Back to SherlockPeoria front page

The Dissecting Room . . . June 1993

Back to the Dissecting Room Index

 

Frankenstein's Sherlock

Every now and again, a magical moment occurs. Two totally unrelated chains of thought intertwine for just an instant, just one fleeting instant, and an idea is born. And an idea, once born, will often cling to life with an incredible tenacity. Even if said idea has no practical purpose and makes no sense at all.

This summer I will have spent sixteen years as a serious Sherlockian. A lot of side issues have come up along the way, but my primary involvement with the master detective has always been, and will always be, his methods of detection, his patterns of thought, and the various logics of what it took to be Sheriock Holmes. All the rest is just icing on the cake.

Last year, partly as a career move, partly as a diversion, I started studying computer programming. Nothing fancy, just plain old business-oriented COBOL, but it was the place I set out to get my feet wet in a massive and rapidly changing field. As my schoolwork in COBOL draws to a close I have begun to look at where to head next in my computer studies. At that point, I have found I am of two minds. One of those minds is serious, career-oriented, and has some very practical goals in mind. And the other?

The other portion of my thoughts is dreamily drifting toward the realm of expert systems and neural nets. For those of you to whom those phrases mean nothing, they have to do with a newer branch of computer programming, a branch science fiction readers might be more familiar with under the title artificial intelligences.

I have only one interest in artificial intelligence, and one only, when you come right down to it: I want to build a Sheriock Holmes.

We Sherlockians have been making Sheriock Holmes real for decades. It might also be argued that we didn't need to make him real, because he was real when we found him. There is much truth in that. Sherlock Holmes has always been one of the most real human beings I have ever met, and I have met a great number of very real human beings. The lad who spent his days wandering the realms of fiction began wandering through reality a few years back and found it every bit as exciting, given the chance.

When  you  think  about  it,  some  of  the  greatest Sherlockians throughout our history have been real people persons. They weren't fascinated by Holmes because they lived their lives in dream worlds. Edgar Smith didn't become a vice president at General Motors by having his head in the clouds. Yet Smith, and all his peers, accepted Sherlock Holmes as a very real person.

The only thing is, Holmes always seemed to lack a certain “presence" in the day-to-day world. Conversations with him seem out of the question at present, and to chat with the master detective is something we'd all like to do.

So why not give this very real person a new sort of reality? Why not program a Sherlock Holmes? With technology at its current level, every memory of every case Watson recorded could be provided to our electronic Sherlock by simply hooking up a CD-ROM drive. True, the quick wit, the flair for the dramatic, and detective skill may be a real challenge to come up with algorithms for, but it took well over twenty years and much study to produce the first Sherlock Holmes. Why should the second come any quicker?

Okay, so it's a very ambitious idea. The current level of computer science may not even be up to the challenge, and personally, my own skills are nowhere close to the current level of computer science. But it's been my experience that the really ambitious ideas are often the best ones. One such interest dogged me for ten years and I got a book out of it. This one could probably take about thirty years or more, but the result could be so much more rewarding.

I'd be interested in hearing from any other Sherlockians who have an interest in such a thing. I'd also be interested in hearing from anyone who thinks the very thought is a crime against God, nature, and Conan Doyle.

Who knows? After forty years of programming Holmes, we might even be able to spend two or three more and get a Watson to go with him.