Entelenomic Search Theorem

Entelenomic Search Theorem.



Let an adaptive system occupy state sts_tst​ in an environment ete_tet​.

Suppose the environment has temporal structure such that I(et;et+Δ)>0I(e_t;e_{t+\Delta})>0I(et​;et+Δ​)>0,
and suppose the system contains an internal state hth_tht​ satisfying I(ht;et+Δ)>0I(h_t;e_{t+\Delta})>0I(ht​;et+Δ​)>0.

If state transitions are biased by hth_tht​ so that expected future viability increases,
then the expected hitting time to a future-adaptive basin is lower than under
an otherwise identical blind mutation-selection process.
Thus, predictive developmental architecture can generate directional adaptive
search without invoking external teleology.