The last thirty years have significantly advanced our understanding of the descriptive and typological properties of evidentials (see inter alia Chafe & Nichols Reference Chafe and Nichols1986; Aikhenvald Reference Aikhenvald2004; Speas Reference Speas2008). Attention to the social and cultural aspects of evidentials in use has been less focused (although see Jaffe (Reference Jaffe2009) for an important sociolinguistic collection on the broader category of ‘stance’). Given the profoundly indexical nature of evidentiality as a category, Nuckolls & Michael thus provide a welcome contribution to work on the topic by bringing together a number of studies on a typologically diverse range of languages, united by a perspective that is grounded in language in use. This volume anchors an exploration of evidentiality in contexts as diverse as Kosovar electronic news sources and Quechua oral narratives, as well as in ordinary conversations in Nanti, Garrwa, and English.

This diversity of contexts is matched, in this collection, by a breadth of scope with respect to the notion of evidentiality itself. Early work in the domain (e.g. Palmer Reference Palmer1986) had suggested that evidentiality be considered as a form of epistemic modality (for a more recent perspective on the relationship between the two, see Cornillie Reference Cornillie2009), but Aikhenvald's insistence that the notion be confined to grammaticalization of information source (Reference Aikhenvald2004:6–7) has proven influential in recent typological and descriptive studies. But, as Hanks points out in his foreword to this volume, studying evidentiality in its home base of interaction virtually compels a broad perspective on evidentiality beyond discrete lexical and morphosyntactic categories to encompass evidential practices in general. In drawing on the Bakhtinian concept of voicing and Goffman's notion of footing to consider the case for evidentiality as a form of deixis, Hanks clearly plants the evidential flag on pragmatic territory. Within this broad domain, the collection as a whole reflects the entire spectrum of analytic positions—from those for whom the core and limit of evidentiality is grammatical, to those for whom evidentiality is more usefully reframed in terms of epistemic positioning.

So, for example, Michael's chapter on self-quotation in conversations in Nanti, spoken in the lowland of south-eastern Peru, investigates both grammaticalized quotatives and complement-taking verbs of saying; his analysis suggests that Nanti formally distinguishes reported speech constructions that serve to express speaker commitment from those that serve narrowly evidential communicative goals. This suggests that those reported speech constructions involve identification of informational source, but not illocutionary source per se (p. 183), thus upholding Levinson's (p. Reference Levinson, Drew and Wootton1988) proposal of a distinction between the two. Michael thus suggests that the data of Nanti is consistent with a cross-linguistic tendency for languages to functionally differentiate lexical quotatives from their grammaticalized evidential counterparts.

The chapters of both Nuckolls and Howard also contribute analyses of South American languages, the former investigating the evidential enclitics -mi, -she, and -cha in the Pastaza Quichua dialect of eastern Ecuador, the latter on a range of evidential and epistemic modal enclitics in the Huamalíes Quechua dialect of the central highlands of Peru. Nuckolls, consolidating her groundbreaking work on the relationship between evidentiality, deixis, and reported speech (Reference Nuckolls2008), uses the concept of the ‘speaking self’ (-mi) and ‘other’ (-shi) predominantly in narratives. She suggests that these enclitics ‘are not distinguishing between direct and indirect experience; rather, they signal the perspectives, whether self or other, from which the utterances are articulated’ (69), and so reiterates her position that an adequate understanding of these suffixes must bring speaker subjectivity and deixis into the analytic framework. In corroborating Nuckolls's finding for Pastaza Quichua, Howard's study of the Huamalíes Quechua focuses on a performed narrative involving interaction between a human protagonist and a spirit being, and tracks how evidential and epistemic markings grammatically structure the transition from conjecture to direct witness. She concludes that a speaker perspective framework has proven to be a productive way of moving beyond the ‘source of knowledge’ versus ‘validational’ debates; furthermore, she suggests that the patterning of evidentiality and epistemic modality in discourse may be primarily a matter of ‘the constitution of voice, identity, and social relations in interaction’ (p. 89). This emphasis on speaker perspective is also reflected in Friedman's chapter, which makes the rather ambitious claim that ‘the pragmatic deployment of a grammatical category… participated in the project of creating a politically independent Kosovo’ (p. 21–22). Using the data of Kosovar electronic news sources during three periods from 1994, Friedman analyzes the deployment of the Albanian admirative as well as evidential particles kinse ‘allegedly’ and gjoja ‘supposedly’. He contrasts these markers of nonconfirmativity with marked systems such as Macedonian and Turkish, and shows how the Albanian languages news sources used the admirative ‘to construct their perceived opponents as unreliable themselves as neutral and trustworthy’ (p. 53); furthermore, the use of the dubitative ‘invites the addressee to share in the speaker's disbelief of a third party, thus creating solidarity between speaker and addressee against a third party’ (p. 54).

The papers by Mushin and Sidnell fit the title of the collection most closely, dealing as they do with the data of ordinary interaction, and both are characterized by material showing a distinct lack of overt evidential strategies in contrast to, for example, some of those described in South American languages. Mushin notes that Garrwa, spoken in the southwestern Gulf region of Northern Australia, has no extensive system of evidential coding, because indexing source of information is not a primary concern in Garrwa culture. Mushin shows that issues of epistemic authority come into play only when knowledge is in dispute, with authority being justified on the basis of claims to witness. She has some fascinating ethnographic observations regarding how ‘the out-of-doors lifestyle has consequences for how epistemic authority is managed’ (p. 109) and, with so much of social life enacted in view, ‘particular significance is placed on witnessed happenings’ (p. 109). Mushin shows how the role of ‘witness’ must be negotiated and socially sanctioned, the analytic upshot being that we must seek a more precise analysis of where the boundaries lie between knowledge that requires a licensed witness and is accountable, and knowledge that is not accountable, such that anyone may be authorized to speak on it.

This role of knowledge, and public entitlement to knowledge, plays a central role in the chapter by Sidnell, which takes as its point of departure the seminal work by Heritage & Raymond (Reference Heritage and Raymond2005) on epistemic claims in conversation. Using the data of English interaction, Sidnell claims that epistemic modulations do not so much reflect the speaker's degree of certainty so much as the asymmetry that speakers assume to exist between what they know and what their recipients know. So Sidnell shows that, in one data fragment, ‘I suppose may be treated as a request for confirmation, and apparently may be a difference, not in epistemic access, but in rights to know (p. 129–30). Sidnell shows how epistemic positioning is ‘typically brought off en passant, in the course of whatever else it is that the participants are doing’ (p. 128). In this article the central grammaticalized versus nongrammaticalized evidentiality dispute falls away entirely: what evidential marking exists is to be found, not in the grammar but in the sequence itself, and in the construction of social action. Whatever one's stance on evidentiality, such a perspective displays a shift towards an increasing acknowledgement that there are sources of evidentiality and epistemic modality to be found beyond grammar.

It is increasingly the case that linguistic evidence from a range of languages is compelling us to reconsider the traditional distinction between grammatical and interactional phenomena; indeed, recent work by researchers such as Evans & Levinson (Reference Evans and Levinson2009) and Blythe (Reference Blythe2013) show how much social factors may shape grammaticalization. In bringing a diversity of data to bear on the issue of evidentiality, and in its focus on social meanings in its interactional contexts, this volume thus makes a timely and significant contribution to work in this domain.