LINGUIST List 12.2155

Tue Sep 4 2001

Review: Floyd, Evidentials in Wanka Quechua

Editor for this issue: Terence Langendoen <>

What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Simin Karimi at or Terry Langendoen at


  1. Bernd Ruppel, Review of Floyd, Evidential categories in Wanka Quechua

Message 1: Review of Floyd, Evidential categories in Wanka Quechua

Date: Tue, 04 Sep 2001 17:37:39 +0200
From: Bernd Ruppel <>
Subject: Review of Floyd, Evidential categories in Wanka Quechua

Floyd, Rick (1999) The structure of Evidential Categories in
Wanka Quechua. Summer Institute of Linguistics
(Publications in Linguistics 131), 206pp, ISBN 1-55671-066-6,

Review by: Bernd Ruppel, Department of Linguistics,
University of Erfurt


The book by Rick Floyd presents an investigation of the
evidential markers of Wanka Quechua (or Huanca Quechua), a
language spoken in the central Andean highlands of Peru.
The intended audience are students and researchers who are
interested in evidentiality or epistemic modality or people
studying Quechuan languages. The book can be divided into
three parts

1) Theoretical and methodological prerequisites (pp. 1-55;
chs. 1-3)
2) Semantic analysis of the three evidential markers '-
-m(i)', 'chr(a)', '-sh(i)' (the vowel of each morpheme is
dropped when it is followed by an open syllable) (pp. 57-
159; chs. 4-6)
3) An application of concepts developed within the framework
of cognitive grammar to the analysis of the Wanka
evidential markers (pp. 161-194; ch. 7).

Some parts of the book (some parts of chs. 4-6) seem to
have been published separately earlier according to
footnotes on pp. 57, 93, 123.


Chapter 1
A very brief overview of the discovery and research on
evidential categories is given followed by an outline of
the present work, and a list of texts that are used for
analysis in this study. Floyd (henceforth F.) decided to
use mainly conversations assuming that this kind of texts
represents instances of "prototypical language use" (p. 6).
The end of the chapter consists in a brief introduction
into the Wanka language. Important for the present study is
an information given only at the end of the book:
evidential marking is not obligatory in Wanka Quechua,
around 31% of the utterances in F.'s sample don't show
evidential marking (p. 189).

Chapter 2
In agreement with Anderson (1986) F. defined the core
notion of evidentiality as "the encoding of the information
source on which a speaker bases a proposition" (p. 13).
Following the divisions made by Willett (1988) different
kinds of evidential marking are explained. F. draws a
semantic map of evidential meanings with three main parts:
a) source of information is the speaker's direct
b) source of information is the speaker's inference,
c) source of information is based on reports.
Evidential markers may have validational meanings too
(expression of the speaker's commitment to the truth of a
proposition). Ordered on a scale from high commitment to
low commitment F. argues that inference implies lower
commitment than direct or reportative evidence, which
doesn't say anything about the speaker's commitment in
Wanka Quechua. At the end of this chapter F. gives an
outline of the syntactic behaviour of Wanka evidential
markers. They may be attached to almost all lexical classes
and tend to appear preverbally after the first major
constituent. As for the etymology of this markers, it is
suggested that they are borrowed from the neighbouring
language Aymara.

Chapter 3
F. is using notions and models of cognitive grammar in his
semantic analysis of the Wanka evidentials. Fundamental
notions such as 'base', 'profile', 'domain' and 'construal'
are introduced. He aims to describe the meaning of an
evidential marker as a semantic network. A semantic network
is composed of several uses of a lexical item which are
interrelated through partial meaning resemblance. Within
the network one meaning is termed prototypical. The
prototypical meaning of a lexical item is one which is most
"salient", i.e. it comes up most likely in neutral
contexts. A feature which is present in all uses of a
lexical item is termed the schematic meaning of this item.

Chapter 4
F. deals with the Wanka evidential marker '-m(i)'. Its
prototypical use according to F. is the coding of direct
experience of an event. This comprises visual, auditory and
other kinds of perception. Even events occurring in dreams
are construed as perceived directly. Since direct evidence
points to great reliability, '-m(i)' is also used, when the
speaker wants to emphasize the truth of the proposition he
is making. That means, schematically 'm(i)' is denoting the
speaker's certainty. F. gives some examples where speakers
are telling about an event which they were not able to
perceive directly, but still are using the direct
evidential marker, in order to exemplify this usage.
A second related non-prototypical use is the marking of
firm predictions about the future (mainly first person) and
the use of this morpheme for making strong suggestions to
the addressee. Compare F.'s example from p. 84:
"shramu-nki-m" 'come-2-DIR' (DIR = direct evidential)
"You will come (directive)".
Finally '-m(i)' is used in direct questions. F. explains
this usage as a shift of the observer position from the
speaker to the addressee. Whereas in assertions the speaker
is prototypically the one who perceives the event, in
questions the addressee is construed as potential
experiencer. The marker of yes/no questions 'chun' is
analysed as 'chu' + 'm' (NEG + 'm(i)').

Chapter 5
This chapter is devoted to the study of the marker
'-chr(a)'. Its prototypical meaning is the coding of
inferences (the speaker infers an event from his
observation of another event) and guesses (the speaker
simply speculates). This use is the one which occurs most
often with '-chr(a)' in F.'s database (67% out of 201
instances, [p. 95, n. 32]). This is the prototypical use.
F. lists five non-prototypical uses:
a) the marking of mild exhortations (this meaning
occurs, when the event in consideration lies in the
future and the actor is a second person)
b) the marking of acquiescence (the speaker concedes
something, he isn't able to avoid),
c) use marking of questions. According to F. the
occurrence of '-chr(a)' in questions implies that the
speaker doesn't think that the addressee could give
an exact answer. Or it is used in questions that
aren't expected to be answered anyway, e. g. (p.113)
"may-chruu-chra gasta-y-pa paawa-alu-n"
where-LOC-CONJ spend-NOM-GEN finish-ASP-3
"I wonder where he spent it all."
(NOM = Nominalizer, LOC = Locative, CONJ =
Conjecture, GEN = Genitive, ASP = aspect),
d) use in ironical remarks, where the speaker suggests
something he doesn't believe
e) use in "first person inferences" (p. 118), a usage
which resembles very the ironical meaning, e.g.
ya'a trabaju-u-ta-chra-a upya-ka-mu-u
I work-1P-ACC-CONJ-TOP drink-REF-AFAR-1
(REF = reflexive; AFAR = translocative (at a
Lit: (I suppose) I drink my work (i.e.,what I earn).
This sentence is interpreted by F. as "I drink, and
what's it to ya?".

Chapter 6
This chapter examines the reportative suffix '-sh(i)'.
F. lists four contexts, where this suffix is used: a) to
mark hearsay, b) in folktales, c) in riddles, d) in
challenges. The prototypical use of this suffix is the
marking of hearsay. Its occurrence in folktales
therefore is not a surprise. Rather puzzling, however,
are the uses c) and d). By the latter term F. designates
a formulaic expression by which the speaker asks the
addressee to join him/her in some action. This is
exemplified by the following sentence spoken by someone
who invites the addressee to join a vigil contest (p.
148, E 188):
maa mayan-ninchik-shi waala-shrun
PART which-12P-REP dawn-12FUT
Let's see which of us lasts till morning.
(PART = Particle, 12P = first inclusive possessive, REP
= Reportative, 12 = first person inclusive)
In riddles (use c) the reportative evidential occurs in
the opening line (p. 142, E 182),
cp. ima-lla-sh ayka-lla-sh
what-LIM-REP how_much-LIM-REP
What is it? How much?
(LIM = limitative (just))
F. tries to explain the use in riddles by the similarity
between folktales and riddles, for both belong to the
oral tradition of this community. Furthermore F. thinks
that '-sh(i)' bears some mirative meaning (expressing
surprise about some event), since the outcome of a
riddle is totally unexpected. This sense is according to
him the link to the use of '-sh(i)' in challenges (see
above). In challenges, too, the outcome is unexpected
and thus the use of '-sh(i)' could be explained.
In the last section of this chapter F. considers the
question, whether '-sh(i)' bears some validational
meaning as well. He draws the conclusion, that there may
be contexts in which such a meaning is implied but he
sees this purely as a by-product of the evidential

Chapter 7
In this chapter F. applies some notions of cognitive
grammar to Wanka Quechua, e.g. the notions of directness
and proximity. By proximity the relative distance
between two construed entities is meant, e.g. between
now and some past event. Using the notion of directness
F. explained the crucial difference between '-m(i)',
which implies direct experience and the other two
evidential markers. The validational meanings of '-m(i)'
and '-chr(a)' are discussed in terms of proximity. The
direct evidential marks a proposition which is
incorporated into the speaker's reality, whereas
'-chr(a)' marks a proposition as being outside the
speaker's reality. Reality is here understood as the
domain the conceptualizer has access to.

1. Methodology
F. tries to comprehend the meaning of each of the Wanka
evidentials in terms of a semantic network. This
approach is useful and allows for a general description
of the meaning of an grammatical item. There are,
however, some limits to this approach. Especially the
implicit claim that each synchronic use of a certain
morpheme is related to another by some semantic
resemblance, seems to be unwarranted. Imagine some
diachronic change such that a meaning 'B' develops out
of a meaning 'A', and 'C' out of 'B'. There may be no
connection between 'A' and 'C', but as long as 'B' is
present, one can readily establish the resemblance
between 'A', 'B' and 'C'. But if 'B' disappears, it can
be impossible to see, how 'A' and 'C' are related to
each other. This situation can be exemplified by the
French negative marker 'pas':
'pas' = step 'pas' as second 'pas' as negation
part of negation e.g., "Pas si vite!"
The connection between 'A' and 'C' can only be
constituted, if one knows 'B'. If 'B' disappears, there
is no resemblance between 'A' and 'C'.
This means, that one should allow for the possibility
that synchronically may exist some uses of the Wanka
evidentials which aren't related to other current uses.

2. Analysis of '-sh(i)'
This morpheme has two uses that are clearly correlated
with each other: marking of hearsay and folktales. In
both situations the speaker says something, he has heard
from other people. Riddles, too, can be said to be part
of the oral tradition of a community and therefore are
marked by '-sh(i)'. F. assumes that in riddles the
hearsay evidential bears some mirative meaning(for this
category, see DeLancey 1997), because something new and
unexpected is revealed. According to F. this is also the
clue for the understanding of the use of this marker in
challenges, too.

I think this analysis is debatable. If there were a
mirative meaning in 'sh(i)', it should be present in
declarative sentences, too. F. doesn't explain why the
supposed mirative meaning is present only in the opening
question of riddles and in challenges (cf. above to
chapter 6) introduced by 'maa'. The highly restricted
occurrence may be a hint, that we are dealing with
relicts of some older meaning of this morpheme which
needn't be related to its current usage.

3. Open questions
As F. himself points out (p. 193) there remain some
points which would be worth to investigate further:
a) which utterances needn't be marked by an evidential
marker at all?
b) what relationship does exist between the evidential
markers and other modality markers (e.g. the
potential marker)?
c) how does the position of the marker affect the
meaning of the sentence?

DeLancey, Scott, Mirativity, "Linguistic typology", 1:33-52

Palmer, Frank R., 1986, "Mood and modality", Cambridge

Willett, Thomas, 1988, A cross-linguistic survey of the
grammaticization of evidentiality, "Studies in Language"

Bernd Ruppel is graduate student at Erfurt University. He
is writing his thesis about epistemic modality in Latin.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue