Evidentiality and Involvement in the native languages of South-America: teasing apart semantic categories

In many South American languages some kind of evidentiality is expressed. However, in the existing grammatical descriptions of these languages the category of evidentiality is not always adequately described or even recognized. One of the problems is that the term evidentiality is used both with a broad interpretation and with a narrow interpretation. In order to grasp the general nature of this category, it is very important to determine its constitutive parts and tease out the different semantic categories that are involved in the expression of evidentiality. The relationship between evidentiality and epistemic modality has been quite well studied, whereas the relationship between evidentiality and involvement of the speaker has hardly been studied at all. In this paper I will make a first attempt to first establish and unravel this relationship.

In this paper, I use the terms ‘evidentiality’ and ‘speaker involvement’ as follows: semantically, evidentiality is used in a narrow sense to mean the marking of the source of information. When it comes to the form of this marking, the use of the term is much more broad. That is, not just grammatical marking of the source of information is included in this study, as Aikhenvald (2003) defines the term; any type of material, such as morphology, periphrastic constructions, syntactic constructions etc. can be used in order to express the source of information. The term involvement (of the speaker) is used as a semantic category to express the physical, mental or social participation of the speaker in the event he or she is describing. This can also be expressed formally in many different ways.

A first indication of the relation evidentiality and involvement of the speaker can be found in what Aikhenvald (2004:219) calls the ‘first person’ effect. These effects consist of the reduced involvement of the speaker when a less direct evidential is used. This is illustrated by example (1) from Tucano. Because of the use of the non-visual form –asï, the speaker shows that he was not consciously involved in the breaking of the plate. When the speaker applies a visual evidential as in example (2), he or she is suddenly consciously involved in the breaking of the plate.

This lack of involvement is not only found with first person subjects. In his doctoral thesis, Michael (2008) shows how ‘information responsibility’ and ‘event responsibility’ are tied together in the Arawakan language Nanti. That is to say, in Nanti indirect evidential forms are not only used to express that the speaker does not have direct information, but also to express that the speaker has no responsibility pertaining to the event; he or she was not involved in the event.

Both the ‘first person’ effects of evidentials and the use of evidentials to express (the lack of) event responsibility indicate that there is a relation between evidentiality and the involvement of the speaker. This paper will address different phenomena that indicate this relation. Understanding the interrelatedness of the different semantic components which constitute different aspects of evidentiality will undoubtedly give us a tool by which we can more readily identify and analyze evidentiality both conceptually and formally.


(1) bapá bopê-asï. (Ramirez 1997:133)
     plate break-REC.P.NVIS.NON3
‘I have broken a plate unintentionally.’

(2) bapá bopê-apï. (Ramirez 1997:133)
     plate break-REC.P.VIS.NON3
‘I have broken a plate intentionally.’



Aikhenvald, Alexandra Y. (2004). Evidentiality. Oxford University Press, Oxford.

Michael, L.D. (2008). Nanti evidential practice: Language, knowledge, and social action in an Amazonian society. PhD Dissertation. University of Texas, Austin.

Ramirez, Henri (1997). A fala tukano dos yepâ-masa. Tomo I: Gramática. Inspectoria Salesiana Missionária da Amazônia, CEDEM, Manaus.