LONG Re: QATAL C/RT [Statistics]
furuli at online.no
Tue Aug 1 12:08:03 EDT 2000
Henry Churchyard wrote:
>[Behind in reading B-Hebrew, as usual...]
>> Subject: Re: QATAL C/RT
>> From: Rolf Furuli <furuli at online.no>
>> Date: Tue, 25 Jul 2000 19:31:12 +0200
>> WEQATAL 6087 % QATAL 13922 %
>> PAST 357 5,8 7450 53,5
>> PRESENT 192 3,15 2505 18
>> FUTURE 4100 67.35 965 6,9
>> PERFECT 55 0,9 2605 18,7
>> MODAL 147 2,41
>> IMPERATIVE 643 10,56
>> FINAL 31 0,5
>> COND,PROT 312 5,12
>> COND,APOD 123 2,02
>> GNOMIC 48 0,78
>> OTHER 79 1,29
> [Adds up to 13525, 397 forms missing?]
>> These statistics are problematic for the traditional of view of
>> QATAL as a preterit, the perfective aspect, or a combinations of
>> both. The fact that almost half of the QATALs have non-past
>> reference, makes it very difficult the claim that QATAL is a
>> grammaticalized past tense. I believe the differences can be
>> explained on the basis of syntax and linguistic convention, that is,
>> QATAL and WEQATAL are one and the same conjugation which is used in
>> different syntactic environments according to the general agreement
>> between the persons having the same presupposition pool.
>Hmmm, I actually read these statistics differently. If you say that
>"PAST" and "PERFECT" meanings are the proper province of QATAL, and
>the other meanings are the proper province of WEQATAL, then at most
>((13922-(7450+2605))/13922) or 28% of QATAL forms have "unexpected"
>meanings, while only ((357+55)/6087) or 7% of WEQATAL forms have
>"unexpected" meanings. In addition, I'm slightly suspicious of an
>aspectually-underdifferentiated "PRESENT" category. Also, it has to
>be taken into account that the category "WEQATAL" to some degree is
>ambiguous, containing forms which belong to a special WEQATAL
>conjugation semantically distinct from plain QATAL, as well as
>containing forms which are simply conjunction + ordinary QATAL. So
>all in all, considering these factors, together with the inevitable
>residual "noise" which is to be expected when conducting research of
>this kind, I'm actually surprised by how strongly your statistics
>distinguish WEQATAL from QATAL.
>If you ran a statistical "level of significance" test on the following
>table (view with non-proportional font), it would probably come out
>fairly strong (the "odds ratio" is about (5675*10065)/(412*3700), or
>roughly 35 to 1):
> WEQATAL QATAL
> | | |
> PAST & | | |
> PERFECT | 412 | 10065 |
> | | |
> | | |
> | | |
> Other | | 3470 ? |
> Meanings | 5675 | or |
> | | 3867 ? |
> | | |
Statistics can be of course read in different ways, particularly if the
premises are different. Good science is made up of arguments and
experiements that can be tested by others. This means that our premises
must be easy to find, and that we need definite points of reference, by
which our conclusions can be tested.
I take the English system of tense and aspect as a point of departure,
because it can be clearly defined. I am not aware of a better explanation
than the one given by Mari Broman Olsen, and I therefore use her model. As
a background to my arguments, let me give a sketch of the English system
based on Broman Olsen:
Simple past is marked for past tense as seen by (1), because only (1a) is a
grammatical clause. In English we therefore have a past tense consisting of
forms with an uniform interpretation (+past).
(1a) Matt ran yesterday.
(1b) *Matt ran as I am speaking.
(1c) *Matt ran tomorrow.
Future is marked for future tense as seen in (2), because only (2a) is a
grammatical clause. In English we therefore have a future tense consisting
of forms with an uniform interpretation (+future).
(2a) Matt will work tomorrow.
(2b) *Matt will work as I am speaking.
(2c) *Matt will work yesterday.
English simple present is not marked for tense because (3a,b,c,d) are
(3a) So Paul works all yesterday to finish. (Past)
(3b) Paul works from now till 4. (Present)
(3c) Paul works tomorrow. (Future)
(3d) Paul works. (Omnitemporal)
Usually it is quite easy for a native speaker to know the tense. In a dead
language we need a scheme, and we have a very simple one based on reference
time (RT), event time (ET) and the deictic point (C):
Past: RT comes before C
Present: RT coincides with C
Future: RT comes after C
(Note that RT coincides with C in English perfect, so it must be taken
together with present rather with past.)
The imperfective aspect is expressed by the present participle, and the
perfective aspect by perfect (not simple past!).
The Imperfective aspect indicates that RT intersects ET at the nucleus
(before the end) as seen in (4). The perfective aspect indicates that RT
intersects ET at the coda (the end) as seen in (5).
(4) While John was reading the paper, Peter arrived.
(5) Peter has arrived at your place.
The combination of tense and aspect gives quite unambigous interpretations.
The tense shows where the event is located in relation to a deictic point,
while the aspect shows whether the event was finished or not at the
Let us look at QATAL in the light of tense. We can make the following figure:
RT comes before C in 7450 cases (53,5 %)
RT coincides with C in 5110 (2505+2605) cases (36,7 %)
RT comes after C in 965 cases (6,9 %)
The picture we see is the same as that of English present, which is
unmarked for tense. So QATAL is unmarked for tense.
Let us look at QATAL in the light of aspect. We can make the following figure:
RT intersects ET at the nucleus: 2505 (18 %) = "imperfective"
RT intersects ET at the coda: 2605 cases (18,7 %) = "perfective"
(I can add that less than 10 % of those with future reference are "future
perfets" and must be added to the possible perfective group. The group of
7450 cases with past reference is more difficult to combine with an aspect,
although past reference in most languages with aspects are combined with
the perfective aspect. But this is open as far as Hebrew is concerned.
The crucial factor which distinguishes the aspects in English, is the the
coda (the end). It is impossible that the perfective and imperfective
aspect both can include and exclude the end. When QATAL does that in a
significant number of cases, it is impossible that it can be an aspect in
the English sense of the word. Given that my interpretation of how many
verbs have perfect and how many have present meaning is close to correct,
there is a very strong case against interpreting QATAL either as a tense or
an aspect (in the English sense).
We can make the following figure:
RT comes before C in 357 cases (5,8 %)
RT coincides with C in 295 (192+55+48) cases (4,8 %)
RT comes after C in 4100 cases (67,3 %)
The picture we see is again that of English present, which is unmarked for
tense. The definition of tense is clearcut, and we cannot just say that the
5,8 % with past reference and the 4,8 % with present reference are "noise".
If English was a dead language and I should make tests to see if compounds
with "will"+ verb were marked for future tense, I would reject the
possibility if I found 357 cases similar to "*Matt will work yesterday."
and 295 cases similar to "*Matt will work as I am speaking."
Before we reject the hypothesis that "will"+ verb is marked for future
tense (in our hypothetical example), we will do some research, and ask if
there are special reasons for the anomalous examples. In Norwegian we can
say "This was a cute baby." when we mean "This is a cute baby." It can be
explained as a polite way of expressing oneself. Can we find something
similar in Hebrew? So we have to look at the environments where these
problematic forms occur. And we find just normal environments!
So we can try another path and say that "will" in the clauses *Matt will
work yesterday." and *Matt will work as I am speaking." is not the same
word as "will" in the clause "Matt will work tomorrow." But then we must
ask for reasons for this claim, and show that it is not just an ad hoc
claim to uphold a hypothesis. If we say there are two different QATALs, how
do we distinguish them? You yourself have pointed out that a QATAL with
different stress can not be established by a diachronic study as you claim
is the case with YIQTOL (*YAQTUL). Should we say that all forms that
graphically are written as WEQATAL which do not fit our definition of the
semantic group WEQATAL, are normal QATALs? This would seem to be an ad hoc
explanation, and I would ask: What is the characteristics of the WEQATAL
group? A particular stress pattern is only seen with 1st. and 2nd person
singular and is not consistently applied even with these persons. Is future
reference the criterion or future+imperative+modal? Is it subordination
versus coordination (Waltke/O'Connor)? The high percentage of imperatives
is in itself a problem, because imperative and future normally do not
collocate. So again, how are we to distinguish graphic WEQATALs from
RT intersects ET at the nucleus: 295 (4,8 %) = "imperfective"
RT intersects ET at the coda: 55 cases (.9 %) + (less than 10 % future
perfect) = "perfective"
The percentage is not as high as in the case with the QATALs, but the
English aspects do not allow *any* exceptions regarding the relationship to
the nucleus/coda, except those which can be viewed as particular idioms.
WEQATAL can therefore not signal aspect in the English sense of the word.
COMBINING WEQATAL AND QATAL
If we view both as one group, we can make the following scheme:
RT comes before C in 357+7450 =7808 cases (39 %)
RT coincides with C in 295+2505+2605 = 5405 cases (27 %)
RT comes after C in 4100 + 965 = 5065 cases (25,3 %)
RT intersects ET at the nucleus: 295+2505 = 2800 cases (14 %) = "imperfective"
RT intersects ET at the coda: 55+2605 (+ fut. perf?) =2660 cases (13,2 %) +
If our hypothesis is that WEQATAL and QATAL constitute one semantic group,
we have to account for the high percentage of simple future among the
WEQATALs compared with the QATALs, 4100, 67.35 % and 965 6,9 %
respectively (and a similar relationship between
WAYYIQTOL/YIQTOL/WEYIQTOL). If Hebrew neither has grammaticalized tenses
nor aspects with a uniform relationship between RT and E, but only
subjective aspects with the other aspectual characteristics, a plausible
model can be made. If this is the case, the "naked" verb form does not tell
us anything about the relationship between C and RT or the relationship
between RT and ET. But listeners and readers are interested in just these
relationships. When the verb *form* does not signal such relationships,
patterns will be made to the effect that readers and listeners with the
same presupposition pool can gather these relationships by help of the
One such pattern that we see is that HINNE + participle allways (according
to some grammars) have future reference. WEQATAL and WEYIQTOL have
something in common, particularly future reference and modality. (See my
table of WEYIQTOL meaning below). The past and present reference is about
the same in WEQATAL and WEYIQTOL, the difference being future reference and
modality versus imperative. 10,5 % of the WEQATALs are used as imperatives
but this is lacking in WEYIQTOL. On the other hand do we find 64,3 % of
other modals among WEYIQTOLs but only 10 % among WEQATALs. We see a pattern
where one form was used for modality and the other for imperative. In
addition to these small patterns, we have the two great ones, WAYYIQTOL for
past narrative and WEQATAL for future reference. In both cases the verbs
are preceded by the conjunction waw, and the conjunctive idea is present in
all cases of WAYYIQTOL and WEQATAL. I am working on this in more detail,
but it seems logical to me that there were two such big patterns which
would help the hearers/readers in addition to the context (e.g a YIQTOL
with prefixed waw most likely had past reference if not forbidden by the
context and a QATAL with prefixed waw most likely was future (or
imperative) if not forbidden by the context.). We can also illustrate the
case with the participles. I have not completed my study, but Genesis seems
to be representative for the whole Tanach. The spread of the participles
can be compared to the spread of WAYYIQTOL/WEYIQTOL/YIQTOL (taken as one
group) and QATAL/WEQATAL (taken as one group). The uses of the participle
show that one form can be used for both past, present, future, and for
modality. If we test the participle for aspect, at least 30,8 % of the
functions are compatible with the English imperfective aspect, and .6 %
with the perfective aspect. Even the low percentage of perfect reference
would exclude an aspectual understanding of the participle.
Let me add that while the finite verbs do not describe the relationship
between C and RT and RT and ET, they have other properties (aspectual
ones), which, in combination with Aktionsart and different syuntactic
factors, can be used to signal a host of fine nuances. So, while YIQTOL,
WAYYIQTOL, WEYIQTOL, QATAL and WEQATAL in principle can be used for almost
any description in the past, present, and future (except a few cases), they
are not used haphazardly but according to particular patterns.
IND.FUT 244 20,04%
IND.PERF 9 0,73%
IND.PRES 39 3,20%
IND.GNOM 68 5.58%
IND.QUEST 21 1,72%
IND.PAST 53 4,35%
434 434 35,66%
MODAL 426 35,00%
FINAL 195 16,02%
PURPOSE 75 6,61%
MODAL OTH. 87 7,14%
783 783 64,33%
PARTICIPLES IN GENESIS
PRE-PAST 4 = 37,3 % (Pa+Pre-P together)
PERFECT 2 30,8 % (Pr+Perf together)
FUTURE 19 6,1 %
MODAL 12 4 %
NOMEN AGENTIS 62
University of Oslo
More information about the b-hebrew