[b-hebrew] Paper uploaded: Tagging Hebrew Tense, Aspect, Mood

furuli at online.no furuli at online.no
Tue Dec 16 03:52:44 EST 2003

Dear Ken,

>Dear Rolf,
>Thank you for your feedback, and especially for phrasing your concerns in a
>friendly way.
>>  Note 4. "used differently".  What is the meaning of "differently"?
>>  Because you mention the basic TAM factors in your introduction, the
>>  reader may think that "differently" refers to one or more of these.
>>  If you for instance are saying that the *meaning* of a verb, say a
>>  WAYYIQTOL is different in different genre situations, I disagree with
>>  your assumption. A verb form has the same meaning in any context, but
>>  it can be used differently.
>I am puzzled by your last line (above) stating that that a verb form can be
>"used differently," yet it was my own use of this phrase that you object to.
>Does this mean that we agree?

I understand that you are puzzled, because my last phrase should not 
have been phrased in this way. My intention was to make a suggestion 
for a clarification of your note, and now I make the same suggestion 
to myself. What I tried to say was that a YIQTOL or another verb form 
signalled a concept in the minds of native Hebrews. If the concept 
signalled by YIQTOL is identical with the imperfective aspect, a 
YIQTOL will be imperfective in any context, and can only function 
inside the constraints of the concept of imperfectivity that it 
signals. For instance, a YIQTOL will in no context be perfective. The 
consequence of this is that a YIQTOL in any context can be used to 
find the meaning of all YIQTOLs.

>I am not ready to agree that "a verb form has the same meaning in any
>context." My research may ANSWER the question of whether "a verb form has
>the same meaning in any context," but I will not assume an answer at the
>outset. The question reminds me of the recent B-Hebrew discussion of "core
>meanings" for vocabulary words. Would you also say the word "strike" has the
>same meaning in every context?

We cannot start from scratch as did Descartes and say "Cogito ergo 
sum," but we all need to use particular assumptions. However, the 
point is that we always keep our assumptions in mind in order to see 
whether any of them can disturb the interpretation of our data.  One 
basic assumption in my study is that a difference in morphology 
signals a difference in meaning. Exceptions must be demonstrated by 
the context (e.g. the conflation of Qatal and Hiphil in Qumranic 
Hebrew). I cannot see how a profitable study of a Hebrew corpus is 
possible if this assumption is not made. Would you for instance say 
that the meaning of an English preterit is different in different 
genres or contexts? My assumption is that a preterit has the same 
meaning (reference time comes before the deictic center) in any 
context, even in hypothetical conditional sentences, but the *use* of 
the preterit can be different.

>I am not _ASSUMING_ the meaning of a wayyiqtol is different in different
>genre situations, but I want to ANSWER the question of whether a wayyiqtol
>conveys something different in different situations.
>>  p 9.  WAW.  I am not sure what the terms "T" and  "F" refer to. If
>>  the application of these letters is based on the Masoretic pointing,
>  > I see no problem here. Because then the database only give data as to
>>  Masoretic pointing. But if the letters are based on the judgement of
>>  the compiler, then their use is circular. The purpose of the database
>  > is to give a basis for the interpretation of Hebrew verbs, so the
>  > compiler should  not assume *before* everything is compiled that WAWs
>>  have a converting (or some other) power and that there are four
>>  rather than two conjugations in Classical Hebrew. Conclusions
>>  regarding this should be drawn when everything is compiled.
>You are right. I intend the tag to be entirely objective, with no judgement
>on the part of the compiler. I originally designed the tags for use on
>unpointed texts, so I could not distinguish between waw with shewa and waw
>with patach and doubling. To make the tags more intuitive, I will change my
>proposed tags: "W" for a prefixed waw with shewa, "Wa" for a prefixed waw
>with patach/doubling, and "-" for no waw prefix. For unpointed texts, if
>there is a waw prefix, the tag will be "W", and "-" for no waw prefix.
>I do not assume there ARE four conjugations, but I want to be able to TEST
>whether there are four conjugations.

Very fine, I have exactly the same approach.

>>  pp 17, 23. ASPECT. "temporal constituency (complete /progressive
>>  recurring/ event)". This is the most problematic part.
>You make some perceptive comments about tagging aspect, especially:
>>  How can you by observation know
>>  that a verb is what you call "durative" and that others are not? I am
>>  afraid that both terms at the outset entail circularity. If the terms
>>  are not circular (it has already been decided that the perfective
>>  aspect is complete and the imperfective is durative), there must be a
>>  clear observational distinction between "complete" and "completed"
>>  and between "durative" and "non-durative" events.
>You're right that observation is the key. What aspect-related features CAN
>we observe? To tag tense, we can observe the time at which the event
>referenced occurs (relative to the speech time, for example).

I see some problems here. The first problem may only relate to a 
different use of terms. You define tense as  "grammaticalized 
location in time, " in which I agree. But this means that tense 
cannot be tagged, because you cannot at the outset know whether any 
verb form is grammaticalized. What we can tag is past reference, 
present reference, and future reference". If we say we tag tense, 
then we must have decided beforehand that tense is a part of the 
Hebrew verbal system. The conclusion on the basis of my data base is 
that tense is non-existent in Classical Hebrew.

The same problem exists with the use of reference time as a basis for 
tagging. We cannot use reference time if we have not beforehand 
decided that a particular verb codes for a particular tense or 
aspect. But then we use circular reasoning if we first decide the 
nature of a verb form and then try to confirm this by comparing the 
examples of this form. An example: In a section of my dissertation I 
analyze all the examples of )MR ("say"). Most modern studies (e.g. G. 
Hatav) would say that the reference time of a WAYYIQTOL form  of this 
verb either intersects its event time at the end or intersects its 
whole event time (the perfective aspect is viewed as a blob); thus we 
get the translation "and he said". R.S. Driver, however, claimed that 
the reference time intersected the event time at the beginning; thus 
he suggested the translation "and he proceeded to say". My research 
in this area corroborates Driver's suggestion. In order to use 
reference time as a tool for tagging, we must beforehand have decided 
whether the WAYYIQTOL is perfective or imperfective, and that is 
circular reasoning. In my database, therefore, I use event time for 
mapping rather than reference time. This gives results which are not 
as accurate as if reference time was used, but they are accurate 
enough, and I avoid circularity.

>To tag gender,
>we would observe the sex of the referent (if biological). To tag grammatical
>number, we would observe how many of the referent there were. So by analogy,
>what can we tag for aspect? It seems to me the most helpful question to ask
>is whether there was some focus on the progression of the event or not,
>i.e., whether the reference time is between the beginning and end of the
>event, or at or after the end of the event. Do you have a better suggestion?

As shown above you cannot use reference time without being circular, 
and how in the world can you know that the focus of a clause is on 
its progression? Let us use a few English examples:

1) She was reaching the peak.
2) She has reached the peak.
3) She reach the peak. #

4) She was knocking at the door.
5) She has knocked at the door.
6) She knock at the door.#

Examples 1), 2), and 3) represent achievements. The reason why we 
give 1) a progressive interpretation ("she was on the point of 
reaching the peak.") is that we know that the English participle is 
imperfective. And similarly, the reason why we do not interpret 2) as 
progressive, is that we know that perfect is perfective. But what 
about 3) where the verb is not marked? How can we know whether the 
focus is on progression or not? The same is true with the 
semelfactive situations of 4), 5) and 6). As to 6), we cannot know 
whether progression is focused upon or not.

If we start our tagging of Hebrew verbs without any decision as to 
which verbs are imperfective and which are perfective, each verb we 
analyze is similar to those of 3) and 6). Therefore we cannot know 
anything about whether the focus of a clause is on progression or not 
and we cannot know where reference time intersects event time, except 
in a few instances - see below).  The conclusion, therefore, is that 
*aspectual* properties for the most part cannot be tagged.

I say "for the most part," because there are clauses where we can 
know where reference time intersects event time. One such example is 
1 Kings 6:1.  In the WAYYIQTOL form of BNH  in this verse, reference 
time intersects event time at the beginning (what is made visible is 
the beginning and a small section of progressive action). Thus this 
WAYYIQTOL is imperfective ( as I claim that all WAYYIQTOLs are). Two 
sets of information helps us pinpoint reference time in this case: 
the adverbial and the fact that many years was needed to complete the 
temple.  Such clear-cut situations where we definitely can see that 
WAYYIQTOL is imperfective exists in between 5 and 10 percent of the 
14. 000 clauses where WAYYIQTOL occur, and these situations are what 
we need to map.

Because the nature of narrative where most WAYYIQTOLs occur is that 
the event time of one verb is mentioned and then the event time of 
the next follows and then the next, and so forth, it gives very 
little new information to tag or map all these WAYYIQTOLs. These 
WAYYIQTOLs simply *must* have past reference, given the nature of 
narrative, but whether the stress is on the progressive or not we 
cannot, in most instances. know. So the least likely place to find 
the true meaning of the WAYYIQTOL form is in narrative.

When I claim that aspect cannot be tagged, I do not say that the 
situation is hopeless. I have developed three parameters for a 
systematic analysis of aspect, and because there are two aspects, 
aspectual properties can be analyzed  and compared in six different 
areas. About 50 per cent of my dissertation outline  the analysis of 
the hundreds of instances in the Tanakh  where we can be quite 
certain where reference time intersects event time. The conclusion 
drawn is that YIQTOL, WEYIQTOL and WAYYIQTOL all represent the 
imperfective aspect and that QATAL and WEQATAL represent the 
perfective aspect. This holds for the late books of the Tanakh and 
for Qumranic Hebrew as well.

>Ken Penner, McMaster/DSS

Best regards

Rolf Furuli

More information about the b-hebrew mailing list