In the last decade, sporadic attempts have been made to bridge the gulf between cognitive anthropology and various modes of symbolic/interpretative anthropology. It is no surprise that these efforts have achieved only limited success, given the seemingly irreconcilable differences between the philosophical/epistemological premises of scholars on each side of the divide. On one side, cognitive anthropologists mainly share with other cognitive researchers assumptions that approximations to truth regarding the organisation and processing of knowledge (in our society or in others) can be achieved by applying inductive and experimental methods: by collecting and rigorously analysing data. On the other side, symbolists take anthropology's task to be mainly interpretative, more akin to literary criticism than physics; and they regard the empiricism and scientism of cognitive anthropology as inappropriate to the study of humans and their cultural traditions.

The divide, with its historical origins in nineteenth-century continental social thought, remains wide. However, some productive discourse has been taking place across the chasm, as cognitive science becomes less narrow and in some respects, less confidently messianic; some artificial-intelligence researchers and linguists now look to phenomenology for insights.

Anthropological attempts to apply recent developments in cognition and semantics to complex and symbolically salient cultural categories, going beyond the older preoccupations of cognitive anthropology with folk classification, invite further dialogue, in both directions. Here, I will suggest that recent developments in the study of categorisation (especially the work of Lakoff on conventional metaphor and prototypy) give anthropologists the means to make more systematic and theoretically grounded sense of ‘symbolic’ meanings. I take as texts the semantics and symbolic ramifications of two categories in Kwaio, an Oceanic Austronesian language spoken on Malaita, Solomon Islands.