Round 1

Reviewer 1 Report

The paper is well written and soundly argumented. I deem it fit for publications after a few and very minor adjustments. The remarks I have can be found on the attached pdf document in the form of side comments. The English in which the paper is written could benefit from some revisions here and there (mostly it is just typos, but there are a few unhappy formulations: I have highlighted some of them, but I am not a proof-reader nor a native speaker, so some possibly remain).

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Introduction:

Please double check manuscript for typos, word choice, and formatting.

In general, the language used to present the research summary is author focused as opposed to concept focused. For example, on line 151, the sentence starts “The author argued…” This frame for presenting the information is distracting. Reframing the introduce to emphasize the associations between presented information would benefit the manuscript.

Stating there is a lack of research on a given topic is not sufficiently compelling rationale for new research. What is the problem being addressed or significance of the line of inquiry?

On the whole, introduction would benefit from inclusion of examples to illustrate described linguistic units and processes. (e.g., gender markers, determiners, etc.) earlier than line 158.

Need to define and provide examples for central concepts, such as opaque and transparency. Especially because these criteria are used for stimuli selection in the experimental design.

Line 64, 66- Morphemes, are by definition, the smallest contrasting units of meaning in a language. How can a morpheme not have semantic value?

The term morphophonological refers to the intersection of morphology and phonology. Therefore, the lack of phonological criteria in defining morphophonological features is questionable. (line 55)

Introduction would benefit of the inclusion of a discussion of the factors that may mediate language knowledge transfer in bilinguals, including the modality of language presentation (aural, text), distributed patterns of language exposure, and developmental patterns of acquisition.

In the summary of research involving cross-linguistic comparisons, it is not always clear what languages L1 and L2 refer to (line 91). Additional information about the age and task type of participants in this study would aim the interpretation and comparison of relevant results.

Methods:

The lack of reporting data cleaning procedures, missing data analyses, tests of data normality, and analytic procedures is a major weakness of the manuscript.

Participants:

A major flaw of the manuscript is that there is no statement informing the reader if the study was approved by an Institutional Review Board or if Informed Consent was obtained.

The participant sample needs to be described in greater detail to interpret the results, especially considering (1) the literature describing the impact of age, relative L1 and L2 language exposure, education background, context of L2 language use, etc. on L2 acquisition (2) the heterogeneity of the participant sample for these potentially predictive factors.

How was language proficiency determined for the participants? Were participants screened for hearing or vision impairment before participation?

The MINT scores for the participants should be presented by the groups outlined 265-266.

The MINT measure is a vocabulary measure. Using the MINT measure as an omnibus measure of language proficiency weakens the strength of study. Inclusion of measures of phonological and syntactic knowledge would strengthen the study, especially in light of the research questions focus on the intersection of morphological, syntactic, and phonological knowledge.

Materials:

The absence of reliability coefficients in the description of experimental tasks is problematic. There is also a lack of description of scoring procedures.

The experimental measure protocol needs to be explained in sufficient detail to replicate. How were stimuli presented? How were responses recorded? What feedback was provided?

Results:

Please refer to style manual for formatting of tables and figure

Abbreviations are used for analyses that were not introduced. Absent a rationale or description of analytic procedures, it is difficult to draw meaning from the results. A lack of criteria drawn from the literature to evaluate statistical findings (e.g., standard error, betas) is a drawback.   

Discussion:

The task used in the study to elicit language knowledge is text based, however the literature review does not review any relevant research on the difference between oral/aural language processing and text-based processing.  For example, in English the consistent orthographic representation of morphemes in text and inconsistent phonological representation in oral language impacts morpheme acquisition across the lifespan. Furthermore, Spanish and Hebrew have different orthographic systems, which may contribute to the patterns of results observed. 

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper presents an original contribution to the study of L2 gender agreement acquisition by comparing the contrasting Spanish and Hebrew systems. I would like to recommend that it be accepted after minor revisions.

 

Major edits:

 

Since the study’s participants are L1 speakers of English, it would be helpful to include in the literature review some information about the gender agreement system of English as it compares to Hebrew and Spanish, as well as a review of the literature on acquisition of gender in L2 Spanish. I recommend adding a new subsection or dividing up section 2.1 into Spanish and Hebrew.

 

I recommend that this paper be revised by a native English speaker before resubmission.

 

 

Minor edits:

            -lines 35-37: it seems that this study will compare Spanish and Hebrew but this is never explicitly stated. I suggest the author(s) provide a more clear explanation of the gender agreement system in each language and why they are good comparison languages. Currently, only Hebrew is described.

-line 38: which two linguistic systems?

-line 101:  “facilitated” should be “facilitative”

            -line 131: I would replace the word “wonders” with “investigates”

            -lines 134-136: It’s not clear here why this study compares Hebrew to Spanish.I would include an explanation or reminder here of the benefit of this language comparison

            -line 279: this is repetitive, the author(s) already mentioned that the L2 data was collected on the East Coast and this sentence can be removed.

            -lines 280-282: it is stated that the L2 speakers had been studying the language for “some” years. Can the author(s) be more specific? How many years, on average, precisely? Furthermore, can we get an idea of the participants’ overall abilities in the L2? If mean scores on the Mint were around 25/68 (if I understand correctly), were these beginners/intermediate students? Can the author(s) add what class level of Spanish and Hebrew the students were in?

            -Line 292: Materials section—please change all uses of first person “I” to “we” or a passive third person

            -Lines 321-336: Experiment 1 section – it is unclear in this description what the participants had to do. It is only explained that participants read sentences. Then what did they have to do? Please describe in as much detail as possible.

            -Line 354: please remove “For example” and instead add “See Figure 2” at the end of line 353

            -Line 365: In the Methods section, Experiment 1 referred to the FCT and here Experiment 1 seems to refer to the EPT. Please update for consistency. Which is experiment 1?

            -Line 378: the plot is blurry and very small. Please make this clearer. Also, the titles for each group are very unclear. One side seems to say “h” and the other “s”. Are these the languages? How can we tell which triangles and circles belong to the learners and the baseline comparisons? Can the groups be color coded? This plot needs updating.

            -Line 408: Again, the plot is difficult to read. I cannot tell which group is which, color coding would help, even on grayscale. In this plot the top labels are clearer, HB and SP are capitalized and seem to be Hebrew and Spanish. Can both plots show the full range from 0-1 on the lefthand scale?

            -Lines 424-426: Can the author(s) describe what these main effects mean?

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Review of Morphosyntactic features versus morphological features in L2 gender acquisition: a cross-language perspective

 

This study investigates whether determiners (morphosyntactic feature) or noun endings (morphophonological feature) are more reliable for acquiring gender in L2. In order to do so Spanish and Hebrew are chosen as target L2s (the L1 always being English which has not grammatical gender).

This is an interesting and relevant question, but it should be specified from early on in the paper how only non-transparent nouns will be used for this study, as Spanish has a high degree of transparency.

 

The main issue of the study is how it is presented: it lacks in clarity, especially in the introduction and background. I will list my concerns below:

 

1. it should be clear form the get go that plural nouns are analyzed in this study, as in lines 32-37 the Hebrew plural system is mentioned, but the reader might think that Hebrew plurals will be compared to Spanish singulars which is problematic as these data would not be comparable: plural nouns may be acquired at a later stage. Adding a sentence simply saying “testing PLURAL NOUNS ABD whether the presence…”
2. More importantly: the study compares these feature within PLURAL OPAQUE NOUNS, for Spanish the gender cue is the determiner, for Hebrew it is the plural ending. But if the plural ending is gender marked- is it really opaque? All of this should be thoroughly explained in the introduction so that the reader is able to follow what comes next.
3. The article would benefit from restructuring. The descriptions of the gender systems should precede the background sections regarding the features. This is because as the data from studies on L2 acquisition is presented, we are still unsure what the gender systems of these two languages is. The author also says that the study is about opaque nouns, but we need to know about the general transparency of the system, to understand the extent of the opaque nouns. For example, in line 145 the two systems are compared based on gender in plural, but nothing of these systems has been explained yet. This makes it hard to follow.
4. It should be mentioned how Spanish has a highly transparent gender system, even if there are opaque nouns within the system. But the regularity of the system may make the gender of the determiner more salient and thus acquired easily so that it can then be applied to the opaque nouns. Perhaps if the overall system was less transparent, the determiner may not be acquired with such ease and thus used as a cue for opaque nouns.
5. Karmioloff-Smith (1981) is a classic for gender acquisition and it analyses the role of determiners and noun endings. True, this is a reference for L1, but it could still be mentioned how attentiveness for these cues is different even for children.
6. Lines 64-66: very vague explanation and it is unclear why transparent nouns only have a semantic value, but if they are inanimate, they have both syntactic and semantic values.
7. Lines 96-99: missing citation for what is said here.
8. English has no grammatical gender, rather than a natural gender system as we do not see the gender on any English noun.
9. The author refers to a study conducted By Raviod and Schif 2012 and then draws comparisons with the current study. Based on the description of the study, I am however not sure that the participants were focusing on nominal suffixes, but rather the adjectives. I am thus not sure if the comparison holds. This needs to be either further elaborated or reformulated so the prediction follows from the previous study.
10. The explanation of the two target gender systems is very weak. As mentioned in point 3, it should come before the literature review section, but this section has also other issues. Firstly, the word used in example (1) is auto, but in Spanish the word coche can be used for car, and this word is opaque. So, choosing an example that is clearly transparent would be better. Also, the full gender system should be explained, thus also the feminine paradigm. Perhaps also in form of a table listing all the articles declined for gender and number, so it is clear to the reader that the determiners are transparent, even when the nouns are not. Crucially, the reader needs to see the Spanish plural agreement as the test items in the study are plural. Thus, the example provided in the article does not depict any of the factors the reader needs to further understand the study. Secondly, the adjectival paradigm needs ot be explained as the article paradigm (perhaps a table). Do mention that there are adjectives which do not show gender agreement, but ultimately these were not sued in the study, so they can be mentioned but disregarded after that.
11. The same needs to be done for the Hebrew gender system: explain the number of genders and provide examples of the full paradigm, so they can see how the nouns are transparent. The reader also needs to see the determiner and how it does not agree with the noun (thus I imagine the determiner will not vary based on whether the noun is M or F).
12. Regarding opaqueness: I couldn’t help but notice how the nouns in the two languages are differently opaque. In Spanish there is a category of nouns ending in -e or in a consonant which can be either M or F, but from what I understood from the text, in Hebrew there are some nouns that morphologically seem masculine but are feminine. Did I understand this correctly? This should be addressed together with potential implication this may imply for opaqueness.
13. In the current version in line 191 the author starts describing the plural system of Hebrew without ever having mentioned how this works for Spanish. The reader thus lacks the necessary tools for comparing the systems- this needs to be restructured.
14. To summarise the previous points: restructure the backgrounds section by explaining the gender systems first. Explain the gender systems more thoroughly so that the reader has an insight on the gender and number paradigms of the nouns and the elements relevant for this tasks: determiners and adjectives. A lot of this can be done with tables.
15. The bilinguals were not mentioned in the abstract: the abstract should be more representative of the study.
16. The L1 Spanish group are not early L2 learners like the Hebrew group, they are L2 learners that might have even started learning English as an L2 later than the target group started learning the target Spanish. I am not sure this group can give us any valid comparison in the current study. One option would be trying to focus on the participants that started learning English earlier, before the age of 10 and excluding the others from the analysis.
17. With regard to the items for the FCT, Spanish items contain more mass nouns or abstract nouns. Do you believe that this could have influenced the task. Did the results for these items differ from the rest (namely leches, muertes, carnes – incidentally all F). You should look at the individual results for these items and tun the stats with and without these in order to make sure that there are no other factors influencing the responses. This does not have to be part of the article, but if they are no differences it should be reported that this has been tested and that the results were not different, which si why you choose to include data form these items in the final analysis. But if there are differences, then these items should be excluded.
18. Methodological question: why were different adjective used? Were you not testing the effect of determiners/morphological markers? Why add complexity to the task with low frequency adjectives. How can you then be sure that a lower accuracy for low frequency nouns is due to the frequency of the noun? Please explain why this choice was made, as it is not methodologically ideal to do so.
19. It is not specified whether the two experiments were done one after the other. Was this randomized? Please specify.
20. Another issue of structure is why are results of experiment two presented before the results of experiment 1. If the author believes that the discussion is more coherent that way, then invert the order of experiments so that there is symmetry in how things are presented.
21. More space should be given to discussing the fact that the accuracy of non-frequent nouns in Hebrew is at floor level. Why does this happen? The same goes for the results of experiment 1. As far as I recall these nouns were all F. Is masculine considered the default gender for Hebrew? If this is the case, could it be the participants recert to the default due to the low frequency?
22. The statistical analyses need to be explained more clearly. For example: was the model ran separately for the two languages?
23. In the discussion: why is el agua an exceptional noun in Spanish?
24. In the discussion, the overall transparency of the gender systems should be addressed. It might be that it is not only due to the determiner, but since Spanish is transparent as a system, that even the gender of opaque nouns is easy to acquire. This should be part of the discussion.

 

Minor comments

1. Year of reference missing for Oliphant in line 76
2. Line 80: subject missing. I would also reformulate this sentence to “when they allow for an accurate inference…”
3. Line 115: Yonata Levy is a She.
4. Levy (1980) not present in the reference list, only the 1983 publication is listed
5. The paragraph in lines 248-253 is irrelevant
6. Refer to the participants as participants not as students.
7. Table 3: in column 3 there is a spelling mistake: DS instead of SD

Author Response

Dear reviewer:

Thank you for your very helpful feedback. I believe that I have addressed all of your questions and comments regarding the literature review, methodological issues, and statistical analysis. Please find my answers below

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

To concerns regarding the lack of reported pre-analyses procedures in the methods, the author replied (emphasis mine):

"Thank you for this observation. I am not sure what you mean by this observation in the methods section. However, I would like to draw your attention to a section called "supplementary material" under the conclusion. There is a link to find all the data and statistics scripts used in the study."

The link provided for supplementary data is broken, making it not possible to evaluate the pre-analytic and analytic procedures.

Consequently, the lack of reporting data cleaning procedures, missing data analyses, tests of data normality, and analytic procedures continues to be a major weakness of the manuscript.

Please find links below, for your convenience. 

Data cleaning and tests of normality: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3693611/

Missing data analyses: https://journals.sagepub.com/doi/pdf/10.1177/1536867X1301300407

Comments for author File: Comments.pdf

Author Response

Dear reviewer:

Thank you for your very helpful feedback. I believe that I have addressed all of your questions and comments regarding statistical analysis. Please find my answers in the attached document

Author Response File: Author Response.docx

Reviewer 4 Report

I would like to thank you the author for the clear answers that were provided in the commentary and for addressing my concerns in the manuscript. I enjoyed the commentary as it clarified some of my concerns, but ultimately not all the issues were explicitly addressed in the manuscript; I have read the manuscript with a lot of attentiveness, probably more than the average reader, so please do consider clarifying some of the issues I had  for the wider readership, as the average reader is likely to be less attentive than the reviewer and might thus face the same concerns. There is no need to include the thorough explanations that were provided for me in the commentary but do use them as a guide for clarifications.

 The manuscript has now improved considerably, and it is much easier to read. I do still have some concerns that should be addressed. I list these below.

 

1. The background section still needs some clarifications and overall coherence. Properties of Spanish and Hebrew are not presented symmetrically. Ideally you want them to contain the same type of tables: table for Spanish nouns/table for Hebrew nouns. Since there is a table for Hebrew nouns (which makes it clear what are the transparent assignment rules), there should be a table listing the Spanish noun types- focus on the ones ending in -a and -o and label them as transparent, and the ones ending in -e labelled as opaque. The opaque nouns should be explicitly and clearly listed as these are the test items in the study. As in the table for Hebrew nouns, add the examples for each category directly in the table.
2. In line 70: mano and diá do not have the -e suffix. The Spanish noun for street is missing.
3. You can express the adjectives with a table as well- choose an adjective and then have the form declined for gender and number as you have in the table for the articles. You do not need to focus on the adjectives that do not agree in gender as these were not part of the study. You may want to mention it in a footnote, but there is no need in having an example dedicated to these.
4. In table 2: Hebrew transparency, you could also add the plural version of these nouns, and the non-transparent nouns- as I have suggested for Spanish. This way the two tables will be as similar as possible, and the reader will be able to compare them at a glance. This will allow you to exclude what is now in example 2 and 3, as the information will be integrated in the table.
5. Just a note that the information which is now in example 3 is not clear as we do not know that these opaque nouns are assigned M/F- all the naïve reader sees is that the suffixes match as the noun with no overt suffix takes the same plural morphemes as other nouns with no overt suffix, but the noun in 3a happens to be feminine. This means that we need some overt agreement to see how this noun is indeed feminine. This is not crucially relevant, as my advice is to not have these examples and have the information about transparent and opaque nouns expressed in a table.
6. Again, place the declension pattern of adjectives in a table, but you can still have example 4 in order to demonstrate how these nouns are opaque as they are declined with the feminine form of the adjective.
7. The information about the definite determiner should come before, I am aware that it is not relevant for gender in Hebrew, but it should be discussed at the same (sequential) point as it is discussed in Spanish.
8. In line 206 it is not clear if L1 or L2 acquisition of gender is being discussed.
9. Predictions in line 269-273: your predictions are two sides of the same coin. You should set it as: “if the determiner plays a bigger role: A, if the pl morpheme then B. Based on the previous literature we predict A.”
10. In line 302 it is unclear to which groups the word “participant refers to”. I read it as the L2 learners, but since participants is the terms used also in table 3 to describe both targets and controls, it is confusing.
11. There are some issues with the numbers of tables and their reference in the text. Please make sure to re-number your tables- after you add the new tables.
12. Lines 397-401: repetition. Consider deleting.
13. I don’t see that the performance at floor level by Hebrew L2ers is discussed. This is an interesting result and explanations or speculations of why it occurs should be addressed in the discussion. In your commentary you say that this is discussed in lines 703-709, but in the manuscript file there are already references on these lines, so I could not find how this was included in the discussion. I believe that the floor level for the non-frequent opaque nouns in Hebrew is due to the fact that these nouns look morphologically like transparent nouns, and thus need to be rote learned, which takes more time for non-frequent nouns. The strategy may be different for Spanish as there is a class of opaque nouns which can be either M/F and it is thus more salient to the learner that assignment is more difficult than for the transparent nouns.
14. Related to my previous point, I would like to address your comment to point 11 from my previous revision. My response to your comment is: The acquisition strategy might be different: with the Spanish -e class learners have a 50% chance to guess the gender, whereas for Hebrew it is a matter of learning the exceptions (words that seem masculine but are not). Could this potentially explain the ´high error rate in Hebrew? Also, you mention that there are some nouns in Spanish that seem masculine but are not (which would make them more comparable to your test items et in Hebrew) but these were not tested in Spanish, so we cannot really know. Thus, the discussion on the employed strategies must be discussed based on the -e nouns.

Author Response

Dear reviewer: thanks a lot for all your comments and suggestions. You can find my answers in the attached document.

Author Response File: Author Response.pdf