Biblical Hebrew Forum

["Nir cohen - Prof. Mat." <nir AT ccet.ufrn.br>]

dear colleagues, 

i would like to respond to some of the remarks made here, especially by rolf
and karl, concerning tense and time.

------------------

PART I: GENERALITIES on tams and models

linguistics is an empirical field. it is difficult to say 
if it is a science or not. true, it cannot be compared 
to physics, where measurements are far more precise and
much more information is available. but it is more a science than, say,
psychology and economics where every possible statement is contested. indeed,
two individuals spaking the same mother tongue are expected to use similar
sentences; however, each person has a speaking style, and so linguistics can
at best hope at an approximation of "common talk".

but even physics, as an empirical science, has its problems. karl is wrong if
he believes that physics consists of a single model. see the quest for the
so-called "unified theory". 
not even every physical prediction is unanimous. time and distance of the same
object, in relativity theory, are measured differently by different observers.
physics also has many unanswered questions, like black matter, black energy,
etc etc. physics is capable to verify its predictions only up to a certain
precision.

------------------

why TAM is not a single model, asks karl.

the point of view of most linguistic works, including those inclined towards
theory, is the practical one: constructing a simple model, based on a small
number of parameters, for 
imitating the bulk of speech, i.e. "standard use". there is no hope to
describe all the deviations utterable by all the individuals. thus, at best we
have an approximation of reality.

this point of practicality results in terms being used by different
researchers in slightly different ways. of course, 
karl, we could have incorporated them all into a single notational
monster system, which would require constant updating. then it would fit your
criterion for a "good basis for debate", but would be worthless for every
practical purpose.

pluralism is a better way: each person creates his model, and let the best-fit
model win. by evolution, linguistics always get closer to its objectives.
compare today and 300 years ago.

fitting the same model to all languages is 
like fitting the same shoe to all feet. even if you extend your definition of
shoe with as many parameters as you wish, it will never fit all feet. 

The existence of different TAM approaches has two sources: 
variation between the languages, dialects, genres studied,
and variation between the different scholars' perception of them.
this creates a certain pluralism, which is acceptable and even
necessary, given the complexity of the subject. hopefully, one day in the
future we will have a single model.

------------------------

PART II, BH TEMPORALITY

following bryant's sage remark, let us leave generalities and focus on BH.
unfortunately, complex issues do not have simple answers.

of course, in spite of rolf's opinion, almost everybody
agrees that BH uses time as an important determinant of 
the verb form. however, not in the same way english does, for example. or
chinese. these diferences should be made precise.

in order to show that BH does not consider time as verb form parameter, rolf
defines time in a very narrow and non-compromising sense, imposed by some
theories. this, of course, is the wrong direction: in an empirical science, it
is fact which determines theory. BH indeed uses time in more 
ways than he envisages.

even if we define time as an abstract, rigid "past/present/future", it is
clear that there are nuances:
remote past? pluperfect? recent event? vespernal? etc etc

BH indeed uses qatal and yiqtol for remote past and future,
never the other forms (for a panchronic discussion on this, 
see andrason 2011 on yiqtol).

in BH there is distinction between absolute and relative time. tania notarius
has an excellent paper from 2011 where she provides scores of examples for
this. she bases her analysis on a careful study of temporality expounded in
several recent papers of carlota smith, but the basis was laid by reichenbach
1947. 

of course, we can argue if this use is tense or aspect: rolf also uses the
same deictic mechanism of reichenbach but calls it aspect. but he cannot use
one side of the stick ("aspect") to show that the other side ("time") does not
exist. we have to admit, as pointed out by some in this debate, that an
intersection exists.

aspect, too, is divided under most theories into viewpoint aspect and
situation aspect. for some reason, in the BH context the
perfect(ive)/imperfect(ive) division is called by many "aspect", ignoring the
other half, i.e. the distinction between (in first approximation, see wendler,
smith etc) state and event. 

so, what rolf calls "aspect" is really only "half aspect".
where is the other half?

karl, even the distinction between sate and event is not
unanimous in linguistics:  some see it as durative vs instantaneous, others as
stative vs transformative, etc etc. 
see vendler, smith.

this distinction is language-related. in my opinion,
the BH definition is this: "an eventuality which includes an impactive moment,
though possibly extending beyond it". 

for example, when moses climbs the mountain we find twice WY(L ("and he
climbed") in the same sentence, and i imagine there have been many attempts at
settling this. but this is simple: the first marks the beginning of climbing,
the second marks the end of climbing (read the text!). both are, therefore,
events, as they include an impactive moment. and so both WY(L use wayiqtol.
see below.

of course, karl, english too has a clear distinction between state and event.
but the distinction is mainly durative/instantaneous. thus, it would be
difficult to put them under one basket.

it is my understanding that time, plus the state/event 
division, is much more important to BH verb units than 
"aspect" (perfect(ive)/imperfect(ive)). for example,

a) wayiqtols and weqatals describe events and not states.
unless the reason for the waw-prefix is SYNTACTIC (i.e. there was no
alternative). 

b) all wayiqtols and weqatals have the value "relative future",
compared with their event predecessor verb form. with the same caveat.

c) all the sentence-initial temporal anchors, and states, when followed by an
event, also use wayiqtol and not qatal. the 
exceptions are few, and can be explained one for one.

this waw is odd: it is not justified by syntax, as conjunctive
(in english there is no "and" there), yet the waw is justified syntactically,
as disjunctive. again, karl, it would be impossible to use the same model for
BH and english. their use of the waw is 
completely different.

d) a sentence-initial event (weqatal/wayiqtol is expected) uses
yiqtol/qatal. the synchronic reason is because SYNTAX does not allow the waw.
the panchronic explanation is, of course, based on the fact that qatal and
yiqtol (canaanite) existed, morphologically, long before the waw-prefixed
forms, hence are used as default in some situations.

another important distinction is between verb forms which LABEL time and those
which DO NOT LABEL time (i am not sure i use the correct technical term for
this). when we say in english "having nothing to do/having finished my
chores/work completed, I left", the first clause is not labeled, but then
"left" makes everything clear: past tense, providing BOTH clauses with a past
timing.

the use in BH is different, and often includes finite verb forms: an initial
verb form may indicate
the absolute time (past, present, future) and then the second verb form may
only add a relative future reference, accepting the absolute temporality from
the first form. e.g. LK W)MRT: go (indicating immediate future) and then
(perhaps later) say.
the absolute future is only encoded in the first one (TIME!).
the second is TENSED, i.e. a formal operation which follows yiqtol by weqatal.
it hikes on the future indication of LK but also adds relative future: AFTER
you go, you will say.

a sentence-initial wayiqtol, similarly, indicates that this continues a
previous discussion. indeed, a sentence-initial qatal normally indicates a new
topic.

i have already written to karl last week about two additional facets of time:
the repeated event, which uses yiqtol-weqatal
(weqatal for the followup, UNLESS SYNTAX INTERVENES), and the gnomic
(imperfective/atemporal/modal) which abolishes all labeling and so allows
qatal and yiqtol to be used for the same semantic value.

as to STATES, most of them, in both past, present and future, use the
participle. so, here again we have a verb form (we can argue whether 
the participle is finite or not; in the panchronic view, it was moving towards
a finite status) which as a state does not LABEL time. but 
used on events, it does LABEL two specific situations: present 
time, or a future plan (analogous to "i am going to" in english).

all the above only refer to the way i see the relation between
form and immediate verb semantics. of course, one then may ask for an inner
logic which has produced these relations as surface values. most of the
document i posted in researchgate is devoted to this. 

my main conclusion, in this debate, is that temporality is also multi-faceted,
and its correct understanding in BH 
requires an act of flexibilization of our definitions, rather
than sticking to definitions wrigged in an indo-european context. 

again, syntax needs to be addressed as exercising a veto
over tense, in many situations.

----------------------------------

PART III: BH TENSE

are the BH verb forms TENSES? here we need to flexibilize our definitions not
only of time, but also of tense. 

in BH, the preterite is a combination of qatal+wayiqtol, jointly but not
separately. 

the future indicative is yiqtol+weqatal, jointly but not separately. 

the subjunctive, ditto. 

the repeated event, ditto. 

the volitive, ditto. LK W)MRT

in all of them, the waw-prefixed form is used as sequel, UNLESS SYNTAX
INTERVENES, and the wawless forms are used for non-sequel
(initial, parallel, remote past and future, syndetic), with the same caveat. 

nir cohen