At the International Conference on the History of Computing held in Los Alamos in 1976, R.W. Hamming placed his proposed agenda in the title of his paper: "We Would Know What They Thought When They Did It." He pleaded for a history of computing that pursued the contextual development of ideas, rather than merely listing names, dates, and places of "firsts". Moreover, he exhorted historians to go beyond the documents to "informed speculation" about the results of undocumented practice. What people actually did and what they thought they were doing may well not be accurately reflected in what they wrote and what they said they were thinking. His own experience had taught him that.
Historians of science will recognize in Hamming's point what they learned from Thomas Kuhn some time ago, namely that the practice of science and the literature of science do not necessarily coincide. Paradigms (or, if you prefer with Kuhn, disciplinary matrices) direct not so much what scientists say as what they do. Hence, to determine the paradigms of past science historians must watch scientists at work practicing their science.  We have to reconstruct what they thought from the evidence of what they did, and that work of reconstruction in the history of science has often involved a certain amount of speculation informed by historians' own experience of science.
But Hamming's point has special force in the history of technology, as I learned several years ago when I first started teaching the subject. As an historian of science I had been accustomed to teaching from primary sources, that is, the works of the scientists of the period under study. Hence, in drawing up a syllabus I cast about for the primary sources of technology since the Renaissance. I had great difficulty finding them. Indeed, I never did find them. I stopped looking when it dawned on me that I was looking for the wrong thing in the wrong place. What I needed for my students was not a library, but a museum. They shouldn't be reading great books, but examining great things. Or, to put that last point another way, the great ideas we were seeking did not lie in books. They lay in objects. Understanding those ideas meant learning to "read" in a new way.
That notion took some time to penetrate my literary instincts, but the more I have thought about it the less novel or strange it has seemed to me. Humanists of various stripes have long been using objects as "texts". For example, neither music nor the visual arts is essentially a literary activity, and so we listen and we look, and we use written texts to inform and support our listening and looking. Historians have long been doing something similar. First, they have found that, when read in certain ways, documents drawn up for day-to-day purposes, with no eye toward history or toward saying anything other than what they contained prima facie, in fact reveal much about the people who wrote them and about the societies that used them as instruments. Legal and social relations, economic conditions, famine and pestilence, literacy and family life: information about these things have emerged from inventories, tax rolls, censuses, parish registers, bills of sale and lading, and so on. In particular, historians learned to look for what was not said in such documents but simply taken for granted.
Among the things often taken for granted were the technical exigencies and preconditions of the events and transactions being recorded. When mounted armor rides into the records, its appearance bespeaks new levels of know-how in the control of horses, in the working of metal, and in the organization of men. When people divide up their fields in new ways and when they begin to raise certain crops, then the protein level of their diet is changing. Here interpretation rests heavily on inference and indirect documentation. The Middle Ages was one of the most creative periods in the history of technology, yet the creativity is cloaked in anonymity, and the technology is hidden by silence. The historian looks about for chinks of light, and usually the machines and processes themselves (or at least illustrations of them or passing descriptions of them) provide those glimpses. It is rather like finding the question given the answer.
Yet that is a negative way of putting it. It suggests that, if only craftsmen or (later) engineers had written about what they were doing, we would not have to rely on artifacts viewed either first or second hand. And hence we look forward to the appearance in the Renaissance of engineering literature and to the growing literacy among technicians in the early modern era so that we can get back to more familiar ways of doing history. Viewing the artifacts as substitutes for missing texts puts the historian on the track of ignoring the artifacts when the texts become available.
Along that track lies the danger of derailment. For technology is not a literate enterprise. Certainly until the present century, design and invention proceeded largely by trial and error, and the thinking that they involved went on not in words but in visual and tactile patterns and relations. The extent to which the wedding of science to technology has acted to transform that thinking is an interesting and important question for historians of modern engineering. But whatever its answer, it should not be projected into the past. The historian of technology will not discover what inventors and builders had on their minds by reading what they wrote. At best, their writings serve as guides to understanding the real expressions of their thoughts, namely their works.
It is important to recognize that the artifacts of technology -- I mean, of course, the products of conscious design by inventors and engineers-- embody thinking, different in kind, perhaps, from what humanists are used to, but not in degree. Anthony F.C. Wallace makes the point quite well in a digression in the middle of Rockdale:
The complexity of thought required to understand mechanical systems would seem in no way inferior to what is required for the trains of reasoning in mathematics or the common language. Thinking visually or tactilely has an inherent disadvantage, however, in comparison with thinking in language. Those who think in words --on subjects which are thought about effectively in words-- can think a sentence and then utter it for others to hear. If one visualizes a piece of machinery, however, and wishes to communicate that vision to others, there is an immediate problem. Speech (and writing) will provide only a garbled and incomplete translation of the visual image. One must make the thing --or a model, or at least a drawing-- in order to ensure that one's companion has approximately the same visual experience as oneself.
He goes on immediately to bring out a fateful implication of this problem.
In the Western world, an effect of this special problem in communicating technological information has tended to be the growing isolation of those who think in mental pictures. Theologians, humanists, even scientists can converse freely because the thinking is done with the same system of symbols as those used in communication. Indeed, it has become conventional to assume that thought itself is merely a kind of internal speech and to disregard almost completely those kinds of cognitive processes that are conducted without language, as though they were somehow more primitive, and less worthy of intellectual attention. Those who think about machinery have tended to undervalue their own accomplishments, or to deny that the process is intellectual at all, and to belittle "intellectuals" in turn.
More is involved here than social values. To think about a machine --I here use the term generically for the products of technology-- means ultimately to think about its use in society: what it will be used for, who will use it, what its use will require of the user, how it will fit in with other machines, and so on. To think about those things is to think about society and the humans who constitute it. Disdain for visual and tactile modes of thinking can mean ignorance of ideas that vitally affect the human condition.
Daniel C. Calhoun recognized this in his provocative attempt to gauge the intelligence of Americans in the nineteenth century. After examining the direct evidence of educational theory and practice (needless to say, a highly verbal record), he looked for indirect evidence in the demands that three major activities --preaching, shipbuilding, and bridgebuilding-- made on their participants, both actors and audiences. In shipbuilding he found that the record of American innovation in the design of hulls lay neither in equations nor in blueprints, but rather in models. "We can most of us whittle better than we can draw," observed Norman S. Russell in the Transactions of the Institution of Naval Architects in 1861. New designs emerged from the collaboration of hand and eye working in three reduced dimensions, and the result was transmitted to artisans through the models, often sliced in parallel sections and sized upward to full scale by a few proportions and another trained eye. What designers thought ships were for, who would man them, what they would carry, how well they would behave: all these ideas were carried in the hand-carved shape of a hull, rather than in the literature of sailing.
Consider in this light two passages in Henry Ford's My Life and Work:
I do not consider the machines which bear my name simply as machines. If that was all there was to it I would be doing something else. I take them as concrete evidence of the working out of a theory of business which I hope is something more than a theory of business --a theory that looks forward toward making this world a better place in which to live.
There is an immense amount to be learned simply by tinkering with things. It is not possible to learn from books how everything is made --and a real mechanic ought to know how nearly everything is made. Machines are to a mechanic what books are to a writer. He gets ideas from them, and if he has any brains he will apply those ideas.
Few things have so shaped the human experience in the twentieth century as have Henry Ford's machine and the theory of business it embodied. As Robert and Helen Lynd showed in their study of Middletown (1927), no realm of human relations, values, and aspirations remained untouched by the automobile; it loomed large in the very image people had of themselves. Much has been written about it. Henry Ford himself wrote six books. Yet, from the man himself we have the admonition that the central text is the machine itself. That is where we shall find his ideas.
Let's see if we can read the Model T, then, and do so on two levels: as a machine in itself and as a component of a mechanized process of production. Ford's theory of business encompassed both.
Obviously, it would be best if I could conjure up before the reader a full-scale Model T and a tool kit and begin to take the vehicle apart. Even when lecturing in person, however, the best I can muster is a scale model which is reasonably accurate in detail, including the way it goes together. Still, it is not the real thing. Nor would the real thing alone suffice. We today are not the audience for whom the Model T was conceived. For the most part, we do not know the things Ford expected his customers to know, and, unaided by contemporary evidence, we may not see what he intended them to see and to appreciate.  So we must turn to other evidence in the form of advertisements and of the Model T owner's manual. Such may hardly seem the stuff of humanistic inquiry, yet it is a reader's guide to the text. Similar things may be said of the guide to the process of production, Horace L. Arnold and Fay L. Faurote's classic 1915 study of Ford Methods and the Ford Shops. and Fred M. Colvin's series of articles in Engineering Magazine in 1913.
Like Edison's electric light, Ford's Model T originated in an analysis of cost. In 1905 Ford decided that, if the automobile were ever to become more than a toy or hobby for the rich, it would have to be available to the nation's lawyers, doctors, small businessmen, and well-to-do farmers. Making around $2500 a year, such people would be willing, he estimated, to spend $500 on a car, provided that it met their needs and demands.
But the design that met their needs and demands had to meet Ford's as well. Simply put, it had to be a design he could profitably produce for $500. The Model T, introduced in 1908, initially fitted only the first part of the equation: it was the design its intended users wanted, but it cost $850. Completing the equation meant extending the design into a system of production. By the middle of 1914, the first year of full assembly-line production, the touring car cost $490; at the end of 1924, the year in which the ten-millionth T was built, the price reached a low $290. When the fifteen-millionth T came off the line at the end of 1926, it cost $380, but the price included a self-starter and balloon tires. Except for such minor changes, the T of 1926 was essentially that of 1908. The system of production was in fact an extension of the car's original design. That is what Ford meant when he called the vehicle a "working out of a theory of business."
What were the main features of the machine's design? It was a "standard" or "universal" car. It had four main components, each of which had only one form: there was only one engine and transmission, only one chassis, only one front axle, and only one rear axle. Each component constituted Ford's best solution to the mechanical problems involved. "The universal car," he later wrote,
had to have these attributes: (1) Quality in material to give service in use, (2) Simplicity in operation --because the masses are not mechanics, (3) Power in sufficient quantity, (4) Absolute reliability --because of the varied uses to which the cars would be put and the variety of roads over which they would travel, (5) Lightness, (6) Control, (7) Inexpensive operation [largely because of lightness].
In his listing Ford gave pride of place to the materials he used, foremost among them vanadium steel  for the moving parts of the engine and transmission. He came across the alloy literally by accident, while scouting in the wreck of a French racing car in 1905. To secure the steel he had to guarantee the existence of the one company in America at the time that could produce it. (That is a point worth bearing in mind. Its importance will become clear below when we consider the machine of production.)
Tough steel meant a car that would not break down from internal strains, especially strains in the transmission. Here reliability, control, and simplicity interacted. Early sliding-gear transmissions required considerable strength and dexterity to get proper mesh without damaging the teeth. That, perhaps more than anything, had placed the automobile in the hands of male chauffeurs . Ford's planetary transmission kept all the gears engaged at all times; which ones actually turned the drive shaft depended on bands surrounding the outer rims of the combinations. When the band on the braking gear was tightened by the brake pedal, the transmission became a brake on the wheels. Requiring neither strength nor dexterity, the planetary transmission made everyone a chauffeur, everyone including the nation's women. "Anybody can drive a Ford," went the saying. The design of the transmission was one reason. Ease of steering was another.
But simplicity had another meaning for Ford. He meant to build a car that anybody could fix. Each of the four components was easily accessible, he later pointed out,
and they were designed so that no special skill would be required for their repair or replacement. ...The parts could be made so cheaply that it would be less expensive to buy new ones than to have old ones repaired. They could be carried in hardware shops just as nails or bolts are carried.
Here is where a process of mechanical design involves a concept of society. Despite what Ford said about the masses not being mechanics, the Model T as a machine rested on the supposition of mechanical skill in its owner. Simplicity was part of reliability. The car might break down, but when it did you could fix it: you yourself, on the spot. Its design facilitated access to every component. The hood lifted away completely from the top and sides, the engine compartment was completely open underneath, a removable panel on the floor gave access to the transmission to allow adjustment of the bands. That concept of design tells us much about the Model T, but it tells us even more about the people --the many, many people-- for whom Ford intended the car. He counted on their mechanical skill, so much so that he never came right out and said it. The hardware store was ubiquitous in America, especially in rural America. To place automobile parts there, next to nuts and bolts, was to place the car owner at home in the store.
So it is that the T came with an owner's manual that took as a matter of course the trimonthly reseating of the valves, a process that involved removing the cylinder head (which is one reason why the T's engine had a removable cylinder head), lifting the valve heads out of place, and using a grinding tool and paste. Indeed, there is nothing about the mechanics of the T that the manual thought lay beyond the capacities of the owner. A readily available Parts List contained pictures of every component, keyed by number to the stock code for ordering. Owners used to fixing machinery but not versed in the terminology of engines, drives, and chassis could identify the broken part they held in their hand by visual matching; they did not have to try to describe it verbally. The reader should look at the illustrations taken from the manual and the parts list and compare them with the manuals that accompany today's new cars. Better still, readers might ponder how well they would do if they owned a T and had to use its manual.
Ford was not alone in presupposing mechanical skills in Americans. Myriads of small companies also banked on those skills. Things did go wrong with Fords: crankcase supports cracked, the engine could be hard to start in cold weather, the electrical system occasionally failed, seals leaked, and parts simply wore out. Nor was everyone satisfied with Ford's mechanical tastes. The T invited tinkering: different transmissions for the hardy, different bodies for the racy, gas gauges for the timid, and so on. The design made all this possible because it was part of the design. The mechanically minded historian can see that in the car itself; in its absence, the manual and the advertisements tell the same tale.
The Model T came apart easily for other reasons besides repairs. It was designed as a utility vehicle. Its popularity as a tractor (with suitable rear wheels) prompted Ford ultimately to design the Fordson tractor, which was built in large numbers both here and abroad, especially in Soviet Russia. Jacked up and shed of its rear wheel, the Model T provided a power plant for a mechanical saw or a water pump. Its front wheels might give way to skis, its single rear axle to a bogey with cater pillar treads. Its uses appear to have been limited only by the imagination of the owner.
Lightness (the car weighed only 1200 pounds) contributed to that versatility, as it did to the car's economy of operation and to its adaptability to extreme conditions. The Model T was designed for unimproved roads --or no roads at all. It bounced its way out of some trouble, and four strong men could lift it out of other trouble. Other features of the design make sense in this context: the high clearance, the triangular suspension which allowed the front and rear axles to pivot independently of the chassis, the steel crankcase and transmission cover which could withstand rocks and stones. Popular as this car may have been in the towns and cities of America, its design shows the setting for which Ford intended it. It belonged on the farm. It was, after all, he once summarized, just a one-horse shay with a motor in it. 
This last point warrants further discussion, but that discussion in turn requires an examination of the Model T as a component of a larger machine. Several of the features discussed so far are pertinent. Durability calls for special alloys, simplicity involves readily available spare parts, reliability and versatility mean ease of assembly and disassembly. Consider some other features more closely. The cylinder block is cast in a single piece enclosing the camshaft and valve rods and surrounding the upper halves of the cylinders with a water jacket. (Early automobiles usually had machined cylinders bolted individually or in pairs to the crankcase and left the lifters exposed.) The steel pan covering the bottom of the crankcase and transmission is stamped, not forged.
On the one hand, such features mark the T as a cheap car. What the single-cast block gains in serviceability, it loses in refinement. One does not expect to have to pay much for machinery of that sort. On the other hand, such features are quite expensive to produce. One needs a foundry, not a machine shop. One needs stamping presses and annealing furnaces, not blacksmiths and forges. Designing a machine that anyone can fix because it is easy and cheap to replace parts means having (or being ready to acquire) the capacity to turn out large numbers of wholly standardized, interchangeable components and to turn them out while using a portion of them to construct the original machine itself. In short, the cheap car is not cheap. As noted above, Ford had to guarantee a company's survival to get his vanadium steel.
Or rather, the car is cheap only if it is produced in volume. And the only way to produce in volume is to produce by machines. That is an essential meaning of the Model T read as a technological text: it is a machine built by machines. Its design makes sense only if it is built by machines. Hence, to have designed it is to have had in mind a machine-based system of production, in scheme if not in detail. Indeed, in 1908 Ford had only a scheme. The assembly line was not in place until the end of 1913; in concept as well as in time, it was the last element of Ford's system of production.
To appreciate the role of machine-building machines in Ford's design, one need only look at the stages an engine block passed through on its way from the foundry to the assembly room. No fewer than twenty-eight machines were involved, each carrying out a specific task. In 1914, a block took forty-five minutes (exclusive of travel time) to pass through the process. The design of the Model T was built into those machines, almost all of them built to Ford's specifications, almost half of them of his and his staff's original design. Although the Model T appears to have changed very little between 1909 and 1927, it in fact changed considerably. The changes, most of them reflected in the car's decreasing price, occurred in the machinery that built the car.
Can one read that machinery, both in its parts and as a whole? Yes. A cursory glance alone makes one thing clear. These are not tools. That is, they are not devices conceived of or acting as extensions of human skills or capacities. In the paragraph just preceding the two quoted from Rockdale, Wallace speaks of the "grammar of the machine or mechanical system." Human actions determine the grammar of tools, even tools large enough to be mechanized. Mechanical actions determined the grammar of the machines at work in Ford's shops. 
Early textile machinery offers an example of this notion. The first mechanical spinning machine was the jenny. It worked in imitation of the human actions of spinning, just as had the spinning wheel, which in essence was a tool for twisting the roving while the operator drew it out between thumb and forefinger, controlling its fineness. Basically, the jenny put many threads in the spinner's hand, connecting them from individual spindles to a common bar which the operator drew back by means of a lever or a wheel. But the frame took a mechanical path to the goal of twist-and-draw. It drew out the roving by passing it through three pairs of rollers, each pair rotating at a faster rate than its predecessor and hence pulling on the yarn. The mechanism took up little space and hence allowed a significant increase in the number of threads a single machine could spin. That increase in turn meant a machine so large that it required a water wheel (whence the name "water frame") or a steam engine to power it.
One ordinarily thinks of the machine's size as the factor that took its direct operation out of the worker's hands. Actually, it was the drawing mechanism that removed the machine from the realm of tools that extend human skills. Human hands cannot draw yarn that way. Hence, the operator does not use the frame to exercise his skill; he supplies the machine while it carries out its function. That is the real meaning of the replacement of human skill by mechanization. The transition between jenny and frame is the transition between spinner and machine tender. Historically, it was also the transition between male and female operatives in the early textile factories.
None of the machines in the stages of finishing the cylinder block of a Model T imitated human actions. Each took full advantage of the grammar of mechanical actions, and what it did determined the next stage in the sequence. So too the car as a whole went together in a syntax determined by mechanical parsing.  As Ford himself put it:
In the chassis assembling are forty-five separate operations or stations. The first men fasten four mud-guard brackets to the chassis frame; the motor arrives on the tenth operation and so on in detail. Some men do only one or two small operations, others do more. The man who places a part does not fasten it --the part may not be fully in place until after several operations later. The man who puts in a bolt does not put on the nut; the man who puts on the nut does not tighten it. 
So wrote the man who designed a car for people at home in a hardware store, where nuts and bolts naturally go together with a wrench to tighten them.
It is a matter here neither of contradiction nor of perversity, but rather of irony arising out of the unforeseen consequences of creativity and innovation. The design of the Model T was predicated on both the skills of its owner and the mechanization of its production. Together they made the concept feasible. As a machine the Model T expressed Ford's idea of a small, cheap, serviceable car that would foster the independence of its owner, especially its rural owner. The aspiration matched his own experience as a farm-bred, trained machinist undertaking an independent business venture in the mythic American way. The Ford Motor Co, capitalized initially at $100,000, took in fact only $28,000 cash to get it started.
However traditional the aspirations behind it, the car could not be built by traditionally skilled workers. It would have been too expensive for the sort of car it was, and even then the workers would have received a pittance for their labors. To be cheap, the T had to be produced in large numbers by few people. Without a breakthrough in productivity, more workers would not help. Without the new machinery of production, no number of men, however skilled, could have produced in sufficient volume and to requisite standards of precision the original components and spare parts required to build the cars and keep them running.
Hence, Ford's design entailed a system of production that measured its capital costs in the millions and its workers in the tens of thousands. It paid those workers an unprecedentedly high salary, while demanding of them only that they do precisely what they were told to do and do it for eight hours.  Mechanized production required mechanical skill not from those who assembled the cars but from those who designed and arranged the machines that built the cars. However, that was mechanical skill of a new kind. Although Ford required some machinists of the traditional sort, he had even greater need of engineers and managers, who before long were coming to him not up from the shop floor but out of colleges and business schools. 
Thus, Ford's idea when fully expressed drove little people like the Henry Ford of 1905 out of the market. Where $28,000 had sufficed, $5,000,000 would scarcely reach. People with Ford's inventive talent less often struck out on their own and more often went to work for others. Invention itself acquired a name better suited to its new setting: research and development. As the nation adopted the automobile, new generations of Americans began to lose touch with the mechanical skills Ford had assumed among his customers. His form of industrial production did not demand them. In 1927 Ford replaced the T with a car that did not demand them either, the Model A.
In basic outline, much of this is a familiar story, and one may ask the purpose of pursuing it in such detail. The detailed reading is prerequisite to understanding the cultural or human meaning of the technology that put Americans on the road. Only by reading the machine can one appreciate the assumptions and choices that were built into it. They determined a technical system that, once in place, foreclosed some choices and redefined others. Although the Model T could well have failed (it was an immensely courageous venture), it could not have been only moderately successful. Henry Ford did not design a car, find that a lot of people wanted to buy it, and then figure out a way to make more cars, as in a series of independent steps. That is, of course, the classic American entrepreneurial myth. Yet the design of the car belies the myth. A careful look at its structure and components, at the unchanging specifications of parts over a decade or more, at the implicit relation between design and production shows that the first Model T contained in embryo the millions of Fords that forged roads across America in the 1920s, together with the system of production that, as it spread from Ford's plant throughout the automotive industry, brought Americans off the farm and into the factory. 
The point to emphasize here is that one must look. As Ford said, books do not suffice to tell us how things are made and work. Mechanics must learn to read machines. So too must historians of technology, and indeed historians of science. The technical and cultural meaning of the Model T lie in the automobile as an artefact. That is where Ford expressed his ideas, addressing himself to an audience he clearly understood well. To understand his audience and the meaning they attached to the car, historians must see what they saw and feel what they felt, for their experience of the car was visual and tactile, not verbal or literary. Reading a machine means determining what the artifact says about the people who designed it, the process of its design, the assumptions made about its purposes, the expectations held of its putative users, and the ways it could actually be used. Most interesting and revealing are the points at which these overlapping questions do not have coinciding answers. For it is there, rather than at the level of theory, that the dialectic of technology is carried out.