By the time of the Gilchrist murder, the Holmesian method of rational inquiry, in which observed facts rather than reflexive prejudices dictate the solution, was well established, at least among fictional detectives. Holmes was so skilled at this way of working that Conan Doyle’s stories anticipate the use of similar methods by actual police forces. “To-day criminal investigation is a science,” the distinguished forensic pathologist Sir Sydney Smith wrote in 1959. “This was not always so and the change owes much to the influence of Sherlock Holmes.”
As early as 1932, the true-crime writer Harry Ashton-Wolfe could declare:
Many of the methods invented by Conan Doyle are today in use in the scientific laboratories. Sherlock Holmes made the study of tobacco-ashes his hobby. It was a new idea, but the police at once realised the importance of such specialised knowledge, and now every laboratory has a complete set of tables giving the appearance and composition of the various ashes, which every detective must be able to recognise. Mud and soil from various districts are also classified much after the matter that Holmes described….Poisons, hand-writing, stains, dust, footprints, traces of wheels, the shape and position of wounds, and therefore the probable shape of the weapon which caused them; the theory of cryptograms, all these and many other excellent methods which germinated in Conan Doyle’s fertile imagination are now part and parcel of every detective’s scientific equipment.
In the Gilchrist investigation, alas, these techniques were either irrelevant or of little help, a circumstance that worked to Slater’s cost. But even with scant scientific means at their disposal, the Glasgow police had access to one powerful forensic tool, though they rarely seem to have used it: logical reasoning. That, after the rigorous sifting of empirical evidence, is the next step in the Holmesian method and in many ways its soul. Though Holmes himself often describes this brand of reasoning as deductive, it actually entails no deduction.* It hinges, more properly, on a logical process known as induction—or, still more properly, abduction.
“Abduction” was first used in this sense by the American philosopher Charles Sanders Peirce. A polymath whose work had profound implications for philosophy, logic, semiotics, mathematics, psychology, anthropology, and other fields, Peirce (pronounced “purse”) was born in Cambridge, Massachusetts, in 1839; his father, Benjamin Peirce, a Harvard mathematics professor, had helped establish the Smithsonian Institution. After graduating in 1859 from Harvard, where he studied chemistry, Charles took a job as a surveyor with the United States Coast and Geodetic Survey, work that for the next thirty-two years would support his wide-ranging philosophical investigations: at his death in 1914, he left a written legacy of some twelve thousand published pages and eighty thousand manuscript pages.
Abduction, or “retroduction,” as Peirce also called it, is much like Huxley’s “retrospective prophecy.” Presented with a set of effects—animal tracks, medical symptoms, crime-scene clues—the investigator uses abduction to pinpoint their most logically probable cause.
“A given object,” Peirce wrote, “presents an extraordinary combination of characters of which we should like to have an explanation. That there is any explanation of them is a pure assumption; and if there be, it is some one hidden fact which explains them; while there are, perhaps, a million other possible ways of explaining them, if they were not all, unfortunately, false. A man is found in the streets of New York stabbed in the back. The chief of police might open a directory and put his finger on any name and guess that that is the name of the murderer. How much would such a guess be worth?” (The Glasgow police, of course, did essentially this in fingering Slater.)
The abductive method permits no such precipitate conclusions. “Abduction makes its start from the facts, without, at the outset, having any particular theory in view, though it is motivated by the feeling that a theory is needed to explain the surprising facts,” Peirce writes. “Induction makes its start from a hypothesis which seems to recommend itself, without at the outset having any particular facts in view, though it feels the need of facts to support the theory. Abduction seeks a theory. Induction seeks for facts.”
Abduction, as a group of British scholars explain in an article on medical diagnosis, takes the following form:
FACT C is observed.
IF A were true, C would be a matter of course.
Therefore there is reason to suspect that A is true.
That process is the mirror image of deduction: In deduction the investigator reasons forward, from cause to effect. When Holmes says, as he does in his debut appearance, “In solving a problem of this sort, the grand thing is to be able to reason backward,” he is singing the praises of abduction. To illustrate the differences among deduction, induction, and abduction, Peirce invoked a trio of syllogisms like these:
Deduction
RULE: All serious knife wounds result in bleeding.
CASE: This was a serious knife wound.
THEREFORE [the deduced result]: There was bleeding.
Induction
CASE: This was a serious knife wound.
RESULT: There was bleeding.
THEREFORE [the induced rule]: All serious knife wounds result in bleeding.
Abduction
RULE: All serious knife wounds result in bleeding.
RESULT: There was bleeding.
THEREFORE [the abducted case]: This was (likely to have been) a serious knife wound.
In assembling their case against Slater, police and prosecutors were working deductively, to the detriment of justice. If their preposterous reasoning were schematized, it would look much like this:
RULE: All murders are committed by undesirables.
CASE: Oscar Slater is an undesirable.
THEREFORE, Oscar Slater committed the Gilchrist murder.
Abduction, like the reconstructive sciences of the Victorian age, generates narrative. Zadig used precisely this method in spinning an etiological thread that would account for the observed facts of a case. So, more than a century later, did Sherlock Holmes:
“We are coming now rather into the region of guesswork,” Holmes’s client Dr. Mortimer protests in The Hound of the Baskervilles.
“Say, rather,” Holmes replies, “into the region where we balance probabilities and choose the most likely. It is the scientific use of the imagination, but we have always some material basis on which to start our speculation.”
In case after case, Holmes uses abduction to solve mysteries, reasoning backward until, he said, “the whole thing is a chain of logical sequences without a break or flaw.” In “The Adventure of the Six Napoleons,” a 1904 story, London is plagued by a string of bewildering crimes: the theft and smashing, one by one, of a set of identical plaster busts of Napoleon. To the dependably dim Inspector Lestrade of Scotland Yard, the obvious explanation is that the thefts are the work of a madman, someone “who had such a hatred of Napoleon the First that he would break any image of him that he could see.”
But to Holmes, Lestrade’s theory accounts for the facts only trivially. If a madman were indeed moved to smash Napoleon’s every image, then, why attack those particular busts, “considering,” Holmes points out, “how many hundreds of statues of the great Emperor must exist in London”? And why, Holmes asks further, did the culprit, having made off with one of the busts, wait to smash it until he reached a particular spot in the street? “Holmes pointed to the street lamp above our heads,” Conan Doyle writes. “ ‘He could see what he was doing here, and he could not there. That was his reason.’ ”
These rational observations, combined with empirical legwork, allow Holmes to construct, as he tells Lestrade with no little ego, a narrative of the crime “by a connected chain of inductive reasoning.” The real object of stealing and smashing the busts, he correctly concludes, was to find a priceless jewel concealed inside one of them.
“Holmes…operates like a semiotician,” the critic Rosemary Jann has written. “He ‘reads’ crimes like literary texts, as if they were systems of signs. The true significance of each sign is determined by its relation to others in a particular network of meaning….He is able eventually to recognize the one relationship capable of accounting for all the clues.”
In contrast, the police of the period—in the Holmes canon and all too often in life—tended to think not in terms of a subtle web of contingencies but of a straight line, drawn in unambiguous black and white. Holmes acknowledges this danger in an 1891 story, “The Boscombe Valley Mystery,” when he declares, “There is nothing more deceptive than an obvious fact.”
“Many men have been hanged on far slighter evidence,” Watson concurs.
“So they have,” Holmes replies. “And many men have been wrongfully hanged.”
* Conan Doyle, per common parlance, tended to use “deduction” as a general rubric denoting any type of logical inference.