Biblical Hebrew Forum

["Dave Washburn" <dwashbur AT nyx.net>]

Peter wrote:
> DW: ...Nope.  That wasn't the criterion.
> 
> PK: Well then, what was it? You did write earlier "Syntactically, it 
> only makes sense to begin with the most basic form of the verb, which 
> is the qatal.  All other forms, including those with W-, are made by 
> adding something to it..." I do not accept that yiqtol is derived from 
> qatal plus an affix, see the following answer.

Then where does it come from?
 
> DW: ...Then what would you say is the foundational form of the verb?
> 
> PK: The only foundational form I would recognise is the triconsonantal 
> root. All surface forms are derived by adding various affixes 
> (including infixes) to that.

At the level of x-bar syntax, I quite agree.  In fact, one of the really 
exciting things about using transformational-generative grammar to 
look at Hebrew is how well it fits, I would almost say exemplifies, 
the x-bar criterion.  However, see below.

> DW: ...Rare in all possible types of ancient Hebrew, or just in the 
> specialized type of literature that we work with?  The nature of the 
> corpus has to be considered as well.
> 
> PK: Our corpus is not all that narrow, it does consist of a wide range 
> of different genres. (Out of interest, what size and variety of corpus 
> would usually be considered sufficient for a linguistic study 
> determining the basic word order of a modern language?) To say 

I wish the TG grammarians would give me a straight answer to that 
very question!  All they usually tell me is that the BH corpus is too 
small.  I don't buy it.  There's a greater amount of extrapolation and 
generalization (read, speculation) involved because of the size of 
the corpus and the lack of native speakers, but I still believe it can 
be done and done effectively.

> anything beyond this corpus is pure speculation. I would accept that 
> any conclusion has to be somewhat tentative because of the limited 
> corpus. 

Agreed.  However, for that very reason a certain amount of 
speculation is necessary and unavoidable.

In any case the size of the corpus is irrelevant to this 
> particular question as I am simply suggesting that clauses with an 
> initial we- prefix should be considered alongside those without, 
> simply because the alternative of looking only at clauses with no 
> initial we- would give us a far more limited set of examples.

Precisely!  That's the whole idea.  From the medieval grammarians 
on there has been a perceived distinction in syntax/usage/whatever 
one calls it between the simple qatal and the qatal with we 
(whether we call it waw-consecutive, waw-conversive, waw-relative 
[with Waltke-O'Connor] or whatever).  If for no other reason than 
that history, we have to begin by dealing with them separately.  If 
they show a fair to large amount of homogeneity in terms of usage, 
context etc. then we can be justified in lumping them together.  But 
to begin with we separate the various conjugations from each other 
precisely to reduce the size of the field and give us a better chance 
of figuring out what is going on.  If they are two allomorphs of the 
same conjugation, fine.  But if they aren't, we can't lump them 
together just to get a bigger field of examples for studying the qatal.
 
> PK: Concerning qatal and weqatal: I am not sure that I have a theory 
> in any technical sense. But my thinking, which is unchanged as far as 
> I know, is something like this: On a morphological or syntactic basis 
> (the distinction I suppose is based on whether we- is taken as a 
> prefix or a separate word, which is I think a rather arbitrary choice) 

Arbitrary how?  Either it is a separate word or it isn't.  The question 
isn't what we think it is, but what they thought it was.

> the form weqatal is made up from two morphemes we- and qatal (and 
> qatal is further made up from the root q-t-l and one or more infix 
> morphemes giving the vowels). 

Well, yes and no.  We have to be careful to distinguish conjugation 
from inflection.  I see two distinct levels of formation going on here: 
the choice of conjugation ("tense") determines which set of 
inflections will be used.  If we just lump it all together in the form of 
affixes, it's likely to give us a skewed picture of things.  As for the 
triliteral root, x-bar principles are what determine the "infixes" 
(which are actually phonological realizations of the syntactic LF 
"verb," "noun" etc. and can only marginally be considered true 
morphemes).  So there's a definite hierarchy of operations that we 
can discern from the way an S-level verb (or other part of speech) is 
built.  To determine the functions of the various verb forms, 
especially the stems and "tenses", it's important to keep them 
distinct.

This is NOT the same as I have been 
> arguing concerning wayyiqtol, which I view as more than we- plus 
> yiqtol. 

So do I.

On the other hand, from a semantic viewpoint I do see the 
> meaning of weqatal as being different from the sum of the two parts 
> we- and qatal. 

The question is, how does this come about?  That's the question 
I'm hoping to answer (eventually).

This is hard to account for; on the other hand, such 
> situations are common in language e.g. in the English present perfect 
> "I have gone" the word "have" does not have its normal meaning 
> "possess" (any attempt to explain as "I possess the attribute of 
> gone-ness" would be rather far-fetched except perhaps as a theory of 
> how this form arose in the distant past) but its meaning has been 
> modified within this construction. 

I think that what's going on with "have" (as well as avoir in French, 
haben in German etc.) is pretty clear as far as their dual status of 
"full verb" and "auxiliary verb."  As the parallel examples in several 
Romance languages show, certain verbs can perform this dual 
function.  However, it's quite a jump from that to saying that a 
simple conjunction such as we- or English "and" can do something 
similar to a verb.  I don't say it's impossible, but I'd have to see 
some heavy-duty backing for the idea.  The other possibility is that 
it's an inflectional morpheme that happens to be homonymous with 
the conjunction; in the case of wayyiqtol this is exactly what I've 
argued, based partly on the different vocalization.  It's harder to 
explain why such a morpheme built onto the qatal would have 
developed in such a way that it looks and sounds (we assume) 
EXACTLY like the simple conjunction so that we have to look 
beyond the word itself to decide which it is.

If we keep syntax and semantics 
> strictly separate, as I think Dave prefers to do, 

Yes I do.

this semantic shift 
> should not affect the arguments for basic word order; but perhaps we 
> do need to take care that there is not some confusing interaction.

Ya lost me here.  Please develop this "confusing interaction" idea 
for me some more?

Dave Washburn
http://www.nyx.net/~dwashbur
A Bible that's falling apart means a life that isn't.