[1] Hans-Cees Speel is research assistant Policy Analysis at the Delft University of Technology, School of Systems Engineering, Policy Analysis and Management. Jaffalaan 5 PO box 55015 2600 GA Delft, The Netherlands. E-mail hanscees@hanscees.com

[2] In fact this is just one of the ways variation in physical genes can arise. Other processes involved are re-combination of already existing variation by means of sexual reproduction for instance, or mutations in the mould resulting from mutagens, uv-radiation, etc., etc.

[3] If I use the word physical gene, I am referring to a chunk DNA that codes for a protein, not to a gene as defined by Williams (1966) . He, as well as Dawkins, defines genes as the entities that have effect in the selection process: a gene is 'any hereditary information for which there is a favourable or unfavourable selection bias equal to several or many times its rate of endogenous change'.

[4] According to the Concise Oxford dictionary of Ecology (Allaby, 1994 ) gene pool is: "The total number of genes or the amount of genetic information possessed by all the reproductive members of a population of sexually reproducing organisms'.

[5] In describing non-selective situations, the concept niche is useful. A biological niche is characterized by a relative lack of selective pressure for a specific species. Similarly a religion that has no opponents has discovered an empty niche.

[6] This is a weakness of the words replicator and interactor. These words seem to refer to active entities, thus to matter, and not to symbols (Pattee, 1974 ). However, in a selective event or in a case of interaction, it can be far from clear whether interactors actually do anything. In the same way replicators like viruses do very little. Their DNA or RNA is replicated and translated by the cell-machinery of the hostcells. In fact we can ask if DNA in general does anything. To do anything DNA needs enzymes, that actually do things. The fact that DNA codes for these enzymes does not change this. If we separate DNA from these enzymes, nothing will happen. In the same way, if we write a book and nobody reads it, nothing happens.

p [7] In most cases evolution in biology means evolution by natural selection, involving a kind of learning-process. For example May (1978) begins an essay on the evolution of ecological systems with the phrase: 'Strictly speaking, ecological systems as such do not evolve. As has been stressed repeatedly in this issue, natural selection acts almost invariably on individuals or on groups of related individuals. Populations, much less communities of interacting populations, cannot be regarded as units subject to Darwinian evolution. ....Therefore in a sense constellations of species can be viewed as evolving together within a conventional Darwinian framework.'dt It is clear that May takes 'evolution' to mean evolution by natural selection, and this is probably common for most biologists. This evolution acts on interactors, producing adaptations in the species; only the species evolves in this sense, being the selective retention system.

[8] An example of a strategy with a body of knowledge. The approach I have developed was with a real life situation in mind. The strategy concerning environmental policy on a national level in the Netherlands has been published under the name National Environmental Policy Plan (NEPP) by the Ministry of Housing Planning and Environment (1989) . Since this first version there have been two more generations.