Exploring Grammatical Gender Agreement in Russian Learners of Greek: An Eye-Tracking Study

Round 1

Reviewer 1 Report

This study examines if Russian L1-Greek L2 learners are sensitive to gender incongruencies in  determiner-noun and adjective noun pairings using an eyetracking paradigm while participants  read passages for comprehension. Both Greek and Russian feature grammatical gender, but they  differ in how gender is marked, and therefore implications about how L1 processing can affect  L2 parsing are discussed. Nevertheless, this study is innovative in that it employs a Greek L2  learner corpus to motivate the design of their stimuli, uses contextualized short passages, and  triangulates the online data with an offline post-reading questionnaire to explore if longer  fixations on agreement violations could be associated with explicit awareness of the study’s  manipulation. Results revealed that Russian L1-Greek L2 learners are sensitive to adjective-noun  gender manipulations in the late measure of total duration time, and to noun-adjective gender  manipulations in both late measures (total duration and selective regression path). This  sensitivity was not affected by phonological conditions. Results from the post-reading  questionnaire showed that only 33% of participants reported noticing grammar mistakes and only  9% of them was able to recall the type of error. When the type of error was provided, only seven  participants indicated determiner-noun errors and only one marked adjective-noun errors.  Overall, these findings suggest that intermediate Russian learners of Greek are aware of the  gender feature and its respective morphological computations, and that the study design (with  contextualized passages and post-reading questionnaire) suppressed awareness of the critical  stimuli. 

General comments: 

The manuscript is very well-written and easy to follow. I commend the authors for spending so  much time and effort designing the stimuli, which I believe increases the ecological validity of  the data obtained. I do not see any major concerns with the study design, or theoretical  argumentation, but I have a few suggestions to improve the manuscript: 

Abstract 

This sentence (lines 19-21), particularly the underlined section is hard to parse: “…the gender  agreement data suggests that the abstract gender feature is present in the developing grammar of  Russian learners of Greek, while the agreement computations are operating for the same  participants of the study.” Do you mean that learners possess both: (a) knowledge of lexical  gender, and how it should agree with nouns and adjectives? Please rephrase to ensure clarity. 

Introduction 

Line 78: Please spell out GLCII for the first time. It is spelled out once before, but in the abstract. 

Syntactic background: gender 

Line 219: Does Ds stand for determiners? Please spell out. 

Study 

Participants

According to more contemporary terminology, it is not accurate to say that the Russian speakers  were monolingual (line 327), I recommend the use of the term “monolingually raised Russian  speakers”, since these participants are, to some extent, proficient in Greek. 

Preliminary considerations on the design 

Similar to the current study, Beltrán-Lopez & Dussias (2023) in Bilingual Approaches to  Bilingualism (forthcoming) used a corpus to create their eyetracking stimuli. If I recall correctly,  they used sentences verbatim, but I believe this study is worth citing as it sets a precedent  together with the present study for future studies to adopt such an approach when creating  critical stimuli in sentence processing research. 

https://www.researchgate.net/publication/372336249_Heritage_speakers%27_processing_of_the _Spanish_subjunctive_A_pupillometric_study?fbclid=IwAR2En8vu5Re5Io9fyMIm3- iVmKbBEmm8cgUpIK6ObVNkj-JZ_8T-7qTvQd8 

Because of the unconventional stimuli design approach, several aspects of the stimuli are not  counterbalanced as traditionally expected (e.g., nouns not being equally dispersed by formal  aspects such as gender). I wonder if the authors could control for noun frequency, since  frequency has been found to affect language processing, particularly with L2 learners. Hopp  (2014, 2018) examined factors at the world level and how they can have implications for  processing at the phrase and sentence level. Frequency is one of them. If there is a Greek corpus  out there with frequency data, it may be a good idea to examine the frequency of the nouns used  in the study to determine if there is a significant difference between those used with determiners  and adjectives. If there is a difference, the authors can always add frequency as a covariate in the  model (as in Fernández Cuenca & Jegerski, 2022). 

Fernández Cuenca, S., & Jegerski, J. (2022). A role for verb regularity in the L2 processing of  the Spanish subjunctive mood: Evidence from eye-tracking. Studies in Second Language  Acquisition, 1-30. doi:10.1017/S027226312200016X  

Hopp, H. (2014). Working memory effects on the L2 processing of ambiguous relative clauses.  Language 

Acquisition, 21, 250–278. 

Hopp. H. (2018). The bilingual mental lexicon in L2 sentence processing. Second Language, 17,  5–27. 

Materials and methods 

Typo in line 370 (e.ye tracking) 

Could you elaborate on what “attribute settings” mean in line 397? 

In lines 401-402, the authors say “The ratio of critical to non-critical items was 25%.” What did  the non-critical items look like? Were they distractors, fillers? It is hard to decipher given the  innovative stimuli design. It would be important to clarify this by providing a clear description of  how this was done for future researchers to use this study as a reference point.

For those of us unfamiliar with Greek and Russian, it would be helpful to have English  translations or linguistic annotations for table 4 (after line 411). This table is fundamental to  understand the study’s conditions, but the way it is set up right now, forces the reader to go back  to table 1 to remember which suffixes correspond to which grammatical gender. 

If I understood correctly, there is a comprehension-based question after each passage. Could you  provide an example of a passage and a comprehension question to illustrate a critical item and to  show that the comprehension questions did not target the critical stimuli directly? 

In line 418, the authors mention that a minimum of 80% accuracy in responses was established  to determine if participants had properly read the texts (passages). Were the passages with an  inaccurate response included as part of the final data pool that was analyzed? Please specify, if  they were discarded indicate and what percentage of the total data pool it was. 

Despite the post-reading questionnaire being a great method to triangulate level of awareness and  validity of the online data collected, the terminology used (e.g., illicit use of certain nouns or  tense-adverb violations) seems to be very technical. Did you check to see if participants were  familiar with this terminology? Were they linguistics students? This could be a limitation  explaining why only 3 of the participants (out of the 8 that reported seeing errors) were able to  identify which particular kind of error they were. 

Results 

Eyetracking registration 

It would be very helpful to see the distribution of words in a critical region of interest. Was each  word a region of interest? Were the determiner and noun (for example) two adjacent areas of  interest? Was there no analysis of a spill-over region? 

Line 529-530, please elaborate on why these types of measures were chosen for the particular  target constructions examined in the study. For example, total dwell time is a good late measure  to examine sensitivity to higher-level factors related to syntax, semantics, and pragmatics,  whereas first fixation is often used to examine lexical factors such as word frequency, polysemy,  etc. Clifton et al (2007) and Godfroid’s eyetracking book (2020) are good references for this. 

Clifton, C., Staub, A., & Rayner, K. (2007). Eye movements in reading words and sentences. In  R. P. G. Van Gompel, M. H. Fischer, W. S. Murray, & R. L. Hill (Eds.), Eye-movements:  A window on mind and brain (pp. 241–372). Elsevier Science. 

Godfroid, A. (2020). Eye tracking in second language acquisition and bilingualism: A research  synthesis and methodological guide. Routledge. 

What software was used to run the statistical analysis? R, SPSS? This information is relevant for  how the most effective mixed effects models are built. 

In lines 544-545, when discussing the adj-noun results, the authors state that there “ is an  increase in the fixation times for both phonological and non-phonological conditions”. I  encourage the authors to rephrase this and be more specific by stating that the increase could be 

observed only with late measures (selective RP, and total duration fixation), but so much with  the early measure of first pass (only in the right direction in the non-phonological condition). I  would also use the same terminology for the det-noun results, in lines 570-571, where the  difference in the expected direction (ungrammatical higher than grammatical) is present with  both early and late measures. The timing of the sensitivity is relevant, since these results suggest  that Russian learners of Greek were sensitivity to gender incongruencies faster (first pass) with  noun-determiners than noun-adjectives pairings. 

It is not clear why pairwise comparisons were run for what I believe is grammaticality and  phonological conditions in lines 588-589, since there was not significant interaction (p=0.677). 

Vocabulary and gender assignment post-test 

In line 621, the authors say “the respective data were not included in the analyses”, is the  respective data the vocabulary accuracy scores? And is “the analyses” referring to the  eyetracking analyses? Meaning that they did not discard nouns that were familiars for specific  participants? Please specify.  

Results 

In line 640, “the results indicate…” I would specify which tasks these results come from. Are  these results from the post-reading questionnaire or the vocabulary and gender assignment task?  Similarly, should “using an eyetracking paradigm” be added after “to explore the computations  operating at the syntactic level” to clarify the task these results refer to? 

In line 659, the authors state that “the violation was detected earlier in the Det-N condition”. My  suggestion for establishing a difference between early and late eyetracking measures when  discussing the descriptive stats (in lines 544-545 and 570-571) would connect very well with this  observation. Also, in line 663, the authors mention “word integration” making reference to a  particular eye movement measure (an early measure, in fact). It is for this reason that I suggest a  more elaborated description in section 5.1 explaining why these three eyetracking measures were  chosen for this study. 

One of the most interesting points mention in the discussion can be found in lines 674-675,  where the researchers point out that production learner data from the corpus matches the  processing data found in the study. Perhaps the authors could elaborate more on this. For  example, does the fact that Russian learners of Greek have more difficulty computing adj-noun  incongruencies signify that this is a processing issue (at the input-intake stage) that affects the  acquisition of how this feature works in Greek ? How does the post-reading questionnaire inform  this proposal if only one participant was able to identify adj-noun errors? How does this relate to  established theories in the field such as shallow processing or the feature reassembly hypothesis? 

There is a no mention about proficiency level affecting sensitivity to gender agreement violations  in the discussion, but this is briefly mentioned in the literature review. Was the proficiency level  of the Russian learners of Greek comparable to that of those participants in previous studies? 

Finally, the authors make a strong point towards the end of the discussion highlighting the  importance of contextualizing the stimuli to increase the ecological validity of the data 

elicitation. Although not exactly the same, there are a couple of studies that examined how the  distractor task used to examine language processing can alter how much attention participants  pay to the stimuli. I leave the references here in case the authors see the value in mentioning  these studies to strengthen their argument: 

Leeser, M., Brandl, A., & Weissglass, C. (2011). Task effects in second language sentence  processing research. In P. Trofimovich & K. McDonough (Eds.), Applying priming  methods to L2 learning, teaching, and research: Insights from psycholinguistics (pp. 179- 198). John Benjamins. 

Tokowicz, N., & MacWhinney, B. (2005). Implicit and explicit measures of sensitivity to  violations in second language grammar: An event-related potential investigation. Studies  in second language acquisition, 27(2), 173-204

Author Response

Dear Reviewer 1.

We would like to express our gratitude for the time and effort you dedicated to reviewing our manuscript. Your insights and constructive suggestions have been instrumental in enhancing the quality of the paper. We have taken each of your comments into account and have made the necessary revisions to address them. Below, we elaborate on some of the comments you made for which we feel clarification is warranted.

Regards.

Alex Tantos, Nikos Amvrazis & Konstantinos Angelou

Author Response File: Author Response.pdf

Reviewer 2 Report

See attached file.

Comments for author File: Comments.pdf

There are several places where the English needs work.  I am not opposed to having features of speaker dialect or L2 features in an article, but in many places this article is difficult to read and understand.

Author Response

Dear Reviewer 2.

We would like to express our gratitude for the time and effort you dedicated to reviewing our manuscript. Your insights and constructive suggestions have been instrumental in enhancing the quality of the paper. We have taken each of your comments into account and have made the necessary revisions to address them. Below, we elaborate on some of the comments you made for which we feel clarification is warranted.

Regards.

Alex Tantos, Nikos Amvrazis & Konstantinos Angelou

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

This revised draft is improved and most of my concerns have been met. There are still a few lingering issues that I feel could be addressed to strengthen the paper. See the attached document. 

Comments for author File: Comments.pdf

Author Response

We would like to extend our sincere appreciation to the reviewer for the meticulous and insightful comments during the second-round review. We are pleased to inform that all the recommendations have been incorporated into the revised manuscript. We believe that with these modifications, the paper now presents a more robust and refined contribution to the field.

Thank you for your invaluable guidance and support.

Author Response File: Author Response.pdf