Biblical Hebrew Forum

[David Kummerow <farmerjoeblo AT hotmail.com>]

Hi Rolf,

Thank you for replying. Some good points are raised below (by you, that
is, not necessarily by me!).



> Dear David,
>
>
> snip
>
>
> I think it is hard for you to grasp my basic point. What I criticize
> is that so many claims regarding Hebrew verbs are made in grammars
> and textbooks but they can not be tested. For example, when I refer
> to particular passages with YIQTOLs and WAYYIQTOLs with past
> reference and ask why we know that the action of the YIQTOL with past
> reference is progressive (or something else), and the same is not the
> case with the WAYYIQTOLs, the answers are interpretations of the
> particular verses where the YIQTOLs are taken as progressive or modal
> and the WAYYIQTOLs are not. But this begs the question why, since
> progressive interpretations could equally well have been given to the
> WAYYIQTOLs. I would go so far as to say that most of the explanations
> of verbs in Hebrew grammars are circular subjective judgements, they
> are just claims that cannot be tested. This was the situation that
> really bothered me when I came from the natural sciences to Semitics.

True. But semantic tests of the like of which you have performed can be 
made. So would not rule out that it is impossible for a wayyiqtol to 
occasionally be progressive, esp. in poetry. However, I am able to say 
this because I do not operate with a theory of uncancellable semantics. 
But I still maintain that this is not the prototypical meaning of 
wayyiqtol. So your "uncancellable semantics" and my "prototyical 
meaning/semantics" do not match.

> While some of your questions are based on assumptions which takes
> them outside my Popperian approach, some are not. One such example is
> grammaticalisation.

None of my questions are really addressed by you directly. Again, you 
seek to answer them by stating your basic premises and methodology. But 
these are the very things I am questioning. Also this post:

https://lists.ibiblio.org/sympa/arc/b-hebrew/2007-March/031766.html


> Grammaticalisation means that a word form gradually developes into
> another form (e.g., a substantive which becomes a preposition; this
> may particularly be seen in Ge´ez), or that a word form with multiple
> functions gradually looses functions until it only has one function.
> A grammaticalization process from aspect to tense is absolutely
> possible. How can we detect a grammaticalisation process. The ideal
> situation would be to have a corpus of texts covering a particular
> time span, where we could point to a word form having one particular
> or multiple functions in the oldest texts, then we could follow the
> gradual change, until we in the yougest texts could see the result of
> the grammaticalisation process.  The ideal situation is not often
> obtainable, but the minimum requirement to speak of a
> grammaticalisation process is to have texts indicating a beginning
> and other texts indicating the process.

Note that grammaticalisation can also EXTEND meaning. There's plenty of 
stuff on this in the linguistic literature. Regarding BH verbs, this is 
likely to have occurred with, say, the hithpael. The range of meanings 
is neatly described by Anstey 2005: 74-76, ie reflexive, grooming/body 
motion, naturally reciprocal, anticausative, and generic passive. It 
likely that meaning has been extended over time towards the passive, but 
it has not lost its reflexive function. Some, eg W-O, claim that it 
expresses even the passive and not just generic/gnomic passive. But do 
you see the problem if I were to follow your method? What is the 
uncancellable semantics denoted by the hithpael binyan? The more we move 
to "passive" the less "reflexive" the semantic, but the more "passive" 
the semantics, the less "reflexive". However, similarity exists between 
anticausative and generic passive and similarly reflexive and body 
motion and reciprocal. But if I were to strictly follow your approach 
you have applied to the verbal system, I am indeed very hard pressed to 
find a common denominator between all of these "functions". I see this 
to be the same problem with your approach to the verbal system and this 
issue really needs addressing in my opinion.


> I will apply my linguistic theories to this situation with the
> following prediction: If the WAYYIQTOL form in the Tanakh is in a
> grammaticalisation process from aspect to tense, we will expect  to
> find the WAYYIQTOL (or its predecessor) in the older texts as an
> aspect. Then we expect to find in texts of younger and younger age
> that the WAYYIQTOL looses its aspectual properties until it, in the
> youngest texts is fully grammaticalised as a past tense. This
> prediction can logically only be applied to one language, but to be
> generous, it is granted that it can also encompass the cognate
> languages. Now, according to scientific thinking, if the prediction
> is fulfilled, what have we proved?  Nothing, because there may be
> other reasons for the pattern (cf. the problem of induction)! But we
> have made the hypothesis of a grammaticalisation process more likely.
> If the prediction is contradicted, what have we then showed? We may
> have falsified the hypothesis, but to be sure we must take Duhem´s
> problem into consideration. Let us apply the prediction.

True.

> For classical Hebrew we have the WAYYIQTOL form, which is taken as a
> fully grammaticalised past tense, or as being partially
> grammaticalised (here we cannot speak of "tense," since a "tense" is
> not partially grammaticalised). Some would say that the
> grammaticalisation process started with a short preterit prefix form,
> and others would say the process started with an aspect. Our first
> task is to find the original form. The only candidates in classical
> Hebrew are YIQTOLs with past reference interpreted as preterits  in
> Psalms believed to be old. But how can we know whether these are
> tenses or aspects? It is exactly the same situation as when
> Joûon-Muraoka says that some YIQTOLs with past reference are
> progressive and others are punctual. This is an excellent example of
> circularity, and there are no controls! This means that the start of
> a supposed grammaticalisation process cannot be found in the Tanakh.

I agree that it is difficult to see evidence of grammaticalisation in 
our texts regarding wayyiqtol. The FORM itself, whichever opinion you 
side with regarding its makeup, is a grammaticalised form. My opinion is 
the meaning of the form is fairly uniform, ie it is predominantly the 
narrative verb.

The possible arenas for possible grammaticalisation, however, which I 
have mentioned before, are not really discussed in your dissertation. a) 
the effect of the introduction of qatal into the paradigm of prefix 
verbs. b) qatal used as something of a narration verb in direct speech. 
c) the use of qatal in LBH vis-a-vis classical/standard BH (this last 
one is an exceedingly difficult question due to the dialectical issues 
in the later books, ie Judean works strongly resemble Judean Hebrew etc).



> Then one has to seek recourse in the cognate languages. But
> linguistically speaking, this is a very big leap, since we cannot
> take for granted that any form or use in one language is similar to a
> form or use in another language. Two candidates have been used: IPRUS
> and IPARRAS  in Akkadian and YAQTUL and YAQTULU in Ugaritic. The
> Akkadian connection is very problematic for two reasons, 1) the great
> difference between Akkadian and Hebrew, and 2) the difference between
> the forms is a vowel between first and second radical. This means
> that if the Akkadian contrast was found in Hebrew, we would expect a
> Qal YIQTOL and YEQATTAL or YEQATTEL in Hebrew, but we do not find
> such a contrast. The Hebrew YIQTOL is morphologically similar to the
> short Akkadian form IPRUS, and the same is true in the WAYYIQTOLs as
> well.  The second connection is between Ugritic and Hebrew. Since
> Ugaritic has only three vowels expressed by three different forms of
> aleph, any possible difference between prefix forms can only be seen
> in roots with final aleph. These are few, and there is no way to know
> with certainty that there are two prefix forms, one YAQTUL and
> another YAQTULU. The prefix forms are used with both past and future
> reference, and apparently as modals as well. The circularity in the
> interpretation of Ugaritic verbs where a short preterit is postulated
> is even greater than in Hebrew, if that is possible, since vowels are
> lacking.

I agree that Akkadian is problematic. Ugaritic I would disagree on. And 
you miss out Amarna.


> There are also other very important problems. If we could identify a
> possible predecessor of WAYYIQTOL in Ugaritic, how could we know
> whether this form was a preterit, a perfective or an imperfective
> aspect, since all three can have past reference.  Only circular
> answers are possible. The same is true in the Tanakh itself, for how
> can we know that a WAYYIQTOL is a partly grammaticalised past tense,
> or a fully grammaticalised past tense, or a perfective or an
> imperfective aspect just by looking at one particular form in its
> context?  But back to the prediction. Do we see any pattern in the
> books of the Tanakh?. We can to some extent identify the younger
> books, but the dating of the other books by linguisticAL means (the
> only ones possible) are fraught with problems. What is perfectly
> clear, however, is that it is not possible to say that the meaning of
> WAYYIQTOL in the younger books is different from the meaning in the
> other books. Some, such as Kutcher, has argued that the frequency of
> WAYYIQTOL is lower in the younger books, but a statistical test of
> this in my dissertation showed that this was not the case. Clearly,
> the prediction above has not been fulfilled.

Good and reasonable questions. You will know by running your semantic 
tests on the verbs and determining prototypical meaning. Determining 
uncancellable meaning will lead you astray in that likely either tense 
or aspect (as you have done) will have to be modified some how in order 
for it not to be cancelled. This is assuming that what we find in these 
texts are indeed non-grammaticalising verbs that are not used in any way 
for rhetorical impact etc where tense or aspect is deliberately "cancelled".



> Any approach to the verbal system of a dead language must entail
> subjective judgements. But in order to reduce these and to make
> testable results, I havechosen todealwith the text and not with lofty
> hypotheses regarding what may have happened in the past. this
> appraoch deals with semantic meaning, and the functions of the verbs
> are only tools that are used in order to isolate the semantic
> meaning. At the outset this is a very simple approach. A general
> prediction regarding tense can be: If one or more of the Hebrew
> conjugations code for tense, we will expect that verbs with one
> morphology refers to the past or to the future and not that all the
> conjugations can have past, present, and future reference. This
> prediction is not fulfilled, and my conclusion is that tense is not
> grammticalised in Hebrew. It is interesting that Waltke and O´Connor
> in principle accepts my prediction. The book says (p. 460):
>
> "How can forms each of which "represent" all three English major
> tenses have a primarily temporal value?"

Yes, and I have put the question back to you in reverse and I did not 
receive a satisfactory answer: "How can verb forms each of which 
'represent' two aspects have a primarily temporal value?"

W-O's question is only a problem if a methodology of uncancellable 
semantics is adopted. I have already drawn your attention to Matthew 
Anstey dissertation where through a constructional analysis he provides 
the very to W-O's question. The answer is there.

In regards to my question I put to you, the answer I find in your 
dissertation is that "they can't". Well, at least without modifying 
"aspect" to suit. But why this choice? Couldn't this be seen to be 
methodological arbitrary? Why not modify "tense" to suit so that some 
sort of modified tense is found to be "uncancellable" between all uses 
of a verb?

I remain at a loss to see how your methodology is able to work in 
existent multifunctionality other than in pinpointing the meaning of the 
use of each individual verbs. The overarching methodology of 
"uncancellable semantics" is, in my opinion, unable to determine 
"linguistic convention" from "semantic meaning" in a model of language 
where these are separate.


> And as mentioned above, there are not clear data for a
> grammaticalisation process. As already mentioned, exceptions that can
> be linguistically explained are accepted, but not exceptions with ad
> hoc explanations. In the discussion of exceptions  the question of
> multifunctionality will have a place.
>
>
>
> Questions:
>
> 1. Explain how your methodology is able to operate in areas of 
> incomplete grammaticalisation? Explain how it is able to explain the
>  movement from an aspectual verb to a tense verb? How is it able to 
> explain when a verb is, say, only partly grammaticalised as a tense?
> (Do you see the problem here? Once a verb has begun to grammaticalise
> as a tense, it is no longer wholly aspectual and it is not yet wholly
> tense. Hence there are two uses, with some correspondence between the
> two, eg past-perfective, future-imperfactive, etc. However, there
> will be some used where there will no longer be overlap, eg
> past-imperfective, future-perfective, etc. So in this situation I am
> unable to see how a methodology of semantic uncancellability is able
> to operate, ie the movement of grammaticalisation seems to be to
> necessarily involve semantic cancellability and bleaching.)
>
> 2. Explain how your methodology is able to operate in areas of 
> linguistic multifunctionality? Eg, nouns, verbs, etc? Explain how it
> is able to operate when there is no intrinsic similarity between the
> two uses?
>
> 3. Explain how your methodology is able to operate if "linguistic 
> convention" does not have any impact on "semantic meaning"? Actually,
>  you don't have to explain this one as it is already answered by your
>  dissertation: you have to redefine aspect to fit all the exceptions!
> You know, you could have made an hypothesis about tense and tried to
> prove or disprove that - and would have had to arrive, I strongly
> suspect, at a redefinition of tense!
>
>
> I will
> > quickly add that in the examples with non-past tense we must be
> > open for textual corruption, exceptions, special contexts, special
> > conventions in different genres, diachronic questions, and that a
> > language is a living medium etc. But these issues should be studied
> > in a scientific way, and the examples should not be brushed away by
> > the argument that languge is fuzzy, or the WAYYIQTOL is not yet
> > fully grammaticalized as apast tense. This prediction was a basic
> > tool for my thesis of 1995. The NWT had caused me to be critical
> > towards the modern theories of WAYYIQTOL, but its grammatical
> > theory did not influence me at all, since this was a study of text,
> > of all the WAYYIQTOLs of the Tanakh on the basis of the
> > falsification principle. Grammatical theories therefore were
> > unnecessary.
>
>
> Rolf, I have not attempted to brush them aside. Regarding qatal, I
> have sought to give an indication how a tense-prominent position is
> not incompatible with non-past uses. Time and again, I have sought to
>  interact with your ideas, whereas I see exceedingly little from you.
> You brush my arguments aside by saying I have brushed things aside!
> Yet you are doing this repeatedly yourself! Again, you have failed to
> interact with what I have said. Further, I sought to question your
> methodology in that I have doubts, as Peter raised also, as to
> whether it is able to
>
> [snip]
>
>
>
>
> Regards, David Kummerow. 
> _______________________________________________
>
>
> Best regards,
>
> Rolf Furuli University of Oslo

Anstey, Matthew P. 2005. “Towards a Typological Presentation of Tiberian 
Hebrew.” Hebrew Studies 46: 71-128.


Regards,
David Kummerow.