By Stephen H. Levinsohn
The Balochi particle ki is a “conjunction of general subordination” (Delforooz 2010:16), found in adverbial, relative and complement clauses (sec. 1). This paper concentrates on its presence before complement clauses and, in particular, those that report a speech.
Roberts (2009:295, 300) claims for Persian that the “clause linkage marker” “ke is used primarily in spoken texts to give prominence to speeches that the author considers are important to the story”. In the pre-defence version of his dissertation, Delforooz (ibid. 224) likewise suggested that, when Balochi ki introduces a direct speech, it “has a highlighting function. … The marked speeches push the story forward to its goal”.
However, further research suggested that, when ki introduces a reported speech in Balochi, it is the consequences of the speech that are highlighted, rather than the speech itself (see sec. 3). In particular, “when a question is answered, the answer is usually more important than the question, which is why the effect of marking a question with ki is to highlight the answer” (Delforooz 2011:227). Furthermore, ki may introduce a speech, not to highlight anything, but to indicate that its words represent the gist of what one or more people said, are to say or could have said, rather than being uttered on a particular occasion (ibid. 229).
This paper argues that the reason that ki can have such diverse effects is that it is an “indicator of interpretive use” (Blass 1990:104). This means that, when introducing a reported speech, the speech is to be understood not as a description of what was said on a particular occasion, but rather as a representation of an utterance or thought (sec. 2).
1. Ki as a Conjunction of General Subordination
Linguists usually posit three basic types of subordinate clauses: adverbial, relative, and complement (see Whaley 1997:247) and ki is found in all three in Balochi. This section briefly illustrates its typical use in the three types.
1.1 Adverbial clauses subordinated by ki
In adverbial clauses, ki most often occurs immediately before the subordinated verb or verb phrase, and after the subject, if present.
When the clause is post-nuclear, it is usually of reason or result (if realis) or purpose (if irrealis) (Delforooz 2011:200), though other relations are also found. In (1), the subordinated clause (1b)
gives a reason for the event of the main clause (1a).2
(1)a nam=ay gua galaw-a išt-ant xarmizza
name=PC.3s then melon-O leave.PST-3p xarmizza
b ki mizzag=ay awal xar burt
SUB taste=PC.3s first donkey take.PST.3s
Then they named the melon ‘xarmizza’, since it was a donkey that tasted it first. (XM 109-110)3 mWhen the clause is pre-nuclear, it is usually of time, though other relations are also found.
In (2), the subordinated clause (2a) gives the time for the event of the main clause (2b-c).
(2)a badiša ki eši-ra dist
king SUB DEM-O see.PST.3s
b zant understand.PST.3s
c ki diga kišwar-ay mardum=e
SUB other country-GEN person=IND
When the king saw him, he understood that (he was) a person from another country. (BP 27-29)4
What is noteworthy for the present paper is that, whether pre-nuclear or post-nuclear, the information in adverbial clauses that are subordinated with ki can easily be related to information that has recently been stated in the discourse (see the Conclusion). Thus, in (1), the hearers already know that it was a donkey that first ate the melons (XM 83-87). Likewise, in (2), the hearers already know that the traveller has arrived in a country whose king is out hunting (BP 24-26).
1.2 Relative clauses subordinated by ki
Relative clauses are commonly divided into two types: restrictive and non-restrictive.
Restrictive (identifying) relative clauses serve to delimit the potential referents (Comrie
1989:138). In restrictive relative clauses in Balochi, ki occurs immediately after the head NP that it modifies. In (3), the head NP is har _azau har mewage ‘every kind of food and fruit’, and the relative clause ki dist limits the referent to that ‘which he saw’.
(3) am=e har _aza=u har mewag=e ki dist
EMP=DEM each food=& each fruit=IND SUB see.PST.3s
He bought and ate every kind of food and fruit which he saw. (MG 58-60)5
Non-restrictive relative clauses serve “merely to give the hearer an added piece of information about an already identified entity, but not to identify that entity” (ibid.). Non-restrictive relative clauses in Balochi begin with ki. In (4), for instance, the head NP is yakk tiri bar_e ‘a light post’
(4a), and the relative clause of (4b) gives additional information about it.
(4)a e badiša bi=m=e wat-i šar-ay wasat-(t)a yakk tir=i bar_=e dašt
DEM king in=EMP=DEM RFL-GEN town-GEN middle-OBL one pole=IZ electricity=IND have.PST.3s
b ki harci am=e tilipunan-i sim=at-ant bi am=eši wasl=at-ant
SUB whatever EMP=DEM telephones-GEN wire=COP.PST-3p to EMP=DEM.OBL connected=COP.PST-3p
This king had a light post in the centre of his town, to which were connected whatever phone
wires there were. (XM 3-5)
(5b) is different, as ki introduces the presupposition (established information—see (5a)) in the equivalent of an “it-cleft structure” (Levinsohn 2011a:65).
(5)a Finally, he became the (biggest) trader of the world, the trader of the entire world.
b FOCUS presupposition
marg na(y)-at ki ta ar but
death NEG-come.PST.3s SUB trader become.PST.3s
(It was through) death not coming (contrary to expectation)6 that he became a trader. (MG 94-97)
1.3 Complement clauses subordinated by ki
In complement clauses, ki typically introduces the complement, as in (2c) above and (6).
(6) b(y)-a ki e rang eš-i mas e rang gu””o=at=o
SBJ-come.PRS SUB DEM manner DEM-GEN mother DEM manner strangled=COP.PST.3s=&
It happened that its mother was strangled in this way and… (XM 59)7
When the complement is a reported speech, however, the default is for ki to be absent (see (7) below), so its presence before a reported speech is noteworthy. The following sections discuss the pragmatic effects of introducing a reported speech with ki.
2. Ki as an Indicator of Interpretive Use
“When a speech is reported directly, it usually purports to describe a state of affairs—what was said on a particular occasion (a “descriptive use”—Sperber & Wilson 1986:224-31).
However, some reported speeches do not inform the reader of what was said so much as represent an utterance or thought that resembles it.
“Some languages have ‘an explicit linguistic indicator of interpretive use’ (Blass 1990:104)
whose function is to indicate that the speech concerned is not describing what was said on a particular occasion, but rather represents an utterance or thought. Such markers are often found in speech orienters.
“A variety of circumstances motivate the presence of interpretive use markers.”
(Levinsohn 2011a:115) I now argue that Balochi ki is such an interpretive use marker and discuss some factors that motivate its presence.
When ki does not introduce a reported speech, then the sentence concerned purports to describe what was said on a particular occasion. So, in (7), by not using ki, the storyteller implies that, on a particular occasion, someone said to the king, ‘Lord king, it is a dragon’.
(7) gušt badša saib aždiya=(y)e
say.PST.3s king master dragon=IND
He said, ‘Lord king, it is a dragon’. (XM 15-16)
One of the reasons for introducing a reported speech with ki is to indicate that the words that follow were NOT said on a particular occasion. In (8c), for instance, the words that follow ki are not a report of what Sabzo said on some occasion. Rather, they are hypothetical, representing what she might have said, had Khudanizar Khan not given her a generous dowry.8
(8)a am=e ku
b ki yane sabzo ma-guš-it wat-i dil-ay ta pikr ma-kan-t
SUB means Sabzo PROH-say.PRS-3s RFL-GEN heart-GEN in thought PROH-do.PRS-3s
c ki xudanizar xan mna bi xwar-en mard=e dat
SUB Khudanizar Khan me to lowly-ATTR man=IND give.PST.3s
he collected them (almost 200 sheep and 100 camels) in order for Sabzo not to say, not to think in her heart, ‘Khudanizar Khan married me off to a lowly man’. (KH 110-113)
Similarly, what follows ki in (9c, e) is not a report of what Pirakk said on a specific occasion.
Rather, it gives the gist of what he said a number of times, as the imperfectives (9b, d) imply.
(9)a but=u mardum=e ki b(y)-at-en bi pirakk-ay gis-a
become.PST.3s=& person=IND SUB SBJ-come.PST-SBJ.3s to Pirakk-GEN house-OBL
b bass ša xudanizar kissa=a kurt
just from Khudanizar story=IMPF do.PST.3s
c ki xudanizar pa(m)=man e rang kurt=u
SUB Khudanizar for=me DEM manner do.PST.3s=&
d gua dem-a=a gardent
then face-O=IMPF turn.PST.3s
e ki sabzo xudanizar am=e rang kurt ya na-kurt
SUB Sabzo Khudanizar EMP=DEM manner do.PST.3s or NEG-do.PST.3s
‘It so happened that, whenever someone came to Pirakk’s house, he used to talk so much about
Khudanizar: “Khudanizar did this kind of thing for me”; and then he would turn (to his wife):
“Sabzo, Khudanizar did this kind of thing, didn’t he?” (KH 125-131)
When a well-known folktale is related, it is not usually thought of as something that was told on
one particular occasion, so it is natural that ki should introduce it, as in (10b) (the present is used
b ki yag badiša=(y)e=at
SUB one king=IND=COP.PST.3s
They say that there was a king. (XM 1-2)
Ki may also introduce the gist of a whole conversation, as in (11b). It is most unlikely that three thieves would chorus together, “Let’s go and steal from the king’s treasury”! Rather, this speech represents what they decided after a discussion.
(11)a say duzz irada kurt-at-ant
three thief desire do.PST-COP.PST-3p
b ki b-raw-an badiša-ay xazanag-a b- an-an
SUB SBJ-go.PRS-1p king-GEN treasury-O SBJ-hit.PRS-1p
Three thieves had taken a decision: “Let’s go and steal from the king’s treasury”. (PJ 60-62)
3. Ki and Grounding
Extracts (10) and (11) above provide evidence that the use of ki to introduce a reported speech does not highlight the speech concerned.
In (10b), the copula indicates that the information concerned is of a background nature, as far as the theme line of the narrative is concerned (Levinsohn 2011a:68), even though it is introduced with ki.
In (11a), a pluperfect (translated ‘had taken a decision’) introduces the speech. Pluperfects are associated with back grounding in narrative, as they are used for events that take place prior to the theme-line events (ibid. 70). So the reported speech of (11b) is back grounded with respect to what follows (the later decision of the thieves to pour a lap of gold into the very grave where the hero of the story is hiding—PJ 64-67).
Further evidence that ki does not highlight the speech that it introduces is that, when it precedes a reported question, the ANSWER (which is not introduced with ki) is more important than the question. Such is the case in (12).9
(12) question gušt=i ki piramard baba e ce=(w)ant
say.PST.3s=PC.3s SUB old.man father DEM what=COP.PRS.3p
ANSWER gušt=i e bexi wan-en ciz=ant man=um wart-a
say.PST.3s=PC.3s DEM entirely good-ATTR thing=COP.PRS.3p I=also eat.PST-PSTP
e ar-a=um dat-a=un=o šuma=um bor-it
DEM donkey-O=also give.PST-PP=COP.PRS.1s=& you=also SBJ.eat.PRS-2p
He said, “Dear old man, what are these?” He said, “These are very good things. I’ve eaten them
and also given them to this donkey. You should eat them, too.” (XM 93-99)
A similar pattern is sometimes found when a reported “proposal” is followed by “its non-speech execution” (ibid. 111). The effect of introducing the proposal with ki is to background it in relation to its non-speech EXECUTION. This is seen in (13). 10
(13) proposal gušt=i ki b-ra…
say.PST=PC.3s SUB SBJ-go.PRS
God said: ‘Go (and tell that poor fellow…)’
EXECUTION His Holiness Moses came and gave God’s message to the fellow… (MG 14-24)
Cross-linguistically, indicators of interpretive use often introduce indirect speech (ibid. 116), as it frequently communicates only the gist of the original communication (ibid. 106), and indirect reporting is associated with back grounding (Lowe and Hurlimann 2002:75). Reported speeches in Balochi that are introduced with ki are not classified as indirect because they retain first and second person pronominal references. Nevertheless, they tend to behave like indirect speeches as far as grounding is concerned. Given that one cross-linguistic way of highlighting an event is to background the one that immediately precedes it (Levinsohn 2011a:79), I conclude that, if the effect of using ki to introduce a reported question is to thereby highlight the answer, this is consistent with it being an indicator of interpretive use.
4. Further Evidence that Ki is an Indicator of Interpretive Use
One reason for using an interpretive use marker is to signal that what follows relates back to and interprets something in the immediate context. So, for example, when a demonstrative is used cataphorically to point forward to and highlight what follows, it is cross-linguistically normal to introduce what follows with such a marker. Such is the case in (14): ‘that which I ate at once’
relates back to and interprets ameš ‘this’.
(14) mni rozi am=eš=int ki man yakk war-a wart-un
my ration EMP=DEM=COP.PRS.3s SUB I once time-ADVZ eat.PST-1s
my ration is this: that which I ate at one go. (MG 67-68)11
Similarly, in (15), the material following ki relates back to and interprets allaay pay_aman ‘God’s
(15)a alla-ay pay_aman-a pa bandag-a dat
God-GEN messages-O for servant-OBL give.PST.3s
b ki ay bandag ti rabb e rang gušt
SUB oh servant your God DEM way say.PST.3s
he gave God’s message to the fellow: ‘Oh fellow, your God said like this…’ (MG 24-25)
See also KH 58-59, in which the material following ki (‘I swear by God…’) relates back to and interprets kasam wa ‘took an oath’.12
Extract (16) indicates the most significant points in the second half of a folktale about a camel that has become so exhausted from neglect and ill-treatment that it ‘laid down its neck to die’ (BU 13-14). What is noteworthy is that ki introduces the culminating speech by the camel (16b), which is consistent with Roberts’ (2009:300) claim that, in Persian, the corresponding particle “ke is used primarily in spoken texts to give prominence to speeches that the author considers are important to the story”. However, it is clear from the content of the camel’s speech that it is a direct response to its owner’s request for forgiveness (16a). As such, the presence of ki can be understood as an overt indication that the reply relates to and interprets the owner’s request (see further below).13
(16)a The Baloch nomad told the camel, ‘I have come to my senses now and now I know that [ki] I have been unjust to you. … I want that [ki] you forgive me before you die, forgive me and do not take my negligence into consideration…’ (BU 27-49) b Then this camel, by the order of God the Almighty, turned its face and, in the very agony of death, began to speak and said to its owner ki ‘It doesn’t matter. … If you have loaded me up with heavy loads, I will also forgive you. … I will forgive whatever you have done to me. But I will not forgive one thing, I will not forgive that until doomsday; that is this, that [ki] you didn’t understand anything of my lawful and clean flesh… But one deed of yours I will not forgive, that [ki] … you tied my rein to the tail of a crop-tailed donkey and made the donkey my leader and made the donkey my way-guide. This deed I will not forgive.’ (BU 50-105)
1 A shorter version of this paper was presented at the 4th International Conference on Iranian Linguistics (ICIL 4), Uppsala University, Sweden in June 2011.
2 Abbreviations used are as follows (adapted from Delforooz 2010:15-16): ADVZ adverbializer; ATTR attributive form of an adjective; COP predicate copula; DEM demonstrative; EMP emphatic particle; GEN genitive; IMPF imperfective; IND indefinite; IZ izafa construction; NEG negative; NP noun phrase; O object; OBL oblique; PC pronominal clitic; PP past participle; PROH prohibitive; PRS present; PST past; RFL reflexive; SBJ subjunctive;
SUB conjunction of general subordination (ki); 1s/3s 1st/3rd person singular; 1p/2p/3p 1st/2nd/3rd person plural; & associative connective.
3 See XM 74-75 (reason), MG 61-62 (result) and KH 111 ((8b) below) (purpose). The references are to texts found in Appendix 2 of Delforooz 2011 (pp. 287-392). To try and ensure that the paper covers all the major uses of ki, every instance of ki in texts XM (pp. 288-295) and MG (pp. 296-304) has been cited.
4 See also XM 7, 31, 36, 50, 55, 92, 106; MG 9, 45 (reason). See BP 106-107 for a complement following zant ‘understand, know’ that is not introduced with ki.
5 See also XM 6, MG 90.
6 This ki may be a “proclitic” (Hosseini 2011) that indicates contrastive emphasis.
7 See also MG 107, 114. In MG 72 (man šapi ki mír-in [I tonight SUB die.PRS-1s] ‘(It is sure) that I will die
tonight’), ki appears to introduce a complement (mír-in), even though no orienter is present.
8 Ki also tends to be used to introduce a message that the addressee is asked to pass on to someone else. This is probably because, at this stage, it is hypothetical. See MG 5-6.
9 In the Xarmizza (‘Melon’) folktale (XM), which concerns the discovery that melons are good to eat, three reported questions are introduced with ki, all of them asked by the king. The first leads to the discovery that a dragon has come to ask for help (12-16—see footnote 7) and the second to the carpenter receiving a strange seed as a reward for helping the dragon (29-54). After someone sows the seeds, feeds the fruit to his donkey and, eventually, tries them himself (68-91), the third of the king’s questions ((12) above) leads to him being persuaded to try the fruit, which results in it being called Xar-mizza (‘donkey-tasted’) (95-111—see (1) above).
10 See also MG 29-38. In XM 12-13 (badša dem dat yakk=e=ra ki e ci=(y)e [king face give.PST.3s one=IND=O SUB DEM what=IND] ‘The king sent someone (to check) who it is’), no speech verb is used to introduce the proposal, but the presence of ki again backgrounds this event in relation to its EXECUTION (14-16).
11 See also MG 77-78. In BU 55-56, the material following ki relates back to and interprets anco ‘such’. PJ 3-8 (with ancen sawt… ‘such a voice’) is similar.
12 See also BP 21-22 (interpreting so kan ‘pose a question’). In (11) above, it could be argued that the material following ki relates back to and interprets irada kurt- ‘made a decision’.
13 The following observation about the Koiné Greek interpretive use marker ___ applies equally to the way ki is used in (16b): “When ___ introduces direct speech in Luke-Acts, it not only indicates that the speech concerned “interprets” what has already been said …; it also seems to mark that speech as the culmination of a narrative unit or subunit” (Levinsohn 2011b:266).
14 Compare the use of the interpretive use marker kî in Biblical Hebrew to introduce a subordinate clause of time (e.g. in ‘When you go’—Exodus 3:21c). Follingstad (2002:152) states, “Throughout its uses, kî has this ‘mention’ [interpretive use] function”. However, he goes on to describe its use in general in terms of Mental Space Theory.
Blass, Regina. 1990. Relevance Relations in Discourse: A Study with Special Reference to Sissala. Cambridge: Cambridge University Press.
Comrie, Bernard. 1989. Language Universals and Linguistic Typology. 2nd. edition. Chicago: University of Chicago Press.
Delforooz Barjasteh, Behrooz. 2010. Discourse Features in Balochi of Sistan (Oral Narratives). (Studia Iranica Upsaliensia 15). Uppsala: Uppsala University.
———. 2011. Discourse Features in Balochi of Sistan (Oral Narratives). (Studia Iranica Upsaliensia 15). Online at
Follingstad, Carl M. 2002. Deictic Viewpoint in Biblical Hebrew Text: A Syntagmatic and Paradigmatic Analysis of the Particle Ki. Dallas: SIL International.
Hosseini, Ayat. 2011. Prosodization of Function Words in Persian. Paper presented at the 4th International
Conference on Iranian Linguistics (ICIL 4), Uppsala University, Sweden; June 2011.
Levinsohn, Stephen H. 2010. Some Notes on the Information Structure and Discourse Features of Exodus 1.1-12.15.
Online at http://www.sil.org/~levinsohns.
———. 2011a. Self-instruction Materials on Narrative Discourse Analysis. Online at
———. 2011b. Discourse Features of New Testament Greek: A Coursebook on the Information Structure of New Testament Greek, 2nd. edition (2000), revised. Dallas: SIL International.
Lowe, Ivan, and Ruth Hurlimann. 2002. Direct and indirect speech in Cerma narrative. In Reported Discourse: A
Meeting Ground for Different Linguistic Domains, edited by Tom Güldemann and Manfred von Roncador,
71-90. Amsterdam & Philadelphia: John Benjamins.
Roberts, John R. 2009. A Study of Persian Discourse Structure. (Studia Iranica Upsaliensia 12). Uppsala: Uppsala University.
Levinsohn: Introducing Reported Speeches in Balochi of Sistan with ki p.8
Sperber, Dan, and Deirdre Wilson. 1995. Relevance: Communication & Cognition, second edition. Oxford: Blackwell.
Whaley, Lindsay J. 1997. Introduction to Typology: The Unity and Diversity of Language. London: SAGE