|
PART
dass
that
er
he
vor langer Zeit
a long time ago
einmal
once
einen
a
Vortrag
lecture
vom
by-the
weak
/
/
?
von
by
dem
the
strong
Autor
author
des
the
GEN
Romans
novel
besucht
attended
hatte.
had.
‘Hans discovered a novel about the Hudson in the library. In the process, he
remembered that he had attended a lecture by the author of the novel a long
time ago.’
If we assume that both cases involving the overt and covert relatum arguments
have the structure in (303) (i.e., assume that the relational noun takes the DP the
novel or the index as a complement: [Autor [des
GEN
Romans /1]]), the contrast be-
tween the two cases is unexpected.
Taken together, these points provide strong evidence against the structure in (303)
for covert relatum arguments.
19
An alternative structure would be one that is more
19
Further evidence may come from a closer examination of the role that implicit relatum arguments
can play more generally. Asudeh (2005) presents a puzzle of why relatum arguments can be bound,
269
parallel to that proposed above for regular anaphoric strong-article definites, namely
one where the anaphoric index for the relatum argument is introduced higher up in
the structure, at the level of the DP.
(305)
[
DP
1 [
D
0
D NP]]
Note that this structure arguably is parallel to that of prenominal possessives.
(306)
a. sein
his
Bruder
brother
b. [
DP
sein [
D
0
D
Bruder]]
Interestingly, the prenominal form is obligatory if the possessor is a pronoun, i.e.,
in such cases, the postnominal alternative is ruled out.
(307)
a.
* der Bruder sein
b. der
the
Bruder
brother
des
the
GEN
Mannes
man
Furthermore, the prenominal position is restricted to pronouns and proper names.
(308)
a. Peters
Peter’s
Bruder
brother
b.
* des
the
GEN
Mannes
bother
Bruder
as in (i), but cannot serve as resumptive pronouns in sentences equivalent to (ii) in languages that
allow such structures.
(i)
Every suburbanite knows a neighbor.
(ii)
Every suburbanite who Mary knows that {Xhe / *a neighbor} got arrested vanished.
(Asudeh 2005)
Asudeh takes this as evidence against representing relatum arguments as covert pronominal com-
plements.
On the present account, a possible explanation would be that (i) does not actually involve direct
binding of the relatum argument, but rather is a larger situation use utilizing the type-shifter Π
from chapter 5. Cases like (ii) with a full noun phrase would require a determiner that introduces an
anaphoric index for the relatum argument, as on the analysis of the strong article below. Note that
such cases should then be possible with the equivalent of the
strong
, if there is one in the relevant
languages. I am not in a position to test this prediction empirically at the moment.
270
Both of these points lend support to the idea that relational anaphora and prenom-
inal possessives involve the same structure, since the anaphoric index can be seen as
a null pronoun. The only difference between the two cases then would be that in one,
the possessive pro-form is overt and the determiner is phonologically null, whereas
the reverse holds in the other.
The issue of how the prenominal and postnominal forms are related and how the
restriction on the prenominal position can best be accounted for is not settled in
the literature (for a recent discussion and review of some of the main arguments,
see Hartmann and Zimmermann 2002). With respect to the analysis of relational
anaphora, the most plausible assumption is that the anaphoric index is base-generated
in the position it appears in in (305), rather than having been moved there (an option
that may be worth considering for the pronominal possessive case). The reason is that
in order to account for the contrast in (275) based on a difference between the strong
and the weak article, the anaphoric index should be interpreted as an argument of
the determiner (parallel to anaphoric uses of strong-article definites, but in contrast
to cases like (304) with an overt relatum argument).
20
Assuming the structure in (305), we need an alternative version of the meaning for
the strong article, however. In order for the anaphoric index to serve as the relatum
argument of the relational noun, the strong article has to combine with the relation
denoted by the head noun. The meaning in (309) will yield the desired result.
(309)
Jthe
strong
hs,hhe,esti,he,eiii
K = λs
r
λR.λz.ιx[R(y)(x)(s
r
) & y = z]
20
Angelika Kratzer (p.c.) suggests that movement of the index could account for the article-
contrast if we assume that it blocks contraction after it has moved. While I find this possibility
attractive, in particular in light of the discussion of the relationship between form and meaning in
section 6.2.2, I do not see how this structure could be reconciled with the meaning of the strong
article that I argue for. In particular, we would seem to face the problem that the anaphoric index
introduced by the strong article would have to stand for the author, for which there is no antecedent.
271
The anaphoric interpretation of a strong-article definite in a bridging use, as in
(275), would then come about in the following way.
21
(275)
Hans
Hans
entdeckte
discovered
in
in
der
the
Bibliothek
library
einen
a
Roman
novel
¨
uber
about
den
the
Hudson.
Hudson.
Dabei
In the process
fiel
remembered
ihm
he
Dat
ein,
Do'stlaringiz bilan baham: |
