One Thousand and Twenty Four

Everybody knows that no two fingerprints are alike. That fact, on its own, is almost useless to a police force.

Say you have a print from a burglary and a cabinet holding a hundred thousand cards. Uniqueness tells you that exactly one card matches. It does not tell you which drawer to open. Without a way of filing prints so that a search goes to one small group rather than through the whole cabinet, the discovery is a curiosity.

The filing system is the invention. It was worked out in Bengal in the 1890s, and the man who did the mathematics was recruited by advertisement.

What came before

The method police used was anthropometry, devised by Alphonse Bertillon in Paris. You measured a prisoner: head length, head breadth, the middle finger, the left foot, the forearm, and several more, and filed the card by those numbers.

It worked, after a fashion, and it was slow. Measurements had to be taken carefully by trained men, different men got slightly different numbers off the same prisoner, and searching a large collection took about an hour. It also failed in the way that any system of continuous measurements fails: two different people can be the same size.

Edward Henry, Inspector General of Police in Bengal, wanted something better. In 1892 he wrote to the principal of Presidency College in Calcutta asking him to recommend his strongest student of statistics.

Azizul Haque

The student nominated was Qazi Azizul Haque, born about 1872 in Paigram Kasba in the Khulna district. His parents had drowned in a boat accident when he was a child. He left home at twelve for Calcutta and got himself an education in mathematics and science at Presidency.

He was taken on as a sub-inspector, initially to run the anthropometric system in Bengal, which is to say to operate the method he would shortly replace.

The formula

What Haque produced is a way of turning ten fingers into a single address.

Each finger is scored on whether it carries a whorl, one of the three broad pattern types along with loops and arches. The fingers are taken in pairs and the pairs are weighted: the first pair is worth 16 to each finger, the second 8, the third 4, the fourth 2, the fifth 1. A whorl scores the value of its position. Anything that is not a whorl scores nothing.

Add up the scores on the even numbered fingers and that is your numerator. Add up the odd numbered fingers and that is your denominator. Add one to each, so that a hand with no whorls at all gives you 1 over 1 rather than 0 over 0.

Each half can now take any value from 1 to 32. Thirty two possible numerators against thirty two possible denominators, which is thirty two columns and thirty two rows, which is one thousand and twenty four pigeonholes.

That is the whole thing. Ten fingers, five weights, one fraction. A clerk with no mathematics at all can look at a set of prints, count whorls, do two additions, and walk to the correct drawer.

Hem Chandra Bose, working alongside him, built out the sub classification that divides those thousand and twenty four groups further, so that even a very large collection resolves to a small pile of cards. He also devised a telegraphic code for transmitting a fingerprint classification by wire, which is what turns a filing system into a network.

What it did to the work

By 1897 Haque had collected around seven thousand sets of prints, and the bureau at Calcutta became the first fingerprint bureau anywhere in the world.

The comparison with the old method is the point. Identification that had taken about an hour under anthropometry took about five minutes. Not because the matching was cleverer, but because the search space had been cut by a factor of a thousand before any comparing began.

In 1900 fingerprinting replaced anthropometry throughout British India. In 1901 Henry went back to Britain as an Assistant Commissioner of the Metropolitan Police and set up a fingerprint bureau at Scotland Yard on the same principles. From there it went everywhere, and it is still the skeleton of how prints are filed, including inside the software that now does the filing.

Whose name is on it

It is called the Henry Classification System.

Henry took the credit for it in the years when it spread, and the Indians who did the work were not named. That is the ordinary shape of a colonial administration: a British officer's report, a British officer's system, and staff who appear in it as staff.

What happened afterwards is less ordinary, and it should be recorded. Haque asked the government for recognition of what he had contributed, and Henry supported him rather than resisting. On Henry's recommendation the government gave Haque and Bose an honorarium of five thousand rupees each. Haque was made Khan Shahib in 1913 and Khan Bahadur in 1924. Bose received the corresponding titles.

And in a letter in 1926, Henry wrote that Haque had contributed more than any other member of his staff to perfecting the classification.

That is an unambiguous statement, made by the man whose name the thing carries, in writing, and it has been available the whole time. The system is still called Henry's.

What it is fair to conclude

Not that Henry was a fraud. He set the problem, ran the bureau, backed the method against considerable scepticism, and pushed it through two police forces on two continents, which is real work and not everybody could have done it.

The narrower point is about how names attach to things. The person who commissions a piece of mathematics and the person who does it are not doing the same job, and only one of those jobs leaves a name on the result. Haque died in 1935 in Champaran. The formula he wrote is applied a very large number of times a day, by machines, in every country that keeps a criminal record.

One thousand and twenty four pigeonholes, and the arithmetic to find the right one, from a boy who arrived in Calcutta on his own at twelve.

Sources

• Wikipedia, "Henry Classification System" (the development of the system by Edward Henry, Qazi Azizul Haque and Hem Chandra Bose in British India; the primary classification giving values of 16, 8, 4, 2 and 1 to successive pairs of fingers with whorls scoring and other patterns counting zero; the sum of even numbered fingers forming the numerator and odd numbered fingers the denominator, with one added to each, producing a range from 1 over 1 to 32 over 32 and therefore 1,024 groups; Haque's mathematical formula for sorting into 1,024 pigeonholes and Bose's improvements to sub-classification; the replacement of anthropometry across British India in 1900; and the establishment of the first United Kingdom fingerprint bureau at Scotland Yard in 1901 after Henry returned to Britain).
• Wikipedia, "Qazi Azizul Haque" (his birth around 1872 at Paigram Kasba in the Khulna district and death in 1935 at Champaran; the loss of his parents in a boat accident and his departure from home at twelve for Kolkata; his education in mathematics and science at Presidency College; Henry's request in 1892 to the principal for a strong statistics student and Haque's recruitment as a police sub-inspector to run the anthropometric system in Bengal; his formula organising fingerprints into 1,024 pigeonholes across 32 columns and 32 rows; his collection of 7,000 sets of prints by 1897; the reduction of identification time from about an hour to about five minutes; the titles Khan Shahib in 1913 and Khan Bahadur in 1924 and the honorarium of 5,000 rupees, matched for Bose; and Henry's letter of 1926 stating that Haque contributed more than any other member of his staff).
• Wikipedia, "Hem Chandra Bose" (his improvements to the sub-classification of fingerprints, his work on a telegraphic code for transmitting fingerprint classifications, and the titles conferred on him).
• The Quint, "How Two Indians Finally Won Credit For Henry Classification System" (the establishment of the world's first fingerprint bureau at Calcutta in 1897 largely through their efforts, Henry's initial appropriation of sole credit without reference to either man, and the later acknowledgement and compensation on his recommendation).
• Kolkata Police, "Lal Bazar headquarters" (the building housing the police headquarters at which the bureau was established).