452 
David Zeitlyn 
Anthropos 99.200 4 
discussion of the comparative study of personhood 
(Strathern 1992) and to help understand both sac 
rifice and prayer (Hubert and Mauss 1964). For 
believers prayers are answered (see Goody 1995). 
Conversely, Barber (1981) documents how a per 
ceived failure by the gods to reciprocate led to 
their worshippers threatening them with repudia 
tion: the gods have “failed to deliver” and may 
be spumed as a result! In another work, which 
is significant for my argument, Bourdieu (1977) 
elaborated Mauss’s original argument, emphasiz 
ing both the sensitivity of timing in gift giving 
and the intricacy with which it is involved with 
differential relations of power. As the later discus 
sion of conversational structure will show, there 
are clear sociolinguistic parallels with the social 
correlates of gift giving which Mauss first brought 
to our attention. Although occasional allusions are 
made to the linguistic sphere (e.g., Godbout cited 
above), it has not been discussed in much detail in 
the subsequent literature on the Gift. 
Conversation Analysis: 
Reciprocity in Conversation 
With such ideas in mind I turn to the analysis of 
conversation, of language as it is used by human 
actors. I draw on a body of work within the field 
of pragmatics known as conversation analysis (see 
Levinson 1983 for an introduction). Later, I sketch 
two features of conversation, which have been 
shown to be globally applicable, and discuss how 
these can be helpful to anthropological analysis. 
Pragmatics in general, and conversation analysis 
in particular, give pointers to key moments in 
social action when features of a social system are 
easier to understand. They provide anthropologists 
with a valuable method in our goal to understand 
the outlines of social process. 2 This is principally 
through a view of conversation as structured by 
a series of reciprocities (as explained in more 
detail below). The key assumption is the idea of 
local organisation: the details of conversational 
stmcture are the result of the actors (consciously 
or semiconsciously) responding to what others are 
saying in ways that are principled and predictable, 
so that the fellow conversationalists can use these 
responses to tailor their own contributions to the 
conversation (see Schegloff 1997 for a summary 
introduction). 
2 The debate between Schegloff (1997; 1999a; 1999b: espe 
cially 563) and Billig (1999a, 1999b) has clarified some of 
these issues. 
As an anthropologist rather than a linguist I use 
conversation analysis as a means to an end rather 
than the ultimate purpose. 3 Even a stalwart of con 
versation analysis such as Emanuel Schegloff has 
reflected that “the entire range of non-recognitional 
reference - is still an immense territory, and one of 
deep importance to sociology, perhaps even more 
than it is for linguistics” (1996a: 464 f.). Recall 
that Goody (1978) discusses the profound ambi 
guity of questioning both as linguistic feature and 
as social action: at one level to ask a question of 
anyone is to question their social role. (Hence, the 
practice of anthropology is socially challenging!) 
A casual conversation between people who 
have known each other all their lives contains 
much for which the coconversants can safely rely 
on tacit knowledge. 4 Often it is not easy for by 
standers, overhearers (or anthropologists) to follow 
the interchange. In effect ethnographic knowledge 
is needed order to unpack that tacit knowledge and 
to trace out some of what the speakers understand. 
Without wanting to claim that as anthropologists 
we can have the same or as complete an under 
standing as the conversationalists themselves, we 
can use the interchanges of everyday conversation 
as the foundations for ethnography or, to change 
the metaphor, as pegs on which to hang ethnogra 
phy. Moreover, the conversation can act as warrant 
for the ethnography: the overheard words make 
sense in the light of the ethnography, the speakers 
as social actors use, assume, or tacitly invoke the 
social structures which the anthropologists seek to 
make explicit. 
Two Mambila Examples 
Reflection on reciprocity in conversational ex 
changes led me to reexamine some of my own 
data. Here, I examine two passages from a 48-min- 
ute recording of a conversation that took place in 
December 1990. I have been using the transcript 
of the conversation as part of a research project 
3 I am not the first to take this approach. As well as 
the contributors to “The Ethnography of Communication 
(Gumperz and Hymes [eds.] 1986) other anthropologist 8 
have used ideas taken from conversation analysis. Ae 
should acknowledge Goldman’s analysis of Huli disputes 
(1983), Sherzer’s work on the Kuna (1983), the work of 
the Tedlocks (B. 1982 and D. 1983), Moerman’s work 
on Thai conversation and social structure (1988), Bihu eS 
(1996) also on Thai, Hanks (1990) on Maya deixis and 
Durant! on Samoan village politics (1994). 
4 This is another version of Bernstein’s (1973) distinction 
between “restricted” and “elaborated” codes, although he 
finds them in differences between social class.