Advances in Role and Reference Grammar

(singke) #1

328 LAURA Α. MICHAELIS


h. Ferunt eum adesse.
(they) say him() be present(iNF)
"They say that he is present."
Such sentences as (23a,c,e,g) raise the following question: how can one
account for the fact that the genitive arguments of such predicates as pudet
and paentitet, although lacking salient subject properties (nominative cod­
ing and the capacity to trigger verb agreement in oratio recta), can nonethe­
less "raise to object", as can a bona fide subject? The answer to this ques­
tion will require a more inclusive definition of subject than that provided by
the traditionally recognized case and agreement properties. Such a defini­
tion, it will be seen, is provided within this framework. We might then now
turn to the RRG analysis of the Latin data.


  1. A Role and Reference Grammar analysis of deviant case-marking in
    Latin


The case-marking rules to be posited for Latin are similar to those proposed
by Van Valin (1991) for Icelandic and to those of many nominative-accusa­
tive languages. These coding principles assume an algorithm intended not
only for the mapping of lexical representations into their syntactic manifes­
tations but also for the linking of syntax to logical structure (See "Synop­
sis", sect. 4.6,5 ). They are stated in (24):
(24) a. PrPs take nominative case.
b. Non-PrP macrorole core arguments take accusative case,
 The default case for non-macrorole direct core arguments is
dative.
As in Icelandic, actor outranks undergoer for PrP. No argument can occupy
PrP position without being a macrorole. Further, any solitary macrorole-
bearing argument, whether actor [A] or undergoer [U], must be a PrP. (As
will be discussed, the verbs licensing genitive subjects, shown in (19) and
(23), constitute an exception to this last rule — there, the single macrorole-
bearing argument lacks PrP status, as indicated by its accusative case.)
We might now examine the evidence that the grammatical function
coded by the nominative in fact represents a pragmatic pivot. This discus­
sion will necessarily be divided into two parts. First, it must be shown that
the nominative argument represents a pivot, i.e, that it occupies a
Free download pdf