Elementary, dear Watson!

1. MOTIVATION In “The Sign of Four", Sherlock Holmes makes some very startling inferences about Watson’s whereabouts by informing Watson that he was at the Wigmore Street Post Office earlier that morning. On further inquiry by Watson, Holmes explains that he combined (joined) the observation that “Watson’s shoes had a reddish mud on their instep", with the fact that “just opposite the Wigmore Post Office the pavement was freshly dug up to expose a similar reddish mud and it was so positioned that it would be challenging to enter the office without treading into the mud". Contrary to the popular belief that this is deduction, this method of reasoning is actually an example of abductive inference. Abduction begins from the facts observed and then seeks the simplest hypothesis (“Occam’s razor") that best explains the facts, while deduction finds data to support a hypothesis. Given the big data challenge that we presently face, is it then possible to utilize an abductive model (effect to cause reasoning) to find the best explanation for the observations, as opposed to the traditional method of forming hypotheses and testing them with observations (cause to effect reasoning)? Can our databases extend our understanding of this data automatically by inferring explanations from incomplete observations?