Of Thinkers and Tinkerers

1 Introduction: Thinking and tinkering

In some very colorful and instructive reminiscences of her lifelong work in the history of science, Marie Boas Hall, recipient of the prestigious Sarton Medal, tells an interesting anecdote about the illustrious George Sarton:

Stephenson, always interested in intellectual history, turned seriously to the history of medieval technology, writing an excellent little article, “In Praise of Medieval Tinkers”; regrettably Sarton, to whom he naturally sent an offprint, replied brusquely, indeed rudely, that he was interested in medieval thinkers, not tinkers, a revealing comment. (Hall 1999: S72).

That comment is indeed revealing – not just of Sarton’s ideological stance, but more widely of the then-prevalent disparagement of technological work, as echoed in his remark. With ancient roots, this attitude has deeply influenced Western thought, most notably in the changing philosophical assessments of the relations of theory and praxis.

It seems almost paradoxical that little more than half a century later we witness a sort of volte-face from this anti-tinker outlook, largely thanks to the intimate, progressive entanglement of scientific and technological research in the last few decades. This embroilment is neatly encapsulated in the fashionable term technoscience, promoted by Bruno Latour (Latour 1987) and cogently explicated in Paul Forman’s controversial paper, “The primacy of science in modernity, of technology in postmodernity and of ideology in the history of technology” (Forman 2007, see also Forman 2010).

In this era of technoscience, scientific and technological research and innovation are so closely intertwined, conceptually, economically, and politically, that their respective natures and the intricacies of their interplay are hard to discern. Yet it remains possible to distinguish both their common characteristics and their contrasting differences (i.e. as to goals, cognitive styles, methods, etc.), when they are properly grasped as the outcomes of historically conditioned human endeavors. In this context, some reflections will be advanced on the activity of “tinkering” as a peculiar cognitive instrument in the generation of technological novelty.

This brief paper points to the invention of the triode thermionic valve, a crucial technological innovation of the early twentieth century, as a springboard for a historically informed discussion of the complex interrelations of theory and praxis in the generation of technological novelty. The choice of this episode was dictated by its critical role in triggering a whole chain of developments that culminates in the growing network of technologies and economic infrastructures that underpin our so-called “information society” and the evolving ascendancy of technoscience in all spheres of human endeavor.

In effect, the multiple and, at first, inchoate functions of the triode valve or triode electronic tube (i.e. switch, relay, rectifier, amplifier, detector, modulator) were initially applied to the development of practical wireless radio in both broadcasting and reception. This turned out to be just the initial step in a broadening cascade of innovations: the rise of television, digital computers, and many other contrivances of the ever-expanding web of cyberartifacts that shape our daily existence today. In the decades succeeding its creation, the triode valve was superseded almost everywhere by its ubiquitous solid-state successor, the triode crystal or transistor.

2 Diodes at work

The invention of the triode is a dramatic turning point in a vast and protracted narrative — the ongoing creation and development of technologies for information transmission, retrieval, and storage by means of electro-magnetic apparatuses. This long story, to which we can only allude in passing, traces the evolution and interrelations of numerous artifacts for electric telegraph and telephone communication, among other emergent technologies.

An apt starting point for our story is February 1880, with Edison’s discovery of thermionic emission, which his contemporaries dubbed “the Edison effect.”^[3]

As part of his experimentation for improving incandescent lamps, Edison introduced separate, sealed wires or electrodes into the evacuated bulbs. During one test, he measured the passage of an electric current between the hot filament and the electrode, while the electrode was connected to the positive terminal of a battery. No current at all was registered when the electrode was connected to a source of negative voltage with respect to the filament (see Fig. 1).

Figure 1

Schematic view of Edison’s circuit

According to present understanding, the thermionic effect is produced through the emission of negative charges (electrons) from the hot filament into the evacuated bulb and their capture by the positively charged electrode. The incandescent lamp was functioning as a rudimentary diode (as it was later called), that is, as a thermionic valve that allows the passage of current in only one direction (i.e. a rectifier). This one-way passage explains the name “valve” usually given to electronic tubes outside North America.

Edison was unable to understand the nature of his effect – electrons were discovered some eight years later. He attributed the current to the emission of tiny carbon particles by the filament. Quite uncharacteristically, he failed to recognize the great practical potential of his proto-diode for rectification (namely, conversion from AC to DC power) and for the detection of radio signals. Nevertheless, in 1883 Edison obtained a patent for a device based on his principle, a voltage regulator.

The next important development on the way to the triode was the work of John Ambrose Fleming, a scientist and engineer who had studied under John Clerk Maxwell at Cambridge. At the turn of the century, Fleming was a consultant for Marconi’s telegraphy company, and in charge of designing the transmission facilities of the Pohldu Wireless Station in Cornwall. In 1901, he sent the first transatlantic wireless message, received by Marconi at the Signal Hill station in Newfoundland.

Fleming, acquainted with Edison’s effect, invented in 1904 the first thermionic valve, the diode, and obtained a patent for it that same year (see Fig. 2). The diode rapidly replaced the unreliable crystal rectifiers and several other devices currently in use as wireless signal detectors.

Figure 2

Schematic from Fleming’s patent for the diode as a rectifier

The diode consisted in an evacuated tube with a heating filament and two electrodes: the cathode (which was a metallic element heated by the filament, or often just the filament itself) and a plate or anode. When a voltage difference was established between the electrodes, a current flowed from a negative cathode to a positive plate, but not vice versa.

3 The advent of the triode

The next important invention, and by all means most crucial innovation in our saga, was created by inserting a third electrode — the grid — between the cathode and the anode of a diode. To understand the impact of this development, something must be said about the state of electrical communications in the first decade of the twentieth century.

Figure 3

De Forest’s triode

At the time when diodes were commercially introduced, the expanding enterprises of line and wireless telegraphy were both confronted with a similar difficult problem: the rapid attenuation of the signals at increasing distances from their source. At the turn of the century, practical telephony was only possible by means of line transmission. Cumbersome systems of relays were developed and located in the lines at periodic intervals to counteract the effects of attenuation. The relays worked relatively well for line telegraphy, but in the case of line telephony they introduced unacceptable levels of voice distortion.

By this time, wireless telegraphy had become a successful venture, but telegraphy is (to use anachronistic but apposite language) a digital affair. Switching the current on and off generated signals as sequences of dots and dashes, separated by silences. Wireless telephony, an entirely different undertaking, demands the processing and transmitting of analogue data encoded in the continuously changing oscillations that constitute the human voice.

From our present vantage point we can see that four important inventions were needed to conquer attenuation in line telegraphy and make possible long-distance wireless transmission of acoustic signals:

1. an amplifier capable of boosting the energy of the signals;

2. a modulator, able to incorporate the weak analogue pattern produced by a microphone – a low (audio) frequency alternating current – into a powerful high-frequency wave that would travel long distances with low attenuation;

3. a detector (demodulator) at the receiving end, with the ability to retrieve the original oscillation pattern from the modulated high frequency wave;

4. and a transmission oscillator (i.e. a high-frequency electromagnetic wave generator) to provide a stable, high-energy wave of high frequency suitable to become modulated by the audio signals.

This exercise of hindsight may seem out of place in a historical exposition, but it highlights the unique and extraordinary character of the invention of the triode.

The American inventor Lee De Forest is credited with inventing the triode in 1906. He baptized it the Audion, a name that suggests some audio frequency applications that he may have surmised at that time. For all intent and purposes, this sole device managed to enclose within a single package all of the four aforesaid inventions, as became manifest when the triode was inserted in succession into properly designed circuits.

It was immediately realized that by feeding appropriate signals to its grid, the triode could work as an amplifier, detector, or modulator (see Fig. 4). In fact, once the physics behind the invention became understood some time later, it was recognized that these separate practical functions are essentially implemented via one single physical process. On the other hand, the electronic oscillator had to wait another six years for its experimental realization.

Figure 4

Early schematic of an Audion radio receiver

The historical details in the invention of the electronic oscillator are not at all clear^[4] but most accounts attribute the first implementation of triode oscillatory circuits to the extraordinary inventor Edwin Armstrong, in 1912. With the oscillator, another technological novelty of great importance was explicitly developed: the electronic feedback loop (see Fig. 5).

Figure 5

Original schematic of Armstrong’s triode oscillator

Until the arrival of the Audion, high-frequency continuous wave generation was an extremely expensive undertaking that required gigantic alternators or cumbersome, costly arc generators. The new device changed all that. By the early 1920’s the increasingly more reliable and cheaper triodes were the ubiquitous prime movers that drove the inner workings of both the broadcasting radio stations and the domestic receivers, which were rapidly proliferating in the United States and abroad.

4 Thinking about tinkering

Upon reading various accounts of the activities that led to the invention of the triode (e.g. Aitken 1985; Hong 1998, 2001) one is impressed by two different but related facts. First, De Forest had no idea how the Audion worked in terms of physical theory, neither before nor after arriving at its creation. Second, he reached his invention by a process of protracted, painstaking tinkering upon various realizations of Fleming’s diode.

The first point illustrates a fact that seems more a rule than an exception. Frequently an invention precedes the theory that explains its function. Most remarkably, a new device is often instrumental in achieving the formulation of the theory that will eventually explain how it works. One conspicuous example is the precedence of steam engine development to the discovery of the laws of thermodynamics. But many other examples abound. Given these historical facts, it seems now strange that the long-dominant view of technology as mere applied science had maintained such enduring sway (for a classical locus, see Bunge 1966).

The second point illustrates the main issue that will occupy us here: the role of tinkering in the generation of technological novelty. Before we proceed, it will be useful to distinguish between two related but different meanings of the English word “tinkering.”

5 The meaning of tinkering

The first meaning is practically synonymous with the French word “bricolage.” It refers to an opportunistic employment of whatever elements happen to be at hand for constructing an artifact capable of performing a desirable function. This is the meaning in François Jacob’s famous article “Evolution and tinkering,” on the analogy between the work of natural selection in biological evolution and the actions of a human tinkerer (see Jacob 1977, also Jacob 2001):

[…] natural selection does not work as an engineer works. It works like a tinkerer -^- a tinkerer who does not know exactly what he is going to produce but uses whatever he finds around him whether it be pieces of string, fragments of wood, or old cardboards; in short it works like a tinkerer who uses everything at his disposal to produce some kind of workable object. For the engineer, the realization of his task depends on his having the raw materials and the tools that exactly fit his project. The tinkerer, in contrast, always manages with odds and ends. What he ultimately produces is generally related to no special project, and it results from a series of contingent events, of all the opportunities he had to enrich his stock with leftovers. As was discussed by Levi-Strauss […] none of the materials at the tinkerer's disposal has a precise and definite function. Each can be used in a number of different ways. In contrast with the engineer's tools, those of the tinkerer cannot be defined by a project. What these objects have in common is "it might well be of some use." For what? That depends on the opportunities.

The materials at hand for De Forest were different realizations of the diode, pumps, glass tubes, wires, batteries, and such, recycled from previous devices and experiments.

The second discernible meaning of “tinkering” refers to a peculiar form of experimentation or experimental exploration that differs widely from the experimental activity that is usually the focus in science and philosophy.

Traditionally, under that heading one is largely concerned with the instrumental practices directed at testing (i.e. refuting or corroborating) scientific hypotheses. This is an activity that takes place after their formulation. Experimentation as tinkering in scientific research, on the contrary, not only precedes the formulation of new hypotheses but is also a heuristic practice aimed at their discovery and formulation. Tinkering is also involved in activities directed at designing and implementing the instrumental arrangements needed for testing hypotheses.

In technological research, heuristic experimentation performs functions that are different from those at play in scientific research. In scientific research the goal is mainly explanatory, typically aimed at answering the question of why something happens. Technological research usually aims at finding how something happens, with the goal of revealing new and potentially useful functions in the behavior observed. The “whys” are embedded within a background of well-established “hows,” and vice versa. In the invention of the triode, lack of knowledge of the why did not prevent De Forest from discovering how the insertion of the grid created new useful functions.

In scientific research, how and why are connected logically and sometimes chronologically. To use a familiar example, Kepler’s ellipses are precise descriptions of how planets move, while Newton’s laws explain why they move in that way. The difficulties in separating the theoretical and practical components of current technoscience are partly due, I would suggest, to the increasing difficulty of separating the why from the how in the tangled network of practices that make up contemporary industrialized research.^[5]

6 Tinkering as technological abduction

The last two decades have seen a remarkable revival, development, and extension of Charles S. Peirce’s notion of abduction (for an extensive survey and bibliography, see Magnani 2009). According to Peirce’s view, there are three basic kinds of inference: deduction, induction, and abduction. Only this last one is capable of generating new ideas. Deduction is an inference that preserves truth from the more general to the less general. In contrast, abduction is an ampliative and eminently fallible operation for generating hypotheses that exceed the information given in their premises. Abduction proceeds from the effects to the causes, and often advances through a survey of particular details in hope of striking on a suggestion for an explanatory general idea (i.e. a new hypothesis).

Abduction as traditionally understood is thus a purely conceptual or theoretical operation. I would like to propose that tinkering, as exploratory experimenting, is a form of typically technological, non-conceptual abduction. Although it is not directed at reaching explanations, it is similar to theoretical abduction in being aimed at making things intelligible. In the practice of tinkering, this intelligibility does not spring from subsuming phenomena under general laws. Instead it arises from the realization of how something works, from perceiving its behavior as resulting from the actions and connections of the surrounding items in the particular, concrete artifactual context in which it is inserted. Although he did not understand the physics behind the functioning of the triode, De Forest understood its workings well enough to perceive some of its potential applications and to successfully improve it until they became clearly manifest.

Starting with the classical work of Vincenti (1990), much has been written on the peculiarities of technological knowledge, notably on its similarities to and differences from the knowledge used and produced in scientific research. The discussion of this vast subject and the assignable place of tinkering and abduction within it are beyond the limits of this essay.

Nevertheless, to close this discussion one other observation can be advanced, concerning the role of tacit knowledge in technological research, that may help integrate these ideas with current research in future work.

The conception of tacit knowledge promulgated by Michael Polanyi (see, e.g., Polanyi 1968; Houkes 2009; Nightingale 2009) is eminently applicable to a discussion of technological tinkering.

Polanyi distinguishes between two kinds of knowledge that are in some respects antithetical but function in complementary ways, guiding all human activities. We employ focal knowledge when we attend to a thing in the foreground, in such a way that we can tell what it is and can articulate this knowledge in words. We employ tacit knowledge (or subsidiary knowledge) of something when we rely on it only insofar as it helps us to attend to something else. This knowledge resists articulation in words and is typically displayed in the acquisition of skills (e.g. riding a bicycle through repeated practice and imitation of other riders).

We can tell what the things are which we know by attending to them focally, but we are uncertain, or entirely ignorant, of things that we know only by relying on our awareness of them for attending to something else, which is their meaning (Polanyi 1962: 601).

For our purposes, what is important is that we have mediated access to a large reservoir of knowledge. This knowledge becomes available to us through our dealing with things in the pursuit of goals — goals for whose attainment those things are apt to become instrumental. In technological research, tacit knowledge is in a sense embodied within the items we manipulate. Tinkering explorations may be seen as procedures for releasing that latent knowledge, in the pursuit of bringing about new arrangements of parts capable of performing a desirable function.

Tinkering is important both in experimental science and in the creation of technological novelty. But in the search for invention, tinkering is no longer subservient to the generation of focal, conceptual knowledge. Moreover, it attains its highest fulfillment in bringing into existence new artifacts and processes. The invention of the triode is an archetypical example of this kind of achievement.

7 In guise of conclusion

To bring some sort of closure to this stream of reflections, let us revisit our point of departure in Sarton’s dictum about tinkers. It represents a symptom of the traditional depreciation of technological activity, still prevalent through much of the twentieth century. We considered the current ascendancy of technoscience and noted the extraordinary reversal in the valuation of technology that is enshrined in that word. As a partial explanation, we pointed to the increasing and complex entanglement of theory and praxis in commercialized research.

We considered possibilities of theoretical disentanglement by discerning a difference in goals and in the effect of cognitive practices with dissimilar operational functions in science and technology. We focused on tinkering as exploratory experimentation and illustrated its fruitfulness using the invention of the triode. Finally, we related tinkering to the exercise of abduction in the formation of scientific hypotheses and to the application of tacit knowledge in research.

To pursue these ideas further, it will be necessary to place them within the growing net of new insights recently emerging through the work of historians and philosophers of technology. Among these new realizations one holds particular promise for developing these ideas: the notion of the “hybrid” nature of technological artifacts. These things are amphibious creatures between two worlds. On the one hand, they partake of the nature of inert physical objects, because their defining functions result from the blind necessity of the laws of physics. On the other hand, they belong in the world of living organisms, and are characterized by an intrinsic teleology. These instruments are what they are by being inserted in a relational network of means to ends (Fernández 2008). Technological artifacts are instrumental extensions for enacting the purposive actions of organisms, humans in particular.^[6] The expansion and elaboration of the ideas reviewed here are, of course, a matter for some future work.