I have never received (nor written) a referee report suggesting that the authors consult a native English speaker for copy editing purposes. I suspect I would have a similar reaction as you did, in part due to the assumption that native English speakers are necessarily good writers or copy editors — this is certainly not the case. For the situation you mentioned that happened with you, I hope you felt comfortable talking to your PhD supervisors or other mentors. You mentioned in the comments that you didn’t feel empowered to contact the editor yourself — I could imagine that for a truly egregious report a PhD supervisor could have done so on your behalf. There are reasons for not wanting that either, of course. I just hope you had supportive people to talk to at the time.

There have been many cases where I have commented on language usage in referee reports — either due to grammatical or syntactical errors, which I think is what you are referring to, or due to poor organisation, e.g. of the introduction. I would suggest the following steps:

1. If the writing or organisation makes it impossible for you to evaluate the paper, write a first referee report asking for the language issues to be corrected, withholding judgment of the content until you receive a revised version. (I didn’t know this as a PhD student, it’s certainly possible to have multiple rounds of refereeing and revision, for the same paper at the same journal.) Or, contact the editor directly about this. Personally I would only do this in extreme cases, since/if I can imagine an editor outright rejecting the paper on the basis of writing alone, which is not my goal.

2. Split your referee report into a “content” section and an “exposition” section. State explicitly that any issues with exposition do not detract from the content, assuming this is true.

3. In the “exposition” section, as with the “content” section, it is up to you how fine-grained your comments are. For organisational issues, it would be reasonable to list your concerns and either suggest alternatives or leave it to the authors to address. For obvious writing errors, you could mention some major recurring errors, and leave it to the authors to detect where these have occurred. Once, as a much younger person with much more time on my hands, feeling guilty about not recommending a paper for publication written by friends of mine, I listed every single language error I found — I would not recommend doing this in general.

4. Consider making constructive comments about both content and exposition even if you do not recommend the paper for publication, to ease a future refereeing process.

5. Don’t equate “native speaker” with “competent writer and expositor”. Instead of writing “Please consult a native speaker before resubmitting” consider writing “Please correct the errors in language before resubmitting”, ideally having pointed some out. The first assumes that the authors are not native speakers as well as that any arbitrary native speakers could/would have written better. The second centres the actual issues you have detected with the paper and provides specific examples of what you would like changed.

6. As mentioned in the other answers and comments, be kind. Consider being especially kind to early career colleagues and to those who may already feel isolated within academia.

I have not received (nor written) a referee report suggesting that the authors consult a native English speaker for copy editing purposes, but I suspect I would have a similar reaction as you did, in part due to the assumption that native English speakers are necessarily good writers or copy editors — this is certainly not the case. For the situation you mentioned that happened with you, I hope you felt comfortable talking to your PhD supervisors or other mentors. You mentioned in the comments that you didn’t feel empowered to contact the editor yourself — I could imagine that for a truly egregious report a PhD supervisor could have done so on your behalf. There are reasons for not wanting that either, of course. I just hope you had supportive people to talk to at the time.

There have been many cases where I have commented on language usage in referee reports — either due to grammatical or syntactical errors, which I think is what you are referring to, or due to poor organisation, e.g. of the introduction. I would suggest the following steps:

1. If the writing or organisation makes it impossible for you to evaluate the paper, write a first referee report asking for the language issues to be corrected, withholding judgment of the content until you receive a revised version. (I didn’t know this as a PhD student, it’s certainly possible to have multiple rounds of refereeing and revision, for the same paper at the same journal.) Or, contact the editor directly about this. Personally I would only do this in extreme cases, since/if I can imagine an editor outright rejecting the paper on the basis of writing alone, which is not my goal.

2. Split your referee report into a “content” section and an “exposition” section. State explicitly that any issues with exposition do not detract from the content, assuming this is true.

3. In the “exposition” section, as with the “content” section, it is up to you how fine-grained your comments are. For organisational issues, it would be reasonable to list your concerns and either suggest alternatives or leave it to the authors to address. For obvious writing errors, you could mention some major recurring errors, and leave it to the authors to detect where these have occurred. Once, as a much younger person with much more time on my hands, feeling guilty about not recommending a paper for publication written by friends of mine, I listed every single language error I found — I would not recommend doing this in general.

4. Consider making constructive comments about both content and exposition even if you do not recommend the paper for publication, to ease a future refereeing process.

5. Don’t equate “native speaker” with “competent writer and expositor”. Instead of writing “Please consult a native speaker before resubmitting” consider writing “Please correct the errors in language before resubmitting”, ideally having pointed some out. The first assumes that the authors are not native speakers as well as that any arbitrary native speakers could/would have written better. The second centres the actual issues you have detected with the paper and provides specific examples of what you would like changed.

6. As mentioned in the other answers and comments, be kind. Consider being especially kind to early career colleagues and to those who may already feel isolated within academia.