Semantic Relation Priming Is Not Constituent-Specific—Evidence from Electrophysiology

Round 1

Reviewer 1 Report (New Reviewer)

This paper presents an interesting proposal a priming paradigm to investigate the relation priming of three morphemes level. However, there is room for improvement before the manuscript can be accepted for publication. I recommend that the authors carefully revise the manuscript based on the following feedback:

 

1. Add a table of used symbols in the paper to improve readability.

2.The introduction should be revised to clearly present the main ideas and motivations behind the proposed research. Please ensure that the research question and motivation of the proposed study are clearly stated. It is important to cover the research gap adequately.

3.- A section under the title "discussion" should be added to the article in which the authors examine the presented method from different perspectives with other methods.

4.Provide more information about how the experiments were conducted, including the tools and software used, in the experiment section.

5- The figure 2 is ambiguous, and the authors should explain each one and its impact on output production.

6.  the author miss Pseudocode, please add it.

7. the authors should analyze how to set the parameters of the proposed methods in the framework.  Do they have the “optimal” choice?

 

Author Response

Response to reviewer1

 

This paper presents an interesting proposal a priming paradigm to investigate the relation priming of three morphemes level. However, there is room for improvement before the manuscript can be accepted for publication. I recommend that the authors carefully revise the manuscript based on the following feedback:

 

1. Add a table of used symbols in the paper to improve readability.

Response: Thanks the reviewer for the comments. Please refer to the footnotes and Table 1 for a comprehensive list of symbols and their corresponding explanations.

 

2.The introduction should be revised to clearly present the main ideas and motivations behind the proposed research. Please ensure that the research question and motivation of the proposed study are clearly stated. It is important to cover the research gap adequately.

Response: The first three paragraphs of the Introduction have been rewritten to revise the above issues, as detailed in the manuscript.

3.A section under the title "discussion" should be added to the article in which the authors examine the presented method from different perspectives with other methods.

Response: The third paragraph of the Discussion section has been added to address the aforementioned concerns, as outlined in the manuscript.

 

 

4.Provide more information about how the experiments were conducted, including the tools and software used, in the experiment section.

Response: The detailed description of the experimental procedures, including the software and equipment used, has been provided in the Method section. Specifically, behavioral data was recorded using E-prime software, EEG data was captured using Neuroscan equipment, and statistical analysis was conducted using R software. Please refer to the manuscript for further details.

 

 

5. The figure 2 is ambiguous, and the authors should explain each one and its impact on output production.

Response: Figure 2 depicts the average ERP responses on representative electrodes, highlighting the six conditions (MSRS, MSRD, HSRS, HSRD, NURS, and NURD) using different colors. As our study primarily investigates the semantic relation priming effect, which primarily manifests in the N400 time window (300-450ms), our analysis focused on examining the EEG responses elicited by these six conditions within the N400 window. For a detailed statistical analysis, please refer to the "ERP results" section.

 

 

6. the author miss Pseudocode, please add it.

Response: Please forgive me for not knowing how to write Pseudocode. Psycholinguistics doesn't seem to have a tradition of writing Pseudocode, and we attach the R script I used for statistical analysis of data, see "script" for details.

 

7. the authors should analyze how to set the parameters of the proposed methods in the framework. Do they have the “optimal” choice?

Response: The parameters used in our study are based on previous literature, specifically the work of Gagné and Shoben (2009) and Jia, Wang, Zhang, and Zhang (2013). These studies employed these parameters to successfully elicit a typical relation priming effect. Therefore, we adopted the same parameters in our experiment. The corresponding references have been added to the relevant sections of the manuscript.

 

Reviewer 2 Report (New Reviewer)

The paper present an interesting topic of psycholinguistic research into compounds and semantic priming in Chinese. It is interesting as an interdisciplinary study.

In my opinion, the main weakness of the paper is the limited theoretical presentation of key concepts and points. For example, it needs to be theoretically stated what 'a relation priming phenomenon' is, possibly with examples in language context. Moreover, some more linguistic details need to be provided for the morphemes in Chinese and the particular morphemes used in the analysis, what is their status, their productivity in the language, etc. For people unfamiliar with Chinese (I assume standard Mandarin? - this also needs to be stated), these remain unclear.

The authors seem to mix two different levels of analysis of compounds - form (lexical / grammatical) and meaning (semantic level), e.g. in the statement "the meaning of compound words is not simply the addition of the meanings of two sub-concepts". First, in terms of form, compound words have components (not sub-concepts). Second, even in a free phrase the meaning is rarely a sum of the components' meanings. But since within compounds components are not independent lexical units, so they don't have independent conceptual meaning (so they are not sub-concepts meaningwise either).

This statement is not entirely correct: a focus of controversy about compounds is whether it is represented as the whole word or decomposed into constituents in a mental dictionary. This is not controversial and besides, not all compounds are decomposable, as you stated, so it is not always possible to use decomposability.

There are some repetitive statements which need to be removed, e.g.

1. Raffray et al. (2007) found that relation priming can occur without morpheme repetition, but morpheme repetition can boost relation priming (Raffray, Pickering, & Branigan, 2007). Therefore, we speculate that relation priming can occur without morpheme repetition.

2. ...it can be combined more freely. Hence Chinese should have greater freedom of combination...

This is a generally accepted theoretical statement, so it is not correct to say 'some researchers believe': Some researchers believe that for the modifier + head structure, the head represents the conceptual category to which the entire word belongs, while the modifier only modifies details, so the head is more critical for whole-word processing.

The experiments need more precise presentation. For example:

1. The statement 'Using a 5-point scale (1 means not in the same category, 5 means in the same category)' implies a binary categorisation task which conflicts with the introduction of a 5-point scale.

2. In 'the prime was interpretable, but the target was nonsense' define 'nonsense', or better give an example of a 'good' or 'nonsense' pair.

3. This is unclear: 15 people from the same subject population but did not participate in this experiment rated the prime and target's semantic relatedness. By rating they in fact participate in the experiment.

4. What is a 'filler pair'?

5. It would be helpful to show a figure of what is seen on screen. The way of presentation of the information may be significant for evoking certain results.

Also there is research showing that semantic relations of words in context might be different than those evoked by isolated words. The paper says that 'Later, another experiment found that syntactic priming exists independently of the repetition of the concepts (such as verbs) expressed in the sentence'. It remains unclear whether the same experiments was performed by authors and what the results have shown.

Some reported conclusions are unclear, e.g. Above, we have proved that relations are represented independently.

The statement in the introduction contradicts that in the discussion:

This priming phenomenon is bound to morpheme repetition, that is, relation priming will only occur when there is morpheme repetition between the priming word and the target word.

and

Initially, we thought that relation priming would only occur in the case of morpheme repetition, but now we have found that it could exist independently of morpheme repetition.

The paper needs to present consistent findings, so if the introduction presents only 'initial' beliefs which are later found to be unsubstanciated, this needs to be clearly stated. 

 

Minor:

most concerned problems > most concerning / critical / key problems

Moreover, Chinese only has morpheme boundaries but no word boundaries, so that it can be combined more freely.

> it? 

with a distance of 0.7 m > at a distance of 0.7 m

Some formatting (spaces before opening brackets, etc.) is needed.

The paper will benefit from a thorough proofreading.

Author Response

Response to reviewer2

 

The paper present an interesting topic of psycholinguistic research into compounds and semantic priming in Chinese. It is interesting as an interdisciplinary study.

1. In my opinion, the main weakness of the paper is the limited theoretical presentation of key concepts and points. For example, it needs to be theoretically stated what 'a relation priming phenomenon' is, possibly with examples in language context.

Moreover, some more linguistic details need to be provided for the morphemes in Chinese and the particular morphemes used in the analysis, what is their status, their productivity in the language, etc. For people unfamiliar with Chinese (I assume standard Mandarin? - this also needs to be stated), these remain unclear.

Response: Thanks the reviewer for the comments. The relevant issues have been revised in the manuscript and are expounded upon in the introductory section of the paper, specifically within the initial three paragraphs of the Introduction.

 

2. The authors seem to mix two different levels of analysis of compounds - form (lexical / grammatical) and meaning (semantic level), e.g. in the statement "the meaning of compound words is not simply the addition of the meanings of two sub-concepts". First, in terms of form, compound words have components (not sub-concepts). Second, even in a free phrase the meaning is rarely a sum of the components' meanings. But since within compounds components are not independent lexical units, so they don't have independent conceptual meaning (so they are not sub-concepts meaningwise either).

Response: Relations of multi-morpheme words can be divided into grammatical and semantic relations (Cui et al., 2018; Ji & Gagné, 2007). At the grammatical level, unlike the compound words in Indo-European languages, modifier-noun is the dominant structure. There are five grammatical structures in Chinese. For example, subordinate (e.g.,野鸡, meaning pheasant), which corresponding to modifier-noun structure in English; coordinative (e.g.,灯火,meaning lights); verb-object (e.g.,开学,meaning school opens); supplement (e.g.,延长,meaning extended); subject-predicate (e.g.,地震, meaning earthquake). And in Chinese, we find that both grammatical and semantic relations play an important role in compound word processing (Cui et al., 2018; Ji & Gagné, 2007; Jia, Wang, Zhang, & Zhang, 2013). What is discussed in this article is the relation priming at the semantic level. Our experimental framework draws upon the Competition Among Relations in Nominals Theory, which posits that the meaning of a compound word transcends the mere summation of its sub-conceptual meanings. Instead, it emerges through a series of intricate computations guided by the underlying sub-conceptual meaning. Notably, Gagne et al. have conducted extensive research in this domain, as evidenced by their significant contributions (see Gagné & Shoben, 1997; Ji & Gagné, 2007;Gagné & Spalding, 2009 for further details).

 

3. This statement is not entirely correct: a focus of controversy about compounds is whether it is represented as the whole word or decomposed into constituents in a mental dictionary. This is not controversial and besides, not all compounds are decomposable, as you stated, so it is not always possible to use decomposability.

Response: This issue is not as uncontroversial as initially perceived, as it challenges the conventional understanding that high-frequency words have direct access to semantics, while low-frequency words require decomposition access. Additionally, it challenges the notion that opaque words have direct access, while transparent words necessitate decomposition access. Contrary to these beliefs, Gagné et al.'s experiments indicate otherwise. They propose that even for familiar and lexicalized compound words, their meaning is unlikely retrieved directly as a whole. Instead, there appears to be a process of semantic composition, wherein the meaning of the entire word is constructed by utilizing relation information to connect the meanings of its constituent parts. For further insights, please refer to Gagné and Spalding (2004).

 

4. There are some repetitive statements which need to be removed, e.g.

1). Raffray et al. (2007) found that relation priming can occur without morpheme repetition, but morpheme repetition can boost relation priming (Raffray, Pickering, & Branigan, 2007). Therefore, we speculate that relation priming can occur without morpheme repetition.

2). ...it can be combined more freely. Hence Chinese should have greater freedom of combination...

This is a generally accepted theoretical statement, so it is not correct to say 'some researchers believe': Some researchers believe that for the modifier + head structure, the head represents the conceptual category to which the entire word belongs, while the modifier only modifies details, so the head is more critical for whole-word processing.

Response: Revisions have been made to the corresponding positions in the manuscript.

 

5. The experiments need more precise presentation. For example:

1). The statement 'Using a 5-point scale (1 means not in the same category, 5 means in the same category)' implies a binary categorisation task which conflicts with the introduction of a 5-point scale.

Response: The text has been revised to‘Using a 5-point scale (1 means completely unrelated, 2 means unrelated, 3 means unknown, 4 means correlated, and 5  means highly correlated)’, please refer to the manuscript for details.

2). In 'the prime was interpretable, but the target was nonsense' define 'nonsense', or better give an example of a 'good' or 'nonsense' pair.

Response: 'Nonsense' is a combination of the first and last characters of a true word, such as '技究', it does not constitute a true word, has no meaning, and is not homophonic to any true word, but its physical properties completely match the true word. For example, "雪人(means snowman)" is 'good', while "技究" is 'nonsense'.

3). This is unclear: 15 people from the same subject population but did not participate in this experiment rated the prime and target's semantic relatedness. By rating they in fact participate in the experiment.

Response: Revisions have been made to the corresponding positions in the manuscript.

 

6. What is a 'filler pair'?

Response: The filling materials in psychological experiments refer to the auxiliary or supportive materials added to the experiment in order to complete the conditions set in this experiment. The filling materials in this study were mainly used to balance the reactions of the participants' left and right hands.

 

7.It would be helpful to show a figure of what is seen on screen. The way of presentation of the information may be significant for evoking certain results.

Response: The parameters used in our study are based on previous literature, specifically the work of Gagné and Shoben (2009) and Jia, Wang, Zhang, and Zhang (2013). These studies employed these parameters to successfully elicit a typical relation priming effect. Therefore, we adopted the same parameters in our experiment.

 

8.Also there is research showing that semantic relations of words in context might be different than those evoked by isolated words. The paper says that 'Later, another experiment found that syntactic priming exists independently of the repetition of the concepts (such as verbs) expressed in the sentence'. It remains unclear whether the same experiments was performed by authors and what the results have shown.

Response: This paragraph in the manuscript aims to illustrate the potential equivalence between syntactic priming and relation priming, highlighting that both phenomena are essentially forms of structural priming. However, they operate at different levels, with syntactic priming occurring at the sentence level and relation priming taking place at the lexical level. Both types of priming can be understood as manifestations of the underlying language composition rules that govern language processing. The discussion on structure priming at the sentence level refers to the research conducted by Martin Pickering and his team, while the author's own experiments did not specifically investigate this aspect.

 

9.Some reported conclusions are unclear, e.g. Above, we have proved that relations are represented independently.

Response: This passage serves as a bridge between the past and the future in the manuscript. Based on our experimental results, we observed a significant decrease in amplitude in the relation same (RS) condition compared to the relation different (RD) condition at the levels of non-repeated (NU), modifier-shared (MS), and head-shared (HS). This finding provides clear evidence of a relation priming effect. The observation of relation priming at the morpheme-repeated levels (MS and HS) aligns with the findings of Ji and Gagné (2007) and Jia et al. (2013). However, our new contribution lies in the discovery that relation priming can also occur independently at the non-repeated level (NU) without the presence of shared morphemes. This result indicates that relation can be represented independently and are not necessarily tied to specific morphemes.

 

10. The statement in the introduction contradicts that in the discussion:

This priming phenomenon is bound to morpheme repetition, that is, relation priming will only occur when there is morpheme repetition between the priming word and the target word.

and

Initially, we thought that relation priming would only occur in the case of morpheme repetition, but now we have found that it could exist independently of morpheme repetition. The paper needs to present consistent findings, so if the introduction presents only 'initial' beliefs which are later found to be unsubstanciated, this needs to be clearly stated.

Response: In 1997, Gagné et al. initially observed that relation priming occurs only when there is morpheme repetition between the priming word and the target word. However, subsequent research yielded contrasting findings, suggesting that relational priming is essentially the activation of a specific pattern and can manifest irrespective of the presence or absence of morpheme repetition or semantic correlation. As a result, there is currently no consistent consensus on whether relation priming can occur independently of morpheme repetition. The aim of our study is to address this unresolved question by employing Chinese materials. Our findings demonstrate that relational priming can indeed occur independently of morpheme repetition. This logical progression is consistent throughout our manuscript, establishing a coherent link between the Introduction and the Discussion sections.

 

 

 

 

 

Minor:

 

most concerned problems > most concerning / critical / key problems

 

Moreover, Chinese only has morpheme boundaries but no word boundaries, so that it can be combined more freely. > it?

 

with a distance of 0.7 m > at a distance of 0.7 m

 

Some formatting (spaces before opening brackets, etc.) is needed.

 

The paper will benefit from a thorough proofreading.

Response: Thanks the reviewer for the comments. The above issues have been revised in the manuscript.

Reviewer 3 Report (New Reviewer)

The present work seems to have enough interest and many possibilities, according to the writing team and the existing literature. We quickly noticed that the text has many changes according to the reviewers' comments, and it seems that the writing team is interested in presenting the best result following the journal's standards.

Below I note some comments that you may find helpful to improve your work. Some acronyms in the abstract need to be immediately broken down and defined so that someone reading the text can better understand and follow it. The text of the abstract seems longer than usual, and the structure (introduction, methodology, results, conclusions) is not clear as is typical when writing an article.

The bibliographic support of the introduction, especially in the first two pages, is fragile and does not cover as much as it should the existing knowledge about the subject. It has points presented as given knowledge without bibliographic reference in the text, e.g., "the traditional view is that compounds are processed as a whole." Someone reading the article will not have a correct picture descriptively of the theoretical background on the topic and will not have a natural cascading flow in reading it and connecting the data to the research gap.

In many places, the bibliography is older, e.g., "in recent years researchers... on page 4. but the references are over ten years older, especially in 1997. On page 5 in "Gagne et al.....well," the text is more like a part of the discussion as written as it presents some theoretical views and comparatively opposite and similar associations with these views.

In other cases, the writing team states, e.g., that the researchers cite and have a single citation in the entire repeated paragraph. The issue under investigation has literature in the Chinese language that concerns the use of ERPs and semantic network analysis. Still, not all of them exist within the text, while the team states that their study is the first to study the semantic web in parallel ERP use. Due to the topic's specificity and the language limitation that the authors mentioned below, it would be good to stick to bibliographies and comparisons related to the Chinese language.

In the methodology, it is seen that we are talking about a small sample to experiment with and have clear correlations, which is later seen in the group results. Some decisions have been adopted regarding the study process, which, however, are not supported in the literature as following a documented protocol or some common good practice in the research of the specific fields, for example, why 72 linguistic stimuli (how have they been selected, with what criteria?), 3 *2 within-subject design, 0.7 m distance from the screen)

On page 9, the team mentions part of the piloting of the process and provides results of that process which should not be at this point. It would be better suited to refer to piloting and weighing the process.

On page 9, it again has acronyms to be defined by the writing team, e.g., MSRS, MSRD, etc.

In addition, I would like to express concern about the sample that may affect the subsequent process. Why hasn't the selection of participants been assessed on their vocabulary so that if a participant has a lack of language, they may also have semantic difficulties, or directly the team could determine the participants with a weighted semantic test, even with a Ravens type cognitive test, to rule out semantic, linguistic, or even cognitive weaknesses? The data could also be helpful in later correlations.

 

The group cites their previous study, but there is no literature citation. Verbs like speculate should be avoided as they usually have harmful content. The team also states that the language materials used in the study are transparent words with relatively low frequency, perhaps this creates serious issues with the study results, and the research team should address the case during the design or piloting of the survey.

More attention to the expression in places where the first plural is used and in places where a question is used to express the research questions.

Author Response

Response to reviewer3

 

The present work seems to have enough interest and many possibilities, according to the writing team and the existing literature. We quickly noticed that the text has many changes according to the reviewers' comments, and it seems that the writing team is interested in presenting the best result following the journal's standards.

 

1. Below I note some comments that you may find helpful to improve your work. Some acronyms in the abstract need to be immediately broken down and defined so that someone reading the text can better understand and follow it.

The text of the abstract seems longer than usual, and the structure (introduction, methodology, results, conclusions) is not clear as is typical when writing an article.

Response: Thanks the reviewer for the comments. The Abstract section has been revised to reduce the word count and symbolic callouts have been added. Additionally, structural notes have been included in the corresponding sections. For detailed information, please refer to the manuscript itself.

 

2. The bibliographic support of the introduction, especially in the first two pages, is fragile and does not cover as much as it should the existing knowledge about the subject.

It has points presented as given knowledge without bibliographic reference in the text, e.g., "the traditional view is that compounds are processed as a whole." Someone reading the article will not have a correct picture descriptively of the theoretical background on the topic and will not have a natural cascading flow in reading it and connecting the data to the research gap.

Response: We have rewritten the first three paragraphs of the Introduction and the corresponding questions have been revised, as detailed in the manuscript.

 

3. In many places, the bibliography is older, e.g., "in recent years researchers... on page 4. but the references are over ten years older, especially in 1997. On page 5 in "Gagne et al.....well," the text is more like a part of the discussion as written as it presents some theoretical views and comparatively opposite and similar associations with these views.

Response: The conditions for the occurrence of the relation priming effect have been a topic of debate among researchers since its initial proposal by Gagne et al. in 1997. From approximately 2001 to 2007, several studies (Estes, 2003; Estes & Jones, 2006; Raffray et al., 2007; Spellman, Holyoak, & Morrison, 2001) questioned whether morpheme repetition between the priming word and the target word is necessary for the manifestation of relation priming, but no consensus has been reached. In the context of Chinese language research, little attention has been given to this phenomenon, with the first study conducted by Jia et al. in 2013. Since then, there have been no recent updates on this issue, which justifies the inclusion of older literature. Page 4 of the manuscript has been revised accordingly to address these points.

Page 5 of the manuscript simply describes the lack of consensus among researchers regarding the conditions for the emergence of the relation priming effect, specifically whether morpheme repetition between the priming word and the target word is required. This lack of consensus motivates our study to further investigate this phenomenon.

 

 

 

4. In other cases, the writing team states, e.g., that the researchers cite and have a single citation in the entire repeated paragraph.

The issue under investigation has literature in the Chinese language that concerns the use of ERPs and semantic network analysis. Still, not all of them exist within the text, while the team states that their study is the first to study the semantic web in parallel ERP use. Due to the topic's specificity and the language limitation that the authors mentioned below, it would be good to stick to bibliographies and comparisons related to the Chinese language.

Response: Because the main purpose of this study is to clarify the conditions under which relation priming occurs, that is, whether relation priming only occurs when there are repeated morphemes. However, the relational effect is much weaker than the semantic effect; Without sharing morphemes, it is difficult to extract independent semantic relation themselves. And our experiment used electrophysiological technology (ERP), which is more sensitive to capturing cognitive processes than traditional reaction time technology and easier to detect independent relational effects. However, this issue has received relatively little attention in Chinese, and our team is the first to attempt this, so we are unable to find more Chinese literature for comparison.

 

5. In the methodology, it is seen that we are talking about a small sample to experiment with and have clear correlations, which is later seen in the group results. Some decisions have been adopted regarding the study process, which, however, are not supported in the literature as following a documented protocol or some common good practice in the research of the specific fields, for example, why 72 linguistic stimuli (how have they been selected, with what criteria?), 3 *2 within-subject design, 0.7 m distance from the screen)

Response: The parameters used in our study are based on previous literature, specifically the work of Gagné and Shoben (2009) and Jia, Wang, Zhang, and Zhang (2013). These studies employed these parameters to successfully elicit a typical relation priming effect. Therefore, we adopted the same parameters in our experiment. The corresponding references have been added to the relevant sections of the manuscript.

 

6. On page 9, the team mentions part of the piloting of the process and provides results of that process which should not be at this point. It would be better suited to refer to piloting and weighing the process.

Response: Apologies for the confusion. The content on page 9 is not a pilot study but rather a description of the experimental materials, the formal experimental process, and the EEG recording parameters. It provides detailed information about the materials used in the experiment, the procedures followed during the experiment, and the specific parameters employed for EEG data recording.

 

7. On page 9, it again has acronyms to be defined by the writing team, e.g., MSRS, MSRD, etc.

Response: Please refer to Table 1 for the meanings of these abbreviations. Table 1 provides a comprehensive overview of the abbreviations used in the manuscript along with their corresponding explanations and definitions.

 

8. In addition, I would like to express concern about the sample that may affect the subsequent process. Why hasn't the selection of participants been assessed on their vocabulary so that if a participant has a lack of language, they may also have semantic difficulties, or directly the team could determine the participants with a weighted semantic test, even with a Ravens type cognitive test, to rule out semantic, linguistic, or even cognitive weaknesses? The data could also be helpful in later correlations.

Response: The selected participants for this study were Chinese college students, all of whom were native Chinese speakers. These students were chosen based on their high academic achievements, as they had successfully passed the college entrance examination in China. As such, they possess strong language skills and cognitive abilities, and they have already acquired knowledge of compound words during their high school education. Therefore, the concerns raised by the reviewers regarding language ability and cognition do not apply to our participant sample.

 

9. The group cites their previous study, but there is no literature citation.

b.Verbs like speculate should be avoided as they usually have harmful content.

c.The team also states that the language materials used in the study are transparent words with relatively low frequency, perhaps this creates serious issues with the study results, and the research team should address the case during the design or piloting of the survey.

Response:

a.References have been added to relevant sections.

b.The word (speculate) has been replaced in all relevant places.

c.Gagné et al. propose that the semantic access of compound words, including familiar high-frequency words, may involve a calculation process rather than direct extraction. However, it is evident that low-frequency transparent words undergo such a calculation process. In our study, we selected low-frequency transparent words as materials to ensure typicity and to enhance the visibility of the relation priming effect. Nonetheless, due to the constraints of material selection, our conclusions may only be generalized to low-frequency transparent compound words.

 

 

10. More attention to the expression in places where the first plural is used and in places where a question is used to express the research questions.

Response: Thanks the reviewer for the comments. We have paid attention to these issues and made appropriate modifications in the manuscript.

 

Round 2

Reviewer 1 Report (New Reviewer)

I have gone through the revised paper. All my concerns and requests have been carefully addressed by authors.

Author Response

Thank you for your help.

Reviewer 2 Report (New Reviewer)

Significant improvements have been made on the content of the introduction, the methodology presentation and the discussion of results, which outlines some theoretical aspects and justifies the approach used in the study and the significance of the results produced.

Table 1 is missing. 

The paper still would benefit from a more thorough analysis of results and clearer conclusions.  

Minor proofreading is still necessary and formatting, e.g. spacings before/after brackets, etc.

Author Response

Revise accordingly. Thank you for your help.

Reviewer 3 Report (New Reviewer)

I am happy to note that my suggestions have been taken into account, and the text has been improved significantly, mainly in the documentation and a better flow, which leads the reader more naturally. I see the methodology, the choice of tools, and the process of documenting and connecting the study with existing knowledge, and I assess it as having significant progress.

Author Response

Thank you for your help.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This paper has serious problems in English grammar and writing. Please see the attachment. I suggest that the authors find a well-experienced native speaker of English to edit it. 

Comments for author File: Comments.pdf

Author Response

Thank the reviewers for the comments. Please see the attachment for our specific reply.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper is very interesting and well written. The argument is clear and the scientific result based on excellent experiments.

I would suggest to accept the paper as it is because it shows important results about relation priming and its role in word  and morpheme processing. The research is well presented and also the methodology is solid enough to consider the scientific importance of this work.

Author Response

Thank the reviewers for the encouraging comments, which gave us great confidence.

Reviewer 3 Report

This study investigates priming effects on Chinese compounds. Thirty-two Chinese speakers completed a behavioral self-paced priming paradigm and ERP task.

While this is an interesting study that could show the process of the construction of meaning in a language that is not orthographic and has no word boundaries, the manuscript needs significant changes. 

Here I list my comments, per section.

Abstract

The abstract does not have enough information about the study, and it’s not well organized. This section should be improved

 

Intro

The authors suggest that the mental dictionary has a limited capacity, and having a high vocabulary will be a memory burden. While it seems reasonable to believe that storing compound words in a decomposed way is a more efficient way, it does not mean that having individual entries for the compound words represents a burden. Who says this?

 

In the sentence “Other creatures can only use a limited single vocabulary to communicate information and express ideas”, What do the authors mean by “other creatures”?

The authors define relation priming in the abstract but do not provide a definition for RELATION within the text before using this term. Relation priming should also be defined within the introduction, have it in the abstract is not enough.

Provide examples for the different relations, grammatical and semantic.

There is no clear reason to start talking about Chinese compared to English relations as the authors have not mentioned the population and why this comparison.

 

As the authors did not provide an example or clear differentiation between grammatical and semantic relations, is difficult to see why student accusation (accusation BY a student) is a semantic relation.

What do you say that the semantic classification used in Gagne et al (2005) was vague? Provide more detail about this if you are going to build an argument on this issue.

What are the characteristics you are referring to in the last paragraph of page 5?

Provide examples of morpheme repetitions.

Define meaning-spelling.

Is eight thousand much fewer semantic bases than other languages?

What consistency of the rate between subjects reached above 95%?

 

It is stated that the experiments use morpheme repetition (Ms or HS) and non-repetition (NU). The not-repetition is also a condition. Therefore,  the sentence we expect that relation priming will appear under BOTH conditions is not clear. This paragraph needs to be reworded.

 

The authors should explain ERP , N400, N200, at least briefly. A novice reader may not know what you mean.

 

Methods

It’s not until this point that the number of conditions is clearly stated. This is ok as this is the method section. However, the authors have constantly referred to the different conditions in this study through the introduction in a confusing way.

 

The information and conditions in this section could be in a table that would allow for more details about the relations. For example, MSRS which part of this word primes/matches wristwatch?

 

The sentence “15 people from the same subject population but not participate in this experiment rated the prime and target's semantic relatedness” is not clear.

 

Are the conditions named MSRS and MSRD or RS and RD?

What is the rationale for so many fillers?

 

What value was assigned for each key> was F sense and J non-sense?

How long was the task?

How many target items per condition were in each part/list?            

Did participants need to press the letter “Q” before each trial in the task?

 

Were the participants tested with EEG (which needs to be spelled out) while completing the judgment task? This is not stated, and the reader must infer this.

 

The authors have not stated how they record the response time. Again, the reader must infer this information

 

 

 

Results & Discussion

 

Given the different confusing aspects of the tasks is difficult to evaluate the accuracy of the results. Apparently, the analysis seems correct, but I don’t feel confident in interpreting the results.

Author Response

Thank the reviewers for the comments. Please see the attachment for our specific reply.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Please see the attached.

Comments for author File: Comments.docx