WAW the conjunction

Rolf Furuli furuli at online.no
Tue Jan 11 11:04:54 EST 2000

Peter Kirk wrote:

>Dear Rolf,
>I'm not sure your challenge is a fair one as I have not been making
>any claims, but only examining your ones. Also I have not completed a
>long research programme into this subject and so cannot give any
>definitive answers. But I will take up your challenge to put forward
>some definitions. Now I know that these definitions are far from
>perfect and will need to be refined. I also know that there are
>counter-examples, though perhaps not as many as you might think once
>the texts are properly understood. Furthermore, I cannot prove these
>as rigorously semantic according to the criteria you use. Another
>objection is that of course anything as brief as this is
>over-simplistic, especially that it does not take into account
>discourse factors. Also these defintions are intended to apply to
>"standard" rather than "late" Biblical Hebrew, and their applicability
>in poetry needs to be examined further. But (bearing in mind these
>caveats) here goes with some definitions - rather, some distinguishing
>features of the Hebrew verb forms:
>QATAL:          relative past, non-sentence initial*
>WAYYIQTOL:      relative past, sentence initial
>YIQTOL (long):  relative non-past, non-sentence initial
>YIQTOL (short): relative non-past, modal?
>WEQATAL:        relative non-past, sentence initial
>WEYIQTOL:       relative non-past, modal?, sentence initial
>* QATAL may be sentence-initial at the start of direct speech only, in
>this context it is used instead of WAYYIQTOL.
>I need to study further modality in relation to the short YIQTOL and
>WEYIQTOL; I am not ready to offer any definition of modality.
>As for "relative past" and "relative non-past", I am referring to the
>relationship between the end of the event time and the reference time;
>or you might want to call "deictic point" what I am calling "reference
>time". One might use the terms "perfective" and "imperfective" here,
>as the essence of my "relative past" is that the event is completed at
>the reference time, and of my "relative non-past" that it is not
>completed. Now I realise that this definition does not tie in fully
>with Galia's idea of WAYYIQTOL creating a new reference time. I need
>to think more about this, to see if there is a real difference of
>thinking or simply one of terminology.
>Genesis 2:4-7 is a good example for this: the end of verse 4 sets up
>the reference time, verse 5 gives events which took place after this
>reference time (YIHYEH, YICMAX), hence relative non-past, YIQTOL;
>verse 6 gives events which had already started at this reference time
>and which continued until later (YA`ALEH, WEHI$QAH), hence again
>relative non-past, YIQTOL and WEQATAL; in verse 7 we come back to
>events which took place and were completed at the reference time
>(WAYYICER etc), hence relative past, WAYYIQTOL.
>You might object that "sentence initial" is not a semantic category. I
>could rewrite this as saying that the verb is non-sentence initial if
>it is negative or if some other sentence component is fronted for
>Over to you. You are welcome to comment on this, but you have also
>promised to give your definitions.

Dear Peter,

Thank you for accepting my challange. Looking at your table and your
explanation of it, your definition seems to exclude tense and be wholly
aspectual (as you yourself imply, position in the clause is just an
observation, not a definition of meaning). You also show that your
definition of aspect is different from Comrie's and similar to Broman
Olses's definition. That is, you say that the perfective aspect
(encompassing QATAL and WAYYIQTOL) is not only "complete" but also
"completed", and the imperfective aspect (encompassing long and short
YIQTOL,  WEQATAL, and WEYIQTOL) is not completed. Modality (not indicative)
is associated with the imperfective aspect. In my eyes your definition is
quite clear.

As for my approach to Hebrew, I will give a *sketch* of it below. I started
without any linguistic presuppositions regarding Hebrew (at least I tried
to do that; we all have our presuppositions).
What I wanted, was to "measure" the Hebrew verbal system in light of the
principal traits of language in general, namely tense, mood, and aspect
(TAM). To achieve that I needed a method by which the "semantic" parts of
the TAM-structure could be singled out, and I found such a method by help
of the work of Mari Broman Olsen. Any model of the verbal system of a dead
language will entail some circularity, because some assumptions must be
made along the way; but it is important to find a model that reduces
circularity as much as possible.

You mentioned that you had some problems with understanding my fundament,
and I will express the difference between "semantic meaning" and
"conversational pragmatic implicature" with the principle of H.P. Grice:
"Semantic meanings may not be canceled without contradiction or reinforced
without redundancy". Broman Olsen illustrates the point by following

(1) (a) Elsie plodded along,#but not slowly.

    (b) Elsie plodded along,#?slowly.

(2) (a)  Margaret plodded along, although she wasn't tired.

    (b)  Margaret plodded along, she was very tired.

Her point is that "tiredness" is associated with plod by conversational
pragmatic implicature (it can be canceled), while "slowly" is a part of the
semantic meaning of "plod",i.e. "slowly" can under no circumstances be
separated from "plod".

Modality is an elusive characteristic that is difficult to define with
reference to objective criteria. Webster's defines it thus, "(3) in
grammar, mood", but it is not very informative. Modality is often combined
with will and whether something is certain or not, so because it is a
subjective viewpoint, it is very difficult to define it by objective
reference points and say that this or that is the *semantic meaning* of
modality. Therefore I looked at the other two TAM members in order to find
definite points of reference.

But first it was logical to take a look at what is called "the Vendlerian
categories" or "procedural traits" connected with Aktionsart. Of the four
fundamental properties connected with the mentioned categories, Broman
Olsen has shown convincingly that three are semantic and one is pragmatic.
The one that is pragmatic is statitivy, while durativity (ongoingness),
dynamicity (change), and telicity (end included) are semantic. *Any* verb
that is viewed as expressing stativity can also be fientive, but *every*
verb that expresses durativity, dynamicity , or telicity can never loose
this property. This knowledge is a very important tool in the analyzis of

In contrast with mood, tense is semantic. Past tense always occur before
the vantagepoint (C), present tense occur contemporaneous with it, and
future tense occur after the vantagepoint (C). Aspect, as far as English is
concerned, is also semantic. The perfective aspect is a coda view of a
situation and the imperfective aspect is a nucleus view of a situation.
That is, the perfective aspect shows definitely that the event was
completed and the imperfective aspect shows that it was not completed.

While I could not assume that any of the TAM elements were found in Hebrew,
I had a definite framework with a clear definitions of the basic factors
that Hebrew could be measured by. Based on the framwork (or model) several
tests  of the Hebrew verbs were quite simple.

(a) To find out whether tense is the semantic meaning of any of the verb
forms, what I had to do, was to look at all the examples of the form and
find out whether RT (reference time) and C (the deictic point) always had
the same relationship (RT>C - past tense), (RT=C - present tense), and
(C>RT -future tense). Textual criticism had to be taken into account, as
well as a diachronic study, and special uses of the verb forms (such as
hypothetical conditional clauses) had to be taken into account, and also a
reasonable number of exceptions had to be accepted. But the numbers below
seem to say clearly that tense is not a semantic part of the verbal system
of Classical Hebrew.

The 836 examples (6,4%) of QATALs with future meaning, and the 1372
examples (10.5%) of QATAL  with present meaning excludes that QATAL is a
past tense.
The 357 examples (5,8%) of WEQATALs with past meaning shows it is no future
Regarding WEQATAL, YIQTOL, and WAYYIQTOL  I have not completed my research,
but about 5% of the YIQTOLs have past meaning, excluding it from being
present or future tense. Between 5 and 10% of the WAYYIQTOLs have non-past
meaning which would exclude it from being past tense.

b) To find out whether one of the verbal groups coded for modality while
others did not, it was necessary to see if subjunctive could be expressed
by all the groups, and my finds were positive. All agree that YIQTOL,
WEYIQTOL,  and WEQATAL  can express subjunctive, and there are several
examples of QATAL  and WAYYIQTOL that have modal characteristics as well.
Therefore I have drawn the conclusion that no particular verb form codes
for modality.

c) Because two members of TAM was excluded as representing semantic meaning
in Hebrew, only the third remained, namely aspect. Even though the TAM
system is quite universal, neither of the members need to be present in
Hebrew, but when two was excluded, a logical step was to test the system
for the third member, namely aspectuality.
Aspect has something in common with mood, and there are different
definitions of aspect. There is, however, a common denominator between the
differing viewpoints which can be expressed in three points:
(1) Aspect is a subjective viewpoint; i.e. it shows how the reporter
chooses to report an event.
(2) Aspect somehow makes visible or accentuates, either a part (the
imperfective aspect) or the whole (the perfective aspect) of event time
(the non-deictic time it takes to complete an event,from the beginning to
the end).
(3) The beginning and/or the end of an event is important for
distinguishing between the imperfective and the perfective aspect.

If we start with QATAL  and YIQTOL, it seems that all three points are
compatible with their use. The subjectivity of (1) is seen in the fact that
the same event can be described both by a QATAL  and a YIQTOL without any
visible difference in meaning. Aspect is non-deictic and YIQTOL and QATAL
can be used for past, present and future, and the fact that semelfactive
and telic verbs expressed as YIQTOLs make the action iterative, habitual,
or frequentative, while QATALs do not have this effect, is both compatible
with (2) and suggests that YIQTOL has imperfective characteristics and
QATAL has perfective characteristics. The end of an event evidently has
some importance, because most QATALs include it while most YIQTOLs do not.
Based on these observations I see a basic similarity between YIQTOL and
QATAL and the general definition of aspects, so I proceded with the view
that one of the TAM members exists in Hebrew, namely, aspect. (Past,
present, and future meaning can of course be expressed, and subjunctive as
well, but not by morphosyntactic means, except possibly by the ekstra long
and short YIQTOL.)

A very important question would then be: Can we find any semantic
(uncancelable) property in YIQTOL and QATAL? In the English aspectual model
of Broman Olsen we have such properties. The perfective aspect shows
definitely that an event has ended at C (the deictic point) while the
imperfective aspect shows definitely that it has not ended. This inticates
that the two aspects are different ways of viewing (or presenting) the
internal time of an event, as Comrie defines them.
To test YIQTOL  and QATAL for this, we can make the following prediction.
If there is a semantic difference between YIQTOL and QATAL as it is between
the English imperfective and perfective asopect, we expect to find that
YIQTOLs *only* are used with events and states that are not terminated at
C, and that QATAL is used for events and states that are terminated. A test
shows that the prediction holds in many, but far from all cases. (Alviero
is right when he says that the verbs do not describe the events as they
objectively are, but only the way the reporter chooses to depict them.)

Because a nucleus view/coda view is the only *semantic* (uncancelable)
property of aspects, we are forced to draw the conclusion that Hebrew
aspects do not have a *semantic* difference. This is not problematic,
because neither can we point to a property that is so definite and
restricted inside modality that we can say it is semantic. And what about
participles and infinitives, what are their semantic properties? Or, what
about stativity, which is very important in Hebrew? Yet stativity is only
pragmatic. There is no problem in dealing with stative verbs, with
infinitives and participles or modals, even though we cannot point to a
definite property that *allways* is found in one of the groups. The reason
is that we are able to indentify them by their forms, and we have a general
understanding of their meaning.

Then back to the aspects of YIQTOL and QATAL. To get an understanding of
their meaning we must,as in the case with the other mentioned forms, try to
make a generalization on the basis of their role in the verbal system and
how they are used. The following observations should be considered, (1)
They are both finite verbs which translate events and states into words,
(2) Both are subjective viewpoints, (3) Combinantions of YIQTOL with
semelfactive and telic Aktionsart signal iterative, frequentative, and
habituaal events, and conative events are only grammatically expressed by
YIQTOL. This suggests that YIQTOL makes visible details of situations as
imperfective aspects tend to do, and QATAL does not make details visible.
But what about the end which is so important in English and French and
other aspectual systems? It certainly has some importance because QATAL to
a great extent describes events which are terminated, and YIQTOL to some
extent describes situations that are not terminated. But QATAL describes
uncompleted situations as well (836 events with future meaning (a few of
these future perfect) and 1373 events with present meaning), and YIQTOL
describes completed events.

(The real course of events as such remains outside the verbforms and
outside the text. In order not to fall into the same trap where I claim
that those who try to construe semantic meaning on the basis of discourse
function, have fallen, I basically use the *semantic* properties
+dynamicity, +durativity, and +telicity when I test YIQTOL  and QATAL
regarding their relationship to the end of a situation.)

The most narrow generalisation that accounts for all these observations
associated with YIQTOL  and QATAL is one centered around distance and
scope, and this is my definition of aspect:


For the most part the narrow scope is directed toward a part of the
situation *before* the end, but it can also include the end, as it does in
the resultative and factitive situations in the Piel binyan (but in such
cases we can still say the aspect is "open" because the end of the
resulting state is not reached). For the most part the broad scope includes
the end, but there are many instances where the situation continues after
it is presented by the perfective aspect.

How useful is such a definition? When a definition is not semantic, there
is always the danger that it can be too vague and therefore be useless. On
the other hand, we cannot make a definition that is more definite than the
data. So perhaps we need to do some reflection. While I view the model of
Alviero as completely circular as far as the *meaning* of the verb forms is
concerned, I find its description of the *use* of the different forms
extremely valuable. Alviero has done a descriptive work that deserves
admiration. I think that a study of the use of the different forms to a
great degree can compensate for the lack of *semantic* meaning of YIQTOL
and QATAL. Exactly the same is true regarding modality, a lack of semantic
meaning is compensated by a description of use.

Let us use an illustration. The infinitive in Hebrew is said to be a verbal
noun, but such a definition does not tell us much. How is it "defined"
further? By its use (syntactic features, with prepositions, with L, verbal
uses etc). The participle is said to be "a non-finite verb form used as a
noun". How is it defined further? By its use (substantical, adjectival,
predicate etc). The basic definitions of the participle and the infinitive
are quite similar. However, for the most part the infinitive and participle
are used differently, although they can, and are used for the same
events/situations as well. These facts  show that two forms without a
definite semantic factor that differentiates them and with several factors
that make them similar can be used inside a language (without problems) to
make visible particular nuances.

Everybody who has worked with Hebrew knows that while infinitive
constructus and the active and passive participle can be used in many
different ways, there are clear patterns in most of their uses. This is due
to linguistic convention. In a language lacking tenses but having aspects,
we naturally will expect to find patterns in the use of the verb forms. We
should allways keep in mind Alviero's basic principle that there is not a
one-to-one relationship between the event and the description of it (the
verb). Therefore, when an author wants to stress a particular side of a
situation (e.g. that evens are conative, iterative,habitual, gnomic etc) by
grammatical means *only*, there need to be conventions for this, and here
is where the use of aspects can be a distinguishing factor. Gnomic
situations can be expressed by both aspects, but conative, iterative, or
habitual situations which *only* are signaled by the grammar, are in Hebrew
expressed by telic/semelfactive verbs and the imperfective aspect.

Apart from situations where particular properties are signaled, the two
aspects can in principle be used for any situation. YIQTOL and QATAL  are
used with past, present, and future meaning, and with indicative and
subjunctive force. But looking at the MT landscape, there are particular
patters that lead most workers to see four or five different conjugations
and not just two. As I see it, this is a wrong view, which is inherited
from the "fathers" right back to the Middle ages, and it is based on a
misunderstanding of the use and function of the conjunction WAW, and on the
lack of a distinction between semantic and pragmatic factors. What then, do
the patterns represent? The same that is the reason for the pattern of
infinitives and participles, namely, on linguistic convention. Several
grammars view YIQTOL and WEQATAL on the one hand, and QATAL and WAYYIQTOL
as similar, or quite similar in meaning, The reason is a similar use, and
that is true. But this need not be explained as a converting force of WE-
and WAYY- or that these prefixes signal an opposite meaning from the forms
without prefixes. Given my definition of aspects, the reason for the use of
QATAL  and WAYYIQTOL in past contexts and YIQTOL and QATAL in future
contexts, can be that both aspects can be used in these "unmarked"
situations without any problems of meaning. And the reason why past
contexts start with QATAL and continues with WAW+YIQTOL while future
contexts starts with YIQTOL and continues with WAW+QATAL can be explained
as convention, a way to order the parts of speech in a language lacking
tenses but having aspects. The use of YIQTOL with past meaning and QATAL
with future meaning can be explained by discourse function, word order,
genre (e.f. many QATALs with future meaning in prophetic texts) and
foreground, background etc.
(Even when persens have defined four (or five) conjugations solely on the
basis of aspect, tense has been lurking in the background. One good advice
for those working to find the number of conjugations, is to look at the
role *tense* (and I mean "tense", not "time") plays in one's model. Either
it should be discarded completely if one thinks it is absent in Hebrew, or
one should allways keep in mind the influence it has on the conclusions if
it is used. The way in-between, when one uses tense without admitting it or
being conscious of it, is very common - and dangerous.)

I will like to repeat that the description above is just a sketch, and many
questions remain to be answered. The two most pressing questions regarding
the model is:

(1) How can imperfective forms be the basic forms bringing the action
forwards in past narrative? Answer:The argument against this is based upon
the "English" aspect where aspect describes event time in an objective way.
Hebrew aspect is different. The prefix form is used for narrative in
Aramaic and Ugaritic as well.

(2) How can WAYYIQTOL be viewed as identical with YIQTOL when it is so
different in form? Answer: In unpointed texts the difference is minimal and
it is not demonstrated any morphological difference before the Masoretes.
Nobody has ever conclusively demonstrated that the WAYY-element change the
meaning of the following YIQTOL. It need not have any semantic meaning.


Rolf Furuli
University of Oslo

More information about the b-hebrew mailing list