Two new(ish) works on the Biblical Aramaic verbal system

The Verbal System of Biblical Aramaic

The Verbal System of Biblical Aramaic
A Distributional Approach
Studies in Biblical Literature (Peter Lang series) – SBL 116
by Michael B. Shepherd
Peter Lang, 2008
xiv + 178 pages, English
ISBN: 9781433102011
Your Price: $64.95

The Verbal System of the Aramaic of Daniel
An Explanation in the Context of Grammaticalization
Studies in the Aramaic Interpretation of Scripture – SAIS 8
by Tarsee Li
Brill Academic Publishers, Forthcoming June 2009
xvi + 199 pages, English
ISBN: 9789004175143
List Price: $132.00
Your Price: $125.40

The Semitic verbal system has long been a source of debate, and much effort has been expended in the attempt to explain the functions of the verbal forms. Most of the work has concentrated on Biblical Hebrew (BH), but in the last year two works have come to print dealing with the verbal system of Biblical Aramaic (BA), though with vastly different methodologies. Li adopts the grammaticalization approach exemplified by the work of Bybee, Perkins, and Pagliuca (see The Evolution of Grammar, University of Chicago Press, 1994) and previously applied to BH by John Cook in his U-W Madison dissertation. Shepherd, on the other hand, adopts a text-linguistic approach following Wolfgang Schneider and John Sailhammer among others. It should further be noted that Shepherd deals with all Biblical Aramaic, while Li limits his study to the Aramaic found in the book of Daniel.

Like BH, the primary verbal forms in BA are the prefix conjugation yqtl (traditionally the “imperfect”) and the suffix conjugation qtl (traditionally the “perfect”). The BH system is further clouded by the existence of the so-called “waw-consecutive” forms, but while Old Aramaic (OA) does know an “imperfect consecutive” (better, “narrative preterite”), it had fallen out of use by the time of BA. Thus, the principal parts of BA are yqtl, qtl, and the imperative. The infinitives do not participate in the verbal system proper, but the role of the participle, which occurs both on its own and in combination with הוה, is less clear.

As in BH, the chief difficulty in describing the BA verbal system is that both the prefixed and suffixed forms can occur in past, present, and future contexts implying that the basic opposition is not one of tense. At the same time, the distribution of forms is not spread equally among the tenses. The suffixed form primarily refers to events in the past, and the prefixed form primarily refers to events in the present or future, neither form seeming to have much aspectual nuance in these cases. Thus it is difficult to argue that the basic opposition yqtl :: qtl is based on aspect. Therefore, whether an attempt is made to explain the basic opposition as tense or aspect, numerous exceptions remain unexplained.

Shepherd argues that the failure to find a satisfactory explanation based on tense or aspect is due to the fact that tense, aspect, and Aktionsart are functional categories derived from Greek and Latin, but foreign to BA. In order to determine the functional opposition of qtl :: yqtl within BA itself, a bottom-up approach must be employed using a text-linguistic methodology.  His working thesis is that the opposition is not based on tense or aspect, but narrative :: discourse. This follows Schneider’s conclusion that in BH the imperfect consecutive wayyiqtol is the narrative form, while yiqtol (and weqatal) is the discourse form.

Shepherd’s method is to divide the text at the clause level, assigning each clause to one of six levels: 1) Narration; 2) Discourse; 3) Narration in discourse; 4) Reported discourse; 5) Narration in reported discourse; and 6) Reported discourse in reported discourse. It should be noted that Shepherd has only included clauses that he has deemed to be on the main line of narration or discourse. He has therefore omitted clauses which are themselves constituents of higher-level clauses such as relative or purpose clauses (note that this does not seem to be a pragmatic distinction of foreground and background, but merely a syntactic distinction of the head phrase from complements and specifiers). Each clause is also marked for its verbal form (qtl, yqtl, imperative, or nominal clause including participles), whether the clause begins with the conjunction waw or edayin, and whether the first position in the clause is occupied by a verbal form or shows an inverted word order. His results seem to confirm his thesis as qtl occurs in 65% of main narrative clauses (levels 1, 3, and 5) in contrast to 6% for yqtl, while yqtl occurs in 53% of main discourse clauses (levels 2, 4, and 6) compared to 0% for qtl.

Unfortunately, Shepherd’s book is disappointing overall. While there are many unnecessary digressions (such as a lengthy defense of BA as a dead language), there is a glaring omission of any justification for the bifurcation of texts into narrative and discourse, no formal definition of his proposed six levels, and therefore no explanation of how he assigned clauses to these levels without relying on any assumed tense or aspectual value for the verbal forms. In fact, in his examples, Shepherd consistently translates qtl as a simple past while yqtl is translated either as a future or modal when he deems them to be on the main line. This begs the question whether the primary opposition is narrative :: discourse, or whether the distribution follows the tense and aspect of the verbal forms.

Among others, Hopper and Thompson (“Transitivity in Discourse”, Language 56/2 1980), argued that the distinction of foreground from background is universal and that languages often have several features for this purpose. For instance, many languages have a special narrative tense that is reserved for the foreground in narrative, such as the wayyiqtol narrative preterite in BH (note that qtl can occur in background clauses in BA, and is therefore not a narrative tense). Another important feature is aspect since punctual action is more likely to be in the foreground than durative or habitual/iterative actions. Thus, if narrative is loosely defined as a sequence of past events, then the distribution of qtl and yqtl may merely confirm that qtl often functions as a simple past tense and is therefore the natural choice for main foregrounded clauses in narrative, while yqtl in the past has a durative or habitual/iterative aspect and therefore does not occur in foregrounded clauses (but may occur in “main” clauses based on Shepherd’s definition). If the remaining clauses are considered discourse, that leaves non-past uses of yqtl for the foreground, in which case it must not have any special aspectual nuance. In this way Shepherd’s approach, with a better refined methodology, could be a helpful guide for the analysis of the BA verbal system by setting limits on the interpretation of tense-aspect. However, I am not convinced that the opposition of narrative :: discourse is a suitable alternative to a system based on tense and aspect.

In contrast to Shepherd, Li argues that the failure to explain the BA verbal system is not due to the unnatural imposition of the categories of aspect and tense, but the structuralist assumption that the system must be based on binary oppositions. Note that such an assumption lies at the heart of Shepherd’s study as he is motivated to find a different basic opposition since neither tense nor aspect seem to work.

Grammaticalization is the process by which lexical items come to have a primarily grammatical function, or grammatical items obtain a new grammatical function. A good example from English is “going to”, which moved from primarily indicating movement in space, “I am going to school”, to also express future tense, “I am going to read”. Grammaticalization seems to be a uni-directional process with some common paths of development cross-linguistically. Within a language there is constant development of new forms through grammaticalization so that at any one point in time, there may be more than one form available for expressing similar functions, while traces of older functions may also remain. These two concepts are labeled layering and persistence.

Thus at any given time, it is not unexpected that we should find more than one form serving similar functions or that a single form may have multiple functions rather than a neat system of oppositions. Proponents of grammaticalization argue that such a situation is better explained by locating each form along its path of grammaticalization rather than trying to relate the forms to each other. Li has chosen Daniel as the subject of such a study because it sits at the crossroads of two major periods of Aramaic, Imperial Aramaic and Middle Aramaic. This has contributed to making the verbal system of Daniel among the most difficult to explain.

Li begins his analysis with the suffix conjugation which seems to be on the path from an anterior to a simple past. In Daniel it is primarily a simple past, but it has not completely lost the old anterior/resultative function which can occur in past, present, and perhaps even future tense contexts. For example, Dan 4:28-29a:

לך אמרין נבוכדנצר מלכא מלכותה עדת מנך 29 ומן־אנשא לך טרדין

“To you it is said, king Nebuchadnezzar, the kingdom has departed from you, and you will be driven from men…”

This is the beginning of a pronouncement of future judgement, thus Li argues that עדת cannot be a simple past, but must be anterior to the following string of participles. 

The prefix conjugation, on the other hand, is on the path from an earlier imperfective to a future. The imperfective use is retained in past tense contexts with a progressive or habitual/iterative function. The prefix conjugation can also express general present and various modalities, which  probably develop from the future function.

The main wrinkle in the Aramaic of Daniel is the beginning of the incorporation of the participle into the verbal system. The active participle is developing from a progressive to a general imperfective which has begun to take over the imperfective and present from the prefix conjugation (helping to limit it to future and modal uses). It may also occasionally express modality. The active participle can be used in past tense contexts for progressive, habitual/iterative, and perhaps inceptive functions. For instance, Dan 5:5 demonstrates the past progressive:

ומלכא חזה פס ידה די כתבה

And the king was watching the back of the hand that was writing.

Unlike the prefix conjugation, which only is used for general present, the active participle can express both the general and real present. More difficult to analyze is its alleged function as a historical present. Li argues that the use of the participles ענה and אמר to introduce direct speech within narrative is formulaic, and the remaining cases can be explained as past imperfectives or regular presents. 

Li also argues that the passive participle has not yet fully entered the verbal system but is a verbal adjective which is developing into a resultative participle (and which will later become the base for the past tense in Neo-Aramaic). At the same time, the T-stem participle has become the true passive counterpart of the active participle and thus is also an imperfective.

In the Aramaic of Daniel, the participle has also been incorporated into a complex verb phrase with a conjugated form of הוה (and sometimes איתי). Here הוה functions as a tense-marker, but Li argues that the complex is in the early stages of grammaticalization as the new imperfective. Since the participle on its own can function as an imperfective, the majority of the cases are with the suffix conjugation of הוה which is being reanalyzed from הוה [past] + participle [imperfective] to הוה + participle [past imperfective]. Thus in some later Aramaic dialects, the active participle loses its imperfective function and is restricted to present tense.

While I agree substantially with Li’s conclusions, there are two main weaknesses to his study. The first is his almost exclusive reliance on the work of Bybee, Perkins, and Pagliuca. While they are certainly the leaders in the field, I am a bit wary of a methodology that takes the latest linguistic theory and demonstrates how it can be applied to the biblical text, rather than beginning with the biblical text and pulling from linguistics more broadly to explain the features. Here I can appreciate Shepherd’s hesitation in imposing foreign categories onto the text, but I think he has pushed this too far. The work of Bybee, Perkins, and Pagliuca is cross-linguistic and in my opinion demonstrates strongly that the categories of tense, aspect, and modality are indeed universal.

The second weakness is common to works on the verbal system of BH and BA. Too often justification for assigning a particular verbal form to a specific category of tense and/or aspect is based on the translation into English (or German, etc), and not on any formal criteria within the language itself. Of course, this is something of a catch-22 if the tense and aspect marking is primarily within the verbal morphology and we are left without any native speakers. Again, I think Shepherd has overstated this objection. As I stated earlier, I think his text-linguistic methodology could be helpful for establishing limits on our analysis of tense and aspect based on what type of forms we expect in which level of discourse, and I think that the grammaticalization approach is also very powerful for explaining the apparent unevenness of the system.

Explore posts in the same categories: Aramaic, Li, Tarsee, Semitic Verbal System, Shepherd, Michael

7 Comments on “Two new(ish) works on the Biblical Aramaic verbal system”

  1. Carl Says:

    How does Li deal with the discourse or text linguistic aspects? While Shepherd doesn’t draw back from the larger picture very often, your review makes it sound like Li doesn’t endeavor to address these larger issues in his work.

    • Peter Bekins Says:


      Li notes in his introduction that text-linguistic/discourse approaches are complementary to morphosyntactic explanations, but he does not deal with the former in this work as you note. Obviously he is well aware of them since his thesis dealt with sequentiality in Semitic. Shepherd tries to explain everything by text-linguistics (which he sharply distinguishes from discourse analysis by the way since discourse tends to be developed from spoken language) and seems to be agnostic about the possibility of morphosyntactic explanation.

  2. Chip Says:

    It is important to note that there has been serious debate in linguistic circles over whether grammaticalization is truly unidirectional or not. Certain changes have been shown to move in both directions on the cline of less grammatical > more grammatical, thus questioning this assumption; though in general this assumption is true at least on a morpho-syntactic level.

    Good resources are available for those interested:

    “Up and down the Cline: The Nature of Grammaticalization.” Typological Studies in Language 59 (Benjamins, 2004)

    • Peter Bekins Says:

      Yeah, that’s why I think Li may have relied a bit too much on Bybee, Perkins, and Pagliuca where I wish he would have brought a broader linguistic perspective.

  3. Carl Says:

    You are right to point out Shepherd’s strong position on Text-linguistics over Discourse linguistics. On the whole, I thought that his book was overly committed to a conclusion that I never thought he justified through his argument. What do you thing about arguments like his that are very much dependent on statistical data? What are the pitfalls?

    Are you referring to the existence of a lexicalizing cline alongside a grammaticalizing cline, or is this something else? I know that lexicalization and grammaticalization are two important concepts for language contact, but I had not thought about how they affect the diachronic development of a single language.

    • Peter Bekins Says:

      Carl, technically speaking he doesn’t use statistics, he just presents the raw data as % verb form / clause type. I guess since the findings are so overwhelming there is no need to check if they are statistically significant. Of course, that also shows that the findings are somewhat trivial.

      You have to use statistical methods if you are analyzing large amounts of data, so I don’t mind such an approach in general. The danger is garbage in, garbage out. Your answer is only meaningful if your question was meaningful. If Shepherd wants to test whether the choice of verbal form is dependent on the text-linguistic categories of narrative and discourse independent of tense-aspect, then he should have spent the bulk of his work demonstrating how to categorize clauses independent of tense-aspect.

  4. Carl Says:

    Well reasoned.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: