Discourse Studies

Discourse Studies http://dis.sagepub.com

Stance-taking in Hebrew casual conversation via be'emet (`really, actually, indeed', lit. `in truth') Yael Maschler and Roi Estlein Discourse Studies 2008; 10; 283 DOI: 10.1177/1461445608090222 The online version of this article can be found at: http://dis.sagepub.com/cgi/content/abstract/10/3/283

Published by: http://www.sagepublications.com

Additional services and information for Discourse Studies can be found at: Email Alerts: http://dis.sagepub.com/cgi/alerts Subscriptions: http://dis.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from http://dis.sagepub.com at University of Haifa Library on May 31, 2008 © 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

A R T I C L EMaschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 283

Stance-taking in Hebrew casual conversation via be'emet (‘really, actually, indeed’, lit. ‘in truth’)

YAEL MASCHLER AND ROI ESTLEIN

Discourse Studies Copyright © 2008 SAGE Publications (Los Angeles, London, New Delhi and Singapore) www.sagepublications.com Vol 10(3): 283–316 10.1177/1461445608090222

U N I V E R S I T Y O F H A I FA , I S R A E L

ABSTRACT In this article, we investigate the functional itinerary followed by Hebrew be'emet (‘really, actually, indeed’, lit. ‘in truth’), through a close exploration of its synchronic uses in the contemporary spoken language. Since this utterance, derived from the noun 'emet (‘truth’), is so profoundly tied in with the speaker’s beliefs and attitudes towards his or her discourse, we consider issues of metalanguage, modality, evidentiality, and stance. Be'emet is traditionally classified as ‘adverb’, but in our corpus of naturally occurring Hebrew conversation, only 22 percent of all tokens function in this role. Whereas these tokens function referentially, the great majority of tokens (70%) function in the interpersonal realm of discourse, manifesting properties of discourse markers: 44.5 percent of all tokens function as full-fledged discourse markers (both semantically and structurally, Maschler, 1998) serving mirative (DeLancey, 2001), reprimanding, or negating any doubt roles; 22.5 percent function to ratify a previous stance; and three percent function to latch onto a new topic, requesting its elaboration. An intermediate category (8%) is composed of tokens functioning both referentially and interpersonally, mid-way between an adverb and a discourse marker, in a way that is particularly illuminating for understanding the changes undergone by be'emet as a result of its employment in discourse. The study thus lends support to previous studies of subjectification and intersubjectification in the process of grammaticization of discourse markers (Traugott, 2003; Traugott and Dasher, 2002).

adverbs, evidentiality, grammaticization, Hebrew discourse markers, (inter)subjectification, metalanguage, stance

KEY WORDS:

1. Introduction The Hebrew word be'emet (lit. ‘in truth’) is comprised of the noun 'emet (‘truth’) preceded by the preposition be- (‘in’), a common way of deriving adverbs from nouns in Hebrew (cf. bekitsur ‘anyway’, lit. ‘in shortening’; Maschler, in press b).


284 Discourse Studies 10(3)

Two meanings are provided for be'emet in Even Shoshan’s 1986 dictionary. The first is indeed adverbial and is found already in Biblical Hebrew: ‘honestly, in trustworthiness, wholeheartedly’. The second comes from a later period of the language (Talmudic period and the Dead Sea Scrolls): ‘but, it is true that ('omnam), in contrast to what was said before’. We adopt Biber and Finegan’s definition of stance as ‘the overt expression of [. . .] attitudes, feelings, judgments, or commitment concerning the message’ (1988: 1). In both usages of be'emet provided by Even Shoshan, some degree of stance the speaker is taking towards his or her talk can be discerned. Even Shoshan classifies be'emet as an adverb, regardless of which of the two meanings are concerned. However, a brief look at the second meaning reveals that it is more of a discourse marker than an adverb. This diachronic information alone provides support for recent grammaticization studies of discourse markers (e.g. Maschler, in press b; Traugott, 1995a, 1995b, 2003; Traugott and Dasher, 2002) on the development of discourse markers from adverbs. Traugott and Dasher argue that over time, these words become first more subjective and metatextual, and finally more intersubjective. Their scope is initially within the clause, and later over larger discourse segments. This process is paralleled by the movement of the adverb from within the clause to a syntactically marginal position with respect to the clause. In SVO languages, such as English and Hebrew (Ravid, 1977), this implies initial position, whereas in SOV languages, such as Japanese, the movement is to final position (Traugott and Dasher, 2002). In this article, we investigate the various uses of be'emet in casual Israeli Hebrew conversation, not all of which are covered by Even Shoshan’s definitions. Our goal is to understand the functional itinerary1 this particular word has followed in Hebrew through a close exploration of its synchronic uses in the contemporary spoken language. We relate to position not so much within the clause, but rather within the basic unit of spoken language – the intonation unit (Chafe, 1994). Since this particular adverb, derived from the noun 'emet (‘truth’), is so profoundly tied in with the speaker’s beliefs and attitudes towards his or her discourse, our journey takes us to consider issues of metalanguage, modality, evidentiality and stance. Since so much in this article depends on the distinction between adverb and discourse marker, it is necessary to clarify what is meant by these two categories from the outset. 1.1 ADVERBS Ramat and Ricca have shown that the category of adverb is a scalar, or even radial one, ‘with prototypical instantiations and less typical or even marginal ones’ (1994: 289). In one of two approaches they present, based mostly on the study of English -ly adverbs but taking into account adverbs in many other languages (Latin, Italian, French, Spanish, German, Dutch, Norwegian, Irish, Greek, Modern Greek, Sanskrit, Armenian, Lithuanian, Latvian, Russian, Polish, Rumanian, Estonian, Turkish), prototypical adverbs are those identified


Maschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 285

as manner predicate adverbs (e.g. quickly). They share a functional feature (Verb Phrase modifier) and a semantic feature (Manner). They ‘can be viewed as ‘‘verb or verb-phrase modifiers’’, much like adjectives function as ‘‘noun and noun-phrase modifiers”’ (Ramat and Ricca, 1994: 307). Ramat and Ricca’s characterization of prototypical adverbs in parallel to adjectives is reflected in Hebrew grammatical terminology: to'ar hashem (‘modifier of the noun’, i.e. ‘adjective’) versus to'ar hapo'al (‘modifier of the verb’, i.e. ‘adverb’). Next in their classification come the degree adverbs (very, extremely), modifying a modifier (i.e. an adjective or another adverb). Then come the sentence adverbs (frequently, strangely, frankly, probably), whose semantic scope embraces some aspect of the entire utterance, many times expressing some attitude of the speaker. Even further away from the center are ‘setting’ adverbs of space and time, such as today, now, here, recently. Next, focalizers such as only, also, even, exclusively, which are highly deviant since they typically modify noun phrases rather than verb phrases or sentences. At the very margin of Ramat and Ricca’s classification are what they call ‘text adverbs’ ‘(or ‘‘conjuncts’’, ‘‘connectives’’, and so on) like firstly, consequently, nevertheless’ (p. 308), pertaining to the level above the sentence. In many languages, their syntactic behavior clearly distinguishes them from conjunctions. In the present study, these ‘text adverbs’ are called ‘discourse markers’ and semantic as well as structural (prosodic and turn-taking) criteria are used for determining whether an utterance constitutes a discourse marker (see later). In the case of be'emet, we will see a situation of pragmatic polysemy, such that the same lexeme is characterized in some contexts as Ramat and Ricca’s prototypical adverbial category (manner predicate adverb), in others as sentence adverb, and in others yet as the most marginal category in their classification; namely, ‘text adverb’ (discourse marker). Ramat and Ricca do not relate to Hebrew (or to any other Semitic language) among the many languages considered in their study. In an article entitled ‘On Verbal Structure and the Hebrew Verb’, Goldenberg (1998) argues that the verb is not a single unit but rather a complex of three components: person marker (the obligatory subject morpheme bound to every Hebrew finite verb), attribute (the lexeme signifying the predicate), and predicative nexus (the syntactic link between the subject and the lexeme signifying the predicate, which creates the sentence). Hebrew adverbial complements apply explicitly to a certain member of the above three components comprising the predicative complex. The circumstantial qualifier ('ani mele'a halaxti [‘I went away full’, Ruth i 21]) applies to the person within the predicative complex. Qualitative adverbs apply to the verbal lexeme and can usually also be expressed as adjectives qualifying an inner object (vayexerad libo me'od [‘and his heart trembled greatly’, 1 Sam. xxviii 5]), cf. vayexerad yitsxak xarada gedola 'ad me'od (‘and Isaac trembled a very great tremble’, Genesis xxvii 33) (p. 186). Finally, ‘it is the ‘‘circumstantial adverbs’’ (adverbs of time, place, cause, purpose, comparison, etc.) that function as complements of the nexus. These refer only to the nexus and cannot, by definition, have nominal referents’ (p. 186). Considering the oldest usage of be'emet attested in Even Shoshan (1986), Biblical Hebrew ‘im be'emet uvetamim



'asitem (‘if you acted honestly and wholeheartedly’, Judges ix 162), we see that this token of be'emet constitutes one of the circumstantial adverbs and refers to the nexus, describing the manner in which the action was accomplished (‘honestly’). This, then, is also Ramat and Ricca’s prototypical adverb – the manner predicate adverb. 1.2 DISCOURSE MARKERS There have been many approaches to discourse markers over the past two decades, Schiffrin (1987) being the most notable one.3 The approach taken here follows Maschler (1994, 1998, 2002, in press b) and differs from other approaches in two ways. First, it anchors the definition of discourse markers to the process of metalanguaging involved in their employment. Discourse markers are viewed here as linguistic elements employed for metalanguaging – languaging (Becker, 1988) about the interaction, as opposed to languaging about the extralingual world. In other words, rather than referring to the world perceived by speakers (through language) to exist beyond language, discourse markers refer to the text itself, to the interaction among its speakers, or to the cognitive processes taking place in their minds during verbalization. (Maschler, in press b)

The second way in which the present approach to discourse markers differs from previous approaches is that it supplies an operational definition of discourse markers. This definition is based, among other things, on the concept of the intonation unit and on prosody in general. 1.3 DEFINITION OF DISCOURSE MARKERS In order to qualify as a discourse marker, the utterance in question must fulfill both a semantic and a structural criterion: a) Semantically, the utterance must have a metalingual interpretation in the context in which it occurs. In other words, rather than referring to the extralingual world, it must refer metalingually to the realm of the text (‘textual discourse marker’, Maschler, 2002), to the interaction between its participants (‘interpersonal discourse marker’, Maschler, 1994), or to their cognitive processes (‘cognitive discourse marker’). b) Structurally, the utterance must occur at intonation-unit initial position, either at a point of speaker change, or, in same-speaker talk, immediately following any intonation contour other than continuing intonation. It may occur after continuing intonation or at non intonation-unit initial position only if it follows another marker in a cluster (Maschler, 1998). These two criteria have been found to coincide for 94 percent of the discourse markers throughout the corpus (Maschler, 1998). The discourse markers comprising the remaining 6 percent category fulfill only the semantic criterion above (i.e. metalinguality) and share certain functional features. In particular, they construct a boundary between conversational actions (Ford and Thompson, 1996) which is of a lower level in comparison to the boundaries constructed by discourse markers fulfilling both criteria (Maschler, 2002).



We will see that the tokens of be'emet in the present corpus vary in the degree to which they fulfill these two criteria. Accordingly, they vary in the degree to which they can be considered discourse markers. This variability can be employed in order to conceptualize also the category ‘discourse marker’ as a scalar category, with those members fulfilling both criteria constituting the prototypical instantiation of the discourse marker category.

2. Data and quantitative overview The study is based on the Haifa Corpus of Spoken Israeli Hebrew (Maschler, 2004), comprising, at the time of this study, 91 conversations between friends and relatives, constituting approximately 270 minutes of talk among 223 different speakers. The corpus was collected over the years 1994–2004 and transcribed following Chafe (1994). Altogether, 63 tokens of be'emet were employed. The initial classification here is semantic, according to whether the particular be'emet token relates to some action or situation in the extralingual world (‘referential’ be'emet, 14 tokens [22%]), or whether it relates to the world of the interaction between discourse participants (‘interactional’ be'emet, 44 tokens [70%]). Five tokens (8%) relate to both the extralingual realm and the realm of the interaction simultaneously, thus constituting a third, intermediate category. These categories will become clearer as we illustrate them in the excerpts below, but for now Table 1 presents an initial, quantitative sketch of the findings. Below we show that this semantic categorization of be'emet tokens is paralleled by structural and functional properties. We will see that referential be'emet is found within the intonation unit and functions as a predicate adverb, modifying the predicate or predicate-phrase.4 On the other hand, interactional be'emet is found intonation-unit initially, finally, or as a separate intonation unit, and manifests properties of a discourse marker. Thus, the scope of referential be'emet is limited to the predicate of the clause, whereas the scope of interactional be'emet is wider, often spanning large discourse segments.

3. Referential be'emet Altogether, 14 be'emet tokens (22%) function referentially in this corpus. We classify them into two sub-categories. 3.1 REFERRING TO AN ACTION OR SITUATION IN THE EXTRALINGUAL WORLD We begin with be'emet referring to an action or situation in the extralingual world. In Excerpt 1, Daniel, a property confiscator, tells Debbie how, as part of his day at work, he ran after a guy whose tractor it was his job to confiscate. Once the guy spotted Daniel, he began escaping on his tractor. Daniel describes himself running after the slowly-moving tractor in the muddy fields of the Arab village Tamra, in pouring rain. Finally, he catches up with the guy and confiscates his tractor-keys, but the guy attempts yet another escape:5


6 (9.5%)

8 (12.5%)

5 (8%)

14 (22%)

referring to interaction in the extralingual world

be'emet both referential & interactional

referential be'emet

1 . Functions of be'emet throughout the database

referring to action/situation in the extralingual world

TA B L E

4 (6.5%)

10 (16%) Total = 63 (100%)

uttered by interlocutor

uttered by speaker

14 (22.5%)

ratifying the stance taken in a previous utterance

2 (3%)

latching onto new topic requesting further development

28 (44.5%)

discourse marker be'emet

19 (30%)

5 (8%)

4 (6.5%)

(ad)mirative reprimanding negating any doubt

44 (70%)

interactional be'emet



Maschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 289 Excerpt 1 (‘Tractor Chase’): 73

Daniel:

74 75 76

77 78

Debbie:

79 80

Daniel:

81

Debbie:

82

Daniel:

83 84 85

86

87

88 89 90 91

92

.. hahu motsi li mavreg, that one takes out to me a screwdriver the guy gets a screwdriver out in front of me, .. maxnis letox haswitch, puts [it] in the switch [of the tractor], .. mesovev, turns [it around], hatraktor matxil la'avod. the tractor begins working the tractor starts. ke'ilu, like, ken? yeah? (laughter) lema tsexim mafteax? {laughing} for what is needed a key who needs a key? nu. go on. nikrati mitsxok, I was torn up by laughter I laughed like hell, lo yadati ma la'asot, I didn’t know what to do, ke'ilu hu matxil livroax li 'od pam. like he begins escaping from me again. ... ra'iti she'harbe brerot 'ein li, I saw that many choices are not to me I realized I didn’t have much of a choice, ... lamrot she'ze 'asur li la'asot, even though it is forbidden to me to do, although I’m not allowed to do, ma she'asiti, what that I did what I did, ... ke'ilu, like, .. 'od pam ratsti 'axarav. again I ran after him. .. hotseti sakin, I got out a knife, ... pincharti lo pashut ta'tsmigim. I punctured to him simply the tires I simply punctured his tires. .. ve'az hu be'emet nish'ar taku'a sham. and then he really remained stuck there.



Unlike the previous time, after which the guy began escaping on his tractor, this time Daniel made sure it wouldn’t happen again, by puncturing the tractor’s tires with his knife (lines 90–1). Daniel reports in line 92: ve'az hu be'emet nish'ar taku'a sham (‘and then he really remained stuck there’). In other words, unlike the previous time, in which the guy was somewhat successful, this time he did not succeed in escaping and was be'emet (‘really’) stuck in the muddy fields. This, then, is the manner adverb be'emet, translated here as ‘really’, functioning referentially by modifying the action nish'ar taku'a (‘remained stuck’) in the extralingual world. However, the extralingual world is never completely detached from its observer. The action nish'ar taku'a (‘remained stuck’) is described by some speaker, who also conveys his own stance towards the action via be'emet, by alluding to a comparison with a previous, more successful attempt to escape.6 Structurally, we see that this token of be'emet occurs in medial position within the intonation unit and its scope is the predicate of the clause, as is the case with Hebrew circumstantial adverbs, which, as we have seen (Goldenberg, 1998), refer to the nexus. Here be'emet precedes the verb it modifies, although in Hebrew the adverb could also follow the verb (and indeed does in about half the instances here). However, even in this manner predicate adverbial use of be'emet, we find a trace of a stance-taking function pertaining to a larger discourse segment; namely, the speaker’s alluding to a comparison with a previous attempt to escape. This stance-taking function will become considerably more pronounced in the interactional uses of be'emet. As seen in Table 1, in this corpus, seven be'emet tokens (11%) function in this referential category. 3.2 REFERRING TO AN INTERACTION IN THE EXTRALINGUAL WORLD Six more be'emet tokens (9.5%) function referentially, but refer to an interaction that occurred in the narrated world. For example, in Excerpt 2, a conversation between a couple in their early 20s, Hadas relates an incident from second grade, in which her teacher had sent her to the principal for what the former interpreted as impertinence: Excerpt 2 (‘I’m Not a Porter’) 148 Hadas: 149 150 151 152 153 Idan: 154 Hadas: 155

.... vehi 'amra li, and she told me, .. telxi lamenahelet, go to the principal, .. tagidi la, tell her, ma 'amart. what you had said. .... ve'ani hitxalti livkot. and I started crying. ... 'oy [tsk] [tsk] [tsk]. .. vehalaxti lamenahelet, and I went to the principal, .. veyashavti sham, and I sat there,


Maschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 291 156

157 158

159

172

173

174

175

176 Idan:

177 Hadas: 178

179 Idan: 180 Hadas: 181 182

183

.. velo yaxolti ledaber. and not I could speak and I couldn’t speak. ratsiti leha I wanted to 'aval lo yaxolti ledaber, but not I could speak but I couldn’t speak, pashut hitxalti livkot. simply I started crying I simply started crying. {12 intervening intonation units} .... ve-- 'ex shehi ra'ata 'oti, and how that she looked at me and just as she was looking at me, .. kaxa boxa, so crying crying so, .. hi nora nilxatsa venivhala. she terribly got pressed and scared she became terribly alarmed and scared. ... ve'az bidyuk higi'a hamora. and then exactly arrived the teacher and exactly then the teacher arrived. ... laxeder shel hamenahelet? to the room of the principal at the principal’s office? .. ken. yeah. ... lir'ot ‘im be'emet halaxti. to see whether actually I went. to see whether I had actually gone [to the principal]. .. nu? go on? .. ve'az 'eh, and then uh, .. hi 'amra, she [the teacher] said, .. 'at shamat, you heard did you hear, ma hi 'amra li? what she said to me?

Following a long description of what happened to her in the principal’s office, Hadas recounts how at some point, the teacher who had sent her there showed up (line 175). Idan requests a clarification of where she had shown up (laxeder shel hamenahelet?, ‘at the principal’s office?’, line 176), and Hadas responds affirmatively (ken, ‘yeah’, line 177) in sentence final intonation. She then adds the increment (Schegloff, 1996) lir'ot 'im be'emet halaxti (‘to see whether I had



actually gone.’). This is the manner adverb be'emet, translated here as ‘actually’, functioning referentially and modifying the action halaxti (‘I had gone’) in the extralingual world. In other words, following the teacher’s order telxi lamenahelet, (‘go to the principal’, line 149), Hadas had two options: she could have either obeyed and gone to the principal’s office, or she could have disobeyed. The teacher had apparently gone to check on her and verify whether or not she had indeed gone. As we saw in Excerpt 1, many such instances of adverbial be'emet are employed against the backdrop of two contrasting options. Following this short verification + response sequence, uttering nu 7 (line 179), Idan encourages Hadas to continue her narrative (line 179), which she indeed does beginning at line 180. We see that this token of be'emet functions in relation to a stance taken by one of the participants in the storyworld; namely, ‘the teacher’s order telxi lamenahelet’ (‘go to the principal’). Since the interaction between these participants occurred in the storyworld, rather than in the world in which the interaction between Hadas and Idan took place, we classify it as referential be'emet. The sub-classification reflects the fact that we are dealing with an interaction in this storyworld. Thus, speakers can employ be'emet in order to ratify a stance taken by some participant depicted in the discourse.

4. Interactional be'emet Far more tokens – 44 (70%) – function in the realm of the interaction among participants. They are employed to express amazement ([ad]mirativity), reprimand, to negate any possible doubt, in order to ratify the validity of a participant’s previous utterance, or in order to latch onto an interlocutor’s mention of a new topic and pursue its further development. In that all tokens in this category refer to the interaction between discourse participants (and in the case of the latter category – also to the realm of the text), they all function metalingually and thus fulfill the semantic requirement in the definition of discourse markers above (section 1.3). In what follows, we shall see that the majority of these tokens fulfill the structural requirement as well. Those that do not nevertheless fulfill the structural requirement of occurring intonation-unit marginally (however they occur finally, as opposed to initially). 4.1 (AD)MIRATIVE BE'EMET ‘The term ‘‘mirativity’’ refers to the linguistic marking of an utterance as conveying information which is new or unexpected to the speaker’ (DeLancey, 2001: 369–70). This is the most common function of be'emet, with 19 tokens – 30 percent – functioning miratively. Examine, for instance, Excerpt 3, in which Orna is telling her sister how her husband-to-be proposed marriage: Excerpt 3 (‘Marriage Proposal’): 75

Orna:

.... ve--'é--h, and u–h,



.... ve'àz beshalav mesuyám, and then at stage specific and then at a certain point, .. hu pit'òm 'omer li kazé, he suddenly says to me like, .. 'é--h, {change of tone} u–h, .... titxatní 'itì? you will marry me will you marry me? ... 'ani 'omeret lo, I say to him, ma? what? ... 'az hu 'omer li, so he says to me, .. 'ani rotse shetitxatni 'iti

77 78

79

80 81 82 83

84

Alona:

85

Orna:

86 87 88 89 90

91

bevakasha. {–laughing–} I want that you marry with me please I would like you to marry me please. be'emét? really? ken! yeah! ... 'az 'amarti lo, so I said to him, .. ma z'tomeret, what d’you mean, ... ta'rotse lehitxaten 'ití? you wanna marry me? .. 'ata kolet? [are] you comprehending? .. 'ata .. ha'oznayim shelxa shom'ot, you .. the ears yours hear you .. are your ears hearing, 'et ma shepixa 'omer? what that your mouth says what your mouth is saying?

By means of change in tone, constructed dialogue (Tannen, 1989) and the conveying of suddenness, the marriage proposal (lines 78–9) is depicted here in a highly involving fashion and constructed as occurring ‘out of the blue’ to the recipient. Orna tells of her surprise and disbelief in response to the proposal: 'ani 'omeret lo, ma? (‘I say to him, what?’, lines 80–1). (Her surprise is further elaborated in lines 86–91.) In response, her husband-to-be repeats the proposal, somewhat exaggerating its explicitness and politeness: 'ani rotse shetitxatni 'iti bevakasha (‘I would like you to marry me please’). Orna’s laughter towards the end of this intonation unit supports the ‘exaggerated’ nature of this explication.



Perhaps mirroring Orna’s surprise, comes her sister’s response to this telling: be'emet?, translated here as ‘really?’, line 84, in rising question intonation. In response to her sister’s telling, Alona constructs here a stance of amazement and disbelief. In other words, she marks her utterance as a response formulated following the receipt of information which is new or unexpected to her. It is in this sense that we term this usage ‘mirative’, perhaps somewhat differently than DeLancey (2001), who employs the term to refer to the conveying of information which is new or unexpected to the speaker. The term ‘(ad)mirative’ (Friedman, 1986) may thus be better suited, particularly as his approach involves the concept of ‘truth’: ‘The admirative is traditionally defined as a mood expressing surprise [. . .] Surprise results from a past state during which the speaker would not have accepted the truth of some subsequently discovered fact or event’ (1986: 180). Of course, we know that the sister isn’t all that surprised, particularly since the recording begins 75 intonation units earlier with her requesting Orna to tell her how exactly the marriage proposal transpired. This be'emet?, then, is a direct response to the involvement creating strategies employed immediately preceding it by the interlocutor, and it occurs precisely at the climax of the story. It is employed to construct a stance of amazement and disbelief, even if the information conveyed isn’t all that unexpected to the recipient. As is often the case, the participant who provided the ‘surprising’ information answers ken (‘yeah’, line 85), and continues telling the story. This is be'emet functioning in the interactional realm rather than in the extralingual one. Thus, it fulfills the semantic requirement for discourse markerhood. Structurally, it occurs intonation-unit initially (constituting an entire intonation unit, in fact) at a point of speaker change, thus fulfilling the second requirement for discourse markerhood as well. 4.2 REPRIMANDING BE'EMET We have seen that (ad)mirative be'emet occupies the entire intonation unit. Two more sub-categories of be'emet share this property; however, they differ in their intonation contours. While (ad)mirative be'emet ends in sentence-final rising intonation, both reprimanding be'emet (five tokens, 8%) and be'emet negating any doubt (four tokens, 6.5%) end with sentence-final falling exclamatory intonations. We begin with reprimanding be'emet. In Excerpt 4, a family conversation, Shmulik is about to begin a story: Excerpt 4 (‘Pimples’): 10

Shmulik:

11

12 13

Merav: Mother:

'a.. 'amartem ptsa'im? y.. you said pimples? .. yesh li sipur yafe 'al ptsa'im. there is to me story nice about pimples I’ve got a nice story about pimples. (laughter) 'at maklita? you[’re] recording?



Shmulik:

15

16

Mother:

17 18

Noga: Shmulik:

19 20 21 22 23

24

Merav:

yesh li sipur there is to me story I’ve got a story yesh li sipur yafe 'al ptsa'im. there is to me story nice about pimples I’ve got a nice story about pimples. 'at maklita? you [’re] recording? shshsh. lama 'at horeset? why you ruining why are you ruining [it]? .. 'axrei ze afterwards .. 'axrei ze keshemaklitim, afterwards when one records, 'az shom'im 'otax sho'elet. then one hears you asking. .. be'emet! really! .. lo kol hazman 'ani maklita. not all the time I’m recording I’m not recording all the time. .. midei pa'am. [only] occasionally.

This family interaction begins humorously with the youngest sister, Noga, asking her older brother Shmulik to tell his story about pimples. Shmulik begins at line 11. The mother interrupts twice (lines 13, 16), asking Merav whether she is recording. At this point, Shmulik reprimands the mother with lama 'at horeset? (‘why are you ruining [the recording]?’, line 18). He further explains that her questions ruin the naturalness of the recording (lines 19–21), and concludes with be'emet! (line 22) in a reprimanding tone. That Merav interprets Shmulik’s utterance as a reprimand is evidenced by her positioning herself in justification of her mother’s questions with lo kol hazman 'ani maklita. midei pa'am. (‘I’m not recording all the time. Only occasionally’, lines 23–4), particularly in light of the fact that we know she was in fact recording these past few seconds. Through reprimanding be'emet, the speaker signals to the hearer that the common ground and attitude he thought they shared were being fractured by the latter. Reprimanding be'emet – in a way that reminds us of (ad)mirative be'emet – reveals that what has been said contradicts the speaker’s expectations or assumptions. Adopting Chafe’s (1986) broad definition of evidentiality as the linguistic marking of attitudes towards knowledge, Downing, in her study of English surely, contemplates evidentials as ‘interactive devices or resources for redefining common ground between interlocutors’ (2001: 251). According to Haviland, evidentials encode for the speaker ‘what an addressee can be taken to know, or should know, or apparently (perhaps culpably) fails to know’ (1987: 343, emphasis in the original). From Shmulik’s point of view, his



mother failed to know that she must not reveal the fact that the conversation wasn’t all that ‘natural’, and her questions were therefore uncalled for. His reprimanding be'emet aids in constructing this stance. Reprimanding be'emet thus has a moral character, which, according to Haviland, is typical of evidentials. The reprimanding function of be'emet is magnified when it appears in a cluster together with the discourse marker nu (Maschler, 2003): nu be'emet! It is often heard in causal Israeli Hebrew conversation, sometimes in a sarcastic key. However, the only instance of this cluster found in our corpus is unfortunately not clear enough for illustration. 4.3 BE'EMET NEGATING ANY DOUBT Another use of be'emet as a discourse marker is illustrated in the following excerpt. Anat begins a story about an 85-year-old Polish survivor, somewhat senile, who went on a trip back to Poland with his family. During the visit to Auschwitz, where he had been imprisoned about 60 years earlier, he disappeared. It turned out, that he had become tired of waiting there for his sons, who were walking through the place rather slowly, and decided to go back – by means of three trains and a taxi – to what he thought was his home, the little Polish village he was originally from. Anat recounts an amazing story of what they did in order to find him. Her story opens with: Excerpt 5 (‘Trip to Poland’): 2

Anat:

3

4 5

6

Dana:

7 8

Anat:

9 10

..'ani 'asaper 'et hasipur, I will tell the story, ... ve'ani xoshevet shehu be'emet sipur mat'im and I think that it really story fit and I think that it’s really a story fit shehu mat'im lihiyot, that it[’s] fit to be, ... seret katsa--r, film short a short film, .. seret katsar. film short a short film. (laughter) ... be'emet! really! ... 'ani hayiti 'osa mimeno, I would have made out of it, .. seret katsar. film short a short film.

Anat announces that she believes the story she is about to tell is fit to become a short film (lines 3–5). Her interlocutor, Dana, echoes her with seret katsar (‘a short film’, line 6), accompanied by laughter (line 7). This humorous echoing by the interlocutor casts some doubt over the fact that the following



story is fit to become a short film. Anat responds to this humorous doubt with be'emet!, translated here as ‘really!’, negating any doubt that this is indeed the case. A paraphrase would be: 'ani be'emet mitkavenet laze, ‘I really mean it’. This interpretation is supported by her following utterance: 'ani hayiti 'osa mimeno, seret katsar. (‘I would have made out of it, a short film’), which strengthens her position by claiming that not only is the story fit to become a short film, she would have made a short film out of it. It is also a strengthening by repetition of the stance expressed by the referential be'emet token of line 3: the story is really fit (hu be'emet8 sipur mat'im ‘it’s really a story fit’) in the speaker’s mind, to become a short film (lines 3–5). 4.4 BE'EMET RATIFYING THE STANCE TAKEN IN A PREVIOUS UTTERANCE Another major sub-category of be'emet, with 14 tokens (22.5%) throughout the database, involves ratification of stance. In most of the cases here (10 out of 14, or 16% of all be'emet tokens), be'emet is employed to ratify the stance of one’s own previous utterance; in four (6.5%) of the cases, the stance of another participant is being ratified. Structurally, six out of these 14 be'emet tokens do not appear at intonation-unit initial position, unlike the case of the previous three categories (sections 4.1–4.3). However, the majority of them (five out of six tokens) still occur intonation-unit marginally at intonation-unit final position. For example, in Excerpt 6, a conversation between two young women, Miri narrates to her best friend Tirtsa a conversation which took place between herself and her partner, Leonardo: Excerpt 6 (‘Mutuality’): 1

Miri:

2 3

4 5 6 7

8 9

10

Tirtsa: Miri:

... 'aval mitsad sheni, but on side other but on the other hand, hu-he, .. be'eizeshehu shlav, at some stage at some point, pa'am /'amar li/ once said to me {coughing} ... ya, wo, .. 'eize kef lada'at, what fun to know it’s so good to know, she'at tihyi, that you’ll be, 'ima shel hayeladim sheli. mother of the children my the mother of my children. ... ve'ani, and I,


298 Discourse Studies 10(3) 11

ma ze nivhalti, what this got alarmed I became really alarmed, 'amarti lo, I told him, halo Le'onardo, hey Leonardo,

12 13

14

15

Tirtsa:

16

Miri:

17 18

19

20 21

22

23 24

25

26

.. le'at le'at, slowly slowly not so fast, 'eize kef ze lishmoa davar what fun is to hear thing it’s so good to hear something

kaze. like this like this. 'o, there,

.. 'at ro'a? you see? ... ze me'od hevix 'e it very embarrassed me it very much embarrassed me ze ma ze hilxits 'oti, it what this stressed me it put me under real pressure, be'oto rega--, at that moment, .. ha'eme--t. the truth truth is. .. vebe'emet 'amarti lo, and in truth I said to him and indeed I said to him, le'onardo! Leonardo! ... le'at le'at, slowly slowly not so fast, .. lo kavanu 'od, not fixed yet we haven’t yet decided, shum davar. anything.

Miri tells Tirtsa how Leonardo’s implied long-term commitment to her (lada'at, she'at tihyi, 'ima shel hayeladim sheli: ‘to know, that you’ll be, the mother of my children’, lines 7–9) put her under tremendous pressure (ve'ani, ma ze nivhalti: ‘and I, I became really alarmed’, lines 10–11), as she was not ready for this sort of commitment at the time. She begins to tell how she asked him to ‘slow down’ (le'at le'at: ‘not so fast’, line 14), but Tirtsa interrupts at this point with a statement from the opposite stance: 'eize kef ze lishmoa davar kaze (‘it’s so good to hear



something like this’; line 15). With 'o, 'at ro'a? (‘there, you see?’; lines 16–17), Miri begins to convey that this may be what other women feel, but to her this was actually too much pressure: ze me'od hevix 'e ze ma ze hilxits 'oti, be'oto rega-(‘it very much embarrassed me it put me under real pressure, at that moment’, lines 18–20). She concludes her response to Tirtsa with a related expression, the definite noun ha'emet (‘the truth’, line 21), underlining her previous point, in contradiction to Tirtsa’s, quite parallel to English ‘truth is’ (see note 6). Following this response, and in a way that continues her disagreement with Tirtsa, she proceeds with her story which was interrupted at line 14, continuing to describe how she indeed asked Leonardo to ‘slow down’: 'amarti lo, Le'onardo! le'at le'at, lo kavanu 'od, shum davar (‘I said to him, Leonardo! not so fast, we haven’t yet decided, anything’; lines 22–6). Miri’s actions are depicted in these lines by way of illustrating the extent of her alarm in response to his implied suggestion, describing what she proceeded to tell him as a result. This continuation of the narrative is preceded by two discourse markers: 9 vebe'emet (‘and be'emet’) – the first (ve-: ‘and’) employed to add the next conversational action (Ford and Thompson, 1996) in the least marked way (Maschler, 1998; Schiffrin, 1987), the second to ratify the stance taken in her previous, interrupted utterance. A paraphrase of this instance of be'emet could be 'kfi shebe'emet kodem 'amarti’ (‘as I was indeed saying earlier’, or simply, as translated here, ‘indeed’).10 We see that this be'emet is employed to ratify the stance of the speaker’s previous utterance. However, Haviland (1987) has taught us that evidentiality, the linguistic marking of attitudes towards knowledge, is always interactive. Similarly, Kärkkäinen (2003) has shown that stance is always constructed in interaction. One’s stance is always alive against the background of stances of other discourse participants. In ratifying her own stance, Miri also further disagrees with Tirtsa’s. Thus, be'emet is used here also to relate to her interlocutor’s stance. An example of ratifying the interlocutor’s stance is given in Excerpt 7. Two young women, Ela and Irit, are discussing the recent Seder (traditional Passover meal). Ela says her family’s wasn’t a good one, partly because it was celebrated at a hotel: Excerpt 7 (‘Encounter at the Seder’): 273 Ela:

274 275

276

277

... 'ani mitsta'eret shenasati, I sorry that I traveled I’m sorry I went, .. ve'im shana haba'a, and if next year, ... hahorim sheli maxlitim shuv, the parents my decide again my parents decide again, .. laxgog beveit malon, to celebrate at house hotel to have the Seder at a hotel, ... 'ani, I,


300 Discourse Studies 10(3) 278 279 280 Irit: 281 Ela: 282 283 Irit:

284

285

286

287 Ela: 288 289 290

291

292

293 294 295 Irit:

.. lo, won’t, .. yavo! come! .. 'avo. come. {correcting interlocutor’s verb conjugation} .. (chuckling) ... 'ani nishba'at! I swear! ... lo nora. not terrible never mind. ... ta'asi 'itanu 'et hasedder, you will do with us the seder have the Seder with us, ... ki yihye ma ze kef ! because will be what this fun because it’ll be loads of fun! ... ki haben dod sheli yavo! because the cousin my will come because my cousin will be there! .. nenene--e! {imitating Irit’s prosody} .. eh ken! uh yes! .. ken. yes. .. /'ani 'a'ase kax/! I will do so /I'll do that/! ... ra'ayon lo ra. idea not bad not a bad idea. ... 'ulay 'ani ya'ase 'et ze be'emet. maybe I will do it actually I might just actually do that/ maybe I’ll actually do that. ... 'axla, great, .. toda. thanks. ... vaksha! welcome!

Ela concludes by saying that if next year’s Seder will also take place at a hotel, she won’t participate (lines 274–9). Irit suggests to her: ta'asi 'itanu 'et hasedder, (‘have the Seder with us’, line 284), enticing her with the fact that her cousin, whom Ela had apparently expressed some interest in in the past, will be there (line 286). Following some bashful imitation of Irit’s prosody when speaking about the cousin (line 287), Ela tentatively accepts this potential match-making offer (lines 288–91), concluding that she may very well join Irit’s family’s



Seder next year: 'ulay 'ani ya'ase 'et ze be'emet (‘I might just actually do that’ or ‘maybe I’ll actually do that’, line 292). This instance of be'emet relates not so much to the verb ya'ase ‘will do’) describing an action in the extralingual world. It occurs following the entire clause, rather than around the predicate. Ela is not saying here that she will ‘do so honestly’ or ‘wholeheartedly’. Rather, she is relating to the entire offer of joining Irit’s family’s Seder next year, thus accepting the stance taken by Irit in her offer. A paraphrase would be 'ulay 'ani ya'ase 'et ze be'emet, kmo she'at matsi'a (‘maybe I’ll actually do that, as you suggest’). Since this be'emet ratifies a suggestion made by the interlocutor, we believe this token of be'emet functions mainly in the interactional realm. Of course, in that an offer is generally not unrelated to some action in the extralingual world, traces of referential be'emet may also be discerned. However, they are far more marginal than those manifested by ve'az hu be'emet nish'ar taku'a sham (‘and then he really remained stuck there’) of Excerpt 1. Indeed, be'emet here occurs externally in relation to the intonation unit, as well as to the clause, as it did in the previous excerpt (vebe'emet 'amarti lo: ‘and indeed I said to him’, Excerpt 6, line 22). The difference is that in Excerpt 6 it was employed intonation-unit initially, whereas here we find it intonationunit finally. 4.5 BE'EMET LATCHING ONTO NEW TOPIC AND REQUESTING ITS FURTHER DEVELOPMENT Another category which occurs consistently at intonation-unit final position in our data is be'emet uttered by the recipient in order to latch onto a topic mentioned by the interlocutor and invite him/her to further develop it. For example, in Excerpt 8, Roi and his partner are discussing the 40-minute walk they took that day: Excerpt 8 (‘Walk’): 32

Galia:

33 34 35 36

37

38

39

Roi:

.... 'az beseder, so alright, .. 'arba'im dakot, forty minutes, ze beseder. is alright. .. gamur. completely [alright]. .... venelex beyom shabat shuv. and we’ll walk on day Saturday again and we’ll walk Saturday again. ... shabat shuv, {–singing–} Saturday again, ... shabat shuv, {–singing–} Saturday again, venir'e li maxar, and seems to me tomorrow,


302 Discourse Studies 10(3) 40 41 42

Galia:

43

Roi:

44 45 46

47 48 49

50

51 52

'ani 'esa le'Amir, I will go to Amir, batsohorayim. in the early afternoon. .. ma hu 'amar 'al ze be'emet? what he said about this be'emet? what did he say about this be'emet? ... beseder, fine, .. ve'az be'eyzeshehu shalav hu--, and then at some point he, .. hevin she, realized that, lo yaxol lihiyot, {---acc---} not can be, it can’t be, she'ani 'ekba hakol. that I will set everything, 'az hu 'amar, so he said, ... tov, {imitating} fine, 'od nedaber batsohoraiyi--m, {---imitating---} still we will talk in the early afternoon, we’ll talk in the early afternoon, 'az 'amarti, so I said, ken. yes.

In conclusion of the previous topic, Galia verifies that they had walked long enough that day (lines 32–5), and expresses her enthusiasm for another walk on Saturday (lines 36–8). The conversation takes place on Thursday. The mention of Saturday probably brings Roi to consider his plans for the rest of that weekend:11 on Friday afternoon he plans to go to his brother Amir (lines 40–1). To this Galia responds with ma hu 'amar 'al ze be'emet? (‘what did he say about this be'emet?’, line 42), ze (‘this’) being Roi’s suggestion that he visit Amir on Friday. The scope of this instance of be'emet is Galia’s entire utterance, rather than the predicate ‘say’. With this be'emet, Galia invites Roi to develop an issue he mentioned (without necessarily intending to develop it). Roi indeed does so in the following lines, elaborating Amir’s response (43–52). Thus, be'emet is employed here in order to latch onto a topic just mentioned by the interlocutor and invite him to further develop it. A paraphrase would be 'im kvar he'eleta 'et hanose, bo be'emet nedaber 'alav, kmo she'ata 'ulay matsi'a (‘since you’ve raised the matter, let’s be'emet (‘then’?) talk about it, as you may be suggesting’). It is difficult to find an English ‘equivalent’.



We see that this be'emet relates to the interaction (as well as to the text) rather than to the extralingual world and expresses the speaker’s stance towards a topic potentially proffered by the interlocutor; that is, her interest in pursuing this topic. Only two such instances were found in the database (3.5% of all be'emet tokens), both occurring in sentence-final rising intonation, at intonation-unit final position. We have already mentioned Traugott’s finding that discourse markers originating from adverbs occur in syntactically marginal positions in the clause (Traugott, 1995a; Traugott and Dasher, 2002). Maschler (in press b) has explained this phenomenon by reference to the phenomenon of projection in interaction (Auer, 2005; Hopper and Thompson, in press): [Discourse markers] carry a strong interactional projection of an immediately pending frame shift [Goffman, 1981], as well as a projection concerning the type of frame shift that is about to occur. Since, as I have argued [Maschler, 1994], frame shifts tend to occur at clause boundaries (a point of maximal shift in constraints of linguistic structure (among other constraints)), the strong interactional projection of an imminent frame shift carried by discourse markers constitutes motivation for verbalizing them in syntactically marginal positions relative to the clause. (Maschler, in press b: 267)

We have also seen Traugott and Dasher’s (2002) claim that the exact position of the discourse marker depends on the typology of word order in the language: SVO languages, such as English and Hebrew, favor initial position; whereas SOV languages such as Japanese, favor final position. However, in a recent article concluding a special issue of the Journal of Historical Pragmatics on historical change in Japanese with special focus on subjectivity and intersubjectivity, Traugott writes: Being an SOV language, Japanese appears to favor the right periphery for many intersubjective functions that are of an illocutionary type [. . .] Of interest is the fact that, according to the papers in this issue, the very left periphery appears to be the locus mainly of interjections. Other intersubjective elements are recruited to ‘near-sentence-initial’ [. . .]. An important research question that arises [. . .] is precisely what different subtypes of functions are served by left and right periphery in Japanese. (Traugott, 2007: 303)

The intersubjective elements mentioned by Traugott all refer to the interaction among discourse participants and therefore all fulfill the semantic criterion for discourse markerhood here. The present study shows that just as in the case of Japanese, also in an SVO language (Hebrew), some elements carrying intersubjective functions appear in the ‘opposite’ position; namely, intonation-unit finally as opposed to initially. Similarly, in their study of English in fact (which, as we have seen (note 6), is not unrelated to Hebrew be'emet), Schwenter and Traugott note that in fact signaling that what follows is a stronger argument than what precedes (such as what’s more, indeed) is a discourse marker occurring at clause-initial position. In conversation, they note, it may occur clause-finally (2000). However, Schwenter and Traugott do not specify the conditions under which in fact



might occur clause-finally in conversation. We can begin to answer Traugott’s research question above (2007) for Hebrew be'emet: latching onto a new topic and requesting its further development, as well as many instances of ratifying be'emet, tend to occur at the ‘opposite’ position.12

5. Be'emet which is both referential and interactional The blending of interactional and referential functions is far more pronounced in the intermediate major category of Table 1. We find this blending in five tokens throughout the corpus – 8 percent of all be'emet tokens. For example, in Excerpt 9, two non-observant Jewish friends, Beni and Uri, tell Michal, also non-observant, of their experience at a Purim party they attended at an Orthodox Yeshiva in Jerusalem. Michal is strictly opposed to their attending such events, as she believes that the only goal of the organizers of such events is to convert the visitors into Orthodox Judaism: Excerpt 9 (‘Born-Again Judaism’): 39

Michal:

40 41

Beni:

42

Uri:

43 44

Michal:

45

46

47

Beni:

48

Michal:

49

Beni:

50

'al telxu lamifgashim ha'ele, don’t go to the meetings these don’t go to these meetings, .. nu! (urging and belittling token13 ) lo holxim. [we] don’t go. eize mifgashim. what meetings. ... halaxnu lehagid shalom lexaver. we went to say ‘hi’ to a friend. lo, no, pa'am 'axat le'Amnon Yitsxak, time one to Amnon Yitsxak once to Amnon Yitsxak, vepa'am 'axat leze. and time one to this and once to this [meeting]. lo, no, ma 'atem holxim? {---p---} what you going why are you going? le'amnon yitsxak be'emet lo hayinu. to Amnon Yitsxak really not we were to Amnon Yitsxak we really weren’t. ... banu, we came,



.. 'aval lo ra’inu 'oto. but not we saw him but we didn’t see him.

In the preceding lines, Michal had expressed her unambiguous objection to their participation in these religious events. Beni and Uri defend themselves, claiming they do not go to these events (e.g. Beni: lo holxim ‘[we] don’t go’, line 41; Uri: eize mifgashim. halaxnu lehagid shalom lexaver. (‘what meetings. we went to say ‘hi’ to a friend’, lines 42–3). Michal, however, does not accept their objections (line 44). She elaborates that once they went to a meeting with Amnon Yitsxak (a famous born-again rabbi who is very active in convincing nonobservant Jews to become Orthodox), and another time they went to the party they were just describing to her (lines 45–6). It is not, of course, possible for Beni to deny their participation in the party, but it is possible to object to their participation in the Amnon Yitsxak event. Beni thus says: lo, le'amnon yitsxak be'emet lo hayinu (‘no, to Amnon Yitsxak we really weren’t’, lines 47, 49).14 He further explains that they got there, but didn’t see him (lines 50–1). We find here adverbial be'emet modifying the predicate lo hayinu (‘we weren’t’) of the clause, and attempting to construct a situation in the extralingual world – their not having gone to see Amnon Yitsxak. However, this utterance clearly refers back also to Beni’s earlier utterance, lo holxim (‘[we] don’t go’, line 41), further constructing and ratifying the stance he had taken up earlier, that they do not attend such events. In this sense, this be'emet is not only referential, but relates also to the interaction, and can be paraphrased le'amnon yitsxak be'emet lo hayinu, kmo she'amarti kodem (‘to Amnon Yitsxak we really didn’t go, as I’ve said before’). Furthermore, as it objects to Michal’s stance expressed in the earlier statement (that they are constantly going to such events), this be'emet constructs not only the speaker’s subjectivity, but also his intersubjectivity in relation to the interlocutor. The intersubjectivity is constructed also in relation to Uri, who had constructed a similar stance in lines 42–3, thus strengthening the bond of agreement between them. We see that on one level, be'emet here is adverbial, relating to the extralingual world; but on another level, it is interactional – ratifying a speaker’s previous stance, objecting to the stance of one interlocutor, and bonding with the stance of another. We find it intonation-unit internally (as in the case of referential be'emet), but its scope spans a much larger discourse segment than the predicate of the clause and refers back to utterances occurring in this case up to eight intonation units earlier. Interestingly, almost all of the contexts in which be'emet functions both referentially and interactionally involve an argument, or at least some kind of disagreement. It seems that the stance-taking function of the adverb be'emet is intensified the more argumentative the discourse situation is. In such contexts, participants are drawn to construct their stance more strongly, in opposition to that of the other participants. In other words, as the primary, adverbial function becomes contextualized, it gains interactional functions as well. The more argumentative the context, the more pronounced the stance-taking function



of be'emet. Later on in the development of be'emet, the stance-taking function is what ‘lives on’, and we find it unaccompanied by the adverbial function, as in the case of, for instance, (ad)mirative be'emet. The five tokens of be'emet comprising this intermediate category, then, are particularly illuminating for deciphering the functional itinerary of be'emet. Close examination of the contexts in which they occur suggests how a particular utterance might come to serve both referential and interactional functions.

6. Discussion: the functional itinerary of be'emet In Hopper and Traugott’s definition, grammaticization is the change ‘whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions, and, once grammaticalized, continue to develop new grammatical functions’ (2003: xv). Hopper and Traugott mention also the principle of unidirectionality in grammaticization, according to which a typical path of grammaticization follows a cline of de-categorialization (2003). The starting point of this process is typically a full category (e.g. noun or verb), and over time, the lexical item becomes more grammatical and tends to lose the morphological and syntactic properties that would identify it as a full member of a major grammatical category. [. . .] In its most extreme form such a change is manifested as a cline of categoriality, statable as: major category (> adjective/adverb) > minor category (Hopper and Traugott, 2003: 107)

where ‘minor category’ includes categories such as preposition, conjunction, auxiliary, pronoun, demonstrative (Hopper and Traugott, 2003) as well as the category of discourse marker (Ferrara, 1997). The utterance be'emet came to function as an adverb according to a productive process of deriving adverbs from nouns via prefixation with the preposition b(e)-.: Preposition (b(e)-) + Noun ('emet) ––> Adverb (be'emet).

The prototypical adverb, as we have seen (Goldenberg, 1998; Ramat and Ricca, 1994) is the manner predicate adverb. The earliest documented (Biblical) tokens of be'emet are of this type. Once be'emet became an adverb, it continued to develop according to Hopper and Traugott’s schema and came to function in the minor grammatical category of discourse marker. We have seen that already the information found in Even Shoshan’s (1986) dictionary concerning be'emet supports the grammaticization path from adverb to discourse marker: Biblical usages are adverbial; whereas later, Talmudic and Dead Sea Scroll usages (‘but, it is true that ('omnam), in contrast to what was said before’) are more of the discourse marker type. Upon completion of the analysis for this study, a more recent edition of Even Shoshan’s dictionary became available to us. The 2003 edition adds three



more meanings to those already noted in the 1986 edition, specifying that they originated at a later period of the language – Modern Hebrew – and that they are colloquial. They are supplied in the dictionary along with punctuation marks: 1) be'emet? ‘is this true?’, and be'emet! with two meanings: 2) ‘it’s true, there’s no doubt of it!’ and 3) ‘please, will you stop it already!’. These are, of course, precisely the admirative, negating any doubt, and reprimanding uses, respectively, which we have found in our analyses of the excerpts. In other words, uses of the quintessentially discourse marker type are listed in the dictionary as having developed at a later stage of the language. Note that the information found in Even Shoshan (1986, 2003) does not map directly onto our findings. The functions of 1) ratifying a stance taken in a previous utterance and 2) latching onto a topic proffered by the interlocutor and requesting its further development are found neither in Even Shoshan 1986 or 2003. On the other hand, the second meaning mentioned by Even Shoshan, ‘but, it is true that ('omnam), in contrast to what was said before’, from Talmudic and Dead Sea Scroll periods, was not found in the present data and strikes us as archaic and incomprehensible in terms of modern day usage.15 However, traces of the adversative meaning can be found in some of the functions we have discovered in our corpus. We have seen that the concept of adversativity is sometimes involved in referential be'emet (Excerpts 1, 2) as well as when ratifying previous utterances, particularly in argumentative contexts, when stances are constructed in opposition to one another (Excerpt 9). The connection between adversativity and the concept of ‘truth’ is not surprising: ‘the truth’ will often be provided (particularly in an argumentative setting) in opposition to what is perceived by the speaker as ‘not the truth’. Can the new information from the 2003 dictionary be taken as proof of the later stages of the grammaticization path we are suggesting here? It is theoretically possible that, since no recordings of spoken discourse are available from Biblical or Talmudic times, (ad)mirative, reprimanding, and negating any doubt uses were already found at these earlier periods of Hebrew, but there was of course no way of documenting them. We tend to think otherwise, because had such uses been common then, written renditions of spoken discourse (such as written dialogues) would have illustrated such uses. For the discourse marker tov (‘okay, fine’) derived from the adjective ‘good’, for instance, we find the discourse marker use already in Biblical Hebrew (Maschler, in press b). We would like to suggest that from a prototypical manner predicate adverb relating to the extralingual world (i.e. with referential functions) and occurring close to the predicate, be'emet developed into an adverb having both referential and interactional functions (third column from the left in Table 1) particularly in contexts in which participants’ stances were presented against one another. In this role, be'emet maintained its position intonation-unit (and clause-) internally. It then developed into an utterance mid-way between an adverb and a discourse marker, whose interactional functions gradually became more prominent than its referential functions (namely, the functions of 1) ratifying a stance taken in a previous utterance (fourth and fifth columns, Table 1) and 2) latching onto a new topic pursuing its further development (sixth column)). The intermediate



position of these latter two functions in terms of the grammatical category they belong to (adverb vs discourse marker) is mirrored on the one hand by the fact that they usually do not occupy a position close to the predicate (i.e. they do not occur intonation-unit internally), and on the other by their only partial fulfillment of the structural criterion in the definition of discourse markers (i.e. many of them occur at intonation unit final position). Finally, by association with certain metalingual utterances (see later), be'emet developed into a full-fledged discourse marker fulfilling both the semantic and the structural requirements for discourse markerhood. We have seen that almost half the be'emet tokens in this corpus – 28 tokens (44.5%) – function in this role. The present corpus shows that discourse marker be'emet came to have three different interactional functions: (ad)mirative, reprimanding, and negating any doubt. The functional itinerary from the referential use, as in, for example, ve'az hu be'emet nish'ar taku'a sham (‘and then he really remained stuck there’, Excerpt 1, line 92) to (ad)mirative be'emet?, ‘really?’ (line 84) may be puzzling upon first sight. However, it is explainable once we view Alona’s be'emet? as part of a longer, metalingual utterance vaguely in the background, such as ze be'emet ma shekara? (‘is this really what happened?’ or ‘is this what actually happened?’) in response to the climax of her interlocutor’s narrative. Similarly, in the case of reprimanding be'emet, we can posit a metalingual utterance such as 'ani be'emet mitpale 'alayix! (‘I’m really surprised at you (for not sharing my assumptions)!’). In the case of be'emet negating any doubt, we have seen that a metalingual utterance such as 'ani be'emet mitkavenet laze! (‘I really mean it!’) could be posited. The functional itinerary from referential be'emet leading to be'emet ratifying the stance taken in a previous utterance, as in vebe'emet 'amarti lo (‘and indeed I said to him’, Excerpt 6, line 22) can also be explained if we view be'emet as part of a longer, metalingual utterance, this time: kfi shebe'emet kodem 'amarti, 'amarti lo (‘as I was indeed saying earlier, I said to him’). Similarly, in ratifying the stance taken by another discourse participant, the metalingual utterance kfi shebe'emet kodem 'amarta (‘as you were indeed saying earlier’) can be posited. Note that in all five posited metalingual utterances above, be'emet occurs clause-internally, close to the predicate. We suggest that over time, as such metalingual utterances are repeated over and over again in interaction, it becomes unnecessary to verbalize the entire utterance, and its most modal component – be'emet – is all that remains.16 This view is supported by studies, such as those in Bybee and Hopper’s 2001 collection entitled Frequency and the Emergence of Linguistic Structure, but of course, without a spoken corpus from an earlier period of the language, this can only be hypothesized. This grammaticization process is accompanied, according to Traugott, by subjectification in the semantic-pragmatic process of change: Subjectification is the mechanism whereby meanings come over time to encode or externalize the speaker/writer’s perspectives and attitudes as constrained by the communicative world of the speech event, rather than by the so-called ‘real-world’ characteristics of the event or situation referred to. (2003: 126)



This subjectification can be seen already in the first stage of grammaticization, in the move from referential be'emet to be'emet which is both referential and interactional. One would anticipate such a process particularly in the case of a word involving the concept of 'emet (‘truth’), which is intertwined with the speaker’s beliefs and attitudes towards his or her discourse. In fact, it is quite possible that already many of the adverbial, referential uses of this word are subjective. After all, when a speaker describes some action in the extralingual world as being performed ‘honestly, in trustworthiness, wholeheartedly’ (Shoshan, 1986), his/her subjective judgment concerning these values can hardly be separated from the comment concerning the extralingual world. Following the process of subjectification in grammaticization, intersubjectification, a further process of semantic-pragmatic change occurs: [i]ntersubjectivity is the explicit expression of the speaker/writer’s attention to the ‘self ’ of addressee/reader in both an epistemic sense (paying attention to their presumed attitudes to the content of what is said), and in a more social sense (paying attention to their ‘face’ or ‘image needs’ associated with social stance and identity). (p. 128)

We have seen how stances are constructed in interaction via be'emet, resulting in the attending not only to one’s previously constructed stance, but also to that of other participants in the discourse situation (Excerpts 6, 7 and 9). This attending to the interlocutor is evident also in (ad)mirative be'emet, when Alona constructs a stance of amazement and disbelief via be'emet (‘really?’, Excerpt 3, line 84) following her interlocutor’s delivery of the climax of the narrative, even though the information conveyed isn’t all that unexpected to her. Figure 1 summarizes the grammaticization path suggested here.17 From referential be'emet as in ve'az hu be'emet nish'ar taku'a sham (‘and then he be'emet remained stuck there’, Excerpt 1), grammaticization developed to be'emet which is both referential and interactional, as in le'amnon yitsxak be'emet lo hayinu (‘to Amnon Yitsxak we be'emet didn’t attend, as I’ve said before’, Excerpt 9). In both cases, be'emet is close to the predicate. Because of the process of subjectification, be'emet came to be used more and more in order to construct the speaker’s ‘perspectives and attitudes as constrained by the communicative world of the speech event, rather than by the so-called ‘‘real-world’’ characteristics of the event’ (Traugott, 2003: 126). Grammaticization continued, then, into interactional (ratifying) be'emet, as in vebe'emet 'amarti lo (‘and indeed I told him, as I’ve indeed said before’, Excerpt 6) or 'ulay 'ani ya'ase 'et ze be'emet (‘I might just be'emet do that, as you indeed suggest’, Excerpt 7). The metalingual utterances referring to the communicative world of the speech event (‘as I/you’ve be'emet said before’) became redundant, so that there was no need to repeat them over and over again. Just as one can accept an interlocutor’s suggestion concerning a matter in the extralingual world (such as the invitation to participate in next year’s Seder, Excerpt 7), so can one accept, in the metalingual realm, an interlocutor’s potential suggestion to raise a new topic (such as the topic of visiting Amir, Excerpt 8). Thus we see another interactional be'emet – that of latching onto a new topic pursuing its





FIGURE

1.

Grammaticization of be’emet

RATIFYING

↓

INTERACTIONAL be'emet 

↓

INTERACTIONAL (DISCOURSE MARKER) be'emet

('ani be'emet mitkavenet laze! NEGATING ANY DOUBT ‘I be'emet mean what I’ve just said’!)  be'emet! (5) ('ani be'emet mitpale 'alexa! REPRIMANDING ‘I’m be'emet surprised at you be'emet! (4) for what you’ve just said’!)  (ze be'emet ma shekara? (AD)MIRATIVE ‘is this be'emet what happened, be'emet? (3) as you’ve just said?’) 

POSITED METALINGUAL USAGES (NOT IN CORPUS)

ma hu 'amar 'al ze be'emet? (8) ‘since you’ve raised the matter, let’s be'emet talk about it, as you may be suggesting: what did he say about it?’.

NEW TOPIC

le'amnon yitsxak be'emet lo hayinu. (9) vebe'emet 'amarti lo (6) ‘to Amnon Yitsxak we be’emet didn’t ‘and be'emet I said to him, as I’ve be'emet said before’ attend, as I’ve said before’ 'ulay 'ani ya'ase 'et ze be'emet (7) ‘I might just be'emet do that, as you be'emet suggest’

↓

be'emet BOTH REFERENTIAL & INTERACTIONAL 

ve'az hu be'emet nish'ar taku'a sham (1) ‘and then he be'emet remained stuck there’

↓

REFERENTIAL be'emet



further development. In these last three cases, be'emet occurs intonation-unit initially or finally. In order to understand the final grammaticization stage, we posit additional metalingual utterances vaguely in the background, for example: 'ani be'emet mitkavenet laze (‘I be'emet mean what I’ve just said!’), 'ani be'emet mitpale 'alayix! (‘I’m be'emet surprised at you for what you’ve just said’), or, in rising question intonation: ze be'emet ma shekara? (‘is this be'emet what happened, as you’ve just said?’). With time, only the most stance-taking component of these metalingual utterances remains, along with its intonation. This component occurs intonationunit initially, occupying in fact the entire intonation unit. What remains, then, is either be'emet! or be'emet?. We see that just as the category of adverb is a scalar, or even radial one (Ramat and Ricca, 1994), so the category of discourse marker can be viewed in a similar fashion: prototypical be'emet discourse markers (those which are highly metalingual (interactional) and have a very minor referential dimension) fulfill both the semantic and the structural requirements in the definition of discourse markers, whereas less prototypical discourse markers (those operating metalingually [interactionally] as well as referentially) fulfill the requirements only partially. Of this latter category, discourse markers having a higher interactional dimension (ratifying and latching onto new topic) occur intonationunit initially or finally, while those which are both interactional and referential occur intonation-unit internally. The answer, then, to Traugott’s question concerning the different subtypes of functions served by the left and right periphery (2007), at least in the case of Hebrew be'emet, has to do with the degree of metalinguality of the discourse marker in question. ACKNOWLEDGMENTS

We are very grateful to Paul Woodward, Tamar Zewi, Galia Estlein, and an anonymous reviewer for discussion and constructive criticism of this work. This is a written version of papers presented at the 2005 ICA annual meeting in New York City and at the IPrA 2005 meeting in Riva del Garda, Italy. Our thanks to the participants for their comments. N OTE S

1. The metaphor of ‘functional itinerary’ is used here as in Fleischman and Yaguello (2004: 139), and is of course not meant to suggest that grammaticization (see later) is an independent process, as one of the reviewers has cautioned us. As will be demonstrated throughout, rather, we take a usage-based approach, according to which participants act in strategic ways that result in grammaticization, so that speakers drop off frequently used bits of language that ultimately come to be regarded as uninformative. 2. In the English Standard Version of the Bible: ‘if you acted in good faith and integrity’. 3. For surveys, see Brinton (1996), Fischer (2006), Jucker and Ziv (1998), Maschler (in press a, in press b), Schourup (1999), Schiffrin (2001). 4. Hebrew predicates can be either verbal or nominal.


312 Discourse Studies 10(3) 5. Transcription method: Each line denotes an intonation unit and is followed by an English gloss. In the cases in which this gloss is not close enough to an English utterance, it is followed by a third line supplying a usually literal (but sometimes functional) translation. Utterances under consideration (generally be'emet) are given in boldface. Transcription basically follows Chafe (1994), with a few additions. Conventions are as follows: ... – half second pause (each extra dot = another 1/2 second) .. – perceptible pause of less than half a second (3.56) – measured pause of 3.56 seconds , – comma at end of line – clause final intonation (‘more to come’) . – period at end of line – sentence final falling intonation ? – question mark at end of line – sentence final rising intonation, ‘appeal intonation’ (Du Bois et al., 1992). ! – exclamation mark at end of line – sentence final exclamatory intonation ø – lack of punctuation at end of line – a fragmentary intonation unit, one which never reached completion. -- two hyphens – elongation of preceding vowel sound [ square bracket to the left of two consecutive lines indicates overlapping speech, two speakers talking at once alignment such that the right of the top line is placed over the left of the bottom line indicates latching, no interturn pause /??????/ – transcription impossible /words within slashes/ indicate uncertain transcription pp – pianissimo, very softly ff – fortissimo, very loudly [xxxxx] – material within square brackets in the gloss indicates exuberances of translation (what is not there in the original). {in curly brackets} – transcriber’s comments concerning paralinguistics and prosody, which do not have an agreed upon symbol in this transcription system. 6. The link between be'emet and adversative meaning is reminiscent of Schwenter and Traugott’s analysis (2000) of English in fact in the following excerpt: They think he’s crazy, (but) in fact he’s quite sane. Schwenter argues that in fact in this example expresses an adversative meaning in a dialogical viewpoint (Schwenter, 2000). Interestingly, the translation of this utterance into Hebrew involves another derivative of the noun 'emet (‘truth’); namely, the phrase ha'emet (hi she-) (‘the truth [is that]’) (either with or without the following copula [hi, ‘is’] and complementizer [she-, ‘that’]): hem xoshvim shehu meshuga, ('aval) ha'emet (hi she)hu dey shafuy. they think that he[‘s] crazy, (but) the truth (is that) he [is] quite sane. Compare this use of ha'emet with the one of Excerpt 6, line 21. 7. For the functions of nu in Hebrew talk-in-interaction, see Maschler (2003). 8. Notice that this token of be'emet indeed occurs intonation-unit internally preceding the predicate of the clause, as in the case of referential be'emet. 9. Notice that this instance of be'emet also fulfills the structural requirement for discourse markerhood, as it appears intonation-unit initially, in a cluster of discourse markers.


Maschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 313 10. The translation of ‘indeed’, however, is lacking in that it is much higher in register compared to Hebrew be'emet. There exists another Hebrew word, 'axen, which is a closer ‘equivalent’ of English ‘indeed’ in this context, as it, too, belongs in a fairly high register. There is no English equivalent in both function and register for be'emet of line 22. 11. Israeli weekends span Friday through Saturday. 12. Another Hebrew discourse marker which sometimes appears intonation-unit finally is 'aval (‘but’), though no examples were found in the corpus. Interestingly, like some instances of be'emet, this discourse marker expresses adversativity. 13. See note 7. 14. The syntax of line 49 warrants further explanation. The Hebrew utterance is just as ungrammatical as its English rendition here. Following the preposition le- (‘to’) beginning this intonation unit, we would expect the verb lo halaxnu (‘we didn’t go’) rather than the verb lo hayinu (‘we weren’t’), as we find here. This is an instance in which the speaker begins a certain syntactic construction but completes it as if s/he had begun a different one. Since the message was understood by the recipient, the speaker does not bother with repair at mid-utterance in spite of its ungrammaticality. However, this switching constructions at mid-utterance also aids Beni to construct reality in a way that better suits his argument. It would be a lie to employ the verb lo halaxnu (‘we didn’t go’) here, since the truth is that they did go to the event. Beni began saying this intonation unit (le'amnon yitsxak be'emet lo [‘to Amnon Yitsxak we really didn’t’]), dialogically repeating (Du Bois, 2001) the construction lalexet le- (‘to go to’), various forms and parts of which are found throughout the previous intonation units (lines 39, 41, 43, 45, 46, 48). Presumably, as he began verbalizing intonation unit 49, he realized that employing lo halaxnu (‘we didn’t go’) would not make a truthful utterance. A quick way not to lie here is to continue the utterance with a different verb – lo hayinu (‘we weren’t’), even if this verb doesn’t fit too well with the preposition already verbalized. This makes for a truthful utterance: they indeed went, but the Rabbi didn’t get there (as he explains in lines 50–1), so, technically speaking, one could say that they therefore did not attend the event, and the verb lo hayinu (‘we weren’t [at the Amnon Yitsxak event]’) is suitable. The fact that the first syllable of hayinu is identical to that of halaxnu also makes this a good choice, because it renders the incompatibility with the preposition le- (‘to’) already uttered less jarring. 15. The example provided in Even Shoshan for the Talmudic meaning is incomprehensible in terms of the modern language with respect to any of the three options provided by Even Shoshan: velo yikra le'or haner, be'emet 'amru: haxazan ro'e heyxan hatinokot kor'im (‘and he shall not read by candlelight, be'emet (‘but/it is true that/in contrast to what was said before’) it was said: the cantor sees where the youngsters read’, Shabbat a, 3). 16. A similar process has happened for other Hebrew discourse markers: nu (‘yeah, go on’), bekitsur (‘anyway’), ke'ilu (‘like’), and tov (‘okay, fine’). Maschler argues, in fact, for a general process of crystallization of metalingual utterances into discourse markers (in press b). 17. Boldfaced English utterances in Figure 1 indicate translations of the metalingual utterances hypothesized throughout the study in order to explain the functional itinerary. Numbers to the right of each Hebrew utterance refer to the excerpt from which the utterance is extracted.


314 Discourse Studies 10(3) REFERENCES

Auer, P. (2005) ‘Projection in Interaction and Projection in Grammar’, Text 25: 7–36. Becker, A.L. (1988) ‘Language in Particular: A Lecture’, in D. Tannen (ed.) Linguistics in Context: Connecting Observation and Understanding (Lectures from the 1985 LSA/TESOL and NEH Institutes), pp. 17–35. Norwood, NJ: Ablex. Reprinted in Becker (1995). Becker, A.L. (1995) Beyond Translation: Essays Toward a Modern Philology. Ann Arbor: University of Michigan Press. Biber, D. and Finegan, E. (1988) ‘Adverbial Stance Types in English’, Discourse Processes 11: 1–34. Brinton, L.J. (1996) Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin: Walter de Gruyter. Bybee, J.L. and Hopper, P. (2001) Frequency and the Emergence of Linguistic Structure. Amsterdam: John Benjamins. Chafe, W. (1986) ‘Evidentiality in English Conversation and Academic Writing’, in W. Chafe and J. Nichols (eds) Evidentiality: The Linguistic Coding of Epistemology, pp. 261–72. Norwood, NJ: Ablex. Chafe, W. (1994) Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago, IL: University of Chicago Press. DeLancey, S. (2001) ‘The Mirative and Evidentiality’, Journal of Pragmatics 33: 369–82. Downing, A. (2001) ‘ ‘‘Surely You Knew!’’: Surely as a Marker of Evidentiality and Stance’, Funcions of Language 8(2): 251–85. Du Bois, J.W. (2001) ‘Towards a Dialogic Syntax’, manuscript, Department of Linguistics, University of California, Santa Barbara. Du Bois, J.W., Cumming, S., Schuetze-Coburn, S. and Paolino, D. (1992) Discourse Transcription: Santa Barbara Papers in Linguistics, Vol. 4. Santa Barbara: Department of Linguistics, University of California, Santa Barbara. Even Shoshan, A. (1986) Hamilon Hexadash [The New Dictionary]. Jerusalem: Kiryat Sefer. Even Shoshan, A. (2003) Milon Even Shoshan [The Even Shoshan Dictionary]. Jerusalem: Hamilon Hexadash Ltd. Ferrara, K. (1997) ‘Form and Function of the Discourse Marker Anyway: Implications for Discourse Analysis’, Linguistics 35: 343–78. Fischer, K. (2006) ‘Towards an Understanding of the Spectrum of Approaches to Discourse Particles: Introduction to the Volume’, in K. Fischer (ed.) Approaches to Discourse Particles, pp. 1–20. Amsterdam: North Holland/Elsevier. Fleischman, S. and Yaguello, M. (2004) ‘Discourse Markers in Comparative Perspective: A Contribution to Cross-language Pragmatics’, in C.L. Moder and A. Martinovic-Zic (eds) Discourse Across Languages and Cultures, pp. 129–47. Amsterdam: John Benjamins. Ford, C.E. and Thompson, S.A. (1996) ‘Interactional Units in Conversation: Syntactic, Intonational, and Pragmatic Resources for the Management of Turns’, in E. Ochs, E.A. Schegloff and S. Thompson (eds) Interaction and Grammar, pp. 134–84. Cambridge: Cambridge University Press. Friedman, V.A. (1986) ‘Evidentiality in the Balkans: Bulgarian, Macedonian, and Albanian’, in W. Chafe and J. Nichols (eds) Evidentiality: The Linguistic Coding of Epistemology, pp. 168–87. Norwood, NJ: Ablex. Goffman, E. (1981) Forms of Talk. Philadelphia: University of Pennsylvania Press. Goldenberg, G. (1998) ‘On Verbal Structure and the Hebrew Verb’, in Studies in Semitic Linguistics, pp. 148–96. Jerusalem: Magnes.


Maschler and Estlein: Stance-taking in Hebrew casual conversation via be'emet 315 Haviland, J.B. (1987) ‘Fighting Words: Evidential Particles, Affect and Argument’, in J. Aske, N. Berry, L. Michaelis and H. Filip (eds) Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society on Grammar and Cognition, pp. 343–54. 14–16 February 1987, Berkeley, CA. Hopper, P.J. and Thompson, S.A. (in press) ‘Projectability and Clause Combining in Interaction’, in R. Laury (ed.) Crosslinguistic Studies of Clause Combining: The Multifunctionality of Conjunctions. Amsterdam: John Benjamins. Hopper, P.J. and Traugott, E.C. (2003) Grammaticalization, 2nd edn. New York: Cambridge University Press. Jucker, A.H. and Ziv, Y. (eds) (1998) Discourse Markers: Descriptions and Theory. Amsterdam: John Benjamins. Kärkkäinen, E. (2003) Epistemic Stance in English Conversation. Amsterdam: John Benjamins. Maschler, Y. (1994) ‘Metalanguaging and Discourse Markers in Bilingual Conversation’, Language in Society 23: 325–66. Maschler, Y. (1998) ‘Rotse lishmoa keta? ‘‘Wanna Hear Something Weird/Funny [Lit. ‘a Segment’]?’’: The Discourse Markers Segmenting Israeli Hebrew Talk-in-interaction’, in A.H. Jucker and Y. Ziv (eds) Discourse Markers: Descriptions and Theory, pp. 13–59. Amsterdam: John Benjamins. Maschler, Y. (2002) ‘The Role of Discourse Markers in the Construction of Multivocality in Israeli Hebrew Talk-in-interaction’, Research on Language and Social Interaction 35: 1–38. Maschler, Y. (2003) ‘The Discourse Marker nu: Israeli Hebrew Impatience in Interaction’, Text 23: 89–128. Maschler, Y. (2004) ‘The Haifa Corpus of Spoken Israeli Hebrew’, available online at: http://hevra.haifa.ac.il/com/maschler/, accessed August 2004. Maschler, Y. (in press a) ‘Discourse Markers’, in International Encyclopedia of Communication. New York: Basic Blackwell. Maschler, Y. (in press b) Metalanguage in Interaction: Hebrew Discourse Markers. Amsterdam/ Philadelphia, PA: John Benjamins. Ramat, P. and Ricca, D. (1994) ‘Prototypical Adverbs: On the Scalarity/radiality of the Notion of ADVERB’, Rivista di Linguistica 6: 289–326. Ravid, D. (1977) ‘Mispar heibetim shel be'ayat seder hamarkivim be'ivrit yisra'elit modernit’ [Several Aspects of the Problem of Element Order in Modern Israeli Hebrew], balshanut, 'ivrit xofshit [Hebrew Computational, Formal, and Applied Linguistics] 11: 1–45 (in Hebrew). Schegloff, E.A. (1996) ‘Turn Organization: One Intersection of Grammar and Interaction’, in E. Ochs, E.A. Schegloff and S.A. Thompson (eds) Interaction and Grammar, pp. 52–133. Cambridge: Cambridge University Press. Schiffrin, D. (1987) Discourse Markers. Cambridge: Cambridge University Press. Schiffrin, D. (2001) ‘Discourse Markers: Language, Meaning, and Context’, in D. Schiffrin, D. Tannen and H.E. Hamilton (eds) The Handbook of Discourse Analysis, pp. 54–75. Oxford: Blackwell. Schourup, L. (1999) ‘Discourse Markers: Tutorial Overview’, Lingua 107: 227–65. Schwenter, S.A. (2000) ‘Viewpoints and Polysemy: Linking Adversative and Causal Meanings of Discourse Markers’, in E. Couper-Kuhlen and B. Kortmann (eds) Cause, Condition, Concession, Contrast, pp. 257–81. Berlin: Walter de Gruyter. Schwenter, S.A. and Traugott, E.C. (2000) ‘Invoking Scalarity: The Development of in Fact’, Journal of Historical Pragmatics 1: 7–25.


316 Discourse Studies 10(3) Tannen, D. (1989) Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse. Cambridge: Cambridge University Press. Traugott, E.C. (1995a) ‘The Role of the Development of Discourse Markers in a Theory of Grammaticalization’, 12th International Conference on Historical Linguistics, Manchester, August 1995 (version of 11/97), available online at: http://www.stanford. edu/~traugott/papers/subject2intersubject.pdf, accessed 22 January 2004. Traugott, E.C. (1995b) ‘Subjectification in Grammaticalisation’, in D. Stein and S. Wright (eds) Subjectivity and Subjectivisation: Linguistic Perspectives, pp. 31–54. Cambridge: Cambridge University Press. Traugott, E.C. (2003) ‘From Subjectification to Intersubjectification’, in R. Hickey (ed.) Motives for Language Change, pp. 124–39. Cambridge: Cambridge University Press. Traugott, E.C. (2007) ‘(Inter)subjectification and Unidirectionality’, Journal of Historical Pragmatics 8: 295–309. Traugott, E.C. and Dasher, R.B. (2002) ‘The Development of Adverbials with Discourse Marker Function’, in Regularity in Semantic Change, pp. 152–89. Cambridge: Cambridge University Press.

teaches Linguistics at the Department of Communication and at the Department of Hebrew Language, University of Haifa. Her research interests focus on the grammaticization of discourse patterns into linguistic structure. She has published articles on how grammar is both constitutive of and emergent from interaction in book collections and journals such as Language in Society, Research on Language and Social Interaction, Journal of Pragmatics, Discourse Studies, Text, Pragmatics, International Journal of Bilingualism, and Balshanut 'Ivrit (Hebrew Linguistics). Her book Metalanguage in Interaction: Hebrew Discourse Markers will be published by John Benjamins later this year. A D D R E S S : Department of Communication, University of Haifa, Mount Carmel 31905, Haifa, Israel. [email: [email protected]] YA E L M A S C H L E R

R O I E S T L E I N holds a Masters degree from the Department of Communication, University of Haifa. His research interests focus mainly on human communication processes investigating how they shape social roles and construct identities, and on how people create and maintain power, intimacy, and solidarity in interaction. He plans to start his doctoral studies this year at the Department of Communication at Rutgers University, concentrating on interpersonal communication in general and on family communication in particular. A D D R E S S : Department of Communication, University of Haifa, Mount Carmel 31905, Haifa, Israel. [email: [email protected]]
