Questions for Rolf

Rolf Furuli furuli at online.no
Fri Apr 27 09:34:18 EDT 2001




Dear Greg,


Se my comments below.


>Rolf, thanks for your answers. Comments below.
>
>> Because we have no informants, the task to define the Hebrew conjugations
>> is a real challange. The greatest problem is that to substantiate a model,
>> we need clearcut examples which must fulfill many requirements, and
>because
>> of the strict requirements, the examples *must* be few. For instance, when
>
>Yes, the issue is substantiating a model. I am still not understanding
>entirely:
>could you spell out the kinds of structures that would falsify your model,
>if they were to exist (but which you have checked and don't exist)?

I worked with the  natural sciences before I started with languages, and I am
influenced by Popper's requirements for falsification. I see no difference,
in principle, between the approaches of the natural sciences and
linguistics/philology. However, the requirements of falsification is more
difficult to meet in connection with dead languages than in the natural
sciences. For instance, when we claim that the Hebrew participle is a
verbal noun, which tests could be designed to falsify this claim,
particularly when we keep in mind that the participle can both have verbal
and nominal functions?

I am not hiding myself by the last question, but I am rather pointing to a
difficulties in the study of a dead language. In connection with models, the
tendency is that one that has a great potential for explanation has a weak
potential for falsification, vice verse. I claim that the model I use in
connection with Hebrew verbs can explain the whole verbal system without
exceptions, and this would suggest that its falsification potential is low.

The greater part of my work in connection with my dissertation has a
negative aim, namely, to show what the Hebrew conjugations are not. On this
basis I then try to show what they are. There are no final conclusions in the
study of a dead language, so everything is tentative. But it is easier to
show what a form is not than what it is. My first interest is to
distinguish between "semantic meaning" (characteristics of a form that
can not be cancelled) and "conversational pragmatic implicature"
(cancellable meaning applied to a form by means of the context).

The falsification test for the traditional explanations of Hebrew grammar is

simple. Two of the predictions are:

(1) If WAYYIQOL *is* preterite, all examples of it (save those which can be
systematically explained as special cases) must have past reference.
(2) If WAYYIQTOL is the opposite of YIQTOL, having the same meaning as
QATAL, the intersection of event time by reference time will be
systematically similar in cases of WAYYIQTOL and QATAL, and systematically
different in cases of WAYYIQTOL and QATAL.

I claim I can demonstrate by the patterns of verbs in Classical Hebrew that
both predictions are falsified. Thus there are two cunjugations and not
four. But because of the problem of induction and Duhems problem we cannot
at the ouset know the real meaning of these two conjugations.

I have analysed the conjugations with the most sophisticated paramaters of
which I am aware, and the conclusion drawn as to their meaning, is that
three traits of interplay of event time and reference time are similar with
English aspects and three are different. The QATAL/WEQATAL will in most
cases, but not in all include the end of a situation, and the lack og
conclusiveness prevents me from making a prediction which can be falsified
regarding this form. YIQTOL/WAYYIQTOL/WEYIQTOL are allways "open", in the
sense that they never include the end of a situation (the end of an event
can be included in resultative situations, but then the end of the
resulting state is not included).

After reading this, I can hear several members of the list with a smile say
"But then the WAYYIQTOLs cannot be imperfective, because their essence is
to portray consecutive, terminated past events." Sorry boys! This is
traditional thinking and bad linguistics! The absolutely worst place in
Classical Hebrew to learn the the real nature of verbs is in narrative
contexts. Because, as Comrie has correctly pointed out for languages
generally, the factor of these contexts which signal past, terminated,
consecutive events is *not* the verb form but the narrative itself! Leaving
intrusions alone, it would be a contradiction of terms to say that
narrative consisted of anything but past,terminated, consecutive events
regardless of which verb forms were used. So there is no way to sort out
semantic meaning from conversationsl paragmatic implicature in narrative.
This means that the predictions which can be falsified must be restricted
to situations where semantics can be distinguished from pragmatics, and
this leads to the following prediction:

(1) To express conative situations, resultative situations, and situations
where one clause interrupts another in the middle ("When Peter came,Roy was
reading.") only YIQTOL/WAYYIQTOL/WEYIQTOL are used. The use of
QATAL/WEQATAL in such situations will falsify my definition of verbs.


A note on "resultative situations": Any event can be said to have a result,
but the term "resultative situations" refer to those where the result is
stressed. Because it is not easy to say when the result is stressed, the
falsification possibility must be restricted to situations where this is
clearly the case. Clauses with the adverbial "until this day" are such
cases.

So far I have studied the Tanach in order to classify each form and make my
statistics. Now I am reading the text slowly to look for peculiarities. So
far I have seen many examples which corroborates the prediction, but I have
not studied the whole text from this point of view of the prediction. The
prediction,therefore, is just as big a challenge to myself as it is to
others. I am eager so see if it holds.


GD:
>Would you comment on a fourth alternative (not that I'm advocating it): it
>draws
>on Speiser long ago who argued biblical Hebrew was an amalgamation of
>two systems. I probably don't have Speiser understood right, but no matter:
>the point is -- hypothetically -- it could be simple historical accident
>that

>there are two conjugation systems, and *nothing* is being marked by form,
>i.e. its all pidgin with meaning understood from helping words (and some
>conventions or patterns have evolved).

This fourth alternative was also advocated by Sperber from the late 30s on,
and he was a dilligent student of the text, not only of other's gramatical
work. The possibility is there, and at least dialect differences in the
text has been proposed. However, much of this is speculative, such as
Dahood's work with Ugaritic and the Psalms. The arguments should be taken
seriously, but in my view the arguments are too weak for the following
reasons:

1) The text of the Tanach has been remarkably stable during several hundred
years (only minor developments in the youngest books) while colloquilal
language undoubtedly changed.
2) The problem of seeing a clear pattern in the verbal system (the hundreds
of exceptions) is not the fault of the text, but rather the fault of
traditional grammar which never has been tested against the whole text. I
claim that an alternative grammatical explanation can account for the whole
system. So there is no longer any need to explain the language as
anamalgamation (an influence of other languages, particularly Aramaic, is
however seen).
3) The overall system of Hebrew verbs is very similar to the cognate
languages. The Aramaic of Daniel, for instance, has many YIQTOLs with past
meaning, and YIQTOL  and QATAL can be used for past,present, and future.
The "preterite" of Accadian can have present and future reference, and the
"present" can have past reference.

>
>That is, the elimination of P and M I am not sure how much
>of a positive argument that is for A.
>
>Could you cite examples where the waw-consecutive expectation is
>interrupted in such a way as to illustrate the intrusion of considerations
>related to your aspect considerations--in those cases in which it does
>matter for meaning?

I am working to gather *all* such examples, but because I am among "wolves"
(a friendly remark for those who follow a critical method and are eager to
pin me to the wall). I will wait to present such examples until I find the
clearest ones.

>
>> As to patterns, conative situations are only expressed by
>YIQTOL/WAYYIQTOL,
>> and in situations like "When Peter entered (simple past), Roy was reading
>> (imperfective), the first is expressed by QATAL and the second by
>WAYYIQTOL.
>
>I don't understand how the empirical observation of these two patterns
>(no objection to that) provides a positive argument for your
>proposition that forms are marking aspect, i.e. I don't see how the
>conclusion follows from the fact cited. Could you maybe restate
>or make this clearer?

This is a classical linguistic test for imperfectivity. Provided that
imperfectivity means that the focus is on the nucleus of event time, and
the end is not focused upon, the clause "Roy was reading" must be
imperfective. If such a clause is intersected by another clause, we can
draw the conclusion that the clause being intersected is imperfective
(provided that aspect is found in the particular language).

>
>> The factor which create most confusion and misunderstanding in Hebrew
>> verbal studies, is that the aspects are ascribed a role in communication
>> that is exaggerated. I will go so far as to say that in most (more than
>> half) of the passages the aspects conmtribute very little to the
>> communicated meaning; so they could have been dropped altogether. And
>there
>
>Interesting. Then to prove that form is marking aspect will have to focus
>on cases within the other half.

Yes.

>
>> is no wonder that sentences with different aspects seem to have the same
>> meaning, provided that my claim is true. This is illustrated by the 470
>> doublettes in the Tanach (a few triplets). There are at least 69
>> differences in the use of verbs, and these can hardly indicate a
>difference
>> in meaning, because the verses are similar. The list below show which
>forms
>> alternate:
>>
>> WAYYIQTOL - QATAL = 24
>> WAYYIQTOL -YIQTOL = 7
>> WAYYIQTOL -PARTICIPLE =2
>> WAYYIQTOL - INFINITIVE = 7
>> WAYYIQTOL - WEYIQTOL =2
>> YIQTOL - IMPERATIVE =4
>> YIQTOL - INFINITIVE =1
>> YIQTOL - QATAL =6
>> YIQTOL - WEQATAL = 4
>> YIQTOL - PARTICIPLE =1
>> WEYIQTOL - WEQATAL =2
>> WEYIQTOL - INFINITIVE = 1
>> QATAL - PARTICIPLE = 2
>> OTHER DIFFERENCES = 8
>
>(First reaction to above is: sort of looks like they could have
>drawn them out of a hat and gotten the same meaning, which
>would be determined by contexts and helping words.)

I understand the reaction, but an important point is that the aspects are
not equipollent but rather privative. They are not exact opposites, and
some of their characteristics can overlap. This is shown by the empirical
fact that both aspects (expressed by YIQTOL, WAYYIQTOL, WEYIQTOL, QATAL,
and QATAL) can be used  for indicative and subjunctive and have past,
present, and future reference. Generally, however, there is a clear
preference for the use of each form bacuse of linguistic convention. This
is shown on the discourse level as well. Alviero Niccacci has shown a great
consistency in the use of particuler forms in particular discourse
environments. He interprets the discours functions as the "meaning" of the
forms, I interprete them in pragmatic terms.

>
>> What basically communicate, are lexical meaning and Aktionsart,
>singularity
>> and plurality, definiteness and indefiniteness of nouns, whether a noun is
>> countable or non-countable, the syntactic role of adverbials, whether the
>> verb phrases are durative, dynamic, and telic or not, etc. These and
>others
>> are objective factors in the clauses which together create mental pictures
>> of events and states. And these objective factors are there without the
>
>Good point. You almost make a good case that biblical Hebrew could be
>a well-functioning language without any distinction between yqtl and qtl
>beyond word-sequencing conventions. What in the data causes you to
>stop short of this?

All languages have different conventions. In English, word order helps us
differentiate between subject and object, and in Hebrew will the definite
article help us distinguish between adjectives which are predicative and
which are attributive. These are conventions, but word order could be free
in English if there were cases etc. My point is not that the aspects (the
Hebrew conjugations) are unnecessary, but rather that the basic meaning of
propositions is expressed by  lexicon and Aktionsart, and it is modified by
syntax. The morphosyntactic forms aspects simply are peepholes  of
different size and focus through which we see a particular part of the
meaning which already is there. Therefore, in many cases aspects do not
contribute anything extra to the conveyed meaning, so in principle both
aspects can be used. However, the aspects are not used haphazardly because
a particular convention has evolved, just as in the case of word order in
English. In particular cases the nature of the aspects works very well in
the interplay of other factors to create particular nuances. So aspects are
important for the language.

For example, the imperfective aspect is used in cases where the stress is
on the beginning and the continuance of an action (e.g. "Solomon bagan
(WAYYIQTOL) to build the temple"). This accords well with its focus on a
small area with details visible. The same aspectual logic can hardly be
applied in all the cases where QATAL is used to mark the entrance into a
state, which is comparable as respects the area of stress. This is,
however, linguistic convention, and one possible explanation for the
choice, is that because the slot of nascent action (logically) was occupied
by the imperfective aspect, the perefective aspect was chosen to express
the same stress in states, i.e. the entrance into the state.


>> In the imperfective (12) and (13) the event was terminated a long time
>> before "this day", but the resulting state lasted "until this day". In
>> several of the imperfective clauses the subject did not live "to this
>day".
>
>Maybe I'm misunderstanding, but in all of the examples you cited the forms
>seem to be predictable and explicable on the basis of routine word-order
>sequencing. Therefore since the forms are well-accounted for in terms
>of a pattern that we know is operable (which has no known demonstrated
>relationship to aspect), where is the necessary invocation of some other
>factor (aspect) to explain the data? That is, how does one know any
>other factor is operating than word-order conventions?

It is true that word-order is a part of the linguistic convention of
Hebrew. In narrative, an element comes before a QATAL  and the sequence is
continued by WA(Y)+YIQTOLs. A future context often starts with a YIQTOL
with a preceding element followed by one or more WE+QATALs. Subjective
expressions such as direct speech (QATAL) and jussive (YIQTOL) are often
sentence initial. However, word order is no strait jacket, as seen in (1).
When the author wanted to use an adverb in order to stress a certain
element, it was no longer possible to use a WAYYIQTOL, and therefore a
YIQTOL  was chosen (See  2Chr. 21:10 where the same words are found). Some
grammars have noted that there are several instances where )FZ is used
before YIQTOLs with past reference and vague hints are given that in such
cases it is understandable that the YIQTOL has past reference. But )FZ can
refer to the future as well. Nearly all the 762 YIQTOLs with past reference
in the Tanach and the 965 QATALs with future reference has a syntactic
element preceding, and in several cases, as in (1) below, would we expect a
WAYYIQTOL in that position if this particular element was removed. And
similarly with WEQATAL. There are instances where we would expect a WEQATAL
but where we find a preceding element which prevents the use of WAW.


(1) 2Kings 8:22 "To this day Edom has been in rebellion (WAYYIQTOL) against
Judah. Libnah revolted (YIQTOL) at the same time."

(NB. the NIV's rendering "has been in rebellion" does not catch the
resultative situation. The adverb )FZ hardly refer to the whole period
"until this day", but rather to the first part of the period when the
rebellion occured; the resultant state holds at "this day".)



>
>> (10) 1 Samuel 12:2 "I have walked (QATAL) before you from my youth until
>> this day"
>>
>> (11) 1 Samuel 8:8 "In accord with all  they have done (QATAL)from the day
>> of my bringing them up out of Egypt until this day"
>>
>> (12) Joshua 8:28 "Then Joshua burned Ai  and reduced it to an eternal
>> lasting mound (WAYYIQTOL) , as a desolation down to this day. "
>>
>> (13) Joshua 8:29 "and raised (WAYYIQTOL) a great heap of stones over him,
>> down to this day. "
>>
>> Resultativily is widespread in Hebrew, not only in Piel, but also in the
>> other stems when the imperfective aspect is used. To convey this notion
>the
>> imperfective aspect is important.
>
>(Could you explain how resultativity is related to imperfect, by your
>definition, rather than perfect, by your definition? Focus on the event
>as a whole, and focus on a part of the process in the event--don't
>both give results?)

In Chinese there is one resultative conjugation which is imperfective and
one which is perfective, so both can work, depending on linguistic
convention.
All events expressed by both the imperfective and perfective aspect has a
resultant state, the event leads to something. In connection with
communication, the question is what is made visible, and what is not. If
the imperfective aspect is a small closeup view where either progressive
action or an ongoing state is made visible, we will allways expect to find
this somewhere when the imperfective aspect is used. This is just as when a
letter in a word falls away, because then we expect to find a compensation
somewhere, a doubling of a consonant, a lengthening of a vowel, or (in
Aramaic, nasalization). When the perfective aspect is used with past
reference, in most instances the whole event is seen including its end, and
there is no need to stress a resulting state. If a past reference to the
event "rebel" is expressed by the imperfective aspect, the Hebrew people
would understand this as progressive action or resultant state, even though
we, who do not have the same presupposition pool and from childhood do not
know the meaning of the aspects, do not see it. However, in situations
where an event occurred and was terminated at some time in the past, and it
is said to last "until this day" or similar, an event resulting in a state
is clearly seen. In such clearcut cases the nature of the imperfective
aspect would suggest that it was used. The focus is then not ingressive,
continuative, or egressive, but it is on the details of a small area
including the end. The end is in the middle of the peep-hole as it were,
and because the imperfective aspect is not closed, the focus is on the end
and what follows; and what follows is a resultant state.

>
>> Origen in his Hexapla does not distinguish between WE- and WA(Y) as
>> prefixes to YIQTOL; both are transcribed as OU-. Several manuscripts with
>> Palestinian pointing differ from the Masoretic text as to pointing;
>> WEYIQTOLs are pointed as WAYYIQTOLs vice versa. I think there is good
>> reason to believe that while the Masoretes were extremely faithful to what
>> they heard in the synagogue, and would not dream of inventing anything
>new,
>> the difference between the "consecutive" forms is graphically their
>> invention. There is evidence that the default pronunciation of shewa in
>> Masoretic times was an "a"-sound, and I think that the Masoretes
>> "semantically" would use shewa and patah indiscrimately but their phonetic
>> rules distinguished between them. My suggestion, therfore, is that the
>> difference between WAYIQTOL and WEYIQTOL  was invented by the Masoretes,
>> not on semantic grounds (they were not grmmarians and did not even know
>> about the triradical nature of roots, and their Massora is not
>> grammatical), but partly because of their phonetic rules, and partly
>> because they saw little or no "semantic" difference between patah and
>> shewa. However, from the tenth century onward, in the infancy of hebrew
>> grammar, the graphic and phonetic differences in the text were interpreted
>> as *semantic* differences, and the four-component model of Hebrew
>> conjugations was born. This model have survived to the present.
>>
>> If there is an important semantic difference between WEYIQTOL  and
>> WAYYIQTOL, why did not the scribes at Qumran mark this,in order to avoid
>> ambiguity, when they else used so many plene vowels?
>
>This is very interesting and your last point seems to be a very good point.
>
>In some Aramaic
>> inscriptions, shewa is written as YOD. Are you aware of any examples from
>> Qumran where patah is written plene?
>
>Yes (though it seems very uncommon). MT xqt (with patah second syllable)
>appears
>several times in 11QT(a) as xwqwt. At 4QpNah 3-4 ii 3 and 1QH 3.30  lhwb
>corresponds (presumably) to MT lahab. (But the pN case is a variant produced
>by scribal error from lhwb wbrq of the Nahum quotation.)
>Compare a brief discussion of this very point at D. Schwartz, _Studies in
>the
>Jewish Background of Christianity_ (Tübingen: Mohr): 86-87n27.

Thank you for the information and the reference. This shows that to write
patah plene in cases of WAYYIQTOL (if they distinguished between WEYIQTOL
and WAYYIQTOL) would be inside their system of writing. Are there examples
when shewa is written plene?

>
>Well thanks,
>Greg
>


Regards

Rolf


Rolf Furuli
University of Oslo





More information about the b-hebrew mailing list