I'm teaching A Study in Scarlet this semester, for the first time in several years, and so I am contemplating the utter strangeness of Part II ("The Country of the Saints"). Which has led me to further contemplate the fictional authorship of A Study in Scarlet. (Incidentally, I'm not a member of the Baker Street Irregulars, and this is not part of the Game. If you want to see how someone playing the Game solves this question, see Ben Vizoskie, "Who Wrote the American Chapters of A Study in Scarlet?" [PDF].)
1. A Study in Scarlet may or may not be A Study in Scarlet--that is, within the novel's world, we may not be reading Watson's actual narrative of the case, but instead his memoir of how he came to write it up.* After all, A Study in Scarlet is supposedly a reprint from the reminiscences of Dr. John H. Watson, MD, late of the Army Medical Department. If A Study in Scarlet is A Study in Scarlet, then the novel forms a nice circle (since it concludes with Watson announcing that he is about to write...the text in front of us).
2. But then, is "The Country of the Saints" in A Study in Scarlet--that is, in the narrative that Watson publishes (again, within the novel's world)? Doyle tells us explicitly that Watson is not the author: "As to what occurred there [London], we cannot do better than quote the old hunter's own account, as duly recorded in Dr Watson's Journal, to which we are already under such obligations" (ch. 5). It appears that our fictional author has a fictional editor, and that "The Country of the Saints" has been interpolated into the text at a "later" date. (Moreover, this interpolation also suggests that we aren't reading Watson's A Study in Scarlet.)
3. To make matters even odder, there's nobody in the novel who could have written "The Country of the Saints." The closest candidate is Jefferson Hope--dead by the time A Study in Scarlet is "published." Everyone else involved has passed on into the fictional next world; moreover, "The Country of the Saints" has an omniscient narrator, which disqualifies...well, it disqualifies everybody else in the novel, including Watson and Holmes. An omniscient narrator suggests that there's a novelist somewhere about.
4. In the midst of a fiction staged as a memoir, we have a sensationalist inset narrative overtly written as fiction--and that fiction purports to explain the murders. On the narrative's own terms, "The Country of the Saints" inserts a pointedly unanswered question into the text (just who is writing this, anyway?) in the act of supposedly answering another question (why did Jefferson Hope kill two men?). So much for clarifying the mystery. In a sense, the blatant fictionality points up Holmes' limitations: his powers stop short of full-blown authorial omniscience. Watson, meanwhile, has no claims to omniscience, even when writing in the past tense. Doyle winks at the audience, metafictionally speaking. (Whether this is an intentional effect is an entirely different issue; as Iain Sinclair notes in the introduction to the Penguin Classics edition, Doyle was desperately trying to make his book longer in order to increase its saleability [viii].)
*--In dramatized versions of the Holmes stories, what appears on the screen is not, in fact, whatever Watson is "writing." (The Jeremy Brett adaptations make this point explicitly, since the Watson we see has a much better grasp of Holmes' methods than the textual Watson.)