From: Vincent DeCaen (email@example.com)
Date: Sat Oct 21 1995 - 21:49:06 EDT
re INFL projection.
I take a view of grammar that can be found in Ray Jackendoff "Semantic
Structures" 1990 (ch 1, especially 1.4) and Jerrold Sadock
"Autolexical Syntax" 1991. it's not clear how this connects with
Chomsky's so-called Minimalist program; but I think it's pretty much
the same, except they differ on how best to implement the
"computational system": I prefer parallel processing.
essentially, a grammar consists of autonomous and simplified
subgrammars including phonology, morphology, syntax, semantics and
probably discourse-pragmatics. a well-formed utterance will be
well-formed in all dimensions. the dimensions interact by
"correspondence rules" vs transformations; principles of "economy" are
constraints on correspondence between levels. the complexity of language
arises from the number of ways the components interact. the "interface
protocol" is the lexicon. all lexical entries will be specified for properties
in the different subgrammars; but crucially, they may be
"underspecified", or even have nothing to do with a given component.
1. it is standardly assumed that in the syntax, "phrases" can be parsed
into heads plus "objects" and "subjects". e.g., a prepositional phrase
has a head preposition (P) plus usually an object, which is in turn
usually a noun phrase, headed by N (noun), etc. the all important
verb phrase will also have formal positions governed by a V head.
here's where it gets tricky: it is further assumed that the sentence has
two "functional" projections INFL (inflection; formerly AUX
"auxiliary") governs the VP (verb phrase), and COMP (complementizer;
more or less "subordinating conjunction") takes IP.
all "inflected" tenses will have an INFL node, or better TENSE node,
in the syntax, governing the VP.
2. the lexical entry of tense-inflection formatives will have the
syntactic entry I or T head. the entry will have the morpho(phono)logical
properties (e.g., English past /-t/). the entry will also have
semantic specification. the English past will have PAST in its entry.
if Mari is right, maybe the nonpast will not have an entry
so what I was getting at with Greek "present" is that its entry will
look much like English's in the semantic specification. since
perfective is the privative selection, the imperfective reading will
be supplied by pragmatic defaulting for Grk. it will also have the morphology
and phonology given. and crucially, it will be a TENSE head in the syntax.
3. no matter how you slice it, you're going to need a theory of "verb
movement". within the frameworks I use, it is assumed that the verb
"moves" to the tense head for morphological reasons.
In Hebrew I assume that verb moves to tense, and that a constituent is
preposed ("topicalization") giving Hebrew its characteristic verb
second (V2) flavour (cf. German matrix clauses). further, the verb-tense
complex moves to complementizer for the verb initial ordering.
in my analysis of Biblical Hebrew, the short-form prefixed
(traditionally jussive/preterite) had the completely underspecified
form. both prefixed forms were as a class nonpast: "unmarked". the "marked"
suffixed form was specified PAST. in brief there are three TENSE/INFL
heads, but traditionally only two were recognized as "tense"; the third was
treated under mood if at all. since Ewald, there's been this nonsense about
aspect: don't believe a word of it.
4. I'm sorry if this is still blibber blabber. but I can't go much
simpler without giving a LIN 100 course on line. all I can say is that
if this stuff interests you, go to your Univ library and start
reading. maybe you can get a reading list from someone teaching intro
courses. I don't know what's used currently, especially in the
Vincent DeCaen firstname.lastname@example.org
Near Eastern Studies, University of Toronto
Religion and Culture, Wilfrid Laurier University
I really do not know that anything has ever been
more exciting than diagraming sentences.
This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:37:30 EDT