biological systems or economics) but unfortunately its exponents are rarely equipped with the hard background (mathematics, advanced systems theory) that would be required to tackle such complexity.
   
   Both parties should exert more humility and approach the others' field with rigor -as opposed to just passion, as it is often the case.
   
   --
   Edited by paolomagrassi at 05/10/2008 1:42 AM

   Reply | Report Abuse | Link to this

2. 2. paolomagrassi 02:07 PM 5/13/08

   See http://science-community.sciam.com/blog/Paolomagrassis-Blog/580004022

   Reply | Report Abuse | Link to this