<<

Evidentiality in Tibetic

Scott DeLancey

1 Introduction

The unusual and unusually prominent grammatical expression of epistemic status and information source in , especially ,1 has played a prominent role in the study of evidentiality since that category was brought to the attention of the general linguistic public in the early 1980’s. Work by many scholars over the past two decades has produced a substantial body of more deeply-informed research on Lhasa and other Tibetic languages, only a small part of which I will be able to refer to here (see also DeLancey to appear). In this chapter I will mostly discuss Lhasa data, but concentrate on the ways that this system is typical of the pan-Tibetic phenomenon. The system will be presented inductively, in terms of grammatical categories in Tibetic languages, rather than deductively, in terms derived from theoretical claims or typological generalizations about a cross-linguistic category of evidentiality.

1 The terms “Standard” and “Lhasa” Tibetan are sometimes used interchangeably, but they are not exactly the same thing. The early publications on the topic (Goldstein and Nornang 1970, Jin 1979, Chang and Chang 1984, DeLancey 1985, 1986) were based on work with educated speakers from Lhasa, and those works and the data in this paper represent the Lhasa dialect, not “Standard Tibetan”. 2

The first known Tibetic language is , the language of the Tibetan

Empire, recorded in manuscripts from the 7th century CE. In scholarly as well as general

use, ‘Tibetan’ is often applied to all descendants of this language, as well as to the written standard which developed from it. Since the modern descendants of OT are fully as divergent as, say, the Romance languages, it is preferable to distinguish them in our

terminology; following Tournadre (2014), I will refer to the family as Tibetic, and to the

individual languages as, e.g., rather than ‘’. ‘Tibetan’ is appropriate in reference to varieties shared across the Tibetan cultural sphere, i.e. ,

Written Tibetan, Standard Tibetan.

Evidentiality is not a feature of Classical Tibetan grammar. All the modern languages have some grammaticalized evidential constructions, and almost all (with the exception of Balti, see Bielmeier 2000) have a complicated system in which typical evidential categories such as inferential interact with the very unusual system of grammaticalized knowledge category which will be discussed below. But forms in these systems are often not cognate, and combined with the lack of evidence for evidentiality in older stages of the written language, this tells us that the modern systems represent a recent development, so that its prevalence across the family represents horizontal spread rather than common inheritance.

3

2 Evidentiality in

Evidential phenomena in Lhasa (Central Tibetic) had already been noted in the literature

before the concept and the associated term became a topic of wide interest in the 1980’s

(Goldstein and Nornang 1970, Goldstein 1973, Jin 1979). From the mid 90’s (Tournadre

1996) we have had more and better primary descriptions, including a growing body of

work on other Tibetic languages.

2.1 Previous work

The notoriety of Tibetic in evidentiality studies is due to an unusual grammatical

interaction with person. In almost all Tibetic languages there is a set of forms which

typically occur only with 1st person statements and 2nd person questions; following

current usage I will call these Egophoric forms (glossed EGO). In the following

exchange, from a folktale, we see the Egophoric Imperfective verb ending -gi.yod used in a 2nd person question in (1a), and a 1st person statement in (1b):2

(1a) rang nam.rgyun lto.byad ga.re za-gi.yod=zer

you ordinarily food what eat-IMPF.EGO=QUOT

‘‘What food do you usually eat?’ [he] said’

2 Examples are presented in standard Tibetan orthography, transliterated according to the Wylie system. Hyphens represent relevant morpheme breaks. Periods represent morpheme breaks in compounds, and in the composite verb endings described in §2.3. 4

(b) tsa zas chu ‘thung byas-kyi.yod

grass ate water drink do-IMPF.EGO

‘Eat grass, drink water is what I do.’ (Rabbit Eats the Baby)

In quotative complements these forms signal coreference of the subjects of the

complement and main clause, and this was the basis for Hale’s (1980) terms CONJUNCT and DISJUNCT for a similar phenomenon in Newar. In earlier work (DeLancey 1985,

1992) I adopted these labels, but, as this terminology grossly oversimplifies the

fundamental nature of the Tibetic phenomenon (and is not apt even for Newar, see

Hargreaves to appear), it was not universally popular with Tibetanists. Most scholars now

use Tournadre’s (1996) term EGOPHORIC for the 1st-person-associated category which I had called ‘conjunct’. (For a more detailed history of the concept see Hargreaves 2005,

Tournadre 2008).

2.2 The structure of information in Tibetic

The verbal systems of modern Tibetic languages grammatically mark a distinction among

three kinds of knowledge. The characteristic which has made Tibetic languages so

celebrated in the world of evidential studies is a set of forms which mark a statement as

representing certain kinds of personal knowledge. Following widespread current practice

we may call these forms Egophoric, and the category which they express, personal

knowledge. (I will use capitalized labels for specific grammatical constructions, and a

distinct set of labels, uncapitalized, for the functional categories which they express). 5

Tibetic languages are often described in terms of a two-way distinction between

Egophoric and other forms, but the latter can be further subdivided. A set of what I will call Factual forms express a category of assumed knowledge, i.e. assertions which are presented as not requiring any kind of evidential support. Finally, in perfective aspect, assertions which are presented as requiring evidential support are marked for Evidential status, i.e. as based on direct observation of the asserted fact or on inference from secondary evidence.

This distinction has not been much discussed outside the Tibetic literature, and there is no standard or consensus terminology for it. Hein (2001: 35) speaks of ‘focus on speaker’s involvement’, ‘focus on speaker’s unspecified knowledge’, and ‘focus on

speaker’s perception’; Zeisler (2004) of ‘personal self-evident’, ‘generic’, and

‘immediate perception’, DeLancey (2012) of personal, generic, and immediate forms,

Hill (2013) of Personal, Factual, and Testimonial categories, Tournadre and LaPolla

(2014) of Egophoric, Factual, and Sensory. The different labels emphasize different aspects of the same overall idea, distinguishing information which the speaker is the only possible source for, information which the speaker feels no need to provide any source for, and information which the speaker takes responsibility as the source for. Here I want to carefully distinguish between grammatical forms and the cognitive categories which they express. I will refer to the latter as personal, assumed, and contingent knowledge, and to the forms which express them in Tibetic languages as Egophoric, Factual, and

Evidential (Direct vs. Inferential). 6

The distinction is encoded in the copular system, and many of the verbal endings

which encode it are based on copulas and derive their evidential force from them.

Consider the Lhasa examples:

(2) nga-‘i nang bod-la yod

1sg-GEN home -LOC exist.PERSONAL

‘My home is in Tibet.’

(3) bod-la g.yag yog.red

Tibet-LOC yak exist.FACTUAL

‘There are yaks in Tibet.’

(4) bod-la moṭa mang.po ‘dug

Tibet-LOC auto many exist.DIRECT

‘There are lots of cars in Tibet.’

Examples like (2), with the Egophoric Existential yod, are often described as

expressing personal knowledge, but are perhaps better thought of as self-representation.

This statement, for example, does not necessarily describe any objectively verifiable fact, such as the existence of an ancestral homestead somewhere in Tibet; it simply expresses the speaker’s personal view of her place in the world. In (3), the use of the Factual form yog-red (sometimes written as yod-pa red) suggests that this statement is based on general knowledge, rather than personal experience or direct perception. The speaker 7

feels no need to justify the claim, and asks the addressee to simply take it as given. As we

will see below (§4.1), statements in this form do not necessarily represent generic

knowledge, but this is the appropriate form for statements which do. In (3), the Direct

Evidential ‘dug indicates that the statement is based purely on direct perception. It might

be said, for example, by a Tibetan who has lived abroad for many years, and returns for a

visit to discover that nowadays one sees many cars, which were much less common in the

old days. A speaker who had read or heard this information from others prior to visiting

the country could instead use yog-red, indicating that this is a generally-known fact, but

such a speaker could also use ‘dug as a way of expressing an acute awareness of how

much things have changed. In either case the choice of ‘dug indicates that the statement

expresses an impression based on immediate perception, rather than on any other basis.

2.3 The Lhasa verbal system

The written language includes a small paradigm of opaque verb inflections which

distinguish four stems, traditionally labeled Present, Past, Future and Imperative. In

Classical Tibetan these constitute finite verb forms. The Classical system does not

express evidential or egophoric categories. In the modern languages the verb usually requires further marking through innovative paradigms of suffixes and auxiliaries in order to function as finite, and this is where evidential categories are marked. Some

Tibetic languages have finite forms which are simple nominalizations, but in all the largest part of the verb paradigm consists of endings of two kinds: erstwhile serial and nominalized clause + copula constructions (DeLancey 2011). 8

Lhasa verb inflections are of these two kinds. The core set of endings, which

encode the Egophoric and Factual categories, originated and are written as a

nominalizing suffix -pa/-ba or -gi/-gyi/-kyi, followed by an equational (yin, red) or

existential (yod, ‘dug) copula. The Evidential forms were originally serialized verbs, song

‘went’ and bzhag ‘put [someplace]’ (DeLancey 1991):

Egophoric Factual Evidential

Direct Inferential

Perfective -pa.yin -pa.red -song -zhag

Perfect -yod -yog.red -‘dug

Imperfective -gi.yod -gi.yog.red -gi.‘dug ~ -gis

Future -gi.yin -gi.red

Table 1: Verb endings in Lhasa Tibetan

We will not discuss the Future forms further as they are not relevant to questions of evidentiality. Two other Egophoric forms, -byung and -myong, will be discussed in §3.3.

9

3 Egophoric forms

Most of the substantial body of work on Tibetic evidentiality has concentrated on the

typologically unusual Egophoric category. In previous work I have treated this category

as part of the evidential system (DeLancey 1985, 1990), with a special relationship with

(DeLancey 1997, 2001). Some Tibetic specialists have expressed discomfort

with this (e.g. Zeisler 2004, Tournadre 2008, Hill 2012), and with reason. The Tibetic

Egophoric category is not part of the evidential system; it is an independent, and more

fundamental, category which affects evidential meanings that come under its shadow.

Rather than an evidential category, Egophoric is a category to which evidentiality is not

applicable.

3.1 Self and Other

The most striking characteristic of Tibetic evidential systems is a fundamental distinction between a set of forms which report information to which the speaker has privileged access (in the sense of Hargreaves to appear), and those which report information which the speaker finds in the world outside – Denwood’s (1999) ‘self’ vs. ‘other-centered’ (see also Sun 1993, Hein 2001, inter alia). In earlier work I and a few others have referred to the distinction with the labels conjunct vs. disjunct. This opposition is fundamental to the verbal system of most Tibetic languages:

In Tibetan, the category of person constitutes an important factor which

determines much of the verbal morpho-syntax. The relevant oppositions,

however, are not the well-known trichotomy of first person (speaker), 10

second person (interlocutor) and third person (other referents), but rather a

referentially fluid dichotomous distinction between self-person and other

person. In rather vague terms, self-person sentences are marked as

utterances produced by oneself. (Sun 1993: 955-956)

As many scholars have noted, the fundamental distinction between these categories is that direct statements are subject to verification, and thus must be specified for evidentiality, while personal statements do not allow evidential marking:

A simple test allows us to distinguish the two categories: the evidential

morphemes are in general not compatible with first person. In effect, a

speaker cannot observe himself (except in particular situations: mirrors,

photographs, dreams, mental illness, daze, etc.) while the use of the direct

form poses no problem with first person … (Tournadre 1996: 201, my

translation, emphasis original)

This way of characterizing the distinction requires some care in defining the personal category, as some kinds of things which one can say about oneself are treated grammatically as Evidential rather than Egophoric, as we will see in §§3.2, 5.

3.2 Egophoric and Evidential

The basic pattern of interaction between the Egophoric and Evidential categories can be neatly illustrated with examples from Amdo (Eastern Tibetic; Sun 1993: 955-959, glosses 11

modified for consistency), where we find three different evidential forms in

complementary distribution based on person. The particle =nə marks the event as volitional on the part of the subject, =thæ indicates a report based on direct perception,

and =zəg one based on inference. With a volitional predicate such as xabda ndʑo ‘go

deer-hunting’, an affirmative statement with 1st person actor can only take the first of

these:

5a) ŋæ xabda shoŋ=nə

I deer.chase went=EGOPHORIC

‘I went deer-hunting.’

b) *ŋæ xabda shoŋ=thæ

c) *ŋæ xabda shoŋ=zəg

In contrast, with a 3rd person actor, volitional =nə is impossible, while the direct and

indirect evidentials are fine:

6a) dordʑe xabda shoŋ=thæ

Rdo.rje deer.chase(DAT) went=DIRECT

‘Dorje went deer-hunting.’ [direct evidence]

b) dordʑe xabda shoŋ=zəg

Rdo.rje deer.chase(DAT) went=INFERENTIAL 12

‘idem.’ [indirect evidence]

c) *dordʑe xabda shoŋ=nə

Rdo.rje deer.chase(DAT) went=EGOPHORIC

The impossibility of the volitional form with a non-1st person actor can be interpreted in

evidential terms: assuming fundamentally evidential senses for all forms in the paradigm,

=nə can be interpreted as indicating direct evidence of volition, which one can only have

with respect to oneself (DeLancey 1986, 1990, Hargreaves 2005, to appear). Here we are

concerned with the converse problem, the impossibility of direct or indirect evidential

specification with a 1st person volitional actor. Nonvolitional predicates with 1st person

actors behave as with non-1st person, that is, they disallow the =nə which indicates volitionality, and accept evidential marking:

7) ŋə ndaŋ hȵədɕed ji=zəg

I.ERG last.night sleep.talk do=INFERENTIAL

‘I talked in my sleep last night.’

8) ŋə ma-sæm shæ=ni der tɕag=taŋ=thæ

I.ERG NEG-think place=LOC dish break=AUX=DIRECT

‘I broke the dish by accident.’

13

Examples like these are easily interpreted in terms of evidence: I can only know about

something that I did in my sleep by indirect report, while in the ordinary circumstance I

would be aware of accidentally breaking something at the moment it happened, and thus

have direct evidence for it.

3.3 Subdividing the Personal category: Further egophoric forms

Lhasa has elaborated the Egophoric category more than some other Tibetic languages,

subdividing it according to the nature of the experience being reported. Like Amdo =nə in the preceding section, the Lhasa Egophoric forms listed in Table 1 are used only to report volitional actions. In some languages, including some other Tibetic languages, non-volitional actions with a 1st person actor require Evidential forms, and this occurs in

Lhasa as well, as we will see in §5. But Lhasa also has two forms marking particular

types of 1st person statement where volition is irrelevant or not involved. Both occur with

all persons as independent verbs or auxiliaries, but as verb suffixes occur only with a 1st

person argument (see DeLancey 1991).

The first of these, -byung, occurs freely as a main verb meaning approximately

‘happen, come to pass’:

(9) .'dra byas=nas sdad.bzod mi-bde.ba gtan.gtan

thus did=NF peace NEG-peaceful really

byung-yog.red

happen-PERFECT.FACTUAL 14

‘[They] having done that, [he] came to have no peace.’

As a verb suffix, -byung occurs only with a 1st person argument functioning as some kind

of Goal (i.e. a Patient, Theme, Recipient, or spatial Goal). We can see the contrast

between Egophoric Goal and non-Egophoric Evidential constructions in (10a-d):

(10a) kho.tsho-s nga.cag spun.kya drug btang-byung yin.na’i

they-ERG we kin six send-EGO.GOAL but

(b) am.khag-la seng.ge gzhan.dag chen.po cig thug=nas

way-LOC lion other big a meet=NF

(c) nga-‘i spun.kya gzhan.dag lnga bsad-song

I-GEN kin other five killed-PERF

(d) nga gcig.po bsad-ma-byung

I alone kill-NEG-EGO.GOAL

‘They [the Animals’Committee] sent six of us kin, but meeting another big lion

on the way, [it] killed the other five of my relatives, only me [it] didn’t kill.’

(Rabbit Fools Lion)

The (fictitious) killing of the narrator’s companions is presented in (10c) with the direct 15

evidence form -song, telling us that the source for this information is his witnessing the

event. The Animals’ Committee’s sending of the narrator and his (fictitious) kin, and the

(fictitious) lion’s sparing of the narrator, however, are marked with -byung, presenting

these not as a observation, but as personal experience.

The other Egophoric form is Experiential -myong, which expresses the state of

having had the experience described by the rest of the clause. This is rare in the spoken

language as an independent verb, but occurs freely as an inflected auxiliary:

(11) khye.rang rgya.gar-la phebs myong-zhag

you -LOC go(HON) experience-PERF.INFERENTIAL

‘You have been to India, I see.’

(12) kho bod-la ‘gro myong-ba.red

he Tibet-LOC go experience-PERF.CONTINGENT

‘He has been to Tibet.’

Like -byung, it occurs as a verb suffix, rather than an inflected auxiliary, only with a 1st

person experiencer argument:

(13) nga rgya.gar-la las.ka byed-myong

I India-LOC work do-EXPERIENCE

‘I used to work in India.’

16

(14) rang-gi thog-la=ya, nga-s mgo ‘khyer

self-GEN lifetime-LOC=CONTR I-ERG head take

yongs-pa thengs.ma gnyis mthong-myong

came-NMZ occasion two see-EXPERIENCE

‘In my own lifetime I have on two occasions seen a head taken.’ (Head)

4 Assumed and contingent knowledge

Some treatments of Tibetic languages describe a simple binary distinction between

Egophoric forms and everything else. But the ‘other’ category is better understood as bifurcated into what I am calling the Evidential category, expressing contingent knowledge which requires evidential specification, and the Factual category, expressing propositions which are to be taken as assumed without justification. I will outline the use of each category in this section; we will see how the two categories are related in narrative in §6.

4.1 Assumed assertion and the Factual paradigm

The formal category which I will gloss as FACTUAL (following Hill 2013 and Tournadre and LaPolla 2014), has attracted less attention then the Egophoric and Evidential forms, and is less well-understood. Goldstein and Nornang (1970) call these forms ‘Narrative’;

Zeisler (2004) labels them ‘Generic’. Both narrative style and expression of generic knowledge are prominent functions of the form, but neither is in fact a basic meaning, 17 which is simply the absence of any specification of source of knowledge. The Factual verb endings are the only forms in the system which neither assert nor imply anything about the source of information.

For this reason attempts to define this category tend to be rather vague:

The unmodalized declarative construction is found mainly in generic

statements, proverbial sayings, and stories, but is ill-suited for informative

reports. (Sun 1993: 951)

The category is difficult to describe explicitly precisely because it has no evidential value whatever. As Sun and others note, this is the grammatical form used to express generic knowledge:

The general fact copula is for those very generally known facts about the

world, such as sugar being sweet. It is not frequently used in daily

interaction. (Gawne 2014: 78)

But this is also the form used in narrative, whether fictional, historical or autobiographical. Essentially assumed knowledge is just that, information which the addressee is asked to take for granted, without worrying about source or justification:

18

The morphemes of this category are mainly used with non-speaker agents

and patients … to express the speaker’s knowledge of the verbal action

without specifying how this knowledge is/was gained. (Hein 2001: 43)

Emphasizing the use of this form to express ‘generally known facts’ is thus misleading. It is used in that function, because, by definition, one can always assume that one’s interlocutor shares one’s attitude toward such facts, and so their evidential status is not in question. And this is the only function which can be easily described. But sometimes in conversation, and generally in narrative, claims which do not have this status may be assumed for the immediate purposes of the discourse, and the basic function of the

Factual verb forms is to mark an assertion as having this status.

In searching for a positive evidential meaning for this set of forms, it is tempting to set aside as an oddity the fact that these are the forms used in narrative, including much autobiographical narrative. But this establishes the true function of the Factual category: it simply disregards the question of evidence. In narrative, including both personal and witnessed, and both true and fictional, most of the story is told in the Factual form.

Consider examples (15-16), from an autobiographical narrative (which we will see more of in §6). The two sentences both refer to the same individual, the main character in this episode, and illustrate the Perfective (15) and Perfect (16) Factual forms:

(15) de-’i rjes-la=ya thengs.ma gsum.pa de

that-GEN after-LOC=CONTR occasion third that

19

drug.cu=re.gnyis lo-la yin.na, re.gsum lo-la yin.na

sixty=two year-LOC maybe three year-LOC maybe

de.’dra gcig-laya kho tshur log yongs-pa.red

thus a-LOC he back return came-PERF.FACTUAL

‘After that, the third time was in ’62 or 3 maybe, he came back again. (Head)

(16) rnam.lha gtan.gtan sdad.bzod mi-bde-pa

Namla really peace NEG-peaceful-NOMZ

byung-yog.red

get-PERFECT.FACTUAL

‘Things had gotten really difficult for Namla.’ (Head)

These events are related as part of a story in which the narrator herself is a player, and so could in principle be expressed with Direct or Inferential Evidential forms. But they are not major events in the story, and the narrator had no direct connection with them, so their evidential status does not need to be specified.

20

4.2 Contingent assertion and the Evidential paradigm

I will use the term ‘contingent’ to refer to the domain of knowledge for which Tibetic languages mark the cross-linguistically typical evidential distinction between events directly perceived by the speaker and those inferred from indirect evidence. This is information which the speaker does not ask the addressee to take for granted; rather, it is

information which the speaker presents as requiring some additional specification as to

source in order for the addressee to accept it. Statements which are neither assumed nor

asserted on the basis of the speaker’s personal authority must be marked as Direct or

Indirect. Although some languages elaborate the system, e.g. the non-visual direct

evidential rak in Ladakhi (Bielemeier 2000) or the ‘acquired knowledge’ category of

Dzongkha, (van Driem 1992:169), the fundamental distinction in all languages is

between direct evidence and inference.

In Lhasa and other Central varieties the Direct Evidential forms are

perfective -song, perfect -‘dug, and imperfective -gi.‘dug. The Inferential, indicated by -

zhag, is marked only on completed events, since it entails that the speaker knows of the

event through perceiving its aftermath. The difference is illustrated in examples (17-18),

adapted from Denwood 1999: 159–161 (see also DeLancey 1985).

(17) de.ring char.pa btang-‘dug

today rain fall-PERFECT.DIRECT

‘It has been raining today.’

21

(18) de.ring char.pa btang-zhag

today rain fall-PERFECT.INFERENTIAL

‘It has been raining today.’

Example (17), with -‘dug, entails that the speaker has seen the rain directly. In (18), with the inferential -zhag, the speaker has not actually seen the rain, but infers it from secondary evidence, e.g. fresh puddles on the ground.

As further illustration compare (19) (previously presented as 10c) and (20):

(19) nga-‘i spun.kya gzhan.dag lnga bsad-song

I-GEN kin other five killed-PERF.DIRECT

‘[He] killed the other five of my kin.’ (Rabbit Fools Lion)

In (19) the speaker is relating an event to which he was a witness. In his story (which is in fact a lie), a lion set upon him and his five brother rabbits and killed all but him. In (20), from another folktale, the protagonist (rkun.ma-gi gźu.gu ‘Thieving Tail’) wants to drive off a band of thieves who are dividing up their loot, which he wants to take for himself.

He calls out an alarm, and the thieves, thinking that the shout is coming from a police patrol, react:

(20) kho.tsho-‘i sems-la=ya,

they-GEN mind-LOC=CONTR

22

‘da rgyal.po-s rkun.ma-gi gźu.gu=ya rjes.’ded btang-zhag’

now king-ERG Thief-GEN Tail=contr pursuer send-INFERENTIAL

bsam-rjes … bros phyin-pa.red think-NF fled went-PERF.FACTUAL

‘They, thinking ‘Now the king has sent pursuers after the thieves,’ … fled.’ (The Hungry

Dried-up Goat Tail)

The thieves express their conclusion with the inferential -zhag because they did not have

direct experience of the king dispatching pursuers, but came to this conclusion only by

inference from hearing someone shouting. (If the thieves said, ‘Pursuers are after us!’, this could be said with the Direct Evidential. But even if I directly perceive that I am being pursued, that is not direct evidence that someone else sent pursuers after me; that can only be an inference from the fact that they are after me).

Because the Direct forms assert that the speaker directly perceived the event, they generally occur only in contexts where that direct perception is the speaker’s only warrant for the statement. Normally this will only be true of relatively recently acquired knowledge:

The use of this what may be termed the immediate evidential indicates that

the speaker’s basis for his assertion comes solely from perceptible

evidence directly present in the immediate speech-act situation. What is

crucial here is that the speaker implicitly denies having any information 23

regarding the situation prior to the current perceptual experience; in other

words, this knowledge is entirely novel for the speaker. (Sun 1993: 996-7)

In my earlier work I therefore treated these forms mirative, but this is misleading, as their basic meaning is unquestionably fundamentally evidential, not mirative (DeLancey 2012,

Hill 2012).

5 Evidential and Egophoric

There is an obvious relationship between the cognitive category of personal knowledge and the grammatical category of first person, but the distinction between Egophoric and other verb endings is clearly not verb agreement or argument indexation. There are a few situations where statements about non-1st persons can be in Egophoric form, and

numerous types of statement which one can make about oneself which are not marked as

Egophoric.

5.1 Personal ≠ Egophoric

Statements about others, or about material objects, can sometimes represent personal

knowledge. For example, statements about the speaker’s close family members may use

Egophoric forms:

(21) nga-‘i ama bod-la yod

1sg-GEN mother Tibet-LOC exist.PERSONAL

‘My mother is in Tibet.’ 24

(21) would be a normal way for a Tibetan visiting abroad to speak about his family, or, evocatively, for a Tibetan expatriate to talk about his family and homeland. On the other hand, if I am not Tibetan, and my mother is an inveterate tourist, such that I can’t keep track of her wanderings, and I just found out her whereabouts this morning when she phoned me from , I would more appropriately use an Evidential form.

Conversely, information about oneself that is not part of one’s self-presentation can be presented as contingent information, using Evidential forms. If I am in the habit of carrying money with me, or if I made sure that I had some when I left the house, then when it comes time to pay for refreshments I would typically say:

(22) nga-r dngul tog.tsam yod

I-DAT money some exist.PERSONAL

‘I have some money.’

But if I don’t think I am carrying any money, and reach into my pocket and find some there, I could indicate the unexpectedness of the fact by using the Direct Evidential form:

(23) nga-r dngul tog.tsam ‘dug

I-DAT money some exist.DIRECT

‘[I see] I have some money.’

25

5.2 Evidentiality and volition

One of the most intriguing phenomena connected with Egophoric forms is their strong association with volitionality. The Egophoric forms in Table 1 always entail that the clause describes a volitional act on the part of the speaker. An intrinsically non-volitional verb such as rjed ‘forget’ requires Evidential marking:

(24) kho-'i ming brjed-zhag

he-GEN name forgot-INFERENTIAL

‘I’ve forgotten his name.’ (Head)

Even imagining that there is some specific moment at which a fact disappears from one’s memory, one is never conscious of that moment, and can only infer it from the present fact that one has searched one’s memory and failed to find the fact one was looking for.

The same is systematically true for verbs of involuntary perception:

(25) nga mthong-gis

I see-IMPF.DIRECT

‘I see [it].’

It has been argued (DeLancey 1986, 1990, Garrett 2001) that this is evidence that the

Egophoric forms are evidentials, expressing direct experience of the act of volition which leads to an action. Events such as seeing and forgetting have no such antecedent act of volition, and therefore cannot be described with Egophoric forms. 26

5.2 Endopathic states

Since they are inherently nonvolitional, endopathic states such as ‘hungry’ and ‘sick’

likewise do not typically take volitional Egophoric markings. With a 1st person argument

they may take either the Egophoric Goal form -byung, or else Evidential marking:

(26) nga na-byung

I sick-EGO.GOAL

‘I got sick.’

(27) nga na-‘dug

I sick-PERFECT.DIRECT

‘I’ve gotten sick.’

Interestingly, such verbs may appear in the Imperfective Egophoric form to indicate a chronic state:

(28) nga na-gi.yod

I sick-IMPF.EGO

‘I’m chronically sick.’

It is not clear why this verb form, which is otherwise strongly associated with

volitionality, occurs in this function. It may be that we need to reconsider the volitional

component of these Egophoric forms as just one manifestation of a more basic sense of 27 self-presentation – a current state or feeling is not part of who I am, while a chronic condition is.

6 The system in use

Let us look at some examples from an autobiographical narrative. The speaker is telling a story of an atrocity which she witnessed during one of the drives to bring the nomads of

Tibet into the commune system in the early 1960’s. Most of the narrative is in the Factual form, expressing assumed knowledge, as is typical in narrative. At a few points the narrator shifts to Egophoric or contingent forms, and here we can see the system in use.

The story concerns a nomad who fled and rejoined the commune several times, finally resulting in Red Guards chasing him out into the countryside, beheading him, and bringing the head back as a trophy. Early in the story, after one return to the commune, he is summoned to speak with commune officials. The narrator has to translate, because the nomad speaks only Amdo. Everything recounted up until now has been presented as assumed knowledge, as in exx. (15-16), although much of the story was also personally witnessed and experienced by the narrator. But only the nomad protagonist, the two officials, and the narrator are present at this event. Thus events in this episode are either contingent, and thus specified for Evidential value:

(29) gnyis-kyis kho-r=yang skad.cha bshad-song

two-ERG he-DAT=CONTR speech spoke-PERF.DIRECT

‘[Those] two [officials] spoke to him.’

28

Or personal, and thus marked as Egophoric:

(30) de-'i skabs-la=ya, nga-s skad.gyur byas-pa.yin

that-GEN time-LOC=CONTR I-ERG translation did-PERF.EGO

‘That time I translated [for them].’

At the conclusion of this episode, the narrative reverts to Factual forms. Example (31) is the end of the discussion between the nomad and the officials, and thus is still contingent, marked for Direct evidentiality:

(31) kho-s mtsho.dmar nang-la sdad-kyi.yin=ze

he-ERG Red Lake in-LOC stay-IMPF.EGO=QUOT

zer-song

say-PERF.DIRECT

‘… he said ‘I will stay in Red Lake.’

The next sentence, wrapping up the episode, returns us to the main narrative track, and hence uses the Factual form:

(32) kho mtsho.dmar-gyi shang nang-la bsdad-pa.red

he Red Lake-GEN village in-LOC stayed-PERF.FACTUAL

‘He stayed in Red Lake village.’ 29

Although some authors suggest that the difference between examples like this these is

that (32) does not report direct evidence, this is not exactly correct. The narrator did

directly observe most of the events in the story, including the nomad’s return, and the fact

that he then took up residence in Red Lake. (If she did not actually witness the moment

when he entered Red Lake village, she could have used the Inferential form). The choice

of verb form is a matter of presentation, not a report of the objective circumstances under

which the speaker came to know the information. By the shift from Factual to Direct the

narrator presents the episode in which the officials interrogate the nomad as one in which

she was directly involved. The rest of the story is presented as a sequence of events in

which she was not directly involved, even though, like other people living in the

commune at the time, she did witness them directly or indirectly. Although the speaker

does have direct knowledge of the nomad’s taking up residence in Red Lake, she is

presenting it as simply one more event in the narrative, and therefore the source of her

knowledge is irrelevant.

It is important to understand that these categories are not objectively recoverable.

The alternation in exx. (31-32) reflects the speakers’ desire to present certain events in

the narrative as part of her autobiography, as opposed to others which are presented as

generally available information. There is no way that one could consistently predict

choice of verb form from some ‘objective’ characterization of the situation.

Contrast this with the narrator’s comment after the climactic point when the Red

Guards return, brandishing the recalcitrant nomad’s severed head as a warning to others:

30

(33) de nga-s nga-ra-‘i mi.tshe ‘di-‘i

that I-ERG I-REFL-GEN lifetime this-GEN

thog-la=ya mgo mthong-ba dang.po yin

in-LOC=CONTR head see-NOMZ first be.EGO

‘That was the first time I ever in my life saw a (severed) head.’

This is presented neither as assumed nor contingent, but as personal, to emphasize that the sentence is about her personal experience. It is worth noting that examples like this demonstrate that the association of the Egophoric forms with first person is not a syntactic rule, but a matter of presentation. Here the first person narrator is a syntactic argument only of the nominalized clause ‘I in my lifetime seeing a head’, not of the matrix clause, and thus has no syntactic relation to the main verb yin. The syntactic arguments of yin are de, referring to the preceding story, and the nominalized clause ‘I seeing a head’ modified by dang.po ‘first’. We find the personal form here not because of any syntactic trigger, but to emphasize that the speaker is presenting this as part of her personal story, not as assumed knowledge. A few lines later, marking the end of this entire episode and the transition to the next, the narrator says;

(34) mgo mthong-ba dang.po de red

head see-NOMZ first that be.FACTUAL

‘That was the first [occasion of] seeing a [severed] head.’

31

This sentence refers to and says the same thing as (33) a few lines earlier, but it has a

different narrative function, and thus is marked as assumed rather than personal

knowledge.

Abbreviations

AUX Auxiliary

CONTR Contrast

DAT Dative

EGO Egophoric

ERG Ergative

GEN Genitive

HON Honorific

IMPF Imperfective

LOC Locative

NEG Negative

NF Nonfinal

OT Old Tibetan

PERF Perfective

QUOT Quotative 32

REFL Reflexive

References

Bielmeier, Roland 2000 Syntactic, semantic, and pragmatic-epistemic functions of

auxiliaries in Western Tibetan. Linguistics of the Tibeto-Burman Area 23.2: 79-125.

Chafe, Wallace, and Johanna Nichols (eds.) 1986 Evidentiality: The Linguistic Coding

of Epistemology. Norwood, NJ: Ablex.

Chang, Betty Shefts, and Kun Chang 1984 The certainty hierarchy among Spoken

Tibetan verbs of being. Bulletin of the Institute of History and Philology,

Academia Sinica 55: 603–635.

DeLancey, Scott 1985 Lhasa Tibetan evidentials and the semantics of causation.

Proceedings of the Eleventh Annual Meeting of the Berkeley Linguistics Society,

65-72.

DeLancey, Scott 1986 Evidentiality and volitionality in Tibetan. In: Chafe and

Nichols, 203-13.

DeLancey, Scott 1990 Ergativity and the cognitive model of event structure in

Lhasa Tibetan. Cognitive Linguistics 1.3: 289-321.

DeLancey, Scott 1991 The origins of verb serialization in Modern Tibetan. Studies

in Language 15.1: 1-23.

DeLancey, Scott 1992 The historical status of the conjunct/disjunct pattern in

Tibeto-Burman. Acta Linguistica Hafniensia 25: 39-62.

DeLancey, Scott 1997 Mirativity: The grammatical marking of unexpected

information. 1.1: 33-52. 33

DeLancey, Scott 2001 The mirative and evidentiality. Journal of Pragmatics

33.3: 371-384.

DeLancey, Scott 2012 Still mirative after all these years. Linguistic Typology

16.3: 529-564.

DeLancey, Scott 2011 Finite structures from clausal nominalization in Tibeto-

Burman languages. In Yap, Foong Ha, Karen Grunow-Hårsta & Janick Wrona

(eds.), Nominalization in Asian Languages: Diachronic and Typological

Perspectives, 343-62. Amsterdam/Philadelphia: John Benjamins.

DeLancey, Scott to appear Evidentiality and mirativity in Tibeto-Burman. In:

Hans Hock, K. V. Subbarao, and Elena Bashir (eds.), The Languages and

Linguistics of : A Comprehensive Guide. Berlin: Mouton de Gruyter.

Denwood, Philip 1999 Tibetan. (London Oriental and African Language Library

3). Amsterdam/Philadelphia: Benjamins.

Driem, George van. 1992. The grammar of . Thimphu: Dzongkha Development

Commission, Royal Government of .

Garrett, Edward 2001 Evidentiality and assertion in Tibetan. Ph.D. dissertation,

Department of Linguistics, UCLA.

Gawne, Lauren 2013 Lamjung Yolmo copulas in use: evidentiality, reported

speech and questions. Ph.D. dissertation, Department of Linguistics and Applied

Linguistics, University of Melbourne.

Gawne, Lauren 2014 Evidentiality in Lamjung Yolmo. Journal of the Southeast

Asian Linguistics Society 7: 76-96.

34

Goldstein, Melvyn 1973 Essentials of Modern Literary Tibetan: A Reading Course and Reference Grammar. Also published 1991: Berkeley/Los Angeles: University of

California Press.

Goldstein, Melvyn, and Nawang Nornang 1970 Modern Spoken Tibetan: Lhasa

Dialect. Seattle/London: University of Washington Press.

Hale, Austin 1980 Person markers: Finite conjunct and disjunct verb forms in

Newari. In: Ronald Trail (ed.), Papers in South-East Asian Linguistics no. 7, 95-106.

(Pacific Linguistics Series A 53). Canberra: Australian National University.

Hargreaves, David 2005 Agency and intentional action in Kathmandu Newar.

Himalayan Linguistics Journal 5: 1-48.

Hargreaves, David to appear ‘Am I blue?’: Privileged access constraints in

Kathmandu Newar. In Simeon Floyd, E. Norcliffe and L. San Roque (eds), Egophoricity.

Amsterdam and Philadelphia: Benjamins.

Häsler, Katrin 2001 An empathy-based approach to the description of the verb system of the Dege dialect of Tibetan. Linguistics of the Tibeto-Burman Area 24.1: 1-34.

Hein, Veronika 2001 The role of the speaker in the verbal system of the Tibetan dialect of Tabo/Spiti. Linguistics of the Tibeto-Burman Area 24.1: 35-48. 35

Hein, Veronika 2007 The mirative and its interplay with evidentiality in the

Tibetan dialect of Tabo (Spiti). Linguistics of the Tibeto-Burman Area 30.2: 195-214.

Hill, Nathan 2012 ‘Mirativity’ does not exist: ḥdug in ‘Lhasa’ Tibetan and other

suspects. Linguistic Typology 16.3: 389-434.

Hill, Nathan 2013 ḥdug as a testimonial marker in Classical and Old Tibetan.

Himalayan Linguistics 12.1: 1-16.

Jīn, Péng 1979 Lùn Zàng yǔ Lāsà kǒuyǔ dòngcí de tèdiǎn yǔ yǔfǎ jiégòu de

guānxì (On the relations between the characteristics of the verb and the syntactic

structure in Spoken Tibetan (Lhasa dialect)). Minzu Yuwen 1979.3: 173-81.

Sun, Jackson T.-S 1993 Evidentials in Amdo Tibetan. Bulletin of the Institute of

History and Philology 1993: 945-1001.

Tournadre, Nicolas 1996 Comparaison des systèmes médiatifs de quatre dialects tibétains (tibétain central, ladakhi, dzongkha et amdo). In: Zlatka Guentchéva, (ed.)

L'Énonciation médiatisée, 195-211. Paris and Louvain: Peeters.

Tournadre, Nicolas 2001 Final auxiliary verbs in Literary Tibetan and in the dialects.

Linguistics of the Tibeto-Burman Area 24.1: 49-111. 36

Tournadre, Nicolas 2008 Arguments against the concept of ‘conjunct’/’disjunct’ in

Tibetan. In: Brigitte Huber, Marianne Volkart, and Paul Widmer, (eds.) Chomolangma,

Demawend und Kasbek: Festschrift für Roland Bielmeier zu seinem 65. Geburtstag, 281-

308. Halle: International Institute for Tibetan and Buddhist Studies GmbH.

Tournadre, Nicolas 2014 The Tibetic languages and their classification. In: Nathan

Hill and Thomas Owen-Smith (eds.), Trans-Himalayan linguistics, historical and descriptive linguistics of the Himalayan area, 105-130. Berlin: Walter de Gruyter.

Tournadre, Nicolas, and Randy LaPolla 2014 Towards a new approach to evidentiality: Issues and directions for research. Linguistics of the Tibeto-Burman Area

37.2: 240-263.

Woodbury, Anthony 1986 Interactions of tense and evidentiality: A study of Sherpa and English. In: Wallace Chafe and Johanna Nichols (eds.), Evidentiality: The Linguistic

Coding of Epistemology, 188-202. Norwood, NJ: Ablex.

Zeisler, Bettina 2004 Relative Tense and Aspectual Values in Tibetan Languages.

Berlin/New York: Mouton de Gruyter.