Epistemics
| Encyclopedia of Terminology for CA and IL: Epistemics | |
|---|---|
| Author(s): | John Heritage (UCLA, USA) (https://orcid.org/0000-0002-8603-6447) |
| To cite: | Heritage, John. (2026). Epistemics. In Alexandra Gubina, Elliott M. Hoey & Chase Wesley Raymond (Eds.), Encyclopedia of Terminology for Conversation Analysis and Interactional Linguistics. International Society for Conversation Analysis (ISCA). DOI: [ ] |
Epistemics is a topic of conversation-analytic inquiry, focusing on the knowledge claims that are asserted, contested and defended in social interaction. These claims are observable in the ways that persons design and implement social actions and sequences of them (Heritage 2024, see also Wootton 1989). The term derives form the Greek episteme meaning knowledge or understanding, and is the subject matter – epistemology – of a branch of philosophy. Research on epistemics is, together with the analysis of other dimensions of stance, especially deontics and affect (Stevanovic and Peräkylä 2014; Wu, Heritage & Raymond 2025), a central element drawing together the analysis of social interaction with sociological conceptions of identity, role, institutions and social structure.
Epistemic issues are omnirelevant and ubiquitous features of social interaction (Heritage 2018) and are universal across the societies of the world. (The capacity to track the knowledge and social connections of others has also been argued to underly the evolutionary expansion of brain size relative to the size of social groups in hominids and other primates [Dunbar 2003]). From the point, normally reached between the ages of four and eight, that children are able to recognize differences between their own and others’ knowledge of the world (Astington 2006), they enter a world structured in terms of epistemic distinctions. (It seems likely that the representation of others’ beliefs can be traced to the functioning of single neurons [Jamali et al. 2021] where the latter may be an evolved feature of human sociality.) For example, in referring to a neighbor a person might simply use a first name, Mary, or a variety of more complex reference forms: Mary Smith; Mrs. Smith; John’s Mom; Jane’s daughter; Linda’s aunt; David’s girlfriend; my mother; your aunt; Fido’s owner; the woman who drives the white BMW; that neighbor of ours… and so on. While some or many of these expressions may be correct, the choice between them is most centrally shaped by what the recipient is presumed to know about ‘Mary’ and how she knows it (Schegloff 1972). The choice, in other words, is one made on epistemic grounds.
The kinds of choices just described involve speakers in taking account of what others know in designing their turns at talk. Other kinds of epistemic choices involve the evidential calibration of knowledge claims in terms of, for example, their sources in firsthand knowledge, inference or hearsay (Aikhenvald 2004, 2021). 57% of the 400+ languages surveyed in the World Atlas of Language Structures Online (WALS) impose this distinction as a grammatical choice on their speakers (de Haan 2013). Another variation involves epistemic modality (Palmer 2001) in which claims can be adjusted as more or less certain as in ‘He may have robbed the bank’ (van der Auwera & Ammann 2013). And in egophoric languages the relative access to knowledge of the participants in relation to some state of affairs is expressed in grammaticalized form (Grzech & Bergqvist 2025; Sandman & Grzech 2022; San Roque, Floyd & Norcliffe 2018; see also Evans, Bergqvist & San Roque 2018a, 2018b).
Within CA, interest in epistemic questions mainly focuses on circumstances in which interactants have asymmetric epistemic status or access to knowledge. Interactants are conceived as differentially informed about the topics their conversations touch upon. The prototypical case is one in which A asks B a question, reflexively framing A as lacking some form of knowledge (K-) and reciprocally constituting B as the knowledgeable (K+) party (Heritage 2010). These asymmetries may be rooted in different life experiences, including personal memories, expectations, desires and hopes, or in institutionally sanctioned knowledge and roles that are found in the medical and other professions, or simply that A saw an event whereas B only heard about it, or saw it more recently than B, or even something as ephemeral as A just saw something a moment ago that B didn’t. The differential may not be restricted to actual knowledge, but may also involve rights to know it, or even to articulate it. For example, it is well documented that patients may possess a good deal of knowledge about their medical conditions, but may find themselves unable to articulate it in the doctor’s office with any degree of authority.
A general feature of social interaction is that persons design their epistemic stance in talk to match their epistemic status. This epistemic stance is indexed using a very wide-ranging set of methods that are fitted to the actions and sequential contexts in play. The fit between epistemic stance and status is relied upon by the recipients of talk as an element in determining the social action to which they should respond. For this reason, epistemic issues absolutely permeate other domains of conversational organization including, most importantly, action formation (Heritage 2013, 2022) where arguably the most fundamental action distinction in interaction – between requesting and giving information – hinges upon epistemic status (Heritage 2012a), which can readily override the epistemic stance otherwise conveyed by grammatical form and other aspects of turn design. Relative epistemic status is arguably a vital support to the rather unreliable (Levinson 2012) use of intonation to mark questions that is depended upon by about 18% of the world’s languages (Dryer 2013a), not to mention those languages for which there are no attested means of marking actions as questioning. It may be added that, in a context in which early determination of the social action being implemented by an incoming utterance is a central feature of response construction (Levinson 2013; Levinson & Torreira 2015), epistemic status may be a vital early clue in the process (Heritage 2022). Similar considerations may apply to the 314 languages in which question particles are deployed to signal polar questions in sentence-final position (Dryer 2013b) – long after the construction of response has been initiated.
Our understanding of the significance of epistemics for action formation and sequence in interaction progressed rapidly following the publication of work by Heritage and others (Heritage 2012a, 2012b; Stivers et al. 2011). In particular, the recognition that epistemic status is central to participants’ ability to discriminate between assertions and questions – the two primary conversational actions undertaken in human interaction – has become important in the analysis of action formation, sequence organization and repair. Critics have claimed that the concept of epistemics is ‘cognitivist’ and refers to a “hidden order”, discernable only to the analyst (e.g. Lynch and MacBeth 2016), critiques which have been effectively rebutted in publications demonstrating that epistemic phenomena are a basic order of social organization (Drew 2018; Heritage 2018, 2024; Whitehead et al. 2025) alongside turn-taking, sequence organization, repair and others identified by Schegloff (2006). Thus, these critiques notwithstanding, epistemic phenomena continue to be a primary and heavily research topic of investigation among interaction researchers.
For example, from the earliest studies in CA onwards, epistemic issues have been seen as crucial to a wide variety of interactional processes. For example, Lerner (2003, see also Stivers & Robinson 2006; Raymond, Robinson and Bolden 2027) showed that epistemic status is implicated in next-speaker selection: K+ recipients are selected to speak next in contexts where there is differential knowledge access among potential recipients. In 1979, in a classic paper, Charles Goodwin showed that speakers will design and re-design turns to accommodate different recipients’ presumed knowledge. Again, numerous analysts from Sacks onwards have explored the centrality of epistemic issues in news delivery sequences, under the auspices of a general norm that bars speakers from telling things that recipients already know (Maynard 2003; Pomerantz 1980; Terasaki 2004). As noted above, Sacks and Schegloff showed the centrality of identifying the knowledge of a recipient as a means to design a recognizable reference to persons (Sacks & Schegloff 1979; Schegloff 1972, 1996, see also Heritage 2011). Epistemic issues have similarly been found to infuse many aspects of repair sequences, including self-repair (Drew, Walker & Ogden 2013; Raymond & Gill 2025) and responsibility for other-initiated repair (Robinson 2006, 2013). Finally, epistemic issues are central to many aspects of a very wide range of interactions in institutional contexts (Drew & Heritage 1992; Heritage and Clayman 2010).
Additional Related Entries:
Cited References:
Additional References:
