Verbal Argument Structure in the War Scroll

This essay presents the results of an extended study of verbal argument structure in the War Scroll (1QM). I first establish a method based in generative linguistic theory. I then illustrate this method with a discussion of the argument structure of Qal יצא in 1QM and other Dead Sea Scrolls. Following this case study, I present the data from 1QM on verb argument structure—specifically, instances where 1QM adds evidence that is not covered in previous studies of the Dead Sea Scrolls. 1QM presents few develop-ments from earlier Hebrew; I argue that such continuity is significant. I conclude with reflections on the implications of argument structure in 1QM for the study of ancient Hebrew.


Introduction
This study presents the results of an extended survey of verbal argument structure in the War Scroll (1QM). After articulating a method based in generative linguistic theory, I discuss an example case from my corpus. I then present the general data, highlighting where 1QM adds evidence not covered in previous studies of the Dead Sea Scrolls. Finally, I conclude with some reflections on implications for the study of Hebrew-specifically, the value of an analysis showing that the Hebrew of the Scrolls does not differ much from earlier Hebrew. I have considered and compiled data for the argument structure of all verbs occurring five times or more in 1QM. The end of this analysis of 1QM marks the beginning of broader considerations that place 1QM's data in the context of all the Scrolls and the even larger context of ancient Hebrew's development over time. Both of these steps are necessary: beginning with a narrow analysis in just one text-in order not to lose the trees-and from that basis going to a broader view of the forest.1

Linguistic Framework and Method
The concept of a verb's "argument structure" is similar to "verb complementation," "verb valency," and "transitivity." A number of scholars have considered verbal argument structure in ancient Hebrew. Takamitsu Muraoka,2 Michael Malessa,3 and John Cook4 published dedicated studies to the issue that take the Hebrew of the Bible as the main point of departure; the recent contributions by Malessa and Cook are based on the perspectives of Dependency Grammar and Generative Linguistics, respectively. Many reference grammars include discussions of verb complementation5-often under the guise of discussing the "accusative"-while recent volumes in the Baylor Handbook on the Hebrew Bible address the theory of argument structure and apply it to particular texts from the Hebrew Bible.6 Also of note is the nine-volume Dictionary of Classical Hebrew,7 which includes extensive data on lexemes and parts of speech that frequently occur with verbs (though not set within an explicit argument structure framework), including in the Dead Sea Scrolls. Discussions of argument structure in the Hebrew of the Dead Sea Scrolls can be found in Qimron's reference grammar8 and Muraoka's "Morphosyntax and Syntax of Qumran Hebrew."9 Studies dedicated to argument structure in the Scrolls have been presented by Muraoka-who catalogues the complements of verbs in 1QS, 1QSa, 1QSb, and 1QH, with further discussion of particular cases10-and Rey-who focuses on the use of bet to mark "objects" of the verb in the Dead Sea Scrolls.11 In my analysis of verbal argument structure, I follow generative linguistics instead of traditional grammar.12 As a result, I use terms and concepts like complement, argument structure, and theta-role, instead of concepts like indirect object, direct object, and transitivity.13 Depending on the verb being used, the verb phrase may require other constituents to complete the verb phrase. These constituents are complements-in contrast to adjuncts, which are not required. Complements can be noun phrases (NP s), but are often other types of phrases such as prepositional phrases (PPs).14 Besides complements, most verbs also require subjects. The requirements of each specific verb for certain arguments (subjects and complements) is its argument structure. To a native English speaker, for example, the phrases "I hit" (without any further context) and "hits the ball" are ungrammatical. The phrase "she hits the ball," on the other hand, is grammatical, because it contains all the necessary arguments. Arguments can be null/covert, as in the following example (with its preceding context): "Did she go to the lecture? Yes, she went." In the second sentence, the phrase "to the lecture" is null/covert. The fact that a particular type of phrase is used often with a verb does not entail that such a phrase is a required argument; the phrase may be adjunct to the verb. Verbs are not the only words that take argument structure; adjectives, for example, can require complements.15 11 Rey, "Prepositional Object with bet." On Rey's concept of object, see footnote 13. 12 See Carnie,Syntax,Cowper,Syntactic Theory,[48][49][50][51][52][53][54][55]Aarts,English Syntax and Argumentation,[94][95][96][97][98][99] The notion of "object"-the thing acted upon in a predication-is more about syntax than about the semantics and argument structure of the verb; the subject of a passive verb is not its object, though it is the thing acted upon (e.g., "the car was washed"). As John Cook has noted, analyses of Hebrew based on these kinds of categories result in descriptions that are possibly misleading: "the transitivity approach of traditional grammar leads to awkward discussions about so-called accusative noun phrases that function as something other than direct object and other 'objects' of the verb as mediated by prepositions;" Cook, "Valency," 55. Such approaches are commonly applied in the study of Biblical Hebrew and Dead Sea Scrolls Hebrew; e.g., Rey (1) *I hit.
(2) I hit the ball.
(10) I walked my mother home.
(11) *I went the sky. (4) *Walked to the store. (12) I went to the store. (5) I walked to the store. (13) I went for a run. Because a verb's argument structure typically includes subjects, not just complements,16 I avoid the term "complementation" as a way of referring to the entire phenomenon.17 I also avoid "valency" and related terms ("monovalent," "bivalent," etc.) because they are redundant with "argument structure" and related terms ("requiring one argument," "requiring two arguments," etc.). Argument structure depends on the semantics of the lexemes involved.18 This is most apparent for the verb, where the number and type of arguments depends on the meaning of the verb. The ability of an argument to fulfill its role also depends on semantics-the argument's semantics. "I walked to the store," for example, is grammatical, but "East walked to compassion" is notdespite the same number of arguments being present in the same syntactic constructions (subject, PP). The concept of thematic relations accounts for the semantic aspects of argument structure (see the table of thematic relations in Figure 219). The verb "to walk" requires a complement, but not just any complement: it must be a complement whose thematic relationship to the verb is LOCATION. Similarly, the subject required by "to walk" must have the thematic relationships of both AGENT and THEME. Argument structures require these thematic relations to be present; because different parts of speech can take the same thematic relation, argument structure is not about the requirement for particular parts of speech.20 16 The use of different binyanim in Hebrew, moreover, causes significant overlap between the two. 17 Note that in some linguistic theories, the subject is referred to as a complement; Malessa,"Valency," 895. 18 Cf. Malessa,"Valency," 893. 19 Adapted from Cowper,Syntactic Theory,[48][49][50][51] For example, one can "walk to the store" (PP) or "walk home" (NP); one can ‫י‬ ‫ע‬ ‫ש‬ ‫ה‬ ‫כ‬ ‫א‬ ‫ש‬ ‫ר‬ There is not a one-to-one relationship between thematic relations and a verb's argument structure.21 In the phrase "Tim sold me his saxophone," "Tim" is both AGENT and SOURCE, while "his saxophone" is PATIENT and GIFT. Moreover, a single argument might be allowed to take one thematic relation out of several options: the verb "to go," for example, requires a complement that can be either a LOCATION or a GOAL: I went to the store and I went for a run.22 Because thematic relations do not map isomorphically onto argument structure, generativists have posited the theta-role: the bundle of possible thematic relations assigned to one position in a verb's argument structure.23 Each argument of the verb has one theta-role. For convenience, theta-roles are often referred to by the most prominent thematic relation in that theta-role. In "Tim sold me his saxophone," "Tim" is both AGENT and SOURCE, but fills just one theta-role, which we could call the AGENT theta-role. In the discussion that follows, I will refer to arguments using thematic-relation terms like AGENT, PATIENT, and GOAL, as shorthand terms for theta-roles. In my analysis of argument structure in 1QM, what I am ultimately after is the identification of thetaroles. I still pay attention to what parts of speech these theta-roles take (e.g., LOCATION as NP, ‫-ל‬PP, ‫ע‬ ‫ל‬ -PP, etc.); the differences in such cases, however, are not differences in argument structure-they are semantic nuances.24 Though further conceptual nuance has been developed to describe verbal argument structure,25 the model I have just sketched is sufficient for our purposes. A few more points will help to clarify what I aim to accomplish (and what I do not aim to accomplish) in this study. First, verbs-whether in ancient Hebrew or modern languages-are used in various ways, with and without different types of arguments.26 "I sold my car" (two arguments) and ("act thus"; adverb complement). This point marks a major point of difference between my study and past studies of argument structure in the Scrolls; Muraoka's "Verb Complementation in Qumran Hebrew," for example, lists complements by their part of speech, not their thematic relation or theta-role. On the distinction between syntactic role and thematic relation, as well as the place of semantics in argument structure, cf. Aarts, English Syntax and Argumentation, 97-99; Qimron, Hebrew of the Dead Sea Scrolls, 428. 21 Cowper, Syntactic Theory, 51. 22 Hebrew verbs of motion (like ‫ב‬ ‫ו‬ ‫א‬ ) are similar: ‫ב‬ ‫ו‬ ‫א‬ usually takes a LOCATION complement-whether a PP or NP-but the same role can also be fulfilled by a GOAL. 23 Carnie, Syntax, 233-34; Cowper, Syntactic Theory, 55. 24 Muraoka's semantic distinction between ‫ב‬ ‫ו‬ ‫א‬ ‫ל‬ ("approach") and ‫ב‬ ‫ו‬ ‫א‬ ‫ב‬ ("enter"), for example, does not entail a difference in argument structure; "Verb Complementation in Qumran Hebrew," 110, n33. 25 See, e.g., Jackendoff, "Thematic Relations." 26 Cf. Malessa, "Valency," 895.
Dead Sea Discoveries 27 (2020) 392-409 AGENT-entity that initiates or does an action INSTRUMENT-entity with which an action is performed PATIENT-entity that undergoes an action CAUSE-entity that makes an AGENT initiate or do an action GOAL-end result of action (including a verb of motion) LOCATION-entity in or toward which motion takes place SOURCE-entity from which motion/transaction takes place THEME-entity that moves with a verb of motion RECIPIENT-entity on the receiving end of a transaction GIFT-entity that changes possession in a transaction EXPERIENCER-entity that undergoes or perceives the event or state of being PERCEPT-entity that is experienced or perceived BENEFACTIVE-entity for whose benefit the event takes place figure 2 Some thematic relations "I sold Becky my car" (three arguments) are both grammatical; similarly, "I walked to the store" (with a LOCATION complement), "I walked my dog" (with a PATIENT/THEME complement), and "I walked my mother to the store" (with LOCATION and PATIENT/THEME complements) are all grammatical. Though these variations are possible, this does not entail that there are no rules governing the argument structures of "to sell" and "to walk." Given various aspects of context and the semantics of the individual words, these verbs take specific argument structures in each case, entailing ungrammaticality if the argument structure is contravened. Second, argument structure is one of the more difficult aspects of language to access in a dead language. Because arguments of the verb are often null/covert, native-speaker grammaticality judgments are crucial for determining argument structure with certainty.27 Given our lack of access to native speakers of ancient Hebrew, we must tread carefully and recognize the limits of our investigation. Hard and fast rules for determining argument structure based on ancient Hebrew evidence are inadvisable. A single token, for example, may show us a grammatical instance of the argument structure of a verb; and though it only occurs once, such an analysis may be preferable to positing a mistake, corruption, or a completely different meaning for the verb. Although we cannot make grammaticality judgements, this does not negate the reality that there was grammaticality in ancient Hebrew-like any living, spoken lan-guage. To the native speaker of ancient Hebrew, grammaticality mattered, and we must account for this in terms of our theory.

Analysis
In this section, I begin with an example ( ‫י‬ ‫צ‬ ‫א‬ ) that illustrates my approach, and then present and discuss my data for all verbs in 1QM. As noted above, studies of argument structure in the Hebrew of the Scrolls have been published by Muraoka,Qimron,and Rey.28  almost always occurs in a ‫-ב‬PP in 4QInstruction "when it is the object" of six "verbs of instruction" ( ‫ל‬ ‫ק‬ ‫ח‬ , ‫ג‬ ‫ל‬ ‫ה‬ , ‫י‬ ‫ד‬ ‫ע‬ , ‫נ‬ ‫ב‬ ‫ט‬ , ‫ש‬ ‫כ‬ ‫ל‬ , and ‫ב‬ ‫י‬ ‫ן‬ ); "Prepositional Object with bet," 189. He then seeks to explain these uses as developments in verb complementation, analyzing the use of ‫-ב‬PPs with these verbs throughout the Dead Sea Scrolls. Generally speaking, I might prefer to analyze the phenomena Rey discusses in terms of the semantics of ‫,ב‬ rather than the argument structure of the verbssince (in my methodology) a PATIENT is a PATIENT whether it is a NP or a ‫-ב‬PP; cf. the discussion of ‫נ‬ ‫פ‬ ‫ל‬ in section 3.2. In his discussion of ‫ג‬ ‫ל‬ ‫ה‬ , Rey argues that ‫ב‬ does not indicate instrument but instead the thing "revealed"; 209. The thesis in this case seems to require a development in the semantics of the verb ‫ג‬ ‫ל‬ ‫ה‬ , from "to uncover [something]" to "to reveal [something]  The data from 1QM and from Muraoka's corpus (1QS, 1QSa, 1QSb, and 1QH), however, suggest that his analysis is incorrect or at least not representative (see Table 1). In 1QM, ‫י‬ ‫צ‬ ‫א‬ regularly occurs with constituents indicating a LOCATION or GOAL: one goes out for battle, or goes out to a certain position.34 In cases where there is no LOCATION or GOAL overt, it is not difficult to reconstruct from context what the LOCATION or GOAL is, and sometimes context demands it. The question is whether the verb requires a complement of this type-a LOCA-TION/GOAL argument (I use "LOCATION/GOAL" as shorthand for the theta-role fulfilled by a LOCATION or a GOAL)-or whether these are simply adjuncts to the verb. In my view, then, ‫י‬ ‫צ‬ ‫א‬ 's main argument structure in the Hebrew of the Scrolls requires a THEME (the person or thing going out) and a LOCATION/GOAL (the place to which they go out or the goal of going out), with no requirement for a SOURCE. Given the problems inherent in analyzing argument structure in a dead language (see section 2), I hold loosely to this particular analysis of ‫י‬ ‫צ‬ ‫א‬ as requiring a LOCATION/GOAL complement. It is also important to note that other argument structures, though not the mainly used structure, are in use (see below).
One brief passage, 1QM 7.9, 13-14, provides a good example of the use of ‫י‬ ‫צ‬ ‫א‬ in 1QM. The LOCATION/GOAL and SOURCE arguments are bracketed for clarity and labelled with ⇤ and ↤, respectively. ‫ל‬ , "to a-place-thatis-between the battle lines." In 7:13-14, the verb is used two more times; in neither case is the SOURCE overt, and only once is the LOCATION/GOAL overt (   ‫ה‬  ‫מ‬  ‫ע‬  ‫ר‬  ‫כ‬  ‫ו‬  ‫ת‬ ‫ב‬ ‫י‬ ‫ן‬ ‫א‬ ‫ל‬ ). However, given the context, the audience clearly knows that the SOURCE and LOCATION/GOAL are the same as in line 9. Indeed, it would be redundant to state them again overtly. With the words "when the priests go out into the space between the battle lines," the reader assumes from the middle gate; similarly, with the words "seven Levites will go out," the reader assumes from the middle gate, to the place between the battle lines. From a generative perspective, these constituents are truly present in the syntax, even though they are null/covert. In 1QM 7:9, 13-14, and in most instances of ‫י‬ ‫צ‬ ‫א‬ in 1QM, context suggests that the point of arrival is in focus, not the point of departure. In the language of 1QM and possibly 1QH, 1QS, 1QSa, and 1QSb, then, ‫י‬ ‫צ‬ ‫א‬ can take an argument structure that requires a THEME and a LOCATION/GOAL, without a SOURCE.36 35 Translations are mine unless noted otherwise. 36 This judgement depends on how willing I am to presume a null SOURCE in many placesothers may make a different judgement as a result.
Dead Sea Discoveries 27 (2020) 392-409 Though this looks like the dominant argument structure, another argument structure (on which see below) is also used in a few cases-requiring a THEME and SOURCE with no other complements.37 When we consider earlier periods of Hebrew, reflected in the Hebrew of the Bible, we find that there, too, Qal ‫י‬ ‫צ‬ ‫א‬ often takes a LOCATION/GOAL complement. A full-scale analysis of ‫י‬ ‫צ‬ ‫א‬ 's argument structure in the Hebrew of the Bible is outside the scope of this study; however, a survey of the evidence and the lexicons is generally conclusive on the following point. Though semantically the lexeme indicates an action involving departure,38 the verb's complement is not always the point of departure (SOURCE); the point of arrival (LOCA-TION) or purpose of departure (GOAL) are also prevalent as complements.
There seem to be at least two argument structures in use for Qal ‫י‬ ‫צ‬ ‫א‬ in the Hebrew Bible (as in many of the English examples in Figure 1 above) Other possible argument structures include: subject but no complements (e.g., 2 Chr 21:15); and subject with two complements (SOURCE and LOCATION/GOAL; e.g., Gen 11:31). The former seems likely; in examples of the latter, it seems more likely that either the SOURCE or LOCATION/GOAL is an adjunct, i.e., one of the argument structures discussed in the body of this essay. , "Ben-Hadad went out to him" Because Hebrew lexicons do not explicitly address argument structure and complementation, it is impossible to do more than intuit what argument structure(s) might be behind their semantic analyses. Even so, the lexicons give an understanding of Qal ‫י‬ ‫צ‬ ‫א‬ that is compatible with the two argument structures I define above. The first paragraph of BDB's entry on ‫י‬ ‫צ‬ ‫א‬ frequently cites uses with a SOURCE ( ‫מ‬ ‫ן‬ -PP).43 The second paragraph, though much smaller than the first, is entirely focused on uses with a destination (LOCATION) specified.44 HALOT does not provide much access to the types of complements most often appearing with Qal ‫י‬ ‫צ‬ ‫א‬ , nor the distinction of cases focused on departure versus those focused on destination.45 Nevertheless, HALOT's entry cites some examples with PPs that clearly fit the SOURCE or the LOCATION/GOAL analysis. DCH cites extensive data on what types of PP are used with ‫י‬ ‫צ‬ ‫א‬ .46 We should not equate particular parts of speech with particular theta-roles (see section 2). However, in most of the cases DCH cites with a ‫מ‬ ‫ן‬ -PP, the ‫מ‬ ‫ן‬ -PP indicates a SOURCE and is often a complement. Similarly, many cases with ‫-ל‬PP s, ‫א‬ ‫ל‬ -PPs, ‫ע‬ ‫ל‬ -PPs, and directional ‫ה‬ indicate a LOCATION/GOAL, and a good number of these are complements rather than adjuncts. While DCH's main gloss for this section, "go out from," seems to suggest a required SOURCE, the frequency of ‫-ל‬PP s, ‫א‬ ‫ל‬ -PPs, ‫ע‬ ‫ל‬ -PPs, and phrases with directional ‫ה‬ is striking; there are about 2.25 columns on these, compared to 1.75 columns for ‫מ‬ ‫ן‬ -PP s. In summary, the evidence presented in these lexicons supports the analysis of Qal ‫י‬ ‫צ‬ ‫א‬ as employing at least two argument structures in the Hebrew of the Bible-one with a SOURCE complement, and another with a LOCATION/GOAL complement.47 43 Brown,Driver,and Briggs,  When we set the evidence from the Dead Sea Scrolls in the context of earlier Hebrew, it is important to consider how stylistic and other factors play a role in shaping the evidence.48 Critically, the prevalence of LOCATION/GOAL complements in 1QM and 1QS (where most of the data is found) is conditioned by the specific contents of those texts: soldiers going to battle, and initiates joining a group. If we instead had evidence of ‫י‬ ‫צ‬ ‫א‬ from texts that discussed other topics-laws for bodily emissions, for example-the argument structure with a SOURCE complement would likely be more prevalent. Using Judg 20-another text that describes soldiers going to battle-as a point of comparison, we find similar patterns of complementation for ‫י‬ ‫צ‬ ‫א‬ .49 The particular language use in the Dead Sea Scrolls does not transparently reflect the language system that lies behind the text: the balance of one argument structure versus the other in the data does not mean the same balance exists in the language. Though they are much fewer, cases of ‫י‬ ‫צ‬ ‫א‬ in the Scrolls with a SOURCE and no LOCATION/GOAL likely represent an argument structure that was common and grammatical in the language system. Comparing the overall picture in the Dead Sea Scrolls to the overall picture in the Hebrew Bible, the argument structure with SOURCE makes up a much larger proportion of cases in the latter than in the former. This difference does not, however, reflect a diachronic change; rather, the texts from the Scrolls that make up our data set just happen to address topics that more often require the LOCATION/GOAL argument structure of ‫י‬ ‫צ‬ ‫א‬ . In summary, it appears that Qal ‫י‬ ‫צ‬ ‫א‬ has a well attested argument structure where a LOCATION/GOAL argument is required, in contrast to Muraoka's entry with an argument structure only a SOURCE is required. Both argument structures are evidenced in 1QM, other Scrolls, and in the Hebrew Bible. The example of Qal ‫י‬ ‫צ‬ ‫א‬ demonstrates that profiling verbal argument structures in one particular text is an important task. It can provide important information about diachronic development (both evidence for and against development), and it can illuminate synchronic variety of which we previously were not aware. adjuncts. Thus, the verb's only argument for Cook is the subject; Cook, "Verb Valency Lexicon," ‫י‬ ‫צ‬ ‫א‬ . At the same time, he notes that "destination" is more frequently specified than "origin." 48 Screnock, "Hebrew Diachrony," 792.

3.2
The Argument Structure of Other Verbs in 1QM The evidence from 1QM adds new verbs, binyanim, and argument structures that do not appear in Muraoka's "Verb Complementation in Qumran Hebrew"; these are presented in Table 2.50   table 2 Argument structures not present in Muraoka's database Most of 1QM's verbs, then, behave in the same way they behave in the Hebrew Bible. Rather than hiding the continuity with preceding Hebrew and focusing on ways in which the Hebrew of the Scrolls is different, I think the continuity is worth highlighting-indeed it must be highlighted. Though the prevailing focus in Dead Sea Scrolls scholarship falls on what is different in the Scrolls, it is equally important to note and chart out continuities and similarities. 63 In addition to presenting the evidence clearly and accurately, a focus on continuity can have significant impact on how we view the language of 1QM and the Scrolls-a point I return to in my conclusion.

Implications and Conclusions
More research on the argument structures of verbs in ancient Hebrew textsin the Bible and the Dead Sea Scrolls-is needed. Concerning the Scrolls in particular, Muraoka's important study included data from only a handful of Scrolls, and there is some flexibility in the interpretation of the data. And while the present study has argued for a different method and provided data and an example, much more space is needed to fully flesh out the profile of a single text like 1QM. Such studies are how we will fully chart the nature and development of the Hebrew language.
In the areas of language that I have considered extensively in 1QM-the syntax of the noun phrase,64 word order,65 and verbal argument structure, as well as numeral syntax, which I have considered in all of the Scrolls66-I see little development from earlier periods of Hebrew. There are few cases where verbal argument structure has developed from the Hebrew found in the Bible to the Hebrew of the Scrolls, correlating with what I have seen with other areas of language. This is a positive outcome. Each text, and its language, is worthy of study in its own right, even if the results do not give us anything different to contrast with biblical texts and language. More importantly, the linguistic continuity of 1QM with earlier texts must be explained. Either the native language of 1QM's author(s) is substantially the same as earlier Hebrew, or 1QM is classicizing. While I think the classicizing explanation is more likely, it remains important to recognize and highlight the many ways in which the Hebrew of the Scrolls is in line with Hebrew we find in the texts of the Bible. It is also interesting to