Jump to ContentJump to Main Navigation
Understanding Industrial and Corporate Change$

Giovanni Dosi, David J. Teece, and Josef Chytry

Print publication date: 2004

Print ISBN-13: 9780199269426

Published to Oxford Scholarship Online: September 2007

DOI: 10.1093/acprof:oso/9780199269426.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. Subscriber: null; date: 16 October 2021

Heroes, Herds and Hysteresis in Technological History: Thomas Edison and ‘The Battle of the Systems’ Reconsidered

Heroes, Herds and Hysteresis in Technological History: Thomas Edison and ‘The Battle of the Systems’ Reconsidered

Chapter:
(p.309) Heroes, Herds and Hysteresis in Technological History: Thomas Edison and ‘The Battle of the Systems’ Reconsidered
Source:
Understanding Industrial and Corporate Change
Author(s):

Paul A. David (Contributor Webpage)

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199269426.003.0011

Abstract and Keywords

This chapter looks at the development of electric lighting and power supply networks in terms of the battles waged around 1887-1892 between proponents of direct and alternative current systems of electrical supply in order to raise questions about technical progress as a continuous flow. Utilizing the physics concept of hysteresis as the persistence of an altered state when the force that caused alteration ceases, the chapter concentrates on the critical moments or ‘points of bifurcation’ in the dynamic of technical change that are prone to appear at the early stages of an incremental process. The chapter concludes from its reading of the ultimate victory of alternative over direct currents of electrical supply that innovation implies not so much the work of unique creative attitudes in the manner of the classic Schumpeterian entrepreneur, as it does the occupation of a pivotal situation during comparatively brief moments of industrial development when the balance between choices can go either way.

Keywords:   technological change, electric supply, hysteresis, points of bifurcation, industrial innovation

The history of this development is worth following—as an example of scientific and technological co-operation, of multiple invention, of progress by an infinitude of small improvements, of creative entrepreneurship, of derived demand and unanticipated consequences.

David S. Landes (1969, p. 284) on electric power.

A master's advice always is compelling. The story to be told here will be about the early period in the development of electric lighting and power supply networks. Since this tale is one whose broad outlines probably are familiar to many readers from previous and more skillful recountings,1 I must have a special reason to return to it here. My motivation lies in the bearing which some less known details of the rivalry between direct current and alternating current systems of electricity supply will be seen to have upon a far broader, and more complex matter. That subject concerns the precise place to be accorded individual economic actions in the evolution of modern technological systems.

What causal role properly can be assigned to particular decision-makers in the build-up of those complex production systems that form the material (p.310) infrastructure of modern societies? When do specific actions, taken purposively and implemented by identifiable agents (say, entrepreneurs) have the power significantly to alter the course of technological history? Ever? Or should we hold that the economic analyst of technical change, being occupied for the most part with the progress of armies which have few generals and need none, will find only misleading metaphors in the heroic deeds chronicled by traditional political and military historians?2 Rather than being ‘made’ by anyone at any particular point in time, does not technology simply ‘accumulate’ from the incremental, almost imperceptible changes wrought by the jostling interactions among the herd of small decision-makers—the crowd of myriad dreamers and scientists, tinkerers and engineers, inventors and innovators, tycoons and managers, artisans and operatives, all of whose separate motions somehow have been coordinated and channelled by markets under the system of capitalism? Should we then not insist that the nature of technical progress is best conveyed by conceptualizing it as a continuous flow? Shall we therefore foreswear the reading and writing of narratives which would dwell on a succession of unique actions and discrete, historical ‘events’? Inasmuch as recognition of such actions, and events, is required if we are to speak in a meaningful way about the occurrence of ‘changes’ in technological practice and economic organization, does it not appear that there is a high intellectual cost entailed in adopting a mode of discourse that suppresses them? These questions are daunting, indeed, and no definitive answers to them can be expected to come from this quarter. But, surely, they cannot be inappropriate ones to consider, especially now that ‘innovation’, ‘entrepreneurship’ and, indeed, ‘intrapreneurship’ are words on the lips of those in business and politics who concern themselves with preserving ‘competitiveness’ by fostering technological dynamism.

1. Views of Technological Change—Herds or Heroes? Continuities or Events?

From the way that my larger agenda of concerns has been phrased, one might surmise that I am preparing to plunge into the thick of the recurring debate over Schumpeter's representation of entrepreneurs as individuals who form a distinct sociological category—those heroic few to whom in each era is assigned the function of innovation, and without whom the capitalist (p.311) economic system soon would settle down to a dismal stationary state.3 But, I confess, I am too wary for that; I do not mean to focus squarely on the vexed question: what is it that entrepreneurs and ‘bosses’ do, really do, or may have done in the past?4 Instead, I want to take up a different problem which, being more general, is logically antecedent to that subject of ongoing controversy among economic historians. When, if at all, in the development of society's technological knowledge and apparatus, could the actions of some individual participants in a market process be expected actually to exert a perceptible influence over the eventual outcome—in effect, determining the characteristics of the particular trajectory of diffusion and development along which some branch of technology becomes channeled?5 For this, one must admit, there is no need that the vital actors be cast as visionary heroes, or essential contributors to enhancing economic efficiency. Scoundrels and fools, or merely men with weak stomachs for the perils of entrepreneurial voyages, might serve just as well in teaching us when it is that big consequences are most likely to flow from historical events petty enough to be the work of ordinary mortals.

Oddly enough, it would be quite uncongenial for many who now are studying technological and economic history to be asked to concentrate attention on the detailed actions of specific individuals, and so to emphasize the role those persons had in producing changes that were consequent upon ‘events’.6 At least two causes can be found for this state of affairs. For the first, there is the dominant historiographic tradition in regard to the progress of technology. As elsewhere in the writing of social science history, this tradition perpetuates (p.312) Darwin's approach to the study of evolution. Its followers cling rather perversely to an essentially ahistorical concept of the modus operandi of change: natura non facit saltum. Changes, on this view, are produced by slow, continuous adaptations in an ‘eventless’ world.7 It is, of course, a strong tradition that can draw upon Karl Marx, A. P. Usher, and C. S. Gilfillan for elaborations of the view of technical progress as a social process, consisting of the steady accretion of innumerable minor improvements and modifications.8 It therefore supports the conceptualization of ‘technology’ as a powerful flux, a broad and mighty river reshaping the economic and social landscape and, in the process, sweeping puny individual actors—inventors, businessmen, workers, customers, politicians—along in its currents. Thus, although the dominant tradition in the historiography of modern technological change has firmly embraced the Schumpeterian vision of a distinctly capitalist process of development situated in historical time, its adherents paradoxically still resist Schumpeter's emphasis upon the discontinuous nature of technological progress—the driving force he saw at work within that process.

For a second cause we can turn closer to home. In the era of ‘Cliometrics’, especially, economic theory has not been without an influence upon the way that technological history has come to be studied.9 Mainstream economics since ‘the Marginalist Revolution’, and certainly since Alfred Marshall, has subscribed to the characterization of change as being so incremental as to constitute an ‘eventless’ continuum. Moreover, although modern micro-economists show no hesitancy in producing formal theories of human action, they appear to be most comfortable with the assumption that actual individuals (in the past and present) somehow or other manage to cancel each other out, so that changes emerge from the ensemble of their behaviors—that is, from the action of the anonymous masses. As a consequence of the neoclassical economists' efforts to expose the workings of mechanisms that would cause invention and innovation (p.313) to respond like other resource-using activities to signals and incentives emanating from markets, Adam Smith's metaphoric ‘invisible hand’ has now reached deep into the technological historians' bailiwick. The actions of many individual human minds and muscles are thus depicted naturally as being subject to coordination by non-hierarchical social organizations including markets, thereby permitting very substantial decentralization of control in the creation of technological achievements of great complexity.10 On this view it would appear that no techno-economic entrepreneurs of extraordinary stature are required for the creation of such things as railroads, electrical supply systems, or telephone networks.

While prepared to demur from the latter opinion, I seek no quarrel with those who say that technological progress is a collective, social process, many important aspects of which are characterized by continuity. Nor would I dispute the case for treating invention and innovation as endogenous to the resource allocation process conducted by means of markets. But, I do want to counteract or at least to qualify the suggestion—one that is strongly conveyed by some recent contributions to the history of technology—that individuals essentially have no real points of leverage from which to control the outcomes of such a macrocosmic, societal process; that decision-making agents either are rationally moving with the tide of technical change, or, by failing to heed the proper signals, are running the risk of being swept aside by it.

There is another conceptualization of dynamical systems constituted of many elements or ‘actors’, in which the actions of individuals that appear to be autonomous and even extraneously motivated, nevertheless, can turn out to have exercised a profound and persistent influence upon the macrostate. These are dynamical systems in which local positive feedback mechanisms (e.g. in economics, dynamic increasing returns due to learning-by-doing) predominate over negative feedback mechanisms (e.g. increasing marginal costs), and which, therefore, are characterized as ‘self-reinforcing’ or ‘auto-catalytic’. Stochastic systems of this kind in physics, chemical kinetics, and theoretical biology are perhaps not the most suitable paradigms for understanding economies composed of human agents, but they are nonetheless a source of insights into processes of change that are not ‘eventless’, and whose outcomes are path-dependent. As a result of the operation of local positive feedbacks which (p.314) give rise to regions of instability, these systems typically possess a multiplicity of stable ‘attractors’, or possible asymptotic states (sometimes) referred to as ‘emergent structures’. The initial state of the system, combined with random events, or ‘channelled perturbations’ arising from endogenous innovations or periodic fluctuations at the microcosmic level of the individual actors, serves to push the dynamics into the domain of one of these asymptotic states.11 Consequently, even extraneous and transient disturbing events—far from ‘averaging out’, so that their influence may safely be abstracted from—act as historical selection mechanisms, ‘selecting’ the particular persisting structure or configuration that eventually emerges; the nature of the eventual outcome is thus not uniquely preordained, but, instead, is dependent upon the details of the sequence of small ‘events’ along the dynamic path.

In physics the term ‘hysteresis’ refers to the persistence of an altered state when the force that caused the alteration has abated. Positive feedback conditions of the kind just described are the very ones in which the actions of individual agents are most likely to give rise to hysteresis effects. Indeed, in social and economic systems where positive feedback dominates, the character and personality of individual actors takes on much greater importance than classical and neoclassical economic theory, long preoccupied as it has been with the analysis of negative feedback regimes, has been wont to acknowledge. Where interactions among agents tend to be reinforcing rather than counteracting, all the ‘other-directed’ agents end up having to adapt to the ‘inner-directed’ few who remain, for rational or irrational reasons, immune to the influence of their immediate socio-economic environment.12 Under such conditions, the personal ambitions and subjective beliefs of inner-directed entrepreneurs, and the particular ideological convictions of public policy makers, become essential ingredients of the temporary historical contexts within which the long-run course of change may be pushed irreversibly in one direction rather than in another.

Rather than elaborating upon the foregoing general but highly abstract formulation of an alternative vision of the way the world (or at least some parts of the world) works—an approach that I have adopted elsewhere—I shall try here to advance my argument by means of a concrete historical illustration.13 (p.315) I will use the story of a technological rivalry—the ‘Battle of the Systems’, so-called, between the late nineteenth century commercial proponents of direct current and alternating current systems of electricity supply—to suggest that sometimes we should expect to find critical moments, or points of bifurcation in the dynamic process of technical change. Such points, really more like narrow windows in the time domain, are especially prone to appear at early stages of the incremental development of integrated production and distribution systems that are characterized by strong elements of localized positive feedback, or ‘increasing returns’ to the more extensive utilization of one or another among the available technical formulations. In this context, specific technical formulations represent variants of a particular technological system paradigm and tend to undergo further development and elaboration along rather narrowly defined trajectories which lead away from these historical bifurcation points.14 It is at such junctures, I suggest, that individual economic agents do briefly hold substantial power to direct the flow of subsequent events along one path rather than another. Unfortunately, in the very nature of those circumstances it is difficult to gauge just how quickly seemingly similar paths may begin to diverge, or to foresee exactly the eventual alternative outcomes towards which they lead.

Looked at in retrospect, however, things often take on an appearance of having been rather more pre-ordained. In the case at hand, modern commentators have been inclined to downplay as historically insignificant the rivalry that arose in the late 1880s between Thomas Alva Edison and George Westinghouse, then the publically prominent champions, respectively, of direct current and alternating current electrical supply systems in the United States. The actors in these events hardly rank as pygmies on the stage of American technical innovation and entrepreneurship, and their supporting cast in this particular drama included the likes of Elihu Thomson, Nikola Tesla, Henry Villard, J. Pierpont Morgan, and Samuel Insull! Yet, such is the strength of the currently dominant historiographic tradition that some satisfaction has been taken in showing even an Edison to have been utterly without power to prevent the technical ‘progress of civilization’ from pushing (p.316) aside his own previous achievements in the field of direct current, and moving forward in the new form of polyphase alternating current. Edison has been portrayed as foolishly deluded, stubbornly egotistical, and worse, for launching a campaign against the Westinghouse electricity supply system on the grounds that it constituted a menace to public safety.15 Most historians of the episode have supposed that the Wizard of Menlo Park and his colleagues meant to ‘do in’ their rivals by blocking the future commercial application of alternating current in the United States, and so they present Edison's failure to achieve that result (despite the employment of low strategems) as testimony to the proposition that there is not much scope for either individual heroism or knavery where technical change is concerned. This would tell us that the momentum of technological progress is overwhelming, that it raises up its ‘heroes’ and casts down those who try to thwart or redirect it for their own irrational ends. Although the thought may be comforting to some, I reject it as a myth that has encouraged flawed interpretations of the facts and the extraction of quite the wrong moral from the historical episode in question.

The reinterpretation of the ‘Battle of the Systems’ that I offer here argues for a different way of thinking about the process of technological change. We should become more consciously reconciled and, perhaps, even comfortable with its contingent nature, acknowledging that some and possibly many episodes in the progress of technology will be marked by a high degree of ex ante indeterminacy. Not all of them, of course, because the whole of the world of technology does not work that way. I would prefer to see careful qualifications placed around bald propositions such as: ‘There is nothing in the character of a previous event or decision—the choice of one path or another—that implies reversibility. Even an accident changes the future irremediably.’16 But, where localized positive feedback mechanisms are operative, there the element of ‘chance’ or ‘historical accident’—in the form of idiosyncratic personal perceptions and predilections of the actors, as well as extraneous and transient circumstances surrounding their decision-making at such (p.317) junctures—is most likely to acquire sufficient power to shape the eventual outcomes. Such adventitious influences must render the task of prediction an unproductive one for all but the closest and most acute observers.

The point I wish to underscore by example is therefore a modest one, concerning the special scope which the development of network technologies creates for individual actors to influence the course of history, as innovating entrepreneurs, or in other roles. As a result, some important and obtrusive features of the rich technological environment that surrounds us may be the uncalculated consequences of actions taken long ago by petty heroes, and petty scoundrels. The accretion of technological innovations inherited from the past therefore cannot legitimately be presumed to constitute socially optimal solutions provided for us—either by heroic entrepreneurs, or by herds of rational managers operating in efficient markets. Before taking up the promised historical tale, it is important to explicitly consider why this should be particularly true in the case of a generic network technology.

2. Network Technologies, Compatibility Standards, and ‘Accidents’ of History

The recent wave of improvements in information, computing and communications technologies has heightened popular awareness of some of the special opportunities and difficulties created by competitions wages by commercial sponsors of alternative technical formulations for new, complex products and production methods. Brief reference to our contemporary experience therefore can assist in forming an intuitive appreciation for the highly contingent way in which the technology of the electrical manufacturing and supply industries came to be elaborated at the end of the nineteenth, and the significance of the rapid emergence in the United States of de facto technical standards for a universal electrical supply system based on alternating, rather than direct current. Recall that during the early 1980s, the burning issue in the market for video tape recorders (VCRs, more specifically) was whether to commit to the Betamax format developed and sponsored by Sony, or cast one's lot with the VHS format group that had formed around the Japan Victor Corporation.17

Who among us has not wrestled with the question of whether our desktop computer is to be an IBM PC-compatible system, or an Apple system like the Macintosh? Or, paused to consider the call of the International Dvorak Society (p.318) to reconfigure our computer keyboards, learn to touchtype on the DSK keyboard patented in 1932 by August Dvorak and W. L. Dealey, and thereby escape a perpetual bondage to the inefficient QWERTY layout?18 Such quandries seem all too commonplace nowadays.

Two features are significant about the prosaically modern choice-of-technique situations just instanced. First, they involve technologies characterized by increasing returns to scale of use or production, and second, they entail choices where considerations of technical interrelatedness among the components forming alternative systems cannot be ignored. These are attributes of ‘network technologies’, which, because they may give rise to pecuniary and technical externalities, create special problems as well as economic opportunities for private and public decision-makers. It is now quite widely recognized that users of a network technology are at the mercy of the social mechanisms available for maintaining efficient system performance by providing compatibility among all its constituent elements. Network technologies may, of course, be developed in a coordinated way through private commercial sponsorship of an integrated system: all the necessarily interrelated components of a product or production facility can, in many instances, be packaged together by a single provider who develops and maintains the required interface specifications. A difficulty can arise, however, when the entailed fixed capital costs are so large that they pose effective barriers to the entry of competitive formulations of the basic technology; the gains in operating efficiency assured by integrated supply then may be obtained at the expense of the inefficiencies that come with ‘natural monopoly’ and too little diversity to suit the varied needs of end-users. On the other hand, when coordination of a network or technological system is left to the decentralized resource allocation processes of competitive markets, the latter have generally led to an insufficiently high degree of standardization of ‘compatibility’ to avoid efficiency losses in systems operations.19

A different point, which has only lately come to be more fully appreciated among economic analysts, is that when private agents in pursuit of profits set about solving what seem to be reasonably straightforward problems of operating more efficiently with the technologies already at hand, the ensemble of suppliers and users can easily become committed to an unfavorable implicit ‘trade-off’ against dynamic gains from further basic innovations in (p.319) fundamental systems-design. When the relative attractiveness of adopting a technical solution tends to increase as that particular practice or device gains adherents, and durable commitments are being made sequentially by different agents under conditions of imperfect information, the ultimate outcome of the process can turn on the detailed timing of small events—even happenings of a nature so extraneous as to appear random in character.20

Such commitment, moreover, is most likely to occur within a narrow frame of time, positioned near the beginning of the technologic's trajectory of diffusion and development. What defines these critical phases, or windows in time, and causes them to close down quickly, essentially, is the growing weight that attaches to considerations of ‘network externalities’—the shaping of individual technological choices by the contact formed through the prior decisions of other agents. For it is the distribution of the already ‘installed base’ that will come to count more and more heavily in determining choices about interrelated capital investments—which must be made in respect to both supplying and using the network technology.

Governmental intervention, therefore, is not the only potential source of ‘standardization errors’; markers also can make early and costly technological mistakes which subsequently prove policy interventions, the existence of significant increasing returns to scale, or analogous positive feedback mechanisms such as ‘learning by doing,’ and by ‘using,’ can give the result that one particular formulation of a network technology—VHS-formatted VCRs, or QWERTY-formatted keyboards—will be able to drive out other variants and so emerge as the de facto standard for the industry. By no means need the commercial victor in this kind of systems rivalry be more efficient than the available alternatives. Nor need it be the one which adopters would have chosen if a different sequence of actions on the part of others had preceded their own.21

Yet, by the same token it is entirely conceivable that for reasons having little to do with the ability of economic agents to foresee the future, or with anything more than an accidental and possibly quite transitory alignment of their private interests with the economic welfare of future generation, technological development may take a particular path which does prove eventually to have been superior to any available alternative. This is a disquieting message for (p.320) those who normally find comfort in the Panglossian presumption that ‘the invisible hand’ of market competition—assisted sometimes by the visible hand of farsighted management—somehow has worked to shape ours into the most economically efficient of all possible worlds. Yet, recognition of the importance of network externalities naturally predisposes one to be extremely skeptical about the claims of that proposition to any general validity. We must acknowledge that where network technologies are involved one cannot justifiably suppose that the system which has evolved most fully is really superior to others whose development might have been carried further, but was not. Nor should we comfort ourselves with the presumption that the ‘right economic reasons’ were responsible for the emergence of a technological system that has in fact turned out to be superior to any of the alternatives available. These are lessons borne out by the curious events now to be recounted.

3. Some History of ‘Current’ Affairs

The years 1887–1892 witnessed an episode of intense rivalry involving the proponents of two technologies vying for the electricity supply market. The contestants were, on one side, the incumbent direct or ‘continuous’ current system sponsored by the complex of manufacturing and financial interests that had been built up around Edison's 1878 patent for a carbon filament incandescent lamp, and its sequelae;22 and, on the other side, the challenging alternating current technology which at the time was represented on the American scene primarily by the Westinghouse and Thomson-Houston Companies. This was not a Competition of long standing, for the first commercial a.c. system was built during the fall of 1886 in Buffalo, NY, by the Westinghouse Electric Co., an enterprise that had received its corporate charter from the State of Pennsylvania only as recently as January of the same year.23 George Westinghouse, whose fame and fortune already had been established by more than a score of patents which he had secured for railroad airbrake apparatus in the years 1869–1873, was thus a comparative latecomer to the young electrical lighting industry.24 In 1886 the Edison Electric Light (p.321) Company already had been installing isolated plant electric supply systems for 5 years, and, having commercially implemented its system of incandescent lighting in 1882 by opening the Pearl Street Station, in New York City, and the Holborn Viaduct Station, in London, was well launched into the business of building ‘central’ generating plants.25

Although Thomas Edison's inspiration had propelled direct current into an early position of leadership as the basis for a commercially implemented technology for electricity supply and application, the fundamental principles of generating alternating current with dynamos were not new. Indeed, the invention of the first direct current dynamo, by H. Pixii, in 1832, had involved the addition of a commutator to the alternating current dynamo he had built immediately following Michael Faraday's discovery of the ‘dynamo principle’ in the preceding year. Furthermore, knowledge of the fundamental ‘inductance’, or self-induction property of electric conductors (which formed the basis for the technique of distributing alternating current at reduced voltage by means of ‘step down’ transformers) dated from the same era. It had been discovered by Faraday in 1834, and, even before him, by an American professor, Joseph Henry.26 Commercial exploitation of a.c. had awaited the demonstration—by the Frenchman, Lucian Gaulard, and his English business partner, J. D. Gibbs—that by using highest voltage alternating current and step-down transformers it was possible to achieve substantial reductions in the costs of moving energy, while making electrical power available at the low voltages then suitable for application to incandescent lamps.27 The beauty of the thing was that transformers could be used without entailing significant losses of power to substitute voltage for amperage. Doing so would reduce the need for the high conductivity, heavy gauge copper-wire transmission lines that were used to distribute current in the early d.c. lighting systems, thereby saving greatly on fixed capital costs. Step-down transformers could then be employed to reverse the process and make low voltage available for local distribution at points of consumption.28 The critical portion of the detailed (p.322) chronology with which we shall be concerned (see Figure 1) is therefore the part that commences in 1882, when a patent was awarded to Gaulard and Gibbs in Britain.

There were, however, practical problems with the Gaulard-Gibbs transformer system, and it was not until 1885 that remedies were found for these. On the American side of the Atlantic, this was accomplished by William Stanley, an inventor in the private employ of George Westinghouse. In 1884, Westinghouse, having established himself during the preceding decade as the inventor and manufacturer of air-brake systems for railroad passenger trains, had become actively interested in the potentialities of alternating current, and had moved quickly in securing the American rights to the Gaulard-Gibbs transformer patent. Within the same year, 1885, S. Z. de Ferranti was installing an a.c. lighting system with transformers wired in parallel in London's Grosvenor Gallery.29 The commercial use of alternating current for lighting in the United States followed close on the heels of these developments. An experimental lighting system had been set up by Stanley in Great Barrington, MA., and successfully demonstrated for Westinghouse in March, 1886; within six months Stanley's work was being put to use in Buffalo by the Westinghouse Electric Company.30 Thus, on both sides of the Atlantic, Thomas Edison in 1886 was quite suddenly confronted by potential commercial rivals who had began to explore an alternative technological trajectory.

Initially, the two systems did not compete directly. Instead, they staked out distinct portions of the market which were determined primarily by the differences in their inherent technical constraints. Two distance-related problems hampered immediate widespread application of the d.c. technology. First, despite implementation (beginning in 1883) of Edison's three-wire distribution design which reduced the cost of wiring by two-thirds relative (p.323)

 Heroes, Herds and Hysteresis in Technological History: Thomas Edison and ‘The Battle of the Systems’ Reconsidered

Figure 1. Chronology of key events in the ‘battle of the systems’ episode in the US.

Size distribution of industrial enterprise (%)

China (1982)

S. Korea (1981)

Japan (1972)

Yugoslavia (1981)

Hungary (1981)

5–33 employees

59.2

70.6

80.2

6.6

2.2

33–75 employees

19.5

14.4

10.7

15.8

4.8

75–189 employees

12.2

9.2

6.1

32.1

18.7

189–243 employees

8.5

1.5

0.8

12.0

9.2

More than 243 employess

0.6

4.3

2.2

33.5

65.1

Source: World Bank, as reported in The Economist, August 1, 1987, China's Economy Survey, p. 10.

Direct Current System Developments

Alternating Current System Developments

1878: December: Swan carbon lamp exhibited in Newcastle, England.

1878–1882: Different ‘alternators’ (a.c. generators) designed by Gramme, Ferranti, and others.

1879: Edison patents filament lamp in U.S. and Britain; builds bipolar dynamo.

1880: Edison electric lighting ‘system’ at Menlo Park;

: First commercial carbon filament lamps produced.

1882: Edison central stations at Pearl St., N.Y., Holborn Viaduct, London.

1882: Gaulard and Gibbs file British patent for distribution by transformers.

1883: Edison–Hopkinson 3-wire distribution system, patented and installed.

1883: Parallel wiring of generators demonstrated, following Hopkinson.

1885: van de Poele streetcar system in New Orleans, South Bend, etc.

1885: Stanley patents improvement of GG transformer; Zipernowsky, Deri and Blathy also patent improvement.

: Ferranti installs A.C. system with transformers in parallel in Grosvenor Gallery.

1886 November: Edison sanguine about Westinghouse competition, raises ‘safety issue’ in memo to Johnson.

1886 November: Westinghouse Co. completes first A.C. central station in Buffalo, N.Y.

: Villard returns with ‘proposal’.

: March: Stanley demonstrates experimental A.C. system for Westinghouse at Gt. Barrington Mass.

1887: Sprague electric traction system successful in Richmond, Va.

1887: Induction motor research by Telsa, Ferraris.

: Edison and associates engaged in animal experiments at West Orange.

:Thompson-Houston enter production and sale of A.C. lighting system.

: March: Bradley patents polyphase induction motor.

: October: Telsa files first A.C. motor patent

1888: July: Brown demonstration at Columbia: campaign to presuade N.Y. legislature to set voltage limit.

1888: May: Bradley files for patent on rotary converter.

: April–June: Shallenberger develops A.C. meter based on induction motor.

: July: Westinghouse acquires Tesla patents.

1889: January: Edison General Electric Co. organized.

1890: Edison liquidates personal holdings in EGE Co.

1890: Westinghouse Electric and Manufacturing Co. reorganized and refinanced by Belmont syndicate.

1891: Westinghouse power system at Telluride.

: Demonstration of transmission with 3-phase alternators driving synchronous motors, at Lauffen-Frankfurt.

: Niagra projects commits to electricity.

1892: General electric formed by Thomson, Houston & Morgan-backing for acquisition of EGE Co.

1893: Westinghouse demonstrates ‘universal’ 3-phase system at Chicago World's Fair.

: Westinghouse and GE submit A.C. generator designs for Niagra project.

(p.324) to the previous two-wire scheme,31 the cost of copper wire in the transmission lines continued to define what constituted an economically viable distance in competition with illuminating gas. Moreover, even if lower cost conductors were to be found, or new wiring systems devised, the low voltage d.c. system was distance-constrained on technical grounds having to do with ‘voltage drop’. In other words, the further the distance a current traveled, the larger the voltage loss on the lines between the point of generation and the point of consumption. The earliest Edison systems, which typically were confined to a fairly small service are (approximately one mile in diameter), made allowance for this problem by generating at 105–110 volts in order to insure the delivery of at least 100 volts at the point of consumption—the latter being the voltage for which household lamps and other appliances were then being designed.32 Beyond a certain transmission distance, however, this type of adjustment would no longer prove adequate; the system would deliver too broad a spectrum of voltages along the transmission lines.

Like others, George Westinghouse had seen these limitations of the direct current system as leaving an opening through which the alternative, Gaulard-Gibbs formulation could be used to make a significant inroad into the promising electricity supply business dominated by Edison. Transmission of alternating current at higher voltages meant that with a given amount of generating capacity and a given weight of copper wire, the distance over which it remained economically feasible to deliver electric power would increase substantially. Indeed, for a given amount of power input, and a wire conductor of specified material and cross-sectional area, transmission distance increases as the square of voltage.33 The economic implication of this was that expansion of supply from existing a.c. generating capacity—which allowed the further spreading of fixed capital charges—would be comparatively unconstrained by the costs of reaching customers over a wider area. Situating (p.325) generating facilities at locations where ground rents were not as high as those found in urban central business districts, where electric energy demand for lighting, generally was most concentrated, was another potential source of economic advantage. Transmitting alternating current at high voltage and then employing step-down transformers at sub-stations, or ‘bumping down’ at the point of consumption, also greatly ameliorated the voltage drop problem encounter in operating d.c. systems.

Nevertheless, at this stage, Edison's system remained the better positioned of the two to dominate the more densely populated urban markets. Part of the reason was that there peak load lighting demands could be matched with Edison's larger, more efficient generators, leaving the a.c. supply systems to work the surrounding territory. Reinforcing this territorial division were numerous absolute disadvantages which blunted the ability of the early, single-phase a.c. technology to penetrate the urban market in competition with d.c.

At least four generic difficulties with a.c. remained to be resolved at the end of 1886. First was the fact—alluded to already—that the early Westinghouse and Thomson-Houston ‘alternators’ (the a.c. equivalent to dynamo generators) were only 70 percent efficient, whereas 90 percent efficiency was being achieved with d.c. dynamos, especially the large, ‘jumbo’ design introduced by Edison. Second, the d.c. system was able to provide metered electric supply, whereas no a.c. meter had yet been developed—a deficiency that greatly reduced the system's attractiveness to central station operators. A third draw-back for the a.c. technology at this stage of its development was that, although the principle of operating alternators in parallel had been demonstrated in 1883 (following the theoretical work of John Hopkinson), it remained to be translated into practical central station operations. The ability to connect dynamos in parallel rather than in series gave the d.c. system a distinct advantage: its generators could be disconnected and reconnected to the mains in response to varying load, thereby saving power inputs and wear and tear. Further, d.c. generators could be shut down for repair and maintenance—or might even be tolerated to break down—without disruption of the entire system.34

Fourth, at a time when central stations employing the Edison system were beginning to spread fixed generating costs by supplying electricity for power as well as lighting, the a.c. system's ability to compete in urban markets was restricted by the lack of any satisfactory secondary motor available to be used with alternating current.

In addition to depriving electricity supply companies who chose to install alternators of the ability to serve industrial and commercial power customers, (p.326) the lack of a motor became an increasingly pronounced comparative disadvantage with the success of the experimental electric streetcar systems installed during 1885 by Charles J. Van Depoele in New Orleans, South Bend, and Minneapolis. Its ability to deliver high torque at low r.p.m. made the direct current motor particularly well suited in such applications as traction, where continuous speed control was important. The ‘traction boom’ based upon d.c. was well and truly launched in 1887, with the completion of the superior streetcar system developed for the City of Richmond, VA, by Frank J. Sprague—a talented but erratic inventor formerly employed by Edison—in partnership with another Edison associate, Edward Johnson. There would be 154 electric street railway systems in operation in the United States by the close of 1889.35

It should be emphasized that up until 1888 the underlying technological and economic considerations were not such as to allow great scope for the commercial realization of conventional scale economies and positive network externalities in the electricity supply business, as distinguished from the business of manufacturing electrical supply equipment and appliances, such as lamps. Within local territories served by utility companies, of course, there were significant fixed cost requirements for generation and transmission, and these were sufficient to produce some ‘exclusion effects,’ or ‘first-move advantages,’ which created incentive for racing between the sponsors of rival systems.36 But, as has already been noted, rather than being symmetrically positioned with regard to every marker, d.c. and a.c. systems each could find some markets in which they would enjoy some advantages in preempting entry by the other. Furthermore, such economies of scale as each could exploit in the generation of current remained definitely bounded by rising marginal distribution costs, with the geographical bounds due to transmission costs being more tightly constricted around the central generating stations in the case of d.c. systems. On the other side of the ledger, the possibilities of enjoying economies of scale in generation by achieving greater ‘load diversity’, and consequently higher load factors for plants of given capacity, were far more limited for a.c. systems at this stage. The latter were still restricted to serving only the segment of the market (for incandescent lighting) that was characterized by very high peak-load in power usage. In the case of residential lighting demands, the load factor—defined as the ratio of the average load to the maximum load of a customer, group of customers, (p.327) or an entire system during a specified period—typically might be as low as 10–20 per cent.37

All things considered, in 1886 Edison was perhaps quite justified in his rather sanguine view of the challenge represented by a.c. and the recent entry by Westinghouse into the central station business. November of that year found him writing to Edward Johnson:

‘Just as certain as death Westinghouse will kill a customer within 6 months after he puts in a system of any size. He has got a new thing and it will require a great deal of experimenting to get it working practically … None of his plans worry me in the least; only thing that disturbs me is that Westinghouse is a great man for flooding the country with agents and travelers. He is ubiquitous and will form numerous companies before we know anything about it.’38

One cannot be sure of the bases for Edison's sanguine assessment of the situation at this point. Had he been thinking primarily of the commercial advantages that his d.c.-based enterprises derived from their headstart in acquiring experience and expertise in engineering design, component manufacturing, and operation of central stations supplying incandescent lighting, he could have viewed the safety problem as indicative of some of the many practical improvements that remained to be made in the a.c. system. Quite possibly he had in mind also the disadvantage at which the a.c. lighting technology would be placed in competing against a more comprehensive (‘universal’) electrical system which could supply lighting, power, and traction customers—such as the one that was beginning to be implemented on the basis of d.c. under Edison's sponsorship.39

Data on the number of central station and lamps associated with each of the rival formulations suggests that by the years 1888 and 1889 the diffusion of a.c. technology was rapidly catching up with that of the d.c. technology in the United States. As of October, 1888, the Westinghouse Electric Co. already could count 116 central stations (some of which, however, actually were d.c. plants) with a total capacity to run 196 850 lamps, in comparison with the 185 Edison central stations operating 385 840 lamps. By 1891 a.c. had pulled ahead in the area of lighting, but the available estimates of generating capacity for that date indicate that when electricity supply for power (p.328) and traction is included, d.c. remained preponderant.40 The catch-up that occurred during the latter 1880s mainly reflected the a.c. companies' move into the electrification of smaller cities and towns not well served by d.c., rather than penetration into the Edison system's ‘natural’ territory in the spatially more dense urban lighting market.

There surely were some geographical markets in which the two variants were approximately balanced in their advantages, at least from the viewpoint of the cost of supplying electricity for lighting—larger territory for a.c. being offset by the impossibility of load balancing through connections to industrial power users, and traction companies. In such circumstances one might well expect that competition would feature the use of marketing tactics of all kinds, designed to tip the balance in one direction or another between competing system sponsors. And, in fact, it is precisely against this background of an apparent ‘technological stand off’ in the late 1880s that historians of technology and economic historians who have studied the electric supply industry have set the ensuing episode known as the ‘battle of the Systems.’ Nevertheless, the events which marked this period of intense and open rivalry between the proponents of direct and alternating current only become fully comprehensible, in my view, when they are seen to have been precipitated by a fundamental transformation of the underlying techno-economic situation. That disruption of the status quo ante occurred in 1887–1888.

4. The Electricity Supply ‘Battlefield’ Revisited

As it has been recounted by more than one historian of technology and business enterprise, the ‘battle’ that burst into public view during 1887–1892 was remarkable—even bizarre in some aspects—and regrettable in transgressing the normal boundaries for either market competition between different formulations of a new and highly promising technological paradigm, or professional disputation among scientists and engineers with regard to their comparative technical merits.41

Spilling over from the market-place and the academy into the legal and political arenas, the ‘contest of the currents’ between d.c. and a.c. took the form of courtroom struggles over patent rights, attempts to pass anti-competitive legislation, and public relations schemes aimed at discrediting the opposition and frightening their customers. The conjuncture of these events in time has contributed to the impression that they all were facets of a concerted and unbridled (p.329) counter-attack launched by the Edison camp, parts of an irrational effort to turn back the tide of technological progress that had brought an unwelcome influx of competition into the electrical supply business. But, some of the temporal coincidences in this instance are rather misleading. So it is necessary to begin by disentangling two of the major strands that appear intertwined in this fabric.

Throughout the late 1880s claims and counter-claims regarding infringement of electrical patent rights flew back and forth, both between members of the d.c. and the a.c. camps, and among parties belonging to the same camp.42 Virtually from their inception the companies associated with Edison, and their financial backers, had approached the development and commercial exploitation of electricity for lighting and other uses within the paradigm of an integrated system—originally conceived of by the inventor through a conscious analogy drawn with existing systems of lighting, based upon the generation and distribution of illuminating gas.43 The owners of these companies had an obvious collective motivation to block competition from a major variant system. In claiming patent rights to some components that were utilized by the a.c. alternative, the Edison interests might well have hoped to delay the marketing of a rival integrated electrical supply system, if only by inducing would-be competitors to take the time and trouble to ‘invent around’ Edison's patents. Indeed there were several occasions upon which Westinghouse's company was induced to go to great lengths in circumventing Edison patents, particularly those linked to the incandescent lamp.44

Legal contests of this form are not costless, however; time, energy and money were expended by all parties involved, and the costs often were deemed particularly steep by the more talented inventors, who were drawn away from their laboratories in order to defend proprietary rights to previous inventions—sometimes even though they had ceased to hold a major ownership stake in those putative rights.45

(p.330) Edison's personal financial stake in patents he had received relating to lighting diminished greatly from the outset of his incandescent lamp project in 1878, when he had slightly over a 50 percent share; by 1886, after several rounds of raising more capital—first for the lighting enterprise, and then to expand his companies engaged in manufacturing components—he had sold nearly all his shares of the Edison Electric Lighting Co. The latter was the paper entity which legally held the patents and licensed use of the devices they covered to local light and power utilities.46

Throughout 1880–1885, neither Edison or the holding company's directors showed much interest in litigating over patent infringements, even though Edison maintained that his incandescent lamp invention preempted the claims of both Sawyer-Man and Maxim. Edison had been involved in enough patent litigation to know that such battles could stretch out over many years and entail very considerable expenses; it was also true that the Edison lamp at this time was far superior to the others being offered, and that during the pioneering phase of their development of the d.c. electric supply system the Edison people were more concerned with competing against gas than against the sponsors of rival electric lighting systems.

Moreover, the subsidary enterprises engaged in manufacturing dynamos, motors, conducting mains and components, lamps and other appliances required for lighting plants using the Edison system, were proving quite profitable even in these early years. The manufacturing part of the business constituted the primary source of income not only for the inventor, but also for his original colleagues at Menlo Park; Johnson, Batchelor, Upton, and Bergman had become responsible for the running of these subsidiary enterprises in which they had co-ownership interests with Edison.47

Beginning in 1885, however, the Edison Electric Light Co's passive stance vis-à-vis patent infringers was altered. At that time a reorganization of the board of the holding company occurred, which resulted in a weakening of the power of Edison and his original co-workers, particularly Batchelor and Upton. The new board were almost entirely representative of the financiers behind the company, and so were primarily interested in protecting the revenues derived from the proprietary rights of the Edison patents. A policy decision was therefore made to go after the infringers; law suits were initiated and customers were notified that the other sellers of lamps would be prosecuted.48 The patent fights that reached the courts in the late 1880s and dragged on into the early (p.331) 1890s thus had been set in motion well beforehand and bore little direct connection with Edison's own responses to the intrusion of competition from the a.c. suppliers.

Of course, the patent litigation was only an aspect of the conflict and not the one that has caused the ‘Battle of the Systems’ to be characterized as bizarre. The most striking events of this episode revealed the lengths to which Edison and his immediate associates were prepared to go in order to convince the public that alternating current was an unsafe basis for an electricity supply system. Their campaign was waged through a barrage of ‘scare’ propaganda, supported by the grisly ‘evidence’ Edison and his immediate associates produced by experimenting with a.c. at their laboratory in West Orange, NJ. As Edison's biographer, Matthew Josephson, relates:

There, on any day in 1887, one might have found Edison and his assistants occupied in certain cruel and lugubrious experiments: the electrocution of stray cats and dogs by means of high tension currents. In the presence of newspaper reporters and other invited guests, Edison and Batchelor would edge a little dog onto a sheet of tin to which were attached wires from an a-c generator supplying current at 1,000 volts.49

In July, 1888, Harold Brown, a former Edison laboratory assistant, electric pen salesman, and self-styled ‘Professor’, put on a demonstration of the harmful effects of a.c. at Columbia College's School of Mines: the electrocution of a large dog was featured.50 Meanwhile, a scarlet-covered book had been issued under the title ‘A Warning from the Edison Electric Light Company,’ in which competitor-companies were accused of patent theft, fraud, and dishonest financing (including personal attacks on George Westinghouse); it gave descriptive details of the deaths, and in some instances the cremation by electrocution of unfortunates who came into contact with wires carrying alternating current.51

During 1888 the ‘West Orange gang,’ consisting of Edison, Johnson, and young Samuel Insull, aided by Brown and other assistants, succeeded in bringing off a related, stunning achievement in the art of negative promotion: they convinced the State of New York to substitute electrocution by administration of an alternating current for hanging as the means of executing convicted criminals. Edison himself had lobbied for this action before the New York legislature, and Brown—engaged as a consultant to the State on capital punishment by means of electrocution—surrptitiously purchased three Westinghouse alternators which he then announced had been selected as the (p.332) type most suitable for such work.52 The Edison group proceeded to milk the legislature's decision for all the publicity it was worth, circulating leaflets which, in warning the public about the dangers of high voltage a.c., used the term ‘to Westinghouse’ in referring to electrocution by alternating current.53 Now, some warnings concerning the dangers involved with the new technology would not have been unwarranted when alternating current was first suggested as the basis of an incandescent lighting system. The source of the problem, however, lay not in the nature of the current, but rather in the fact that the proposed a.c. system would transmit energy on its mains at a higher voltage; direct current is always more deadly than alternating current at an equivalent voltage.54

Back in the fall of 1886, when the Westinghouse Electric Co. commercially introduced its alternating current lighting system, Edison therefore was only being overly optimistic in predicting that there soon would be some accidental electrocutions of Westinghouse's customers, which would bring about the new technology's natural demise. His view of the danger at the time was supported as reasonable by some experts within the a.c. camp.55 In fact, Professor Elihu Thomson, co-founder of the Thomson-Houston Company, of Pittsburgh, PA, advised his company to refrain from marketing an a.c. lighting system for home use until better safeguards were developed; this, despite his having already employed alternating current in the company's commercial arc-lighting systems, and his development of an a.c. incandescent lighting system concurrently with the one that Westinghouse was proceeding to market. As one of Thomson's biographers has put it: ‘There was enough truth in the contention [that h.v.a.c. is deadly] to hold the Pittsburgh firm back and use up much of its time and money in making counterclaims [to those of the Edison forces].’56 Thus it was that while the Westinghouse Electric Co. and Edison-interests were engaged in head-to-head battle in the incandescent (p.333) lighting field, the Thomson-Houston firm held off for a year and developed a simple method of preventing accidental electrocution in the unlikely event a transformer short-circuited.

Following this safety improvement, however, the Edison group did not relent in their propaganda campaign. Right through to 1889 Edison continued to participate in publicizing the undesirability of high voltage a.c. on grounds of safety, by criticizing the technical means given for rendering a.c. less hazardous. In an article on ‘The Dangers of Electric Lighting’, published originally in the North American Review, Edison stated that the undergrounding of high voltage line (an often suggested means of increasing their safety) would serve only to make them more dangerous; further, he claimed that there was no means of effectively insulating overhead high-voltage wires, and concluded that ‘the only way in which safety can be secured is to restrict electrical pressure (voltage).’57

With the benefit of hindsight, and a seeming pre-disposition towards the ‘Whig’ interpretation of technological history, previous chroniclers of this episode have been almost unanimous in portraying Edison as economically irrational in stubbornly championing the direct current system, and uncharacteristically, but nonetheless deplorably unscientific in his dogmatic public opposition to the rival alternating current technology sponsored by Thomson-Houston and the Westinghouse Electric Company.58

Harold Passer described Edison's transformation bluntly in the following terms: ‘In 1879, Edison was a bold and courageous innovator. In 1889, he was a cautious and conservative defender of the status quo.’59 More severe is the judgement offered by Jonathan Hughes' spirited account, according to which, by 1886–1887, Edison

had lost touch with the rapidly changing technology, or was fast losing touch. He had been a staunch, and then a rabid, opponent of alternating current transmission. … For some reason Edison could not comprehend the a.c. system. He was convinced that, transformers or not, the high-voltage of the a.c. system made it extraordinarily dangerous. … Arguments got nowhere with Edison on this issue. The fact that he considered the main proponents of the a.c. system to be (p.334) common thieves made him even more unwilling to see any virtue in their arguments. … There was no part of Edison's career that was so unworthy of the man and, in fact, sordid.60

Now, the reputation of Thomas Edison in American history-books is in all other respects so little in need of rehabilitation that there could scarcely be much call to repair it in this one regard, even if such a defense were plainly warranted by the facts. The concern here, accordingly, lies less with the way the conduct of the ‘Battle of the Systems’ has colored modern appraisals of Edison the man, and more with the proper appreciation of the strategic goals and constraints which shaped the decisions that he and his close associates made and implemented in this episode. Initial safety concerns about the hazards of a house-to-house incandescent lighting system based on (high voltage) a.c. were not irrational in view of the preceding history of faulty installations and consequent injuries with high voltage arc-lighting systems.61

Yet, practical experience in the industry soon revealed that grounding and improved insulation methods worked quite reliably.62 It is therefore difficult to credit the view that Edison's original safety concerns were completely unwarranted, and that he continued to pronounce a.c. ‘unsafe’ because he remained ill-informed, or foolishly obstinate, or both. Indeed, so far was he from spurning high voltage as an unacceptably risky technology that the Scientific American for July 23, 1887 carried an article describing a new Edison (p.335) high voltage d.c. distribution system. Presumably then he had a purpose in going on with the safety campaign against a.c. until 1889, but not continuing it beyond that year. Not surprisingly, the existence of some deeper, goal-directed strategy has been supposed even by some historians who have expressed their strong disapproval of the tactics used.63

An immediate purpose is not too hard to find. The objective of the animal execution demonstrations, the initiation of the electric chair, and the volumes of published material detailing the actual and potential hazards of a.c. was not, however, that of gaining sales in contested markets by influencing the choices made by purchasers of lighting systems, or their ultimate customers. Rather, the proximate goal was the creation of a climate of popular opinion in which further inroads by a.c. systems in the Edison companies' share of the incandescent lighting market might be prevented by legislative restraint, ostensibly justified on grounds that the public's safety needed protection. The gruesome promotional campaign was designed to persuade lawmakers to limit electric circuits to 800 volts, thereby stripping the rival system of the source of its transmission cost advantage. In at least two states, Ohio and Virginia, these efforts came closer to succeeding than in New York, where, having adopted the electric chair, the legislative balked at the intended sequel.64

If the events just recounted cannot be dismissed merely as the spiteful reactions of a short-sighted inventor who was being left behind in the onward march of technology, or simply as an unseemly tussle over marginal market territories between the respective suppliers of d.c. and a.c. lighting equipment, how should they be understood? What was Edison's underlying purpose, and, when was it formed? If Edison was quite sanguine in the face of the commercial initiation of Westinghouse's lighting system in Buffalo, NY, as his communication with Johnson in November of 1886 seems to indicate,65 what had triggered the launching during the following year of the ‘sordid’ propaganda campaign aimed at crippling the a.c. technology with safety legislation? And, having once committed himself to this mode of competition, why did Edison discontinue it after 1889?

Although definitive answers to these questions about the motives of Edison and his associates are elusive, and may remain so even when the contents of the Edison Archives covering these years have become fully accessible, his actions can plausibly be construed as an economically rational, albeit cynical response on the part of an inventor-entrepreneur whose long-term plan to be the sole-sponsor of a ‘universal’ electrical supply system suddenly had gone away. On this view, Edison, far from being foolishly obtuse, was more likely (p.336) than any of the interested parties to have perceived during 1887 that the world in which he was operating had been abruptly altered; specifically, that a number of unexpected and serious blows had been dealt to his previous hopes of profiting greatly from control of the key technological components of a system that could supply energy for public arc-lamps, residential and commercial lighting, industrial power, and traction. This sudden reversal of his fortunes in the direct current electricity business came from the concatenation of two fundamentally unrelated sets of developments. One was a clustering of technological innovations that opened the way for alternating current to become the basis of an alternative, fully competitive universal electricity supply system; the other was the deterioration of Edison's personal financial position, along with that of the companies in which he was most directly involved. We have now to consider the nature of these critical technological and financial developments, taking them in turn. Coming together, they precipitated a course of action well-designed to salvage for Edison and his immediate associates as much as was possible before he would leave the industry to embark upon other ventures.

5. The Advent of Polyphase A.C. and Its Significance

The original ‘stand-off’ situation that had left d.c. and a.c. in 1885–86 with respective markets within which each held a clear, technically derived advantage was in reality a condition of transient equilibrium, a temporary balance soon disturbed by the realization of a differentially faster rate of technological advance along the a.c. system trajectory.66 In the alternating current field, three new and interrelated innovations abruptly removed the disadvantages that had formerly restrained a.c. supply companies from penetrating the d.c. lighting and power markets: (1) the induction motor, (2) the a.c. meter, and (3) the rotary converter.

Between 1885 and 1888 Nikola Tesla, in the United States, Galileo Ferraris of Italy, and Michael Osipowitch von Dolivo Dobrowolsky, in Germany, each had discovered that two alternating currents differing in phase by 90 degrees, or a three alternating currents differing in phase by 120 degrees, could be used to rotate a magnetic field; that by placing in such (p.337) a field a pivoted bar of iron, or a drum of laminated iron on which are wound self-closed coils of copper wire, one could construct a two-or three-phase induction motor.67 Invention of the motor led directly to the development of a special polyphase generator and system of distribution.

Tesla's first two patents on the alternating-current motor were filed in October of 1887, and three more patents pertaining to the system were filed in November. The Westinghouse Electric Co. acquired these patents in July 1888, by which time Westinghouse already had persuaded Ferraris to permit it to file an American patent on the rotating magnetic field motor which he claimed he had built three years before.68 Coincidentally, in April 1888, Oliver B. Shallenberger, an engineer working for Westinghouse, came upon the polyphase induction phenomenon and applied it to solve the problem of designing an effective meter for alternating current; by June he had shown that such meters could be built in quantity. Hence, by 1889, Westinghouse Electric Company was able to put into commercial production a meter which enhanced the scope for a.c. in central station lighting operations, and the technological basis for penetrating portions of the electric power market.69

While a.c. induction motor and meter posed an obvious threat, they did not immediately give rise to commercial challenges to Edison's manufacturing companies in urban markets where d.c. power supply was already established. Despite the greater than fifty percent market share that a.c. had achieved in the field of lighting by 1891, there was a considerable amount of momentum built into the growth of d.c. supply systems. Two forces were behind this momentum. First, where a local electric supply system did constitute a natural monopoly and there already existed an installed d.c. generating and distribution system, the unamortized investment in d.c. plant was so large that it discouraged replacement with polyphase a.c.; by adopting a mixed system the local utility would lose the scale advantages associated with a single system.70 Second, as long as there was an important portion of the urban market held firmly by d.c.—namely, the market for traction power, and to a lesser degree industrial electrolysis, as distinguished from that for secondary-motor power for elevators, factories and other such applications—the local d.c. central stations meeting this (traction) demand would necessarily also supply the other (p.338) electricity needs in the area; rivals could not penetrate the market because the d.c. system with its load balancing advantage could under-price them. Hence, unless some ‘gateway,’ or interface technology emerged, a partial commitment to perpetuation of the d.c. technology would remain for some time to come.

Viewed from this perspective, it was another technical innovation, part of a new ‘gateway’ technology coincident to and dependent on the development of polyphase a.c., that was critical in delivering the coup de grace to a comprehensive electric supply system based entirely on a d.c. technology.71 The rotary converter was a device which combined an a.c. induction motor with a d.c. dynamo to make possible the connection of high voltage a.c. transmission lines to d.c. distribution networks. A former Edison employee, Charles S. Bradley, who already had applied for a patent on a polyphase generator and a synchronous motor in March 1887 (actually before Tesla), successfully patented the rotary converter in the United States during the following year.72

The major significance of the rotary converter lay in the fact that it enabled the old d.c. central station and traction distribution networks to be coupled with new long distance high voltage a.c. transmission mains. These ‘gateway devices’ thereby made possible the formation of a more flexible ‘hybrid’ system, the advantages of which were recognized in the electrical engineering press as early as 1887.73

By the early 1890s Bradley had set up his own factory to produce the converter in Yonkers, NY, a plant that would soon be acquired along with the patent by the General Electric Company. The potential profitability of the business of supplying rotary converters likewise drew the Westinghouse Electric Company into further development work on such devices.74 Converters were also developed to couple existing single-phase a.c. with the newer and more efficient polyphase technology. In fact, by the middle of the 1890s there would exist devices to convert in any direction. Conversion of d.c. to polyphase a.c. proved an immediately attractive application in some locales, such as Chicago, where Samuel Insull saw it would permit raising the load factors on existing d.c. plant by transmitting current over a much more extensive area in which the load was more diverse.75

(p.339) What had happened during 1887–88, in essence, was the dramatic appearance of a new technological variant—the polyphase a.c. system—induced by the opportunities inherent in the limitations of d.c. and singlephase a.c. as bases for a ‘universal’ electricity supply system. It was the polyphase a.c. technology that would diffuse rapidly, becoming the de facto network standard for electricity generation and transmission in the United States, penetrating the core urban markets for electric light and power, and thereby realizing the greater economies of scale which fostered the emergence of extensive ‘natural monopolies’ in the electric supply industry.

Although Edison may not have foreseen the whole evolution of the network technology that would be created to exploit the transmission cost advantages of alternating current, his course of action from 1887 onwards reflected, in my view, an astute grasp of the precarious, unstable nature of a competitive situation that was unfolding with unexpected speed from the successful experiments with polyphase a.c. motors. That perception would have reinforced whatever other considerations might have disposed him to leave the electricity industry, instead of girding himself to participate in an inventive race against the emerging a.c.-based universal system.76

As Thomas Parke Hughes has emphasized, Edison's talents inclined him toward the invention of devices that were interrelated within a system context, and he naturally preferred to work on components whose improvement would lead to enhanced performance and value throughout a system whose parameters he could control—and thereby draw the greater profit from.77 Yet, in the circumstances of the electricity supply industry during 1887–1888, reasonable expectations of extracting significant rents on any incremental, strictly d.c.-compatible inventions had largely been vitiated by the developments leading to Bradley's rotary converter. Furthermore, the d.c. system elaborated under Edison's sponsorship was now at a disadvantage; additional research and development resources would have to be devoted to reducing distribution costs, just for it to be able to hold its own in competition with (p.340) polyphase a.c. systems. What profit to an inventor undertaking that uncertain mission, when the limiting value of the improvement would be that set by the cost of the rotary converters installed to transform d.c. to a.c. and/or a.c. back to d.c. for local distribution and traction uses? Edison had more tempting projects in which to engage his laboratory staff and his own inventive genius.78

The unorthodox and rather desperate tactics adopted by the ‘West Orange Gang’ in their ‘safety campaign’ against Westinghouse takes on a different appearance when set in this context, especially their focus upon invoking some regulatory intervention to deprive alternating current-based systems of the transmission advantages deriving from use of high voltages. The campaign looks more like a temporary ‘holding operation,’ meant to delay the competition so as to permit Edison to stage a more orderly and profitable exit from the industry, and less like a serious counter-offensive. Had it been Edison's hope and intention to permanently cripple the competitive system of electricity supply, it would hardly have made sense for him to accelerate his withdrawal from the business of manufacturing and selling the key components of the d.c. system that would be left in command of the field. Yet, that is precisely what he was proceeding to do.79 It was a decision, however, that owed something also to the financial straits in which Edison unexpectedly found himself just when he was moving into research and development projects in other areas, pursuing costly undertakings that were still quite far from the commercialization stage.

6. The End of the Sponsored-Systems Rivalry

Edison's decision to get out of the electricity business was significant, because it would lead shortly to the disappearance from the United States' electrical (p.341) manufacturing industry of a commercial sponsor having proprietary interests exclusively in the d.c. technology.80 It had its roots not only in the burst of polyphase a.c. developments just reviewed, but in the evolution of Edison's relationship with the holding company (Edison Electric Light) in charge of his lighting patents and the various entities such as Edison Lamp, and (Edison) Machine Works, that actually manufactured the components of the system and serviced them. Recall that by the mid-1880s Edison and his immediate associates had little stake in, and less control of the holding company that drew royalties on the use of the lighting patents by central station companies and other companies set up to license the construction of isolated lighting plants; whereas, Edison remained the principal owner of the factories, from which he and his immediate associates had been drawing their main income. During 1886–1888, however, the precarious financial situation of the latter group of enterprises came to be perceived as tremendously burdensome to their owners. Much of the equipment supplied previously by the manufacturers to central stations had been paid for with hard-to-negotiate shares in those fledgling enterprises, resulting in severe cash flow problems for the Edison concerns. With the recovery from the business recession of 1885, electrical equipment orders were coming in faster than the factories had resources to produce and deliver. But, it was proving difficult to finance the expansion on short-term credit, and to solve this problem by permitting the Morgan banking group (who already dominated Edison Electric Light) to extend their control over the manufacturing enterprises in exchange for long-term loans was hardly an appealing prospect. Thus, at the same time that Edison was fretting over the possibility that his over-expanded manufacturing businesses might be forced into bankruptcy, he was also sending his financially skillful young associate, Samuel Insull, off to borrow the extra sums needed for the West Orange, NJ laboratory's expanding program of non-electrical researches.81

Consequently, the return of Henry Villard from Germany in 1886, bearing a commission from the Deutsche Bank to negotiate with Drexel, Morgan and Company about the acquisition of holdings in American businesses, came at a most fortuitous moment. Edison was much relieved to be presented soon thereafter with Villard's proposed plan for the consolidation of all the Edison-related enterprises (including the patent-holding company, and Sprague Electric Railway and Motor Company) into one new corporation with backing from the Deutsche Bank, the Allgemeine Elektrizitats Gesellschaft, and the firm of Siemens and Halske, of Berlin. The inventor saw in this a welcome (p.342) opportunity to both extricate himself from the worries and distractions of managerial and financial responsibility for the manufacturing business, and raise sufficient capital to place his laboratory on firmer financial foundations.82 Ultimately, the terms to which J. P. Morgan was willing to agree turned out to be somewhat less favorable to Edison and the manufacturing company owners than those initially proposed, and, not surprisingly, more generous to the holders of the lighting patent with whom Morgan was directly involved. Nevertheless, they gave the inventor $1750 000 in cash, 10 per cent of the shares, and a place on the board of the new company.83

In this way the Edison General Electric Co. came to be organized in January 1889. Within a few months the consolidations were formally effected, and Edison himself no longer had direct influence in the running of the manufacturing side of the business. By 1890 he had largely completed the liquidation of his remaining 10 per cent shareholding in the new company, was taking no active role on its board of directors, and was writing to ask Villard not to oppose his ‘retirement from the lighting business, which will enable me to enter into fresh and congenial fields of work’.84 The propaganda war against the Westinghouse a.c. system, which had been brought to its peak in the midst of the consolidation negotiations, was rapidly wound down in 1889. It would seem to have served the real purpose of supporting the perceived value of the Edison enterprises which were at the time wholly committed to manufacturing the components of a d.c.-based electric light and power system, and thereby improving the terms on which Edison and his close associates were able to ‘cash out’.

Elements within the American financial community, among which the Morgan interests were most prominent, were moving at this time to consolidate and hence ‘rationalize’ another network industry—the railroads.85 The control and consolidation of the electrical manufacturing and supply business represented a parallel undertaking. A major step was effectively accomplished by joining Thomson-Houston Company and the Edison General Electric Company; in 1892 the first of these, under the leadership of Charles A. Coffin, received Morgan's support in buying out the second, thereby forming the General Electric Company.86 This turn of events meant that by 1892 (p.343) Edison—who had earlier been adamantly opposed to the idea of forming a combine with Thomson-Houston and doubtless would have remained so—had withdrawn entirely from the business, so that no solely d.c.-oriented manufacturing entity existed in the American market.

Westinghouse's enterprise, however, was able to elude the Morgan group's aspirations for an all-encompassing ‘rationalization’ of the industry, paradoxically because its shaky condition had forced it to put its affairs in order and line up banking support from other quarters at the very beginning of the decade. Finding his business under-capitalized to weather the aftermath of the Baring ‘crisis’ in 1890, and unable to obtain a half-million dollars on satisfactory terms from the Pittsburgh business banking community, George Westinghouse was obliged to turn for backing to a New York-based financial syndicate headed by August Belmont and to reorganize his company.87 Work on the development of Tesla's induction motor and 3-phase system was brought to a halt during these difficulties in 1890–1891, but, it soon was resumed—once Tesla had been induced not to hold Westinghouse to the terms of the royalty agreements concluded between them in 1888–1889.88

The electrical manufacturing business in the United States thus came to be dominated by two large firms from 1892 onward, but the industry also had become essentially homogeneous with regard to the basic formulation of its technology. By 1893, the General Electric and the Westinghouse Electric and Manufacturing Companies both were marketing some version of a polyphase alternating current system, and both had entered the profitable business of manufacturing rotary converters.89 The era of rivalry between commercial (p.344) sponsors of technologically distinct systems in the United States' electrical supply industry was brought to a close within six years of its commencement.

7. Diffusion of the New Technology—and the Path Not Taken

Yet, the question of the superiority of one form of current over the other remained unresolved within the engineering community. Whether direct or alternating current was to be generated by the hydroelectric power project being undertaken at the Niagara Falls was still very much an open question during 1892.90 The proponents of d.c. at Niagara had argued that for conditions of varying load, as was the case in a lighting system, d.c. was much the more efficient of the two.91 While this may have held true in 1890, following the 1891 demonstration by Oscar Muller and the Swiss firm of Brown Boveri & Co. that polyphase current could be transmitted the 110 miles from Lauffen on the upper Neckar River to Frankfurt-am-Main, Germany,92 and the equally impressive Westinghouse Electric Company polyphase system exposition (including the rotary converter) at the Chicago World's Fair in 1893, it was evident that lighting was no longer the only factor to consider when discussing the load and efficiency characteristics of the a.c. and d.c. system variants.93

With the extension of a.c. to power users, and to traction users as a result of the invention of the rotary converter, the decision between a.c. and d.c. came down to the one which could distribute power over a distance most efficiently and cheaply. And, as the distance from Niagara to the nearest concentration of customers in Buffalo, NY, was 20 miles, a.c. could reduce the loss of power on transmission lines to a far greater extent than was possible with d.c. In 1893, both the Westinghouse Electric Company and the newly formed General Electric Company submitted plans to the Cataract Construction Company (which was pioneering the Niagara development), specifying an a.c. system consisting of generators, transformers, and transmission lines.94

Beginning in 1896, central stations began being converted into substations hooked up to a.c. transmission lines, and by 1898 a constant current transformer had been developed to make possible the linking up of arc lighting (p.345) distribution networks with a.c. transmission lines.95 Hence, the flexibility and capabilities of this new coupling technology led to a rapid diffusion of the a.c. polyphase technology and the integration of smaller urban electricity supply systems into larger networks which eventually formed regional grids.96 Once the systems competition had tipped in this direction, lighting plant also came to be replaced with a.c. technology as the previous d.c. distribution networks wore out; other things being equal, transformers stepping-down high voltage a.c. were a much simpler and cheaper technology for lighting purposes than the use of rotary converters to feed local d.c. distribution networks from high voltage a.c. transmission lines.97 Converters continued to be employed well into the 20th century, however, most notably in the traction field where d.c. remained the current preferred at the point of consumption.98

A de facto standard in the form of alternating current as the basis of a ‘universal’ electrical supply system had emerged by the 1920s, both in the United States and abroad. While diffusion data for the 1890s is not available, the figures assembled in Table 1 show that in America the fraction of central station generating capacity accounted for by a.c. rose from 69 percent in 1902 to 95 percent in 1917. Moreover, one can discern in this the large role played by the rotary converter. This appears from the absolute rise in rotary converter capacity installed and also from the fact that the d.c. share of end-use capacity fell far less sharply than its share in generating capacity. Indeed, the former remained essentially unchanged after 1907. At the engineering level, therefore, the ‘battle of the systems’ did not end with the capitulation and withdrawal of one of the contenders, as the Edison-Westinghouse business rivalry had done. The technological denouement has been described by Thomas Parke Hughes99 as a peaceful resolution to the conflict: no outright defeat for d.c.; rather, a graceful and apparently efficient absorption within a transitional ‘mixed’ system, prior to its gradual disappearance from the American electrical scene.

The perspective of hindsight may impart to this story an impression of inevitability, and even a supposition of optimality, both of which should be resisted. Meaningful global evaluations of efficiency are difficult if not impossible (p.346)

Table 1. Distribution of Generating and End-Use Electric Capacity between Direct Current and Alternating Current, Excluding Power Generation by Electric Railways, in the United States, 1902–1917.

Kilowatt Capacity (thousands)

1902

1907

1912

1917

Dynamos:

d.c., constant voltage

330.1

406.5

432.4

418.6

a.c. and polyphase

736.3

2221.8

4689.2

8557.7

Rotary converters

232.2

363.4

1009.1

1898.6

Transformers

560.4

1693.5

4103.9

9499.0

d.c. share in total generating capacitya

0.31

0.16

0.08

0.05

d.c. share in end-useb capacity

0.53

0.29

0.28

0.26

Ratio of transformer capacity to a.c. capacity

0.76

0.76

0.88

1.11

(a) d.c. share in generating capacity is calculated by dividing d.c. dynamo capacity into the sum of d.c. and a.c. dynamo capacity.

(b) d.c. share in end-use capacity is calculated by subtracting rotary converter capacity from the a.c./polyphase dynamo capacity, adding rotary converter capacity to d.c. dynamo capacity, and then recalculating d.c.'s share of total capacity.

Sources: Dynamo data: Bureau of Census (1920), Table 40, p. 63 for 1907, 1912, and 1917; transformer data: Bureau of Census (1920) Table 110, pp. 170–1, for 1917; Bureau of Census (1915) Table 65, p. 104 for 1912; Bureau of Census (1906) Tables 118 and 119, pp. 134–7 for 1902. Note that the horsepower data from Tables 74 and 75 has been converted to kilowatts according to 1 h.p. = 0.746 k.w.

to make between alternative technological systems whose influences ramify so widely and are so profound that they are capable of utterly transforming the economic and social environments into which they have been introduced.100 As a guard against the strong temptation to suppose in matters of technological advance that ‘what is, ought to have been’, it is always useful to at least notice the existence of other paths that were not taken.

From the technical journals and magazines of the period it is apparent that no immediate consensus emerged on the engineering merits of the two currents; there were well-respected inventors and scientists, many of whom were founders of the industry, who would not testify to the technical superiority of alternating current.101 For, the d.c. technology also showed itself quite capable of being (p.347) further elaborated, in directions that both heightened its special advantages and broadened the range of conditions under which it was economically competitive. As was the case with a.c., the possibility of lowering the cost of the d.c. systems by raising the voltage was being actively explored. As early as 1883 Charles F. Brush had attempted a high voltage d.c. system that could more fully utilize fixed generating capacity and increase the radius of profitable transmission, by using ‘accumulators’ (storage batteries) to handle some of the peak lighting load and to accomplish the reduction of voltage for local distribution. Due to a combination of problems associated with the battery technology available at the time, the dangers of operating a high potential system having dynamos wired in series, and the usual run of financial difficulties, this particular project never took off and the concept was not pursued further in the United States.102

Abroad, however, the English from the mid-1880s onwards attempted to implement a variety of h.v.d.c. electricity supply schemes.103 The earliest of these met much the same fate as that of Brush—too many troubles with expensive primitive batteries and a myriad of financial woes arising from inadequate demand and insufficient capital. But several innovations introduced late in the 1880s did prove successful. Notable among these was the ‘Oxford system,’ an approach first employed in 1889, which transmitted high voltage d.c. to substations where it was ‘bumped down’ to usable voltage levels via either a battery arrangement or a direct current transformer.104 On balance, the major advantages of the h.v.d.c.-battery technology lay in ensuring continuity of supply when generating plants failed or were shut down for maintenance, and in reducing the amount of fixed capacity required to meet peak loads on the system.105 In Britain, then, the story unfolded along lines very different from those in the United States; there, the ‘competition of the (p.348) currents’ was tipped during the 1890s towards d.c. by the possibility of using accumulators in combination with high voltage.106

Concurrently, on the European Continent, M. Thury was developing a system of transmitting direct current at very high voltages from constant current generators worked in series, and commonly coupled mechanically in pairs or larger groups driven by a single prime mover. This offered advantages in easier line insulation than was required at half the voltage with alternating current, and removal of difficulties of line inductance and capacity encountered in high voltage a.c. transmission. High voltage constant current plant lent itself to greater ease of operation in emergencies (over a grounded circuit, for example) and permitted the design of comparatively simple and inexpensive switchboard arrangements. Notwithstanding the fact that the direct current generators used in this system were relatively expensive and their individual output was inconveniently small for large transmission work, around 1910 even a contemporary American authority gave it as his view that ‘the possibilities of improvement in the system have by no means been worked out, and although it has been overshadowed by the enormous growth of polyphase transmission it must still be considered seriously.’107

Unlike the battery-using h.v.d.c., the a.c. version of the ‘universal’ electricity system concept inexorably pushed Samuel Insull and others who pioneered it in the US towards ‘load balancing’ as the way to mitigate the wastage imposed by having to build enough generating capacity to meet peak loads. The search for a diversified load over a wider region, with high fixed costs in place, created problems of natural monopolies which would not have existed to the same degree under the h.v.d.c. battery technology. Of course, it must be acknowledged that without strong increasing returns effects via load balancing, a d.c. battery-d.c. technology might simply have allowed more leeway to the forces making for too great a degree of diversity in voltages, current and a.c. frequencies. Such could be said to have been the experience (p.349) of the industry in Britain. Yet, if regulatory intervention is accepted as a proper solution to the natural monopoly problem which soon arose in the United States in the case of electric utilities, presumably public intervention could have imposed some standardization of d.c. voltages to permit realization of scale economies in the production of motors, lamps, and other end uses. Moreover, further down the road, when social efficiency was deemed to be achieved through the development of a larger network or ‘grid’, the US state and local regulatory structure which by then had actually been imposed in response to the condition of local monopoly would prove to have discouraged local utilities from integrating and supplying still wider geographical markets.

Such skepticism about the long-run economic optimality and consequent inevitability of the de facto standard that emerged in the United States is, of course, reinforced by the foregoing detailed recounting of its historical roots in ‘battle of the systems’. Given the urgency of the utility companies' drive to achieve scale economies in electric supply, their move to the polyphase a.c.-based ‘universal’ system was certainly little affected by weighing the potentialities for long-run technical improvement offered by alternative systems. Just as short-run liquidity considerations had figured prominently in Edison's decisions during the formation of Edison General Electric, so short-run maximization of rents on the existing stock of proprietary technology seems to be the best algorithm descriptive of the course of action pursued by the dominant successor-firms engaged in manufacturing and marketing electric supply equipment in the US during the years before the First World War.

8. Reflections on Network Technologies and Schumpeterian Entrepreneurs

The ‘battle of the systems’ has been presented by previous narrators as a colorful and cautionary tale. Many of them, it seems, would have us find in it the moral that even an individual possessing extraordinarily inventive and entrepreneurial talents may sink to foolish knavery by exchanging the role of technological progress's steadfast champion to become, instead, a defender of the status quo; and a vain fool in the bargain. As we have had to be taught that the progress of invention and technical change is a complex social and economic process which transcends the intentions and efforts of individual men and women, and that technology is best regarded as ‘socially constructed’—a cumulative result of the work of many minds and hands under the guidance of unseen forces, we naturally suppose ourselves to have a clearer view than that held by the individual participants in the process. In a sense this leaves one pre-disposed to fault even Thomas Alva Edison for his supposed hubris in hoping single-handedly to stay this advance.

(p.350) Although there are contexts in which such a moral is well worth remembering, it does not strike me as the best one to recall in conjunction with this particular episode in technological and industrial history. Indeed, the perspective which recent contributions to economic theory offer on the early phases of the evolution of a paradigmatic network industry like electricity supply should prepare us to recognize the degree to which discrete ‘events’ of a largely adventitious nature, among them the specific courses of action chosen by individual agents occupying key decision-making positions, really do have the potential to set important technological parameters defining the industry's future trajectory. This is not to say that great consequences can be expected to follow from every move made by the drama's principal human actors; only that if we are concerned to understand a thing such as how and why one particular technical variant rather than some other eventually came to be widely adopted and further elaborated, recognition of the presence of ‘localized positive feedback effects’ should make us especially skeptical of modes of explanation that presuppose the inevitability of one outcome and the impotence of individual agents to alter it. The latter, in such circumstances may well have the power to take early actions which, in effect, will turn out to have strongly directed the ensuing course of developments.

In the story as retold here, the actions of Edison and the group of his immediate associates during the years 1887–1892 do not stand out as having run counter to an imminent flow of events leading the electrical supply industry in America away from direct current and rapidly toward a ‘universal’ integrated system designed around polyphase alternating current. Rather than vainly seeking to block the further development of the competitive a.c. electrical supply technology and so preserve a monopoly of the field for his d.c. system, Edison, in our account, sees the juggernaut of a competitive technical system bearing down upon his own immediate economic interests with a swiftness that he had not anticipated. And so, he undertakes expedient actions aimed to slow the pace of its advance enough to allow him quickly to get his inventive resources and financial assets safely out of the way. The propaganda campaign thus launched against high voltage a.c. in general and Westinghouse in particular, with its threat of crippling restraints by safety legislation, made considerable economic sense in the context, however unscrupulous it may have been.

Did it also make more sense than the other, more direct, and probably more reliable modes of commercial competition that were available to the Edison enterprises, but apparently went untried? That remains less clear, for it seems that Edison had reasons for seeking to exit from the industry promptly. A supplier of direct current dynamos, lighting and other appliances and traction motors who had sought to hold onto a dominant market share might have (p.351) moved immediately to explore ways in which the new rotary converter technology could be supplied cheaply for application in lowering the costs of transmission (via high voltage a.c. mains) between d.c.-based generating plant and end-users. As has been shown, this kind of mixed, or ‘patched’ system, which was already being discussed in the engineering periodicals in 1887, was entirely feasible and came to be implemented by electric utility companies in the following decade.

Additionally, or alternatively, ‘promotional’ pricing of d.c. generators and compatible equipment by the Edison manufacturing companies might well have sufficed in the late 1880s to further entrench that system in urban markets. The principal immediate beneficiaries from this short-run revenue sacrifice, it is true, would have been the financiers around Morgan, because it was they who held the rights to the patent royalties on sales of d.c. central stations and isolated electrical plants. But, had Edison wanted to remain in the business, the occasion of the negotiations opened by Villard during 1887–1888 might have been used to arrange to share in some of the royalties that would thereby be secured during the remaining 7–8 years of life on his basic lighting patents. Perhaps it would have proved impossible ultimately to negotiate such an arrangement. What is significant is that there are no indications from the published sources based on the relevant archives that Edison and those around him ever considered it, or were actively exploring other ways of using their substantial initial market position to compete against the commercial sponsors of alternating current equipment. Westinghouse, if not Thomson-Houston also, might well have proved vulnerable to attack at just that point. The former enterprise evidently was under-capitalized to meet the head-on challenge of a price war, so much so that George Westinghouse found himself obliged during 1890–1891 to turn for financial help from August Belmont, and to persuade Nikola Tesla to relinquish his legal claim to a fortune in patent royalties.

Yet, far from fighting to drive out the competition or, at least, to curtail their ability to finance research and development directed towards improving the polyphase a.c. technology, Edison responded to the cluster of innovations related to the a.c. induction motor by refusing to engage the opposition in an economically costly market rivalry; instead, he conducted a rapid but orderly financial withdrawal—under the diversionary cover of a ‘talking war.’ I have suggested that it was not unreasonable in the circumstances, much less irrational, for Edison to have seized upon the Villard proposal to consolidate all Edison-related electrical enterprises, as a fortuitous opportunity to exit profitably from the electrical manufacturing business. The evidence presented here, however, does not speak to the question of whether or not, in withdrawing and turning his attention to the improvement of his phonograph, the movies, (p.352) and mining technology, the inventor was following an expected-private wealth-maximizing strategy. Evidently Edison's decision was influenced by the short-run asset constraints, indeed, by the distracting liquidity constraints, under which both the manufacturing operations and the laboratory at West Orange, NJ were perceived to be working. In any event, had he chosen otherwise, or been prevailed upon by others to remain in the industry and sacrifice short-run earnings in an attempt to block the commercial development of universal electricity supply systems based entirely upon polyphase a.c., the outcome could well have been very different from that which transpired.

The bunching of related induction-motor inventions by Tesla, Shallen-berger, and Bradley, which appears as the most probable cause of Edison's precipitate retirement from the electricity business, had a direct bearing on the outcome, which is also worth keeping in mind.108 The induction motor made polyphase a.c. a rival standard around which to develop a ‘universal’ system of electricity generation, transmission, distribution and application, such as was originally conceived by Edison. This was so in large measure because it permitted comparatively inexpensive conversion from a.c. to d.c. for application to high torque–low r.p.m. motors most suitable for traction work. Yet, by the same principle, the rotary converter could have been used more extensively to expand the market territory served by existing d.c. central stations, thereby depriving commercial a.c.-generation technologies of part of the widening basis for incremental improvements that they came to enjoy. Just as d.c. traction motors became a specialty application of electrical energy through the mediation of rotary converters, so converters could have been employed to transform d.c. into a.c. for specialized application in textile mills, mines and other industrial contexts where ‘sparkless’ a.c. motors were advantageous. The resulting system by 1914 would quite probably not have been more efficient in an engineering sense than the one that was in existence in the United States at that date; it would, perhaps, have resembled the situation that obtained in Britain. But, that is beside the main point—which is that the advent of polyphase a.c. generators and motors brought into existence a multiplicity of feasible equilibria in the design and configuration of electricity supply systems.

Might-have-beens are difficult for the historian to articulate, in that they call for very precise specifications of the contingent unfolding structure of counterfactual worlds. There was nothing foreordained, much less evidently optimal, about the selection that Edison's reactions contributed to making from among these possibilities. Holding more closely to what did happen in (p.353) this particular episode, I should simply say that those who seem to view with approval the outcome of the ‘Battle of the Systems’ in the United States have unjustifiably withheld from Edison due recognition for his inglorious part in the avoidance of what could have been a protracted market competition of uncertain result between the contending currents.

Little is taken from this conclusion, even by granting that a.c. really was the economically and technically superior basis for a universal electricity supply system rather than being simply the more cost effective form of current for long-distance transmission work. ‘Good guys’ are not automatically winners when network technologies compete. One may note that if a new network technology would be economically superior to an incumbent system when everyone had switched to it, complete information in the possession of all agents would be sufficient to induce everyone to decide independently to make the necessary switchover. But, given the state of uncertainty and conflicting opinion among the scientists and engineers of the day, complete information simply was not in the cards.109 One may also note that technological ‘sponsorship’ sometimes will be adequate to prevent the installation or retention of an inferior network technology as the unintended consequence of mere ‘accidents of history’. Where patent-holdings give commercial sponsors property rights in particular technical standards, as in the case at hand, they may be able to capture the benefits that otherwise could accrue to producers and users from subsequent network expansion. A firm convinced that the system whose benefits it can internalize will be superior in the future to the presently incumbent system therefore may find it well worthwhile to subsidize the initial adoption of its technological variant by pricing the equipment or service below costs.110 On the other hand, an incumbent confronted by challengers sponsoring a potentially superior technology (the benefits of which it cannot expect to fully share) may successfully defend its position if it has enough financial resources to engage and outlast those rivals in a war of attrition.

In view of the latter possibility, I have thought it important in the foregoing account to indicate some reasons why Edison apparently turned away from such (p.354) a course of action. Although these have been sufficiently idiosyncratic to underscore my emphasis upon the working of chance, one should observe that the strategy of counter-attack in such circumstances generally would require access to financial backers with ‘deep pockets’ and widely diversified portfolios.111 Lacking that, short-run asset constraints may prevent superior technologies from acquiring sponsors with sufficient capital to unseat inefficient incumbents, just as they may precipitate the premature capitulation of an established technology-sponsor faced with the entry of an alternative technology.

Thus, while exercises in applied microeconomics can tell us that rather special circumstances are required for an inefficient formulation of a technological system to become accidently ‘locked in’, we may also see that these were the very conditions which historically obtained when the electricity supply business—and possibly some other paradigmatic network industries in the fields of transportation and communications—were beginning to take shape. Most of the complex, multi-component and multi-agent systems of production with which we are familiar did not emerge full-blown in the forms that they have, respectively, come to assume. Large scale technological systems such as railroads, electrical utilities, and telephone networks, have been built up sequentially, through an evolutionary process in which the design and operation of constituent components undergoes both continuous and discrete adaptations to the specific technical, economic and politico-legal circumstances in which new opportunities and problems are perceived. And those perceptions, in turn, are often formed on the basis of experience acquired through the operation of pre-existing systems having some of the same functions, or ones directly analogous to the technology in question. So it was that Edison in the 1870s had before him the model of then-existing illuminating gas supply systems—with their generators, distribution mains, meters, and lamps—when he conceived of an integrated lighting and power system based upon electricity. To recognize this calls for acknowledgement of the importance of chance factors in the precise timing of events, including, naturally enough, the sequential development of technical and organizational innovations that shape the competitive strategies followed by commercial sponsors of different network formulations. But, as traditional historians intuitively have understood, the timing and character of ‘small events’ is more likely to be capable of exerting real leverage when these occur close to the beginnings of a sequential development process.

Of course, the temporal location and brevity of those critical phases, in which there is greatest scope for individual decision-makers to alter the (p.355) direction of a decentralized process of technology diffusion and development, must be a relative matter. The comparison indicated here is with the full course of the market competition which may ensue as one system or another progresses towards establishment as the de facto universal standard for the industry. Actual temporal durations depend upon the rate at which system suppliers and users become sequentially committed to one technical formulation, with its attendant compatibility requirements, or another; cyclical booms, during which high rates of investment are undertaken to embody specific technologies in long-lived physical facilities, thus can contribute to narrowing the time-window within which truly formative decisions can occur. Viewed from this angle, it is the fleeting context created by an emergent network technology that creates the opportunity for one or another innovating agents to take specific initiatives which can be held to have directed the subsequent course of events.

Although I began with the question of what role individual entrepreneurial actions could have on the social construction of technological progress, I should therefore close with the observation that the relationship of interest now appears to be a reflexive one. A special kind of competitive struggle, involving the formation of technological or organizational systems characterized by localized positive feedback, holds a special role in the creation of Schumpeterian innovators. This is so if only because these ‘battles’ evolve as dynamic processes in which chance actions on a human scale can prove determinative. Moreover, as such actions are more likely to be those which have occurred before battle lines became clearly drawn and large forces were engaged, if archetypal entrepreneurs are to be found anywhere by retrospective observers, surely they can be singled out more readily from among the ranks of the early participants in emergent network industries. ‘Innovation,’ then, in the sense of an unanticipated impulse imparting a cumulative motion to the economic system, is perhaps less a product of uniquely creative individual attitudes and special social incentives, and more a matter of being pivotally situated during those comparatively brief passages of industrial history when the balance of collective choice can be tipped one way or another. Thomas Edison demonstrated this, as much by abandoning the electric manufacture and supply business as by his launching of it.

Acknowledgements

The research assistance of Julie Bunn was indispensible to me in preparing this essay. Much of the historical material presented here has been drawn (without further attribution) from sections of our collaborative paper, David (p.356) with Bunn (1987). Numerous debts incurred in that connection have been acknowledged in the proper place. An earlier version of the present paper was presented at the Conference on Economic Growth, Past and Present, held in honor of David S. Landes, at the Villa Serbelloni, Bellagio, Italy, 30 August–4 September, 1987. It remains a pleasant necessity to record here my thanks to Patrice Higgonet and Henry Rosovsky, the organizers of that conference, and to Peter Temin, among other conference participants, for their excellent advice about the structure of my exposition; to Jonathan Hughes for sharing insights into Edison's nature, and generously welcoming disagreement over points of emphasis; to Thomas Parke Hughes for a fine mixture of encouraging comments, technical corrections, and references to sources that I otherwise would have missed; to W. Edward Steinmueller and Gavin Wright for characteristically perceptive and helpful suggestions of ways to sharpen the formulation of my arguments. Financial support for this research was provided under the High Technology Impact Program of the Center for Economic Policy Research at Stanford University. I wish also to thank Kenneth S. Ryan for a personal gift of pertinent reference materials, which proved to be a considerable convenience to me in this research.

References

Bibliography references:

Akerloff, G. A. and J. L. Yellen (1985), ‘Can Small Deviations from Rationality Make Significant Differences to Economic Equilibria?’, American Economic Review, 75, 708–720.

Alvarado, F. L. and F. H. Lansetter (1984), ‘Methodology for Analysis of Multi-Terminal HVDC Systems’, in C. W. Buyllard and P. J. Womelodoff, eds, Trends in Electric Utility Research, Pergamon Press, New York pp. 210–30.

Arthur, W. B. (1984), ‘Why a Silicon Valley?: The Pure Theory of Locational Choice,’ Technological Innovation Project Workshop paper, Department of Economics, Stanford University, November.

Arthur, W. B. (1985), ‘Competing Technologies and Lock-in by Historical Small Events: The Dynamics of Allocation Under Increasing Returns,’ Technological Innovation Project Working Paper No. 4, Center for Economic Policy Research (Publication No. 43), Stanford University.

Arthur, W. B., Y. M. Ermoliev and Y. M. Kaniovski (1985), ‘Strong Laws for a class of Path-dependent Urn Processes,’ in Proceedings of the International Conference on Stochastic Optimization, Kiev, 1984, Springer-Verlag, Munich.

Arthur, W. B., Y. M. Ermoliev and Y. M. Kaniovski (1986), ‘Path-dependent Processes and the Emergence of Marco-structure,’ European Journal of Operations Research, 30, 294–303.

Barstow, W. S. (1901), ‘Notes on the Alternating Current System of Distribution,’ AIEE Transactions, Volume XVIII, pp. 849–53.

Bowers, B. (1982), A History of Electric Light and Power, Peter Peregrinus Ltd, New York.

Bresnahan, T. R. and P. A. David (1986), ‘Strategic Effects in the Diffusion of Innovations: Testable Implication and Preliminary Findings for the Case of Automatic Teller Machines in U.S. Banking Markets,’ a paper presented to the Conference on Innovation Diffusion, held at the University of Venice, Italy, March 17–22, 1986. (Technical Papers Series of the High Technology Impact Program, Center for Economic Policy Research at Stanford University, Stanford, CA.)

Brittain, J. E. (1977), Turing Points in American Electrical History, IEEE Press, New York.

(p.357) Brock, G. (1975), ‘Competition, Standards and Self-Regulation in the Computer Industry,’ in R. Caves and M. Roberts (eds), Regulating the Product: Quality and Variety, Ballinger Publishing Company, Cambridge, MA.

Bureau of the Census, US Department of Commerce and Labor (1906), Special Reports of the Bureau of Census: Electrical Industries 1902, Government Printing Office, Washington, DC.

Bureau of the Census, US Department of Commerce and Labor (1910), Special Reports of the Bureau of Census: Electric Light and Power Station 1907, Government Printing Office, Washington, DC.

Bureau of the Census, US Department of Commerce (1915), Census of Electrical Industries 1912, Government Printing Office, Washington, DC.

Bureau of the Census, US Department of Commerce (1920), Census of Electrical Industries 1917, Government Printing Office, Washington, DC.

Byatt, I. C. R. (1979), The British Electrical Industry, 1875–1914: The Economic Returns in a New Technology, Clarendon Press, Oxford.

Campbell, S. G. (1938), The Reorganization of the American Railroad System, 1893–1900, New York.

Carlson, W. B. and A. J. Millard (1987), ‘Defining Risk in a Business Context: Thomas A. Edison and Elihu Thomson, and the A.C.-D.C. Controversy, 1885–1900’, in B. B. Johnson and V. T. Covello, (eds), This Social Cultural Construction of Risk, D. Reidel, Dordrecht.

Carlton, D. W. and J. M. Klamer (1983), ‘The Need for Coordination Among Firms, with Special reference to Network Industries,’ The University of Chicago Law Review, 50, 446–465.

Chandler, A. E., Jr (1977), The Visible Hand: The Managerial Revolution in American Business, Harvard University-Belknap, Cambridge, MA.

Cheney, M. (1981), Tesla: Man Out of Time, Prentice-Hall, Inc., Englewood Cliffs, NJ.

Clark, G. (1984), ‘Authority and Efficiency: The Labor Market and the Managerial Revolution of the Late Nineteenth Century’, Journal of Economic History, 44, 1069–83.

Clemence, R. V. and F. S. Doody (1950), The Schumpeterian System, Addison-Wesley Press, Cambridge, MA.

Cusumono, M. (1985), ‘Note on VTR Industry and Market Development: Japan, the U.S., and Europe, ca. 1975–1985,’ Harvard Business School.

Daggett, S. (1908), Railroad Reorganization, Boston.

David, P. A. (1975), Technical Choice, Innovation and Economic Growth: Essays on American and British Experience in the Nineteenth Century. Cambridge University Press, New York.

David, P. A. (1985), ‘Clio and the Economics of QWERTY:’ American Economic Review, 75, 332–337.

David, P. A. (1986a), ‘Understanding the Economics of QWERTY: The Necessity of History,’ in W. N. Parker (ed.), Economics History and the Modern Economist, Basil Blackwell, Oxford.

David. P. A. (1986b), ‘Some New Standards for the Economics of Standardization in the Information Age’, in P. Dasgupta with P. L. Stoneman (eds), Economic Policy and Technological Performance, Cambridge University Press, Cambridge.

David, P. A. Bunn (1988), ‘Path-Dependence: Putting the Past into the Future of Economics’, Institute for Mathematical Studies in the Social Sciences Technical Report 533, Stanford University.

David, P. A. with J. A. Bunn (1987), ‘“The Battle of the systems” and the Evolutionary Dynamics of Network Technology Rivalries,’ High Technology Impact Program Working Paper No. 15, Center for Economic Policy Research, Stanford University.

David, P. A. and J. A. Bunn (1988), ‘The Economics of Gateway Technologies and Network Evolution: Lessons from Electricity Supply History,’ Information Economics and Policy, 3, 165–202.

David, P. A. and T. E. Olsen (1986), ‘The Equilibrium Dynamics of Diffusion when Incremental Technological Innovations are Foreseen,’ Ricerche Economiche (Special Supplement on Innovation and Diffusion), n. 4, October–December.

Dosi, G. (1982), ‘Technological Paradigms and Technological Trajectories. A Suggested Interpretation of the Determinants and Directions of Technical Change,’ Research Policy, 13, 3–20.

Dosi, G. (1984), Technical Change and Industrial Transformation, Macmillan, London.

Eldredge, N. (1985), Times Frames: The Rethinking of Darwinian Evolution and the Theory of Punctuated Equilibria, Simon and Schuster, New York.

Bibliography references:

(p.358) Electrical World, (articles without identified authors):

(1887), ‘Edison vs. Westinghouse,’ January 1, pp. 7–8.

(1887), ‘Incandescent Lights on High Tension Circuits,’ February 26, p. 110.

(1887), ‘The Distribution of Electricity by Secondary Generators,’ March 26, pp. 156–8.

(1887), ‘Report on the Committee on Electrical Distribution by Alternating Currents,’ August 20, pp. 92–3.

(1888), ‘Alternating Current or Storage Battery,’ January 21, p. 25.

(1888), ‘Alternating vs Continuous Current,’ March 31, p. 159.

(1888), ‘The Comparative Danger of Alternating vs. Direct Current,’ September 8, p. 126.

(1889), ‘Mr. Edison on the Dangers of Electric Lighting,’ November 2, pp. 292–3.

The Encyclopedia Britannica. A Dictionary of Arts. Literature and General Information (1910–1911), Eleventh Edition. 29 vols. The Encyclopedia Britannica Co., New York.

Evans, G. T. (1982), ‘Some Phases of the Alternating Current,’ Electrical World, January.

Farrell, J. and G. Saloner (1985a), ‘Standardization, Compatibility and Innovation,’ Rand Journal, 16, 70–83.

Farrell, J. and G. Saloner (1985b), ‘Economic Issues in Standardization,’ Working Paper No. 393, Department of Economics, Massachusetts Institute of Technology, October.

Farrell, J. and G. Saloner (1985c),‘Installed Base and Compatibility, with Implications for Product Preannouncements,’ Working Paper No. 385, Department of Economics, MIT, August.

Fleming, J. A. (1921), Fifty Years of Electricity: The Memories of an Electrical Engineer. Published from the offices of Wireless World, London.

Fudenberg, T. and J. Tirole (1985), Dynamic Models of Oligopoly (Fundamentals of Pure and Applied Economics, Volume 3: Theory of the Firm and Industrial Organization Section), Harwood Academic Publishers, London.

Gilfillan, S. C. (1935a), Inventing the Ship, Foller, Chicago.

Gilfillan, S. C. (1935b), The Sociology of Invention, Foller, Chicago.

Gould, S. J. (1987), ‘The Panda's Thumb of Technology,’ Natural History, 1987, no. 1, 14–23.

Haken, H. (1978), Synergetics, 2nd Ed., Springer-Verlag, Berlin.

Haltiwanger, J. and M. Waldman (1985), ‘Rational Expectations and the Limits of Rationality: An Analysis of Heterogeneity’, American Economics Review, 75, 326–340.

Hannah, L. (1979), Electricity Before Nationalism: A Study of the Development of the Electricity Supply Industry in Britain to 1948. Johns Hopkins University Press, Baltimore.

Hennessey, R. A. S. (1971), The Electricity Revolution, Oriel Press Ltd, Newcastle Upon Tyne, UK.

Hughes, J. (1986), The Vital Few, Oxford University Press, New York.

Hughes, T. P. (1983), Networks of Power: Electrification in Western Society. 1880–1930, The Johns Hopkins University Press, Baltimore.

Jarvis, C. M. (1967a), ‘The Generation of Electricity,’ in C. Singer, E. J. Holmyard, A. R. Hall and T. I. Williams, (eds), A History of Technology, Volume V: Late 19th Century. 1850–1900. The Clarendon Press, Oxford, p. 177–207.

Jarvis, C. M. (1967b), ‘The Distribution and Utilization of Electricity,’ in C. Singer, E. J. Holmyard, A. R. Hall and T. I. Williams, (eds), A History of Technology, Volume V: Late 19th Century. 1850–1900. The Clarendon Press, Oxford, pp. 208–234.

Josephson, M. (1959), Edison: A Biography, McGraw-Hill, New York.

Jones, P. (1940), A Power History of the Consolidated Edison System, New York Consolidated Edison, New York.

Katz, M. L. and C. Shapiro (1985a), ‘Network Externalities, Competition and Compatibility,’ American Economic Review, 75, 424–440.

Katz, M. L. and C. Shapiro (1985b), ‘Technology Adoption in the Presence of Network Externalities,’ Discussion Paper in Economics No. 96, Woodrow Wilson School, Princeton University.

Kindleberger, C. P. (1983), ‘Standards as Public, Collective and Private Goods,’ Kyklos, 36, 377–396.

Kolko, G. (1965), Railroad and Regulation 1877–1916, W. W. Norton and Company, New York.

Lamme, B. G. (1929), Benjamin Garver Lamme, Electrical Engineer: An Autobiography, Putnam's Sons, New York.

(p.359) Landes, D. S. (1986), ‘What Do Bosses Really Do?’ Journal of Economic History, 46, 585–624.

Landes, D. S. (1969), The Unbound Prometheus: Technological Change and Industrial Development in Western Europe from 1750 to the Present, Cambridge University Press, New York.

Lardner, H. A. (1903), ‘Economical and Safety Limits in the Size of Capital Stations,’ A.I.E.E. Transactions, 21, pp. 407–416.

Lineff. W. (1888), ‘Accumulators versus Direct Currents for Electric Traction,’ Electrical World, April 14.

Leupp, F. E. (1919), George Westinghouse: His Life and Achievements, John Murray, London.

MacDonald, E. (1962), Insull, University of Chicago Press, Chicago.

Marglin, S. (1974), ‘What Do Bosses Do? Part I,’ The Review of Radical Political Economy, 6, 60–112.

National Electrical Manufacturers Association (1946), A Chronological History of Electrical Development, NEMA, New York.

Nelson, R. R. and S. G. Winter (1982), An Evolutionary Theory of Economic Change. The Belknap Press of Harvard University, Cambridge, MA.

Parsons, R. H. (1940), The Early Days of the Power Station Industry, Cambridge University Press, Cambridge.

Passer, H. C. (1972), The Electrical Manufacturers: 1875–1900: A Study in Competition, Entrepreneurship, Technical Change, and Economic Growth, Arno Press, New York.

Pfrannkuche, A. (1887), ‘Long Distance Distribution of Electric Energy,’ Electrical World, April 2.

Prigogine, I. (1980), From Being to Becoming: Time and Complexity in the Physical Sciences, W. H. Freeman, New York.

Prigogine, I. and I. Stengers (1984), Order Out of Chaos, New Science Library, Boulder.

Prout, H. G. (1921), A Life of George Westinghouse, The American Society of Mechanical Engineers, New York.

Rosenberg, N. (1969), ‘The Direction of Technological Change: Inducement Mechanisms and Focusing Devices,’ Economic Development and Cultural Change, 18, No. 1 (October).

Rosenberg, N. (1976), ‘Marx as a Student of Technology,’ Monthly Review, 28, 56–77.

Rosenberg, N. (1983), Inside the Black Box: Technology and Economics, Cambridge University Press, New York.

Rosenbloom, R. (1985), ‘Managing Technology for the Longer Term,’ in K. Clark et al. (eds), The Uneasy Alliance, Praeger, New York.

Scott, C. F. (1901), ‘Alternating Current as A Factor in General Distribution for Light and Power,’ AIEE Transaction, XVIII, 843–48.

Sharlin, H. I. (1963), The Making of the Electrical Age, Abelard-Schuman, New York.

Stanley, W. (1912), ‘Alternating Current Development in America,’ Franklin Institute Journal, 68, 568–73.

Stillwell, L. B. (1934), ‘Alternating Current Versus Direct Current,’ Electrical Engineering, 53, 708–10.

Stolper, W. F. (1981), ‘Aspects of Schumpeter's Theory of Evolution,’ in H. Frisch, (ed.), Schumpeterian Economics, Praeger Publishers, New York.

Streissler, E. (1981), ‘Schumpeter's Vienna and the Role of Credit in Innovation,’ in H. Frisch, (ed.), Schumpeterian Economics, Praeger Publishers, New York.

Sumpner, W. E. (1890), ‘Some Peculiarities of Alternate Currents,’ Electrical World, January.

Teggart, F. J. (1925), Theory of History, Yale University Press, New Haven.

Temin, P. (1987), ‘Bosses and Technical Change.’ Paper presented to the Conference in Honor of David S. Landes, In Bellagio, Italy, August 30–September 4, 1987.

Usher, A. P. (1929. 1954), A History of Mechanical Inventions, 2nd edition. Harvard University Press: Cambridge, MA.

Weeks, W. (1981), Transmission and Distribution of Electrical Energy, Harper and Row, Publishers, New York.

Williamson, O. E. (1975), Markets and Hierarchies: Analysis and Anti-Trust Implications, Free Press, New York.

Williamson, O. E. (1985), The Economic Institutions of Capitalism, Free Press, New York.

Woodbury, D. O. (1960), Elihu Thomson: Beloved Scientist, 1853–1937. The Museum of Science (1944, 1960), Boston.

Zorpette, G. (1985), ‘HVDC: Wheeling Lots of Power,’ IEEE Spectrum, June, 30–36. (p.360)

Notes:

(1) Landes (1969), pp. 281–90, gives a concise and pithy account of the main electrical developments in Britain, France and Germany up to 1913, drawing largely on Jarvis (1967a, 1967b) for technical details. Among works appearing subsequently that deal primarily with the British and American parts of the story, in chronological order, see Hennessey (1971), Passer (1972), Byatt (1979), Hannah (1979), Bowers (1982), and Hughes (1983). The latter work by Thomas Hughes, also treats the experience of Germany.

(2) On the argument's other side, Jonathan Hughes (1986, p. 3) writes: ‘… in too many accounts the economy lies silently in the background growing miraculously to support and nourish the actions of the gods and heroes the professional historians so love to study in war and politics. But just as men must be mobilized and led in war, voters organized and persuaded in politics, so economic resources must be mobilized and directed intelligently for economic growth to occur.’

(3) See Clemence and Doody (1950), for a useful account of the Schumpeterian system, and the recent, sympathetic overview by Stolper (1981). Streissler (1981), pp. 65–7, makes the interesting observation that to have emphasized the creative role of the entrepreneur was no revolutionary intellectual departure on Schumpeter's part, nor even on the part of his main teacher, Friedrich von Wieser. Rather, what was novel and somewhat shocking to Schumpeter's contemporaries was the stress he placed on the destruction and disorder that entrepreneurs caused by their innovations. The potential for competing network technologies completely to displace one another in actual use, and the associated properties of randomness and unpredictability that adhere to such competitions, make the novel aspect of Schumpeter's untidy vision of economic development particularly congenial to the major themes of this essay.

(4) The long-standing controversy among economic historians over the nature and essentiality of the entrepreneur's role has most recently been renewed and refreshed by Landes's (1986) critique of Marglin (1974), and Temin's (1987) effort to distinguish questions about the genesis and advantages of large-scale, hierarchically structured production activities, from other historical questions concerning the control and management of hierarchical organizations, including those addressed by Clark (1984).

(5) David and Olsen (1986) provide a formal model of the generation of a ‘diffusion-cum-development trajectory’, along which the adoption and incremental improvement of a capital-embodied technology are intertwined. That analysis, however, leaves aside the mathematically more difficult stochastic aspects of the actual matter, which are the focus of interest here.

(6) I concede that I would have numbered myself among them, at least until 1975. Jonathan Hughes complained of this state of affairs among professional economists in his introduction to the 1965 edition of The Vital Few, and has reiterated it recently, while noticing the recent rejuvenation of interest in the subject of entrepreneurship among business and political circles. See Hughes (1986), pp. ix, 2.

(7) This concept of change, as timeless and constant, of course, holds a strong grip on other areas of the natural sciences, and the social sciences that have made them their model—from whence, ironically, it has come to conflict with the narrative tradition in the writing of history. See the perceptive and too little noticed work of Teggart (1925), especially chapters 7–12, recent use of which has been made by Eldredge (1985), chapter 1 and pp. 141–46. In this respect it may be seen that the pioneering work of Nelson and Winter (1982) on ‘evolutionary models’ of economic change has been well-designed to influence economists at ease with the eventless Darwinian conception of continuous modification through (competitive) natural selection. The incorporation of the Markov property in Nelson and Winter's models has served the same purpose, as has been noted elsewhere, e.g. David (1975), p. 76.

(8) Gilfillan (1935b) explicitly stressed the Darwinian evolutionary analogy in placing ‘the continuity of development’ first among his 38 ‘social principles of invention,’ and, in (1935a), documented it in extenso by references to the merchant ship. For Marx's views of technology, see Rosenberg (1976), who calls attention to their fruition in the emphasis Usher (1929) accorded to social processes in ‘setting the stage’ for invention, as well as in the cumulation of small improvements. This particular chain of influence extends at least one important link further: Landes (1986, p. 602), in a related context, declares himself to be among the descendents of the (intellectual) ‘House of Usher’.

(9) See for example Rosenberg (1983), chapter 1, for a review.

(10) Williamson (1975, 1985), and related work on organizational structure, contracting, and hierarchical control, has, of course, been reactive against the more extreme formulations of this position—on the grounds that they left unexplained the boundaries between firms and the market, and the internal institutional features of modern business enterprises. Chandler (1977), while much occupied with the rise of large and complex business organizations within which were developed modes of resource allocation alternative to the market, casts the emerging technology of large-scale production in the role of the principal exogenous force inducing these organizational adaptations. Where the new technology was coming from, and whether its development was occurring primarily under the guidance of a visible hand (rather than its invisible, market counterpart), remains inexplicit in Chandler's scheme of things.

(11) See for example Haken (1978), Prigogine (1980), Arthur et al. (1985, 1986), for formal mathematical analysis of dynamical systems with these properties; Prigogine and Stengers (1984) offer a less technical and more philosophical interpretation.

(12) This rather general proposition has been arrived at by analysts of quite a few, different economic contexts. For example, it has been demonstrated recently, using different arguments, by Akerloff and Yellen (1985), and by Haltiwanger and Waldman (1985).

(13) See for example David (1975, pp. 6–16, 50–91), and, more recently, (1987) and (1988), for formalizations of the underlying view of the nature of the stochastic process that generates path-dependent technological progress. The heuristic value of a vivid illustration, such as the tale of the QWERTY typewriter keyboard layout recounted in David (1986a), has been masterfully demonstrated by Gould (1987).

(14) This is a highly compressed and necessarily abstract statement of complicated matters that have been set out more fully elsewhere. See David (1975), chapter 1, on localized stochastic learning, the channelling of technical innovations along particular ‘cones’ in factor input space defining available production processes (techniques), and the influence of historical events in causing industries or economies to become committed to a particular path, or trajectory of development associated with such ‘techniques.’ The notion of a technological paradigm, which Dosi (1982, 1984) has introduced and elaborated upon, is quite usefully applied to the concept of a ‘universal’ electrical system, i.e., an integrated network for electricity supply and use. The latter concept, the paradigm, embraces variant technical formulations of such a system, notably, the alternative designs based upon d.c., single phase a.c., and polyphase a.c. Each of these, as will be seen, would carry some peculiar implications for the subsequent experience of the regions and economies that became committed to one or another specific ‘trajectory’ of technological and industrial development.

(15) The following passage from Prout (1921), the American Society of Mechanical Engineers' official biographer of Westinghouse, exemplifies this tradition of interpretation: ‘As soon as it became evident that Westinghouse proposed to exploit extensively the alternating-current system great opposition was developed. Looking back at history, one is surprised at the stupidity and puerility of some of this opposition. Men of great repute gave their names and their help to methods of which they must now be thoroughly ashamed. They know now that if they had succeeded, the progress of civilization would have been delayed—how much and how long we cannot even guess. … It is needless to enlarge upon this [battle of the systems] aspect of the development. Every well-informed human being knows Westinghouse was right, the alternating electric current being now used to generate and convey 95 percent of the electric use in power and lighting in the United States.’ (p. 115). More recent versions, to be examined below, do not even contemplate the possibility that Edison could have succeeded—had it been his intention to block these developments.

(16) Landes (1986), p. 622, speaking of the historical evolution of the factory system.

(17) A sequel now confronts the home video or ‘cam-corder’ enthusiast, and the manufacturers of video-camera and VCR equipment: is the current half-inch (12.7 mm) tape width for VCRs to be preserved as a standard, or is it better to switch to systems based on the 8 mm format pioneered by Sony? See Rosenbloom (1985) on the background: Cusumano (1985) on details of the VCR rivalry, from the producer's side.

(18) See David (1985, 1986a) for the story of QWERTY.

(19) The normative literature on the economics of standardization recognizes that while the evident failure of markets would call out for remedies in the form of governmental efforts to promote the exploitation of economies of scale obtainable via network coordination and system integration, such interventions may result all too easily in choices of technology that prove to be ‘mistakes.’ See Brock (1975), Kindleberger (1983), Carlton and Klamer (1983), Katz and Shapiro (1985a, b), Farrell and Saloner (1985a, b, c). David (1986b) reviews this literature in fuller detail.

(20) Under such conditions—where, as I observed more than a decade ago, ‘marked divergences between ultimate outcomes may flow from seemingly negligible differences in remote beginnings’ (David, 1975, p. 16)—market rivalries among variant technological systems do exhibit a strictly historical character. In the sense that they are non-ergodic, the dynamic process cannot shake itself loose from the grip of past events and its outcome therefore can properly be described as path-dependent.

(21) It should be emphasized that this form of (hindsight) recognition that rational decision-making had nonetheless resulted in an inferior path being mistaken for the ‘right’ one could not arise under conditions of constant or decreasing returns; in the latter cases the eventual outcomes would be path independent. See Arthur (1984, 1985), Arthur et al. (1985) for more formal (mathematical) treatment, and David (1986b) for further exposition.

(22) In 1882, according to Jones (1940), p. 41, this complex of enterprises included: (1) the Edison Electric Light Company, which had been formed to finance the invention, patenting, and development of Edison's electric-lighting system, and which licensed its use; (2) the Edison Electric Illuminating Company of New York, that being the first of the Edison municipal lighting utilities; (3) the Edison Machine works, organized to manufacture the dynamos covered by Edison's patents; (4) the (Edison) Electric tube Company, set up to manufacture the underground conductors for electric power distribution in the lighting system; and (5) the Edison Lamp Works.

(23) See Prout (1921), pp. 94–5, 113; Sharlin (1963), pp. 195–6.

(24) On the pre-electrical career of Westinghouse, see Leupp (1919), Product (1921).

(25) Fleming (1921), p. 225.

(26) See for example Sharlin (1963), pp. 136–47; Jarvis (1967a, b), for convenient reviews. More details of the history of alternating current technology prior to 1880 can be found in Electrical World (March, 1887); Sumpner (1890); Encyclopaedia Britannica (1910–11), vol. 9, pp. 179–203; Fleming (1921), p. 44.

(27) Sharlin (1963), pp. 192–3; T. Hughes (1983), pp. 86–91.

(28) A ‘transformer,’ in modern electrical parlance, refers to a device by which the electromotive force (e.m.f.) of a source of alternating current may be increased or decreased. [For the reader whose recollection of the introductory physics of electricity has grown as hazy as was mine: the difference in potential, which may be thought of as a kind of driving force behind the electrons forming the current, is sometimes called the e.m.f. and is measured by V, the voltage. Electric current is analogous to the flow of water through a pipe, the rate of flow being measured by I, the amperage. ‘Power’, P measured in watts in W/t, the time rate at which energy (W) is developed or expended.] Unlike a dynamo used in generation, a transformer needs no moving parts. It has a primary winding (whose terminals are attached to an a.c. source), and a secondary winding, both of which pass in coils around an iron ‘core.’ The first such device was the ‘closed-core’ transformer made by Faraday, the core of which had the form of an iron ring. ‘Step-down’ transformers have a larger number of turns of wire in the primary winding than in the secondary winding, since the ratio of primary (or ‘impressed’) voltage to secondary voltage is equal to the ratio of the number of turns in the respective windings. Although it may seem, on first consideration, that the boost in voltage achieved by means of a step-up transformer somehow violates the conservation of energy law, such obviously cannot be the case; the power supplied at the primary is just equal to that delivered at the secondary. In general, when the voltage is stepped up (or, as English electrical engineers say, ‘bumped up’), the current is ‘bumped down’ by the same proportion.

(29) See Stanley (1912), pp. 564–5; Passer (1972), pp. 136–8, on Gaulard and Gibbs, Stanley and Westinghouse, and the still more comprehensive account in T. Hughes (1983). pp. 95–105. The challenge of solving the practical problems arising from the connecting of the primary coils of the transformers in series by Gaulard and Gibbs drew an even earlier response from the Hungarian team of Zipernowsky, Deri, and Blathy of Budapest. Like Stanley, they found that transformers can be worked independently if the different primary circuits are arranged in parallel between two high voltage mains (1000 V in their patent), like the rungs on a ladder; the secondary circuits of the transformers were kept isolated, with incandescent lamps placed on them in parallel.

(30) T. Hughes (1983), pp. 101–5; Passer (1972), pp. 131–8.

(31) Josephson (1959), pp. 231–2; Jarvis (1967b), Ch. 10, p. 229; T. Hughes (1983), p. 83–4; Byatt (1979), p. 99.

(32) Jarvis (1976b), p. 229. The 110 volt standard for electricity consumption which came to be established in the United States was selected by Edison on the basis of electrical circuit theory and the maximum resistance lamp filament that he was able to obtain circa 1879. From 1883 onwards, however, the Edison ‘three-wire’ system was employing 220 volts for transmission of direct current. How the European following the Berliners' example, came to establish a 220 volt standard is another story. See Hughes (1983), p. 193.

(33) Using the standard notation introduced in note 26, above, the power equation for direct current is p =VI; but, for the case of alternating current, power has to be thought of in an ‘average’ sense and the equation takes the root mean square form: P = Vrms I rms cos φ. Here Vrms = Vmax/(2)1/2 and Irms = Imax/(2)1/2. The power factor, cos φ, depends on the phase angle f, which measures the amount by which the electron current in the line leads or lags the impressed voltage. For current of any kind, however, Ohm's Law gives the resistance, R, measured in ohms, as R = I/V. Note that the resistance of a wire connecting two points is directly proportional to its length, L, and inversely proportional to its cross-sectional area, A, with the factor of proportionality being the ‘resistivity’, β:R = β(L/A). This relation, combined with Ohm's Law, and the a.c. power equation, above, yields the voltage – distance relationship cited in the text: L = (V2/P)B, where the constant is implicitly defined as B = {A(cos φ)}/2β.

(34) Passer (1972), p. 137–44, 165–7; Byatt (1979), pp. 102–7; Evans (1892), p. 52.

(35) See Passer (1972), pp. 237–49 on Sprague and Johnson Particularly; pp. 216–55 on electric street railways: McDonald (1962), p. 36.

(36) And, also, between the main competitive suppliers of a.c. lighting plant—the Thomson-Houston, and Westinghouse Electric companies. On ‘first-mover advantages’, and incentives for pre-emption by suppliers of new technologies and goods and services based upon them, see the theoretical analysis developed by Fudenberg and Tirole (1985); and Bresnahan and David (1986), for an effort at empirical application.

(37) See, for example T. Hughes (1983), pp 217–19. A 10–12 per cent load factor is given as the average for a simple lighting load in the article contributed by J. A. Fleming on ‘Electricity Supply’ in Encyclopedia Britannica (1910–1911), vol. 9, p. 194.

(38) Quoted in Josephson (1959), p. 346, italics added. Josephson's source was an Edison memorandum to E. H. Johnson, commenting on a report that had been obtained from the engineers of Siemens and Halske (Berlin), evaluating the version of the a.c. lighting system that had been developed by Zipernowski, Deri and Blathy (see note 27, above).

(39) According to Prout (1921), p. 95, Westinghouse had begun installing isolated d.c. lighting plants earlier in 1886, and had practically completed his company's first d.c. central station in Trenton, NJ in August. Direct current installations in other towns were underway. Quite probably, it was this bustle of activity—an incursion into markets that otherwise might ‘naturally’ fall to the Edison Electric Lighting Co.—which Edison felt was more disturbing than the first a.c. central station completed by Westinghouse Electric Co, in November.

(40) See Passer (1972), Table 19, p. 150 for lighting capacity data for the three leading companies in 1891.

(41) See, for a sample of the best modern accounts in each genre, Stillwell (1934); Passer (1972), Ch. 5; T. Hughes (1983), Ch. 6, esp., pp. 106–11; J. Hughes (1986), pp. 192–98. McDonald (1962), pp. 43–6 recounts the affair from the viewpoint of Samuel Insull.

(42) For example, in addition to the patent suits mentioned below, Westinghouse filed claim for patent infringement against Thomson-Houston when the latter firm began production and sale of an a.c. incandescent lighting system. The two companies worked out an agreement under which control of the Consolidated and Sawyer-Man patents in possession of Thomson-Houston was relinquished to Westinghouse. See Passer (1972), p. 139.

(43) See T. Hughes (1983), pp. 22–3, and references therein, on Edison as a ‘systems-inventor’—a point of some further significance below.

(44) Passer (1972), pp. 149–64.

(45) For example, Woodbury (1960), pp. 184–5, gives an account of the inordinate amount of time and worry that Elihu Thomson had to devote to problems connected with patent litigation. Josephson (1959, pp. 354–8) details Edison's vexations with the protracted ‘war’ with Westinghouse over the carbon filament patent of 1878, in which he was brought to testify during 1890. The suit cost the victorious Edison General Electric Co., holder of the patent, about $2 million by the time it was concluded in 1891—when the lamp patent had less than 3 years of life remaining. Sheer vanity aside, economically motivated reputational considerations can account for the willingness of inventors to spend time defending claims to proprietary rights which they have already sold; the value to purchasers of patent rights, which a self-employed inventor might wish to sell in the future, will be directly and indirectly affected by this form of commitment.

(46) See Josephson (1959), p. 351.

(47) See Josephson (1959), pp. 295–97; T. Hughes (1983), pp. 38–41.

(48) See Josephson (1959), p. 299; also Electrical World (January 1, 1887) for details on the eleven suits the Edison Electric Company began against the Westinghouse Electric Company on December 23, 1886 for injunctions and damages for the infringement of electric lighting patents; the patents in question covered the entire central station system.

(49) See Josephson (1959), p. 347.

(50) See Josephson (1959), pp. 347–8; T. Hughes (1983), pp. 108–9; Woodbury (1960), p. 174.

(51) Woodbury (1960), p. 170–1.

(52) George Westinghouse, reportedly, was outraged. One of these alternators was in place at the Auburn State Prison in August 1890, where it produced the current lethal to the first victim of the ‘electric chair’, William Kemmler, a convicted axe-murderer. See Leupp (1919), pp. 132–55, for details of the controversy over this execution, and the unfounded suspicions of the time that Westinghouse was financing the opponents of capital punishment who mobilized on the occasion.

(53) See Josephson (1959), pp. 348–9; Cheney (1981), p. 43; T. Hughes (1976), p. 108.

(54) See Electrical World, August 1887; September 1888; Woodbury (1960), p. 174. The lesser hazard of accidental electrocution from a.c. derives from the tendency of the current to repel a body from the mains, breaking the contact, whereas d.c. does not have this action. Although the text follows contemporary popular discussions by speaking of the harm done by high voltage, V, strictly, the damage done by a current is dependent upon the volume of the flow, i.e. the amperage, I. By Ohm's Law, however, the latter is proportional to the e.m.f., V—the proportionality constant being the resistance of the conducting body, R. See above, note 31.

(55) Josephson (1959), p. 345–5, 349. In addition to the early concern expressed by Elihu Thomson, Franklin Pope, in the US, and both Dr. Werner Von Siemens and Lord Kelvin—recognized ‘scientific authorities’ on electricity—warned against the dangers of the a.c. system of transmission at higher voltages.

(56) Woodbury (1960), pp. 169–72.

(57) Electrical World, November 2, 1889, pp. 292–3.

(58) See the influential account given by Josephson (1959), pp. 313–38, 349–50, 361–2, emphasizing Edison's resistance to the advice of others to follow Westinghouse and Thomson-Houston into the a.c. electricity supply field.

(59) Passer (1972), p. 74. This summation in Passer's 1954 study is quoted approvingly in the influential biography by Josephson (1959), p. 350. Alterations in the inventor's objective economic and institutional circumstances are cited by Passer in explanation of this transformation, whereas Josephson attributes Edison's unwillingness to pursue developments in a.c. technology to ‘the fear that all the effort, equipment, and capital invested in the old system would quickly be made obsolete by the new’—seemingly without crediting Edison with an awareness that if his efforts could produce that result, it also could happen at the hands of others.

McDonald (1962), pp. 32–3, takes a different explanatory tack in suggesting that a profound psychological change had occurred with the death of Edison's wife, Mary Stilwell Edison, in 1884: ‘Edison's creative period as an electric innovator [sic] ended with his wife's death, and in the future he not only contributed little to the success of his electric companies, but sometimes actually impeded their progress [by opposing alternating current applications]. From the greatest single asset a collection of electrical enterprises could possibly have, he suddenly became a burden.’ An awkward fact, omitted from McDonald's account, is that within a year of Mary's death Edison had courted and successfully proposed marriage to Mina Miller, and appeared to his friends to have drawn a renewed vitality and pleasure in anticipating life with his second wife. (See Josephson 1959, pp. 301–8). As for the suggested waning of Edison's inventive powers, McDonald's assertion (p. 33, n. 16) that ‘[a]fter 1884, virtually all Edison's inventions were relatively trivial, and some of them were almost foolish’ is refuted within 10 pages by his own, more accurate statement (p. 43): ‘To be sure, some of his greatest achievements were yet to come—among them the perfected phonograph and the motion picture—but none of them was in the electric field.’

(60) J. Hughes (1986), pp. 193–4. This account bears indications of the influence of Josephson's (1959, p. 349) view that the ‘whole dreadful controversy can be attributed only to an extreme bitterness of feeling towards his [Edison's] opponents that completely overbalanced his judgement.’ Among the many previous histories of these events, the account provided by Thomas P. Hughes (1983) stands out in refraining from depicting Edison as ignorant and technologically reactionary on the subject of alternating current.

(61) Carlson and Millard (1987, pp. 267–77) call attention to Edison's earlier experience with the use of safety concerns as a competitive strategy—from the days when the illuminating gas companies had deployed such ‘fear propaganda’ against his own incandescent lighting enterprise. These authors offer an interpretation of Edison's motives in raising the ‘safety question’ which is quite different from the one put forward here. But nothing in the documentation they cite contradicts my contention that from 1887 onward Edison's actions were driven primarily by considerations of short-run expediency, rather than a long-run commitment to promoting the universal use of electricity by increasing the technology's safety.

(62) Woodbury (1960), pp. 172–3; Josephson (1959), p. 345.

(63) See, for example McDonald (1962), p. 62; T. Hughes (1983), pp. 107–9.

(64) T. Hughes (1983), p. 108.

(65) See the extract from Josephson (1959), p. 346, quoted in the text above.

(66) See T. Hughes (1983), pp. 106–11 for a discussion suggesting that this differential rate of progress was somehow elicited because electrical engineers, having used a.c. to overcome the transmission cost constraints that encumbered the expansion of d.c. distribution networks, quickly encountered numerous obstacles in getting an a.c. electricity supply system to emulate the functions that existing d.c. systems were capable of performing—particularly, the provision of power to secondary motors. While such constraints may have served as ‘focusing devices’ (see Rosenberg, 1969), directing engineering efforts towards projects which were more likely to have high commercial payoff if the technical problems could be overcome, they could not have signalled the possibility of actually solving those problems.

(67) See Fleming (1921), pp. 146–7 for details of motor development and their construction.

(68) T. Hughes (1983), pp. 115–17 gives full details on the filing, issuance and acquisition dates of the key Tesla patents, and others.

(69) See Prout (1921), pp. 128–9. Development of an a.c. meter was regarded as sufficiently important that Westinghouse himself filed patents on one such device, in June 1887, and on another, in October 1887, developed jointly with one of his engineers, Phillip Lange. Patents were issued for these meters in May 1888, but by then they had been superceded by Shallenberger's design.

(70) T. Hughes (1983), p. 120, notes also that the electrical equipment manufacturing companies also ‘remained partially committed to direct current’ by virtue of the specialized facilities, patent positions, as well as the experience and expertise they had built up in that field.

(71) The role of ‘gateway’ innovations in network technology evolution is explicitly considered with reference to the rotary converter by David and Bunn (1988), and is discussed more generally in regard to standardization policy issues by David (1987).

(72) Passer (1972), pp. 300–1; T. Hughes (1983), p. 118. The courts eventually ruled against the patent application by Bradley for the generator and synchronous motor, on the ground that Tesla's patent was fuller and more complete.

(73) See Pfrannkuche (1887) for discussion of an ‘all purpose’ system involving d.c. dynamos and a.c. transmission.

(74) Lamme (1926), Chapter 6, recounts his early work for Westinghouse on rotary converters.

(75) See McDonald (1962), pp. 69–70, on the work of Chicago Edison's chief engineer, Louis Ferguson, circa 1894, who installed rotary converters at both the generating and local distribution ends—using polyphase a.c. for transmission only. On the technology of rotary (or, in some English usage, ‘rotatory’) converters, as well as rotary transformers (used in changing the voltage of direct current), and the Ferranti rectifier (used in transforming an alternating single-phase current into a direct pulsating current for arc lighting), see the article on ‘Transformers’ contributed by J. A. Fleming in Encyclopedia Britannica (1910–1911), vol. 27, pp. 178–9. More generally, on the significance of converters, see T. Hughes (1983), pp. 121–2.

(76) Edison, in a 1908 encounter with George Stanley, the son of William Stanley, whose improvement on the Gaulard Gibbs patent had been the basis for Westinghouse's a.c. system, is said to have remarked: ‘Oh, by the way, tell your father I was wrong.’ Josephson (1959, p. 349) in reporting this interprets Edison to have thereby acknowledged that he had made ‘his greatest blunder’ by not following Westinghouse into the a.c. technology. Putting aside the issue of whether Edison had made a blunder or a justifiable decision, it remains quite unclear exactly what Edison felt he had been ‘wrong’ about. Possibly it was his mistaken expectation, in 1886, that a great deal of experimenting and a long period of further practical development would be needed before a system based on a.c. could be brought to the point of challenging his own.

(77) See T. Hughes (1983), pp. 22–3.

(78) Thus, it is not surprising that after the spring of 1887 Edison should have allowed himself to become increasingly pre-occupied with renewed experimental work on his phonograph—entering an inventive race against workers in the laboratory of Alexander Bell who had undertaken to improve upon the original Edison device of 1877. Josephson (1959), pp. 317–31 relates the story of this project, which was brought to a successful culmination by Edison's famous 72-hour frenzy of non-stop work in June, 1888; and of the disappointing sequel, in which J. Lippincott and Edison's unscrupulous associates swindled the inventor out of a major part of the value of the rights to the new phonograph patent and an exclusive manufacturing license.

(79) In response to this interpretation, Jonathan Hughes in private correspondence with the author (March 5, 1987) writes: ‘I hope you're right about how clever Edison was in getting out of the way of the ac “juggernaut”. So far as actual verbal evidence goes, Edison could have done what he did for no more reasons than ignorance and cunning. … I am willing to grant that he had a certain peasant cunning (Bauernschlauheit) that would make him see that a campaign of dust-in-your eyes about ac would allow him to sell out and get out. It is reasonable. The evidence that he was running from what he did not and could not understand is pretty strong too.’ Bauernschlauheit, no less, and from a fellow of mixed Scottish and Dutch descent!

(80) See Josephson (1959), pp. 350–66, and McDonald (1962), pp. 30–39 for a fuller account of the formation of the Edison General Electric Co., and then, the General Electric Co.

(81) See Josephson (1959), p. 340; McDonald (1962), p. 38.

(82) Josephson (1959), pp. 351–3. On Villard's role, see also, McDonald (1962), pp. 39–40, and T. Hughes (1983), pp. 76–7.

(83) See McDonald (1962), pp. 40–1.

(84) Josephson (1959), p. 361, quoting a letter dated February 8, 1890.

(85) See for example Daggett (1908), Campbell (1938), Kolko (1965), Chapter IV, esp. pp. 64–7.

(86) See McDonald (1962, pp. 48–51. Villard, at the head of E.G.E. previously had sought a consolidation with the Westinghouse Electric Company, but when priority was awarded to Edison's carbon filament patent in 1891, he felt his hand sufficiently strengthened to seek to acquire a supposedly weakened competitor, and opened negotiations with Thomson-Houston. In the end, Morgan agreed with Coffin that Thomson-Houston was in a stronger financial position, and so should purchase E.G.E. There is a nice, but unresolved, and probably unresolvable question: whether E.G.E.'s comparatively weaker financial condition owned something to the effect of Edison's propaganda campaign against a.c., which supported an inflation of the price E.G.E. paid for its constituent d.c.-system companies.

(87) See Leupp (1919), pp. 157–61. Under Belmont's direction, two electric lighting companies that had been controlled by the Westinghouse interests (the United States and the Consolidated) were absorbed into the reorganized firm, and their stockholders were given the new preferred and common stock issued by the Westinghouse Electric and Manufacturing Co. The stockholders of the main company surrendered 40 per cent of their old stock, and were asked to take second preference shares in the reorganized firm in lieu of the rest. By these measures the original outstanding liability of more than $10 million (on which the annual interest charges exceeded $180 000) was reduced to less than $9 million—all in equity.

(88) See Cheney (1981), pp. 48– 9. According to the memos exchanged between Westinghouse and Tesla in 1888 and 1889, the former was to pay the Tesla Electric Co. $2.50 per h.p. of electric power sold. It was said that by 1893 the accrued royalties that would have been owed Tesla under these agreements would have amounted to $12 million—considerably more than the assets of the Westinghouse Electric and Manufacturing Co. at that date. Tesla, by then, had already been talked out of his royalties by George Westinghouse, who, reportedly had told him: ‘Your decision determines the fate of the Westinghouse Company’. Were Tesla not to give up his contract, Westinghouse suggested, there was nothing to assure that his inventions would be implemented commercially: ‘In that event you would have to deal with the bankers, for I would no longer have any power in the situation’.

(89) Passer (1972), pp. 298–303; Sharlin (1963), pp. 187–8. Bradley's patents and facilities were bought up by General Electric and the Westinghouse Electric Company developed their own rotary converters over the same period.

(90) Fleming (1921), pp. 238–9.

(91) ‘One engineer estimated that under such conditions, the d.c. system would be twice as efficient as an a.c. system.’ Sharlin (1963), p. 200, quoting from a statement in 1900 by the British electrical engineering authority, J. A. Fleming.

(92) See Landes (1969), p. 286, and Fleming (1921), pp. 238 ff. for technical details.

(93) On the Westinghouse exhibit at The Columbian Exposition in 1893, see for example Prout (1921), pp. 134–40; Sharlin (1963), pp. 206–11.

(94) See Sharlin (1963), pp. 195–210 for discussion of the choice of technology at Niagara.

(95) National Electrical Manufacturers Association (1946), p. 74.

(96) See Byatt (1979), pp. 114; Bowers (1982), pp. 162; Lardner (1903) for discussion of regional grid development.

(97) See Byatt (1979), p. 76.

(98) See Byatt (1979), chapter 4; Sharlin (1963), pp. 185, 188.

(99) T. Hughes' (1983), pp. 120–21, succinct formulation bears quotation: ‘Because “the battle of the systems” had become far more complicated than a technical problem awaiting a simple technical solution, it ended without the dramatic vanquishing of one system by the other, or a revolutionary transition from one paradigm to another. The conflict was resolved by synthesis, by a combination of coupling and merging. The coupling took place on the technical level; the merging, on the institutional level’.

(100) This is a difficulty with which efforts to quantify the effects of major transport innovations, such as the canal and railroad systems, have had to contend [see for example David (1975), Chapter 6]; I am grateful to Edward Constant, who has reminded me of its relevance in the present connection.

(101) Throughout the ‘Battle of the Systems’ electrical engineers debated the advantages and disadvantages of the variant systems in the pages of the Electrical World [see for example Duncan (1988) and February 26, 1887; January 21, 1888; March 31, 1888]. In addition to Lord Kelvin, and Dr Werner von Siemens, two of the more distinguished industry personalities who argued on behalf of direct current until late in the 19th century, there were many other engineers continuing to debate the merits of the two systems at the turn of the century; see for example Barstow (1901) and Scott (1901).

(102) Stanley (1912), p. 565.

(103) The voltages these involved were much lower than those in modern high voltage direct current systems, which recently have re-emerged as a technological area of active research and development interest to the electricity industry in the United States. See, for example, Alvarado and Lansetter (1984), Weeks (1981), Zorpette (1985).

(104) See Parsons (1940), chapter IV, pp. 52–70, for discussion of early h.v.d.c. systems in England. According to the Encyclopedia Britannica (1910–1911), vol. 9, p. 196, the ‘Oxford system’ was distinguished by the use of ‘continuous’ (direct) current transformers to accomplish the drop in voltage from the 1000–2000 v. range at which the current was transmitted from the generating stations to the 100–150 v. supplied by distributing mains to users. Although 3000 v. came to be regarded as the practical limiting voltage for individual d.c. dynamos (due to the problems of sparking at the commutator brushes when running at faster speeds) two or more such machines could be wired in series in order to secure much greater voltages for purposes of transmission. In France, for example, on the Thury direct current system energy was transmitted a distance of 115 miles (between Moutiers and Lyons) at voltages upwards of 45 000 v., using four groups of dynamos in series, each group consisting of four machines in series. See Encyclopedia Britannica (1910–1911), vol. 8, p. 778.

(105) Byatt (1979), p. 100. One additional advantage of the ‘Oxford system’ was that existing d.c. arclamps for street lighting could be worked off the high voltage mains in sets of 20 to 40. See Encyclopedia Britannica (1910–1911), vol. 9, p. 196. See also Lineff (1888).

(106) Landes (1969), p. 285–6 summarizes this with the statement that ‘The two systems competed fiercely in Britain for many years. In the long run, however, victory lay with centralized generators and long-distance transmission.’ Although the latter is indisputable, there is a point in noticing that the ‘victory’ was not an indigenous evolutionary outcome. It came in the 1920s, with the transplantation to Britain of the electric utility system technology which had become the dominant engineering style in the United States. The interruption of WWI had contributed to leaving Britain's electricity supply industry in a rather dilapidated state, considerably behind American practice in terms of generator size and efficiency, and load factors. Compared with British ‘average practice’ c. 1920, American methods looked far superior. But the long-run outcome of the British ‘contest’ cannot be offered in support of the optimality of the course of technological evolution which the industry had followed in the United States. It was not an independent experiment; indeed, had the a.c.−d.c. rivalry not been resolved in the American market so far before, allowing time for much improvement in the design and actual operation of electricity networks, borrowing technology from the United States might hardly have been so attractive to the British in the 1920s.

(107) Louis Bell, ‘Power Transmission’, in Encyclopedia Britannica (1910–1911), vol. 22, p. 234. Bell was Chief Engineer of the Electric Power Transmission Department of General Electric Co., and former editor of Electrical World. See also Weeks (1981), pp. 267–71.

(108) The significance of the rotary converter as an example of a neutral ‘gateway innovation,’ which may have important effects in tipping the balance between a sponsored rivalry, is further explored by David and Bunn (1988).

(109) See Farrell and Saloner (1985a). The indicated route of escape from being ‘locked in’ to a suboptimal technological system depends upon a rigorous backward induction process, which leads the last decision-making unit to switch, given that all others have switched; and the next-to-last to correctly anticipate the decision of the last, and so to switch, given that all before him have already switched; and so on, back to the first decision-maker, who will rationally switch in the expectation that all following him will do likewise. It is a pretty piece of logic, but incomplete information readily breaks this chain, and therefore would prevent it from even beginning to form.

(110) This point has been developed by Katz and Shapiro (1985b). Westinghouse is said to have acted from just such considerations: submitting a successful but money-losing low bid for the contract to light the Columbian Exposition in Chicago in 1893, conscious that the demonstration value would ultimately pay off, possibly by affecting the outcome of the competition for the contract to build a.c. generators for the Niagara Project. See Leupp (1919), pp. 162–70.

(111) Moreover, just those bankers with the financial resources adequate to the task may be the ones most concerned, as was J. P. Morgan, to avert ‘destructive’ price competition, and to seek the pooling of patent rights as a basis for the cartelization, or outright monopolization of the industry in question.