Discovery of a Prototype Hollerith Machine in Paris
David Anderson, Janet Delve and Hans Pufal
Probably the oldest surviving example of the Hollerith Tabulating Machine is to be found in Paris as part of the collection of the Conservatoire des Arts et Métiers (CNAM). Unfortunately Hollerith’s device is not on general display but is in the CNAM reserves stored on a pallet at a height of almost 2 metres above floor level making examination on a first visit quite difficult.
The authors, all members of the History of Computing Group at the University of Portsmouth, UK, visited the CNAM in January 2005 and were able to see the Hollerith machine at first hand.
In contrast to the “classical” Hollerith machines as used in the 1890 census, this machine is comprised of three elements: the press desk, the counters, and the sorter. Each are wooden constructions but mounted on wrought iron legs. The principle of operation is, however, identical to the census machines.
While dating Hollerith machines is not entirely a straightforward business, it can be said with some confidence that the machine pictured [right] (figure 1) is the oldest surviving Hollerith tabulator in the world, and is in any event better representative of Hollerith's early thinking than any other extant device.
An American Machine in Paris
The tabulator was brought to Paris by Hollerith himself for display in the ‘Exposition Universelle’ of 1889 (the Eiffel Tower was another exhibit) and is a near duplicate of the machine being used with some success since its installation in the U.S. War Department on December 9th 1888 from whom Hollerith attempted to secure:
“some statement in the nature of a certificate, regarding the operation and use of this system… so that I could use the same in connection with my exhibit in Paris”1
Hollerith’s request to the Secretary of War was dated April 8th 1889. Despite the report of Captain Fred C. Ainsworth of the Record and Pension Division “that the machine in question has thus far given satisfaction”, Hollerith’s request fell on deaf ears and he was eventually told by Major Charles R. Greenleaf, Surgeon, U.S. Army :
“Sir, I am directed…. to inform you that the Secretary of War… states that it is not the custom of the War Department to give certificates of the character you asked for.”2
Hollerith left for Europe in the middle of April 1889 going first to Berlin in order to exhibit his machine and to look into securing a German patent. By the middle of May, Hollerith had arrived in Paris and the tabulator was installed with other “Machines à Calcul” in the “Instruments de Précision” Section of the Palace of Liberal Arts Building.
Hollerith did not tarry long in France, returning to the United States on business within a fortnight. It is not clear if Hollerith stayed long enough to see his machine awarded at Gold medal at the ‘Exposition Universelle’ but in any case, neither France nor its capital, which he described as being “like a big lunatic asylum let loose.”3 had endeared themselves to him.
It seems clear that the machine which Hollerith took to Europe dates from no later than the beginning of 1889 and could have been of much earlier origin.. It is an exact or nearly exact copy of the War Department machine and predates the devices used in the 1890 Census for which it may be regarded as a prototype.
Records held by the CNAM show that the Hollerith machine entered their collection in 1889, although they do not indicate if this was as a result of its having won a gold medal at the Exposition Universelle or for some other reason. It is known that Hollerith found the cost of transport of the machine to France relatively expensive and so there may have been an element of financial expediency in his not taking on the cost of its return to the United States. As well as the machine itself, several published articles of various vintage are preserved, as are a few dozen pink punched cards of exactly the type illustrated by Hollerith in the Manufacturer and Builder of April 18904 (see Figure 9) and marked with the year 1888.
Extant Hollerith Tabulators
Original Hollerith tabulators are fairly rare with just three others known to exist. One is to be found in the Smithsonian Institute5. All the dials on that device are marked "THE HOLLERITH (/) ELECTRIC TABULATING SYSTEM (/) PATENTED, 1889. For this and other reasons it is reasonable to conclude that the machine was certainly not made before 1889. The machine also has a tag screwed to the front of it, beneath the counters, that reads "The (/) Tabulating Machine Company (/) Washington, D.C. (/) SYSTEM PATENTED JAN. 8. 1889." which was the date his patent was approved and should not be taken as indicating a date of construction.
A second Hollerith machine dating from around 1894, which was used in the Norwegian census of 1900, survives in the Norwegian Museum of Science and Technology.
The only other original Hollerith device known to exist was held by IBM. This machine, which is thought to have dated from around 1890, was placed on display in Manhattan in 1994. Regrettably, in what must be regarded as an act of vandalism, the original working parts were observed to have been replaced by 1960s components. The current location of the machine is not known.
The origins of the Conservatoire des Arts et Métiers
It is probably worth saying something about the creation and history of the CNAM. In the autumn of 1794, in the immediate aftermath of the French revolution, the French Roman Catholic priest, abolitionist and revolutionary leader abbé (Henri) Grégoire proposed to the National Convention:
«Il sera formé à Paris, sous le nom de Conservatoire des Arts et Métiers, [...] un dépôt de machines, modèles, outils, dessins, descriptions et livres dans tous les genres d’arts et métiers.»6
Grégoire’s ambitious goal was the creation of a resource which would serve as a reference and inspiration to inventors, researchers and the generally curious. With his backing, a decree establishing the new institution was signed less than two weeks after the proposal was first tabled.
Immediately on its creation, the CNAM received to its safekeeping an impressive collection of models and machines, previously assembled by Jacques de Vaucanson7.
Antecedents to the Hollerith Tabulator
Born in Grenoble in 1709, Vaucanson was an engineer and renowned maker of automata. Interestingly we may trace an indirect line of influence from Vaucanson to Hollerith, through the work of the former modernising the French weaving industry. Arising out of this effort, Vaucanson, in 1745, gave the world its first completely automated loom utilising punched cards .
Some fifty years later Joseph-Marie Jacquard famously incorporated punched-card technology into his own looms having discovered Vaucanson’s machine in the CNAM. Variations on Jacquard’s punched cards were adopted by Babbage in 1830 and later in the century by Hollerith for tabulating the 1890 census8.
The CNAM’s early collection also included a variety of scientific objects and “curiosities”. Another example which provides an indirect link to the history of computing was a fire pump, originally developed by the Yorkshireman Joseph Bramah which had only recently been confiscated from the aristocracy. Born near Barnsley in 1748, Bramah was a famous locksmith. He was recognised generally as one of the mechanical geniuses of his day and under his tutelage many other fine engineers received their training. His pupils included such luminaries as Henry Maudslay and, more interestingly in the current context, Charles Babbage’s main engineer, Joseph Clement.
An Ever Burgeoning Collection
Under the influence of abbé Grégoire, the CNAM diligently collected anything of scientific or technical interest and after four years of being stored in a variety of locations in and around Paris, the collection as a whole was brought together and situated in its principal home to this day, the old priory of Saint-Martin-des-Champs,.
In May 1802, the CNAM opened its doors to the public and began demonstrating and explaining the marvels of the age to anyone who cared to take an interest..
The CNAM continued to amass artifacts at a prodigious rate. Inevitably, it outgrew the space available at Saint Martin which was, by now, in the heart of Paris. Around 1988, a major renovation effort, lasting some twelve years, was undertaken, eventually resulting in the splitting of the collection across three sites only one of which, the original, but now much improved Saint Martin buildings, remained open to the general public. The vast majority of the 80,000 artifacts of the collection were moved to a purpose-built repository at St-Dennis on the northern outskirts of Paris (figure 2). It is here that the prototype Hollerith machine is to be found.
The collection quickly overflowed the new repository and a number of items were moved to warehouses not far from the reserves in St Dennis (figure 3) where they languish, almost forgotten. The CNAM collection continues to grow and the warehouses continue to fill, not least in consequence of the recent decision to preserve the scientific and technical heritage of the past 50 years.
The authors wish to express their gratitude to the Conservatoire des Arts et Métiers for permission to use the photographs which appear in this article and to Stephane Mathon for his permission to include Figure 1.
Figure 2 from Library of Congress, Prints and Photographs Division [reproduction number, LC-USZ62-106523].
All other photographs ©David Anderson 2005
1 Austrian G.D. “Herman Hollerith: Forgotten Giant of Information Processing” Colombia University Press 1982 p.47
2 Ibid p.47
3 Ibid p.49
5 The authors are grateful to the Curator of Mathematics at the Smithsonian Institute, Peggy A. Kidwell, for providing details concerning the Hollerith machine held by the Smithsonian.
6 “There will be established in Paris, under the name of Conservatoire National des Arts et Métiers, ... a depot of machines, models, tools, designs, descriptions and books of all aspects of industrial arts and trades”
7 Vaucanson was made a member of the Académie des Sciences in 1746. Following his death in 1782, Vaucanson bequeathed the collection of his works to Louis XV and it was this collection that was absorbed by the CNAM in 1794.
8 It may well be that Hollerith became aware of Jacquard’s loom through his brother-in-law Albert Meyer who was in the silk-weaving business and later became on of Hollerith’s financial backers.