[The discussion here is getting more and more substantive. So, I will perhaps move this thread to the "Remarks" board later.]

It may be more appropriate to prepare a postscript to my 1985 dissertation and one to the 1990 manuscript; otherwise, the postscript to the 1995 NELS paper may become quite long.

When I wrote my 1985 dissertation, I knew that I could not clearly detect the locality effects of Binding Principle A or the anti-locality effects of Binding Principle B in Japanese, at least when we deal with coreference.
(Years later, I conducted multiple-informant Experiments to test whether such effects are detected by other native speakers of Japanese, and the results of such Experiments seemed consistent with my own.)

By the time I wrote the 1990 ms., I knew that I could not clearly detect the effects of Binding Condition C in Japanese, corresponding to English examples such as John praised John's father. I want to leave aside the effects of Binding Condition D in Japanese, corresponding to English examples such as he praised John's father, because of extra complications. (Chapters 2 and 3 of Hoji 1990 (mainly) discuss Condition D effects (using about 200 pages...)
(Years later, I conducted multiple-informant Experiments to test whether such effects are detected by other native speakers of Japanese, and the results of such Experiments seemed consistent with my current judgments. Chapter 6 of Language Faculty Science reports results of multiple-informant Experiments dealing with the English examples.)

By the time I wrote the 1990 ms., I thought that I was able to clearly observe WCO effects and the anti-locality effects in Japanese if we used the "right" A and B for BVA(A, B).

The 1995 NELS paper was written with the "understanding" as indicated above.
(But, I came to understand, in the late 1990s, that (i) the choices for A and B of BVA(A, B) must be much more carefully made than indicated in the 1990 ms. and the 1995 NELS paper and (ii) what is presented in the 1995 NELS paper as the clearest case of the anti-locality effects in Japanese turns out to be much less clear than I thought. My 2003 Lingua paper discusses (i) and (ii) and related issues.)