Indologie & Vergleichende Religionswissenschaft

'Evidentiality' in Tibetic languages and beyond – a closer look

International workshop
16-17 February, University of Tübingen, Fürstenzimmer, Schloß Hohentübingen

⇨  Program      Abstracts   ⇨  Flyer

Evidentiality is commonly described as the marking of the source of information (firsthand vs. non-firsthand) or also as the discrimination between direct knowledge through sense perception, on the one hand, and indirect knowledge, namely inference and hearsay, on the other. The complex of indirect knowledge has also been addressed in the French literature with the term 'médiative'.

The modern Tibetic languages are known to have developed a particular type of ‘evidential’ marking. The basic principles have been described for quite a few of the Tibetic languages, see here the recent volume Evidential Systems of Tibetan Languages, ed. by Lauren Gawne and Nathan W. Hill. (Trends in Linguistics. Studies and Monographs [TiLSM] 302, De Gruyter Mouton, 2017) as well as the earlier collection in Person and evidence in Himalayan languages, ed. by Balthasar Bickel. (Linguistics of the Tibeto-Burman Area, 23.1-2, 2000). One of the key features is the subjective involvement of the epistemic source or ‘ego’ (the speaker in statements and the addressee in questions) in the events reported. The ‘system’ is thus also known under the key terms of ‘egophoricity’ and ‘conjunct/disjunct’, both concepts often mistaken for a somewhat weird syntactic person category (ego vs. non-ego).

However, at a closer look, the ‘evidential system’ is far from being a ‘frozen’ grammatical system, but is extremely flexible, allowing, in principle, all forms for all persons, albeit in different frequencies and for different motivations. It further does not only deal with the source of information (firsthand vs. second-hand/ hearsay) or the access channels (self-centred knowledge, perception, and inferences), but also or even predominantly with the subjective assessment of the situation and/or the socio-pragmatic situation.

In this workshop, we want to discuss the ‘unsystematic’ aspects of this ‘system’. A closer look into the genesis of these systems (Zemp, Widmer) and into the hierarchical structure of access channels (Tournadre) may help to explain the idiosyncrasies. The subjective involvement of the epistemic source and the pragmatic restrictions in the speech situation may define the attitude or Stance the speaker is willing or allowed, and the addressee in questions is expected, to take (Sandman, Oisel, Zeisler). This also includes the question of how much does the addressee already know and how does it matter (Simon, Zeisler). We will also look at languages at the periphery of the Tibeto-sphere, that is at languages that have been under the influence of Tibetic languages (Sandman, Widmer). The contrast with the closest 'evidential' neighbours, the Iranian languages (Pezechki) with their mediative system will help to improve our understanding of the specificness of the Tibetic system.

As suitable for a workshop, there will be plenty of time for the presenters to develop their arguments and for discussions.

Abstracts will be posted soon.





Bettina Zeisler, Universität Tübingen,
DFG-Project Evidentiality, epistemic modality, and speaker attitude in Ladakhi - Modality and the interface for semantics, pragmatics, and grammar


The genesis of evidentiality in Tibetan

Marius Zemp

Universität Bern




The evolution of epistemic categories in Bunan

Manuel Widmer
Universität Zürich




Evidential accessibility hierarchies.

Nicolas Tournadre
Université d'Aix-Marseille, Lacito/CNRS, Institut Universitaire de France;  




Evidential categories in Iranian languages: The core categories of inferential and hearsay.

Homa Lessan Pezechki

Université d'Aix-Marseille






Ego evidential -yek as a stance marker in Wutun: evidence from conversational data

Erika Sandman

Helsingin Yliopisto (University of Helsinki)


Negociating the facts: interactional functions of factual evidential markers in Amdo-Tibetan

Camille Simon
Lacito/CNRS, Universität Heidelberg




Evidential freaks in Lhasa Tibetan – Redefining evidentiality in Tibetic Languages

Guillaume Oisel

Universidad Nacional Intercultural de la Amazonia, Peru




Speaker attitude (Stance) and other 'freaks' in the Ladakhi 'evidential' system
Bettina Zeisler




Further discussion


Oisel  Pezechki  Sandmann  ⇨  Simon  Tournadre  Widmer  Zeisler  Zemp

download abstracts

Oisel, Guillaume

Evidential freaks in Lhasa Tibetan – Redefining evidentiality in Tibetic Languages

This lecture aims to present the results of the paper called “a re-evaluation of the Lhasa Tibetan Evidentials and its atypical functions with control verbs” (Oisel, 2017). I show that Evidentiality is not a frozen grammatical system which only deals with the source and access to information. It is a pragmatic and cognitive system of persuasion. Evidentiality is closely intertwined with the degree of speaker’s intention. Intention can be triggered by different states of mind and/or degrees of awareness of the speaker. Mental states at work here can be the case of retrospection, recognition, novel realization and mental transfer. In function of these parameters, the speaker, as a speech act participant (SAP), is willing to show his own subjective point of view, and/or emotional state in order to get a specific behavioral answer from his/her addressee (non-SAP).

In the first place, I remind the reader the so-called normal evidential system of Lhasa Tibetan and I also suggest to include new categories. Then, I focus on the so-called Evidential freaks or its specific uses. In order to do so, I first present the use of the intentional egophoric with non-SAPs and control verbs when the speaker refers to personal knowledge and I discuss some of its restrictions.

Then, I present other evidential freak behaviors with the atypical uses of the sensorial, factual and inferential. I show that when the degree of intentionality is either not involved, but not unintentional, or is only partly involved, these evidentials can be used with the SAP and control verbs. I also present the notion of intentionality out of focus and lower intentionality to describe these two cases. Then, I treat ‘intentionality out of focus’ in greater detail, showing that one can distinguish five different ways of reducing the focus on intentionality.

Pezechki, Homa Lessan

Evidential categories in Iranian languages: The core categories of inferential and hearsay

The notion of médiativité or ‘evidentiality’ in Persian has not been the subject of many studies. The term médiative was proposed by Gilbert Lazard in 1956 in an article on Tajik. Before this, the Iranian grammarian Kasravi who died in 1946 proposed the term nâ-did-e "non-witnessed" (litt. un-seen) to designate these forms in Farsi.

The terms evidential and médiative are close but the latter is often used in a narrower sense than evidential to refer to a distinct source of information or indirect speech.

In Farsi and Tajik, the evidential paradigm or vajh-e bardāšti ‘inferential mood (Pezechki, 1997) is essentially restricted to the past where it is opposed to the indicative mood, vajh-e exbâri. The opposition between evidential and non-evidential is attested for 4 paradigms in both languages: simple past, the imperfect, the pluperfect and the past progressive.

Sandman, Erika

Ego evidential -yek as a stance marker in Wutun: evidence from conversational data

In my talk I will discuss the stance-taking functions of Wutun ego marker -yek in naturally occurring data. Wutun is a variety of Northwest Mandarin spoken by ca. 4000 people in Qinghai Province, Western China. Due to long term language contact, Wutun morphosyntax has been heavily influenced by Amdo Tibetan, and one of its most prominent Tibetan features is egophoric marking. Wutun egohoric marking system is intertwined with evidentiality, and it consists of ego marker, suffix -yek, and two non-ego markers, sensory-inferential suffix -li and factual auxiliary re. The ego marker -yek is associated with privileged access to the instigation of event in question, and it is often used when the speaker talks about his/her own actions. However, -yek is also used in third person statements when the speaker is in privileged position of making statements on others. For example, speakers often use -yek in performatives when they have influenced the physical or mental states of other people. It can be concluded that egophoric marking in Wutun is intimately associated with stance-taking; speakers employ it to express their knowledge states concerning self and others. The data for my study consists of video-recorded everyday conversations collected among the Wutun speech community in autumn 2018, and in addition to discussing egophoric marking in dialogues, I will also address the possible advantages, limitations and methodological problems of studying knowledge-related grammatical categories in conversational data.

Simon, Camille

Negociating the facts: interactional functions of factual evidential markers in Amdo-Tibetan

This presentation will focus on the functions of two copulas and one TAM marker and Amdo-Tibetan: ཡིན་ནི་རེད། /jənnəre/, factual equative copula, on the one hand, and ཡོད་གི་རེད། /joɣəre/ factual existential copula and V+གི་རེད། /ɣəre/, factual generic tense-aspect, on the other hand. They are respectively paradigmatically opposed to རེད། /re/, factual equative copula, ཡོད་ནི་རེད། /jonəre/ factual existential copula and V+ནི་རེད། /nəre/, factual generic tense-aspect. Morphophonologically, ཡོད་གི་རེད། /joɣəre/ and V+གི་རེད། /ɣəre/ are also charcterised by a specific intonation pattern.

The analysis their context of occurence reveals that they express factual evidentiality, but also their use, also depends on the speaker’s evaluation of the co-speaker’s knowledge, and serve the regulation of the interaction : the speaker may want to present herself as possessing a factual knowledge ignored by the co-speaker, or, on the contrary, insist on the fact that they both already share this knowledge. First, typical examples and contexts of use will be presented for each case, and opposed to the other members of their respective paradigm, in order to show that the interactional function is one of their primary function. In a second part, their co-occurrence with utterance-final discourse particles, related either to a strongly assertive or authoritative stance, or, on the contrary, to phatic functions, will be examined. Finally, some observations will be made on the dialectal variation observed of the occurrence of these forms in Amdo.

Tournadre, Nicolas

Evidential accessibility hierarchies

In rich evidential-epistemic systems such as those found in the Tibetic languages, the speaker often has various types of access to information related to the sensory perceptions of the event or to various inferences and even reported information.

In most cases, the speaker can choose between two or more sensory / inferential types of access or even other sources.

It seems that there is a default preference for the selection of markers which seems to follow general cognitive principles of accessibility. A somewhat similar notion of ‘evidential hierarchies’ has been proposed, independently, by other authors (see e.g Faller, 2002 and De Haan, 1998).

We propose the following accessibility hierarchy for the Tibetic languages:


Abbreviation : ENDO = endopathic, EGO = egophoric, VIS SENS = visual sensory, NVIS SENS= non visual sensory, SENS INF = sensory inferential, NSENS INF= non-sensory inferential, QUOT= quotative.

Widmer, Manuel

The evolution of epistemic categories in Bunan

The epistemic verbal categories “evidentiality” and “egophoricity” play an important role in the verbal systems of many Tibeto-Burman languages of the Himalayas. In the course of the past decades, our synchronic understanding of those grammatical categories has been considerably enhanced by numerous descriptive studies. However, little is still known about the diachronic processes that give rise to evidentiality and egophoricity. The article addresses this gap by discussing evidence from Bunan, a Tibeto-Burman language for whose past tense system the development of evidentiality and egophoricity can be reconstructed in detail.

In a first part, the development of evidential marking in Bunan will be reconstructed based on language-internal evidence. It will be argued that evidential marking emerged in the past tense domain when a former periphrastic perfect construction developed into a synthetic past tense with an inferential connotation (Widmer 2017b). This innovative inferential past tense then came to stand in opposition to an old past tense, which originally did not express any evidential / epistemic categories, but subsequently acquired a direct evidential construal in consequence of a generalized conversational implicature (cf. Atlas & Levinson 1981). In a second part, the further development of this dichotomic evidential system will be investigated. It will be shown that Bunan reanalyzed a former person distinction as an egophoric opposition (Widmer 2015, 2017a; Widmer & Zemp 2017). This diachronic process led to a fundamental restructuring of the past tense domain, which eventually gave rise to the modern Bunan verbal system, which distinguishes between an egophoric past tense, a direct evidential past tense, and an inferential past tense.

The talk will thus offer new insights into the hitherto poorly understood diachronic mechanisms that have given rise to complex epistemic systems in Tibeto-Burman languages of the Greater Himalayan region.


Aikhenvald, Alexandra. 2004. Evidentiality. Oxford / New York: Oxford University Press.

Aikhenvald, Alexandra & Robert M. W. Dixon. 1998. Evidentials and areal typology: A case study from Amazonia. Language Sciences 20(3). 241–257.

Atlas, Jay D. & Stephen C. Levinson. 1981. It-clefts, informativeness, and logical form: Radical pragmatics (revised standard version). In Peter Cole (ed.), Radical Pragmatics, 1–61. New York: Academic Press.

Francke, August H. 1909. Tabellen der Pronomina und Verba in den drei Sprachen Lahoul’s: Bunan, Manchad und Tinan. Zeitschrift der Deutschen Morgen­ländi­schen Gesellschaft 63. 65–97.

San Roque, Lila & Robyn Loughnane. 2012. The New Guinea Highlands evidentiality area. Linguistic Typology 16. 111–167.

Widmer, Manuel & Marius Zemp. 2017. The epistemization of person markers in reported speech. Studies in Languages 41(4), 33–75.

Widmer, Manuel. 2015. The transformation of verb agreement into epistemic marking: evidence from Tibeto-Burman. In Jürg Fleischer, Elisabeth Rieken & Paul Widmer (eds.), Agreement from a diachronic perspective (Trends in Linguistics. Studies and Monographs 287), 53–73. Berlin: Mouton de Gruyter.

Widmer, Manuel. 2017a. A grammar of Bunan (Mouton Grammar Library 71). Berlin: de Gruyter.

Widmer, Manuel. 2017b. The evolution of egophoricity and evidentiality in the Himalayas: the case of Bunan. The rise and development of evidential and epistemic markers, special issue of Journal of Historical Linguistics 7(1-2), 246–275.

Zeisler, Bettina

Speaker attitude (Stance) and other 'freaks' in the Ladakhi 'evidential' system (Vive la parole!)

The so-called evidential markers and the (more) epistemic markers used in Ladakhi have a prototypical usage, as known from the standard descriptions of Tibetic ‘evidential’ systems, but also various non-standard, marked usages. The Ladakhi ‘evidential’ markers thus do not so much indicate access to knowledge than a speaker’s attitude towards the situation and towards the addressee and, in questions, the speaker’s expectations about the addressee’s attitude.

Some of the non-standard usages are rather unsystematic or ‘freaky’, sometimes even contradictory. As they obviously belong to la parole, they cannot be explained on the level of la langue. Some of them and even be explained by assumptions about socio-pragmatic factors. I shall introduce a selection of ‘evidential freaks’, which I encountered in Ladakhi, and I shall also discuss also a few more general problems in the description of the markers.

Among other things, I shall talk about the contradictory properties of the linking verbs yin and yod, their asymmetry, and the unexpected use of yin for yod. I shall further present examples that violate the accessibility hierarchy. I shall also discuss the question whether the unmarked past tense form should be taken as a neutral category, and why rather not.

Zemp, Marius

The genesis of evidentiality in Tibetan

Purik, a geographically peripheral and phonologically archaic Tibetan dialect spoken in the western part of Ladakh (Jammu and Kashmir, India), gives us valuable hints as to how Tibetan came to grammatically encode evidentiality, that is, how the verb came to regularly indicate how one obtained the information conveyed in a sentence. In Purik, evidentiality is only encoded by two contrasting existential copulas, direct sensory evidential duk (WT ’dug) and factual jot (WT yod), and two related pairs of constructions, resultative V-suk (< *V-s-’dug) and V-set (< *V-s-yod) and prospective V-(t/n)uk (< *V-(’d)ug) and V-et (< *V-yod). Of these constructions, only resultative *V-s-’dug, which used to describe the directly witnessed present result of an event, is reflected in all dialects of Tibetan, and I will show that the evidential systems of all dialects indeed derive from that construction. At some point, *V-s-’dug came to infer a past event from its directly witnessed result, and in many Eastern and Central Tibetan dialects, the grammaticalization of this inferential *V-s(-’d)ug ensued the grammaticalization of a contrasting *V-thal ‘went past V-ing’ or *V-song ‘went V-ing’ as indicating a directly witnessed past event. In many Western and Central Tibetan dialects, on the other hand, perfect *V-s-’dug became contrasted by *V-s-yod describing a present state one knows from having closely monitored the past event that led to it