"Hypothesis testing in generative grammar: Evaluation of predicted schematic asymmetries" Journal of Japanese Linguistics 26.

A slightly altered version of the paper is available here.

***Abstract***
This paper explores how the hypothetico-deductive method can be applied to research concerned with the properties of the language faculty. The paper first discusses how we can try to identify informant judgments that are likely a reflection of properties of the Computational System (or properties of the language faculty that are directly related to the Computational System), proposes a method of hypothesis testing in line with the hypothetico-deductive method, and provides an illustration by examining the predictions made under the lexical hypothesis that otagai in Japanese is a local anaphor.
***

Due to the pace limit, the empirical illustration of the proposed methodology is severely limited and the conceptual justification of certain crucial aspects of the proposed methodology is also not as satisfactory as it could have been if I had more space. The abstract of a longer version of the paper, which is at the moment entitled as "Point-Value Predictions in Language Faculty Science: Suggestions and Illustration" is:

***Abstract of "Point-Value Predictions in Language Faculty Science: Suggestions and Illustration" (as of 10/27/2010)***
This paper is concerned with how hypotheses about the language faculty can be made testable. Adopting the general model of the Computational System in Chomsky 1993, a model of judgment-making suggested in Ueyama 2010, and a research heuristic "Maximize our chances of learning from errors," the paper suggests how we can try to identify what is likely a reflection of properties of the Computational System hypothesized at the center of the language faculty. Testability pursued here is in terms of "point-value predictions" rather than predictions about a difference or a tendency. The paper argues, as a consequence of the three starting assumptions, that it is indeed possible obtain categorical judgments in our experiments, but only if we recognize a fundamental asymmetry between a *Schema-based prediction and an okSchema-based prediction. The paper provides some illustration by applying the suggested method to specific hypotheses, and by making reference to results of on-line experiments.
***

The paper starts with:

***
Linguistics is often said to be a scientific study of language, and generative grammar a scientific study of the language faculty. While some people might consider the mention of "scientific" rhetorical, others maintain that it is not. To the extent that we indeed wish to be engaged in a scientific research program that is concerned with the properties of the language faculty, it is worth considering how our hypotheses about the language faculty can be made testable and how we can aspire to make our research program "scientific."
　　　In the generative tradition, the main goal of our research is understood to a discovery of the properties of the Computational System, hypothesized to be at the center of the language faculty. It is furthermore assumed that a major source of evidence for or against our hypotheses is informant judgments. Despite a wide acceptance of this assumption, however, the field has so far failed to seriously consider in what way informant judgments can be revealing about the properties of the Computational System, let alone come up with an answer that the majority of the field can agree upon. Chomsky 1986: 36, for example, remarks, "In general, informant judgments do not reflect the structure of the language directly; judgments of acceptability, for example, may fail to provide direct evidence as to grammatical status because of the intrusion of numerous other factors," and it has remained unclear exactly what should qualify as evidence for or against our hypotheses about the language faculty, how we can identify such evidence, and hence how we can put our hypotheses to rigorous empirical test._FN1
　　　The main thesis of this paper is that we can pursue language faculty science as an "exact science," in which we can extract reproducible phenomena and analyze them quantitatively. This goes directly against the common perception that informants' (including researchers' own) introspective judgments are subject to across-informant and within-informant variation and fluctuation. The paper is thus concerned with how we can extract reproducible phenomena and analyze them quantitatively in language faculty science. One can thus view the main aim of the paper as providing a basic tool in language faculty science in response to Anderson's (20008: 812) "whimper" that "we need better tools of investigation" in language faculty science.

FN1: Since the inception of generative grammar in the 1950s, there has been concern about in what way our hypotheses with regard to the Computational System of the language faculty can be put to (rigorous) empirical test. Chomsky's remarks on the relation between acceptability and grammaticality, which are frequently found in his writings until about the mid 1970s, can be understood to be a manifestation of his own concern. A number of publications in the late 1960s and in the 1970s addressed the issue (e.g., many of the articles included in Cohen 1974, Cohen and Wirth 1975, and Eckman 1977); see Schütze 1996 for an extensive literature review. Concerns seem to have resurfaced in recent years; cf. the 2005 special issue of Lingua on "Data in Theoretical Linguistics," the 2007 special issue of Theoretical Linguistics volume on "Data in generative grammar," Devitt 2006, and Fitzgerald 2010, among many others.
***

What is crucial but is not discussed in much depth at all in "Hypothesis testing in generative grammar: Evaluation of predicted schematic asymmetries" includes the significance of "preliminary experiments." The issue is closely related to the so-called Duhem-Quine thesis and the fact that our testable predictions concerning properties of the Computational System of the language faculty must be made on the basis of, at least, two types of hypotheses, one universal and the other language-particular. The illustration in the paper includes discussion of three (sets of) hypotheses:

(i) the lexical hypothesis about so vs. a demonstratives
(ii) the lexical hypotheses about so-called local anaphors in Japanese
(iii) the universal hypothesis about a necessary structural (i.e., LF c-command) condition for a formal relation that underlies a certain type of dependency intuitions.

The predictions made under each of (i)-(iii) appear to be disconfirmed. See Aoshima, Sachiko, Yoshida, Masaya, and Phillips, Collin. 2009. Incremental processing of coreference and binding in Japanese. Syntax 12.2: 93-134 for such experimental results with regard to (i) and (iii) and "Hypothesis testing in generative grammar: Evaluation of predicted schematic asymmetries" contains experimental results with regard to (ii). One of the main empirical points in "Point-Value Predictions in Language Faculty Science: Suggestions and Illustration" is as follows: the prediction made under (i) is supported once we pay close attention to the designs of experiments; likewise, the prediction made under (iii) is supported once we pay attention to preliminary experiments; but in the case of the predictions made under (ii), an attempt to save them actually results in a worse case of disconfirmation.

Hoji 1985 (Ph.D) Logical Form Constraints and Configurational Structures in Japanese and Hoji 1990 Theories of Anaphora and Aspects of Japanese Syntax can be downloaded here.

"Evaluating the Lexical Hypothesis about Otagai" in Linguistic Research 17.1. pp. 65-119.

The paper contains a summary of the methodological proposal in the book manuscript that I prepared in 2009 (A Foundation of Generative Grammar as an Empirical Science), which I did not get to publish. A more developed version of that work is being prepared as Language Faculty Science: Methodology and Empirical Demonstration, to be published by the Cambridge University Press. Much of the content of Hoji 2006 Otagai is reproduced in the Appendix of this paper.

Abstract
This paper explores how the hypothetico-deductive method can be applied to research concerned with the properties of the language faculty by illustrating how it can be applied to the language-particular hypothesis that otagai in Japanese is a local anaphor. The paper adopts Chomsky's (1993) conception of the Computational System (hypothesized to be at the center of the language faculty) and considers informant judgments to be a major source of evidence for or against hypotheses about the Computational System. Given that informants' acceptability judgments can be affected by various non-grammatical factors, it is imperative, for the purpose of putting our hypotheses to rigorous test, that we have a reasonably reliable means to identify informant judgments as a likely reflection of properties of the Computational System (or properties of the language faculty that are directly related to the Computational System). The paper suggests a means to do so. I maintain that we are led to some version of it once we adopt the basic assumptions noted above, along with the research heuristic, explicitly advocated by K. Popper, that we should maximize our chances of learning from errors; cf. Popper 1963. The paper then examines, in accordance with the proposed method, the predictions made under the lexical hypothesis that otagai is a local anaphor and shows that the predictions are not borne out. If what underlies a local anaphor is closely related to "active functional categories" in the sense of Fukui 1986 and if, as suggested in Fukui 1986, the mental lexicon of speakers of Japanese lacks them altogether, this result is as expected.

"KARE" (1991) in Carol Georgopoulos and Roberta Ishihara, eds., Interdisciplinary Approaches to Language: Essays in Honor of Prof. S.-Y. Kuroda, Reidel, Dordrecht, pp. 287-304.
(A pre-publication version, but there is no substantive difference between the published version and this version.)

This paper was written before Hoji 1990.

"On the so-called Overt Pronouns in Japanese and Korean" (1990) in E.-J. Baek, ed., Papers from the Seventh International Conference on Korean Linguistics, pp. 61-78.

This paper was written after Hoji 1991 "Kare," the final version of which had been written in 1989 but did not appear until 1991.

"Assessing Competing Analyses: Two Hypotheses about 'Scrambling' in Japanese" (2006) in Ayumi Ueyama, ed., Theoretical and Empirical Studies of Reference and Anaphora?Toward the establishment of generative grammar as an empirical science, pp. 139-185.

The paper compares Ueyama 1998, 2003 with Saito 2003. It can be considered as a(n) (almost direct) continuation of Hoji 2003 in Lingua and as a predecessor of Hoji 2015 Language Faculty Science. The paper contains some of the materials discussed in the papers and handouts placed in the Past Posting board under Discussion here (see [21482] "Mayfest at U. of Maryland."). In particular, some/much of the Mayfest2005-related documents, both empirical and methodological, has been incorporated in this paper. The methodological issues addressed in the Major-Object-related documents placed under [24591] "LSA Workshop on Control and Raising: Handouts" and [20194] "The 'raising-to-object' construction in Japanese" are also addressed in this paper, hopefully in a way that is clearer than the works prior to this paper.

There are some postings in the General Remarks board on the issues related to falsifiability in the context of generative grammar. The first paragraph of this paper, if expanded, would have contained some materials in the postings alluded to above.

There is an error in the paragraph right below (104). The paragraph contains (i).

(i) We have also seen that otagai need not have 'its antecedent' in its local domain or is it necessary for otagai to be c-commanded by 'its antecedent'; see (92)-(95).

"or is it" should have been "nor is it."

"Otagai" (2006) in Ayumi Ueyama, ed., Theoretical and Empirical Studies of Reference and Anaphora—Toward the establishment of generative grammar as an empirical science, Report of the Grant-in-Aid for Scientific Research (B), Project No. 15320052, Supported by Japan Society for the Promotion of Science, Kyushu University, pp. 126-138.

The paper demonstrates that otagai in Japanese is not a local anaphor, contrary to a widely-held and yet rather poorly supported thesis. The paper puts forth an alternative analysis of otagai, but its main contribution is its demonstration that otagai is not a local anaphor, despite the exposition given in the paper. A brief historical background of this paper is given in "Postscript in 2006," which I reproduce here.

***
Postscript in 2006

I have been advocating the view in the preceding pages at least since the spring of 1993. Most of the empirical materials are contained in "Otagai," presented at the 16th West Coast Conference on Formal Linguistics, University of Washington, March 2, 1997 and "Movement and Dependency: On the Landing Site of Scrambling," presented at the Stanford University Linguistics Colloquium, May 26, 1995. Some of the arguments are introduced in Ueyama 1998, and Hoji 2003. One might wonder why the hypothesis that has been falsified quite clearly and blatantly has continued to be used in a crucial way in many of the works even up to the present time. The reason, I believe, has to do with the lack of understanding on the part of many practitioners of the significance of negative predictions and falsification in linguistic science, which is addressed to some extent, but admittedly insufficiently in Hoji 2003.
***

Hoji, Hajime (2003) "Falsifiability and Repeatability in Generative Grammar: A Case Study of Anaphora and Scope Dependency in Japanese," Lingua, vol.113, No.4-6, pp.377-446.
Ueyama, Ayumi (1998) Two Types of Dependency, Doctoral dissertation, University of Southern California, distributed by GSIL publications, University of Southern California, Los Angeles.

Hoji 2003: 4.2.2 contains some examples that are not provided in the "Otagai" paper, including:

(126)a.Otagai-ga manzoku nara, boku-wa monku-o iwanai tumorida.
otagai-NOM satisfied if I-TOP complaint-ACC say:not intend
'I will not complain if those two feel happy.'

b.Haru-no atatakana kaze-ga otagai-o totemo siawase-na kimoti-ni sita.
spring-GEN warm wind-NOM otagai-ACC very:much happy feeling-to made
'The warm spring breeze made the two feel very happy.'

One might wonder why the hypothesis that has been falsified quite clearly and blatantly has continued to be used in a crucial way in many of the works even up to the present time. The reason, I believe, has to do with the lack of understanding on the part of many practitioners of the significance of negative predictions and falsification in linguistic science, which is addressed to some extent, but admittedly insufficiently in Hoji 2003.

"Assessing Competing Analyses: Two Hypotheses about 'Scrambling' in Japanese" in Ayumi Ueyama 2006, ed., Theoretical and Empirical Studies of Reference and Anaphora—Toward the establishment of generative grammar as an empirical science, Report of the Grant-in-Aid for Scientific Research (B), Project No. 15320052, Supported by Japan Society for the Promotion of Science, 2006, Kyushu University, pp. 139-185., I believe, provides a better illustration than Hoji 2003 of negative predictions and falsification in linguistic science.

The Methodology Board [27158] "Popper (1983) on falsifiability and falsification" may be of some interest in this connection.

"What Gets Mapped to the Tripartite Structure of Quantification in Japanese," (2004) Hoji & Ishii WCCFL 23 Proceedings, ed. Vineeta Chand et al., 346-359. Somerville, MA: Cascadilla Press.
(Note: The version available here is not identical to the published version in some non-substantive respects, including the page numbers.)

The paper endorses the view, advocated in Gunji & Hasida 1998 and Takami 1998, that the association between a 'floating' numeral-classifier in Japanese (such as 3-nin '3-classifier') and its 'host NP' is not grammatically constrained, unlike the 'standard' view that is suggested in Kuroda 1980 and argued for in Miyagawa 1989. It also supports the thesis, defended in Fukushima 1991, Gunji & Hasida 1998, and Kobuchi-Philip 2003, among others, that a numeral-classifier (#-CL) can be base-generated as an adverbial and a 'floating' numeral-classifier is indeed an adverbial. Miyagawa and Arikawa's (2003) account of 'long distance association' between a 'floating' #-CL and 'its host', in defense of the 'standard' view, is shown to make a wrong prediction. Some empirical materials are then considered where the #-CL and the 'host NP' together serve as the intended 'binder' of a 'dependent element' in terms of bound variable anaphora. It is suggested that both the #-CL and the 'host NP' in such cases must c-command the 'dependent element' prior to the application of LF movement. The Isomorphism Principle of Reinhart (1976) and Huang (1982) enables us to put forth testable predictions by making reference to surface forms, and a few of them are considered. The suggested analysis is based on the hypothesis that an adverbial #-CL and the 'host NP' in such cases as alluded to above get mapped to an operator and its restriction, respectively, of the tripartite structure of quantification. Some consequences of the hypothesis are discussed.

"Reconstruction Effects in Passive and Scrambling in Japanese," (2008) "Reconstruction Effects in Passive and Scrambling in Japanese," in Mutsuko Endo Hudson, Sun-Ah Jun, Peter Sells, Patricia M. Clancy, Shoichi Iwasaki and Sung-Ock Sohn, eds., Japanese/Korean Linguistics Conference Vol. 13, CSLI, pp. 152-166

The conference was held in 2003, and there was a considerable delay in publication.

Because of the page limit, I was not able to include various materials, empirical, conceptual and methodological, in the paper. Virtually no reference is made to the issues re. negative propositions and negative predictions, with which the bulk of the JK presentation was concerned.

"Falsifiability and Repeatability in Generative Grammar: A Case Study of Anaphora and Scope Dependency in Japanese," (2003) Lingua 113/4-6, pp.377-446. Included as Ch. 7 in Hoji 2013, edited by A. Ueyama and Y. Takubo.

You can also obtain the other papers included in the special issue of Lingua if you have access to the Lingua web site.

"Demonstratives in Modern Japanese," (2003) (with S. Kinsui, Y. Takubo, and A. Ueyama) In: A. Li, and A. Simpson (eds.), Functional Structure(s), Form and Interpretation: Perspectives from East Asian Languages. London: Routledge, pp.97-128.

This paper makes an attempt to extend the theory of Ueyama 1998 -- which deals with a-NPs and so-NPs in their non-deictic uses -- to ko-NPs as well as to the deictic uses of the ko/so/a-NPs. It contains an appendix on Kuroda 1979.

The treatment of the deictic so-NPs in the paper is clearly not satisfactory, and I am currently working on it, based on the presentations at ATR and Nanzan at the end of 2000.

"Surface and Deep Anaphora, Sloppy Identity, and Experiments in Syntax," (2003) In Anaphora: A Reference Guide, ed. A. Barss, Blackwell, pp.172-236. Included as Ch. 6 in Hoji 2013, edited by A. Ueyama and Y. Takubo.

This paper is an attempt to articulate what was behind my 1998 LI paper on null objects in Japanese. If you are interested in issues in ellipsis (and would like to look at them from the perspectives of Japanese), I suggest that you read my 1998 LI paper, this paper, Fukaya & Hoji 1999, and Mukai 2003, all of which, except for my LI paper, can be downloaded at this HP. If you are interested in works on ellipsis in general, I suggest that you visit Jason Merchant's HP.

If you are interested in getting the basic idea/claim of the paper, without getting into the details of the paradigms in Japanese, here is the handout of a recent talk I gave, based on the "Surface and Deep ..." paper. It is basically the same as the one I used at a workshop at UNICAMP in Brazil in the summer of 2001, and it only deals with English paradigms.

For the recent presentation, I meant to include the methodological remarks along the lines of the JK presentation of 2003 (the handout of which is posted here), but I was not able to fit the relevant materials in the handout. Neither was I able to include discussion of Jason Merchant's recent proposal on 'fragments' (the paper downloadable at his HP) in the handout.

When I return to this topic, I will try to address the empirical generalization that has motivated (49) (as discussed in Fiengo&May 1994, Dahl 1974 and Sag 1976, Dalrymple, Shieber, & Pereira 1991, Fox 1998, 2000, and Tancredi 2000) and how the effects of (49) can be derived.

(49) *FD(A, B) if B is c-commanded by an NP C, where A and C have the same (indexical) value and C does not c-command A; cf. Fox 1998 and 2000: chap. 4.

"Demonstratives, Bound Variables, and Reconstruction Effects,"(with S. Kinsui, Y. Takubo, and A. Ueyama) Proceedings of the Nanzan GLOW: The Second GLOW Meeting in Asia, (2000), pp. 141-158.

"Stripping and Sluicing in Japanese and Some Implications," (with Teruhiko Fukaya) (1999) In Proceedings of The 18th West Coast Conference on Formal Linguistics, Cascadilla Press. (www.cascadilla.com)

"Sloppy Identity and Formal Dependency" (1997b) in Proceedings of the 15th West Coast Conference on Formal Linguistics, Center for the Study of Language and Information, Stanford. pp. 209-223. Included as Ch. 3 in Hoji 2013, edited by A. Ueyama and Y. Takubo.

"Demonstrative Binding and Principle B" (1995) in Proceedings of the Northeast Linguistic Society 25, pp. 255-271. Included as Ch. 1 in Hoji 2013, edited by A. Ueyama and Y. Takubo.