I have had an inkling for a while now that as a copy editor, I have been enforcing a rule that might not be justified. This post is part confession, part apology to all the authors whose prose I have changed without good cause, and part contemplation on prescriptivism.

For most of my editing life (including nine years as the co-editor of the Journal of English Linguistics), I have had a thing about on the other hand when it does not follow on the one hand. I have had it in my head for all these years that this is one of those points of usage that irks style guide writers and other copy editors. Therefore, as a responsible copy editor, I must enforce the pairing of on the one hand and on the other hand so that authors’ prose will not be judged as being stylistically maladroit—and so that the journal, for example, will not be seen as having lax editorial standards.

As a result, for years, when I have run across on the other hand, I have scanned backward to see if there is a[*] on the one hand; if not, I have replaced on the other hand with in contrast or something similar. And let me tell you, I have done a lot of replacing of other people’s on the other hand’s.[*] For all my descriptive tendencies as a linguist, I was privileging a prescriptive sense of logic (that if there is a second or other hand, there must be a first hand), in the face of the usage of many highly skilled, eloquent writers.

At some point, as the stranded on the other hand’s[*] piled up, I realized that perhaps I was the outlier on this point of usage. Did I actually know whether other people cared as much as I did and were judging writers for having only one contrasting hand, so to speak? I added the topic to a list of points of usage students could research for an essay in my “History of English” course. The essay assignment asked students to investigate the history of a prescriptive usage rule and compare prescription with actual usage.

when I have run across on the other hand, I have scanned backward to see if there is a[*] on the one hand;