According to my EL&U activity statistics, I have asked 15 non-wiki questions and answered 239 non-wiki questions tagged as "phrase-origin"—so obviously I find them both interesting and enjoyable to ask and answer. In fact, I think that first-published-occurrence questions and answers are among the most useful posts that appear on this site, for two reasons.
First, most general-reference resources (such as standard dictionaries) do not focus on the origins of phrases to the same extent that they do on the etymology of individual words and the first published occurrences of individual words. As a result, accounts of phrase origins tend to be less readily available and less thoroughly presented than accounts of individual word origins.
Second, the emergence of searchable online book and periodical databases over the past three decades has made it possible for amateurs (like the vast majority of participants at this site) to find earlier instances of specific phrases than the earliest ones that professional researchers (like those at OED) have reported. Consequently, to the extent that anyone cares what the earliest documented occurrence of a particular phrase is, amateurs can contribute to the body of current knowledge.
The goal of such research isn't to claim to have identified the very first time a phrase was ever used or the very first time it ever appeared in print. Rather, it is to provide documentation that the phrase was in use as of the date noted and, if possible, to clarify how it was being used in one or more particular instances of such early use.
The limitations of this type of research are obvious: it depends on written sources; it extends only to the publication databases that the researcher has access to; and it depends on the researcher's ability or good luck in anticipating antecedent wordings or spellings that might reveal the evolution of the phrase in question before it arrived at the exact form that a posted question asks about.
Since I don't subscribe to any pay-walled book or periodical databases (of which there are many online) and since I don't have access to academic resources that are limited to university researchers, I have no illusions that the relatively small pool of online database content available to me for free is likely to yield the earliest instances of a word or phrase that a more extensive search might produce. Nevertheless, I think that any incremental progress in documenting earlier occurrences of a phrase than were heretofore identified makes the effort worthwhile.
As the posted question above observes, phrase-origin questions are fundamentally questions of history. But to my mind, posing such inquiries under the category head "English Language & Usage" makes much more sense than posing them under the category head "History." Several first-rate dictionaries—including the OED, the Random House Historical Dictionary of American Slang, and A Dictionary of Americanisms—have embraced the notion of formulating and presenting their content "on historical principles." I take it that they see the "history" aspect of English as being both compatible and deeply intertwined with the "language and usage" aspects.
