17.5: Two types of evidentials

Last updated
Save as PDF

Page ID: 138722

$ \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } $ $ \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} $$\newcommand{\id}{\mathrm{id}}$ $ \newcommand{\Span}{\mathrm{span}}$ $ \newcommand{\kernel}{\mathrm{null}\,}$ $ \newcommand{\range}{\mathrm{range}\,}$ $ \newcommand{\RealPart}{\mathrm{Re}}$ $ \newcommand{\ImaginaryPart}{\mathrm{Im}}$ $ \newcommand{\Argument}{\mathrm{Arg}}$ $ \newcommand{\norm}[1]{\| #1 \|}$ $ \newcommand{\inner}[2]{\langle #1, #2 \rangle}$ $ \newcommand{\Span}{\mathrm{span}}$ $\newcommand{\id}{\mathrm{id}}$ $ \newcommand{\Span}{\mathrm{span}}$ $ \newcommand{\kernel}{\mathrm{null}\,}$ $ \newcommand{\range}{\mathrm{range}\,}$ $ \newcommand{\RealPart}{\mathrm{Re}}$ $ \newcommand{\ImaginaryPart}{\mathrm{Im}}$ $ \newcommand{\Argument}{\mathrm{Arg}}$ $ \newcommand{\norm}[1]{\| #1 \|}$ $ \newcommand{\inner}[2]{\langle #1, #2 \rangle}$ $ \newcommand{\Span}{\mathrm{span}}$$\newcommand{\AA}{\unicode[.8,0]{x212B}}$

In §17.3 we mentioned that evidential markers in some languages do not contribute to propositional content but function as illocutionary modifiers. One of the best documented examples of this type is Cuzco Quechua as described by Martina Faller.¹⁰ Faller analyzes the evidential enclitics in Cuzco Quechua as “illocutionary modifiers which add to or modify the sincerity conditions of the act they apply to.” She notes that “they do not contribute to the main proposition expressed, can never occur in the scope of propositional operators such as negation, and can only occur in illocutionary force bearing environments.”¹¹

We present here some of her evidence for saying that the evidential enclitics do not contribute to the propositional content of the utterance, focusing on the Reportative clitic =si. First, the evidential is always interpreted as being outside the scope of negation. In example (6), the contribution of the Reportative evidential (‘speaker was told that p’) cannot be interpreted as part of what is being negated; so (ii) is not a possible interpretation for this sentence.

(6) Ines-qa mana=s qaynunchaw ñaña-n-ta-chu watuku-rqa-n.

Inés-top not=report yesterday sister-3-acc-neg visit-past₁-3

propositional content = ‘Inés didn’t visit her sister yesterday.’

evidential meaning: (i) speaker was told that Inés did not visit her sister yesterday

not: (ii) # speaker was not told that Inés visited her sister yesterday
[Faller 2002, §6.3.1]

Second, the contribution of the Reportative evidential is not part of what can be challenged. If a speaker makes the statement in (7a), a hearer might challenge the truth of the statement based on the facts being reported, as in (7b); but it would be infelicitous to challenge the truth of the statement based on source of information, as in (7c). (This test is sometimes called the assent/dissent diagnostic.¹²) In other words, the contribution of the evidential does not seem to be part of what makes the statement true or false.

(7) a. Ines-qa qaynunchay ñaña-n-ta=s watuku-sqa.

Inés-top yesterday sister-acc=report visit-past₂

propositional content = ‘Inés visited her sister yesterday.’

evidential meaning: speaker was told that Inés visited her sister yesterday

b. Mana=n chiqaq-chu. Manta-n-ta-lla=n watuku-rqa-n.

not=direct true-neg mother-3-acc-limit=direct visit-past₁-3

‘That’s not true. She only visited her mother.’

c. Mana=n chiqaq-chu. #Mana=n chay-ta willa-rqa-sunki-chu.

not=direct true-neg not=direct this-acc tell-past₁-3S.2O-neg

‘That’s not true. #You were not told this.’ [Faller 2002, §5.3.3]

Third, Faller’s statement that the evidential enclitics “can only occur in illocutionary force bearing environments” means that they are restricted to main clauses or clauses which express an independent speech act. This is a characteristic feature of many illocutionary modifiers. In particular, conditional clauses are typically not the kind of environment where illocutionary modifiers can occur.¹³ Faller states that evidential enclitics cannot occur within conditional clauses, as illustrated in (8).

(8) Mana(*=si) para-sha-n-chu chayqa ri-sun-chis.

not=report rain-prog-3-neg then go-1.fut-pl

‘If it is not raining we will go.’ [Faller 2003, ex. 8]

The German auxiliary sollen ‘should’, when used as a reportative or hearsay marker, behaves quite differently. For example, it is possible for sollen to occur within a conditional clause, as illustrated in (9).

(9) F.C.B.F.A.N.: Bei uns soll es heute schneien‼

FAHRBACH: Also wenn es bei dir schneien soll, dann schneit es bei mir auch.

‘If it said to snow near you, then it will snow near me as well.’¹⁴

The assent/dissent diagnostic reveals another difference. German Reportative sollen, like the Quechua Reportative evidential, allows the hearer to challenge the basic propositional content of the sentence. But in addition, it is possible to challenge the truth of a statement with sollen based on the source of information, as illustrated in (10).¹⁵ This is impossible with the Quechua Reportative. Both of these differences are consistent with the hypothesis that German Reportative sollen is part of the propositional content of the utterance.

(10) A: Laut Polizei soll die Gärtnerin die Juwelen gestohlen haben.

‘According to the police, the gardener is said to have stolen the jewels.’

B: Nein, das stimmt nicht. Das ist die Presse, die das behauptet.

‘No, that’s not true. It is the press who is claiming this.’ (Faller 2006)

A number of languages have evidentials which behave much like those of Cuzco Quechua. However, there are other languages in which evidentials seem to contribute to the propositional content of the utterance, like German Reportative sollen. Murray (2010) suggests that we need to recognize two different types of evidential, which we will refer to as illocutionary evidentials and propositional evidentials.¹⁶ Illocutionary evidentials function as illocutionary operators; examples are found in Quechua, Kalaallisut, and Cheyenne. Propositional evidentials are part of the propositional content of the utterance; examples are found in German, Turkish, Bulgarian, St’át’imcets (Lillooet Salish), and Japanese.

These two types of evidentials share a number of properties in common, but Murray identifies several tests that distinguish the two classes. For example, illocutionary evidentials cannot be embedded within a conditional clause (8), while this is possible for propositional evidentials (9). Second, a speaker who makes a statement using a hearsay or reportative evidential of the illocutionary type is not committed to believing that the propositional content of the utterance is possibly true. So it is not a contradiction, nor is it infelicitous, for a speaker to assert something as hearsay and then deny that he believes it, as illustrated in (11).

(11) a. Para-sha-n=si, ichaqa mana crei-ni-chu.

rain-prog-3=report but not believe-1-neg

‘It is raining (someone says), but I don’t believe it.’
[Cuzco Quechua; Faller (2002: 194)]

b. É-hoo’kȯhó-nėse naa oha ná-sáa-oné’séomátséstó-he-⌀.

3-rain-report.inan.sg and contr 1-neg-believe_INAN-mod_ANIM-dir

‘It’s raining, they say, but I don’t believe it.’
[Cheyenne; Murray 2010: 58]

A hearsay or reportative evidential of the propositional type, however, commits the speaker to believing that it is at least possible for the expressed proposition to be true. For this reason, the St’át’imcets example in (12) is infelicitous.

(12) (Context: You had done some work for a company and they said they put your pay, $200, in your bank account; but actually, they didn’t pay you at all.)

*Um’-en-tsal-itás ku7 i án’was-a xetspqíqen’kst táola, t’u7 aoz kw s-7um’-en-tsál-itas ku stam’.

give-dir-1sg.obj-3pl.erg report det.pl two-det hundred dollar but neg det nom-give-dir-1sg.obj-3pl.erg det what

‘They gave me $200 [I was told], but they didn’t give me anything.’
[Matthewson et al. 2007]

Third, illocutionary evidentials are always speaker-oriented. This means that they indicate the source of information of the speaker, and cannot be used to indicate the source of information of some other participant. This is illustrated in the Quechua example in (13).

(13) Pilar-qa yacha-sha-n Marya-q

Pilar-top know-prog-3 Marya-gen

hamu-sqa-n-ta{-n/-s/-chá}.

come-past.ptcp-3-acc{-dir/-report/-conject}

propositional content = ‘Pilar knows that Marya came.’

evidential meaning: (i) speaker has direct/reportative/conjectural evidence that Pilar knows that Marya came.

not: (ii) #Pilar has direct/reportative/conjectural evidence that Marya came. [Faller 2002, ex. 184]

Propositional evidentials, in contrast, can be used to indicate the source of information of some participant other than the speaker. In the St’át’imcets example in (14), for example, the reportative evidential is interpreted as marking Lémya7’s source of information. It indicates that Lémya7’s statement was based on hearsay evidence. The speaker in (14) already had direct evidence for this information before hearing it from Lémya7.

(14) Tsut s-Lémya7 kw sqwemémn’ek ku7 s-Mary, t’u7 plán-lhkan ti7 zwát-en — áts’x-en-lhkan s-Mary áta7 tecwp-álhcw-a inátcwas.

say nom-name det pregnant report nom-name but already-1sg.subj dem know-dir — see-dir-1sg.subj nom-name deic buy-place-det yesterday

‘Lémya7 said that [she was told that] Mary is pregnant, but I already knew that — I had seen Mary at the store yesterday.’¹⁷

A fourth difference demonstrated by Murray is that markers of tense or modality never take semantic scope over illocutionary evidential markers, whereas this is possible with propositional evidentials.¹⁸

There seems to be a strong tendency for illocutionary evidential markers to be “true evidentials” in Aikhenvald’s sense, i.e., grammatical morphemes whose primary function is to mark source of information; and for propositional evidentials to be evidential uses/senses of morphemes whose primary function is something else: perfect aspect in Turkish and Bulgarian; modality in German and St’át’imcets. In terms of the distinction we made in Chapter 11, illocutionary evidentials seem to contribute use-conditional meaning, while propositional evidentials seem to contribute truth-conditional meaning.

¹⁰ Faller (2002; 2003; 2006), inter alia.

¹¹ Faller (2002: v).

¹² Papafragou (2006).

¹³ Ernst (2009), Haegeman (2010).

¹⁴ www.kc-forum.com/archive/index.php/t-45696, cited in Faller (2006). A reviewer points out that this is not a typical hypothetical conditional, but what we will call in Chapter 19 a reality conditional, which has somewhat specialized functions. It is not clear whether the reportative use of sollen is possible within hypothetical conditionals.

¹⁵ Faller (2006).

¹⁶ Murray uses the terms illocutionary evidentials vs. epistemic evidentials.

¹⁷ Matthewson et al. (2007).

¹⁸ See Murray (2010: §3.4.2) for examples.