ghatav at aall.ufl.edu
Mon Dec 27 07:24:40 EST 1999
>I chuckled when you praised Dave's intuition. Knowing that he has
>clearly expressed that he is basing his current views on the qatal on
>your work, of course you would agree with him! (I'm not disagreeing,
I am not sure he does. He accepts my analysis on the modals, but he
does not seem to realy agree with my view on <wayyiqtol> and <qatal>.
>If you had to leave it short because you did not have the time to more
>thoroughly discuss the matter, I understand.
No, it is not a problem of time. I thought it would be more focused
if I kept it short.
But if you do have the
>time, could you give a bit more insight in R-building
The notion of R(eference time) was introduced by Reichenbach (1947)
in his wonderful book ELEMENTS OF SYMBOLIC LOGIC. The theory prevailing at
the time was that of Tense-Logic (TL), which analyzed the tenses in terms
of two points: the event-time (henceforth: E-time) and the speech-time
(henceforth: S-time). According to this approach, the sentence in (1) below
is analyzed as follows: the E-time precedes the S-time; while the sentence
in (2) is analyzed as saying: E-time follows S-time.
1. Mary ate a banana.
2. Mary will eat a banana.
3. Mary has eatn a banana.
Reichenbach shows that, at least for English, TL analysis is not accurate.
E.g., the sentence in (3) above is also understood to state that the E-time
preceds the S-time. So what is the difference between (1) and (3)?
Reichenbach claims that speakers take into account a third point of time,
he labels Reference-Time (henceforth: R-time, or just R). So, every
sentence (in English) is to be analyzed according to THREE times (in
Reichenbach - points of time; Partee and other linguists believe the
relevant unit is intervals): S, E and R. According to him, the difference
between (1) and (3) is in the relationship of the three times. In (1) E and
R associate, both precede S, while in (3) S and R associate, both follow E.
The sentence in (4) clearly shows that there are three times:
4. (when John arrived) Mary had already left.
As Reichenbach determines for the past-perfect sentence in (4), the E-time
of Mary's leaving precedes the S-time, but also the R-time determined by
the adverbial I put in brackets.
Hans Kamp developed DRT (Discourse Representation Theory), which he
shows to be relevant for analyzing temporality in language. In an article
he wrote with C. Rhorer he shows how the notion of R-time helps to explain
the verb forms in French. Kamp and Rhorer show that sentences in
passe-simple introduce R-times into the discourse, while sentences in
imparfait do not. The latter make use of the R already introduced in the
discourse either by a previous passe-simple sentence or by an adverb. So
we may call passe-simple in French an R-building form, as opposed to
imparfait. Note, that the forms in English (I am deliberately avoiding
the term "tense" as we can be dealing in some cases rather with aspect) may
also be classified this way. The simple-past in (1) builds an R-time,
while the past-perfect in (4) does not. I conclude that the past-perfect is
NOT an R-building form (as well as the present and the future perect, but
for the currant discussion this is not important). That is why without the
bracketed sentence, (4) would be ungrammatical. A sentence like "Mary had
eaten" cannot be interpreted without providing an R-time. This can be done
within the sentence itself as in "By five PM Mary had eaten" or in "When
John got home for dinner Mary had already eaten". This can also be
recovered from the context. E.g., via a dialogue where the context or the
interlocuter provide the relevant R-time. The same is to be said about the
progressive in English. The sentence "John was eating" is ungrammatical
untill an R-time is provided (at five PM, when the phone suddenly rung).
Note that the simple past in English MAY also be dependent on the context
to be provided with an R-time. In (1) above the simple past sentence builds
its own R-time, but in (5) below the stative sentence uses the R-time of
the previous sentece:
5. John came in. There were three men in the room.
Now BH. I believe that <wayyiqtol> is an R-building form; it builds its
own R-time, as opposed to <qatal> and <qotel>. In the case of the latter
two forms, we always need to look for the R-time somewhere in the preceding
or following clauses; otherwise we cannot interpret them.
and what you see
>as the derivation process resulting in what is referred to as
This seems to be the most confusing part of my analysis. The term
"sequentiality" seems to be used by different people to mean something like
"ALWAYS in sequence". I use it the same way Kamp does, i.e., "POTENTIAL
sequence". Like the passe-simple in French, the form of <wayyiqtol> has the
POTENTIALITY to form a sequence, since it builds its own R-time. I tried
before to explain this by an analogy to a line people may form. Let's call
such line "space-sequence". People occupy space, and therefore they have
the potentiality of forming a space-sequence, so let's call people
"space-sequential" entities. Whether they make use of their potentiality to
form a space-sequence or not depends on the context. When boarding an
air-plain, people have to stand in line, forming a space-sequence. But when
sitting in a room listening to somebody's talk they do not form a
space-sequence. Does it mean those people lost their ability to form a
space-sequence when they sat in the room? No, they are still sequential,
but they are not making use of their sequentiality. To go back from the
analogy to our <wayyiqtol>, verbs in this form are ALWAYS sequential
because they always build their own R-time. That is, sequentiality is a
derivational property: if the form builds its R-time then it may form a
sequence, if it does not build its own R-time it follows that it is not
sequential and therefore cannot participate in building a sequence. Thus,
while <wayyiqtol> is sequential, <qatal> and <qotel> are not.
However, just like in the case of space-line, the potentiality of
<wayyiqtol> to actually form a sequence is not always realized. Applying
Kamp's analysis for French again, <wayyiqtol> clauses would have to form a
sequence if the text is a narrative.
Also, what difficult (i.e., supposedly non-sequential
>wayyiqtols) instances would Reference-Time-Building solve, in your
First, as I said above, <wayyiqtol> is always sequential; so your
comment in brackets should be paraphrased as follows: "i.e. supposedly
wayyiqtols appearing in a non-sequence text" or something like that. Now,
if a <wayyiqtol> clause appears by itself, or is not preceded/followed by
other wayyiqtol clauses, we have no problem. It still builds its R-time,
although it is not part of a sequence. Such examples I found in
genealogical texts and in direct speech. (Unfortunately there are no native
speakers of BH, so we cannot examine the behaviour of wayyiqtol when it
appears completely alone.)
I hope I did not wear you out.
>You are currently subscribed to b-hebrew as: ghatav at aall.ufl.edu
>To unsubscribe, forward this message to
>To subscribe, send an email to join-b-hebrew at franklin.oit.unc.edu.
More information about the b-hebrew