Sherlock Holmes would be less perceptive in our time

Detail of an artistic representation of Sherlock Holmes -

• The legendary sagacity of the hero of Sir Arthur Conan Doyle would make "shit" nowadays, according to a study published by our partner The Conversation.

• An assertion which is based on the “Bayes theorem”, which allows to combine observation, deduction and a priori.

• This analysis was carried out by Olivier Marre, neuroscience researcher at Inserm.

Sherlock Holmes is the mythical figure of the detective, the one who knows how to establish the truth thanks to his sense of deduction.

For many, Sherlock Holmes' talent lies first and foremost in an exceptional capacity for observation, symbolized by his magnifying glass, which allows him to grasp the detail that has escaped ordinary mortals.

Allow me here to explore another explanation, which is undoubtedly complementary.

Let us take an example among the purest of the talent of deduction of our famous detective, at the beginning of

.

A man lost his hat in an argument and ran away.

Holmes presents this hat to Doctor Watson, gives him his magnifying glass, and asks him to deduce from it the personality of the owner of this hat.

Sherlock Holmes and Doctor Watson, illustration by Sidney Paget for the Strand Magazine, December 1892 © The Strand Magazine / Wikipedia CC By SA 4.0

Inside the mind of Sherlock Holmes

While Watson struggles to find any clue, Holmes gives him a briefing on this person, which strikes the reader with its details, including the fact that this man has arguably suffered setbacks and lost his wife's love. :

“It is obvious that the owner of this hat was extremely intelligent, and that in recent years he has found himself in a situation, which has easily become difficult.

He was far-sighted, but much less so today, it is the proof of a moral retrogression which, added to the decline of his fortune, seems to indicate some vice in his life, probably that of drunkenness.

This sufficiently explains why his wife doesn't love him anymore.

"

Watson, and we with it, are almost shocked by these deductions, which seem to be a miracle, even a bluff, but which will prove to be correct afterwards.

Holmes gives a detailed explanation:

“This hat is three years old;

however, at this time its slightly turned flat edges were in fashion.

Then it's a top quality hat.

Take a look at the grosgrain ribbon that borders it and its neat lining.

If this man had enough to buy a hat of this price three years ago and he hasn't had another since, I conclude that his situation is less today. good than it was.

[…] Have you not noticed that this hat has not been brushed for several weeks?

My dear Watson, when your wife lets you go out with an unbrushed hat and I see you come to my house like that, I will have doubts about the good relationship between your household.

"

Let us stop for a moment on the path of the detective.

How does he deduce that his wife doesn't love him anymore?

The man is no longer loved by his wife because the hat is not well maintained.

For a 21st century reader like me, half of a modern couple, I do my share of the housework.

If by any chance I put on a hat, and this hat was not well maintained, it would more likely be due to my own negligence than to the love that my wife can have for me!

But now, the society I live in has little to do with the society in which Holmes operates.

To better understand how this influences Sherlock's reasoning, let's summon another famous Englishman, Thomas Bayes.

Bayesian reasoning

18th century pastor and mathematician Thomas Bayes is known for having bequeathed to us his Bayes theorem, arguably one of the most important theorems for anyone looking to analyze data with probabilities.

Thomas Bayes bequeathed us a theorem on which are based many algorithms for data analysis and decision making © Wikipedia CC BY SA 4.0

In particular, it enabled what is now called “Bayesian inference”, an optimal way of combining observation, deduction and a priori.

Let us try to give here a simplified and intuitive version, using our example of the hat.

Let's imagine four hypotheses:

• H1

   : I am in charge of the maintenance of my hat, and do it very well

• H2

   : I am in charge of the maintenance of my hat, but am rather negligent towards it.

• H3

   : my wife is in charge of the maintenance of my hat, and takes care of it lovingly, reflecting the love she has for me.

• H4

   : my wife is responsible for maintaining my hat, but no longer takes care of it, because she has lost her interest in me.

Two things will allow you to decide between these four hypotheses.

First, the observation of said hat.

Like Doctor Watson armed with Sherlock's magnifying glass, you observe that he is very poorly maintained.

This makes assumptions H1 and H3 much less likely than H2 and H4.

On the other hand, if you know that we are in a modern society where women and men share household chores and that, in all likelihood, the owner of this hat does not charge his wife to maintain it, then H3 and H4 are less likely than H1 and H2.

So you can see how, on the one hand, observation allowed you to modulate the probability of the four hypotheses, and on the other hand, how your

knowledge

of the society in which you live allowed you to do so.

Bayes formalized these notions.

To simplify, he gave us the method to combine all this in an optimal way, and thus estimate the final probability (called “a posteriori”) of each hypothesis.

It turns out that the hypothesis H2 is the most probable, that is to say that I am undoubtedly a negligent being, but still loved by his wife, phew.

But now let's give that same hat to Holmes.

He's going to observe the same thing as you, and Doctor Watson: this hat is poorly maintained, favoring H2 and H4.

But Sherlock Holmes is a man of the 19th century.

At that time, household chores were less well shared, and it's a safe bet that, for him, a priori, H1 and H2 are much less likely than H3 and H4.

So, combining all of that, he'll deduce that the most likely hypothesis is H4, which he does well in the story.

Consult our “Sherlock Holmes” file

What can we conclude from all this?

That the brilliant detective's fantastic capacity for deduction owes not everything to his keen sense of observation, but also to a keen knowledge of the society around him, and a remarkable ability to combine the two.

This combination of observation with his knowledge of society is even more striking in the other example highlighted above: the man suffered setbacks of fortune, because his hat is an expensive model that corresponds to the fashion of 'a few years ago, but not the current fashion.

Only a rich man a few years ago could buy it, but if he was still rich, he would have bought another more recently, following the fashion of the moment.

Unless you've done a thesis on fashionable hats in the Conan Doyle era, you just can't make that deduction, even with unparalleled observation skills.

Holmes knows his world well, the fashions of each year, and knows how to mobilize this knowledge wisely.

The advantage of a conformist society

Let's go further: what is this society where we can deduce the past and present wealth of a person in his hat?

Obviously, a fairly conformist society, the Victorian society of the day.

Holmes seems to rule out the hypothesis that the person could just keep their hat on regardless of the fashion.

A conformist society has the effect of

.

In our example above, H1 and H2 are unlikely in this Victorian society, of which Sherlock Holmes is well acquainted with the customs and traditions.

He can therefore define his "a priori" very clearly, combine it with observations and, in good Bayesian fashion, deduce the most probable hypothesis.

But one of its secrets is undoubtedly the conformism of its contemporaries, conformism which makes them predictable and makes it possible to greatly reduce possible hypotheses a priori.

This is something Holmes notices, and complains about, because it makes his daily life boring and the mysteries too easy to solve (and encourages him to explore other worlds, such as artificial paradises).

But it is precisely this society so conformist, so deterministic, that allows it to exercise such brilliant deductions.

Imagine a thought experiment.

Let's move our brilliant detective from his home in Baker Street, in the midst of the Victorian era, to Carnaby Street, the vibrant heart of the London capital of the 1960s. Conformism creaks everywhere in the era of the "swinging sixties

. 'dress more in a specific fashion.

Our detective would probably be hard pressed to deduce anything from his owner if we brought him Jimi Hendrix's hat.

At a time when all cultural codes are broken, all hypotheses become possible: the "a priori" do not tell us much, and even the most brilliant deductions would not make it possible to separate the hypotheses.

No, Sherlock couldn't have lived in the 1960s.

Science

Will LSD and hallucinogenic mushrooms ever be authorized drugs?

Science

Black holes (and their many mysteries) explained to your kids

This analysis was written by Olivier Marre, neuroscience researcher at Inserm.

The original article was published on The Conversation website.

• Science

• England

• Investigation

• Sherlock holmes

• Literature