Recently I got corrected by other person from "a bosonic bath" to "an bosonic bath" (neither of us is a native speaker). That sounds weird to me, and does not look according to the rules I learned back in school. However, a Google search reveals that the "an" variant also sees some use. One can also find "an bosonic field theory", "an bosonic integer quantum Hall state", "an bosonic cloud" etc used. What is the motivation for that? And what variant should be considered correct?
There is a related question about "an yearly" used in some old literature, Why is it "an yearly"?. There, as far as I understood, it can be explained as i) before the Great Vocal Shift, "yearly" was pronounced "eerly", so that's why "an yearly" in older writing ii) the one and only form of indefinite article was "an" long ago, and that form has been preserved in formal language for a longer time. The first explanation does not fit "an bosonic" at all. The second one somehow fits, but... It feels then that some people want to sound more formal and archaic, despite the fact that their language is full of new words.
EDIT: There was a request for "authoritative sources" for the "an bosonic" usage and a claim that Google Ngrams does not find this variant (maybe because of that my question was marked as "off-topic"?). As I mentioned, usages of "an bosonic" can be Googled. Of them, I found three interesting occurences:
- "an bosonic integer quantum Hall phase" in the text of Phys. Rev. Lett. 110, 046801 by T. Senthil and Michael Levin, of MIT and University of Maryland respectively. Physical Review journals are published by the American Physical Society and the papers there are edited rather thoroughly (but maybe not ideally).
- "an bosonic bath" right in the abstract of Quantum 6, 691 (2022) by Rahul Trivedi, Kevin Fischer, Shanhui Fan, and Jelena Vuckovic, of whom Kevin Fischer is of Stanford University. Quantum is an international journal, with some of the editors most probably native speakers.
Interestingly, I have also found two cases when a paper preprint published on arXiv contains "an bosonic" which then was edited away in the published version. Both papers are from the European New Journal Of Physics. Also, there is a thesis from the University of California, Berkley with two versions found online, and in one of them "an bosonic gas" is edited away.
Maybe occurrences of "an bosonic" are merely mistakes that accidentally passed through an editor...