(long) b-hebrew digest: December 03, 1998
furuli at online.no
Fri Dec 4 13:55:58 EST 1998
Matthew Anstey wrote:
>I can't remember if I "recommended" Hatav (aspect-based) and Goldfajn
>(tense-based) since I do not know enough, but I did find them interesting.
>From reading these books, and speaking to people at SBL, and the recent
>posts, I am getting more confused than ever about whether the BH verb system
>is tense or aspect or both (in various proportions.) Scholars, and very good
>ones at that, seem divided about this. I don't even know how to assess these
>arguments, which probably means reading more about linguistics (sigh!) and
>I imagine others would point out problems with other steps as well, but
>these are the ones I think of when I read around. Is perhaps the problem is
>that people are trying to solve the enigma of the _verbal_ system, rather
>than the _BH_ system? That is, are we asking the wrong question when we say,
>"Is the hebrew verb system tense or aspect?" We should instead say, "Is BH
>tense or aspect?" So I was wondering if someone could suggest a better
>method if they have one and/or theoretical considerations about this. It
>seems we won't even get off the ground without a proper method, as in all
It is very fine that you ask about approach and method. When we proceed on
this road, the first point to consider in our approach is the role palyed
by induction and deduction. When we use induction we do not have a
particular hypothesis, but we try to draw conclusions from a huge number of
isolated data pointing in the same direction. The "problem of induction" is
that we by help of this method can prove nothing, because we have not seen
all the data. An inspection of 100 000 white swans do not *prove* that all
swans are white; one black swan who has not been through the fire or is
dyed will falsify the hypothesis. Discourse analysis is primarily based on
induction and therefore we cannot by help of this exercise alone identify
the nature of yiqtol, qatal etc.
In the study of dead languages, we often use induction to create a
hypothesis, which is an abstraction on the basis of our interpretation of
the huge number of data. Having created a hypothesis, we continue with a
deductive approach. We ask what our hypothesis predicts, and design tests
to see if that which is predicted turns out to be true. But again, if the
predictions turn out to be true, we have *proved* nothing, because
deduction can prove anything, it only shows degrees of likeliness.
However, deduction can falsify! The previous swan-example is in nature
The situation where induction has the strongest position, is when we have a
limited corpus, such as BHS. So it can and must be used in Hebrew studies,
but its limitations should always be kept in mind.
Deduction yields the best results when the possible answers are restricted.
If there are one hundred possible answers and we falsify one, we have not
proceeded very far. But if there are just three possible answers and we
have falsified one, we have the upper hand. I think it is possible to find
good answers to the meaning of Hebrew verbs, because the possible answers
are very few (not more than three, see below), provided that we all the
time have our eye on our methodology.
I would like to share with you the approach of the joint Hebrew verb
project, which we have in Oslo and Gothenburg. The working hypothesis is
that Classical Hebrew has just two conjugations which are aspects, and that
tense is absent from the system. What is interesting, is that others in the
project doubt the aspect explanation and believe that tense is the primary
factor in the verbal system. I am running the project, so we will test my
hypothesis, but I will have much work to do to persuade others.
The project is based on a scrupulous distinction between semantic meaning
and pragmatic meaning, and this leads to an equally scrupulous distinction
between aspect, Aktionsart and tense. Let me illustrate the distinction
between pragmatics and semantics by an example from Mari Olsen: "A Semantic
and Pragmatic Model of Lexical and Grammatical Aspect, 1997, New York:
Garland Publishing Inc. (p 17), an excellent study of the Greek verb. The
principle is that "semantic meanings may not be cancelled without
contradiction and reinforced without redundancy".
QUESTION: Are "slowness" and "tiredness" a semantic part of the verb "to
TEST: (1) "Elsie plodded along, *but not slowly."
(2) "Elsie plodded along,*slowly."
(3) "Margaret plodded along; although she wasn`t tired."
(4) Margaret plodded along; she was very tired."
CONCLUSION: (1) is contradictory, (2) is redundant, but (3) and (4) are
normal. We may conclude that slowness is a semantic part of "plod", but
tiredness is associated with "plod" by conversational pragmatic implicature.
The project uses this pattern in the study of Hebrew verbs by asking: Is
past tense (preterite) a semantic part of the form wayyiqtol (or Yiqtol),
or is past meaning a pragmatic implicature?
(a) The primary test is to find contradiction or redundancy in clauses with
wayyiqtol, provided that the form is interpreted as preterite. This means
among other things that clauses with present or future meaning containing
wayyiqtols contradict the preterite interpretation provided that linguistic
reasons cannot be found which account for their non-past meaning. If a
reasonable amount of such examples can be found, it is concluded that the
preterite hypothesis is falsified.
(b) We will in addition to the above test also check the short prefix form
of Ugaritic, Phoenician and Accadian and see if the evidence for a supposed
preterite meaning in these forms really stand scrutiny. The same test
differentiating between pragmatics and semantics will be used.
(c) We will also study unpointed texts to see if there is any difference
between yiqtols and wayyiqtols in these. If not, we have an additional
argument (though inductive) that the difference is pragmatic and not
I have gathered a lot of data already, suggesting that a preterite tense is
not grammaticalized in Hebrew (a), that the short forms of the cognate
languages are not preterites (b) and that there is no difference between
yiqtols and wayyiqtols in unpointed texts (c). If this part of the research
is finished with the mentioned results, we have to seek other explanations
for the Hebrew verb forms than tense. Linguists are used to working with
TAM-systems (tense-aspect-mood), and these are *quite* universal. If T is
taken away, A and M remains. I will not go into details, but we are also
able to remove M, and then only A remains. However, a trap would be to
apply the normal aspect definition to Hebrew. By this we would risk forcing
upon Hebrew an Indo-European property. We therefore refuse to use aspect in
a deductive-nomologic way. Let me illustrate this:
In the natural sciences, the laws of nature are used in a
deductive-nomologic way. We can illustrate this by a simple syllogism (5):
Water expands when freezing.
A is water and is freezing.
Here is the law of nature used in the upper part (explanans) and this
demands a certain conclusion (explanandum) regarding A. It is dangerous to
use the traditional definition of aspect as a law, something which can be
illustrated by the syllogism that follows (6):
The imperfective aspect (yiqtol) does not include the beginning and end of
A is imperfective (yiqtol).
A does not include the beginning and end of the event.
Syllogism (5) is true but (6) is false if it is applied to BH.
We therefore use a hypothetic-deductive approach when we try to find
another meaning than tense in the verbal system. This means that we use the
traditional definition of aspects as a hypothesis and see if it fits the
material. The rusult so far is that the core of of the aspect definition
can be upheld, but the definition itself must be readjusted. This new
definition, which we have arrived at by help of induction, is handled as a
hypothesis, and to use it in a deductive way we ask if it can be falsified.
Then we study the material to see if it is falsified.
We can outline one area where it can be falsified: It seems that the
imperfective aspect can be used in any situation whatsoever, so there is no
way to falsify the imperfective definition (hypothesis). However, the
perfective aspect cannot be used in the following situation: The subject
and object are singular and definite count nouns and the context shows that
the action is past and was not completed. A weqatal or qatal used in such a
situation would falsify our aspect definition and the hypothesis.
To analyze our material and differentiate between semantic and pragmatic
factors, we use the interplay of the different semantic planes as
parameters, the most important being: Aktionsart (the lexical content of
the verb), the verbal arguments (subject and object) - whether they are
count nouns/non-count nouns, definite/indefinite, singular/plural and
whether the phrase is telic or not. In addition we study
co-ordination/subordination, genre and discourse. To learn more about this
I recommend Alan S. Creason: "Semantic Classes of Hebrew verbs: A Study of
Aktionsart in the Hebrew Verbal System", Ph.D diss. 1995, University of
By way of conclusion I will stress the importance of studying the smallest
possible parts of speech to find the meaning of the parts of the verbal
system. All other kinds of approaches can give valuable insights, but they
cannot pinpoint the meaning of the conjugations, because the "problem of
induction" prevents that.
All the different conclusions reached in recent studies of Hebrew verbs
seem to indicate that it is impossible to reach a more permanent
conclusion. I do not think this is true, but rather than look at all the
different conclusions, we should look at the methodology of the studies. In
this way we can reject several conclusions simply because the methodology
Lecturer in Semitic languages
University of Oslo
More information about the b-hebrew