The Evolution of Useful Things: How Everyday Artifacts-From Forks and Pins to Paper Clips and Zippers-Came to Be as They Are.

by Henry Petroski

  The jugs themselves were often inscribed with mottoes and verses to taunt the drinker. For example, one jug read,

  From mother earth I took my birth,

  Then formd a Jug by Man,

  And now stand here, filld with good cheer,

  Taste of me if you can.

  Another offered:

  Here, gentlemen, come try yr skill,

  I’ll hold a wager, if you will,

  That you Dont Drink this liqr all

  Without you spill or lett some Fall.

  And still another put it this way:

  Gentlemen, now try your Skill,

  I’ll hold you Sixpence, if you will,

  That you dont drink unless you Spil.

  he variety of taunting verses demonstrates a range of literary solutions to the same verbal problem: to communicate a light-hearted challenge to the jug user. This nonuniqueness in the way language can transmit a single idea is also suggestive of how various forms can accomplish the same function. Indeed, the variety of verses inscribed on puzzle jugs was surpassed by the variety of jugs themselves. In addition to those with tubes protruding from every which place, there were jugs pierced through the middle, jugs with an inside tube passing from the handle down to the bottom, and jugs whose double sides included infundibular cores. The variety illustrates well that no unique form followed the single function of outwitting the drinker. Though it might be argued that the function of these vessels of practical joking was deliberately to have a deceptive form, the very fact that this could be accomplished in so many different ways serves to underscore the options available to designers, and the fun designers can have. Whereas a broad variety of solutions is generally not sought for production in a typical problem of product design, in the creation of puzzle jugs there was a definite premium on a puzzling array of forms. And their designers clearly had little trouble coming up with a bewildering variety of solutions to the same problem: how to trick the drinker into dribbling.

  Earthenware “puzzle jugs,” such as these two examples, were produced by the Wedgwood family in the late seventeenth century. These ale jugs were deliberately designed to be confusing to use and served as a basis for wagering in alehouses. The drinker would bet he could down the ale without spilling any, but to do so he had to cover up the right combination of holes and tubes, lest the jug behave more like a dribble glass. Had a unique form existed, the practice of wagering might not have been so popular. (photo credit 10.1)

  Not all artifacts are designed to trick the user, of course, and user expectations regarding form can actually constrain designers. By the end of the nineteenth century, the configuration of a standard bicycle, like the motorcycle of today, had achieved a rather mature form, which has not changed substantially since. The turn-of-the-century bicycle worked rather well in its context, and the kinds of modifications that have taken place generally have involved mechanical improvements in brakes, gearing, and tires rather than any dramatic alterations in the way the frame, wheels, handlebars, and seat fit together. This is not to say that the bicycle had evolved into a technologically predestined form, for cycling enthusiasts and designers have long found the old standard balloon-tired workhorse wanting in speed and efficiency (and have come up with designs that put the rider in positions ranging from recumbent to prone). Rather, the two-wheeler that we all might sketch if asked what the archetypal bicycle looks like is what has come to be the accepted and expected form of compromise among the competing things a bicycle is expected to provide: inexpensive, speedy, reliable, and relatively comfortable transportation that is faster than walking but less tiring than running.

  But nothing is perfect, of course, and one of the failings of a bicycle might be said to be its requirement that the rider be also its power. This is fine for trips of moderate distance over manageable terrain, or for people looking for exercise along with or even over transportation, but there are clearly situations where a source of power other than human legs is highly desirable. Thus a problem with the bicycle could easily become a problem to design a motorized bicycle or, more concisely, a motorcycle. Although the problem of designing a motorcycle may be prescribed in the positive terms of fitting a motor to a bicycle to give the new vehicle advantages over the old, in fact the problem derives rather directly from a criticism of the existing device, from the failure of the bicycle to go under its own power. The formulation of the design problem is but a structured articulation of the objective of removing a shortcoming from an existing design.

  The very articulation of a problem, such as “fit a motor to a bicycle (so that the rider can be transported faster and more effortlessly),” can strongly suggest a solution. In practice, the nonverbal conception of a solution by an inventive mind is often what prompts the inventor in retrospect to articulate the problem and couch it in the language of a need. After such a rationalization of a nonrational leap of creativity, what remains is how to effect a solution in a way that minimizes objections and introduces fewer inconveniences than it removes.

  The cover illustration for the issue of Science containing Eugene Ferguson’s insightful article on nonverbal thought in design showed but eight possible turn-of-the-century solutions to the problem of motorizing a bicycle. Not only had the motor somehow to be connected to a wheel by a drive mechanism, but a fuel tank and possibly a battery had to be fitted to the bicycle frame. These ideas might have come in a flash of inventiveness, but, as the illustration so graphically demonstrates, what the motorcycle would look like depends very much on how the component parts were fit together. Assuming they are all technically feasible, the eight configurations taken two by two can best be compared by identifying their individual advantages and disadvantages, which are like opposite sides of the same functional coin. Form may be said to follow function only in the sense that heads or tails follows each flip of the coin. The gaming analogy goes only so far, however, because, unlike the gambler, who is bound by the final coin toss, the designer in the end may pick and choose retroactively which tosses to bet upon in the marketplace.

  Among the many imaginable combinations and permutations of the components of a motorcycle, one places the motor away from the rider, thus eliminating any potential interference with the legs. But locating the motor behind the bicycle requires an extension of the frame, thus increasing the cost of vehicle and altering its center of gravity. What constitutes the “best” solution among the various candidate designs is a matter of judgment and compromise; in the final analysis, the detailed form of the motorcycle does not follow its function in any predetermined way, but ultimately rests on a judgment of which choice is least undesirable. What might ultimately come down to an arbitrary choice among competing configurations, as manifested in the location of the fuel tank, for example, in time can become so strongly associated with motorcycleness that, even if functionally relocated in a new (and improved) design, a vestigial tank (a “survival form”) may be retained in what has become the customary location. The design critic John Heskett has noted a striking example:

  The Ariel “Leader” motor-cycle, … produced in Britain in 1957, had a petrol tank located on the rear frame, but retained a dummy-tank of conventional form. This same device was later repeated on the Japanese Honda “Gold Wing 1000”, the dummy tank opening in half to reveal electrical controls. In both cases, producers felt unable to present a visual choice to the consumers in face of the power of the conventional image of a motor-cycle, even though its form had become functionally redundant.

  Just how far one detail of a design problem can affect form is also illustrated by a more recent “radical innovation” in motorcycle design, in which the powerful motor (now an “engine”) is so large as to itself serve as the frame to which wheels, seat, and other equipment are directly attached. This recalls the form of early motorized tractors, in which the engine and transmission casting also served as the frame to which axles, steering wheel, and other barest essentials were attached. A simple iron saddle seat was mounted directly
over the transmission, and the driver’s feet rested on small stirruplike protuberances, thus giving the impression that the horseless machine itself had been harnessed and was being straddled and ridden much like a living steed. Before that, one of the first steam-powered tractors was actually hitched to a team of horses, not for their power but because there was as yet no mechanical means of steering the machine.

  One of Raymond Loewy’s early commissions was to improve the design of an International Harvester tractor, which even as late as 1940 appeared to be little more than an engine on wheels covered by the barest of protection and having a steering linkage that looked remarkably like a rein. The tractor’s high seat was difficult to reach without getting a leg up, its iron-cleated wheels were prone to clogging with whatever mud they did not throw up on the exposed driver, and the tricyclelike arrangement of wheels made the entire machine rather unstable in tight turns. Loewy’s improved design gave the machine four rubber-tired and spokeless wheels, fenders, and the beginnings of a streamlined body that evoked the form of an automobile more than that of a horse. What Loewy did for International Harvester’s tractor, Henry Dreyfuss did for John Deere’s, and though the two share similarities with what has come to symbolize “tractorness,” each also had its distinctive silhouette.

  Everything designed has an element of arbitrariness in its form. Loewy described how groups of his designers used to go about designing a new model automobile. Different groups were given different tasks, such as the front and rear of the car, and the conceptual work began, to be cut off at some predetermined time by deadlines that were imposed at the outset. After a time, there were “piles of rough sketches,” and Loewy saw the design proceed as follows:

  Now the important process of elimination begins. From the roughs, I select the designs that indicate germinal direction.

  Those that show the greatest promise are studied in detail, and these in turn are used in combination or arrangements with one another. A promising front treatment can be tried in combination with a likely side elevation sketch, etc. From this a new set of designs emerges. These are then sketched in detail. After careful analysis, they boil down to four or five.

  The form of the final design continues to evolve through full-scale plaster or wood mock-ups, and even at this stage there can remain a degree of arbitrariness; “when several models are to be shown, it is advisable to paint them all the same color so that color preference will not influence the choice of the management unduly.” Choices are made not to scheme against the consumer but to choose what would seem to be the best design and therefore the best bet for recouping investment in research and development:

  Changes are inevitably suggested and another complete showing is arranged to demonstrate how these alterations have been incorporated in the design. When the final okay for production is given the design cycle is complete. It is up to the engineering and production departments to draft it and detail it.

  The detailing of a design involves translating the final management decision into precise drawings and specifications so that the thing can be produced. Though designers and engineers can present a multitude of solutions to the design problem, and can argue technologically, aesthetically, and economically for this one over that, it is seldom an engineering decision alone that determines the look that comes off a production line. In cases where the roles of engineer and manager are combined in one person, he or she must wear different thinking caps at different times.

  Loewy gave a further illustration of the absence of predestination in design by relating the story of his involvement in a lawsuit over patent rights in which a client of his was suing another manufacturer for design infringement. According to Loewy, it was “a clear-cut case” in which the competitor simply copied the appearance of a product Loewy had designed. The defense argued that the design patent was invalid, because “the product could not possibly be designed any other way and still function properly.” The case had been dragging on for weeks when Loewy was called as a witness for the client. In the exchange that ensued, the attorney asked Loewy if the particular product could be “designed in any other manner and still be practical and function properly,” and whether he could do so. When he answered positively, Loewy was asked if he could demonstrate such alternative designs, and he replied that he could, by making some sketches. He was then asked to do so, and, according to his own report:

  I unfolded my easel, placed the drawing board on it, and started making rapid sketches in large black outline, visible to anyone in the back row. Ten minutes later, I had about twenty-five designs, all different, most of them attractive, all of them practical.

  Loewy’s ego and business interests seem to have led him to stress his successes, arbitrary as their form might have been; in the final analysis, the design selected would have been the one that, in some compromise way, failed least to satisfy both designer and client. The plurality of solutions to a given problem—and their shortcomings—are virtually inescapable in design.

  Designers less gregarious than Loewy, and working with less conspicuous things than locomotives, have tended to call themselves not designers but inventors. Lyndon Burch, an inventor of circuit breakers, electromechanical switches, and waterproof thermostats that enabled electrical appliances like frying pans and coffee makers to be immersed for washing, got his first real break when he was hired as a design engineer by a New Jersey thermostat manufacturer, which obviously expected him to solve problems associated with the company’s business. According to his own description of how he thought about problems, his mind worked basically with shape and pattern:

  Most of my work really involves geometry—simple geometric structures to perform a function. So I’ll start with a geometric pattern in my mind.… After I see the pattern, I’ll try to find fault with it, and nine times out of ten, I can tear it to pieces, so I’ll start again. But when I’ve got the right pattern, somehow I just know it’s right,

  Burch clearly was able to come up with tentative solution after tentative solution to the same problem. Even if he could tear 90 percent of his solutions apart, this is not to say that they did not solve the problem. They simply did not solve it as well as he imagined they could, or they did not show to Burch the promise he was looking for. For example, one of his important inventions of the late 1940s was a metal switch for thermostats. Existing switches had worked on the principle of a disc of metal responding to temperature changes by snapping through from one position to another, following much the same principle that causes a metal noisemaker to respond to thumb pressure, or the recently faddish slap bracelets to curl around the wrist in a snap. Burch dismissed variations on familiar devices and came up with the idea of achieving a large movement in response to a small one by cutting a flat piece of metal into various shapes that responded in a twisting fashion to pushes and pulls. Thus, the same function, that of responding in a big way to small influences, could be accomplished in a new manner, and this enabled manufacturers to make (and patent) new switches and thermostats which they could claim accomplished functions similar to the snapping disk without infringing on the patents of others.

  All patents contain explicit “claims,” which are often seemingly interminable sentence fragments following the colon ending a rubric such as, “What is claimed is,” “We claim,” or “I claim.” The claims come at the end of a patent, and they ostensibly lay out exactly what is being patented. According to patent attorney David Pressman, claims say to the public:

  The following is a precise description of the elements of this invention; if you make, use, or sell anything which has all of these elements, or all of these elements plus additional elements, or which closely fits this description, you can be legally held liable for the consequences of patent infringement.

  Pressman, giving do-it-yourself advice to the independent inventor who wants to write his own patent application, not only instructs the reader in the basics of writing the sentence fragments of claims but also gives, under the heading
of “other tricks in claim writing,” the advice to “use ‘weasel’ words like ‘substantially,’ ‘about,’ or ‘approximately’ whenever possible” in specifying a dimension, for example, “to avoid limiting your claim to the specific dimension specified.” Pressman also explains why “many patent attorneys recommend that a claim not appear too short”:

  A claim that is short will be viewed adversely (as possibly overly-broad) by many examiners, regardless of how much substance it contains. Thus, many patent attorneys like to pad short claims by adding whereby clauses, providing long preambles, adding long functional descriptions to their means clauses, etc. The trick here, of course, is to pad the claim while avoiding a charge of undue prolixity.

  The legal implications of patents may encourage technical writing at its worst, but the phenomenon is nothing new. In a 1906 patent of theirs, Orville and Wilbur Wright, through their attorney, list eighteen claims for their flying machine The first describes what we today would call one of the wings of a biplane, but then it was called what would become the name for the entire machine:

  In a flying machine, a normally flat aeroplane having lateral marginal portions capable of movement to different positions above or b[e]low the normal plane of the body of the aeroplane, each movement being about an axis transverse to the line of flight, whereby said lateral marginal portions may be moved to different angles relatively to the normal plane of the body of the aeroplane, so as to present to the atmosphere different angles of incidence, and means for so moving said lateral marginal portions, substantially as described.