Analysis of a particular conversation held in Fiji suggests that we know what is true primarily by the evidence of our own and others' eyes. Even so, in this Fijian case, truth is not necessarily a priori, it is not always given in the nature of things; in certain cases - witchcraft, for example - what is true remains to be found out. Ditto for whether or not a person's words or acts are mana, that is to say, materially effective. The Fijian idea is that speech or, more generally the word (vosa) as it is spoken or written, may be mana, effective, and thus what is true (dina) may be an outcome rather than already given. The objective of the paper is not to translate Fijian terms, but to render them analytical. The argument here is that because the environing world provides for all the meanings that humans can make, our own understandings (as anthropologists and human beings) are just as amenable to historical analysis as the next person's; it follows that the explanatory power of our ethnographies must be made to reside in rendering our informants' categories analytical. Finally, the paper makes use of the Fijian material to argue that ideas about what language is good for (its moral force) are crucial to understanding what is said and suggests the analytical utility of laying bare the ontogeny of the moral force of language in any given case.
Ontogeny; Knowing; Truth; Mana; Witchcraft