The Military Origins of Quality Control
While reading Foucault, I found myself constantly asking, "by who?" Foucault's themes include the regulation and disciplining of people, but he never seems to say who is driving the trend. It was surreal - like reading a history of Europe written by Jim Garrison. Garrison, you may recall, was the real-life prosecutor played by Kevin Costner in Oliver Stone's JFK. According to his book, upon which the movie was largely based, the conspiracy included Castro-haters, the New Orleans mob, and dozens of other actors all the way up to Johnny Carson of Tonight Show fame. I'd guess over half the country must have been involved (in which case the assassination was merely a recall vote).
But I digress.
One of Foucault's themes was that the regularization of people for economic production began in the 18th century military, specifically the French army. In Discipline and Punish (from The Foucault Reader, p. 210), he says,
Surprisingly, that is also where my tale of Quality Control begins. Quality control and mass production had its origins in the French Army, jumped to North America in the Revolutionary War and its aftermath, and finally culminates in Fordism and Taylorism. At that point, it gave birth to the statistical science of QC in the 1930s which made its way to Japan in the post-war era. WWII also gave birth to continuous improvement, which also migrated to Japan in the post-war era. Both of these movements came back to the US in part through the US Navy, though also clearly through the automobile industry. Finally, some of the ideas which had taken hold in the military culminated in a document which - though abandoned by the US Army - became influential on European industry and finally became the bureaucratic nightmare known as ISO 9000.
For the first part of this, I will draw largely upon David Hounshell's excellent From The American System to Mass Production: 1800-1932. In it, he describes the French origin of the idea that if the parts of weapons were made interchangeable, they would be more economical to make to make and easier to repair. According to Hounshell, French General Jean-Baptiste de Gribeauval "has been convincingly identified as the principal originator of this plan, which in fact had been known for a long time in France as "le système Gribeauval." It seems that Gribeauval had provided patronage to a certain Honoré Blanc to attempt to produce uniform musket locks. Ultimately, he never succeeded, but his friend, an American by the name of Thomas Jefferson, sent Blanc's ideas and memoirs to War Department Secretary Henry Knox. A second channel for the introduction of the idea was through French engineer, artillerist, and de Gribeauval student Major Louis de Tousard. After serving under Lafayette in the Revolution, Tousard joined the Corps of Artillerists and Engineers where he began teaching. He eventually wrote two books: one a proposal for a military school which he sent to the new Secretary of War, James McHenry, who used it as the "blueprint" for West Point, and the other on military principals which became the standard textbook for military officers. Hounshell says, "The importance of Tousard's book, as well as his informal teaching of officers in the Corps of Artilllerists and Engineers, cannot be overemphasized." He summarizes,
The other side of this story begins when W. Edwards Deming learned of statistical process control from Walter Shewhart. After applying the methods to the 1940 census, Deming was brought to Japan by Gen. Douglass Macarthur's Japanese Occupational Force to help with the 1951 census. While there, he was invited to teach statistical process control methods to the war torn economy there. What is little appreciated, however, is that Macarthur also had access to legions of instructors from the Training Within Industry (TWI) service to help teach modern methods to the Japanese.
TWI was a program created by the War Manpower Commission of the War Department for the purpose of helping manufacturers cope with the fact that they were being asked to ramp up production at the same time many of their employees were being enlisted or drafted into the Armed Forces. TWI consisted of 4 programs: Job Instruction, Job Methods, Job Relations (and another version of this for union officials), and Program Development. Together, these taught what would be recognizable today as standard work and continuous improvement or kaizen. That the Japanese programs brought back into vogue in the US had an American origin is not well known, but also beyond dispute. Maazaki Isai says in Kaizen: The Key to Japan's Competitive Success, "It is well known that the initial concepts of statistical quality control and its managerial implications were brought to Japan by such pioneers as Deming and Juran in the postwar years. Less well known is the fact that the suggestion system was brought to Japan about the same time by TWI (Training Within Industries [sic]) and the U. S. Air Force. In addition, many Japanese executives who visited the United States right after the war learned about the suggestion system and started it at their companies." In the intro to Donald Dinero's Training Within Industry: The Foundation of Lean, John Shook writes about a time when he was working for Toyota when he "protested to my Japanese colleague, declaring that the program as configured just wouldn't do and required radical revision before being unleashed on the NUMMI workforce." His colleague, Toyota Master Trainer Iaso Kato, "stormed out and fetched from a back room file a yellowed, dog-eared, coffee-stained copy of the English-language original training manual, just as he had received it .... To my absolute amazement, the program that Toyota was going to great expense (including retranslating from Japanese to English) to "transfer" to NUMMI was exactly what the Americans had taught the Japanese decades earlier. Of course, it was JI, the Job Instruction module of TWI. Toyota still used it in 1984 and continues to use it today ...." These same points are made in Jim Huntzinger's article, "The Roots of Lean; Training Within Industry: The Origin of Kaizen ". To bring it full circle, the American Society for Quality says that the term "Total Quality Management" was first used by the U.S. Naval Air Systems Command around 1984 "to describe its Japanese-style management approach to quality improvement." (incidentally, most of this paragraph is plagiarized from my own edits from Wikipedia, so I will warn anyone trying to confirm these with an independent authority that those entries in Wikipedia are highly influenced by my own research ~~~~).
The final promised tale is that of the evolution of Army quality specification MIL-Q-9858 (1958). This standard was soon adopted by NATO as AQAP-1 (1969). That in turn was adopted by the British Standards Institute as BS 5750 in 1979, which in turn became the basis of the original ISO 9000 standard in 1987. Thus we have a direct line of descendancy from an American military standard toa protectionist European an international industrial standard in about 30 years. From my understanding, it is just about impossible to fail a certification inspection because they are done by for-profit consultants. As a result, the automotive industry has established their own variant, QS 9000, because they need actual standardization and not lots of paperwork and paper tiger certifications.
Upon realizing that history, my initial response was - in Greg Mankiw's terms - to update my priors in response to this information. Confirmation bias should push me to believe "market good, government bad", but to the extent that I think quality control is desirable (and there are good reasons to believe so: for one thing, quality control means less waste, a very green attitude and confirmed by the chapter on lean production in Natural Capitalism) I have more data to support an alternate hypothesis, "markets not perfect, government usually not always bad". My second response was to realize that the world is both more complex and less complex than normally thought: more complex in the sense that there was global influence long before the present "globalized" era, since these ideas originated in France more than 200 years ago, were perfected in the US between 120 and 100 years ago, were exported to Japan 50 years ago, and are currently being reimported from there; and less complex in the sense that Americans who claim credit for the Industrial Revolution are probably less wrong than their supposedly more cosmopolitan and less chauvinistic neighbors think them to be. My third response was to start to rethink my earlier easy dismissal of Foucault. My fourth was to start integrating this knowledge with what I have gleaned out of Human Scale: that mass production has gone hand-in-hand with the rise of the militarism and the centralized state. Correlation, not causality, but curious correlation nonetheless.
technorati tags: "W. Edwards Deming" "kaizen" "TWI" "Training Within Industry" "ISO 9000" "Michel Foucault" Foucault
But I digress.
One of Foucault's themes was that the regularization of people for economic production began in the 18th century military, specifically the French army. In Discipline and Punish (from The Foucault Reader, p. 210), he says,
If the economic take-off of the West began with the techniques that made possible the accumulation of capital, it might perhaps be said that the methods for administering the accumulation of men made possible a political take-off in relation to the traditional, ritual, costly, violent forms of power, which soon fell into disuse and were superseded by a subtle, calculated technology of subjection. In fact, the two processes -- the accumulation of men and the accumulation of capital -- cannot be separated [...] The disciplinary pyramid constituted the small cell of power within which the separation, coordination, and supervision of tasks were imposed and made efficient; and analytical partitioning of time, gestures, and bodily forces [EH: Taylorism?] constituted an operation schema that could easily be transferred from the groups to be subjected to the mechanisms of production; the massive projection of military methods onto industrial organization was an example of this modeling of the division of labor following the model laid down by the schemata of power. [emphasis added]For readers referred here from the Mutualist Blog, it is worth pointing out that Hounshell credits Lewis Mumford of having traced similar themes.
Surprisingly, that is also where my tale of Quality Control begins. Quality control and mass production had its origins in the French Army, jumped to North America in the Revolutionary War and its aftermath, and finally culminates in Fordism and Taylorism. At that point, it gave birth to the statistical science of QC in the 1930s which made its way to Japan in the post-war era. WWII also gave birth to continuous improvement, which also migrated to Japan in the post-war era. Both of these movements came back to the US in part through the US Navy, though also clearly through the automobile industry. Finally, some of the ideas which had taken hold in the military culminated in a document which - though abandoned by the US Army - became influential on European industry and finally became the bureaucratic nightmare known as ISO 9000.
For the first part of this, I will draw largely upon David Hounshell's excellent From The American System to Mass Production: 1800-1932. In it, he describes the French origin of the idea that if the parts of weapons were made interchangeable, they would be more economical to make to make and easier to repair. According to Hounshell, French General Jean-Baptiste de Gribeauval "has been convincingly identified as the principal originator of this plan, which in fact had been known for a long time in France as "le système Gribeauval." It seems that Gribeauval had provided patronage to a certain Honoré Blanc to attempt to produce uniform musket locks. Ultimately, he never succeeded, but his friend, an American by the name of Thomas Jefferson, sent Blanc's ideas and memoirs to War Department Secretary Henry Knox. A second channel for the introduction of the idea was through French engineer, artillerist, and de Gribeauval student Major Louis de Tousard. After serving under Lafayette in the Revolution, Tousard joined the Corps of Artillerists and Engineers where he began teaching. He eventually wrote two books: one a proposal for a military school which he sent to the new Secretary of War, James McHenry, who used it as the "blueprint" for West Point, and the other on military principals which became the standard textbook for military officers. Hounshell says, "The importance of Tousard's book, as well as his informal teaching of officers in the Corps of Artilllerists and Engineers, cannot be overemphasized." He summarizes,
Thomas Jefferson's enthusiasm for Honoré Blanc's experiments with the manufacture of interchangeable musket parts and the influence on the American military of the rationalism of General Gribeauval and his followers firmly established the intellectual and institutional basis for the rise of the American system of arms production. The pure rationalism of "system and uniformity" provided an adequate incentive for the pursuit of this goal. The United States War Department soon found the idea of interchangeability irresistible, and through its own armories and through private arms contracts it encouraged and supported attempts to achieve this end. Eventually the War Department demanded interchangeability. Ordnance officers elevated the idea of interchangeability to an ideal and helped to transform it into a reality. [emphasis in original]Still, despite legends about contractor Eli Whitney's invention of the method (as evidenced by the Wiki entry, which I will soon correct), Whitney himself never actually achieved it either. Arms manufacturers worked at the idea at the federal armories in Springfield and Harper's Ferry without much success. Finally, an inside contractor machinist named John H. Hall worked out the final solution in the 1820s. Drawing on the work of his predecessors, especially Simeon North and his use of a master model, Hall worked out a system of gauges and jigs or "rational fixtures". The real trick was to keep close watch on those gauges to make sure they are still in specification or tolerance. We can see in this the basic science of statistical quality control, the science of figuring out how much error is acceptable. Hounshell takes this story all the way to the pinnacle of mass production, the Ford system for manufacturing the Model T. In the remainder of the story, it is astounding how many companies thought to have developed components of mass production - especially Singer and McCormick - first brought in machinists trained in one of the arms manufacturers to show them how to obtain regularity.
The other side of this story begins when W. Edwards Deming learned of statistical process control from Walter Shewhart. After applying the methods to the 1940 census, Deming was brought to Japan by Gen. Douglass Macarthur's Japanese Occupational Force to help with the 1951 census. While there, he was invited to teach statistical process control methods to the war torn economy there. What is little appreciated, however, is that Macarthur also had access to legions of instructors from the Training Within Industry (TWI) service to help teach modern methods to the Japanese.
TWI was a program created by the War Manpower Commission of the War Department for the purpose of helping manufacturers cope with the fact that they were being asked to ramp up production at the same time many of their employees were being enlisted or drafted into the Armed Forces. TWI consisted of 4 programs: Job Instruction, Job Methods, Job Relations (and another version of this for union officials), and Program Development. Together, these taught what would be recognizable today as standard work and continuous improvement or kaizen. That the Japanese programs brought back into vogue in the US had an American origin is not well known, but also beyond dispute. Maazaki Isai says in Kaizen: The Key to Japan's Competitive Success, "It is well known that the initial concepts of statistical quality control and its managerial implications were brought to Japan by such pioneers as Deming and Juran in the postwar years. Less well known is the fact that the suggestion system was brought to Japan about the same time by TWI (Training Within Industries [sic]) and the U. S. Air Force. In addition, many Japanese executives who visited the United States right after the war learned about the suggestion system and started it at their companies." In the intro to Donald Dinero's Training Within Industry: The Foundation of Lean, John Shook writes about a time when he was working for Toyota when he "protested to my Japanese colleague, declaring that the program as configured just wouldn't do and required radical revision before being unleashed on the NUMMI workforce." His colleague, Toyota Master Trainer Iaso Kato, "stormed out and fetched from a back room file a yellowed, dog-eared, coffee-stained copy of the English-language original training manual, just as he had received it .... To my absolute amazement, the program that Toyota was going to great expense (including retranslating from Japanese to English) to "transfer" to NUMMI was exactly what the Americans had taught the Japanese decades earlier. Of course, it was JI, the Job Instruction module of TWI. Toyota still used it in 1984 and continues to use it today ...." These same points are made in Jim Huntzinger's article, "The Roots of Lean; Training Within Industry: The Origin of Kaizen ". To bring it full circle, the American Society for Quality says that the term "Total Quality Management" was first used by the U.S. Naval Air Systems Command around 1984 "to describe its Japanese-style management approach to quality improvement." (incidentally, most of this paragraph is plagiarized from my own edits from Wikipedia, so I will warn anyone trying to confirm these with an independent authority that those entries in Wikipedia are highly influenced by my own research ~~~~).
The final promised tale is that of the evolution of Army quality specification MIL-Q-9858 (1958). This standard was soon adopted by NATO as AQAP-1 (1969). That in turn was adopted by the British Standards Institute as BS 5750 in 1979, which in turn became the basis of the original ISO 9000 standard in 1987. Thus we have a direct line of descendancy from an American military standard to
Upon realizing that history, my initial response was - in Greg Mankiw's terms - to update my priors in response to this information. Confirmation bias should push me to believe "market good, government bad", but to the extent that I think quality control is desirable (and there are good reasons to believe so: for one thing, quality control means less waste, a very green attitude and confirmed by the chapter on lean production in Natural Capitalism) I have more data to support an alternate hypothesis, "markets not perfect, government usually not always bad". My second response was to realize that the world is both more complex and less complex than normally thought: more complex in the sense that there was global influence long before the present "globalized" era, since these ideas originated in France more than 200 years ago, were perfected in the US between 120 and 100 years ago, were exported to Japan 50 years ago, and are currently being reimported from there; and less complex in the sense that Americans who claim credit for the Industrial Revolution are probably less wrong than their supposedly more cosmopolitan and less chauvinistic neighbors think them to be. My third response was to start to rethink my earlier easy dismissal of Foucault. My fourth was to start integrating this knowledge with what I have gleaned out of Human Scale: that mass production has gone hand-in-hand with the rise of the militarism and the centralized state. Correlation, not causality, but curious correlation nonetheless.
technorati tags: "W. Edwards Deming" "kaizen" "TWI" "Training Within Industry" "ISO 9000" "Michel Foucault" Foucault
Labels: decentralization, police-state
<< Home