The impossibility of the simultaneous BVAs in such OS examples and the paradigms of resumption examples provided in the WECOL handout constitute the most serious empirical challenge to the analyses of the OS construction (i.e., so-called scrambling constructions) in Japanese that have been proposed elsewhere.

During my WECOL presentation, I in fact stated something like:

"Those who are familiar with the existing literature on 'scrambling' in Japanese are hereby encouraged to consider how the resumption facts could be captured in the other analyses proposed elsewhere."

It is noteworthy that it is possible for the two BVAs to obtain simultaneously in examples like (16) in the WECOL handout if we use a QP of a different type from the one used in (16). (I am talking about the sentence-initial QP.) As discussed briefly in my Lingua paper, and discussed more in depth in the few other papers cited there, with a QP of that type, e.g., dono-NP, we do not get clear effects of local disjointness associated with Principle B of the binding theory. The availability of BVA with such a QP as the 'binder' is sensitive to PF precedence, as discussed in depth in Ueyama 1998 and also in my Lingua paper. So, it is crucial that we use a QP for which BVA(QP, B) is possible only on the basis of LF c-command if we are conducting tests that make crucial reference to (LF) c-command. The failure to do so has resulted in a great deal of judgmental instability, to put it somewhat mildly, in regard to the alleged generalizations discussed in the literature for the past two decades, not only in the area of BVA but also in the area of quantifier scope. The same holds true also in the case of the generalization illustrated by (16) in the WECOL handout and the various paradigms of resumption examples discussed there; and in this sense, the paradigms in the WECOL handout can be understood as constituting strong confirming evidence for the theory of anaphoric relations proposed in Ueyama 1998, in which different types of BVA are teased apart carefully.

But once we concentrate on the distribution of FD-based BVA, we can clearly observe that Ueyama's (1998: chap. 2) prediction is the correct one.

One such example is given in (16) in the WECOL handout, placed in [13626] taken from Ueyama 2002:section 3.2.

The handout used at the WECOL is placed here.

The JK 13 Passive abstract is placed here.

(4) NP2-gai (NP1-{ni/niyotte}) t_i V-rare

(5) There are items in Japanese, like English and other languages, that bear a feature responsible for agreement/checking; and agreement/checking plays a central role in the generation of Japanese sentences.

(6) No items in Japanese bear a feature that is responsible for agreement/checking, and hence agreement/checking plays no role in Japanese, unlike English and other languages.

The choice between (5) and (6) in regard to passives in Japanese should thus be contingent upon (i) whether Japanese passives involve movement as indicated in (4), and (ii), if they do, whether the movement has properties of A-movement or 'scrambling'

For those familiar with the literature on scrambling, this disjunction might appear to be somewhat puzzling. After all, they might point out, "Can (at least, clause-internal) scrambling not be an instance of A-movement, anyway?" My response to that question is actually given in the abstract, although not totally directly, but in a fairly transparent way; it is claimed/pointed out in the abstract that the reconstruction effects in scrambling in JP are not the same as those in what is analyzed as involving A-movement in English. Actual examples illustrating the point, including (7)-(9) in the abstract (along with some 'scrambling examples') will be given at the talk.

Indeed, it is stated in Saito 2003: 1 that the hypothesis that Japanese scrambling is uniform makes the task of explaining certain properties of the 'scrambling construction' "more challenging."

What is referred to by "Japanese scrambling" is what is called the OS type construction in Ueyama 1998, 2003, plus the DO IO construction, and they are often regarded as constituting a set of phenomena, so to speak. Given a set of phenomena that exhibit some similarities on the surface, we do not know, a priori, that they are manifestations of, or due to, a single operation/relation/etc. Regardless of whether one might argue that "Japanese scrambling" is indeed a uniform phenomenon, one would need to provide a substantial empirical and conceptual justification for one's conclusion. Which conclusion/hypothesis one might draw/pursue, therefore, shouold not make one's task more or less challenging than the other, as far as I can tell.

Ueyama postulates two sets of numerations, and hence at least two derivations, for the same phonetic string of the form OSV. Just because of this, one might conclude that her analysis of 'scrambling' in Japanese is, at least conceptually, inferior to a uniform analysis of 'scrambling' as proposed, for example, in Saito 2003.

My understanding is as follows. Saito 2003 allows the licensing of the syntactic relation that underlies the availability of FD-based BVA in the terms of Hoji 2003 at any stage of derivation. It therefore predicts that the 'scrambled NP/DP' (i.e., the Deep DL in the terms of Ueyama 1998) could be simultaneously (i) A of FD-BVA(A,B) and (ii) a container of a dependent term that must be beta of FD(alpha, beta). And this is precisely what is predicted to be impossible under Ueyama 1998: chap. 2.

The prediction under Saito 2003 may appear to be borne out if we do not restrict our discussion to FD-based BVA. But once we concentrate on the distribution of FD-based BVA, we can clearly observe that Ueyama's (1998: chap. 2) prediction is the correct one.

What is/are the empirical difference(s) between Ueyama 1998: chap. 2 and Saito 2003 (the paper in Lingua)?

The Lingua paper refers to Ueyama 1998 in a very crucial way. Under this posting, I will therefore provide some discussion on the content of Ueyama 1998.