CC BY-NC-ND
29
Alexey Lyavdansky Moscow, Russian State University for the Humanities
Temporal deictic adverbs as discourse markers in Hebrew, Aramaic and Akkadian 1
The categorial shift from temporal deictic adverb to discourse marker is observed in many languages of the world. There are three Semitic languages — Hebrew, Aramaic, and Akkadian — where similar developments were attested for a temporal adverb with present time reference. This article is dedicated to the comparison of non-adverbial usages for Hebrew (wa)'atta, Aramaic k'tlk'ntlJktn and Akkadian inanna and anumma. The preliminary results of this investigation, based on the findings of Rhetorical Structure Theory and discourse markers research, show that in most of the uses these adverbs function as discourse markers. As is the case with Hebrew (wa)'atta, the specific discourse function is attested also for Aramaic k'tlk'ntlk'n and Akkadian inanna: an adverb with the meaning 'now' marks a transition from assertive discourse unit to directive discourse unit within directive utterances. The range of usage for Aramaic k'tlk'ntlk'n and Akkadian inanna is broader than for Hebrew (wa)'atta. Akkadian anumma is another type of lexeme: not being used as a temporal deictic adverb, it also appears in directive utterances, but, unlike Hebrew (wa)'atta and Aramaic k'tlk'ntlk'n, it usually has an assertive discourse unit in its right co-text, the transition from assertive to directive usually left unmarked.
Keywords: Semitic languages, historical syntax, discourse relations, discourse markers.
It is probably a universal phenomenon that a temporal deictic adverb with the present time reference like English now appears in specific contexts, where it assumes non-temporal meaning. Yet it is only recently that such non-adverbial uses of temporal adverbs have drawn special attention. The interest to these uses is related to the growth of discourse studies over the last three decades. It has been noticed that now and its sister-words in other languages (further designated as "now-words") quite often appear in contexts that are very typical for connectors/connecting particles, or discourse markers. Accordingly, there are two main types of meaning which are usually distinguished for these words: the temporal (adverbial) meaning and the discursive (textual) meaning. This distinction may be demonstrated by the following examples:
(1a) sicut fortis equos, spatio qui saepe supremo vicit Olympia, nunc senio confectus quiescit (A fragment from En-nius, quoted by Cicero, Cato Maior 5)
(1b) redeo nunc ad epistulam tuam (Cicero. Ad Atticum 14.13.5)
(2a) Either do it now or not at all (RHWUD).
(2b) richard plantageneth: Lord Buckingham, methinks, you watch'd her well: A pretty plot, well chosen to build upon! Now, pray, my lord, let's see the devil's writ. What have we here? (Shakespeare. Henry VI, 693-696)
1 I am grateful to Sergey Loesov for his useful suggestions and criticism of the first draft of this paper. Journal of Language Relationship • Вопросы языкового родства • 3 (2010) • Pp. 22-42 • © Lyavdansky A., 2010
Examples (1a, 2a) demonstrate "now-words" (Latin nunc, English now) in a temporal adverbial meaning; in the examples (1b, 2b) "now-words" appear as discourse particles/markers2. Consequently, these two types of meaning can constitute the basis for postulating two distinct lexemes, for which I propose the symbols now I (adverbial) and now II (nonadverbial, discursive).
The non-adverbial usage of temporal deictic adberbs being almost universal, it is, however, difficult to predict the range of the discursive meaning that these now-words assume in a given language. The present article is dedicated to the comparative study of now-words in three Semitic languages: Hebrew, Aramaic and Akkadian. The analysis is concentrated on Biblical Hebrew (including Epigraphic Hebrew of the same epoch) and Egyptian Aramaic. In addition, some parallel examples from Old Babylonian and Standard Babylonian dialects of Akkadian are discussed.
The aim of this paper is to show the preliminary results achieved in the analysis of the development of meaning for now-words in the said three Semitic languages. Within the cadre of this analysis I am trying to understand to what extent it is possible to distinguish between adverbial and non-adverbial uses for dead languages like the chosen ones. Another problem which can be treated only in part at the present state of the research: What more can be said about the comparison between Hebrew (wa)Satta, Aramaic kUtlkHntlkHn and Akkadian inanna and anumma, beyond the fact that they are sometimes similar in usage3?
Although this study is comparative in the sense of "Languages in contrast and comparison", it does not pursue any etymological goals. Nevertheless, its results, if proven viable, may shed a new light on the problems of the origin of Semitic temporal deictic adverbs and the words related to them in meaning.
1. Method and theoretical background
The main theoretical frameworks upon which my investigations are based are Discourse Markers (DM) research4 and Rhetorical Structure Theory (RST) [Mann & Thompson 1988]. I will not go into a detailed presentation of these theories, fairly well-known from many publications. Suffice it just to point out the most important issues for the present discussion. The authors of RST have suggested a catalog of rhetorical (i. e. discourse) relations and provided an analytical instrument that is useful to determine the functions of discourse markers. At a certain point it was understood that some of the discourse particles and functionally similar expressions signal or mark discourse relations5. In fact, connectivity took very firm ground and usually the main position in the definitions of that specific class of words and expressions called "discourse markers" [Fraser 2006; Schourup 1999].
Since one of the main tasks of this paper is to distinguish between adverbial and nonadverbial (discourse) uses of now-words in Hebrew, Aramaic and Akkadian, I need to pay
2 On the criteria of this distinction see par. 1 of the present article.
3 Hebrew w?th and Aramaic kff were compared by many authors, e. g. [Lande 1949; Schwiderski 1997]. As to the comparison between w?th and Akkadian inanna, there is a short remark by Moran [2003: 16].
4 There is no universal theory of Discourse Markers, but there is a certain line of thought represented in the works of Deborah Schoffrin [1997], Bruce Fraser [1996, 1999], Lawrence Schourup [1999] and ADP.
5 As it was put, e. g., by Fraser in one of his earlier works: "...discourse marker, an expression which signals the relationship of the basic message to the foregoing discourse" [Fraser 1996: 186]. "The characteristic figuring most prominently in definitions of DMs is their use to relate utterances or other discourse units" [Schourup 1999: 230].
some attention to the problem of what is to be understood under the term "discourse marker". There are many approaches to DM now, but I will only mention a summary list of DM characteristics suggested by Lawrence Schourup [1999], who tried to survey the most important findings in the field of DM research. According to him, the main characteristics of DM include: 1) connectivity, 2) optionality6, 3) non-truth conditionally7, 4) weak clause association8, 5) ini-tiality9, 6) orality, 7) multi-categoriality10 [Schourup 1999]. The following observations on now according to the above list of DM features may show the applicability of this approach; it was shown that now when used as a discourse particle:
1) marks a transition "from a disputable issue to evaluation of it" [Aijmer 1988: 19]; cf. "l'enonciateur ce sert de now pour marquer une transition avec son discours antérieur" [Brun-aud 1991: 85];
2) is much less frequent in non-surreptitious conversation as compared to surreptitious (recorded using hided devices) conversation [Aijmer 2002: 68],
3) loses its ordinary temporal meaning [Aijmer 1988: 15];
4) is phonologically independent, "constituted most often a (prosodic) phrase on its own" [Aijmer 2002: 59];
5) "has a fixed position in the utterance as the leftmost element in the larger utterance" [Aijmer 1988: 18];
6) is "characteristic of speech rather than of writing" [Aijmer 2002: 70];
7) have as its synonyms expressions from other syntactic classes.
One of the main achievements of DM research is the understanding that the discourse model should account for several layers of discourse coherence, or "planes of talk" [Schiffrin 1987]. It is significant for the discussion of 'now-words' that discourse markers are sensitive not only to rhetorical relations in the sense of RST, they sometimes point to the new speech act in the flow of discourse. In other words, DM may contribute to what has been called "action structure11" [Schiffrin 1987].
Both RST and DM analysis draw attention to the immediate context of DM, not only to the textual unit which immediately follows a DM12, but also to the textual unit which immediately precedes a DM. These units, relevant as the intra-textual context of a DM, may be called "text spans" [Mann & Thompson 1988: 245], "discourse segments" [Fraser 2006: 191] or "discourse units" [Schourup 1999; Redeker 2006]. Consequently, this should be the rule for the presentation of the linguistic material in a discussion of concrete DMs — the uses of a DM should be presented with both adjacent discourse units13, which is rarely done in works on Semitic discourse particles of a connective nature.
6 In the sense that "if a DM is omitted, the relationship it signals is still available to the hearer, though no longer explicitly cued" [Schourup 1999: 231]. The statistical data on the ratio of cued/non-cued discourse relations were presented in [Taboada 2006].
7 DMs usually "contribute nothing to the truth-conditions of the propositions expressed by an utterance" [Schourup 1999].
8 It is often indicated by the phonological independence of a DM.
9 DMs tend to appear at the beginning of a sentence or a discourse unit. Though there are DMs sometimes or even exclusively placed within clauses, e. g. English after all, now, Biblical Hebrew ?epö.
10 The class of DM may include adverbs, conjunctions, interjections, verbs, clauses [Schourup 1999].
11 Analogous to G. Redeker's "rhetorical structure", as suggested in [Müller 2002: 30].
12 Sometimes considered as "host utterance" for a DM, which is problematic because DMs are very often seen as syntactically unintegrated entities [ADP: 8].
13 The problem of limits for these units is not discussed here; see on this [ADP] and [Mann & Thompson 1988].
Finally, one more application of RST to DM research deserves our attention. As has already been shown in a number of RST studies, the text in the final analysis may be looked upon as a complex of discourse units, organized hierarchically and sequentially and related to each other by one (or maybe more than one) of the rhetorical relations. Potentially, most of these relations may be lexically marked in a given language, in a given text type. Therefore, one of the ways to determine the functional distribution of DMs in a language is to parse texts rhetorically, in order to show which relations are marked and to what degree, as has been done by M. Taboada for two English corpora [Taboada 2006]. This is one of the paths —already somewhat trodden —which must lead to tangible results, important for the typology of discourse marking.
As to the languages investigated here — Hebrew, Aramaic and Akkadian — there is obviously quite a lot of work to be done as long as rhetorical relations are concerned. In Biblical Hebrew, where discourse particles have been explored but partially, there is a very promising field of study constituted by the constructed literary dialogue with its plethora of discourse particles (wd-r ki, hen/hinne, (wa)iatta, laken, ial ken, lepo, labal). Imperial Aramaic provides a relatively small, but sufficient (for that sort of study) corpus of letters from Egypt and probably other regions; the field of Imperial Aramaic discourse particles is almost untouched. Finally, besides many interesting corpora in Akkadian, there is the corpus of Amarna correspondence, where numerous discourse particles present serious problems, often avoided by translating these particles automatically by "now" or "moreover".
2.1. Classical Hebrew (wa)fatta.
The Hebrew expression iatta is the main word for "now" in Biblical Hebrew. It occurs 433 times [Jenni 1972: 6], which is quite a figure for such restricted corpus as Hebrew Bible14. In approximately 60% of its usages it occurs with the preceding conjunction wa-, namely waiatta (272 times [Jenni 1972: 6])15. There are also 20 occurrences of wit in epigraphic material, all of them in letters. These statistical data already suggest the general tendency in the usage of iatta: if wa- is attached to iatta, it obviously represents the clause-combining wa-, which means that in almost all of these 272 cases iatta is placed at the beginning of a clause. According to the 'initiality' characteristic of a DM, the clause-initial position of iatta points to its discursive character at least in these cases.
For quite a long time, the string waiatta had been considered a compound with its own specific range of meaning [Laurentin 1964; Brongers 1965]. It was Ernst Jenni who rightly noted that there is no principal difference between the expressions waiatta and iatta. In fact, waiatta is frequent and most conspicuous in its non-adverbial discursive usage, but both waiatta and iatta may have temporal adverbial and discursive functions [Jenni 1972]. Moreover, if we look at the expression wa-iatta from the vantage point of DM analysis, it should be treated as a collocation, a juxtaposition of two discourse markers with similar meanings. Inclusion of the coordinating conjunction wa- into the class of Hebrew DMs is corroborated by my observation that its usage in dialogue differs very much from its usage in narrative. In narrative, wa- is a default coordinating and subordinating conjunction: it opens every clause if it is not (rarely) substituted by other coordinators (lap, raq) or subordinators (lasar, lamaian), or omitted before clauses with specialized function in narrative (author's remarks). In dialogue
14 Approximately 300000 words.
15 On the problems of these statistical data, mostly related to text corruption, see [Jenni 1972].
every appearance of wa- should be accounted for in terms of its specific discourse functions, one of which has been analyzed in [Miller 1999].
2.2. (wa)fatta as a temporal adverb.
First of all, it is important to demonstrate the usage of (wa)Satta in different types of temporal adverbial meaning. Typologically, as shown in [Perennec 2002], the range of temporal reference attested for an adverb may suggest the type(s) of its discourse function.
We already said that Satta is the main word for 'now' in Biblical Hebrew; consequently, it appears in all the most typical contexts for a temporal adverb with present time reference. The adverbial meaning of Satta is highlighted in contexts where the situation of speaking is contrasted with the (more often) past or future situation:
(3) zakarnu ?at haddaga ?asar nokal bamisrayim hinnam ?et haqqissu?im w?et ha?abattihim wa?at hahasir wa?at hab-basalim wa?at hassumim waiatta napsenu yabesa ?en kol bilti ?al hamman 'enenu
We remember the fish we used to eat in Egypt for nothing, the cucumbers, the melons, the leeks, the onions, and the garlic; but now our strength is dried up, and there is nothing at all but this manna to look at.16 (Num 11:5f)
The beginning of the new state of affairs, starting from the present moment, is usually expressed by a prepositional phrase, constituted by the preposition min (me-) and adverb Satta (4); the same phrase is used also in the context of the comparison of the present situation with the past (5).
(4) lamarbe hammisra ulasalom ?en qes 'al kisse dawid wa'al mamlakto lahakin ?otah ulasa'adah bamispat ubisdaqa meiatta wa'ad 'olam
His authority shall grow continually, and there shall be endless peace for the throne of David and his kingdom. He will establish and uphold it with justice and with righteousness from this time onward and forevermore (Isa 9:7).
(5) wa?amara ?elaka wa?asuba ?al ?isi harison ki tob li ?az meiatta
... she shall say, "I will go and return to my first husband, for it was better with me then than now." (Hos 2:9)
Satta may also denote recent past17:
(6) ?anoki'asiti ?at ha?aras ?at ha?adam wa?at habbahema ?asar 'al pane ha?aras bakohi haggadol ubizrol hannatuya unatattiha la?asar yasar ba'enay waiatta ?anoki natatti ?at kol ha?arasot ha?ella bayad nabukadnasar malak babal abdi wagam ?at hayyat hassada natatti lo la'obdo
It is I who by my great power and my outstretched arm have made the earth, with the people and animals that are on the earth, and I give it to whomever I please. But now I have given all these lands into the hand of King Nebuchadnezzar of Babylon, my servant, and I have given him even the wild animals of the field to serve him (Jer 27:5f).
16 The translation of biblical texts is according to New Revised Standard Version if not specified otherwise.
17 The specific nature of the prophetic texts here allows for a double meaning according to time reference: the giving of lands has happened just before the moment of speaking, but it is understood as a decision, or promise of Yahweh, the event itself will happen in (near?) future.
Summing up, iatta as a temporal adverb denotes the moment of speaking, the recent past, the imminent future. The syntactic position of iatta in its temporal adverbial uses is not always non-initial: it may occur at the beginning of a sentence, in which cases it is often preceded by the conjunct wa-18. The adverbial meaning applies here because it is highlighted by the contrast with mostly past situation. The contrast is expressed lexically: zakarnu 'we remember'(4), (me)laz 'then' (5; Josh 14:11), mime qadam "from early days" (2Kgs 19:25). The sense of contrast is sometimes suggested by knowledge shared by the communicators: in (6) it is known that the "making of the earth etc." happened in the remote past. The morphological criterion in this last case and in many other cases does not work: Classical Hebrew verb forms are often ambiguous in their time reference, as is the case with the form nokal 'we used to eat' (3), which, in other contexts, could be translated as 'we eat' or 'we shall eat'. The occurrences with prepositions m- 'from' (13 cases) and id 'until' (Gen 32:4 and 8 more cases) are clearly adverbial; here, this simple syntactic criterion of iatta's adverbial usage works successfully.
2.3. (wa)fatta as a discourse marker.
The most typical context of waiatta (and sometimes iatta without wa-) is when it has in-junctive verb forms in its immediate right co-text:
(7) wayyislah mal?akim ?al ?abimalak batorma lemor hinne gaTal ban Tabad we?ahaw ba?im sakema wdhinnam sarim ?at haTir Talaka waiatta qum layla ?atta wahaTam ?asar ?ittak wa?arob bassada
He sent messengers to Abimelech at Arumah, saying, [DU1] "Look, Gaal son of Ebed and his kinsfolk have come to Shechem, and they are stirring up the city against you. Now therefore, [DU2] go by night, you and the troops that are with you, and lie in wait in the fields (Jud 9:31f).
There are two most important issues to be noted here: first, waiatta appears on the border between two discourse units within direct speech (DU1, DU2); DU1 and DU2 are related, the text before waiatta being a background or motivation for the text after waiatta; second, the texts (discourse units) before waiatta and after it are different in their mood — indicative changes to imperative.
The relatedness of both discourse units is not expressed by any cohesive devices such as pronominal anaphora, it is suggested by the narrative context and may be revealed with the help of semantic analysis: the words 'city' and 'field' (i. e. open space around the city) are related in the mental world, shared by speakers of Classical Hebrew. But there is another type of cohesion between DU1 and DU2: the text immediately preceding (wa)iatta serves as a justification or motivation for the imperative utterance in DU2. In this and many other cases the text before (wa)iatta describes a state of affairs which is supposed to urge the addressee to undertake an action. Or it may be said that the speaker presents his request or order as justified or motivated by the preceding discourse unit19. In terms of RST, the rhetorical relation JUSTIFY is obtained between the two discourse units. It is for that reason that in many cases the adverb (wa)iatta is translated as 'therefore' or 'now therefore, now then'.
As has already been said above, DU1 and DU2 differ in their mood (indicative vs. imperative). For the analysis of (wa)iatta it is more important to note that they differ in their illo-
18 The discursive non-adverbial meaning may be present in these "temporally contrasted" contexts; on this problem see par. 2.3.
19 This type of a relation between two unites is described also as "resultative" [Müller 2005: 82].
cutionary force: DU1 assertive, DU2 directive. In fact, in the overwhelming majority of cases the text before (wa)iatta is illocutionary assertive, while the text after (wa)iatta is in most cases directive20. Therefore, (wa)iatta may be tentatively called a "speech act marker"21. One of the earlier analyzes of Hebrew waiatta comes to a similar conclusion: waiatta is "Illokutionsindi-kator" [Wagner 1997: 236].
Let us now look at the applicability of the main DM characteristics for this use of iatta. The connectivity is suggested by a rhetorical relation between two discourse units discussed above. Non-truth-conditionality is seen in that the removal of (wa)iatta does not affect the meaning of the imperative sentence: the slot of temporal adverbial is filled by layla 'by night'. As to the characteristics of initiality, iatta is posited at the beginning of the sentence, as indicated by the preposed sentence coordinator wa- 'and'. As far as the 'optionality' characteristics is concerned, we need to look at all the potential slots for (wa)iatta. Let us suppose that (wa)iatta appears mainly at the border between indicative DU (assertive illocution) and imperative DU (directive illocution), if they are related by the rhetorical relation JUSTIFY. This is a simplified procedure, not taking into account the more complicated contexts (e. g., with indirect illocutionary force), but it gives the idea of what I mean by the "potential slot" for a DM. All such potential slots were checked in the Hebrew text of the Old Testament Book of Judges and it was found that it includes 16 dialogic utterances (turns), built as a succession of discourse units assertive-directive, where the discourse unit before (wa)iatta is interpreted as a justification or motivation of the following directive22 (Jud 7:2; 9:38; 10:15; 11:36; 13:3f, 7; 14:2; 15:18; 16:10, 13; 18:14; 19:9(2), 30; 20:4-7, 12-13). Among these 16 potential slots for a discourse marker eight are filled with (wa)iatta (7:2; 13:3-4, 7; 14:2; 15:18; 16:10; 18:14; 20:12-13), the rest are unmarked. Here is one of the examples where the would-be slot for (wa)iatta is not filled:
(8) wattomar ?elaw ?abi pasita ?at pika ?al yhwh Tase li ka?asar yasa mippika ?ahare ?asar Tasa laka yhwh naqamot me?oyabaka mibbane Tammon
She replied, [DU1] "Father, you have made a promise to Yahweh; [DU2] treat me as the promise that you have made requires, since Yahweh has granted you vengeance on your enemies the Ammonites." (Jud 11:36; NJB)
Thus, from this rather short but representative number of examples it is seen that (wa)iatta may be omitted in that type of contexts where its appearance is expected, so its usage is optional. It should also be noted that there is no competing DM to fill the potential slots for (wa)iatta. These generalizations surely need to be checked on bigger amounts of text, but I believe that the rate of filled/unfilled slots will not change drastically after the investigation of the whole corpus of Classical Hebrew.
In this analysis of the most typical usage for (wa)iatta I attempted to show that the main DM characteristics are applicable here; (wa)iatta usually marks or cues a certain rhetorical relation (JUSTIFY) and is tightly related to directive utterances, expanded by the preceding as-
20 The directive illocution is expressed, besides imperative including prohibitive (134 times), also by jussive (16 times) and cohortative (7 times). According to A. Wagner [1997: 238], (w3)?atta is never used "vor einfachen Mitteilungen (repräsentativen)".
21 It is not a very widespread term, used, e. g., to describe the functions of English so [Müller 2005]. For French car and puisque as speech act markers see [Delort & Danlot 2005], following [Groupe A-l 1975]. It was noticed also by Helbig [1988] that "discourse particles function as illocutionary indicators" (quoted in [Fischer 2006: 437]).
22 In one of the cases (Jud 20:12) the rogative (question) has assertive force; in Jud 15:18 rogative has directive force. Both cases are interpreted in terms of indirect illocutionary force.
sertives, being the only Classical Hebrew DM with this specific function23. I admit that these characteristics of the discourse function of (wa)Satta are not exhaustive, e. g. I did not discuss some other functions and interpretations assigned to (wa)Satta and other now-words24, but the higlighted features are, in my opinion, most relevant to the present comparative investigation.
Concluding this short representation of (wa)Satta in its discursive usage, I shall discuss some of its controversial uses. There are certain contexts where waSatta appears as a temporal adverbial but its discursive interpretation is not to be excluded. Typologically, it is predictable and fairly well-known from works on German nun and English now [Perennec 2002: 342; Ai-jmer 2002: 59]. The problem is sometimes solved by saying that both meanings apply in such a case, but one of them prevails over another.
(9) udabar ?abner haya 'im ziqne yisra?el lemor gam tamol gam silsom hayyitam mabaqsim ?at dawid lamalak 'alekam waiatta 'asu ki yhwh ?amar ?al dawid lemor bayad dawid 'abdi hosia' ?at 'ammi yisra?el miyyad palistim umi-yyad kol ?oyabeham
Abner sent word to the elders of Israel, saying, "For some time past you have been seeking David as king over you. Now then bring it about; for the LORD has promised David: Through my servant David I will save my people Israel from the hand of the Philistines, and from all their enemies." (2 Sam 3:17f)
The contrast between the moment of speaking and the past appears highlighted here. Since Satta is here in the same extreme left position as in cases without pronounced temporal contrast (exemplified by (7)) where discourse function of Satta applies, it is not focused and it must not be integrated into the clause. If its temporal adverbial meaning were focused, it would be placed at the other end of the clause. On the contrary, the focused item is imperative. The question at stake is not when "to bring it about", but the necessity of the action itself is highlighted here.
2.4. wafatta in Ancient Hebrew letters.
This section may appear to be a curious addendum, since the material observed here is very scarse (21 occurrences of wSt)25, but its significance is increased by the fact that the bulk of Aramaic and Akkadian material discussed in this paper comes from letters. The corpus of Ancient Hebrew letters is very small: it includes 50 letters, only 20 among them in relatively good condition [Schwiderski 1997: 128]. The particle wSt is used very consistently in the letters, always marking the transition from the introductory part, usually containing the name of the addressee and greetings, to the body of the letter26:
(10) ?l ?lysb wit ntn lktym b 1 2 yyn l?rb't hymm w 300 lhm wml? hhmr yyn whsbt mhr ?l t?hr w?m 'wd hms wntt lhm [Ahituv & Mazar 1992: 56]
23 As shown in [Lyavdansky 2007], there is the following functional distribution for different inferential discourse markers in Biblical Hebrew: (wa)'atta marks directive utterances; laken marks commissives; ?epo marks ro-gatives (interrogatives).
24 The Hebrew (wa)'atta was also interpreted as "attention arouser" (Aufmerksamkeitserreger) by D. Schwiderski [1997] and in very similar terms by E. Jenni [1972].
25 Two texts found in the territory of Edom (Horvat Uzza ostracon) and Ammon (Tell Mazar ostracon) are included.
26 This usage is typologically significant and is paralleled by certain usages of English now: according to Hal-liday & Hasan [1976: 268]; quoted in [Aijmer 2002: 69], in 'a transaction situation such as a shop encounter, the transition from phatic communion to transactional relations is often made by now'.
To Eliyashib: And now — give to Kittiim 1 bat and 2 hins of wine for four days, and 300 [loaves of] bread, and a homer full of wine. You should send [it] out tomorrow, do not tarry. Also, if there is any vinegar left, give [it] to them.
The introductory part of a letter may also include a greeting/salutation formula:
(11) ?l ?dny ?lysb yhwh ys?l lslmk wit tn lsmryhw... [Ahituv & Mazar 1992: 74]
To my lord Eliyashib. May Yahweh ask for your peace! And now — give to Shemaryahu...
Lexical marking of the border between the introductory part of a letter and the body of the letter is a widespread phenomenon in Ancient Northwest Semitic epistolography. Thus, the use of wit in letters from Iron Age Judaea may be compared to similar use of Egyptian Aramaic kin/kint/kit, Mishnaic Hebrew s- and Hellenistic Aramaic d-. It is not to be excluded that the ultimate origin of this phenomenon is the Akkadian epistolary style, but a cursory look through different Akkadian letter corpora does not reveal any consistent usage of a lexical marker on the border between the introductory part of a letter and its body, as it happens in Ancient Hebrew and Egyptian Aramaic letters.
The function of wit in letters is similar to its function in dialog in at least two respects:
1. wit is inserted at the transition from the subsidiary part of the text of a letter (address, salutation) to its main part, traditionally called the "body" of the letter. In terms of RST analysis the introductory part of a letter is a satellite, whereas the body of a letter as a whole is a nucleus. The same terminology is applicable to the typical dialogic utterances hosting (wa)iatta: assertive DU1 is a satellite, directive DU2 is a nucleus.
2. In most of its attestations in letters wit marks a transition to an utterance with directive illocutionary force: in 9 cases, before imperative, in 6 cases — before absolute infinitive with the imperative function; together with one prohibitive there are 16 volitive utterances. The rest are 4 assertives with the perfect in the main clause and one commissive (promise). It may be argued that the function of wit in letters is to mark the transition from the introductory part to the body of the letter, irrespective of the illocution of the first utterance in a letter or of the body of a letter as a whole. It is hard to come to any definitive conclusion with this scarse material, but the statistical data given above support the comparison of wit in letters with (wa)iatta in literary dialog.
Consequently, the epistolary usage of (wa)iatta is in line with its usage in dialogue. In my opinion, the genealogy of the usage of (wa)iatta in Classical Hebrew may be presented as follows it is born in spoken interaction and all the other uses (literary dialogue, epistolary usage, liturgical and prophetic poetry) derive from it.
It is beyond the scope of this paper to discuss the uses of (wa)iatta in Psalms and Prophets; let me just refer to the observations of Jenni [1972] who, not contradicting my own findings, did not notice any major deviations from the picture drawn above.
3.1. Egyptian Aramaic kin / kint / kit.
There are three particles with the meaning "now" in Egyptian Aramaic27 which are almost identical in meaning and function and often interchange in the same contexts: kin, kint and kit.
27 The term "Egyptian Aramaic" refers to the language of a relatively large corpus of Aramaic texts from Egypt as represented in TAD, dating from seventh to third century BCE, but the bulk of this material is dated
The nature of this variation is not altogether clear; the attempt to understand it, undertaken in [Folmer 1995: 661-71], is full of interesting observations as to the distribution of kSn, kSnt and kSt in different epistolary archives, groups of texts and in the notation of different scribes, but it did not reach any definitive overall conclusions. One thing that these three expressions all have in common is that they are construed with the preposition k- attached to three different (but probably related) words: k-Sn, k-Snt and k-St. It is not a common opinion that these three words are etymologically related28, but their usage, which will be discussed below, points to that possibility. According to Ribera i Florit [1983], the common etymology of these three expressions may be shown as follows:
*Sadtu/Sidtu > *Sattu > St
*Sadtu/Sidtu > *Sattu > *Santu > Snt > Sn
In light of their usage and probable common etymology, the expressions kSn, kSnt and kSt are treated as allomorphs29. There are also variants kSn / w-kn, kSnt / w-kSnt and kSt / w-kSt. The uses of kSt and its allomorphs with preceding w- are considered as collocations (see 2.1).
3.2. Adverbial kfn.
My observations on the 164 attestations of kSt and its allomorphs in the corpus of Egyptian Aramaic [Porten & Lund 2002] show that there are only 9 clear instances where the Aramaic word for "now" is used adverbially30, and in all of these instances only kSn is attested. They are interpreted as adverbial based on criteria applied to the analysis of (wa)Satta above: in 6 cases kSn appears within a prepositional phrase ( Sd kSn A4.3:7; D2.29:1; D7.19:5, 7; mn zy kSn A4.7:3; A4.8:2); in 2 cases there is a contrast with the past state of affairs (B3.8:41; B6.4:7). In one more case (A6.4:3) kSn is found in a relative clause:
(12) k'n ps]msh brh zy 'hhpy zy kin pqyd 'bd hlpwhy byn bgy? zyly zy b'lyt? wthtyt? s?l lmns? dsn? zky z[y] mn mlk? wmny [y]hb l'hhpy
Now, Psamshek the son of Ahhapi who now has been made an official in his stead in my domains which are in Upper [and Lower (Egypt) asked to carry on] that grant which was given by the king and by me to Ahhapi (TAD A6.4:3-4).
Beyond intuitive contextual considerations, which suggest an adverbial function for kSn in this context, it should be noted that, according to the principles of RST, restricted relative clauses are not considered separate discourse units. In the above example we, however, have an unrestricted relative clause. The problem is solved by the observation that DMs like (wa)Satta and now, with which kSn is comparable, always mark a transition to nucleus. It is natural to assume that the relative clause never constitutes a nucleus. Therefore, I suppose that the appearance of kSn in any relative clause forbids its interpretation as a discourse particle.
around the fifth century BCE. The corpus of Egyptian Aramaic is fairly representative for the larger linguistic entity usually designated as Imperial Aramaic, or Achaemenid Aramaic.
28 On this problem see the discussion and references in [Folmer 1995].
29 See, e. g., [Schwiderski 1997: 132].
30 TAD A4.3:7; A4.7:3; A4.8:2; A6.4:4; B3.8:41; B6.4:7; D2.29:1; D7.19:5, 7. Here and below the indices for Aramaic texts are given according to the TAD edition, divided into four thematic volumes: A — Letters; B — Contracts; C — Literature, Accounts, Lists; D — Ostraca & Assorted Inscriptions.
The list of temporal adverbial uses of kin may be expanded by one more example (TAD C1.1:51), discussed below, where both basic functions — adverbial and discursive — are probably concomitant.
3.3. Egyptian Aramaic kit/kint/kin in letters.
There are approximately 150 attestations of kit and its allomorphs in Aramaic letters included into the corpus of Egyptian Aramaic as represented in TAD. In 60 cases it is placed at the beginning of the body of the letter. Most of the other attestations (excluding 7 clearly adverbial uses) are clause-initial uses of kit/ kint/kin, where it marks a transition from one paragraph of a letter to another, thus having a connective function in the text. For the present comparative investigation it is relevant to note that in 41 cases it marks the transition to a directive utterance; only 5 among them open the body of a letter. These usages are exemplified by the following two texts from different letter corpora:
(13) slm ?wryh kin hlw t?t?zylk rbt? mt?t lmgz Tmr?zylh qdm? mtmrt bkb? kin ?t? wgzh bywm zy tr hmnh tgznh... Greetings, Uriyah! Now your big ewe is ready for shearing. The one you sent over before is being combed. So you can come shear her whenever you please (TAD D7.8:1).
(14) mn wrwhy Tl nhthwr wkndsyrm wknwth wkit tnh ?nh qblt l?rsm Tl ?htbsty pqyd? zyly zy m[nd]t[?] mndTm l? mhyth ly ?h[r ] t mhytyn bb[?l] kit ?ntm ?tnsh[w] whndrzw Tbdw lpqyd? [zy] ly Td mndt [bgy? ?lk yhy]th Tly bb?l From Varuvahya to Nakhthor and Kendasirama and his colleagues. And now, I complained here to Ar-sames about Ahatubasti my official who is not bringing me anything of [the] r[en]t. The[n...] ... they are bringing to Baby[lon]. N[o]w, you, be diligen[t] and issue instruction to [m]y official that he [bri]ng to me to Babylon the rent of [those domains] (TAD A6.14:1-3).
Naturally, there are some local and register-related peculiarities in the usage of kit. The letters from Hermopolis apparently demonstrate a deviation from the more widespread usage. I adduce a rather lengthy example, omitting the original Aramaic text; not a single case of wkit has any equivalent in the translation:
(15) Greetings to the temple of Bethel and the temple of Queen of Heaven. To my sister Nanaiham from your brother Nabusha. I bless you by Ptah — may he let me see you again in good health! Greetings to Bethelnetan. Greetings to Nikkai, Asah, Tashai, Anati, Ati, and Reia.
wkit The tunic you sent me has arrived. I found it all streaked; I just don't like it at all! Do you have
plenty of other kinds? If I knew, I would exchange it for a dress for Ati.
wkit As to the tunic which you brought for me to Syene, I wear it.
wkit Please have some castor oil sent to us, so we can exchange it for olive oil.
wkit Don't worry about me and Makkebanit; let us worry about you instead! Take care of Bethelnetan; keep Habib away from him!
wkit If I can find anyone dependable, I will send you something......(TAD A2.3)
The paragraphs/units that are introduced by wkit are pragmatically of different nature — constative (the paragraph about the tunic), directive (request to send castor oil), commissive (a promise to send something). But what is more important here is that the paragraphs marked by wkit are not understood as being explicitly related to each other. Only the second wkit marks a switch to sub-topic within the paragraph; it is, apparently, a case of the rhetorical re-
lation elaboration in terms of RST. Nevertheless, there is still cohesion in this letter, which we may call "global cohesion": explicitly unrelated discourse units are related by having the same speaker and the same speech situation.
Thus, in the letters from Hermopolis the particle wkSt becomes the default transition marker. It does not matter whether there is any specific discourse relation between the paragraphs of a letter or not; what matters is that every other unit introduced by wkSt represents a new topic in the broadest sense of this word.
3.4. kfn in dialogue.
Naturally, it is problematic to discuss the dialogic usage in a dead language, but it is possible with certain restrictions. Within strictly Egyptian Aramaic material we have only rare instances of reported speech in letters and in the Story of Ahiqar. To these scarse data we may add literary texts, written in the idiom traditionally called Biblical Aramaic (BA). Since BA in many respects follows Egyptian Aramaic31, the Aramaic passages from Daniel and Ezra can be also included into the present discussion.
In the Aramaic Story of Ahiqar32 kSn appears in a context that is typical for a discourse marker, before directive utterance (the Aramaic text is omitted for the sake of brevity):
(16) I am Ahiqar who formerly rescued you from an innocent killing... I brought you to the house of mine. There, I was supporting you as a man with his brother, and I hid you from him. I said, "I killed him," until at [an]other time and many days later I presented you before Sennacherib the King and I removed your sins before him and evil he did not do to you. Moreover, abundantly Sennacherib the King loved me because I let you live and did not kill you.
Now (kin), you, just as I did for you, so, then (?pw), do for me. Do not kill me. Bring me to your ho[u]se un[til] later days (TAD C 1.1:46-52).
The temporal adverbial function of kSn is probably concomitant here with its textual function. There is one more discourse marker (?pw) here; it occurs only twice in Egyptian Aramaic, but in its meaning it follows Biblical Hebrew ?epo. Its appearance in the same sentence as kSn may affect the interpretation of the meaning of the latter, because if kSn and ?pw mark the same rhetorical relation, then the discourse function of kSn will appear redundant.
The text of the following short letter, written on an ostracon, is not without problems, but it is unique, because it includes reported speech that demonstrates one of the rare examples of kSn in dialogue:
(17) 'l hgy ?mrt l?sn ?l ksp mrzh? kn ?mr ly lm l?yty33 kin ?ntnnh lhgy ?w ygdl dbr 'lwhy wyntnhy lkm
To Haggai: I talked to Ashina about the money for the marzeah society. He told me " [If?] there is not, so I will give it to Haggai or to Yigdal." So go see him and get him to give it to you!34 (tad D7-29)
The supposed scenario behind this implies that Ashina is a sponsoring agent for the ritual communal banquet (mrzh?); seeing that there is no (l?yty) money for it, he promises the author
31 Sometimes Biblical Aramaic is included into Imperial Aramaic [Beyer 1986].
32 On the peculiar dialect of this text see [Kottsiepper 1990].
33 RES: l?ytw, this was understood as personal name Ito (see [Lindenbrger 1993: 39]), which is hardly plausible here.
34 Translation follows [Lindenberger 1993: 39], modified according to the emended text in [TAD 4: 177].
of the letter to give the needed money to Haggai or to Yigdal. The older interpretation 'to Ito' (l?yty) would, naturally, not destroy this scenario but would not support the inferential meaning of kin in this context. This reading, which runs into certain problems, was rejected in the newer edition of the text (TAD 4: 177).
The following pragmatically complicated example from Biblical Aramaic cannot be discussed at length here, but it demonstrates some typical problems and ways to overcome them:
(18) Nebuchadnezzar said to them, "Is it true, O Shadrach, Meshach, and Abednego, that you do not serve my gods and you do not worship the golden statue that I have set up? Now (kTn) if you are ready when you hear the sound of the horn, pipe, lyre, trigon, harp, drum, and entire musical ensemble to fall down and worship the statue that I have made, well and good. But if you do not worship, you shall immediately be thrown into a furnace of blazing fire, and who is the god that will deliver you out of my hands?" (Dan 3:13-15)
The textual units before kin and after kin are difficult to interpret pragmatically. I believe that, in this case, the interrogative in the left co-text of kin may be interpreted as an indirect assertive, because Nebuchadnezzar does not seem to expect an answer to his rhetorical question. The immediate right co-text of kin may be interpreted as a directive, because the purpose of the whole utterance is to urge three young men to worship Nebuchadnezzar's idol.
3.5. Results for Aramaic.
It should be taken into account that, generically, the discussed Aramaic textes are different in comparison with Classical Hebrew: they are mostly letters, and examples from reported speech are quite few. If we consider only letters, there are many examples of directives in the right co-text of kit: among 83 inner-body uses of this particle 36 have a directive in said position. The relatively low percentage of directive uses is explained by the observation that kit, at least in some subcorpora (Hermopolis), assumed a new generalized function: it could mark every turn (paragraph) in the letter, irrespective of the type of rhetorical relation. This new function of kit implies that it appears not only at satellite-nucleus junctures, but also at nucleus-nucleus junctures.
The function of kit in the text may be seen from a new perspective if it is compared to and contrasted with other discourse particles in Egyptian Aramaic, e. g. ?p, hlw, h?. This work is yet to be done, but it may be preliminarily noted that hlw and h?, contrary to kit/kint/kin, are used mainly with the indicative (assertive) in the right co-text, which will be relevant for the discussion of Akkadian anumma and inanna.
4.1. Akkadian words with the meaning "now".
Let me start with a quotation from [Moran 2003: 16], discussing particles in the Amarna letters from Byblos: "Note also the phrase u inanna in 102.24-28 and compare Heb. weiattah, where "and now" is not temporal but interjectional."
The term 'interjectional' may look obsolete, but the remark as a whole is right to the point, because the "temporal" function of inanna is here taken for granted, and attention is drawn primarily to its non-adverbial usage.
If we look at what standard descriptions of Akkadian tell us about temporal adverbs, we find that there are two adverbs with the meaning "now": inanna and anumma. CAD, together with many other descriptive sources, suggests that both are used as connectors or discourse particles: the "introduce topic of a letter" [CAD 7: 144]. Thus, the situation in Akkadian looks perplexing: we have two words for "now" which are probably competing not only for temporal adverbial slots, but also for discursive slots. But things like that never happen: there must be some rule(s) of distribution for these two words. In fact, one of the recent treatments of inanna together with anumma by Loesov [2004] forcefully draws these two words apart: as to the core meaning, for inanna it is "now", for anumma it is "here". It is true that the typical adverbial usages for inanna may be easily shown: inanna is used in such prepositional phrases as adi inanna "until now", istu inanna "from now on", but anumma, apparently, never appears in such contexts (e. g., all the examples in CAD are sentence-initial). Thus, it appears that — if we keep the translation "now" for lexicological purposes — inanna is used in both senses, 'now I' (adverbial) and 'now II' (non-adverbial, discursive), while anumma is used only as 'now II'.
Let us now look a little closer at some of the uses of inanna and anumma.
4.2. Akkadian anumma in letters.
The description of anumma in CAD A2 is interpreted as follows: there is one specific usage of anumma in letters ("used to introduce the message, its bearer and what he brings") in different corpora and periods of Akkadian, and there are many other usages that are not classified. All the examples in CAD show that anumma is inserted at the beginning of a sentence; it allows to suppose that it may function (or, probably, always functions) as a discourse particle. The remark in [Huehnergard 1989: 195] that anumma is "a sentence-modifying adverb that introduces a new thought" and the findings of Rainey [1988] for Amarna and Loesov [2004] for Old Babylonian letters from Harmal support this supposition.
Taking into account one of the functions of Aramaic kSt and Hebrew wSt — to introduce the body of the letter — it is tempting to also find such a device in Akkadian, and anumma is one of the probable fillers of this slot. First of all, I must say that the situation with this slot in Akkadian is different when compared with Classical Hebrew and Biblical Aramaic: no one of the Akkadian letter corpora that I have looked through (core Old Babylonian, OB Mari, Amarna, Neo-Assyrian) have a regular lexical device to mark the transition from the introductory part to the body of the letter. The assertion found in CAD that "the Mari letters use anumma to introduce the first topic of a letter" is not altogether wrong (see ARM 5 5:4, 9:4, 13:5, 41:4, 78:5; ARM 4 3:5), but it must not be understood in the sense of the above discussion; in Mari letters the first topic is introduced sometimes by sanitam, sometimes by assum and very often does not have any specific lexeme at the transition to the body of the letter. A similar picture is found in the rest of the letter corpora mentioned above.
If we turn to instances of anumma inside the body of the letter, we find that, most often, it introduces an assertive utterance with indicative verb forms in the main sentence. Looking a little beyond the immediate right context, we find that anumma quite often appears before in-junctive utterances, but in all of these cases anumma is separated from the injunctive by an assertive sentence.
In the letters of Hammurapi, analyzed by Sallaberger, among 170 letters with a directive utterance (ANORDNUNG) in their nucleus there are 30 where anumma is detached from an injunctive by one sentence or clause, which may be considered a discourse unit. Here is one of these letters:
(19) [a-na] dEN.ZU—i-din-nam [qi]-bi-ma um-ma ha-am-mu—ra-bi-ma as-sum p dEN.ZU—ra-bi sa it-ti mu-ur— ist[ar] ta-at-ru-da-as-su p dEN.ZU—ra-bi su-a-ti a-na ma-ah-ri-ia u-se-ri-bu-nim-ma as-sum i-din-dEN.ZU u-lam-mi-da-an-ni a-nu-um-ma dEN.ZU-ra-bi su-a-ti a-na se-ri-ka at-tar-dam p dE[N.Z]U u Msi-i-bi sa i-qa-ab-bu-kum a-na ma-ah-ri-ia tu-ur-dam
Zu Sin-iddinam sprich: also (sagt) Hammu-rabi: Was den Sin-rabi betrifft, den du mit Nür-Istar zu mir geschickt hast, (so) hat man diesen Sin-rabi vor mich geführt und er hat mir betreffs Iddin-Sin berichtet. Hiermit schicke ich jetzt den betreffenden Sin-rabbi zu dir. Schicke den Iddin-Sin und die Zeugen, die er dir nennen wird, zu mir. (AbB II 2)
It is impossible to see from just one example, yet it appears that this usage is close to formulaic, because the sentence immediately following anumma in these contexts is restricted as to its content and lexicon [Sallaberger 1999: 146]. From the discussion in [Sallaberger 1999: 146f] it is seen that the DU introduced with anumma (Sallaberger calls it "Initiative") is dependent on the following directive DU35; thus, they are related as satellite and nucleus. It is also clear that anumma marks the transition from the informative part of the letter ("Informieren") to "Initiative". This metatextual (= discourse) function of anumma is discussed in [Loesov 2004] and I do not go into it here, although it is also relevant for the discussion.
The corpus of Akkadian letters from Mari from the time just before Hammurapi (17921750 BCE), or coterminous with his rule, demonstrates another type of formulaic or quasi-formulaic usage of anumma in accounts of prophecies that are found in letters. Usually at the very end of the letter there appears a phrase, introduced with anumma and often followed by the injunctive:
(20) sinnistum si annetim idbubamma [aw]at pisa ana beliya aspuram anumma sarassa u sissiktasa ana beliya usabi-lam beli teretim lisepisma ana ki ilum beli ippalu li[p]us
This is what this woman said, and I have written her [wor]ds to my lord. I have herewith sent her hair and a fringe of her garment to my lord. My lord should let oracles be taken. Let my lord act according to what the god answers. (Nissinen 2003 27:17-31 = ARM 27 217:27-31)
The scenario behind these phrases implies that the author of the letter attaches prophet's hair and garment "to be used as representing the prophet during the process of authenticating the prophecy by extispicy ("oracles" in our letter — A.L.)" [Nissinen 2003: 16]. Thus, the assertive after anumma is related to the following injunctive, and the relation between these two DUs may be interpreted as JUSTIFY. Among 50 prophetic letters from Mari adduced by Nissinen there are 11 letters where anumma appears in the described type of context.
Both groups of contexts — the letters of Hammurapi and prophetic letters from Mari — are quasi-formulaic, but they are different in their Sitz im Leben and are not dependent on each other; therefore, I conclude that they are based on free or 'natural', non-formulaic usage of anumma.
There are also some anumma + Perfekt contexts, where its usage goes beyond formulaic, because the phrase after anumma speaks about the actual affair (Sallaberger's "Thema") of the letter and is therefore intimately related to the folowing imperative (e. g. AbB XIII, 10).
Note that in the above examples the transition from assertive to injunctive (directive) was never marked. This does not always happen in the Akkadian directive utterances, as will be seen in the following paragraph. It was important to show that anumma in that sort of contexts does not appear in the slot filled by Hebrew wTt and in many cases by Aramaic kit; it appears in
35 "Dieses Teil spielt als ... Übergang vom Informations- zum Aufforderungsteil" [Sallaberger 1999: 146].
the slot filled in Hebrew directive utterances by hinne (Jud 16:10; 19:9; 20:7), in Aramaic sometimes by hlw.
4.3. Akkadian inanna in reported speech and in letters.
A very detailed discussion of inanna in Old Babylonian (OB) by Loesov [2004] seems to me sufficient to demonstrate the various discourse functions of this temporal adverb, though the approach to the description of inanna chosen by Loesov is different from mine. For Loesov's analysis, it was the consideration of concomitant verb forms that served as the decisive factor. I admit that the shift from one type of verb form to another is significant for the discourse structure. At the present stage of my research I do not take into account the shifts within the indicative sphere; I am interested in usages which mark (or are somehow related to) the shift in the verbal mood or in the illocutionary force of the utterance.
If we look at the uses of inanna in reported speech, we find that it mostly appears at the transition from one type of speech act to another. I have checked all the attestations of inanna (eninna) in the Standard Babylonian (SB) version of The Epic of Gilgamesh. The "corpus" of reported speech in The Epic of Gilgamesh is not big, but in relation to all reported speech that may be found in OB or SB literature it is hardly small. According to the edition of A. George [2003], there are 16 unrestored attestations of eninna in the SB version of The Epic of Gil-gamesh, plus one in the Ischchaly tablet. In the majority of cases eninna marks a shift in illocution: 5 times from assertive or interrogative36 to directive (I 96; V 102; V 180= V 238; XI 198), 3 times from assertive to interrogative (VII 59; VIII 55; X 73, 150; XI 207), one time from assertive to declarative (III 122), once from interrogative to assertive (III 47), once from interrogative to commissive37 (VII 139). 3 times it is turn-initial (Ischchaly 11 — directive; X 73, 150 — both interrogatives). Note the slight prevalence of directives in the right co-text; however, the data are too limited to reach any definitive conclusions. There are more attestations of inanna before directives among the examples from OB letters, discussed in [Loesov 2004].
The usage of eninna before directives may be demonstrated by the following text with the injunctive form in the immediate right co-text:
(21) at-ti da-ru-ru tab-ni-[i amela (lu)] [e]-nin-na bi-ni-i zi-kir-su
'You, O Aruru, cteated [man:] now create what he suggests! (I 95f)
A more complicated case is represented by the following passage, where the injunctive verb forms are not in the immediate right co-text of eninna:
(22) 100 am-mi-ni ib-ri pi-is-nu-qis [ta-qa]b-bi
101 u pi-i-ka ir-ma-am-ma tu-lam-[man l]ib-bi
102 e-nin-na-ma ib-ri is-ta-at [(x)]-pi?-[x]
103 ina ra-atu SIMUG e-ra-(a) sd-ba-sd-a
104 tu-u-ru ana 1 DANNA"m na-pa-hu na-pi-ih-tu ana 1 DANNA"m x-7u-u"
105 sd-par a-bu-bu is-tuh-hu la-pa-tu
106 [e] "ta^-as-suh GIRmin-ka e ta-tu-ur ana dr-ki-ka
36 Interrogatives are usually not included within the classification of Speech Acts [Allan 1998]. They need further analysis, e. g. rhetorical questions are often assertives or directives.
37 The phrases in the following passage (140-147) constitute a promise to Enkidu.
107 [.......]x x x mi-hi-is-ka du-un-nin
100 'Why, my friend, [do you] speak like a weakling?
101 With your spineless words you [ make] me despondent.
102 'Now, my friend, but one is [our task,]
103 the copper is already pouring into the mould!
104 To stoke the furnace for an hour? To ... the coals for an hour?
105 To send the Deluge is to crack the whip!
106 '[Don't] draw back, don't make a retreat!
10 7......make your blow mighty!' (V 100-107)
If it is possible to say that a DM is hosted by an utterance or a discourse unit, it should be noted that a discourse unit may include from one up to several clauses or discourse units of lower rank. In the above text, eninna marks a transition to that part of the text where directive force is dominating; in terms of RST, the passage (106-107) with injunctives is a nucleus with two satellites — (102) and (104-105)38. Thus, eninna focuses the hearer's attention upon the whole passage (102-107), organized as a complex directive utterance.
It is also to be noted that assertives are rare in the right co-text of eninna in reported speech, as represented by the SB version of Gilgamesh. The significant amount of interrogatives draws our attention in this case, because they are rare in the right co-text of Hebrew (wd)Satta and Aramaic kSt. Apparently, the functions of inanna are broader compared to the now-words in Hebrew39 and Aramaic.
To conclude with inanna, let me briefly discuss the analysis of its meaning suggested in [Loesov 2004]. I quote the results of this analysis in a slightly shortened manner, omitting the references to examples within Loesov's article:
These and related data (presented below) permit one to posit two inanna-lexemes:
inannaA: temporal deictic adverb, further subdivided into
inannaAl pointing to the moment of speaking, used with the Present, employed in non-future sense and with the semantically "present-tense" Stative: ... (speaker-orientation);
inannaA2 pointing to the moment of speaking; it is used with the Perfect and locates its resultative component: . (speaker-orientation);
inannaA3 pointing to the future, used with injunctive forms and with E[pistolary] P[er]f[ect]: ... (addressee-orientation).
inannaB: metatextual "particle" marking a turn in discourse, i. e. a means of discourse deixis. It is formally set apart from inannaA through the combination of two features: inannaB is used only with the Preterite and, unlike inannaA is incompatible with injunctive utterances in its immediate right context/co-text. [Loesov 2004: 96]
As stated by the author, one of the principles of this classification is "to illuminate the verb usage". I am not against the postulation of two distinct inanna-lexemes, but the distinction between inannaA3 and inannab seems to be exaggerated, because inannaA3 also marks "a turn in discourse". Moreover, inannaB is also orientated to the addressee: if now is metatextual (now II), it is that type of now, which is shared by the speaker and the addressee, denoting textual time which is common to both participants of the communicative act.
38 These are proverbial phrases which serve to enhance the rhetorical effect of Enkidu's speech as the latter tries to urge a frightened Gilgamesh to "swift action" [George 2003: 467].
39 Interrogatives related to the foregoing assertives by the relation JUSTIFY are sometimes marked in Hebrew by ?epo.
4.4. Results for Akkadian.
According to my observations, there are components of meaning that are common to both inanna and anumma, and their frequent appearance together (the collocations inanna anumma or anumma inanna) is one of the facts that prove this. Both anumma and inanna draw the attention of the addressee to what happens in the text. Or, in the words of Deborah Shif-frin (said about English now), they "focus attention on what the speaker is about to do" [Schif-frin 1987: 241]. Another side of this function is that they mark a turn in the discourse, a turn which is marked also by a shift from one verb form to another, with or without a shift in verbal mood. When inanna or anumma introduce an assertive utterance, the difference between them is yet to be clarified40. There are certain contexts where anumma would not usually appear (purely temporal adverbial usages; immediately before directives, interrogatives and commissives). The comparison of anumma to Hebrew hinne is probably restricted to just one type of contexts, where anumma is in the vicinity of a directive DU but is separated from it by an assertive DU. The percentage of such uses for anumma should be checked on a wider textual basis, but it is unlikely to be small. The comparison with Hebrew hinne may also be taken into consideration when solving the problem of the origins of anumma. As to the similarity in usage between Hebrew (wd)^atta and Akkadian inanna (examples from reported speech), I think it is possible to interpret it as a parallel development from a temporal deictic adverb to a discourse particle, which marks a transition from one illocution (usually assertive) to another (usually directive) for discourse units related to each other by the rhetorical relation JUSTIFY. This comparison is made with one reservation: not all instances of inanna are comparable to Hebrew (wd)latta.
5. Conclusion.
As the above investigation has shown, there are some common developments in the usage of Hebrew (wd)Satta, Aramaic kltlklntl k'tn and Akkadian inanna. Used as temporal deictic adverbs with an extralinguistic reference to the present time, they are more often attested in contexts where their adverbial meaning is bleached, combined with discursive (metatextual) meaning or does not apply at all, ceding its place to their discursive meaning. The aforementioned 'now-words' in three Semitic languages are thus included into the class of words with a salient discourse function, usually called 'discourse markers'.
Triggered by the specific usage of Hebrew (wd)latta, the focus of this study primarily concerned directive utterances. It was noticed quite a long time ago by Teun van Dijk [1979], one of the "fathers" of discourse studies in Europe, that directive utterances are often preceded by assertives, helping the speaker to fulfill her/his communicative goals, e. g. to make a request/demand/ more acceptable to the addressee. Most often, the assertive discourse unit is placed before the directive discourse unit; transition from assertive to directive constitutes the slot for a discourse marker. Languages vary not only as to the type of lexemes that can fill this slot, but also as to the percentage of slots filled with a DM. Thus it appears that among the three discussed languages this slot is filled most frequently in Classical Hebrew with (wd)'iatta, less often in Egyptian Aramaic with k'it and (sometimes) in some varieties of Akkadian with inanna. In all three languages there are no lexemes that could be considered as really competing for this slot: the lexemes at the focus of this study are the main fillers of the slot for the di-
40 See the discussion and references in [Loesov 2004], an important step in this direction.
rective speech act marker. This type of contexts was taken as the basis and the starting point for the comparison of the discussed 'now-words'.
An attempt was made to demonstrate in what respect the usage of these words is different. Thus, in Egyptian Aramaic the temporal adverb kSt/kSnt/ kSn assumes specific functions in letters: it is used as a transition marker to the body of the letter41, in some of the corpora it is used as the marker of a new topic (paragraph), irrespective of the type of rhetorical relation between paragraphs. In Standard Babylonian inanna often marks transition from different illocutions to interrogatives, and there are also uses of inanna before commissive utterances and declarations42. In general, the usage of inanna is significantly broader than that of (wd)Satta and kSt/kSnt/ kSn as far as the existing evidence shows us: inanna very often appears before asser-tives, e. g. in the narrative part of letters. The distribution of inanna and anumma, both appearing before assertives, is not altogether clear; at present it appears that anumma, sharing with inanna only the discursive meaning "now" (i.e. now II), is a different type of lexeme. It is never used as a temporal deictic adverb and may be compared in its usage to Biblical Hebrew hinne and Egyptian Aramaic hlw and h?.
Finally, I would like to define the field for further research. The discussed discourse markers in Hebrew, Aramaic and Akkadian should be considered as members of their class; their function may be further clarified by juxtaposing them with other DMs such as hinne, laken, yaSan and ki in Classical Hebrew. The use of discourse particles in Aramaic letters is a promising field of study, especially because it may be compared to their use in Akkadian letters, with which there are clear parallels [Fales 1987].
Literature
AbB — Altbabylonische Briefe (series).
ADP — Fischer, K. (ed.) (2006) Approaches to Discourse Particles. Amsterdam: Elsevier. ARM — Archives royales de Mari (series).
Ahituv & Mazar 1992 — Ahituv Sh. & Mazar, A. (eds.) (1992) Handbook of Ancient Hebrew Inscriptions. Jeruasalem:
The Bialik Institute (Mosad Bialik). Aijmer 1988 — Aijmer, K. (1988). Now may we have a word on this: the use of now as a discourse particle //
Kyto M. et al. (ed.) Corpus linguistics: hard and soft. Amsterdam and Atlanta: Rodopi. Aijmer 2002 — Aijmer, Karin. 2002. English Discourse Particles: Evidence from a Corpus. Amsterdam / Philadelphia: John Benjamins.
Allan 1998 — Allan, K. Speech Act Classification and Definition // J.L. Mey (Ed), Concise. Encyclopedia of Pragmatics. Amsterdam: Elsevier, pp. 922-927. Bayerl 2004 — BAYERL P.S. Disambiguierung deutschsprachiger Diskursmarker: Eine Pilot-Studie // Linguistik online 18.1 (2004).
Beyer 1986 — Beyer, K. (1986) The Aramaic Language. Its Distribution and Subdivisions. Gottingen: Vandenhoek & Ruprecht.
Brongers 1965 — Brongers, H.A. (1965). Bemerkungen zum Gebrauch des adverbialen we'atta im Alten Testament // Vetus Testamentum 15(3).
Brunaud 1991 — Brunaud, N. (1991) "Now" et "then" dans la coherence argumentative // Langages 104, p. 79-91. Delort & Danlot 2005 — Delort, L. & Danlot, L. (2005). Coordination of Causal Relations in Discourse // Aurnague, M.; Bras, M.; le Draoulec, A. & Vieu, L. (eds.), Proceedings of Simposium on the Exploration and Modelling of Meaning (SEM'05). Biarritz, p. 75-84. Fales 1987— Fales F.M. Aramaic Letters and Neo-Assyrian Letters: Philological and Methodological Notes // JAOS 107.3 (1987).
41 Also attested for Hebrew (ws)fatta, but the evidence is too sparse.
42 Sometimes attested for Classical Hebrew (wa)fatta.
Fischer 2006 — Fischer K. (2006). Frames, constructions, and invariant meanings: the functional polysemy of discourse particles // Fischer, K. (ed.) Approaches to Discourse Particles. Amsterdam: Elsevier.
Folmer 1995 — Folmer, M.N. (1995). The Aramaic Language in the Achaemenid Period: A Study in Linguistic Variation (Oreintalia Lovaniensia Analecta 68). Leuven.
Fraser 2006 — Fraser, B. (2006) Towards a Theory of Discourse Markers // K. Fischer (ed.), Approaches to Discourse Particles. Amsterdam: Elsevier.
George 2003 — George, A.R. (2003). The Babylonian Gilgamesh epic: introduction, critical edition and cuneiform texts. Oxford University Press.
Groupe A-l 1975 — Groupe A-l. Car, parce que, puisque // Revue Romane, 10.2, p. 248-280.
Huehnergard 1989 — Huehnergard, J., (1989). The Akkadian of Ugarit (Harvard Semitic Studies 34). Atlanta: Scholars Press.
Jenni 1972 — Jenni, E. (1972). Zur Verwendung von 'attä "jetzt" im Alten Testament // Theologische Zeitschrift 28, pp. 5-12.
Kottsieper 1990 — Kottsieper, I. (1990). Die Sprache der Ahiqarsprüche (BZAW 194), Berlin — New York.
Lande 1949 — Lande, I. (1949). Formelhafte Wendungen der Umgangssprache im Alten Testament. Leiden: Brill.
Laurentin 1964 — Laurentin, A. (1964). We'attah — kai nun. Formule charactérestique des textes juridiques et liturgiques à propos de Jean 17,5 // Biblica 45, pp. 168-97.
Lindenberger 1993 — Lindenberger, J. (1993). Ancient Aramaic and Hebrew Letters (SBL Writings from the Ancient World Series 4), Atlanta: Society of Biblical Literature.
Loesov 2004 — Loesov, S. (2004). T-Perfect in Old Babylonian: The Debate and a Thesis // Babel & Bibel 1.
Lyavdansky 2007 — Lyavdansky, A. Inferential Markers in Biblical Hebrew, presented in the Linguistics and Biblical Hebrew Section at the Society of Biblical Literature Internetional Meeting, July 2007, Vienna.
Mann & Thompson 1988 — MANN, B. & THOMPSON, S. (1988). Rhetorical Structure Theory: Toward a functional theory of text organization // Text, 8(3), pp. 243-281.
Miller 1999 — Miller, C. (1999) The Pragmatics of waw as a Discourse Marker in Biblical Hebrew Dialogue. Zeitschrift für Althebraistik 12,2, pp. 165-191.
Moran 2003 — Moran, W. (2003) Amarna Studies // Collected Writings (Harvard Semitic Studies 54). Winona Lake, Ind. : Eisenbrauns.
Müller 2005 — Müller, S.(2005). Discourse Markers in Native and Non-native English Discourse (Pragmatics & Beyond New Series 138). Amsterdam-Philadelpia: John Benjamins Publishing House.
Muraoka 1985— Muraoka, T. (1985). Emphatic Words and Structures in Biblical Hebrew. Jerusalem-Leiden.
NJB — New Jerusalem Bible.
Nissinen 2003 — NissiNEN, M. (2003). Prophets and Prophecy in the Ancient Near East (Writings from the Ancient World Society of Biblical Literature — WAWSBL 12). Atlanta: Society of Biblical Literature.
Pérennec 2002 — PÉRENNEC, M. (2002). Von Zeitdeiktika zu Text- und Diskurskonnektoren. Überlegungen zur sprachlichen Tämporalität // Pérrenc, M., Sur le text: énonciation et mots du discours en allemand. Presses Universitaires de Lyon, pp. 339-356.
Porten & Lund 2002 — Porten, B. & Lund, J.A. (2002) Aramaic Documents from Egypt: A Key-Word-In-Context Concordance. Winona Lake: Eisenbrauns.
Rainey 1988 — Rainey, A.F. (1988). Some presentation particles in the Amarna Letters from Canaan // Ugarit-Forschungen 20, p. 209-220.
Redeker 2006 — Redeker, G. (2006). Discourse markers as attentional cues at discourse transitions // Approaches to Discourse Particles. Amsterdam: Elsevier.
RÉS — Répertoire d'épigraphie sémitique. 1900-68, 8 vols. Paris: Imprimerie Nationale.
RHWUD — Random House Webster Unabridged Dictionary on CD-ROM. 1999.
Ribera i Florit 1983 — Ribera i Florit, J. (1983) Evolución morfológica y semántica de las partículas kïn y ?ry en los diversos estadios del arameo // Aula Orientalis 1, pp. 227-33.
Sallaberger 1999 — Sallaberger, W. (1999) "Wenn Du mein Bruder bist, ...". Interaktion und Textgestaltung in altbabylonischen Alltagsbriefen. Groningen: Styx.
Schiffrin 1987 — SCHIFFRIN, D. (1987). Discourse Markers. Cambridge: Cambridge University Press.
Schourup 1999 — Schourup, L. (1999) Discourse markers // Lingua 107, p. 227-265.
Schwiderski 1997 — SCHWIDERSKI, D.(1997) 'Wer ist dein Knecht? Ein Hund!'. Zu Aufmerksamkeitserregern und Überleitungsformeln in hebräischen Briefen // Wagner A. (ed.) Studien zur hebräischen Grammatik
(Orbis Biblicus et Orientalis 156). Freiburg: Universitätsverlag; Göttingen: Vandenhoeck & Ruprecht, pp. 127-41.
Taboada 2006 — Taboada, M. (2006). Discourse markers as signals (or not) of rhetorical relations // Journal of Pragmatics 38(4), pp. 567-592.
TAD — Porten, B., & Yardeni, A., (1986-1999) Textbook of Aramaic Documents from Ancient Egypt. Vol. 1: Letters. Vol. 2: Contracts. Vol. 3: Literature, Accounts, Lists. Vol. 4: Ostraca & Assorted lnscriptions. Jerusalem: The Hebrew University.
van Dijk 1979 — van Dijk, T.A. (1979). Pragmatic connectives // Journal of Pragmatics 3, pp. 447-456.
Wagner 1997 — Wagner, A. (1997) Sprechakte und Sprechaktanalyse im Alten Testament (Beihefte zur Zeitschrift fur die alttestamentliche Wissenschaft 253). Berlin: Walter de Gruyter.
Во многих языках мира дейктические наречия времени претерпевают категориальный сдвиг: во многих случаях они употребляются как дискурсивные маркеры. В трех семитских языках — древнееврейском, имперском арамейском и аккадском — отмечены общие черты в употреблении наречий времени, указывающих на момент речи. Данная статья посвящена сравнению ненаречных употреблений (юэ)?аШ в древнееврейском, к'ШкХпИкХп в имперском (египетском) арамейском, а также ¡паппа и апитта в аккадском. Теоретико-методологическую базу данного исследования составляет Теория риторической структуры и традиция изучения дискурсивных маркеров. Предварительные результаты показывают, что указанные наречия в большинстве своих употреблений функционируют как дискурсивные маркеры. На основе анализа употреблений древнееврейского (юэ)?аШ в качестве главного объекта сравнения выдвинута специфическая дискурсивная функция: маркирование границы между ассертивным дискурсивным отрезком и директивным дискурсивным отрезком в директивных высказываниях. Отмечено, что как ШУШМ/Шп, так и ¡паппа употребляются в данной функции, однако сфера их употредления шире, чем у древнееврейского (юэ)?аШ, что требует дальнейшего изучения. Апитта представляет собой лексему иного типа: апитта не используется как дейктическое наречие времени; употребление этой частицы также в ряде случаев связано с директивными высказываниями, но в отличие от (и>э)?аНа, ШУШМ/Шп, правый по отношению к апитта дискурсивный отрезок, как правило, является ассертивным, а переход от ассертивного отрезка к директивному в таких случаях остается немаркированным.
