The Rightward Analysis of Wh-Movement: A Reply to Petronio and Lillo-Martin 1997 Neidle, MacLaughlin, Lee, Bahan, and Kegl Language 74:4, 819-831. Unfortunately, several passages from our article did not appear in the printed version. These sections are provided below, so that the reader may reconstruct the article as originally intended. Passage 1: importance of access to video data exemplifying constructions presented in gloss form Passage 2: inconsistencies in the reporting of Petronio's 1991 results. Passage 3: Section 3.8. SUMMARY. Passage 1 - p. 822 3. DECIDING BETWEEN THE TWO ANALYSES. We now consider the predictions of these analyses in relation to data from ASL. The fact that ASL is a visual language, with no written form, presents certain difficulties for the representation and reporting of data. Glosses omit tremendous amounts of detail, and it is virtually impossible to reconstruct an example based on a gloss alone. For this reason, we make available over the Internet digitized video exemplars of the grammatical sentences we report on. insert: We hope that Petronio and Lillo-Martin, who also acknowledge the inadequacy of gloss representations, will likewise make available video corresponding to the examples they presented in gloss form in P&L 1997. (This would be particularly helpful since we have been unable to reproduce many of the sentences that they report as grammatical, including both WH and non-WH constructions.) Especially given the complexities involved in the elicitation of grammaticality judgments in ASL (see NKMBL 1999), in the absence of access to video exemplars of the data under discussion, it is impossible to assess differences in grammaticality judgments reported in the literature. It is essential that videotaped data be made available for public inspection if issues surrounding the data are ever to be resolved. [back to the top] Passage 2 - p. 822 3.1. WH-OBJECTS IN INITIAL POSITION. We begin with the most fundamental constructions critical to deciding between leftward and rightward WH-movement: sentences containing a single WH-phrase.4 Our analysis predicts sentences like (5), with a sentence-initial WH-object, to be ungrammatical, while they are predicted to be grammatical by P&L's analysis. Such sentences are reported to be ungrammatical by our informants (when signed exactly as glossed).5 insert: Petronio (1991:212) reports that "signers who came from Deaf families where their parents used ASL" usually reported sentences with an initial WH-object to be ungrammatical, while "signers who came from hearing families (i.e. their parents did not use ASL)" sometimes concurred with the native signers in finding them ungrammatical, but sometimes did not. Interestingly, Petronio (1993:99), while presenting no new relevant evidence, summarizes these findings as follows: "In previous work (Petronio 1991), I reported that some ASL signers accept whOSV in direct questions while others reject it." No mention is made in Petronio 1993, however, of that fact that, according to Petronio 1991, native signers generally reject such sentences. Although P&L discuss this construction (1997:50-51), they report "varying judgments" and make no explicit claim about the grammaticality of the critical sentences. They summarize their own previous reports on such sentences as follows: "Lillo-Martin 1990 and Lillo-Martin & Fischer 1992 report them as grammatical, and Petronio 1993 reports that they receive mixed judgments." The misleading characterization in Petronio 1993 of the findings of Petronio 1991 is thus perpetuated in P&L 1997, since, yet again, no mention is made of the fact that the native signers tested by Petronio generally reject such sentences. Thus, notwithstanding the characterization in P&L 1997, sentence-initial WH-objects are generally rejected by native signers. [back to the top] Concluding section - p. 829 insert: 3.8. SUMMARY. We have demonstrated that the data from native signers are consistent only with a rightward WH-movement analysis. In sentences containing a single WH-phrase, that phrase may be moved to a clause-final spec-CP position, but not to a left-peripheral position. This also holds for WH-phrases extracted from within an embedded clause; such phrases may move to the right periphery of the matrix clause, but may not precede the matrix clause. The distribution of non-manual WH-marking follows from generalizations about the distribution of non-manual syntactic marking in ASL, given the rightward movement analysis we have proposed. When manual material is available in the rightward spec-CP position, the spread of WH-marking over the rest of the CP is optional; otherwise, spread is obligatory. We have argued against P&L's claim that right-peripheral WH-material cannot be phrasal. Thus, P&L's postulation of a single C node intended to house "focus" elements (including WH-signs) not only lacks independent motivation, but also cannot account for the occurrence of WH-phrases sentence-finally. We have discussed a major problem with P&L's claim that non-manual WH-marking necessarily spreads over the entire CP. To maintain this claim, they are forced to analyze examples in which WH-marking occurs solely over the final WH-phrase as multisentence discourses. We have shown that this is untenable. P&L's analysis makes incorrect predictions for the relative ordering of topics (adjoined to CP) and left-peripheral WH-phrases (in spec-CP, on their account). In fact, WH-phrases may precede or follow (other) base-generated topics, as correctly predicted by our analysis of left-peripheral WH-phrases as base-generated topics. Finally, we have questioned the validity of P&L's use of variability in grammaticality judgments as evidence to support their distinction between WH-traces and "null WH-elements." Essentially, P&L report variability for the majority of the constructions they discuss. Some of this variability is attributed to the presence of null WH-elements, while some is considered to be idiosyncratic; no principled basis for distinguishing between the two is provided. [back to the top]