At one time I believed it possible to simultateously investigate issues of adaptation that straddle evolutionary, cultural and within-lifetime change, and I developed one simulation towards this purpose (cf. , esp. Sect. 7.3). In part because of this experience, however, I have become convinced that truly ``God dwells in the details'' and little progress can be made searching for (necessarily lowest-common-denominator) features shared across such disparate phenomena as genes, neurons and cultures. As a consequence, my two research foci - adaptation by culturally sensitive IR systems, and interactions between GA evolution and NNet learning - have since remained quite distinct. At the same time, techniques discovered in one arena have often benefited the other, and there are often striking parallels between the two phenomena. In one recent example, we have begun to consider the World Wide Web (WWW) itself as an ``environment'' analogous to those defined within LEE, and potentially admitting similar solutions .
I take successes like these as indications of a useful synergy across examples that I hope to pursue towards a general theory of adaptive representation.