Tense, aspect and Sprechhaltung in biblical Hebrew prose Jan Joosten, Oxford Introduction The verbal system of Biblical Hebrew is famously intractable. A well-known, although today somewhat dated, history of research is titled The Enigma of the Hebrew Verbal System (McFall 198x). In a monograph published in 2012 I have argued that an important part of the problem is that grammarians have not paid sufficient attention to the diachronic factor (Joosten 2012). The Hebrew Bible may have come into being over a period covering five centuries or more. Arguably, then, there is no “verbal system” of Biblical Hebrew — instead two, three, or more systems are to be distinguished according to the various chronolects. In what follows I will limit my remarks to “classical Biblical prose,” the prose chapters in the books of Genesis to 2 Kings. In outline, the system in this chronolect looks as follows: INDICATIVE Past tense MODAL non-tensed taxis anterior wayyiqtol qatal prospective volitive yiqtol/weqatal juss., coh., imp. contemporaneous participle Imperfective perfective subj.-ptcp ptcp-subj. The core verbal paradigm is made up of eight forms representing five functional categories: - Wayyiqtol, locates a process at some point in time in the past. It is the leading tense in narrative, where successive wayyiqtol forms will typically present successive events in the story. But the form as such does not imply sequence; it regularly expresses overlapping events (e.g., 1 Kgs 18.24; 1 Sam. 18.11) or even forms of backtracking or anticipation (Washburn 1994). wayyiqtol also occurs in direct speech (e.g., 2 Sam. 14.5 'I am a widow, my husband died [‫)'] ַוָיָּמת‬. While wayyiqtol positively expresses deictic tense (Kuryłowicz 1973), all other verbal forms are non-tensed. They take their reference time from the context (Hatav 1997): in narrative the past, in direct speech, the present, past or future, according to the context. - Qatal expresses anteriority with regard to reference time. Often the result of the process still obtains at reference time: qatal then expresses the meaning of a perfect. With verbs -1- expressing states this usage may require a present-tense translation in English: ‫ ָזַק ְנִתּי‬zå̄ qantī (Jos. 23.2), “I’m old”; ‫ ָי ַדְﬠִתּי‬yå̄ ḏaʿtī (Gen. 12.11), “I know” (Dobbs-Allsopp 2000). However, qatal is used also to express events that are remote, in which case its function approaches that of the preterit. A special usage is found in “performative” expressions such as ‫ ָנַתִתּי‬nå̄ tattī, “I am hereby giving” (Gen. 1.29) where the notion of anteriority is interpreted in terms of certainty (Rogland 2003). - A compound consisting of the participle and an explicit nominal or pronominal subject is used to express contemporaneity with regard to reference time. The expression of the real, ongoing, present (e.g., Gen. 37.16 “I’m seeking” [‫ ָאֹנִכי ְמַבֵקּשׁ‬ʾå̄ nōḵī məḇaqqēš] my brothers”) and of attending circumstance in narrative (e.g., Gen. 19.1 “And Lot was sitting [‫ְולוֹט י ֵֹשׁב‬ wə-lōṭ yōšēḇ] in the gate of Sodom”) are strictly the domain of the participle and cannot be expressed by other verbal forms. This shows that the predicative participle is a full-fledged component of the BH verbal system. Secondary usages of the participle include reference to the general present or to the imminent future (Joosten 1989). - Yiqtol and weqatal are basically prospective; they present a process as “not yet begun” at reference time. Contextually, this function translates into a multitude of modal nuances, both epistemic and deontic. The use of yiqtol with modal verbs like ‫ ָיכֹל‬yå̄ kol “to be able” and in questions pertaining to the present is compatible with this approach (Joosten 2004). A special usage is the representation of iterative processes, especially in past-tense contexts: in this usage prospective modality is interpreted in terms of repeated action, exactly as in the case of English will/would (e.g., “He will sit like that for hours”; “Sometimes I would see Gerald in the Café Royal”). - The volitives—cohortative, imperative and jussive—are prospective too, but they add a nuance of speaker’s volition. Following another volitive, a question or even a statement, they may be used to express purpose or intended result: ‫ הוִֹציאוָּה ְוִתָשּׂ ֵרף‬hōṣīʾūhå̄ wə-ṯiśśå̄ rēp̄ “Bring her out and (i.e. so as to) let her be burnt” (Gen. 38.24). As expressions of the will of the speaker, the volitives are more tightly linked to speech time than the other verbal forms. The forms occasionally express purpose and intended result in past-tense contexts, however: ‘Any one who wanted he consecrated to be [‫ ִויִהי‬wī-hī] priests for the high places’ 1 Kgs 13:33). The grouping of different morphological items under a single functional heading may seem peculiar: -2- - Yiqtol and weqatal relate to one another as positional variants: weqatal can occur only at the head of the clause, while yiqtol is practically limited to clause-internal positions. They express the same modal functions: prediction, nuances of deontic and epistemic modality, general present, repetition in the past. - Three forms expressing volition of the speaker make up a suppletive paradigm organized by grammatical person: the first person is expressed by the cohortative (asura “let me turn aside”), the second by the imperative (lek “go”) (the jussive in negated clauses), and the third by the jussive (yehi ’or “let there be light”). Such heterosyzygy is not wholly unparalleled (e.g., in Coptic, several tenses have a negative form that is etymologically distinct from the positive). In the time at my disposal I will focus on narrative texts and explore, in the framework proposed to us by the organizers of this conference, how verbal syntax organizes temporality. Aspectual contrasts in classical Biblical Hebrew In the above scheme, aspect is not prominent. An aspectual opposition does exist between two forms of verbal periphrasis, the sequence subject–participle and the sequence participle– subject. The former sequence, subject–participle, expresses imperfective aspect: Num 11:27 ‫ֶאְל ָדּד וֵּמי ָדד ִמְת ַנְבִּאים ַבַּמֲּח ֶנה‬ Eldad and Medad are prophesying (subj. – participle) in the camp! The young man who brings this news to Moses presents the process of prophesying as having begun but not come to an end: the view is internal to the process (or if one prefers, it “sections” the process); the process is followed in its course. In other words, the verbal function is imperfective. This is a constant feature with the sequence subject–participle. The latter sequence, participle–subject, is used when no progress can be observed or when the entire process is presented globally: 1 Sam 19:14 ‫חֶֹלה הוּא‬ He is sick (particple – subj.) -3- Michal’s excuse for her husband is presented as a fact, not an ongoing process. The “punctual” or global viewpoint too is a constant feature with the sequence participle–subject. The linguistic contrast between the two participial sequences is interesting in typological perspective. Most languages that express aspectual nuances by means of verbal forms have the perfective aspect only in past-tense or future-tense forms. Hebrew has a form expressing perfective aspect in reference to contemporary processes. However, the contrast between imperfective and perfective contemporaneity hardly plays out in any biblical texts. There are practically no passages that juxtapose the two sequences in a meaningful way. Aspectual contrasts do occur in narrative. The predicative participle whose subject immediately precedes expresses imperfective aspect, and this function is at times exploited to express contrasts similar to what one may observe in languages where aspect is prominent: Gen 19:1 ‫ַוָיּבֹאוּ ְשׁ ֵני ַהַמְּלאִָכים ְסדָֹמה ָבֶּﬠ ֶרב ְולוֹט י ֵֹשׁב ְבַּשַׁﬠר־ְסד ֹם‬ Ἦλθον δὲ οἱ δύο ἄγγελοι εἰς Σοδοµα ἑσπέρας· Λωτ δὲ ἐκάθητο παρὰ τὴν πύλην Σοδοµων. Les deux anges arrivèrent à Sodome sur le soir; et Lot était assis à la porte de Sodome. The two angels came (wayyiqtol) to Sodom in the evening while Lot was sitting (subj. – participle) in the city’s gateway. The arriving of the angels is presented globally, as a single bounded event, situated at a time when Lot’s sitting has already started but not come to an end. The two processes are simultaneous, but they are also related to one another. Temporally the sitting envelops the arriving. In literary terms, the bounded process is foregrounded while the unbounded one presents the background. In terms of verbal aspect the first form is perfective, the second imperfective. 1 Sam 14:13 ‫ַוַיַּﬠל יוֹ ָנָתן ַﬠל־ָי ָדיו ְוַﬠל־ ַרְגָליו ְוֹנֵשׂא ֵכָליו אֲַח ָריו ַוִיְּפּלוּ ִלְפ ֵני יוֹ ָנָתן ְוֹנֵשׂא ֵכָליו ְממוֵֹתת אֲַח ָריו׃‬ (…) καὶ ἐπάταξεν αὐτούς, καὶ ὁ αἴρων τὰ σκεύη αὐτοῦ ἐπεδίδου ὀπίσω αὐτοῦ. (…) Les Philistins tombèrent devant Jonathan, et celui qui portait ses armes donnait la mort derrière lui. Jonathan crawled up on his hands and feet, with his armor bearer following behind him. Jonathan struck down (wayyiqtol) the Philistines, while his armor bearer came along behind him and killed them (subj. – participle). -4- Aspectual contrasts exist in the texts, but they are not central to the verbal system as such. In Saussurian terms they occur in la parole but they are not fully present in la langue. While the subject-participle phrase expresses imperfectivity, the wayyiqtol form is not properly perfective. Wayyiqtol is neutral with regard to aspect and must often be translated with imperfective forms in aspect-prominent languages: 2 Sam 11:2 ‫ַוְיִהי ְלֵﬠת ָהֶﬠ ֶרב ַוָיָּקם ָדּ ִוד ֵמַﬠל ִמְשָׁכּבוֹ ַוִיְּתַהֵלְּך ַﬠל־ ַגּג ֵבּית־ַהֶמֶּלְך ַוַיּ ְרא ִאָשּׁה ר ֶֹחֶצת ֵמַﬠל ַה ָגּג‬ Καὶ ἐγένετο πρὸς ἑσπέραν καὶ ἀνέστη Δαυιδ ἀπὸ τῆς κοίτης αὐτοῦ καὶ περιεπάτει ἐπὶ τοῦ δώµατος τοῦ οἴκου τοῦ βασιλέως καὶ εἶδεν γυναῖκα λουοµένην ἀπὸ τοῦ δώµατος Un soir, David se leva de sa couche; et, comme il se promenait sur le toit de la maison royale, il aperçut de là une femme qui se baignait. One evening David got up (wayyiqtol) from his bed and walked around (wayyiqtol) on the roof of his palace. From the roof he saw (wayyiqtol) a woman bathing. The aspectual contrasts illustrated above are lopsided. Wayyiqtol becomes perfective only when it is paired with a participle clause. In addition there is a syntactic constraint. The participle expresses imperfectivity in narrative only in subordinate clauses. This is true even in the above examples. The construction consisting of the conjunction waw, a nominal or pronominal subject, and the participle, in that order, is treated as a circumstantial clause in Hebrew grammar. Note that the English translation translates with the conjunction “while”: the notion of circumstantiality implicit in the Hebrew has to be brought out lexically in English. The use of modal forms in iterative and habitual clauses Alongside the aspectual contrast illustrated in the preceding section, classical prose also exhibits a different way of dimensioning narrative. The “Inversionspaar” yiqtol and weqatal is used in clauses referring to repeated or habitual processes: Exod 17:11 ‫ְוָהָיה ַכֲּאֶשׁר ָי ִרים מֶֹשׁה ָידוֹ ְוָגַבר ִיְשׂ ָרֵאל ְוַכֲאֶשׁר ָי ִני ַח ָידוֹ ְוָגַבר ֲﬠָמֵלק‬ Whenever Moses would raise (yiqtol) his hands, Israel prevailed (weqatal), but whenever he would rest (yiqtol) his hands, Amalek prevailed (weqatal). -5- While the preceding and following verses enumerate the successive stages of the battle against Amalek, verse 11 tells of a recurring set of processes. This type of syntax has led many scholars to describe the function of yiqtol (and more rarely that of weqatal too) in aspectual terms. It is true that in aspectual languages, like Greek and French, imperfectives may be used in expressions of iterativity. But the fact that imperfectives are compatible with iterativity doesn’t mean all iteratives are aspectually based. Yiqtol and weqatal are very frequent forms in biblical Hebrew, and they usually (in over 70% of occurrences) express forms of prospectivity: prediction, expectation, obligation, etc. It is better, therefore, to define the core function of these forms as modal (ie. they express nonindicative modality, or irrealis). Aspectually yiqtol and weqatal are neutral. As several general linguists have pointed out, modal (irrealis) forms are used for nuances of iterativity and habituality in many languages of the world. The English “would form” is exemplified in the translation of Exod 17:11 above. Ancient Greek has an iterative use of the optative, which is modal. Modal forms underscore the eventuality of repeated processes. In the story of the war against Amalek Moses’ lifting up of his hand might not be happening at any definite point on the timeline. But because he lifted up his hand repeatedly, it was bound to happen eventually. The verbal forms capture this nuance of eventuality. The iterative-habitual use of yiqtol and weqatal is not limited to narrative. It is also used in direct speech where the time frame is past: Amos 4:7 ‫ְוַגם אָֹנִכי ָמ ַנְﬠִתּי ִמֶכּם ֶאת־ַהֶגֶּשׁם ְבּעוֹד ְשֹׁלָשׁה ֳח ָדִשׁים ַלָקִּציר‬ ‫ְוִהְמַט ְרִתּי ַﬠל־ִﬠיר ֶאָחת ְוַﬠל־ִﬠיר אַַחת ל ֹא אְַמִטיר‬ And I also withheld the rain from you when there were still three months to the harvest; I would send rain on one city, and send no rain on another city. But the usage really comes into its own in narrative. At times it serves to integrate a sequence of recurrent processes into the narrative flow, as in Exod 17:11. More often it comes in descriptions of the initial conditions in which the story develops (e.g. in 1 Sam 1:1-7 to tell of the situation in Elkanah’s family and of their annual pilgrimage to Silo—which will become the occasion for Hanna’s vow and the subsequent birth Samuel, the protagonist of the story), or in descriptions of the end state of a certain episode (e.g. in 1 Sam 16:23 where it describes how David, once he had taken up his position at Saul’s court, would play the lyre every time Saul experienced a bout of depression). -6- The back-and-forth between preterit wayyiqtol and iterative yiqtol/weqatal is an effective device in classical Hebrew narrative. It allows the authors to alternate between reporting single events that make up the backbone of the story and opening up a wider temporal scope and thus giving insight into the background of those events. Erzählte and besprochene Welt in biblical narrative The distinction between Hebrew wayyiqtol and qatal is subtle and complex. As was stated above, wayyiqtol must in principle be defined as a preterit and qatal a perfect: 2 Sam 11:26-27 ‫ַוִיָּלֶּחם יוֹאָב ְבּ ַרַבּת ְבּ ֵני ַﬠמּוֹן ַוִיְּלכֹּד ֶאת־ִﬠיר ַהְמּלוָּכה׃‬ ‫ַוִיְּשַׁלח יוֹאָב ַמְלאִָכים ֶאל־ ָדּ ִוד ַויּ ֹאֶמר ִנְלַחְמִתּי ְב ַרָבּה ַגּם־ָלַכ ְדִתּי ֶאת־ִﬠיר ַהָמִּים׃‬ So Joab fought (wayyiqtol) against Rabbah of the Ammonites and captured (wayyiqtol) the royal city. Joab then sent messengers to David, saying, “I have fought (qatal) against Rabbah and have captured (qatal) the water supply of the city.” The relation between the two is not unlike that between German “stritt” and “hat gestritten,” English “fought” and “has fought,” or French “combattit” (passé simple) and “a combattu.” But, as we all know, the precise relation between preterit and perfect is not exactly the same in our languages. In Hebrew too the relation has many peculiarities. The natural home of wayyiqtol is in narrative, that of qatal in interpersonal speech—as shown in the example. But qatal is also frequent in narrative. Three different usages of qatal in narrative need to be distinguished. I will focus on the third one, which is the most interesting in the framework of this conference. But I want to present the other two rapidly for the sake of completeness. 1) To begin with, qatal will stand in for wayyiqtol whenever the verbal form does not, for whatever reason, occupy the first position in the clause. Wayyiqtol incorporates the conjunction we “and,” which on the clausal level is tied to the first position. Thus, qatal will assume the functions of wayyiqtol in negated clauses, in clauses introduced by a temporal adverb, and in clauses in which focus or topicalization of a nominal or adverbial element relegates the verb to the second position: Gen 1:5 -7- ‫ַוִיְּק ָרא ֱאֹלִהים ָלאוֹר יוֹם ְוַלחֶֹשְׁך ָק ָרא ָלְיָלה‬ God called (wayyiqtol) the light “day” and the darkness he called (qatal) “night.” This “surrogate” qatal in narrative is limited to main clauses. 2) In subordinate clauses, qatal normally expresses anteriority with regard to the time frame of the story. The reference time is indicated in the main clause and qatal expresses its usual, anterior, function: Gen 3:23 ‫ַוְיַשְׁלֵּחהוּ ְיה ָוה ֱאֹלִהים ִמַגּן־ֵﬠ ֶדן ַלֲﬠבֹד ֶאת־ָהֲא ָדָמה ֲאֶשׁר ֻלַקּח ִמָשּׁם׃‬ So the LORD God expelled him from the orchard in Eden to cultivate the ground from which he had been taken (qatal) Gen 38:15 ‫ַוִיּ ְרֶאָה ְיהוּ ָדה ַוַיְּחְשֶׁבָה ְלזוֹ ָנה ִכּי ִכְסָּתה ָפּ ֶניָה׃‬ When Judah saw her, he thought she was a prostitute because she had covered (qatal) her face. Gen 18:33 ‫ַוֵיֶּלְך ְיה ָוה ַכֲּאֶשׁר ִכָּלּה ְל ַדֵבּר ֶאל־אְַב ָרָהם‬ The LORD went on his way when he had finished (qatal) speaking to Abraham. This is the “pluperfect” use of qatal. 3) A third usage of qatal in narrative is when the verbal form signals a transition from the narrative mode of communication to a more discursive mode. This is not a shift of aspect, nor temporal scope, but what Harald Weinrich has called Sprechhaltung. Because wayyiqtol is naturally allied to narrative, and qatal to discourse, the switch from the one to the other may imply a change in the “stance” of the speaker. From an anonymous voice relating the story, the speaker becomes a commentator addressing the audience directly. The usage is rare in mains clauses: 1 Kgs 13:18 ‫ַויּ ֹאֶמר לוֹ ַגּם־ֲא ִני ָנִביא ָכּמוָֹך וַּמְלאְָך ִדֶּבּר ֵאַלי ִבּ ְדַבר ְיה ָוה ֵלאמֹר ֲהִשֵׁבהוּ ִאְתָּך ֶאל־ֵבּיֶתָך ְוי ֹאַכל‬ ‫ֶלֶחם ְוֵיְשְׁתּ ָמִים ִכֵּחשׁ לוֹ׃‬ -8- The old prophet then said, “I too am a prophet like you. An angel told me with the LORD’s authority, ‘Bring him back with you to your house so he can eat and drink.’” He lied (qatal) to him. The last two words do not form part of the narrative, they are like an aside from the narrator to the reader, a comment on what was just told. The use of qatal in authorial comments is more commonly attested in subordinate clauses. As I pointed out above, in subordinate clauses in narrative qatal usually translates into an English pluperfect. Sometimes, however, it does not: 2 Sam 16:23 ‫ַוֲﬠַצת ֲאִחיתֶֹפל ֲאֶשׁר ָיַﬠץ ַבָּיִּמים ָהֵהם ַכֲּאֶשׁר ִיְשׁאַל־ִאישׁ ִבּ ְדַבר ָהֱאֹלִהים ֵכּן ָכּל־ֲﬠַצת ֲאִחיתֶֹפל‬ ‫ַגּם־ְל ָד ִוד ַגּם ְלאְַבָשֹׁלם׃‬ And the counsel of Ahithophel, which he counselled (qatal) in those days, was as if a man had inquired at the oracle of God: so was all the counsel of Ahithophel both with David and with Absalom. In this passage, the expression “in those days” reveals that the author is speaking from his own point of view. In other passages such explicit indications are missing. Only the temporal implications of the statement show the switch in Sprechhaltung. A very interesting exploitation of this syntactic possibility occurs in phrases with ki ’amar “for he said,” introducing authorial comments: Gen 29:32 ‫ַוַתַּהר ֵלאָה ַוֵתֶּלד ֵבּן ַוִתְּק ָרא ְשׁמוֹ ְראוֵּבן ִכּי אְָמ ָרה ִכּי־ ָראָה ְיה ָוה ְבָּﬠ ְנִיי‬ So Leah became pregnant and gave birth to a son. She named him Reuben, for she said (qatal), “The LORD has looked with pity on my oppressed condition.” The qatal form is not pluperfect. Nor is the clause it introduces fully part of the narrative. The formula ki ’amar does not refer to something that was really said at the time Leah gave birth to hear first-born. The clause explains the choice of the name. It is an etiology, and as such does not function on the level of the story but of the communication between the author and the readers. -9- Gen 38:11 ‫ַויּ ֹאֶמר ְיהוּ ָדה ְלָתָמר ַכָּלּתוֹ ְשִׁבי אְַלָמ ָנה ֵבית־אִָביְך ַﬠד־ִיְג ַדּל ֵשָׁלה ְב ִני ִכּי אַָמר ֶפּן־ָימוּת ַגּם־הוּא‬ ‫ְכֶּאָחיו‬ Then Judah said to his daughter-in-law Tamar, “Live as a widow in your father’s house until Shelah my son grows up.” For he said (qatal), “I don’t want him to die like his brothers.” In this example, the formula introduces an inner view of Judah’s thought process at the time he refuses to let Tamar marry his youngest son (after the first two sons have died). Again the clause is not tightly knitted into the narrative, but gives “privileged information” allowing to understand the events better. To recapitulate, in the third usage of qatal in narrative the function is neither preterit, nor pluperfect. Instead, qatal is retrospective, the reference time shifts from the past to the present of the storyteller. This shift implies a switch in Sprechhaltung. In his book Tempus. Besprochene und erzählte Welt, Harald Weinrich argued that languages like French and German do not have one verbal system but two, one for narrative and one for discursive texts. Narrative implies a relaxed attitude, what is told is put at a distance, the speaker and the audience are not direct participants; the natural mode is to use the third person. Discourse implies a more intense attitude, in which the speaker takes a position and involves the audience; the first and second persons are regularly used. Weinrich’s book has proved seminal, but it has also attracted much criticism. Rather than describing verbal functions as such, his theory concerns the way verbal functions may be exploited in certain types of text. But however this may be, Weinrich’s approach is helpful in describing the non-preterit, nonpluperfect usage of qatal in narrative. Conclusions The present paper is not intended as an exhaustive description of verbal syntax in classical Hebrew narrative. The three phenomena discussed were chosen to illustrate how verbal syntax structures temporality in storytelling, but many other phenomena could have been selected to show similar effects. Although classical Hebrew is somewhat poor in verbal forms, and even poorer in verbal categories, its verbal usage is very diverse. Simultaneity and overlapping time frames can be expressed with much precision. The combination of preterit wayyiqtol with a circumstantial participle clause is only the most common of a whole series of interclausal relations expressing notions of synchronicity. Temporal scope—zooming in on single events or zooming out on broader and more complex processes—also rates highly in Hebrew syntax. Alongside the switch to modal - 10 - forms expressing iterative-habitual processes, other constructions too effectuate a diversification of immediacy in the representation of the past. Shifts from narration to authorial comment can be explicitly marked by the use of qatal, but many other forms of language, both syntactical and lexical, can be used to similar effect. As a dead language with a relatively small corpus, classical Hebrew may never give up all its secrets. The verbal system will probably remain enigmatic to a certain extent. But what we can understand is surprisingly vigorous, allowing much precision in the expression of a wide variety of temporal nuances, in several dimensions. - 11 -