Biblical Hebrew Forum

[Rolf Furuli <furuli AT online.no>]

Dear Greg,

Sorry for the delay in answering you, but last week was extremely busy. I
give answers to most of your questions below.



>Rolf,
>
>You have two of three elements of what I consider exciting
>scholarship: (a) the bold claim to overturn existing conceptions;
>and (b) competence to talk about the data. All that remains
>is (c) delivery on the promise. (This isn't meant as a criticism;
>its meant favorably.)
>I have studied the issues with the Hebrew verbal system much less
>thoroughly than you but I would like to pursue a couple of things
>further.
>I asked what problems in conventional explanations of Hebrew
>that you saw your system solving and you listed several. Let
>me make a counterproposal, and you show me where my
>proposal is wrong, OK?
>
>(a) I stipulate with you that there are only two, not four, finite
>conjugations, and that the waw in waw-consecutives is nothing
>but a conjunction or particle.
>
>(b) But this requires no complicated aspect theory to explain.
>There is a much simpler explanation: its all pragmatic. Hebrew is
>showing what happens when two conjugations of finite forms got
>mixed. Whatever these finite conjugations were marking from their
>historic origins, in biblical Hebrew/QH as used they are not marking
>tense or aspect inherently, and are effectively interchangeable. The
>biggest argument for their *complete interchangeability* is the
>very fact of the waw-consecutive, in which the opposite conjugation
>follows with *exactly the same meaning* as if it was the other
>conjugation.

>
>(c) Because there is *no demonstrable difference* (hypothesis/proposal
>here--I invite you to falsify it; maybe I can learn something) in meaning
>between prefix- and suffix-conjugations in Hebrew inherently (i.e. the
>only thing going on is pragmatic conventions in the use of these forms),
>there is nothing more complicated needed to explain the waw-consecutive.
>It accounts for all of your data showing interchangeability in those forms.
>
>(d) On the other hand, the more I think about it, the aspect definitions
>which
>you propose as inherent are exceedingly difficult (for me) to pin down.
>The differences almost seem to vanish upon examination. If you could show
>classes of cases in which the waw-consecutive systematically
>doesn't happen which would be predicted on the hypothesis of your
>aspect system, that would be an argument. That is, what are some major-scale
>Hebrew phenomenae, otherwise difficult to explain (but which are
>systematic phenomenae) which would be predicted on the hypothesis
>of your theory?
>
>(e) Question: could you comment on how your aspect system treats
>wa-yitol forms of MT, 'died'? 'And he ..., and he..., and he DIED, and
>then...' What kind of process or glimpse of progressivity do you see
>in the verb mwt/died in cases such as this? It looks to me like a
>reference to the event (what you describe as the suffix-conjugation
>meaning) without any glimpse of progressivity. I cannot see that any
>continuing result after the person dying is being communicated or
>intrinsic to the communication of the verb. Since I know you must
>have thought of this, how do you treat this?
>
>Let me go through the points you cited:
>
>> than the meaning of morphosyntactic forms. I see the following advantages
>> of my model:
>>
>> 1) It outlines a dipolar system of all the prefix-forms in one group and
>> all the suffix-forms in the other, and explain the function of each form
>> without exceptions.
>
>Mine is simpler. According to what I proposed above there are only
>two systems, and the functions are explained without exception:
>(its all interchangeable, except by certain conventions from pragmatic
>grounds, which hold true except for the cases when they don't. :-)
>(The last smile is not an objection to the pragmatic explanation, since
>real-life speakers can do things for all sorts of reasons pragmatically
>and inconsistently.)
>
>> 2) It explains why YIQTOL, WAYYIQTOL, WEYIQTOL, QATAL and WEQATAL can have
>> past, present, and future reference and express both indictive and
>> subjunctive, and at the same time be memebers of an ordered and functional
>> system.
>
>All of this is equally well explained by my proposal, with a far
>less complicated proposal. Occam's Razor?
>(On 'ordered and functional', that is not an argument against what
>I proposed. A system in which two verbal conjugations without
>marking anything are used in contexts which by pragmatic convention
>portray meaning can be part of a well-ordered and functional
>language.)
>
>> 3) It explains the role of the WAW, which is prefixed to many verbs, on
>the
>> basis of simple syntax without seeking recourse in speculative ideas
>>
>> building on very little evidence.
>
>So does my system, which is far less complicated requiring far fewer
>assupmtions than yours.
>
>>
>> 4) It explains why the general narrative account often starts with a QATAL
>> and continues with WAYYIQTOLs, and why we in future contexts, but not so
>> often, find accounts starting with a YIQTOL and continuing with QATALs -
>> this is explained without equating QATAL and WAYYIQTOL and YIQTOL AND
>> WEQATAL.
>
>By my proposal, this would all be convention--useful conventions that
>speakers
>use and hearers recognize because that is how things get said--but nothing
>more complicated is required to account for these patterns.
>Serious and sincere question: am I missing something here?
>
>> 5) It explains how it is possible that the most important form in
>narrative
>> is imperfective while in non-Semitic languages this is the function of
>> perfective forms.
>
>By my proposal it is no less well explained (because neither form
>is inherently perfective or imperfective). Simpler, accounts for the
>data equally well, has good analogy in how pidgins work...
>Hebrew is different from most pidgins in that it has a longer
>development and history, etc., but is there any good reason to rule out
>that these two-conjugations are the result of language contact at
>some point in historic time, resulting in instant pidgin, and then
>developing into the Hebrew with all of the variant patterns, etc.
>familiar to us?
>Unless there is some major point I'm missing, this seems a lot
>simpler and a lot more intuitively feasible as an explanation than
>what seems a highly abstract and complex system of underlying
>rules which proves very difficult to actually verify. (Or seems so.)
>
>>
>> In short, the Hebrew verbal system is explained as a harmonious and
>> well-ordered system.
>
>Same with my proposal, not one bit less. (And less complicated.
>Occoms Razor again.)
>
>Can you show anything wrong with my proposed explanation
>that yours will account for better? i.e. why is your system a superior
>explanatory model for the five points you named? Are there any
>sixth, seventh, or eighth points that you can name in addition to the
>five you named?  The basic question is: if a simple explanation will
>work, why struggle to create a more complex one?
>
>Greg Doudna

>

----------------

SIMPLICITY VERSUS COMPLEXITY

As an ideal a hypothesis/theory/model should fulfill the following
requirements:
(1) It is simple.
(2) It appeals to our intuition.
(3) It accounts for all the data.
(4) It explains all the data better than other theories.
(5) It is falsifiable.

No theory fullfills all these in a perfect way, but we can evaluate a
theory in light of them. Your suggestion that Hebrew is a mixed language
where the forms are used interchangeably fulfills (1) and (3). That it does
not fulfill (5) is not very problematic, keeping in mind that it is a dead
language, but its clash with (2) and (4) is more problamtic.

I am not aware of any language where all the finite verbs are
interchangeable, and I will not expect to find such a language. I agree
that linguistic conventions plays an important role in classical Hebrew -
much greater than is presently realized. But if pragmatics alone was the
driving force, the result would be a language where it was difficult to
convey finer nuances. Linguistic conventions work on the macro-level, but
the fact that we can find patterns on the micro-level along the axis
YIQTOL/WAYYIQTOL/WEYIQTOL - QATAL/WEQATAL suggests to me that there is a
semantic difference between the forms.

It is true that your suggested model is simpler,but the simplest linguistic
explanations are not necessarily the correct ones. Take Sumerian as an
example. Its verbal system is not fully understood, but is is quite clear
that it is a split-ergative language where half of the verbal system
represents an ergative language (subject in intransitive clauses are marked
similarly to objects in transitive clauses) and the other half is an
accusative language. The system is very complicated but it works. So while
I seek simplicity I do not see the higher complexity of my model as being a
problem.

ARGUMENTS AGAINST INTERCHANGEABILITY

I suppose that you accept that the participle in Hebrew must have another
inherrent meaning than the finite forms; it can for instance be used
attributively, which is not possible with the finite forms. However, when
it functions as a verb it is in different contexts used in the same way as
YIQTOL. We find the same situation in other Semitic languages that nobody
would say are pidgin-languages. The point here is that forms with different
inherrent meanings in some contexts can be *used* interchangably, not
because they are semantically interchangeable, but because *some* of their
meaning is similar.

This is a well-known principle in privative models. For instance,
characteristics such as telicity, dynamicity, and durativity are *semantic*
properties. A verb such as "run" can in no context stop to be durative,
because it is inherrently durative. However, while there are groups of
verbs which are marked (+durative),(+telic), and (+dynamic), there are no
homogenous groups which are marked (-durative),(-telic), or (-dynamic).
Thus there is a clear assymetry where some verbs are (semantically)
*marked* with one or more of the (+)characteristics, but where
interpretations generally attributed to the negative features arise, it is
a result of conversational implicature. This means that unmarked members in
a privative system can in some contexts have the same meaning a marked
member, but the reverse is not true. Look at (1) and (2) below.

(1) Peter Smith ran.

(2) Peter Smith ran a mile.

The verb "run" is not (+telic), but neither is it (-telic). In (1) it
denotes an atelic activity, but in (2) it is a telic accomplishment.Thus
"run" can in some contexts have the same characteristic as verbs which are
marked (+telic).
Applied to WAYYIQTOL/QATAL we learn from this that two members of a system
can have the same function, one because of semantic meaning and the other
because of conversational pragmatic implicature. it is therefore not
logical to draw the conclusion that because two forms function similarly
they are used interchangeably. We therefore need a test which can
distinguish between semantics and pragmatics in narrative comntexts.
Because all verbs in narrative *must* have past reference, the test should
be made in other environments where a distinction is possible to find.

In my model one member of the system is marked positively, namely the
imperfective one. This means that YIQTOL/WAYYIQTOL/WEYIQTOL will allways
make visible action in progress/a state in duration (or modality) while
QATAL/WEQATAL either make visible terminated actions/states or action in
progress/a state in duration. Examples of the privative system where QATALs
have the same characteristic (Unterminated event/state) as the marked
member (YIQTOL) are seen in (3) and (4).

(3) Jer. 8:7 Even the stork in the sky knows (QATAL) her appointed seasons,
and the dove, the swift and the thrush observe (QATAL) the time of their
migration. But my people do not know (QATAL)the requirements of the LORD.

(4) Ex. 10:3 So Moses and Aaron went to Pharaoh and said to him,  "This is
what the LORD, the God of the Hebrews, says:  'How long will you refuse
(QATAL) to humble yourself before me? Let my people go, so that they may
worship me.

In (3) there are three states which were not terminated at reference time;
in (4) we see ongoing action at reference time. We know that all four
QATALs are unterminated for pragmatic reasons: the first two are "gnomic",
the third must be interpreted as unterminated as an analogy to the gnomic
ones, and the fourth is seen as unterminated because of the context. If
terminated action/state was an intrinsic property of QATAL,these examples
were impossible.

The arguments above show that WAYYIQTOL/QATAL *can* have different meaning
even though they to a large extent are used in the same larger contexts.
The primary reason why I think they *are* different is that we find very
different patterns among WAYYIQTOL and QATAL in "smaller" contexts.
Linguistic convention (pragmatics) as a reason for the choice of particular
forms works well in greater contexts such as in narrative, imperative,
jussive etc, but in the hundreds  of other more special situations, a clear
pattern along the imperfective/perfective axis suggests that the meaning of
the forms is the reason for its choice. I illustrate the typical patterns I
have found below.

B= Beginning of event
E= End of event
xx= intersection of event time by reference time (which part of the event
that is made visible)

IMPEERFECTIVE

Conative:       -xx-B-------------E---
Inceptive:      ----Bxx-----------E----
Progressive :   ----B-----xx------E---
Egressive :     ----B------------xxE---
Resultative:    ----B-------------xEx--
Gnomic:             No special stress

PERFECTIVE

Ingressive
(state only)    ----Bxx-----------E---
Constative:     ----BxxxxmxxxxE---
Constative:     ----Bxxxxx------E---
Gnomic:             No special stress
Perfect:        ----B-----------xEx--


I have already mentioned clauses with "until this day" where there is a
systematic difference between the choice of forms -WAYYIQTOL  for
resultative and and QATAL for equal stress. I have also mentioned conative
situations and situations where one event "breaks into" another, which both
are expressed by imperfective verbs. Your model cannot explain these
patterns.

COGNATE LANGUAGES

While each language must be studied in its own right, the fact that we find
a pattern in the other Semitic languages which parallels Hebrew, argues in
favour of a dipolar system in Hebrew. Particularly is Biblical Aramaic
interesting. QATAL is here the normal form used for past reference, and
YIQTOL for present and future. However, both QATAL and YIQTOL can be used
with past, present, and future reference, and many YIQTOLs function exactly
as WAYYIQTOLs in Hebrew.



>(e) Question: could you comment on how your aspect system treats
>wa-yitol forms of MT, 'died'? 'And he ..., and he..., and he DIED, and
>then...' What kind of process or glimpse of progressivity do you see
>in the verb mwt/died in cases such as this? It looks to me like a
>reference to the event (what you describe as the suffix-conjugation
>meaning) without any glimpse of progressivity. I cannot see that any
>continuing result after the person dying is being communicated or
>intrinsic to the communication of the verb. Since I know you must
>have thought of this, how do you treat this?


The reason why we know that the clause "he was reaching the top" means "he
was on the point of..." is that we *know* the nature of the English
participle. Therefore we expect to find progressive action somewhere, and
the only possible place is immediately before the top is reached. If we did
not know the nature of the participle, we would not have looked for
progressive action. This is very important, because it illustrates that in
a dead language such as Hebrew, we will not "see" progressive action every
time the imperfective aspect is used (if aspects exist at all, to take your
suggestion into account). However, a native speaker who had the Hebrew
presuppositon pool, would know the nature of WAYYIQTOL  and YIQTOL, and he
or she would "see" things in all the examples that we do not see, because
we have another presupposition pool. Our task, therefore is to look for the
few examples which come as close as possible to the ideal situation of
minimal pairs, that is, examples where the combination of lexicon,
Aktionsart, syntax etc. reduces the possible interpretations to one or two.
A reasonable number of such situations can help us pinpoint the meaning of
a form.

I take the verb MWT as an "achievement" in the Vendlerian sense, a
situation which is paralell to "reach the top". Can we see any progression
in a WAYYIQTOL of this word? To illuminate this question we should look for
YIQTOLs with past reference which are achievments. As to the verb MWT
itself, take a look at (5)

(5) Jer. 38:9  "My lord the king, these men have acted wickedly in all they
have done to Jeremiah the prophet. They have thrown him into a cistern,
where he will starve to death (WAYYIQTOL) when there is no longer any bread
in the city."

The most likely interpreteation of the WAYYIQTOL is "on the point of
dying". Alternatively, it can be viewed as a simple future or even as
modal. But in no case is the event terminated. If the "egressive"
interpretation is correct, this is evidently not the way most WAYYIQTOLs of
MWT should be taken, but in a good many instances it may be the thought.
The opposite of MWT, namely YLD, may throw some light on our question.


(6) Job 3:3  "May the day of my birth (YIQTOL) perish, and the night it was
said,  'A boy is born!'"

(7) Job 3:11  "Why did I not perish (YIQTOL) at birth, and die (WEYIQTOL)
as I came from the womb?


(8) Job 15:7  "Are you the first man ever born (YIQTOL)? Were you brought
forth before the hills?"

In (6), (7), and (8) there are three YIQTOLs and one WEYIQTOL, all being
achievements and having past reference. The only natural interpretation
that I can see, is to view them as resultative, because the "end" was
reached in all of them.

The very important role of resultative events in Hebrew has been
dramatically underestimated. In Accadian this characteristic can be
expressed in two ways, by stative and by the D-form, or by both - so it
evidently was important. It seems to me that this is the most widespread
characteristic of Piel verbs in Hebrew, and in addition, it can be
expressed by a combination of Aktionsart and the imperfective aspect, as in
(9) and (10).

(9) Josh. 7:6 Then Joshua tore his clothes and fell (WAYIQTOL) facedown to
the ground before the ark of the LORD, remaining there till evening.

(10) 1Sam. 19:24 He stripped off his robes and also prophesied in Samuel's
presence. He lay (WAYYIQTOL) that way all that day and night. This is why
people say,  "Is Saul also among the prophets?"

There is just one verb expressing two different thoughts - to fall, and to
lie. I claim that the meaning "lie" is not lexically but aspectually
conditioned.

So back to MWT. There is absolutely no suggestion of the existence of an
immortal soul or spirit in the OT. Men die exactly the same way as animals
do, and experience the very same condition (Ecc 3:19,20). However, it was
extremely important for the Jews of old to be put in a grave in order to
"be gathered to one's fathers" (As a contrast, see Is 14:19). While
"resurrection" is not explicitly mentioned before Daniel 12:13, it probably
is implied in Isaiah 26:19; Hoshea 13:14, Job 14:13-15. It is also
interesting how Jesus saw the doctrine of resurrection in the way Abraham,
Isac and Jacob were mentioned (Luk 20:37,38). So in my view WAYYIQTOLs of
MWT  with past reference are resultative (compare Deut 31:16, 2 Kings
22:20). What is stressed, is  that people have ceased to live, and that
they (their lifless corpses/bones) rest with the bones of their fathers,
all being written in the scroll of remembrance (Mal 3:16), and waiting for
God's promises to be fulfilled.


CONCLUSION

Lack of time prevents me from answering all your questions thoroughly. By
way of conclusion I will say that I do not view the Hebrew conjugations as
interchangeable in meaning because,
a) the hundreds of smaller patterns indicate that particular forms were
choosen to express different nuances along an axis which conforms with an
aspectual difference (Most important).
b) There are no real signs that Classical Hebrew is a pidgin language
(Sperber's problematic texts can be accounted for).
c) No other language has ever existed where there is no semantic difference
between what seems to be different conjugations.
d) In the cognate languages we find an opposition between prefix- and
suffix-forms which is parallel to the QATAL/YIQTOL opposition of Hebrew. By
seeking aspectual explanations the consecutive forms can be explained in
accord with this opposition.
e) An aspectual difference can explain the whole verbal system.



Regards

Rolf


Rolf Furuli
University of Oslo