In terms of the history of research, the concern about FD grew out of the desire to obtain an empirical generalization (in the form of informant judgments) that is as solid as possible and that can be expressed in terms of basic theoretical concepts. The basic theoretical concept turned out to be FD. From the perspective of conceptual articulation of the character of the research in language faculty science in question, as outlined in Hoji 2015, a rather different picture emerges as to the relation between FD and the empirical generalizations in question. It seems that a more appropriate characterization of the relation is that we search for a solid empirical generalization in hopes that it is revealing about properties of FD. Given the inseparability of facts and hypotheses, it is therefore crucial that we have as rich a deductive structure as possible for prediction-making and as rich a network of judgments as possible that are tightly correlated with each other in accordance with our hypotheses. Otherwise, our theoretical account of the solid empirical generalization would remain to be a mere descriptive statement. It is with such a deductive structure of prediction-making and the network of predicted judgments (in the form of predicted schematic asymmetries) that we can expect to make new predictions and aspire to attain a high degree of testability.^FN

FN: It is based on a clear understanding of this point that we place severe constraints as to what kind of description (or, to put it more accurately, what kind of a theoretical account) we want to assign to a given generalization. Some options lead to a set of new predictions that we already know, or have a pretty good idea about, how to test. Other options do not. Testability-seeking research tries to pursue the former type of options while compatibility-seeking research may not.

　　A network of judgments in question is one of confirmed predicted schematic asymmetries in the terms of Hoji 2015. It thus follows (i) that we should pursue hypotheses (and our choices of SGs and LGs in the terms of Hoji 2015) that lead to the predicted schematic asymmetry that the informant's %(Y) on Schema B is 0 and (ii) that we should aspire to obtain experimental results in line with our predictions about a network of judgments.

　　In Hoji 2015, the correlation is between the informant judgments on the predicted schematic asymmetry crucially based on the lexical hypothesis about FD and those that are crucially based on the LF c-command-related structural hypothesis about FD, leaving aside the informant judgments having to do with the effectiveness of the design of the Main-Experiment.

　　As pointed out briefly in Hoji 2015, there is another structural hypothesis about FD having to do with the anti-locality condition.

　　A fuller discussion would have therefore addressed the correlation of informant judgments among the three "dimensions" (the lexical condition and the two structural conditions on FD).

　　As also pointed out briefly in Hoji 2015, not only certain types of BVA but also a certain type of sloppy-identity readings and a certain type of coreference seem to be based on FD. The network of predicted informant judgments is now significantly wider than what is addressed in Hoji 2015.

　　It is by successfully demonstrating that our hypotheses lead to confirmed predicted schematic asymmetries in these seemingly distinct "phenomena" that the proposed methodology for language faculty science as an exact science will begin to be convincing to those who remain skeptical about the viability of the proposed methodology. (I am only considering those who have interest in understanding aspects of the language faculty by the "Guess-Compute-Compare" method. I have nothing to say about what might be convincing to those who do not share that interest.)