I mean, maybe there's an aspect of pointless mimicry (though I haven't seen evidence of it), but my comment represented my general experience as someone who wor

I mean, maybe there's an aspect of pointless mimicry (though I haven't seen evidence of it), but my comment represented my general experience as someone who works the dynamical systems area and has reviewed for that journal in the past. I have yet to receive a negative review based on using a less standard technique/perspective beyond the occasional "this work doesn't belong to this field, but instead to [insert adjacent field here]" which seems to usually be ignored by the editor. I have, multiple times, had one of the three reviewers write a negative review based on a (to me at the time) seemingly conceivable misunderstanding of contribution of the paper because I had done something like include a simple example to better illustrate an idea. The editor sometimes rejects submissions based on one negative review, even if the other reviews are positive. I think it's relatively easy to speculate that the consistency in form is from the feedback of the system: Assume a researcher wants to maximize his chances of getting the paper accepted, and that, if accurately reviewed, the content of the paper would merit acceptance. Furthermore, assume that there are a finite number of distinct styles of writing in scientific papers, and that the likelihood of accurately reviewing a paper increases monotonically with the number papers you've seen in the given style (the source of which are papers that have successfully made it through the process). Then the researcher would aim to match the most prevalent writing style. Even if they aren't always successful at it, it skews the input distribution towards the most prevalent writing style. This, in turn makes it more likely for that style to be accepted in the future.