Biology Reference
In-Depth Information
and getting away with it without changing the entrenchment are virtually zero,
unless there is an array of changes that are already neutral (or nearly so) at
that level of organization (Huynen et al., 1996; Wagner, 2005). To be neutral
any change will usually have to meet an increasing number of constraints that
generate the upper-level property that must be preserved in the face of microlevel
variations in composition and process. Though the selective consequences of
making changes thus are not quite intrinsic properties of the network (fitness still
is a relation between system and environment), essentially no changes external
to the system in the environment or selection regime will change the outcome:
deep generative entrenchment is thus an 'effectively' intrinsic property of that
node in the network. But as changes in the network can cause major changes
in entrenchment we should be especially interested in structural or dynamical
changes in the network which change the entrenchment of the network element
in question. And we are interested in the connectivity patterns in networks both
in general and in detail. Significant changes in generative entrenchment can
emerge from addition or subtraction of a single connection that moves that part
of the circuit in the direction of local integration or parcellation (Schank &
Wimsatt, 2000).
How general is this? We saw that differential dependencies of components in
structures - causal or inferential - are inevitable in nature. And in the symmetry-
breaking, we saw that those that have some tend to get more. Their natural
elaboration generates foundational relationships. New systems in which some
elements play a generative or foundational role relative to others are always
pivotal innovations in the history of evolution, as well as - much more recently -
in the history of ideas. Mathematics, foundational theories, generative gram-
mars, and computer programs attract attention as particularly powerful ways
of organizing complex knowledge structures and systems of behavior. This is
a principle of great generality, going well beyond biology to evolved systems
generally. Generative systems would occur and be pivotal in any world - bio-
logical, psychological, scientific, technological, or cultural - where evolution
is possible. Generative systems came to dominate in evolution as soon as they
were invented for their greater replication rate, fidelity, and efficiency. We must
suppose that even modest improvements in them spread like wildfire. Combina-
torial generative power like that found in the genetic code, the immune system,
and languages of all sorts (spoken, visual, and written) add another important
dimension of amplification best treated more fully on another occasion. Infor-
mation (contrary to the reductionistic talk of replicator and meme theorists) is a
system property, and thus properly leads back to a properly formulated systems
biology.
But no runaway processes are unbounded for which we should be thankful,
else we would not be here, buried under a heap of Darwin's elephants or some
