Within the Sherlock Holmes stories, Watson has the invaluable role of being the chronicler of Holmes’ adventures. This does introduce a minor problem, though, which is that it’s utterly at odds with Holmes’ assurances of discretion to his clients.

This is not every case, of course; there are cases where Holmes has no duty of secrecy to anyone. A Study in Scarlet does not bring about this problem, and I doubt that The Sign of Four does either (the only person to whom Holmes might owe discretion, there, became Watson’s wife and was thus available to wave her rights to confidentiality). But as soon as The Adventures of Sherlock Holmes begin we run into a problem with A Scandal in Bohemia.

I know that the King said that in two years it wouldn’t matter who knew that he had had an affair with Irene Adler but it is none the less hard to imagine that he would take it kindly to his affair with Adler being made public for no reason other than the publicity of Sherlock Holmes. Mr. Jabez Wilson was, perhaps, not owed any great duty of discretion, so we can pass on the Red Headed League.

The Case of Identity was, I think, a bit cruel for Watson to have published, since it would have created much embarrassment for the woman who came to Holmes. This, however, could be dealt with by Watson changing the names. Few people would be over-likely to recognize the original in the story if her name, as well as the name of her stepfather, and also of the places and businesses were changed.

None of this will work for The Boscombe Valley Mystery. Here, Holmes promised a dying man that for the sake of his daughter he would not reveal the dying man’s former life as a highway robber and murderer. The whole point was to keep from his daughter what her father was. Even if the names had been changed there could be no mistaking who the story was about, especially since she was the one who had called Holmes into the case. There can be no excuse that the case happened many years before, as she was a young woman during the story and would still have been a young woman at the time of publication. (Further, since The Sign of Four took place before The Boscombe Valley Mystery (since it refers to Watson as a married man, and he married in The Sign of Four) the earliest the latter could have taken place would have been about 1887, and The Boscombe Valley Mystery was published in 1891.)

This was by no means the only such case. The Adventure of the Naval Treaty seems equally harmful to ever publish, even with names altered, as does The Adventure of the Second Stain. In both cases the preservation of someone’s reputation requires that no one ever find out what really happened, and the details are utterly unmistakable.

These are just the cases where Holmes either implicitly or explicitly promised to never reveal the damaging secrets to anyone, and Watson’s chronicling them is an inarguable breaking of that promise. Probably the majority of cases Watson chronicles would constitute a violation of privacy and trust to the people concerned.

Stranger still, more than a few people who come to Holmes have heard about his adventures as published by Watson and nevertheless come with a hope of discretion on the part of Holmes.

Ultimately, I don’t think that there’s any solution to this. Just pick up any book trying to reconcile the chronology of the Holmes stories and you’ll discover that Conan Doyle clearly didn’t worry about the details. When Watson’s wife—Mary Morstan—became inconvenient, Conan Doyle basically just forgot about her. (It’s more complicated than that, but not tremendously more complicated).

So, what are we to make of Watson’s chronicling of Holmes’ cases being a contradiction of Holmes’ duty to his clients?

I think that one clue to why this was the case can be found in the titles of most of the Holmes stories. The majority of the stories began with “The Adventure of.” These were not really detective stories in the sense of how the genre would evolve in the 1920s. Conan Doyle obviously thought that the science of deduction was interesting but he equally obviously didn’t think of it as something that the reader would be doing.

(I’m still not sure when the idea of the mystery story being a game between the reader and the author began or, more importantly, when it became commonplace. It was certainly well established by the time of Fr. Knox’s Decalogue, but that was published in 1929. You can see some of this in G.K. Chesterton’s advice on how to write a detective story, which was published in G.K.’s Weekly in October of 1925.)

Since these are adventure stories where the science of deduction is meant only to be interesting but not something where the reader is trying to solve the mystery before the author, I suspect that Conan Doyle did not expect the reader to pay attention to details in the same way that later mystery authors would expect that. I suspect that this is why he was not bothered by the conclusion of a story making it impossible for Watson to have published it.

(Incidentally, I think that the contradictory chronology is more an effect of the stories being written many years apart and Conan Doyle simply having forgotten what he wrote early. I’ve faced this problem in writing my own detective stories. The result is I started keeping a file for all of the miscellaneous facts such as birthday, height, weight, age, siblings, siblings’ age, etc. Annoyingly, I had to re-read my own books in many cases to find out whether I wrote down such details before, though now I mostly can write them into the file whenever I invent a new one. Conan Doyle wrote Holmes stories over the course of nearly forty years; if he was not making a conscious effort to be rigorous about details it would have been shocking for him to not mis-remember things he had written decades before.)