Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Recommendation: Reconsider after major revision

This paper looks at bidirectional cross-linguistic influence between Mandarin and French, focusing on a nice test case: page-turning, where the two languages encode 'forward' and 'backward' in opposite ways. The idea of using a conflict domain like this is smart - it should make any L2-to-L1 effects easier to detect. Sample size is decent (n=261), and I appreciated that the authors included qualitative interviews alongside the quantitative task. The pedagogical discussion at the end is useful.

That said, I have some significant concerns about the methodology and, more importantly, how the results are interpreted. The main issue is that the negative correlation between French and Mandarin scores gets framed as evidence of cognitive restructuring / reverse transfer, but the design doesn't really allow for that conclusion. More below.

Major issues

1. Proficiency grouping

This is probably my biggest concern. About half of the participants (128/261) never took a formal proficiency test - they're grouped entirely on self-assessment. Self-reported CEFR levels are notoriously unreliable, especially among learners who may confuse time spent studying with actual proficiency. The other half have a mix of TFS-4, TFS-8, TCF/TEF, and DELF/DALF - but there's no explanation of how these different tests were mapped onto the A1-C1 scale or combined with self-reports to make the final groupings.

The paper says self-reported levels were 'broadly consistent' with university year (l. 225-227), but no data are shown to support this. Given that the central finding is that proficiency doesn't predict spatial task performance, getting the proficiency measure right really matters. Could you at least show a cross-tab of self-reported level by year of study? Or run analyses with years of French as a continuous predictor instead?

2. Task reliability

The experimental task is only 4 items total - 2 per language. I understand the page-turning domain constrains what you can test, but measuring 'spatial cognition' from two binary choices is really pushing it. With two items you can't compute internal consistency, and chance performance (getting both right) is 25%. Some discussion of this limitation would help. Also - why the 1-3 scoring rather than 0-2? It doesn't affect the stats but it's a bit confusing.

It would be useful to see the item-level data too. Were both Mandarin items equally easy? Did participants do better on comprehension vs production in French?

3. Interpreting the negative correlation

This is where I think the paper overreaches. The correlation between French and Mandarin scores is r = -.34, and this gets interpreted as evidence that adopting French spatial construals 'destabilises' L1 representations (section 5.2). But correlation isn't causation, and several other explanations fit the data just as well:

- Individual differences in flexibility: some people may be more willing to switch perspectives, doing better on French items but worse on Mandarin ones because they're less anchored to either frame

- Task confusion: participants who didn't understand what was being asked might have guessed inconsistently across languages

- Prior instruction: maybe some learners had been explicitly taught about French page-turning conventions and over-applied that logic even in the Mandarin condition

The discussion brings in neuroscience literature on bilingual control (Pliatsikas, Cong et al.) but those studies are about grey matter plasticity and prefrontal activation during language switching - pretty far from whether knowing French changes how you think about turning pages in Mandarin. I'd tone down the cognitive restructuring claims and present this as suggestive evidence that needs replication with a design that can actually test causality.

4. No control groups

There's no monolingual Mandarin group or native French group for comparison. So we don't know if the 83% ceiling on Mandarin items reflects genuine L1 stability or just an easy task, and we don't know what native French speakers would score on the French items. Even a small pilot would help here.

5. Statistical redundancy

Table 3 (Kruskal-Wallis) and Table 4 (Spearman) test the same thing - the relationship between French and Mandarin scores. Presenting both makes it look like there's more independent evidence than there actually is. Pick one, or at least be explicit that they're two ways of looking at the same bivariate association.

Minor issues

- Figure 3 x-axis says 'Chinese questions' but the text uses 'Mandarin' throughout. Please standardise.

- The interviews only sampled 'top-performing learners' (l. 273). That's fine for some purposes but limits what you can learn. Lower performers might have interesting things to say about why they found the task confusing.

- Was the order of questions fixed or randomised? If everyone did Mandarin first then French (or vice versa), that could introduce order effects. Please clarify.

- The Mann-Whitney comparing proficiency groups on French scores gave p = .056 - just missing significance. Worth discussing rather than just saying 'not significant', especially since a continuous proficiency measure might have shown something different.

- Effect sizes would be helpful throughout. With n=261 even small effects will be significant, so knowing the magnitude matters.

- The ethics statement says approval was waived on 5 January 2024, but Table 1 references 2025 university rankings. Could you clarify the timeline? (May just be a typo in the ethics date.)

- English proficiency isn't controlled for, which the authors acknowledge. But since English uses the same page-turning logic as French, this could be a confound worth taking more seriously.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is a very well researched paper. Please find my feedback attached. 

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for these very thorough revisions. I believe the manuscript is substantially improved now, and I am happy to recommend it for publication.