When Pitch Falls Short: Reinforcing Prosodic Boundaries to Signal Focus in Japanese

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

As a way to explore how Japanese speakers use prosodic features to realize syntactic structure, the reported study used a set of two-words genitive noun phrases ‘X no Y’ (X’s Y), where either of the two nouns receives syntactic focus, and measured three prosodic features – silence duration between the two words, degree of jaw opening (as inferred by F1) in the particle ‘no,’ and duration ratios of the two words to examine how well these non-pitch-based features that occur at phrasal boundaries predict the focus structures of the phrase and how these features are influenced by other prosodic features as well as lexical accent of the words.

The study addresses an important topic of syntax-prosody interface and proposes an illuminating concept, which is termed in the manuscript as the “Edge-Reinforcing Strategy.” The manuscript also presents excellent review of previous studies on the topic. However, the current manuscript needs major revisions in the framing of the study, connection of the research questions and literature review, and presentation of the results to support the claims, which requires re-writing the entire manuscript. I recommend the manuscript to be returned to the author(s) for major revisions and be reviewed again at that point. Below are the points of revisions, which I hope to be useful for the author(s).

1. Framing of the study - On RQ2-M3 and other models where the FocusNB is used as a to-be-predicted factor:  It is a production study, where focus structures were given, not a perception study where the perceived focus structure was measured, it is problematic to treat the Focus as a to-be-predicted factor. Some features that systematically vary across given conditions may or may not be perceived and used by listeners to make syntactic judgements. I’d suggest the texts be revised so as to match what’s manipulated as conditions and what’s measured.

2. Analyses:
   1. Line 351: Please indicate how the data were coded for Lexical Accent. Lines 303-306 indicate that there were four levels for this factor.
   2. Please indicate how many tests were conducted. Were there four tests (JawOpening, Silence, Duration, and FocusNB as Dependent Variable) with all factors and interactions as predictors, though only subsets were reported? Or, was a separate test conducted for each model?
3. Models and Results (pp. 9-13):
   1. Please present results of all research questions. For example, we’d like to see the results of Focus (three levels) and/or FocusNB (two levels) as a main factor on JawOpening, Silence, and Duration. I’d suggest the paper to present the full results, including all main effect and all interactions in one table for each dependent variable (, based on one test for each dependent variable (see 2.2. above).
   2. Please present figures to show the directionality of differences in the results to allow readers to verify the texts (e.g., lines 436-437).
4. Connection between literature review and research questions: Connections between literature review on the one hand, and the research questions and models on the other hand, are not always clear. More explicit statements linking the previous findings (or unsettled aspects) that motivates each research question would be helpful for readers.
5. Writing: Overall, writing seems to be a little repetitive. It could be condensed more to allow space for logical connections in Section 1 and 2 (Introduction and Methodology) and more tables and figures for Section 3 (Results)

End of Comments

Author Response

Response to the Editor and Reviewers

Dear Dr. Erickson and Reviewers,

We are very grateful for your detailed feedback and for the chance to improve our manuscript. The reviews prompted us to make substantial changes throughout the paper. In particular, as Reviewer 1 suggested, we re-framed the study, strengthened the connections between the literature and the research questions, and reworked the presentation of the results. Following the Editor’s guidance, we also made it clear that the present study concerns production only, while the perceptual side of the hypothesis is left for future work. Reviewer 2’s suggestions were equally helpful: they led us to expand the methodological description of silence detection and to add a new figure illustrating how silences were coded.

Below we address each comment point by point. Reviewer and editor comments are in bold; our responses follow.

Reviewer 1

“The current manuscript needs major revisions in the framing of the study, connection of the research questions and literature review, and presentation of the results…”

We took this seriously and rewrote the introduction and literature review. Section 1.2 now develops the Edge-Reinforcing Strategy as a production hypothesis, and Section 1.3 presents the four research questions in a more concise way that avoids repetition. In the results, we added summary tables and figures, and in the appendix we now include the full model outputs for RQ1 and RQ2.

“On RQ2-M3 and other models where the FocusNB is used as a to-be-predicted factor…”

We revised the text to clarify that production is our domain. Models that predict FocusNB are framed as diagnostic tools on the production side, while perceptual testing remains future work.

“How was lexical accent coded?”

We added explicit definitions of AccentW1 and AccentW2 in the Methods section.

“How many tests were conducted?”

We now list each model clearly in §2.5 and explain their rationale. Full outputs for RQ1 and RQ2 are provided in the appendix.

“Please show the full results and figures…”

We added Figures 5a–c to display the effects of focus on jaw opening, silence, and duration ratio. The appendix contains the complete models for RQ1 and RQ2. Including all models for all questions would have made the appendix prohibitively long.

“Clarify links between the literature and each RQ.”

Sections 1.2–1.3 were rewritten to map earlier findings directly to the four RQs.

“Condense writing to free space for logic and results.”

We reduced repetition in the background and methods and used that space to expand the presentation of results.

We thank both reviewers and the editor again for their careful reading and helpful suggestions. The paper is, we believe, much clearer, better framed, and more transparent as a result.

Sincerely,

Reviewer 2 Report

Comments and Suggestions for Authors

Please see comments in attached file.

Comments for author File: Comments.pdf

Author Response

Response to the Editor and Reviewers

Dear Dr. Erickson and Reviewers,

We are very grateful for your detailed feedback and for the chance to improve our manuscript. The reviews prompted us to make substantial changes throughout the paper. In particular, as Reviewer 1 suggested, we re-framed the study, strengthened the connections between the literature and the research questions, and reworked the presentation of the results. Following the Editor’s guidance, we also made it clear that the present study concerns production only, while the perceptual side of the hypothesis is left for future work. Reviewer 2’s suggestions were equally helpful: they led us to expand the methodological description of silence detection and to add a new figure illustrating how silences were coded.

Below we address each comment point by point. Reviewer and editor comments are in bold; our responses follow.

Reviewer 2

“You start with ip/AP, then discuss syntactic XP; be explicit how cues map to prosodic boundaries.”

We clarified the text and figure captions to distinguish IP and AP and made sure the description of Figure 1 states that reset belongs to the IP level.

“The introduction frames a cross-dialectal mechanism, but the study itself is only on Standard Japanese.”

We now emphasize earlier that Standard Japanese is a conservative test case, where we expect effects to be weaker than in more edge-prominent dialects like Koriyama or Kobayashi.

“What counts as a silence? Many tokens of W2 begin with /k/.”

We expanded §2.4 to describe the silence decision rule in detail. Pauses flagged by WebMAUS were manually checked: if W2 began with /k/ and the pause was <50 ms, it was relabeled as a stop closure; if ≥50 ms, it was retained as a silence. To make this procedure clearer, we added a new Figure 4 with four panels illustrating the cases.

“Line-by-line corrections.”

We revised the wording to say “educated speakers of standard Japanese,” ensured consistent attachment of no (e.g., umáno), clarified which unit resets belong to in Figure 1, specified which word is lengthened under narrow focus, used “genitive NP/DP” instead of “genitive XP,” described participants’ dialectal backgrounds and how the design elicited educated standard, noted that accent position was controlled (2nd mora in both nouns), and added glosses for all stimuli.

“Results transparency.”

We added Figures 5a–c to illustrate the effects and placed the full models for RQ1 and RQ2 in the appendix.

Finally, we corrected several Erickson citations and added a methodological note on the limits of using F1 of [o] as a proxy for jaw opening, pointing to direct articulatory methods as a direction for future work.

We thank both reviewers and the editor again for their careful reading and helpful suggestions. The paper is, we believe, much clearer, better framed, and more transparent as a result.

Sincerely,

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript demonstrates substantial improvements with clearer logical connections among the sections, especially the previous studies and the research questions, and these questions and the results. It engages with ample relevant studies and make significant contributions to the scholarship. The entire manuscript is carefully written. I believe it is ready for publication in present form.

Author Response

We sincerely thank you for your careful and detailed review. Your insightful comments and constructive suggestions in the previous round were extremely helpful in improving the clarity, coherence, and overall quality of the manuscript. We greatly appreciate the time and effort you dedicated to strengthening our work, and we are very pleased that the revised version now meets your approval.