Biblical Hebrew Forum

["Matthew Anstey" <manstey AT portal.ca>]

Gday Vince,

>re derivations. i take a more parallel-processing approach, assuming
>that phonology and syntax interact at an interface. on this view, the
>syntactic representation would have to be a surface rep, to the extent
>that means anything nowadays (the days of transformations are pretty
>much over).

>
>re derivations of pausal forms. i assume that the output of the
>phonology is *pausal* phonology. i derive contextual forms by simple
>rules from these. so a phonological derivation outputs, say, [kotovo]
>(3fs past tense) from /kotob + ot/; a contextual rule of syncopation
>gives the further [kotvo].

>
>re mapping from syntax. very complicated if you don't have syntax and
>prosodic theory under your belt. but let me explain in brief my
>current project.
>
>1. observation: pausal forms obtain systematically before particle ki
>when *not* marking object of verbs of cognition (fail before ki with
>"know" "hear" etc).
>2. theory: ki as diagnostic of pausal environments.
>3. study: look at forms before ki to describe pausal phenomenology
>4. goal: theoretically characterize pausal phonology; establish nature
>of pause/accent interaction
>5. status: finishing database now (a whole lot of ki's!); draft paper
>should be finished by month's end.
>6. basic findings. two degrees of pause confirmed: major and minor.
>two types of pausal alternations: stress-shifters, quality-shifters.
>while stress-shifters show pause in both major and minor environments;
>quality-shifters show pausal variant only in major pause.
>7. twist. the pathah/qametz alternation in final syllable of verb
>forms patterns with shifters, somewhat unexpectedly.

Are you hoping to create a complete set of rules that will predict the
pausal form on the surface? From what you've written, these rules seem to be
a combination of lexical, syntactic, phonological, morphological, and
semantic factors. These individual rules are then combined algorythmically
into ordered rules and conditional rules (ie expression rules) to generate
the required surface form. If this is what you are doing then I am extremely
interested! I am working on my master's thesis using Dik's Functional
Grammar, which is miles away from GB, but these sort of rules fit in very
well into his system, provided they avoid transformations, intermediate
stages, etc. So I have three questions: 1) are you confident of getting a
100% accurate algorithm? 2) how amenable would it be to fitting into a
functional analysis?. 3) can I see your article when it's finished?

My prosodic knowledge is on the lower side of zero so I'll have to learn
some of this eventually.

With regards,
Matthew