Evidentiality

Since the category of information source is not grammatical in classical Indo-European languages , the concept of evidentiality in linguistics was formed only at the beginning of the 20th century when describing such languages ​​of South America as Quechua and Aymara .

One of the first scholars to talk about the concept of evidentiality as an indispensable indicator of the source of information was the American ethnolinguist Franz Boas . In his 1911 work , Boas discusses the statement Man is sick in Quakiutl :

If the speaker himself did not see the sick person, he must indicate whether he received information about the person’s disease from second-hand hands or whether he dreamed about it. ^[four]

Original text

In case the speaker had not seen the sick person himself, he would have to express whether he knows by hearsay or by evidence that the person is sick, or whether he has dreamed it.

The concept of evidentiality as an obligatory grammatical category was first introduced by Roman Jacobson , who considered it as a sphere of meanings pointing to the source of information ^[5] . Jacobson was also the first to single out mood and evidentiality as two independent categories ^[6] .

Depending on the way in which the speaker found out about the reported situation, different evidential values ​​can be distinguished. First of all, the speaker could have direct or indirect access to information.

Direct evidence implies a direct perception by the speaker of the situation. Often there are visual (the speaker visually observed the situation), auditory (the auditory perception of the situation by the speaker) and other sensory information sources. When the speaker himself saw the situation, they often talk about visual evidence ^[7] . It is worth noting that in different languages ​​the border between direct and indirect witnessing is drawn differently: for example, in the screw language direct witnessing is based only on vision data, and all other subtypes of sensory perception are marked as indirect evidentiality ^[8] .

Indirect evidence suggests that the speaker did not perceive the situation directly, and this information was obtained in some other way:

• Inference ( English inference , from English infer “to make an assumption, make a guess”) - the speaker was not a witness to the situation, but some indirect signs allowed him to draw a logical conclusion . In some languages, the value of inference can be conveyed differentially - as a conclusion a) on the basis of direct material (visible) evidence; b) based on intuition or general knowledge, mental activity.
• Retelling ( English reported evidence, hearsay ) - the speaker was not a witness to the situation and knows about it from other people's words. As in the case of inference, several types of retelling can be distinguished in some languages ​​depending on whether the message was received from a person who witnessed the event (second-hand retelling, English secondhand ) or who was not a witness (third-party retelling, English thirdhand ).

Typological data on various semantic subtypes of evidentiality and their territorial distribution were summarized in the framework of the project of the World Atlas of Language Structures and presented on its website ^[9] .

In a number of languages, the combination of the category of evidentiality with the category (hell) of peacefulness is also widespread.

There are different types of evidential systems. The language can be expressed as a binary opposition - (visual) evidence / eyesiness ( English eyewitness / noneyewitness ) - and a system of six or more elements ^[10] .

In general, one of the leading experts on the typology of evidentiality, Alexander Eichenwald, identifies two types of evidential systems: systems in which the presence of the source of information is expressed, but the type of source is not specified ( type I ); and systems in which the types of information sources are different ( type II ) ^[11] . The Turkologist Lars Johanson , followed by Eichenwald, calls the Type I system a system of indirect evidence ^[12] , while Type II is called evidentiality itself.

Indirect Certification (Type I)

In a number of languages, only indirect evidence is grammatically labeled, that is, the fact that the speaker was not a witness to the described situation, but the type of information source (the speaker’s logical conclusion based on some facts, information from other people's words, etc.) not specified. Indirect evidence marking is typical of Turkic languages , it is also found in Iranian , Finno-Ugric and many other languages ^[13] . Moreover, the indirect information gramme often has an additional meaning, implying that the speaker does not take responsibility for the truth of the transmitted information ^[14] .

So, in the next sentence from the Turkish language, the indicator -mış implies that the speaker refers to an indirect source of information (this can be either a logical conclusion of the speaker himself or his sensation or information from other people's words):

Ali	bu-nu	bil-iyor- muş
Ali	it is ACC	to know - INTRA-IC
Ali	obviously	knows that.

(Aikhenvald 2004: 275)

Evidence itself (type II)

Two-element systems

• A1. Eyewitness vs. blindness (noneyewitness)

In the following example, the actions that the speaker saw with his own eyes are contrasted with actions that the speaker did not see from the Jaravar language (the Aravanian language family):

Wero	kisa-me- no ,	ka-me- hiri -ka
Vero	get down- BACK - IMM.P.NONFIRSTH .m	move- BACK REC.P.FIRSTH .m
Vero	tears from a hammock (I have not seen this)	and went out (I saw it).

(Aikhenvald, Dixon 2003: 24)

A similar system of evidentiality is also found in the Yukagir languages , Godoberin language and some others.

• A2. Indirect evidence (nonfirsthand) vs. everything else (indirect evidence implies conclusions and conclusions based on the information heard or some visible signs and evidence). These include the Abkhaz language , some Nakh-Dagestan and Ugric languages , etc.
• A3. Retelling (testimony from other people's words, hearsay) vs. everything else . The most common evidential system; It is found in Papuan , Tibetan-Burmese languages , Lezgi , as well as in numerous languages ​​of South America .

Three-element systems

• B1. Visual evidences (direct evidence), inference , retelling . Similar evidential systems were found in the languages ​​of Aymara , Shastan languages , languages ​​of Embere , Quechua , etc.
• B2. Visual , non-visual sensory , perceived knowledge (inference) ( washo language ).
• B3 Non-visual sensory evidence , inferences and retelling (retuoran language from the Tukan language family, northern Pomo from the Pomo language family).

Four-element systems

• C1. Visual evidence , non-visual sensory evidence , inference , retelling ( Tarian language, some Tukan languages , oriental help).

An example from the Tariana language ( Arawak languages ): In this language, the phrase “Cecilia scolded a dog” has four different variants depending on the type of evidentiality. If the speaker saw how this happened, the phrase would look like this:

Ceci	tʃinu-nuku	du-kwisa- ka
Cecilia	dog- TOP.NON.A / S	3SGF- scold- REC.P.VIS
Cecilia	scolded the dog	(I saw it).

If the speaker only heard Cecilia scolding the dog, the phrase would look different:

Ceci	tʃinu-nuku	du-kwisa- mahka
Cecilia	dog- TOP.NON.A / S	3SGF- scold- REC.P.NONVIS
Cecilia	scolded the dog	(I heard that).

If the speaker saw a frightened dog, his knowledge that Cecilia scolded the dog would be inferred:

Ceci	tʃinu-nuku	du-kwisa- sika
Cecilia	dog- TOP.NON.A / S	3SGF- scold- REC.P.INFR
Cecilia	scolded the dog	(I assumed it).

If the speaker knows that Cecilia scolded the dog from someone else, then he must say:

Ceci	tʃinu-nuku	du-kwisa- pidaka
Cecilia	dog- TOP.NON.A / S	3SGF- scold- REC.P.REP
Cecilia	scolded the dog	(I found out from someone).

(Aikhenvald, Dixon 2003: 134-135)

• C2. Visual evidence , inference-1 , inference-2 , retelling . A similar system is found in the language of Pawnee and Tsafiki . Visual information (directly witnessed events) is unmarked, there is a suffix for expressing direct material evidence of the reported information (inference-1) and for assumptions based on general knowledge (inference-2), as well as for information received from other persons (retelling).

• C3 Non-visual sensory evidence , inference-1 , inference-2 , retelling ( tongue screw ). In addition to suffixes expressing non-visual sensory evidentiality and retelling, the screw also distinguishes knowledge that is inferred logically (inferential) and supposed knowledge, about which the speaker thinks that it is true, based on his experience with similar situations in the past.

• C4. Visual evidence , inference , retelling-1 , retelling-2 (southern tepeua, Uto-Aztec language family ). Retelling is divided into two types, depending on whether the listener knew the information before or not.

Different Evidentive Subsystems in One Language

In one language there can be several evidential subsystems. The choice of the necessary subsystem may depend on the type of sentence, the time in which the verb is located, or on the mood of the verb. So, in the language of the tarian there are four types of evidentiality in affirmative sentences (C1), but only three in interrogative ones (no retelling, scheme B2), in sentences expressing an order, on the contrary, only one kind of evidentiality is distinguished - retelling (scheme A3), and in the relative clauses of the goal, visual and non-visual information is distinguished (diagram A1).

In addition, combinations of different types of evidentiality are possible in some languages. In Qiang language, visual evidentiality can be combined with logically inferred knowledge (inference), for example, in the following situation:

oh	the:	ʐbə	ʐete- k - u !
about	3 SG	drum	beat- INFR-VIS
ABOUT,	he played,	on the drum!

(the speaker makes the assumption that someone was playing the drum outside the door, opens the door and sees a man with a drum in his hands )

(Aikhenvald & Dixon 2003: 70)

In tsafiki, retelling can be combined with any of the three other types of evidentiality of a given language, including with noneyewitness, as in the following example:

Manuel	ano	fi- nu - ti -e
Manuel	food	yes- INFR.PHYSICAL.EVIDENCE-HEARSAY-DECL
(He said	/ they	said that) Manuel ate (they did not see him, but they have evidence).

(Aikhenvald & Dixon 2003: 9)

In languages ​​with grammatical evidentiality, morphological and lexical-syntactic ones can be distinguished among the means of expressing the latter ^[15] .

Morphological tools

In many languages, specific morphemes exist that point to the source of information given to the speaker. These means of expressing evidentiality are common in the languages ​​of the American Indians, in Sino-Tibetan , African languages.

N. A. Kozintseva also relates to this category the forms of retelling mood in Bulgarian and Albanian , absentee in the Uralic languages , and eye-borne mood in Tajik and Archin languages ​​(see also examples from section 2).

A more detailed description of the morphological means of expressing evidentiality and their territorial distribution is given on the WALS project website ^[16] .

Lexico-syntactic tools

In some languages, the subjunctive may be added to the main sentence with the help of different conjunctions indicating different types of evidentiality. In Rwanda, the speaker must choose between the three unions - ko , if he is neutral about the truth of the information expressed; ngo , if he has direct evidence that makes him doubt the truth; kongo , if he has an indirect evidence of possible false information:

ya-vuze	ngo	a-zaa-za
he- past — talk	what	he- FUT- come
He said that	will come	(but the speaker has direct testimony that casts doubt on his coming).

ya-vuze	kongo	a-zaa-za
he- past — talk	what	he- FUT- come
He said that	will come	(but the speaker has indirect evidence that casts doubt on his coming).

(Kozintseva 1994: 96)

In many languages ​​(including the Russian language ), there are lexical tools for expressing evidentiality: complex sentences with a modus verb (Russian. They say that ... ); modal words ( ekan , emish in Uzbek ); introductory phrases with a modus verb (Russian. As I know ...; They say ... ) and some others.

• Evidence in Tuyuka
• Desidence in Desano

• Kozintseva, N. A. Category of evidentiality (problems of typological analysis) // Questions of linguistics. - M. , 1994. - No. 3 . - P. 93-104.
• Plungyan, V. A. Introduction to grammatical semantics: grammatical meanings and grammatical systems of world languages. - M .: RSUH, 2011.
• Plungyan, V. A. General morphology: Introduction to the problems. - M .: URSS editorial, 2003.
• Khrakovsky, V. S. Evidence in the languages ​​of Europe and Asia. - SPb. : Science, 2007.
• Aikhenvald, Alexandra Y. Evidentiality. - Oxford: Oxford University Press, 2004 .-- ISBN 0-19-926388-4 .
• Aikhenvald, Alexandra Y .; & Dixon, RMW (Eds.). Studies in evidentiality // Typological Studies in Language. - Amsterdam: John Benjamins Publishing Company, 2003. - No. 54 . - ISBN 90-272-2962-7 ; ISBN 1-58811-344-2 .
• Boas, Franz. Handbook of American Indian languages. - Washington: GPO, 1911. - P. 43.
• Chafe, Wallace L .; Nichols, Johanna (Eds.). Evidentiality: The linguistic encoding of epistemology. - Norwood, NJ: Ablex, 1986.
• de Haan, Ferdinand. Coding of Evidentiality = The World Atlas of Language Structures Online // Max Planck Digital Library. - Munich, 2011.
• de Haan, Ferdinand. Semantic Distinctions of Evidentiality = The World Atlas of Language Structures Online // Max Planck Digital Library. - Munich, 2011.
• Guentchéva, Zlatka (Ed.). L'Énonciation médiatisée. Bibliothèque de l'information grammaticale. - Louvain: Éditions Peeters, 1996.
• Jakobson, R. Shifters, Verbal Categories, and the Russian Verb. - Cambridge: Harvard University, 1957.
• Johanson, Lars; Utas, Bo (Eds.). Evidentials: Turkic, Iranian and neighboring languages. - Berlin: Mouton de Gruyter, 2000.