Subjective experience is transformed into objective reality for societal members through cultural idea systems that can be represented with theory and data models. A theory model shows relationships and their logical implications that structure a cultural idea system. A data model expresses patterning found in ethnographic observations regarding the behavioral implementation of cultural idea systems. An example of this duality for modeling cultural idea systems is illustrated with Arabic proverbs that structurally link friend and enemy as concepts through a culturally deªned computational system. Computational systems also generate new concepts, as will be illustrated through a theory model for the structure of a kinship terminology system. This examples accounts for what otherwise appears to be an anomaly in the terminology thereby illustrating the constructive role that modeling of cultural constructs can play in ethnographic research.

Comparing the effect of statistical distributions on the outcome of the “treadmill” model relating average skill level achieved through imitation to demogaphic factors is premature as the model incorporates an invalid assumption. The model incorrectly assumes that the imitation bias remains constant with increasing average skill level and is contradicted by data on hunter-gatherer and oceanic fishing groups showing that for these groups there is no relationship between the interacting population size and tool complexity.

Richerson et al. propose cultural group selection (CGS) as the basis for understanding the evolution of cultural systems. Their proposal does not take into account the nature of cultural idea systems as being constituted at an organizational, rather than an individual level. The sealing partners of the Netsilik Inuit exemplify the problem with their account.

Kinship terminologies consist of the terms used to reference culturally recognized kinship relations between persons. These terms have been assumed to identify categories of genealogical relations (despite ethnographic evidence to the contrary), and kinship terminologies are classified using differences in genealogical referents of kin terms. Recent analysis, however, by building on ethnographically validated procedures for computing kin relations from kin terms without reference to genealogy makes evident the underlying generative logic for the structure of kinship terminologies. Making the generative logic of terminologies explicit provides a more rigorous comparative basis for the study of kinship terminology systems.

A mathematical model purporting to demonstrate that the interaction population size of a group of social learners is a primary determinant of the level of technological complexity achieved by the members of that group through imitation of the most skilled individual in the group has been proposed.  Empirical validation of the model has been attempted with archaeological data from Tasmanian hunter-gatherers and ethnographic fishing data from Oceania, but these data do not support the model. Data from a wide variety of hunter-gatherer groups show, instead, that implement complexity varies with an interaction effect between risk and number of annual moves and not with the interaction population size. Data from the Polar (Inuit) Eskimo and the Angmaksalik Inuit on the east coast of Greenland show that complex implements were part of both group’s technological repertoire even though each had interaction population sizes limited to a few hundred individuals, in direct contradiction with predictions from the mathematical model. The problem with the model lies in an invalid assumption.

Alex Mesoudi observes that the “treadmill” model for relating average skill level achieved through imitation to demogaphic factors implies, incorrectly, that skill levels will increase without bound.   Rather than correcting the structural problem with the model, he applies a “Band-Aid” in the form of an imitation cost to the model to force the average skill level to reach a plateau.  The model, however, incorrectly assumes that the imitation bias remains constant with increasing average skill level and, when corrected, average skill level will reach a plateau even without the added cost factor.