Sunday, July 26, 2020

How to Review a Syntax Paper


The objective of a review for a linguistics journal is to evaluate whether a submission advances the scientific study of human language. If the submission makes such an advance, it should be accepted. If it does not, it should be rejected.


You should write in a professional manner. Since reviews are anonymous, reviewers sometimes speak in a harsh or rude tone that they would not normally use in academic discourse. Please do not engage in this kind of discussion in your reviews. It reflects badly on you, the journal and the field. Also, is not really needed since your objective is to accept or reject the paper, not to harass the author.


If you are unable to review the paper in an objective manner, for whatever reason, recuse yourself. Here are some possible reasons that you might recuse yourself (among others):

a. The paper discusses your analyses negatively.
b. The paper was written by your close colleague with whom you collaborate frequently.
c. The paper was written by your student.
d. The paper was written by your thesis supervisor.
e. The paper is written in a framework different from the one you adopt.


The structure of the review should be as follows:

1. Decision (one line)
2. Summary (one paragraph)
3. Justification (one page)
4. Comments (one or two pages)
5. References (less than half a page)


I checked the length of 28 previous reviews of syntax papers that I submitted. The average length of the reviews was 3.2 pages, and the median was 2 pages. The average was skewed a little by one review that was 11.5 pages long. Based on this, I recommend that reviews be 2-3 pages long.

In other words, reviews should be relatively short, and to the point. A review of a syntax paper is not itself a syntax paper. There is no need to explore issues in depth (that is the author’s job), nor to teach the author how to do syntax, nor to comment on every single point brought up in the paper. You only need to justify your decision in a clear way. Of course, in some circumstances your justification may take more than a page.

If you request that the author make revisions, you are obliged to say what those revisions are, and that might take up an extra page or two. Those requested revisions can be listed in the comment section.


1. Decision
In all cases, you should clearly state whether the paper should be rejected or accepted. The editor should not be forced to read between the lines. Here are some sample decisions:

a. The paper should be accepted as is.
b. The paper should be accepted with minor revision.
c. The paper should be accepted with major revisions.
d. The paper should be rejected, but it may be revised, and resubmitted.
e. The paper should be rejected.

The above list is not meant to be comprehensive. There are other possible decisions (e.g., “Do not publish”). But any decision you give should be short (less than a line) and clear.

2. Summary
Start the review with a short paragraph summarizing the main results of the paper. The summary helps orient the reader of the review (the editor and the author).

3. Justification
Your review should justify your decision. In some cases, even a paragraph may be sufficient.

If you reject the paper, you should explain why are you are rejecting it. A rejection means that you do not think the paper makes an advance in the scientific understanding of human language. So, if you reject the paper, you need to explain why it fails in this regard. If you accept the paper, you need to explain that as well.

Here is a list of some reasons to reject a paper (among others).

a. The argumentation is not sound.
(In other words, the conclusions do not logically follow from the premises.)
b. The author misunderstands or misapplies a particular theory or framework.
c. The empirical generalization proposed is incorrect.
d. The syntactic principle proposed makes incorrect predictions.
e. The paper suffers from methodological problems (e.g., in gathering the data).
f. The paper involves plagiarism.

Here is a list of some reasons to accept a paper (among others).

a. The author discovers a new empirical generalization and accounts for it.
b. The author provides new empirical support for an existing principle/analysis.
c. The author proposes a new syntactic principle/analysis, and justifies it empirically.
d. The author shows convincingly that an existing theoretical proposal is incorrect.

In some cases, you may find reasons both to accept and reject the paper. This is a good candidate for “Revise and Resubmit”. Since the paper makes an advance in the scientific understanding of human language, you would like to accept it, but only on the condition that the author fixes or deletes the part of the paper that causes the problems.

If you find that the paper makes a clear advance in the scientific understanding of language, but has a number of smaller problems or has organizational issues that might prevent a reader from understanding it, then the paper is a good candidate for the category “accept with major/minor revisions”. It that case you should make it clear exactly what revisions are required.

The comment section is variously called “specific comments” or “line-by-line comments”.

Any non-required suggestions for the author can be listed in the comment section. You are not obligated to make such suggestions, but they may be appreciated by both the author and the editor.

If you request revisions in your decision (“accept with major/minor revisions” or “revise and resubmit”), then you can list those revision requests in the comment section as well.

Often, the comments are given in the following form:

Pg. 5. [comment 1]
Pg. 16. [comment 2]
Pg. 22. [comment 3]

These comments may concern the following (among others):

a.         Additional references.
b.         Formulation of the principles.
c.         Definitions of terms used.
d.         Clarification of background assumptions.
e.         Additional relevant data.
f.          Possible alternative analyses.
g.         Technical questions about the analysis.
h.         Questions about the analysis from a cross-linguistic perspective.
i.          Clarification of methodology.
j.          Editorial comments, writing mechanics (optional).

In doing a paper review, you are not an editor. Your job is not to correct spelling errors or to help the author write better. If you decide to make such corrections, the author and the editor may appreciate it, but it is not required. Normally, your publication decision should not be based on spelling, style or writing mechanics. Of course, in the extreme case, if the paper is so badly written that it is impossible to follow the argumentation (and so you are unable to evaluate it), it should be rejected.

5. References
If you cite references in your review, give the full reference at the end of the review (just like in the reference section of a paper). If you write, “Smith 2005 says….” without providing the full reference, it is possible that the author will not be able to locate the reference. This is unfair to the author who is trying to follow up on your comments.

In fact, reviewers are often even less specific, using phrases like “In Smith’s recent papers” or “Given the work of Smith…”.  I am under the impression that more often than not reviewers do not provide full references for works cited at the end of a review. This seems like a bad habit our field has developed.

Other Internet References

D'Alessandro, Roberta. 2018. Facebook post on Revise and Resubmit.

Haspelmath, Martin. 2019. Revise and Resubmit is damaging to science and should be abandoned. Diversity Linguistics Comment.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.