Ever since Darwin's earliest remarks on the uncanny similarity between language change and natural history in biology, there has been a persistent conceptual unclarity in evolutionary approaches to cultural change. This unclarity concerns the units of analysis. In some cases the unit is said to be the language system as a whole. A language, then, is "like a species" (Darwin 1871 et al). One reading of this is that we are working with a population of idiolects that is coterminous with a population of bodies (allowing, of course, that in the typical situation – multilingualism – one body houses two or more linguistic systems). On another view, the unit of analysis is any unit that forms part of a language, such as a word or a piece of grammar. "A struggle for life is constantly going on amongst the words and grammatical forms in each language" (Müller 1870, cited in Darwin 1871:60). In contrast with the idea of populations of idiolects, this suggests that there are populations of items (akin to Zipf's economy of word-tools), where these items are reproduced, and observed, in the context of spoken utterances.

While some of us instinctively think first in terms of items, and others of us in terms of systems, we do not have the luxury of ignoring either. Neither items nor systems can exist without the other, and the challenge is to characterize the relation between the two – this relation being the one thing that defines them both. The issue is not just the relative status of items and systems but the causal relations between them. If the distinction between item and system is just a matter of framing, it is no less consequential for that. We not only have to define the differences between item phenomena and system phenomena, we must know which ones we are talking about and when, and we must show whether, and if so how, we can translate statements about one into statements about the other. In this talk I discuss a causally explicit model for the transmission of cultural items, and approach a solution to the item/system problem that builds solely on these item-based biases. I claim that the biases required for item evolution – never forgetting that "item" here really means "something-and-its-functional-relation-to-a-context" – are sufficient not only to account for how and why certain cultural items win or lose, they also account for the key relational forces that link items and systems.