Transectorial Innovation, Location Dynamics and Knowledge Formation in the Japanese Electronic Musical Instrument Industry

TRANSECTORIAL INNOVATION, LOCATION DYNAMICS AND KNOWLEDGE FORMATION IN THE JAPANESE ELECTRONIC MUSICAL INSTRUMENT INDUSTRY

Timothy W. Reiffenstein M.A., Simon Fraser University 1999 B.A., McGill University 1994

DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

In the Department of Geography

O Timothy W. Reiffenstein 2004

SIMON FRASER UNIVERSITY

July 2004

All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author. APPROVAL

Name: Timothy W. Reiffenstein Degree: Doctor of Philosophy Title of Thesis: TRANSECTORIAL INNOVATION, LOCATION DYNAMICS AND KNOWLEDGE FORMATION IN TKE JAPANESE ELECTRONIC MUSICAL INSTRUMENT INDUSTRY Examining Committee: Chair: R.A. Clapp, Associate Professor

R. Hayter, Professor Senior Supervisor

N.K. Blomley, Professor, Committee Member

G. Barnes, Professor Geography Department, University of British Columbia Committee Member

D. Edgington, Associate Professor Geography Department, University of British Columbia Committee Member

W. Gill, Associate Professor Geography Department, Simon Fraser University Internal Examiner

J.W. Harrington, Jr., Professor Department of Geography, University of Washington External Examiner Date Approved: July 29. 2004 Partial Copyright Licence

The author, whose copyright is declared on the title page of this work, has granted to Simon Fraser University the right to lend this thesis, project or extended essay to users of the Simon Fraser University Library, and to make partial or single copies only for such users or in response to a request fiom the library of any other university, or other educational institution, on its own behalf or for one of its users.

The author has further agreed that permission for multiple copying of this work for scholarly purposes may be granted by either the author or the Dean of Graduate Studies.

It is understood that copying or publication of this work for financial gain shall not be allowed without the author's written permission.

The original Partial Copyright Licence attesting to these terms, and signed by this author, may be found in the original bound copy of this work, retained in the Simon Fraser University Archive.

Bennett Library Simon Fraser University Burnaby, BC, Canada ABSTRACT

This thesis explores the relationships between technological discontinuities, spatial discontinuities and regional industrial concentration through a case study of knowledge formation in the Japanese electronic musical instrument (EMI) industry between 1960 and 1995. The investigation aims to highlight the social and institutional dimensions of innovation that emerge during episodes of radical technological change, specifically in relation to problems posed by transectorial innovation, that is the translation of knowledge from new leading sectors to other industrial sectors. Although largely neglected in economic geography, transectorial innovation is of strategic importance during these episodes because it redefines the technological boundaries of previously stable industrial sectors.

Conceptually, the thesis elaborates on the geographical dimensions of inter- sectoral change in a framework that integrates perspectives from evolutionary economics, especially the idea of techno-economic paradigms; recent research on knowledge transfer including with respect to the idea of 'translation'; and related research in economic geography concerned with problematizing inter-regional innovation networks.

Empirically, the thesis explains how the transition from digital to analog sound synthesis, a technical discontinuity, relates to a spatial shift in the industry's centre of gravity from the US to Japan, in particular to firms located in Hamamatsu, most notably

Yamaha and Roland. Successive chapters analyze the rise and fall of the US industry; the ascendancy of Japanese firms; the geography of patent data; the motives, conflicts and consequences of technology transfer for respectively US and Japanese engineers, and; the structure of the EM1 production system in Hamamatsu.

The research design is multi-perspective in orientation and draws on primary and secondary sources, the most important of which include interviews with Japanese and US

R&D engineers, patent data compiled and analyzed by the author and the translation of

Japanese sources such as industry histories and engineering biographies.

The thesis contributes to the industrial geography of Japan by framing

Hamamatsu as a learning region that has benefited fi-om synergistic diversification. More generally, for economic geography, it stresses the inseparability of tacit and codified knowledge, but emphasizes the underestimated role that codification, qua translation, plays in mobilizing knowledge in ways that disrupt heretofore settled industrial geographies. DEDICATION

I do not find it easy to get sufficiently far away from this Book, in the first sensations of having finished it, to refer to it with the composure which this formal heading would seem to require. My interest in it, is so recent and strong; and my mind is so divided between pleasure and regret - pleasure in the achievement of a long design, regret in the separation from many companions - that I am in danger of wearying the reader whom I love, with personal confidences, and private emotions.

Charles Dickens, David Copperfield, Preface to the 1850 Edition

This thesis is dedicated to my Mom, who long ago signalled my attention to the fundamental relationship between music and mathematics, the underlying theme in music's digital revolution. Musical genes bless my maternal line and whatever umble talents I possess merely channel what I like to think is the Irish part of my ancestry. I am also grateful for the opportunity to have taken piano and guitar lessons when I was a kid.

At mathematics, I am no fool, and my cultivated aversion to the discipline in no way relates to my mother having been a math instructor, as well as my tutor. Indeed, my brother John and I fondly remember our parents' subtle reminders that 'Mathematics is not a spectator sport'. My mother also deserves recognition for turning me on to The

Beatles and Dickens. ACKNOWLEDGEMENTS

First, I would like to thank my senior supervisor, Professor Roger Hayter, for his mentorship over two degrees. Back when I was a Master's student, Dr. Hayter fostered my geographical imagination of Japan and sagely advised me to begin language training courses. This excellent guidance has continued from near and far over the course of the

PhD program. I am additionally grateful for the opportunity to do collaborative fieldwork in Japan and it is symbolic that we visited the site of that country's first modern factory in

Kagoshima together. Through our eight year relationship Dr. Hayter's promptness in reading, commenting on, and returning drafts has been exemplary. Finally, I have fond memories of meetings with Roger in which his nimble mind would segue from insightful critiques of my work to the Canucks enduring goaltending woes.

Thanks must also go to the three other members of my supervisory committee:

Dr. David Edgington, Professor Trevor Barnes and Professor Nick Blomley. Dr. David

Edgington has enhanced my scholastic development as a member of my supervisory committee over two degrees. I presented succeeding versions of my PhD research findings at a number of conferences and Dr Edgington has repeatedly asked questions that challenged me to improve my conceptualization of the problem. Moreover, I have benefited by David's superior knowledge of Japan. Dr. Barnes' work inspires me and I feel honoured that he has been a key part of my supervisory committee. Indeed, a paper he presented at the Canadian Regional Science Association in May 2003 stimulated my detour through the 'science studies' literature. I remember Dr. Blomley's self-

vi introduction to my MA cohort in the fall of 1996 as he livedlworked through the ideas that appear in Unsettling the City (2004). At critical junctures Dr. Blomley has improved my project by questioning the 'taken for granted' assumptions of my research design. I would also like to acknowledge my examiners, Professor J.W. Harrington and Dr.

Warren Gill. Dr. Harrington has long supported my efforts as a graduate student and I would like to thank him for travelling to my defence armed with some tough questions.

Dr. Gill's research inspired the musical theme of my thesis so I am honoured that he took part in my defence. As the musician on the panel, he was right to remind me of all the aesthetic drawbacks of digital instruments and I am amazed at how he found a way to relate these critiques to the industry's geography.

I would like to thank Professor Yoshitaka Ishikawa for his supervision over the two years I spent as a research student at Kyoto University. I am especially grateful for his encouragement that I present a paper at the Jimbun Chiri conference, a highlight of my sojourn in Japan. In terms of the practicalities of fieldwork, Professor Jun Nishihara needs to be commended for his tireless enthusiasm and assistance during my stays in

Hamamatsu. Special thanks must also go to Dr. Kazuhiro Uesugi and his family in

Kokubunji for their hospitality.

I would like to thank all my interviewees for their cooperation and insights, and for their patience when explaining the intricacies of electronic engineering to a layperson.

This research has been financially supported by a SSHRC doctoral fellowship and by a Monbusho research fellowship.

I would like to acknowledge my friends (students, staff and faculty) in the

Geography Department at SFU. In particular, I need to thank Marcia Crease, Dianne

vii Sherry, Ravinder Multani, Hilary Jones and Gary Hayward for all their help over the years.

Finally I would like to thank my family and especially my partner Christina for their love, support and patience.

.. . Vlll TABLE OF CONTENTS .. Approval ...... 11 ... Abstract ...... 111 Dedication ...... v Acknowledgements ...... vi Table of Contents ...... ix ... List of Tables ...... XIII List of Figures...... xv CHAPTER ONE ...... 1 1.1 Introduction: From Telharmonium to Keitai Ring Tone ...... 1 1.2 The Research Problem ...... 7 1.2.1 Situating the Study Relative to the Contemporary Literature in Economic Geography ...... -8 1.2.2 Electronic Musical Instruments (EMI) Defined ...... 10 1.2.3 The Significance of Digital EM1 as Radical Innovations...... 11 1.3 The Electronic Musical Instrument Industry and Japan ...... 13 1.3.1 The Digitization of the Musical Instrument Industry and the Transfer of Knowledge to Japan ...... 14 1.4 Theoretical Approaches...... 16 1.4.1 Techno-economic Paradigms, Transectorial Impacts and Spatial Implications ...... 16 1.4.2 The Geographical Formation and Transfer of Technological Knowledge ...... 18 1.4.3 Patenting as Practice ...... 20 1.4.4 Spatial Innovation Systems of Technological Change ...... 21 1.5 Research Questions ...... 24 1.6 Research Design ...... 25 1.7 Outline of the Thesis ...... 28 CHAPTER TWO: THE GEOGRAPHY OF TRANSECTORIAL INNOVATION: A CONCEPTUAL FRAMEWORK ...... 31 2.1 Introduction ...... 31 2.2 The Significance of Transectorial Innovation ...... 36 2.2.1 Transectorial Innovation. Creative Destruction. Uneven Development ...... 36 2.2.2 Techno-economic Paradigms and the Ascendancy of 'Core' Regions ...... 40 2.2.3 Transectorial Innovation and the Evolution of Firm Strategy and Structure...... 42 2.2.4 The Role of Large and Small Firms ...... 44 2.3 The Transfer of Knowledge in Space ...... 48 2.3.1 Tacidcodified Knowledge (Re)Considered ...... 48 2.3.2 From Knowledge Formation in Place to Spatial Innovation Systems...... 50 2.3.3 Conduits of Knowledge Transfer ...... 52 2.4 Science in Action ...... 54 2.4.1 Translation ...... 55 2.4.2 Topological Space ...... -60 2.4.3 Patents ...... 62 2.4.4 Tinkering ...... 66 2.4.5 Reverse Engineering ...... 68 2.4.6 Trade Shows ...... 71 2.4.7 Interlinking Practices ...... -72 2.5 Transectorial Innovation Systems, Engineering Biographies and the Formation of Core Regions as 'Meeting Places' ...... 73 2.5.1 A Local Model of Transectorial Innovation Across Space ...... 77 CHAPTER THREE: THE RISE AND FALL OF THE ELECTRONIC MUSICAL INSTRUMENT INDUSTRY IN THE USA FROM 1890-1980 ...... 88 3.1 Introduction ...... -88 3.2 Instrumental Antecedents to EMI: Pianos and Organs ...... 95 3.2.1 The Piano Industry ...... -95 3.2.2 Organs...... 99 3.3 The Evolutionary Trajectory of Electronic Musical Instruments...... 102 3.3.1 Spatio-temporal Clusters in the Innovation of EM1 ...... 104 3.3.2 The Hammond Organ ...... 109 3.3.3 Into the Laboratory ...... 110 3.3.4 The Audio Engineering Society (AES): A Community of Engineers Looks Forward to the Digital Age ...... 113 3.3.5 Analog Days ...... 122 3.3.6 Towards the Digital Era ...... 130 3.3.7 MIDI [Musical Instrument Digital Interface] ...... 131 3.3.8 Legends of the Fall ...... 134 3.4 Conclusion ...... 136 CHAPTER POUR: THE MUSICAL INSTRUMENT INDUSTRY IN HAMAMATSU JAPAN: CORPORATE STRATEGY. STRUCTURE AND RIVALRY 1880-2000...... 138 4.1 Introduction ...... 138 4.1 .1 Vignette: ICT. Yamaha. Roland and Hamamatsu ...... 141 4.2 Hamamatsu: Industrial City ...... 144 4.2.1 Entrepreneurial Foundations and Transectorial Legacies ...... 146 4.3 Harnamatsu's Musical Instrument Industry ...... -148 4.3.1 The Origin and Pre-war History of Musical Instrument Manufacturing in Hamamatsu ...... 148 4.3.2 The Post-War Period ...... 150 4.3.3 Learning to Mass Produce, Enrolling the Consumer ...... 154 4.3.4 A National Institutional Basis for Mass Consumption of Musical Instruments ...... 156 4.4 Japanese Firms Enter the Electronic Musical Instrument Field ...... 159 4.4.1 The Case of Yamaha ...... 159 4.4.2 Yamaha's LSI Program ...... 164 4.4.3 Yamaha as a Mark I1 Innovator ...... 174 4.5 Competition from an Entrepreneurial Upstart - the Case of Roland ...... 175 4.5.1 The Origins of Kakehashi. Ikutaro . as a Transectorial Entrepreneur ...... 175 4.5.2 Entrepreneurial Invention in Japan ...... 178 4.5.3 First Products Bypass the Home Market ...... 179 4.6 The Case of Kawai ...... 183 4.6.1 EM1 Development within Kawai: the Lack of Local Linkages ...... 185 4.7 The Swarming of Japanese Firms into the EM1 Sector...... 186 4.7.1 The Casio Challenge ...... 189 4.7.2 Post-MIDI Consolidation ...... 191 4.8 Conclusion ...... 192 CHAPTER FIVE: PATTERNS IN THE PATENT RECORD: GEOGRAPHICAL AND TECHNICAL SHIFTS ...... 196 5.1 Introduction ...... 196 5.2 Conceptualizing Patents ...... 197 5.2.1 Patents as Proxy for Innovation (as a Social Process) ...... 197 5.2.2 Patents as Problematic Inscriptions ...... 199 5.2.3 Patents. . as Property ...... -200 5.3 Patenting in Japan ...... 201 5.3.1 The Patent Precedes the Territorialization of Industry ...... 201 5.4 Patterns in Patents and the Spatial Dynamics of Inventive Activity ...... 204 5.4.1 A Brief Note on Regions ...... 204 5.4.2 Regional Trajectories in Patenting Activity ...... 205 5.4.3 National Systems of Innovation v . Corporate Strategies: Comparing the Patenting Performance of US and Japanese Firms...... 207 5.4.4 The Geography of Significant Patents ...... 211 5.5 Discussion ...... -214 5.6 Conclusion ...... -216 CHAPTER SIX: THREE WISE-MEN FROM THE EAST: VECTORS IN THE TRANSFER OF TECHNOLOGY FROM CALIFORNIA TO JAPAN ...... 217 6.1 Introduction ...... 17 6.2 Foreign Experts and Innovation in Japan: Historical Antecedents ...... 221 6.2.1 Technological Learning in Japan and the Role of Transectorial Yatoi ...... 223 6.3 Ralph Deutsch ...... 225 6.3.1 The Deutsch - Rockwell relationship...... 229 6.3.2 The Deutsch .Yamaha relationship ...... 232 6.3.3 The Deutsch - Kawai relationship ...... 233 6.3.4 Ralph Deutsch: An Essential but Controversial Actor in the EM1 Transectorial SIS ...... 234 6.4 John Chowning, FM Synthesis and the Birth of the Stanford-Yamaha Relationship ...... -240 6.4.1 US Firms: FM Falls on Deaf Ears ...... 242 6.5 Dave Smith ...... 246 6.5.1 Approaches to Patenting ...... 246 6.5.2 A Brief Note on MIDI ...... 248 6.5.3 The Demise of Sequential Circuits and the Collaboration with YamahaKorg ...... 249 6.6 Conclusion...... 255 CHAPTER SEVEN: JAPANESE ENGINEERING PERSPECTIVES ON THE TRANSFER OF KNOWLEDGE AND THE COLONIZATION OF TECHNOLOGICAL SPACE ...... 260 7.1 Introduction ...... 260 7.2 Origins: From Tinkering Amateurs to an Innovation System ...... 264 7.2.1 Enrolling Colleagues ...... 266 7.2.2 Linking (Domestic) Colleagues and Allies ...... 269 7.3 Roland ...... 271 7.3.1 Responding to the Commercial Crystallization of Competitors...... 275 7.4 The View from Kawai ...... 278 7.5 Discussion ...... 280 CHAPTER EIGHT: BEYOND THE GREAT DIGITAL DIVIDE: THE TAKE OFF, CONSOLIDATION AND MATURITY OF THE EM1 INDUSTRY - Increasing returns for Hamamatsu ...... 284 8.1 Introduction ...... 284 8.2 Take-off ...... 287 8.2.1 Sampling and the Emergence of a Division of Labour to Exploit Software Synthesis...... 291 8.3 Consolidation ...... 295 8.3.1 The Evolving Spatial Divisions of Labour: The Production System ...... 296 8.3.2 The Local Production System...... 300 8.4 The EM1 Sector as Part of Hamamatsu's Industrial Mosaic ...... 303 8.5 The Maturation of ICT: Implications for the Yamaha - Roland Relationship and by Extension, the Health of the EM1 Industry in Hamamatsu...... 307 CHAPTER 9: CONCLUSION...... 313 9.1 Hamamatsu's Changing Place in the World ...... 313 9.2 Contribution of the Study to the Theory of Innovation and Formation of Knowledge ...... 16 9.3 Contributions to the Industrial Geography of Japan Literature...... 322 9.4 Agenda for Future Research ...... 324 9.5 A Final Word ...... 327 BIBLIOGRAPHY ...... 329 APPENDICES ...... 344 Appendix A: List Of Interviews ...... 344 Appendix B: Ethics Approval ...... 346

xii List of Tables

Table 2.1 A Comparison of Long Wave Periodizations Table 2.2 The Evolution of a Transectorial Spatial Innovation System - The case of an industrial shift from the US to Japan Table 3.1: Estimates of Piano Production 1870-1984 (1,000 units) Table 3.2: Key Innovations in the 2othCentury Development of Analog Electronic Musical Instruments Table 3.3: Session Participant Affiliations at the AES Conventions 1964-1972 Table 3.4: Sample Career Path Profiles of Authors Publishing Papers on Electronic Musical Instruments in the Journal of the Audio Engineering Society (1 969-72) Table 3.5: A Chronology of Moog Music Table 4.1 : Hamamatsu's Industrial Pioneers Table 4.2: Manufactured Goods Produced in Shizuoka Prefecture Ranking First in Japan Table 4.3: Establishment of Musical Instrument Parts Producers in Hamamatsu 1944-1977 Table 4.4: Share of Piano Production within Hamamatsu 1965-1984 Table 4.5: Electronic Musical Instrument Development at Yamaha (1959-1985) Table 4.6: Semiconductor Technology Development at Yamaha Table 4.7: The History of Electronic Musical Instruments at Yamaha Table 4.8: A History of Inter-Firm Relations, Mergers and Acquisitions in the EM1 Sector Table 5.1 US Patents for Electronic Musical Instruments by Region of Assigning Company Table 5.2: US Patents for Electronic Musical Instruments by Region of Inventor's Residence Table 5.3: Patenting Characteristics within the Top Five Japanese Manufacturers 1965-94 Table 5.4: Patenting Characteristics within the Top Five US Manufacturers 1965-94 Table 5.5: Significant Patents:US Patents for EM1 Receiving 25 or More Citations by Inventor's Residence Table 5.6: The Top 20 Most Cited US Patents for EM1 1965-94

... Xlll Table 6.1 : A Synopsis of Ralph Deutsch's Patenting Record 226 Table 6.2: Ralph Deutsch and the Story of the Rockwell Patent 228 Table 6.3: John Chowning and the Story of the FM Synthesis Patent 239 Table 6.4: The Deepening Relationship between Stanford and Yamaha 244 Table 6.5: Dave Smith's Career Chronology 245 Table 6.6 Career Trajectories of Transectorial Yatoi to Japanese EM1 Manufacturers 254 Table 8.1 : Production and Sales Figures for Selected Electronic Organs and Synthesizers 289 Table 8.2: Yamaha's Geography of Factoryllab Establishments Since 1959 298 Table 8.3: Roland's Geography of Factoryllab Establishments 299 Table 8.4: The Social Division of Labour in Hamamatsu's Musical Instrument Industry in 2001 3 00 Table 8.5: Yamaha's Social Division of Labour within its kyoryokokai (Supplier Cooperative Association) in 200 1 301 Yamaha's Functional Division of Labour within its kyoryokokai (Supplier Cooperative Association) in 2001 302 Table 8.7: A comparison of Hamamatsu's social division of labour for EM1 in 1986 and 2001 303 Table 8.8: The Ranking of Japan's Technopoli According to Several Criteria 306 Table 9.1 : The Evolution of the EM1 Transectorial Innovation System 3 17 List of Figures

Figure 1.1 : A Typology of Innovation Systems 23 Figure 2.1 The Circulatory System of Scientific Facts 57 Figure 3.1 The Distribution of EM1 Innovation over 120 Years 103 Figure 4.1 : My Computer's Default Audio Control Panel 142 Figure 4.2: Map of Japan Showing the Location of Hamamatsu 143 Figure 4.3: Map of Hamamatsu City Showing Location of Highways, Railways and Other Features 145 Figure 4.4: Musical Instrument Factories (Final Assembly) in Hamamatsu 1948-1977 152 Figure 4.5: The Agglomeration of Musical Instrument Manufacturers in Hamamatsu in 1960 153 Figure 4.6: The Agglomeration of Musical Instrument Factories in Hamamatsu in 1975 154 Figure 4.7: Piano Production in Japan 1965-2001 158 Figure 4.8: Sales of Keyboard Units by Japanese Manufacturers 1965-2000 158 Figure 4.9: Yamaha's Product Timeline 172 Figure 4.10: Roland's Product Timeline 181 Figure 4.1 1: Kawai's Product Timeline 185 Figure 4.12: Electronic Musical Instrument Manufacturing in Japan: A Timeline of Industry EntranceExit 188 Figure 4.13: Hamamatsu EM1 Factory Location in 1986 189 Figure 4.14: Sales of Keyboard Units by Japanese Manufacturers: The Casio Effect 190 Figure 4.15 : A comparison of US and Japanese EM1 Manufacturers According to the Duration of CEO Tenure and Technological Basis 193 Figure 5.1 A Comparison of the Top Five Japanese EM1 Manufacturers' Patent Outputs 1965-94 209 Figure 5.2: A Comparison of the Top Five US EM1 Manufacturers' Patenting Outputs 1965-1994 210 Figure 8.1 : Imports of Electronic Keyboard Musical Instruments to the US 1989-2001 296 Figure 8.2: Imports of Electronic Keyboard Musical Instruments to the US 1989-200 1 (excluding Japan) 297 Figure 8.3: The Synergistic Development amongst Manufacturing Industries in Hamamatsu 304 CHAPTER ONE

1.1 Introduction: From Telharmonium to Keitai Ring Tone

Thadeus Cahill took out a patent for his telharmonium, the first significant electronic instrument, in 1897'. This invention drew on three scientific and technological developments of the 19th century, namely Helmholtz's overtone theory of 1862, electric generators (1 83 1) and the telephone (1 87612. First, Helmholtz's theory demonstrated that a complex tone could be produced by summing individual sine waves. Second, electric generators produced alternating currents in a sine wave pattern and thirdly, the telephone converted sound into fluctuations in electricity thereby enabling transmission by wire and re-conversion by the receiver. The telharmoniurn consisted of a series of 145 modified dynamos which produced alternating currents of different audio frequencies. Multiple polyphonic velocity sensitive keyboards of seven octaves (thirty-six notes per octave, instead of the standard twelve) controlled these signals which required two performers to operate3.An advertisement of the time explained that, "ground tones are mingled with overtones to produce eflects similar to those of orchestral instruments" (Weidenaar 1995:

18 1, emphasis in original). The musical content performed on the telharmonium tended towards the 'respectable' music of the day. This often meant that selections of Bach and

1 US Patent # 580,035:'Art of and apparatus for generating and distributing music electrically' (Weidenaar 1995). 2 Elisha Gray invented the musical telegraph, a single note oscillator, in 1874 when by accident he discovered that a vibrating electromagnetic circuit could control sound. Gray might also have been credited with the invention of the telephone had not Bell submitted his claim to the patent office an hour prior. After a lengthy period of litigation Bell prevailed and Gray went on to establish the Western Electric Manufacturing Company in 1872, the forerunner of Western Electric (Weidenaar 1995). This combination of equal temperament and just intonation split the 36 keys for each octave into three sets of 12: 12 for the standard equal-tempered tone, 12 slightly sharp and 12 slightly flat (Weidenaar 1995: 63). Chopin were transmitted over telephone lines to the homes and offices of subscribers who had receivers equipped with specialized acoustic horns (Armbruster 1984:18).

One subscriber was Mark Twain who, a little before midnight one New Years eve, told his guests how he;

Had generally been enthusiastic about inventions which had turned out more or less well in about equal proportions. He did not dwell on the failures, but he told how he had been the first to use a typewriter for manuscript work; how he had been one of the earliest users of the fountain- pen; how he had installed the first telephone ever used in a private house, and how the audience now would have a demonstration of the first telharmonium music so employed. It was just about the stroke of midnight when he finished, and a moment later the horns began to play chimes and 'Auld Lang Syne' and 'America'(Paine 1912: 1364-1365, quoted in Weidenaar 1995: 139).

Upon its commercial introduction it was described in the press as the harbinger of

'democracy' in music since it enabled the delivery of music to ". ..towns, villages, and

even farmhouses up to a hundred miles or more from the central station" (quoted in

Armbruster 1984:6). On March 9, 1907, ScientiJic American featured Cahill's invention

on its cover.

The telharmoniurn amounted to no less than an electric power generating plant,

the dynamos of which rotated at the frequencies of the musical scale. As such it clearly

reflected the technological tenor of the times, in which the key invention of alternating

current and the key industry of electrical engineering shaped the paradigm through which

industry and society viewed technology (Hall and Preston 1988:21). Cahill boasted that it

had the power to supply as many as 20,000 subscribers. These numbers were never

realized and its commercial demise related in part to idiosyncratic qualities, not the least

of which was what one performer described as, "its own special character which pervaded everything, and which in time grew highly irritating to the nerves" (quoted in

Arrnbruster 1984:7). In this sense it is no stretch of the imagination to consider the

telharmonium as an antecedent of muzak, the ersatz 'piped' elevator music. That its

transportation also required six rail cars surely limited its market outside of the larger

cities, as did the fact that it had a propensity to interfere with regular telephone signals.

One story has an outraged businessman, upset with the static on his line, breaking into the building that housed the Telharmonium in New York, destroying the instrument and throwing pieces out of the window into the Hudson IXiver4.

If there was any doubt that the telharmonium would not overcome these

substantial constraints in delivering music to the home, it was removed with the advent of

radio in the 1920s. Yet the technological principles embodied in the telharmoniurn

endured to emerge 30 years later in the Hammond organ, the first mass produced

electronic musical instrument. Instead of a series of dynamos, a small electric motor

powered and regulated a series of tone wheels, a mechanism derived from Laurens

Harnmond's experience as a maker of electric clocks (Vail 1997). It may be argued that

the telharmoniurn announced the arrival of the Third Kondratieff (Freeman 1987) - a 50-

60 year long cycle of economic development beginning in the late 19th century - to the

world of music. However, it quickly exited the stage. On the other hand, the Hammond

organ kept the audience listening and is generally considered as the signature keyboard

instrument of its time, particularly in the jazz, blues and gospel idioms.

A little over a century after the birth of the telharmonium, Cahill's amalgam of

music and telephony has been practically realized on a massive scale in the cellular phone

4 Source: httv://www.obsolete.com/l20years/machines/telharmonium/index.html ; (Accessed 2 1/6/04). industry. This technology's pervasiveness is no more apparent than in Japan where cell- phone (keitai) users amount to over 72 million, or roughly 60% of the population

(Ministry of Posts, Telecommunications etc. 2001). Perhaps the most infuriating aspect of the cell-phone, in Japan or elsewhere, is the invasiveness of its ring tone which, however cute and personalized, is still disruptive, particularly in the realm of public space.

Twenty-second snippets of virtually any song you care to imagine are now available for download from service providers who charge a nominal fee of a few dollars per tune to ward off copyright concerns. Telus, Western Canada's largest telecom service provider, runs advertisement campaigns during the NHL Stanley Cup play-offs. The product is a package of ring tones, the content of which includes the theme to CBC's Hockey Night in

Canada. In Japan, 80 billion yen was spent on ring tones and mobile karaoke in 2002

(Dodson 2003). Globally, consumers downloaded US $3.5 billion worth of ring tone songs in 2003, accounting for 10% of the global music economy (New York Times,

January 18,2004).

Sound bites such as these are stored in the phone's memory as digital binary data.

Their reproduction is via hardware consisting of a digital synthesizer miniaturized within the circuitry of a microchip. A device embodying technological knowledge from diverse sectoral sources (telephony, computers, musical instruments and most recently photography), the cell-phone is a thoroughly emblematic technology of the Fifth or information and communications (ICT) Kondratiev - 1980s onwards (Freeman 1987)'.

Looking specifically at its sound synthesizer function, the antecedent trajectory of

In the digital era, the introduction of technologies that perform in the sonic dimension precede those in the visual realm. Thus DVD followed CD audio, digital camera technologies lagged digital musical instruments and in the cell-phone industry, musical phones preceded those with camera functions. technological development traces a similar arc to the one connecting the telharmonium and the Hammond. The first programmable synthesizer that showed the potential for computers in music was borne in RCA's research labs at Princeton, NJ in 1952 (Chadabe

1997). This instrument, like the telharmonium before, proved conceptually fascinating yet practically challenged. Consequently the potential that ICT technologies offered to instrument design was not realized in any pervasive manner until Yamaha, a firm based in Hamamatsu Japan, introduced the DX-7, the first all-digital programmable synthesizer, in 1983 (again, 30 years after the first of its kind). There is a further important connection between this instrument and the synthesizer-in-the-cell-phone.

Yamaha was the first firm in the world to supply application specific integrated circuit

(ASIC) chips to the cell-phone market. Before pioneering this market, Yamaha, which for the first 65 years of the 20~Century was primarily a musical instrument manufacturer, had developed the economies of scope necessary to produce and sell ASIC chips to the rapidly booming 'multimedia' market for in the computer and video-game industries

(Yamaha 1987). Its realization of this technology can be traced directly to a sequence of events that took place between the late 1960s and the mid 1980s - a sequence initiated by its organ division's decision to develop a more cost-effective way of producing tones electronically.

As the above narrative examples show, the electronic production of music has been closely intertwined with technological developments over the last 100 years or so.

Geography has been crucial to these developments, both in setting their original contexts and in shaping the patterns of their adoption elsewhere via technology transfer. To take one example from the start of this trajectory, the telharmoniurn was conceived and developed in Cahill's lab. Other sites like the US Patent Examiners Office, Mark Twain's parlor, the pages of Scientzjk American, even the Hudson River further connect this technology to various 'heterogeneous interests' (Latour 1987): the state, consumers, publishers and competitors. This constellation of interests was highly localized, for the instrument proved immobile beyond the hearth of fin de siecle New York.

Today, digital musical instruments are so commonplace in contemporary society that we barely recognize them as such. Unlike the telharrnonium, they are geographically pervasive. However, because they have become so fused with other technologies of the information age we scarcely consider that, until recently, musical instruments and computers used to occupy entirely different domains. A common digital logic brings these two technological spaces into alignment. This logic, now embedded in the micro- circuitry of various machines, was at one time completely radical to engineers in the musical instrument industry. However, through a process that Rosenberg (2000: 80) calls

"the diffusion of the engineering disciplines", knowledge spawned in leading post-war sectors such as aerospace, was eventually transferred into the realm of music. For example, North American Rockwell, the firm that supplied the navigational systems to the Apollo 11 mission, assigned the most widely cited patent in the field of electronic musical instruments. Piatier (1988) calls the outcome of these inter-industry transfers of technology 'transectorial innovation' (see also Rosenberg et al. 1994). The transectorial introduction of digital electronics to the substantially analog world of the musical instrument industry radically altered the technological trajectory of this sector. 1.2 The Research Problem

This thesis examines the geography of transectorial innovation in the musical instrument industry as an evolutionary, social process. More specifically, it demonstrates why the contours of the post-war half-century long technological shift from analog to digital instruments need to be interpreted in the context of a spatial shift in the core of this industry from the United States to firms located in Hamamatsu in central Japan. As such, it investigates the relationships between a technological discontinuity, a spatial discontinuity and the subsequent regional concentration of an industry. In drawing these processes into the same frame, my objective is to highlight transectorial innovation as both a situated process, and one that is mobilized through the connections between places.

Conceptually I frame the problem through the synthesis of three bodies of literature. First, Freeman and Perez's (1 988) idea of techno-economic paradigms provides the broad evolutionary-economic framework to situate the transectorial impacts of information and communications technologies on the musical instrument industry.

Second, emerging theories in innovation studies detailing the geography of knowledge formation and information transfer (Bryson et al. 2000, Howells 2002) are extended by using the approaches from Bruno Latour's (1987) science studies, especially his concepts of translation and topological space. Third, Oinas and Malecki's (2002) spatial innovation system (SIS) framework is discussed as a way to problematize the relations between regions during times of technological change. These literatures inspire a conceptual model titled the 'transectorial spatial innovation system' which is used to guide the empirical chapters of the thesis. Empirically, the thesis has drawn upon both primary and secondary data sources

collected during two years of fieldwork between 2001 and 2003 when I was based in

Japan. Primary data include interviews in Japan and the US with key informants, principally engineers, directly engaged in the innovation process. The former interviews

were conducted in Japanese by the author, transcribed and subsequently translated into

English by the author. These sources provide a nuanced account of transectorial

innovation as a situated practice. Secondary sources are manifold and include patent data

compiled by the author to assess the aggregate and longitudinal profile characteristics of both the entire data set as well as the a narrower class of widely cited, influential patents.

Another key secondary source is the author's translation of Japanese documents such as

firm histories and other interpretive accounts including the existing geographical

literature of the musical instrument industry in Hamamatsu (Ohtsuka 1980, Takeo 1989).

1.2.1 Situating the Study Relative to the Contemporary Literature in Economic Geography

The geography of technological learning has been a central theme of recent research

(Patchell 1993 Gertler 1995, Storper 1997, Glasmeier 2000). Indeed, in an age in which

Lundvall(1998: 33) has asserted that "knowledge is the most strategic resource and

learning the most important process", geographers have increasingly recognized

industrial learning as a situated process that is territorially and socially embedded. This research has extended the insights derived from the heterodox traditions of institutional

and evolutionary economics (Nelson and Winter 1982), as well as economic sociology

(Granovetter 1985) to reveal, for example, that forms of governance (Christopherson

1999), norms and conventions (Storper 1997) and corporate cultures (Saxenian 1994, Schoenberger 1997) are influential in shaping learning trajectories at local, regional and national scales. This study further elaborates on this research trajectory.

Within economic geography, the theme of innovation has been closely examined, classified in various ways and the basic distinctions between incremental and radical innovations (as well as 'in-between types) widely recognized. As critics have argued, however, recent scholarship has focused largely on articulating the geographies of incremental innovation, that is relatively minor, progressive modifications to existing products and processes. Moreover, much of this work has focused on specific regions.

The literature on 'industrial clustering' is particularly rich in insight as to why 'place' is crucial to innovation (for example: Markusen 1996, Maskell and Malmberg 1999, Pinch and Henry 1999). In contrast, the theorization of inter-regional innovation networks is in its infancy (Amin and Cohendet 1999, Bunnel and Coe 2001, Oinas and Malecki 2002,

Bathelt et al. 2003), while the geography of radical innovation is poorly understood

(Asheim 1999, Hudson 1999). Radical innovations are those discontinuities that appear rather infrequently, precipitating new products, markets and ensuing capital investments.

Hudson (1 999: 346), for instance, notes that in economic geography, "the issue of how radically new knowledge is produced and redefines 'best practice' as radical innovations are created, is left largely unexplored." Asheim (1 999: 346) echoes these sentiments in suggesting that: "one problematic aspect of the 'learning economy' has been its focus being mainly on 'catching-up' learning (i.e. learning based on incremental innovations) and not on radical innovations requiring the creation of new knowledge."

As Freeman (1 982) asserted, radical technological change is difficult to understand because the underlying processes are new and scientifically and institutionally complex, crisscrossing established sectorial and regional boundaries. Indeed, radical innovations represent major discontinuities in technological trajectories and are defined by their re-definition of productivity and by significant and pervasive impacts across sectors and regions. In other words, radical and transectorial technological innovation has to be situated in vastly different, dynamic contexts. This research addresses this challenge by an analysis of the transformation of the Japanese musical industry by absorption of radical innovations initiated in the American electronics sector.

The introduction of digital technologies to the musical instrument industry across the Pacific Ocean from the USA to Japan provides an intriguing focus by which to examine the geography of radical transectorial innovation. Before I address this claim it is necessary to define what is meant by the term electronic musical instruments.

1.2.2 Electronic Musical Instruments (EMI) Defined

Kakehashi's (2002: 165) makes a distinction between electric and electronic musical instruments in the following manner. Electrical instruments, such as the electric guitar, pick up and amplify vibrations of acoustic sound sources. In contrast, electronic musical instruments are defined as those that have a sound source that derives from electrical circuits. Examples include electronic organs, synthesizers, and drum machines. A further distinction within the electronic domain is between analog and digital instruments. The former are capable of exhibiting continuous fluctuations which correspond in a one-to- one fashion with (that is analogous to) the audio output. Both control signals and audio signals may take on any voltage value in a spectrum so that, for instance, a one volt increment is equivalent to a semi-tone. Digital instruments employ computer type binary operations performed by micro-processors to store and retrieve information about sound. Continuous fluctuations in value (such as amplitude) are divided into discrete quantized steps.

Digital keyboard instruments that perform like computers seem a long way from their antecedents, the piano and organ. Indeed there has been a technological leap so significant that basically the only shared property of the piano and organ is that they are played via a keyboard. Yet it is this simple property that has industrially tied together the fates of piano, organ and synthesizer families. Japanese firms, in particular Yamaha and

Roland have been instrumental in this convergence. Moreover, it is uncommon to find a piano retailer who does not stock acoustic pianos and their digital counterparts in the same showroom. It is this definitional challenge that digital instruments pose to their conventional counterparts which is most radical when set against the historical backdrop of the latter. In 1985 Keyboard magazine gave voice to this sentiment when it published a forum article entitled, "The Piano: Can it survive the electronic age?"6

1.2.3 The Significance of Digital EM1 as Radical Innovations

Theberge (1 997) suggests that the technical basis of digital instruments, which include synthesizers, samplers, drum machines and technologies that combine these three functions, is radically different from that of traditional musical instruments. He argues that principles of design rely more on electronics and digital logic than on acoustics. In addition, their tone producing mechanism is entirely distinct from the user interface, the device such as a keyboard that allows them to be played. Within the field of electronic instruments, digital technologies represent a clear discontinuity from their analog counterparts. Instruments that are based on a digital logic are necessarily hybrid devices,

6 Keyboard Nov/Dec 1985: 76-94. for performance on a synthesizer, drum machine or sampler involves the production of melodic or rhythmic patterns which, in turn, hinges on a process of technical reproduction. In practice, this characteristic means that,

"popular musicians who use new technologies are not simply the producers of prerecorded patterns of sound (music) consumed by particular audiences; they, too, are consumers of technology, consumers of prerecorded sounds and patterns of sound that they rework, transform and arrange into new patternsV(Theberge1997:3).

Theberge's thesis is that the introduction and diffusion of digital musical instruments altered the relationship between the production and consumption of technology, producing a fundamental change in the way that musicians practiced their art. To illustrate this point, it is worthwhile to compare two examples that demonstrate how evolving technology has enabled and limited the musical imaginations of composers and performers.

The first example quotes Karlheinz Stockhausen (Poschardt 1998: 353, emphasis added), a German composer who is widely regarded as one of the pioneers of electronic music for his application of mathematical logic to musical composition. Working with a sine generator in a basement studio in 1950s' Paris, "I systematically produced the first sound spectrums by overlaying sine tones, an endlessly arduous task (there was no tape recorder in the studio and I had to copy each sine tone on record, and copy it on from record to record! !)" Stockhausen's then novel sampling of himself, that has become widespread since the 1980s in the hip-hop genre and beyond on the one hand is enabled by his systematic application of mathematics. On the other hand, it is limited by the unease with which existing technology facilitates reproduction. In comparison, Kraftwerk

(ibid, p. 348), pioneers in the idiom of 'techno' music, speak of their own use of

synthesizer technology in 1991: "Recording possibilities keep pace with our ideas.. .the idea and the product are to a large extent identical. Only now has software developed to the point that it can appropriately convert the software in our heads."

Clearly the technical palette for artistic composition had evolved profoundly between the 1950s and 1990s. At the start of this technological trajectory, musicians such as Stockhausen and other avant-garde artists were compelled to wear the hats of applied mathematicians and engineering technicians. By 1977, Boulez (quoted in

Poschardt 1998:354) referred to the challenge facing collaboration between musical and scientific realms when he claimed that, "musical invention will have somehow to learn the language of technology and even to appropriate it. The full arsenal of technology will elude the musician, admittedly; it exceeds, often by a big margin his ability to specialize." In the 1990s technological advances afforded musicians this precise advantage for specialization as they could now buy the applied math and engineering technique off the shelf. Theberge's study which deftly explains the contours of a story that links Stockhausen and Kraftwerk is primarily concerned with the musician- technology interface; hence consumption. My research builds on Theberge's work by examining this revolution from the supply side, in particular via a focus on the engineers and firms that gave impetus and direction to this technological trajectory. In this manner, musical instruments are interpreted as capital goods for artistic production. Electronic instruments employing digital logic can similarly be interpreted as Computer Numerical

Control (CNC) machines which enable the flexible production of music.

1.3 The Electronic Musical Instrument Industry and Japan

Though most of the early moments in the technological trajectory from analog to digital instruments during the 1960s and 1970s took place in the United States their ultimate occurred later and hinged on the agency of Japanese firms. As much as this is a historical study of an important technological change, its primary focus is concerned with the transfer of technology between places and the processes which served to re-embed the core of this industry in Japan during the 1980s. This re-embedding is not well understood.

1.3.1 The Digitization of the Musical Instrument Industry and the Transfer of Knowledge to Japan

In Japan, the musical instrument industry has always concentrated in Hamamatsu, an important centre within the industrial belt linking Tokyo and Osaka. At present the three largest musical instrument manufacturers in the world (Yamaha, Kawai and Roland) are based in Harnamatsu (The Music Trades 2000). These firms have been the leaders of the global music industry in absorbing new innovations generated within the electronics sector, most notably the key early innovations that originated in the US during the 1960s and 1970s. Indeed, these firms led a decisive spatial shift in inventive activity from the

US to Hamamatsu, that in turn foreshadowed the latter's production dominance.

The production of western musical instruments in Hamamatsu, Japan, began in

1887, when Torakusu Yamaha established the Yamaha Organ company7. It is instructive that the origins of the largest musical instrument manufacturer in the world lay in the reverse engineering efforts of its founder. Yamaha was employed as a traveling repairman of medical appliances for a company based in Osaka. During a stop-over in

Hamamatsu, Yamaha was asked if he could repair the local elementary school's imported

7 At that time reed organs in Japan were known asfukin (m%) which is composed of the kanji for wind and koto, a Japanese zither. This term has largely disappeared from the lexicon, replaced by the katakana- ized orugan (ifhf>I. -. reed organ. In order to fix the organ, it was necessary to take the instrument apart and at some point during this process, Yamaha's fascination with the technology inspired him to build his own organ. Piano production at the Yamaha Company was initiated shortly thereafter in 1897. In 1927, Koichi Kawai, one of the original seven Yamaha employees, an engineer and head of the piano action section, established the Kawai Musical

Instrument Company. The fierce rivalry between these two firms has continued to this day, interrupted only by the war years when both firms were assigned to the production of aircraft components, such as propellers (Ohtsuka 1980). Roland's story begins with an

Osaka entrepreneur, Ikutaro Kakehashi, who moved his small electric organ concern to

Hamamatsu in 1968 to begin design and production of Hammond organs under an OEM contract (Kakehashi 2002). Roland has since established itself as a specialist in electronic musical instruments and accessories such as rhythm machines and guitar effects pedals.

The firm's corporate motto, 'To be best not the biggest', reflects an appreciation of its role vis-a-vis the two market leaders that Roland chose as neighbours.

All three of Hamamatsu's firms achieved their prominence in the EM1 field through ambitious efforts to absorb and apply forms of transectorial knowledge. For instance, much of Kawai's early knowledge was cultivated through partnerships with another of Japan's major companies, Toshiba. Roland's first analog rhythm devices were designed with the collaboration of the electronic firm Pioneer (Colbeck 1996: 99).

Yamaha's commitment to the emerging sector was signaled by its decision in 1967 to design and manufacture its own application-specific integrated circuits (ASIC), making it the smallest firm in the world to produce micro-chips (Nakazawa 1984, Johnstone 1999).

This project, besides shaping the trajectory of its electronic keyboard division, put Yamaha in a position to supply chips to others sectors, in particular the emerging market for multimedia sound cards, which are used in computer and video game applications. In

1994, Yamaha's sound chips accounted for 95% of sales to the $1billion market for

sound cards (Johnstone 1999). In this study, I devote particular attention to the examination of the various avenues of exogenous learning that linked firms in the

Hamarnatsu region with sources of knowledge outside of Japan in the USA, conceptually framing these processes as spatial innovation systems (Oinas and Malecki 2002).

1.4 Theoretical Approaches

This section reviews a number of relevant theoretical debates and literatures that have

informed my thesis investigation

1.4.1 Techno-economic Paradigms, Transectorial Impacts and Spatial Implications

According to evolutionary economists (Freeman 1982, Rosenberg 1982, Dosi l988),

long-run, technological change evolves, for the most part, along trajectories that are

ordered and incremental. Rather less frequent are 'major' changes that disrupt these trajectories by introducing radical discontinuities (Dosi 1988). Freeman and Perez (1988),

with a nod to Kuhn (1 962), term these 50 year cycles techno-economic paradigms

(TEPs). The idea of TEPs, which highlights both technological and institutional factors,

adds depth to Kondratieff s long wave theory. TEPs assert their character via the

diffusion of carrier technologies that emerge in key industries. The influence of these

carrier waves on other industries is pervasive because they redefine the long-run cost

curves of much of the economy and give shape to both the pace and direction of

technological trajectories in formerly un-related sectors and places (Hall and Preston 1988). Within these trajectories, a given industry has a prescribed range of innovative options or "cognitive limits" (Oinas and Malecki 2002: 106) that are governed by the prevailing state of technology as well as by a set of social conventions that vary between localities.

According to Freeman and Perez (1 988) the current era, is often known as the

Information and Communications Technology (ICT) paradigm. It began with the micro- electronics revolution in the 1960s and gathered steam in the 1970s with advances in the processing capacity and memory capability of integrated circuits. The widespread applicability of these technical advances propelled the diffusion of ICT throughout the economy spawning accompanying institutional changes. Indeed, much economic, cultural and social activity in everyday life, from writing an essay, to composing a film score, is now mediated by these technologies. Implicit in such accounts, is the idea that TEPs gain their momentum through a cascade of transectorial impacts from carrier to dependent sectors. As Rosenberg (1992: 75) puts it,

"Technological innovation typically exercises its impact upon the economy through a process of inter-industry transfer of technologies, originating at any time in the small cluster of industries that are generating a large flow of new technologies."

These impacts are most acute at the transition between TEPs, when the 'rules of the game' (North 1990) are sufficiently changed to the degree that the basis for competition within a sector, indeed the very definition of the sector itself, is highly uncertain. Though transectorial knowledge is continually introduced to and absorbed by sectors throughout the course of a paradigm, its influence is most radical and thus particularly pivotal for regional development early in the process, in a manner consistent with theories of path dependence (David 2000). Indeed, the way these threshold moments unfold is affected by both the persistence and adjustment of institutions as this knowledge is put into practice.

1.4.2 The Geographical Formation and Transfer of Technological Knowledge

Geographers have long been interested in patterns and processes of innovation diffusion over space and through time (Hagerstrand 1968, Brown 1981, for a review see Gregory

2000). A number of scholars have recently sought an alternative framework to interpret the ways in which knowledge formation and information transfer influence innovation and technological change (Nonaka and Takeuchi 1995, Howells 2002, Malmberg and

Maskell 2002). These latter efforts derive their conceptual nuance from Polanyi's (1966) distinction between codified and tacit knowledge. The former comprise scientific or engineering knowledge such as technical publications, patents, and blueprints that exist in textual form and can therefore be acquired in the absence of direct experience of that knowledge. The latter, on the other hand, is highly contextual as it is "embodied in skilled personal routines or technical practice" (Asheim 1999:348) at scales specific to the machine, factory and firm. In Polanyi's (1967) words, "we know more than we can tell."

Recently geographers have emphasized tacit rather than codified knowledge in understanding technological transfer (e.g. Maskell and Malmberg 1999). Perhaps because tacit knowledge is so intangibly rooted in particular situations, geographers have devoted far more attention to attempting to capture the ways that its formation allows clusters of related firms to become "sticky places in slippery space" (Markusen 1996), than to understanding codified knowledge. Also, with the recent emphasis devoted to interactive learning processes (Lundvall 1992) that occur within industrial districts, clusters, and regional innovation systems - hence within place - codified knowledge is dismissed as an object of study precisely because it is transferable and subject to processes of cosmopolitanization (Storper 1997) and ubiquification (Maskell and Malmberg 1999).

This dismissal of codified knowledge underestimates its importance, particularly vis-a-vis the tacit dimension.

Nonaka and Takeuchi (1 995) suggest that tacit and codified knowledge should be seen not as a dichotomy but as poles of a continuum, noting, in addition, that it is the cycling between them that propels learning processes and therefore innovation. Asheim

(1999) has picked up this line of reasoning and called attention to the idea of disembodied knowledge, or forms of codified knowledge that can only be put into practice in specific epistemic communities that Wenger (1998) calls communities of practice. In this thesis I address the interrelatedness or cycling of knowledge between tacit and codified domains as it unfolds through technological change. My main focus is to consider the institutional aspects that shape both the process of knowledge codification and responses to that codification by other actors during these episodes.

Compared with tacit knowledge, codified knowledge can be seen as the 'recipe' for industrial learning. Consequently the strategic importance of formal reports published in scientific and engineering journals and patents is recognized through various conventions, such as intellectual property regimes designed to ensure its 'appropriate' production and dissemination. Moreover, the manner in which knowledge is published has immediate and long-lasting implications for how it is put into practice. Journal articles and technical reports advance the state of knowledge in a discipline. Patents perform a similar function in that they are published in order to claim novelty and will figure in a significant way in my thesis. Their supplementary legal status, on the other hand, limits who is allowed to make practical use of this knowledge. In tracing the emergence of the patenting system in 1gth century Britain, Stirk (200 1: 475) notes how the lobbying efforts of key inventors and industrialists such as James Watt, "prompted detailed arguments about both property rights in general and the position of manufacturers within society in particular." This example also points to the idea that novelty is socially constructed and the definition of what constitutes 'the new' may be manipulated to serve powerful interests. Much later in Japan in the 1950s and 1960s, the patent regime was structured in such a way that importers of US musical instruments could patent these technologies 'as is' (Kakehashi 2002), implying that domestic claims to novelty could overwrite extant foreign knowledge. As I will show, these inscribed contours of the knowledge economy underwrite the dynamics of industrial location.

1.4.3 Patenting as Practice

Because an analysis of EM1 patents features in this thesis, a brief review of the patent process is pertinent. Patenting is an important institution in both the legal and technical sense; a covenant between an inventor and the government whereby the latter provides the former with an incentive to make new industrial knowledge publicly available through the provision of a legal framework to safeguard that intellectual property. Ceh

(2002) notes that patents are both outputs of, and inputs to, the innovation process. Thus, they are widely considered as an output measure by economists (Schmookler 1966,

Freeman 1974), geographers (Pred 1966, Ceh 1996, Mizuno 200 I), and regional scientists (Jaffe 1989, Jaffe et al. 1993, Acs et al. 1994), who use patent counts as a proxy for the measurement of innovative performance in a manner that allows for the mapping of patterns of concentration and dispersal in inventive activity over time. On the other hand, Jaffe et al.'s (1 993) investigation of localized knowledge spillovers considers patents as an input to innovation by looking at patterns of citation. For the most part, geographers have tended to use these data in only the narrowest of manners. Quite often, the failure to distinguish between significant and insignificant patents turns patent analysis into something of a blunt instrument, running roughshod over the important distinction between incremental and radical innovation. Also, while it is frequently noted that firms display different propensities to patent new knowledge that vary both within and between sectors and between countries (von Hippel 1988, Granstrand 1999), few have investigated how geography might shape this practice. Exceptions include a recent strand of analysis drawing on actor-network theory has redirected attention away from what patents represent towards the work that patents perform (de Laet 2000, Doe1 and

Rees 2003). Other qualitative analyses highlight the way that firm structures and strategies and, by extension, regional economies evolve in correspondence with the patenting regimes (Winder 1995,2001). These approaches are complementary to my own perspective which considers the examination, publication, maintenance and contestation of patents as practices that are embedded in institutions at various scales. Such perspectives derive their inspiration from the science studies approach of Bruno Latour

(1987) which gives emphasis to scientific and engineeringpractices.

1.4.4 Spatial Innovation Systems of Technological Change

In response to the wealth of research on industrial districts (Storper and Scott 1992,

Harrison 1992, Takeuchi 1992, Park and Markusen 1995, Markusen 1996; Patchell 1996; for a review see Amin 2000), regional innovation systems (Braczyk et al. 1998) and learning regions (Morgan 1997), that emerged in the discipline from the late 1980s, a number of authors have noted that innovation is not only a process that occurs within places, but also between them (Bunnel and Coe 2001, Coe and Bunnel2003, Amin and

Cohendet 1999). Oinas and Malecki (2002) provide a framework by which these spatial processes may be better understood. Their spatial innovation systems (SIS) approach builds on the National Innovation System (Lundvall 1992, Nelson 1993, see also

Freeman 1974, Britton and Gilmour 1978) and Regional Innovation System (Braczyk et al. 1998) literature. These studies suggest that the institutional characteristics of places shape innovation propensity, potential, and performance. However, the broader SIS approach of Oinas and Malecki (2002) transcends these territorially defined frameworks by focusing on how technological systems (Carlsson and Stankiewicz 1991), which are sets of technologies in use in specific interlinked industries, evolve in space and time and how innovation can be limited between regions and even between national systems.

Figure 1.1 sorts the various types of innovation systems according to their geographic scale and degree of sectoral concentration. SIS can thus be viewed as problem-solving networks comprised of overlapping and interlinked international, national, regional and sectoral systems of innovation. The boundaries of an SIS are not explicitly spatial, but rather are demarcated according to the hctional relations between actors. This approach emphasizes, first, the external relations between the various actors concerned with a common problem, and second, the variability of the relative weights of different places or regions as centre points of particular technological paths in time. Moreover by recognizing that "spatial discontinuities may relate to specific phases of technological cycles"(0inas and Malecki 2002: 11 O), such as at the frontier between paradigms, the SIS concept is particularly relevant for my analysis of EM1 links between the USA and Japan. The adoption of the SIS approach also affords the opportunity to contribute in novel ways to the already rich literature on Japanese innovation systems. That Japan, as a nation, is a consummate industrial learner is well established, even if explanations for its post-war emergence as an economic powerhouse vary (Johnson 1982, Best 1990,

Lazonick 1991). Indeed, Japan was the inspiration for Freeman's (1987) early formulations on national innovation systems. At the other extreme, i.e. at the scale of the factory, Japan has also served as a model for students of industry (Dore 1979, Fruin

Figure 1.1: A Typology of Innovation Systems .------\

Local Urbanization 0 Economy .I C, A e I I C, Regional I I I 8 Innovation i Spatial : 0 System E Innovation i S I System I National a Innovation .I *a System ma International

Diversified 4 Specialized Sectoral Concentration

Geographers have made their most insightful contributions to this literature through the analysis of Japanese industrial networks, with particular attention being paid to the orchestration of vertical relations between manufacturers and suppliers that effectively harness the learning potential of the latter (Sheard 1983, Kenney and Florida 1988,

Patchell 1993, Edgington 1999; see also Asanuma 1989), through collaborations where proximity does matter (Mizuno 2002). Others have broadened the focus to consider transnational learning networks established by core firms via manufacturing transplants

(Mair et al. 1988, Mair 1997, Liker et al. 1999, Edgington and Hayter 2000, Hayter and

Edgington 2004) research centres (Angel and Savage 1997) or sourcing practices

(Reiffenstein et al. 2002).

Since one of the key characteristic traits of Japanese business networks is the patience of capital with a view to realizing long-term strategic goals (Christopherson 1999), an emphasis on the phenomenon of kuizen or continuous learning has drawn attention away from the way Japanese industry has dealt with technological discontinuities and radical change (although see Glasmeier 2000). The SIS is therefore an appropriate heuristic for problematizing this dimension.

1.5 Research Questions

The general research problem of the thesis is to understand the connections between a technological discontinuity, a spatial discontinuity and the (re)concentration of an industry away from its original hearth, with specific reference to the case study of the electronic musical instrument (EMI) industry.

The research questions are as follows:

1) How did radical technological change, via the introduction of digital

information and communications technologies (ICT) to the musical instrument

industry reshape this sector's geography?

2) As new technological knowledge was introduced to the sector, how did

processes of knowledge creation, transfer, interpretation and application

contribute to the formation of a spatial innovation system that lost its footing in the US and became increasingly centred on Hamamatsu? What were the

avenues by which transectorial knowledge which was developed in the US

entered the Japanese EM1 industry?

What was the role of American engineers and innovators within this

innovation system? Specifically, how did their interests come into alignment

with Japanese firms?

What was the role of different Japanese EM1 firms?

How has the spatial innovation system for EM1 evolved since these

discontinuities?

1.6 Research Design

My research design adopts a 'multi-perspective' approach to capture the depth and scope of these processes which cross the boundaries of firms, sectors and continents. I first examine the uneasy tension between organizational and individual agencies, whereby the motivations of the latter (e.g. engineers) imperfectly align with those of the former (e.g. companies) and vice-versa. The way these different interests are made to 'work together' is problematized with respect to Latour's (1 987, 1999) concept of 'translation', which is the "process, strategy or method by which actors (individuals, organizations) attempt to enrol each other into a network" (Bryson 2000: 159). To situate these processes in an evolutionary framework, corporate strategies and structures with respect to innovation are examined over time in a longitudinal manner from 1960 to 2000 (Laulajainen 1982). To balance these organizational perspectives I focus on individual inventors and engineers in order to understand innovation as a social process (Freeman 1982). It should be mentioned that relative to the wealth of literature on corporate managers, very little geographical research has investigated the role of the engineer (for exceptions see

Shapira 1995, Schoenberger 1997, Pinch and Henry 1999, Henry and Pinch 2000a); a lacuna which this study endeavours to remedy. A further focus of my approach highlights patents. From an initial extensive analysis of patent data, I reveal patterns of inventive activity at the individual, firm and regional levels. These quantitative data are complemented by a qualitative examination of stories surrounding specific patents which proved significant in shifting the direction of the technological trajectory. In sum, these multiple levels of analysis (firms, individuals, patents) are complementary. However, quite often the distinctions between these units of analysis are permeable, such as when the inventor of a particular technology that may or may not be represented by a patent document enables the transfer of that technology beyond the boundary of the organization andlor region from which it originated. Finally, these intersecting units of investigation are analyzed at a variety of spatial scales of resolution in a way that reconciles space- based perspectives, such as the SIS model with place-based perspectives that highlight the embedded nature of specific practices found within sites.

This research has drawn on primary and secondary sources of data. Primary sources consisted of semi-structured interview with key informants (engineers, managers, entrepreneurs) that were conducted in both English and Japanese during the period fiom

2001 to 2003, while secondary sources included the US Patent and Trademark Office

Database, trade journals, firm and industry histories and memoirs of key individuals.

These data sources are complementary and provide a variety of vantage points for quantitative and qualitative analysis. Industry histories identify, for instance, key informants while providing a description of their technological contributions in lay terms.

It is also possible to identi6 key inventors from patent data.

For the quantitative analysis of patents, my data set consisted of all US patents

(2375) in the field of electronic musical instruments published between 1965 and 1995.

These data were obtained from the US Patent and Trademark Office's Database

(www.uspto.~ov)by conducting an all field search using the term "electronic musical instrument".

Twenty individuals in total were selected for semi- structured open ended interviews and/or e-mail questionnaires (See Appendix). These respondents are classified into two groups. The first group comprised American inventors and engineers who acted as vectors in the transfer of knowledge to Japan. In these interviews I was particularly interested in uncovering the motivations, practices and outcomes of what Oinas and

Malecki (2002) refer to as the external relations of actors. The second set of respondents were primarily engineers but also managers of innovation employed by Japanese musical instrument manufacturers. In these cases, the intention was to speak with individuals who were active during the pivotal decades of the 1970s and 1980s, the period when Japanese firms engineered their dominance of the industry. My focus on historical events means that many of my respondents have retired, or are nearing retirement, and thereby occupy senior positions in their respective firms. In all cases except for one, these interviews were tape recorded, transcribed, and in the case of interviews conducted in Japanese, translated into English by the author. One of my respondents, an American inventor, the sector's most prolific patentee and also one of the key individuals responsible for the transfer of knowledge to Japan requested that our interview not be recorded. The reason for this is that this individual's accomplishments have been of such strategic importance in shaping the fortunes of a number of different firms (American and Japanese), that his mobility has proven controversial to the point that he has been involved in a number of lawsuits. Given his position with respect to these highly charged events, I was not permitted to record our conversation.

Throughout my representation of these interviews, firm names and the identities of respondents have not been altered. This convention was permitted by my respondents.

My rational is that since much of my objective is to uncover the chains of contingency that link the mundane, messy and inspirational aspects of engineering practice with the patents, papers and products that register in the public record, it is necessary to position these individuals within a narrative that can bridge the two. This strategy is consistent with a geography that endeavours to highlight how practice and instance give shape to the world (Lee 2002). In other words it accentuates how the mundane everyday geographies of engineers intersect with one another and with other actors such as firms in highly contingent ways.

1.7 Outline of the Thesis

The thesis is organized in the following manner. Chapter two establishes the conceptual framework of the thesis. Three literatures inform my perspective. I first make explicit the recurring theme of transectorial innovation in the evolutionary economics literature.

Specifically, I argue that transectorial innovation allows techno-economic paradigms

(Freeman and Perez 1988) to become pervasive. I next highlight the complementarities between the economic geography literature on knowledge transfers and Latour's (1987,

1999) Science Studies approach. In this manner I am able to explain the 'diffusion of the engineering disciplines' (Rosenberg 2000) as a chain of 'translations' (Latour 1999:3 1 1).

The third literature I discuss is interested in the way technology transfers unfold within what Oinas and Malecki term Spatial Innovation Systems (2001). I modify this concept to highlight the transectorial dimension. These three perspectives contribute to the development of my heuristic model, a table which details the evolution of transectorial spatial innovation systems. The rest of the thesis is structured roughly to fit with the sequence presented in the model.

The following two context chapters deal respectively with the rise and fall of instrument design and manufacture in the United States (chapter three), and the trajectory of industrial development in Japan (chapter four). Chapter five is an analysis of spatial and temporal patterns in inventive activity in both the USA and Japan, derived from the patent records of the USPTO. A key finding of this chapter is that transectorial knowledge, particularly from sources such as the aerospace industry, proved significant in shaping the direction of the subsequent technological trajectory of the EM1 sector. The following two chapters depart from the quantitative analysis of chapter five to interpret transectorial innovation as a social process. In chapter six, I look closely at the career trajectories of three American inventors who acted as vectors in the transfer of key technologies to Japan. Chapter seven presents Japanese engineering perspectives on the formation, application and contestation of technological knowledge. Specific attention is devoted to a comparison of Yamaha, a firm that was already a large international firm at the dawn of the electronic age, and Roland, an entrepreneurial upstart which by rivalling its neighbour, has grown into a global corporation specializing in electronic musical instruments. Chapter eight investigates the evolution of the EM1 production and innovation systems in Japan following the moment of radical innovation. Through the phases of take-off, consolidation and maturity, Hamamatsu firms have locked-in increasing returns. These technical and commercial developments are juxtaposed with an analysis of the contemporary division of labour in production that is both globalized and localized. I focus on Hamamatsu as a particular type of learning region that offers lessons for the way that geographers think about the relationship between patterns and practices of industrial transformation and regional development. I draw on Takeuchi's (1 996) interpretive scheme of industrial development in Harnarnatsu as a 'local model' (Barnes

1996) that illustrates how transectorial impacts drive local dynamics. Hamamatsu, as

Takeuchi argues, is probably the most successful medium-sized city in the world, as far as performance in a range of industries is concerned. It dominates the global musical instrument industry, is known as the 'Motorcycle City' of Japan and it is home to the most successful Technopolis in that country. Only a handful of authors, all Japanese, have written about industry in Hamamatsu from a geographical perspective (Ohtsuka

1980, Takeo 1988, Yarnashita 1990, Oda 1992). I build on these studies by illustrating the way that industry in Hamamatsu has evolved via transectorial and trans-regional linkages. In chapter nine I conclude the thesis by reflecting on the major empirical findings and relating these to the transectorial spatial innovation system model. CHAPTER TWO: THE GEOGRAPHY OF TRANSECTORIAL INNOVATION: A CONCEPTUAL FRAMEWORK

2.1 Introduction

Transectorial innovation (Piatier 1988) is the transfer of knowledge from ascendant and leading sectors of the economy to other sectors. This process induces a radical technological discontinuity in the 'receiving' sector that rewrites the qualitative limits of

'engineering common sense' (Freeman 1982) and decisively alters the long-run basis for competition. Examples of transectorial innovation, in which Japanese industry has played a central role, include the translation of computer logic to the camera and musical instrument sectors. Both of these industries have restructured significantly in response to the advent of digital technologies and this thesis focuses on the latter. Digital age inventions "transform numbers, images and texts from all over the world into the same binary code inside computers" (Latour 1987: 228), allowing for the enhanced handling, combination, mobility and conservation of data; features that makes this general pattern of innovation pervasive. Economic geographers (Dicken 1992, Hayter 1997, Malecki

1997, Glasmeier 2000) have drawn on the Freeman and Perez's (1988) techno-economic paradigm (TEP) model, which strongly implies transectorial impacts, to frame their

analyses of long-run patterns in the spatial dynamics of industry. However, little is known about the geography of transectorial innovation. An important exception is Steven Pinch and Nick Henry's work on 'Motor-Sport

Valley' (Pinch and Henry 1999, Henry and Pinch 2000a, b) that analyzes a number of key inter-sectoral knowledge transfers from the aerospace sector to the motor-sport industry. The origin and destination of these transfers took place largely at an intra- national scale and manifest a significant spatial overlap between sectors. Beyond these studies, there remains a lacuna in our understanding of how transectorial innovation overspills regional boundaries to operate at a distance. Transectorial innovation needs to be more extensively studied for it is the key to understanding the pervasive effects of radical technological change, another lacuna in the literature (Asheim 1999, Hudson

1999).

Geographically, transectorial innovations destabilize the hitherto entrenched advantages held by core industrial regions, while simultaneously providing a platform from which other regions can leapfrog these cores. However, the contingent geographies of people and place define the necessary conditions for radical technological change in a given industry8.As Henry and Pinch (2000a: 126), citing Storper (1995), suggest,

"radically new technologies require different input chains in terms of knowledge and material inputs". Moreover, the people, places and circumstances that introduce the spark of transectorial knowledge may differ from those that carry the torch later on, implying an important connection, or more correctly a chain of connections bridging the two positions. Various forms of knowledge transfer drive these spatial discontinuities and precipitate the shift in an industry's core.

8 As Dicken (1998: 145, emphasis in original ) notes, "in one sense, then, technology is an enabling or facilitating agent: it makes possible new structures, new organizational and geographical arrangements of economic activities, new products and new processes, while not making particular outcomes inevitable. But in certain circumstances, technology may, indeed, be more of an imperative." In episodes of radical change, transectorial innovation becomes such an imperative. Technology transfers follow a spatial form that is, "topologically complex"; a geography "where regions intersect with networks" (Law 1999:7) to bring spatially distant places into connection via the circulation of people, texts and artifacts. The

'region', here, is a 'meeting place' for exogenous and endogenous forces (Holmen 1995) in the networked space of flows. Within this space, Law and Hetherington (2000: 39) make an important distinction between 'knowledge at a distance, or surveillance' and

'action at a distance, or domination'. The earliest surveyors were the explorers, and their forays into the unknown precipitated colonialism, which was orchestrated from the centre but enacted in the periphery, on the ground (Harris 2004). Similarly, ascendant 'core' industrial regions transform their status from peripheral to central because they leverage knowledge and action to orchestrate the logistics of connection in such a way that they

"force the world to come to the centre - at least on paper."(Latour 1987: 233).

Consequently, the mobilization of codified knowledge (patents, diagrams and other texts) in particular, enables individuals and organizations located in these centres to "act at a distance, that is to do things.. .that sometimes make it possible to dominate.. .the periphery" (Latour 1987: 2 19). Borrowing from Baudrillard's (1 994: 1) oft repeated phrase, "the map precedes the territory", it is the patent which precedes the colonization of industrial space.

This chapter interprets the geographical implications of transectorial innovation from two perspectives: the evolutionary (institutional) economics rooted in veblen9and

Schumpeter and the science studies of Latour (1987, 1999). While these perspectives on

- 9 In various works Veblen viewed the engineer as the hero of capitalism, whose relation with business interests were continually held in tension (Tilman 1993) technological change have different starting points and remain independent of one mother they overlap significantly regarding their interpretation of innovation.

Recently, within the context of evolutionary economics, Freeman and Perez

(1988) have articulated the model of techno-economic paradigms (TEP) (see also

Freeman et al. 1982, Nelson and Winter 1982, Dosi 1982, 1983, Perez 1983, 1985, Pavitt

1986, DeBresson 1989). This thesis adopts TEP theory as a point of departure for analyzing transectorial innovation. TEP theory places analytical priority on innovation as the basis for understanding long-run patterns of industrialization. In this model, evolution is both path dependent and subject to change, principally because of radical or paradigmatic innovations that have pervasive effects across all sectors. Indeed, the stimulus for this approach (e.g. Freeman 1982, Rosenberg 1982) was the desire to unravel the 'black box' of technology, traditionally regarded as a given exogenous condition in mainstream economics. For Rosenberg (2000: SO), "the study of technological change needs to devote far greater attention to the emergence and diffusion of the engineering disciplines", the engineer's role in technological change, and the geographies through which this role is enacted.

In this model, technological and institutional innovations that underpin economic evolution are interpreted as social, uncertain, and contingent processes(Hayter 2004:

103). Thus innovations are created by individuals, increasingly formally trained as scientists and engineers but many not, acting on their own or within groups in a variety of organizational circumstances and in specific places at particular times. Some ideas and innovations fail because they do not work, are not supported or cannot compete.

Moreover, many ideas are highly contested. Indeed, Freeman and Perez (1 988) label their TEP model as 'economic7 to formally recognize that evolution is characterized by choice and some choices fail. In this regard, innovation is also contingent, for example, upon the imperatives imposed by crises such as war and recession (see, for e.g. Mensch 1979, cJ:

Freeman et al. 1983), as well as by organizational and geographical context. Innovation is fundamentally a learning process involving the creation of different codified and tacit knowledges that have practical value. Moreover, innovations build upon existing ideas and learning and in turn become a source for the transfer of knowledge.

The problematic nature of knowledge transfers, whether to other firms, sectors, people and places, connects to Latour's (1987, 1999) 'science studies' perspective and especially to the concept of 'translation' which is defined as the 'modifications',

'displacements' and 'persuasions7 (Latour 1999: 3 11) that comprise the, " process, strategy or method by which actors (individuals, organizations) attempt to enroll others into a network" (Bryson 2000: 159). Evolutionary economics and science studies both interpret technological learning as a social process. Both are concerned with time and space. Moreover they share a similar objective - for Latour7sscience studies also aims to pry open technology's 'black box' with a view to unravelling the chains of contingency that animate, stifle and maintain the problem-solving networks of scientists and engineers. On the other hand, science studies is explicit in its network ontology'0. It also opposes binary categorizations, and for this reason it has been used by geographers to recast dualisms such as local/scientific knowledge (Murdoch 1994) and culture/economy

(Barnes 2004) as hybrids. Additionally, science studies is more overtly anathema to what it terms 'diffusionist accounts' of technoscience (Latour 1987: 133). Especially on this

10 Hence science studies and actor-network theory (ANT) are often taken to be synonymous, even as key authors remain cautious regarding their identification with the latter (Law 1999, Latour 1999, see also Thrift 2000) last point, science studies and the TEP 'evolutionary' model sit uneasily in juxtaposition, at least until the latter perspective is cleared of the charge of technological determinism.

Nevertheless, the key overlap between these two conceptual frameworks that this thesis highlights is in the attention devoted to the social processes that distinguish 'core' regions' approach to innovation.

This chapter develops a framework for analyzing the geographic implications of transectorial innovation in three main stages. In the first part, the significance of transectorial innovation is underlined and its nature as a social process, especially the roles played by large corporations, small firms and individual 'inventors', is elaborated.

The second section critically assesses the duality routinely made between codified and tacit knowledge, especially in relation to transectorial innovation. This assessment then critically incorporates Latour's concepts, especially that of 'translation', to work through the tacitlcodified dichotomy. In doing so, it discusses four important 'practical', albeit problematical, forms of technology transfer, namely patents, tinkering, reverse engineering and trade shows. The last section focuses explicitly on the geography of transectorial innovations by interpreting Holmen's (1995) idea of regions as 'meeting places' for exogenous and endogenous forces in terms of the exchange of know-how.

2.2 The Significance of Transectorial Innovation

2.2.1 Transectorial Innovation, Creative Destruction, Uneven Development

Schumpeter (quoted in Freeman 1987: 130) first articulated the notion that the mobilization of 'new combinations' propels the 'creative gales of destruction' in economic development. He further pointed out that innovations are, "not at any time distributed over the whole economic system at random, but tend to concentrate in certain sectors and their surroundings", and are thereby prone to be, "lop-sided, discontinuous and disharmonious by nature" (Fels 1964, pp.75-7, quoted in Freeman et al. 1982:33, emphasis added). In this and subsequent theory, transectorial innovation is a recurring, if implied theme - from Perroux's (1956) 'junction effects'" to Breshnaham and

Trajtenberg's (1 995) 'innovational ~omplementarities"~.Within this tradition, Freeman and Perez's (1 988) TEP model offers the most comprehensive treatment for capturing both the scope and depth of the 50-60 year cycles of economic development first observed by ~ondratieff'~.Briefly, the TEP model posits that key technologies developed in leading sectors diffuse through the rest of the economy affecting both the technological basis and institutional architecture of adopting industries. This diffusion is not gradual but uneven, precisely because the social aspects of technology introduce so many contingencies into the process. Nevertheless, TEPs accumulate a momentum that makes them all-encompassing - for a time. This sequence proceeds until the potential of the key technology is exhausted or rivaled by compelling opportunity costs of a new key technology, at which point a new TEP emerges. In other words, TEPs become pervasive precisely because of their transectorial impacts. To understand what goes on in a contemporary music recording studio (Theberge 1997), or sawmill (Rosenberg et al.

11 Perroux is interpreted as advocating the spatial concentration of junction effects amongst an integrated constellation of industries around a leading sector or industry ('industry motrice') (Lee 2000: 326-327). 12 Distinct from, but closely allied and complementary to Perez's (1985) idea of key technologies is the concept of General Purpose Technologies or GPT (Bresnahan and Trajtenberg 1995, see also Lipsey et al. 1998) which provide technological capabilities that could be adopted by a large number of application sectors. Historical examples include steam engines, machine tools and electricity, and more recently in the post-war era, transistors and computers. The pervasiveness of a GPT is fueled by, "innovational complementarities, that is, the productivity of R&D in a downstream sector increases as a consequence of innovation in the GPT technology" (Ibid: 84). I thank Martin Andresen for suggesting this reference. 13 Since the TEP model is well known and reproduced in a number of economic geography textbooks (Knox and Agnew 1989, Dicken 1992, Hayter 1997, Malecki 1997) - it is merely outlined here. Hayter (2004: 8-10) provides a justification of the TEP model for institutional approaches in economic geography. A number of authors have drawn on the idea of TEPs, in whole or part, for further studies of particular industries (Pavitt 1986, Hall and Preston 1988, Freeman 1990, Glasmeier 2000) 1994), for that matter, one needs to appreciate the ways in which information and communications technologies (ICT) - the 'key technology' (Perez 1985) of the fifth

Kondratieff - have transformed work in those settings.

Scholars' interpretations of the turning points, and hence periodization of each

Kondratieff wave, vary considerably (Table 2.1) (Freeman 1987 c.f: Hall and Preston

1988). These differences take on a cumulative nature such that Hall and Preston view the

4fi Kondratieff wave ending in the year 2000, while Freeman chooses to wrap up this wave by 1980. Despite these differences, the various forms of long wave models all recognize that the technological and institutional dynamics within and between industries, and by extension firms, are driven by the pacing and placing of developments in the relationship between science, engineering and industry. Moreover, they posit a restless geography of leading core regions in successive long waves - in each version of

Table 2.1 the bottom rows of the table could be placed on top to highlight the geographic rather than historical specificity of each long wave. In short, over the long-run, responses to the challenge of transectorial innovation vary between firms, regions and nations. As

Dosi (1 983: 79) notes,

If paradigms and trajectories present strong technological and economic interdependencies between different industrial sectors, a correspondence between discontinuities in the emergence of new technologies and discontinuities in patterns of economic growth becomes quite plausible. Table 2.1 A comparison of long wave periodizations Characteristics of the Four ~ondratkvwaves. Source: based on Hall and Preston (1988: 21) I ~onawave 1 IS'Kondratiev 1 2ndKondratiev 1 3rd Kondratiev 1 4'h Kondratiev 1 !jthKondratiev 1 Dates beriodicitv) 1 1787-1845 (58 yrs) 1 1846-95 (49 yrs) 11896-1947 (51 yrs) 1 1948-2000 (52 yrs) 1 2004-? I Key lnnovations I power loom; puddling Bessemer steel; alternating current; electric transistor; computer, CIT I steams hi^ liaht: automobile I Key lndustries cotton; iron steel; machine tools; ships cars; electrical electronics; computers; Convergent Information engineering chemicals communications; Technologies aerospace; producer services Industrial Small factories; laissez Large factories; capital Giant factories; 'Fordism': Mixture of large 'Fordist' Organization faire concentration; joint stock cartels; finance capital and small factories company (subcontract); MNCs International Britain, workshop of world Germany; American USA, German leadership; American hegemony; competition; capital export colonization Japanese challenge; rise of NICs; NlDL Characteristics of the Techno-economic Paradigm!5. Sources: based on Frleeman (1987:68-75) and Hayter (1997: 21-22) Long wave Early mechanization Steam power and I Electrical and heavv I Fordist mass lnformation and (1'') Kondratiev railways (2nd) engineering (3rd) - production (4Ih) communication (5Ih) Kondratiev Kondratiev Kondratiev Kondratiev Dates I 1780s-1830s (50 yrs) 1830s-1880s (50 yrs) 1890s-1930s (40 years) 1930s-1980s (50 years) 1980s to? Key Factor lndustries cotton, pig iron coal, transport Steel 1 energy - oil 'Chips' (micro-electronics) Main 'carrier branches' Textiles (incl. chemicals, Steam engines, Electrical engineering, Autos, armaments, Computer, electronic and infrastructure machinery), iron working steamships; railway heavy engineering (incl. , aircraft, consumer capital goods, software, and casting, water power; equipment, machine tools, armaments), heavy durables; synthetic telecommunications, fibre- potteries. Trunk canals iron. Railways, world chemicals. Electricity materials and optics, robotics, ceramics. shipping. supply and distribution. petrochemicals. Digital telecom networks Highways, airports. and satellites. Other sectors growing Steam engines, Heavy engineering Autos, aircraft, Computers, television, Third generation rapidly from a small machinery telecommunications, , radar, NC machine tools, biotechnology products base radio, aluminium, drugs, nuclear weapons and processes, space consumer durables, oil and power, missiles activities, fine chemicals plastics Key organizational Entrepreneurs and small Limited liability joint stock Giant cartels, trucks. Oligopolies, MNCs. Arms- Networks of large and features firms. Local capital companies Middle management. length subcontracting small firms. Quality Stable utilities control, trainina. JIT Technology leaders UK, France, Belgium UK, France, Belgium, Germany, US, UK I US. Germany. UK Japan, US, Germany, I Germany, US Sweden 2.2.2 Techno-economic Paradigms and the Ascendancy of 'Core' Regions

Christopher Freeman's (1987) Technology Policy and Economic Performance which is subtitled, Lessons from Japan is the clearest (if again primarily implicit) articulation of the geographical impacts of TEPs. In outlining the case of how Japan, the first and most successful model of non-western industrial development, leveraged ICTs as a means of engineering its post-war ascendancy in a range of sectors (machine tools, cars, electronics, musical instruments), Freeman raises geography as a powerful defense against the criticisms of technological determinism. Simply put, technological performance is spatially uneven and hence contingent. If technology, in and of itself is deterministic, how does one explain why Japan's post-war trajectory was routinely underestimated by US industries competing in the same sectors? For that matter, how many East Asian (and western) nations have sought and failed to emulate the lessons posed by Japan? Freeman's interpretation is neither technologically nor culturally deterministic, but it does paint a compelling picture of Japan as a unique 'core' industrial region.

Freeman argues that a set of characteristics define Japan's national system of innovation (NIS): the role of MI TI'^; R&D at the enterprise level, specifically reverse engineering; education, training and related social innovations, and; competition, the keiretsu and industrial structure. These features of the NIS enabled a range of different industries to deftly harness the potential of ICT technologies. The precise articulation of these characteristics varies according to sectoral context. However, in a general sense, across a range of industries, these features proved, "well suited, both to the timely

14 MITI - The Japanese Ministry of International Trade and Industry was renamed the Ministry of Economy, Trade and Industry (METI) in 2001. identification of the crucial importance of this all-pervasive technology and to its rapid and efficient diffusion throughout the economic system." (Freeman 1987: 54).

Importantly, this diffusion was not a forgone conclusion, for "Japanese firms often seemed to take a relatively long time taking the first decisions about design and development, because it involved a great deal of internal debate, discussion, experiment and training." (ibid: 44). Less clear, from Freeman's account is how particular industries and specific firms negotiated and determined strategy and structure to orchestrate their particular manifestation of this general process or, how sub-national geographies shaped the course of development.

I extend Freeman's research trajectory by providing an interpretation of one industry's pathway to global supremacy. By focusing on the accumulation and concentration of power in a particular 'core' region, this thesis highlights various instituted processes, geographic scales and network configurations through which firms based in the core enacted transectorial innovation. The argument I develop in the rest of the thesis builds on Theberge's (1997) observation that transectorial innovation, via the introduction of digital electronics to a substantially analog world, profoundly altered the technological trajectory of the musical instrument industry and so transformed the realm of musical practice. My contribution is to illustrate why geography is critical to this process, in particular by focusing on the transfer of knowledge to Japanese industry, a topic which receives little consideration from Theberge and others (Chadabe 1997, Pinch and Trocco 2002). Though my conceptual framework is tailored to capture the trajectory of the musical instrument industry, this specific example is put forward to illustrate a more general process which as yet is poorly understood. 2.2.3 Transectorial Innovation and the Evolution of Firm Strategy and Structure

The tension within Schumpeter7sown work over the question of agency in innovation

(Phillips 1971) - entrepreneurs or endogenous science within large firms - highlights the complementary and conflicting motivations of individuals and organizations. Freeman's

(1987: 44, emphasis in original) account of Japan appears to privilege the role of large firms: "The Japanese success.. .seems to have been based far more on an integrative approach within large firms." Missing in his account, is a recognition of the role that smaller firms and entrepreneurs play in the innovative process that is independent of their function as suppliers to larger enterprises (Asanuma 1998, Patchell 1993). These smaller firms must be brought to the foreground to share the stage with larger enterprises.

Since enterprise geography (Hayter and Watts 1983), the firm has been of fundamental categorical relevance within economic geography (Schoenberger 1997;

Yeung 2000,2001,2002). Geographers now have a nuanced and pluralistic conceptualization of the firm, interpreting the enterprise as a 'nexus of treaties' (Aoki et al. 1990) situated within "the context of wider social relations, political economic processes and environmental change.. .the firm in industrial geography goes beyond being an economic entity, it is also a socio-spatial construction embedded in broader discourses and practices" (Yeung 2000: 301). These discourses and practices include

"buyer-seller linkages, subcontracting ventures, local business cultures, conventions and institutions, as well as various types of competitive capital and labour transactions"

(Rigby, 2000: 213). Advances in our appreciation of the firm as a 'complex' unit of analysis have significant implications for how we evaluate the geographic dimensions of the innovation strategies of firms.

Firms possess differential visions of the technological landscape (Fransman

1998). They also possess varied capacities for response. In short, radical transectorial innovation poses a challenge for corporate strategy and structure. Corporate strategies are a firm's long-term plans in relation to the market (Hayter 1976, Harrington 1985) while corporate structures are the organizational architectures that firms put in place to realize strategic goals. Once corporate strategies and structures are in place, firms develop a set of routines through which they accumulate enterprise specific skills (Kioke and Inoki

1990)15.Faced with the discontinuities posed by radical transectorial innovation, these routines are altered, while the coupling between individuals and organizations is subject to strain. Theories of business segmentation (Galbraith 1967, Taylor and Thrift 1983,

Hayter et al. 1999) offer an interpretive frame to make sense of the ways in which industrial structures and strategies shape and are shaped by transectorial impacts. These theories originate in the Schumpeterian debate concerning agency in innovation (Phillips

1971). As Freeman et al. (1982) argue, Schumpeter's views on the innovation process shifted as the leading sectors he observed changed. Thus, initially he emphasized the heroic entrepreneur (Mark 1 innovation), while by the late 1930s, sectors like chemicals were dominated by the endogenous science taking place within massive engineering departments (Mark I1 innovation).

15 "Kioke and Inoki (1990) emphasize the skill formation of workers through on-the-job training over time.. .[creates] what they term 'enterprise-specific skills'. Thus enterprise-specific skills are acquired by workers over time, and bind firms and workers together according to the unique characteristics of these skills" (Hayter 1997: 3 12, see also Patchell and Hayter 1995). 2.2.4 The Role of Large and Small Firms

In the case of transectorial innovation, especially in relation to 'receiving' industries, it is more difficult to discern apriori the degree to which economy of scale effects (and hence the role of larger firms) matter. Large firms do not necessarily possess an advantage over individual entrepreneurs for introducing transectorial innovations to their respective industry. Indeed, internal diseconomies in large firms may dampen their abilities to colonize new technological niches. On the other hand, nascent industries that derive their existence from transectorial impacts are likely to have low barriers to entry for small firms. In this regard, as ideal types, Mark I1 and I innovation posit distinct entry strategies to new industries and the dynamics of interrelation that enfold as a sector matures.

Large firms enter new sectors by strategies of diversification that emphasize either the purchase or internal development of technology - a 'make-or-buy' decision.

Whole firms may be acquired, individuals hired or discrete technologies such as patents licensed. Regardless of the form this diversification-by-acquisition takes, technologically, powerful economies of scope are possible if the purchase is complementary to existing or planned areas of skill. Such synergistic outcomes are also engineered through the enactment of internal strategies. On the other hand there are the cases where financial motives trump the logic for technological synergies in a manner that pits Veblen's (1904,

Tillman 1993) business and industrial interests against each other.

Overwhelmingly, diversification in North American and European firms is enacted through acquisition, often via conglomeration, that is a business model that spreads risk by lumping together unrelated businesses under common ownership. Indeed, for American industry in the 1970s, conglomeration attained considerable currency, albeit a debateable one (Bluestone and Harrison 1982). In fact, conglomeration as a 'prevailing habit of thought' was brought to bear on the musical instrument industry with disastrous results (see Chapter Three). This example also points to the way in which national context shapes the transectorial strategies of firms. Conglomeration, in the US has often been a poor strategy for implementing transectorial innovations that relies on some sort of technological affinity. In contrast, various authors have noted the effectiveness with which Japanese firms have realized internal economies of scope via synergistic diversification (Kodama 1995, Granstrand 199916) and the musical instrument industry presents a unique example of this learning style.

Given that small firms face lower barriers to entry during the formative phases of an industry's development, these business opportunities attract entrepreneurs. Key sectors or disciplines that provide the foundation for TEPs are also likely to be important sources of entrepreneurial talent. In this manner the lead entrepreneurs and intrapreneurs of

'receiving' sectors are likely to be technological migrants. Piatier (1988), almost in passing, refers to transectorial migrations of individuals - human vectors who transfer knowledge between spheres in the course of their career. In the case of the electronic musical instrument sector, Theberge (1 997) relates a few anecdotes about individuals who honed their engineering common sense while working in the computer industry and then migrated to the nascent synthesizer industry. Their histories animate Rosenberg's

(2000) 'diffusion of the engineering disciplines'.

The supply of entrepreneurial talent enables the creation of a population of small firms, some of which are autonomous of larger enterprise, others of which establish relations with the latter. The business segmentation literature again suggests that geography matters. For instance, relative to the United States, small firms play a much

-- - -- 16 Granstrand (1999) cites Yamaha as one company that demonstrated this skill.

45 larger role in the Japanese economy, especially as suppliers to larger enterprises (Patchell

1992), and constitute an important part of production systems. These constitutions notably include their capacity to perform innovation via the cultivation of relation- specific skills in serving core firms (Asanuma 1989, Patchell 1993)17.In this sense, a reasonable hypothesis is that small firms in Japan may be more predisposed to hone their transectorial learning through strategies of technological bonding with larger companies.

Whether or not they persist in the long-run, these relations give smaller firms opportunities to enhance enterprise-specific skills (Kioke and Inoki 1990).

Neither large nor small firms appear to enjoy a particular advantage at the birth of new sectors (Von Hippel 1988). However, over time the evolution of industrial trajectories rewards strategies that effectively leverage transectorial impacts in a sustained manner. Large firms possess greater financial resources to enact technological strategies which, if channelled directly, are subject to increasing returns. For instance, they can usually afford to hire more engineers who in turn accumulate enterprise-specific skills. Financially, large firms are also better able to arrange the purchase of newer technologies, whether these are in the form of technological licenses or the latest products. In general, one would expect that their resources to enact search strategies evolve to become quantitatively and qualitatively distinct from smaller firms. However,

strategically savvy smaller enterprises can persist in the face of these constraints if they too deepen their technological specialization to exploit niches and/or draw on external economies. This perspective brings into focus a non-static view of industrial evolution

17 The relation-specific skill (RSS) provides a conceptual basis for interpreting both the dynamism and stability of manufacturer-supplier relationships in Japanese production system (Asanuma 1989, Patchell 1993). The RSS "defines how suppliers serve the specific needs of core fms(Hayter 1997: 359), while offering an analytical framework for understanding how this 'technological bonding' transforms production systems into learning systems (Patchell 1993). that transcends dual models of business segmentation. Thus, small firms prosper by morphing into 'threshold firms' (Steed 1982) or the type of enterprise whose evolutionary trajectory enables them to become 'big firms locally' (Hayter 1997).

As industries evolve and economy of scale effects become more pronounced, rapidly growing small firms cum medium-sized enterprises can rival larger enterprises if they are able to cultivate areas of specialty in the new technological landscape. According to Porter (1 990) the nature of rivalry, which is an extension of collective strategies and structures plays a determinative role in creating or dampening regional advantage.

Competition within regions does not have to be a zero-sum game, if populations of firms of varying sizes outperform territorial production complexes elsewhere. Patchell (1 996) suggests that the success of these 'core' regions hinge on their particular kaleidoscopic arrangement of the forces of competition, cooperation and control.

Corporate strategies are discerned by researchers in various ways, notably via longitudinal profiles (Laulajainen 1982) and biographies (Schoenberger 1997,2000).

Longitudinal profiles catalogue and analyze temporal patterns in a firm's sales, product mix, CEO tenure and succession, and industrial location in order to paint a picture of that corporation's strategic trajectory relative to its competitors. Biographical approaches, on the other hand, aim for a closer examination of the everyday lives of corporate decision makers, including their perceptions of time, space and technology. In economic geography, Schoenberger's (1997,2000) research has enhanced our understanding of the negotiation of power and knowledge, and hence the contested nature of corporate strategies and organizational cultures, within large firms. Transectorial innovation forces organizational change and this thesis discusses such strategic re-alignments both longitudinally and biographically.

Given paradigmatic shifts in technology and the implications these episodes pose for geographies of industrial structure and firm strategy, the discussion now moves to an examination of knowledge transfers in the innovative process.

2.3 The Transfer of Knowledge in Space

2.3.1 TacitJCodified Knowledge (Re)Considered

It is now widely recognized that technical knowledge forms the basis for industrial learning, and by implication regional development (Lundvall 1992). If information is comprised of bits or strands of data, then knowledge involves a considerably broader process hinging on cognitive structures which possess the capability to assimilate and contextualize information (Howells 2002). The creation, transfer, acquisition, interpretation, protection, contestation and application of knowledge as they are marshaled to the 'practical mastery of technology' (Storper and Walker 1989) are inherently socio-spatial processes.

Although many commentators point out the flaws of relying on a rigid distinction between tacit and codified knowledge, this dichotomy is regarded as the most useful starting point for more complex arguments concerning the geography of knowledge (e.g.

Asheim 1999, Arnin and Cohendet 1999, Howells 2002, Bathelt et al. 2003). Codified knowledge, also known as formal, explicit and articulated knowledge, concerns those forms of scientific or engineering knowledge that exist in textual form and are thereby potentially tradable. Blueprint schematics, circuit diagrams, chemical formulas, patents, operating manuals and textbooks are all examples of codified knowledge. The manner in which this knowledge is conveyed - for instance whether it is published as part of legal documents, such as patents, or made freely available in a textbook - has important implications in limiting how that knowledge may be put to work. Tacit or informal knowledge, on the other hand, in Polanyi's (1967) words is the highly personalized and contextual realm of knowledge in which, "we know more than we can tell." Tacit knowledge is "embodied in skilled personal routines or technical practice" (Asheim

1999:348).

However, viewing tacit and codified knowledge as dichotomous leads to potentially dangerous generalizations regarding their respective geographies. Typically it is asserted that in an age of instantaneous modes of communication, codified knowledge is increasingly transferable, and this potentially 'ubiquitous' quality makes it (literally) difficult to keep in place (Maskell and Malmberg 1999). On the other hand, tacit knowledge is deemed to be context-dependent and 'sticky' and thereby prone to be temporally and spatially bound within specific locations. Asheim (1999) modifies the tacitlcodified dichotomy by suggesting an intermediate form of contextual knowledge,

'disembodied knowledge' that is comprised of locally sticky forms of codified knowledge. Disembodied knowledge are those forms of codified knowledge that, rather than being universal in their interpretability, are only capable of being discerned within localized communities of, "experience-based, tacit knowledge and competence, artisan skills and R&D-based knowledge, and are thus an outcome of positive externalities of the innovation process.. .collective technical culture and a well developed institutional framework" (Asheim 1999:348). For example, only those firms who are acutely aware of their local competitors' technical capabilities through long-standing competition and co- location can assess the significance of, say, a new patent which in theory might be universally available. In addition to Asheim's idea of disembodied knowledge, other authors posit a more nuanced relationship between tacit and codified knowledge that hinges on various forms of proximity, whether spatial or technical (see, for example

Nonaka and Takeuchi 1995, Lawson and Lorenz 1999, Arnin and Cohendet 1999).

Lawson and Lorenz (1 999:3 1I), citing Nonaka and Takeuchi's (1995) account of knowledge formation in Japanese firms, point out that the, "cycling between tacit and articulated knowledge.. .is a key component in the product innovation process." I would add that this cycling extends beyond the bounds of the firm and is the basis for the reproduction of 'core' places as regional innovation systems (Braczyk et a1 1999). In particular, and in reference to Dosi's (1988: 288, see also Storper 1997) 'un-traded interdependencies', which comprise conventions, rules, practices and institutional norms specific to a given industry andlor place, May et al. (2001) suggest these relational forms are akin to a sectoral tacit knowledge or, to paraphrase Lawson and Lorenz (1999) an industry's organizational memory. Thus, innovation within cores depends on the interplay between these tacit knowledges and their articulated counterparts. A modifying element to this local cycling is the 'disrupting' (Lee 2002) influence of geography for innovation systems extends beyond territories as knowledge is mobilized (Bunnel and Coe 200 1,

Oinas and Malecki 2002).

2.3.2 From Knowledge Formation in Place to Spatial Innovation Systems

Much of the geographical literature on industrial learning has implicitly adopted a place- based perspective in its treatment of innovation. The national innovation systems

(Lundvall 1992), learning regions (Morgan 1997) and regional innovation systems (Braczyk et al. 1997) literatures have all gained currency for helping us understand the institutional basis of innovation in a highly scalar manner However, a recent strand of literature has started to reflect on how processes of innovation and industrial learning unfold not within discrete scales but among these scales over space (Amin and Cohendet

1999, Gertler 2001, Bunnel and Coe 2001, Oinas and Malecki 2002, Coe and Bunnel

2003). Regions are indeed different, and economic geographers must be sensitive to the substantial and subtle variations between places. Yet, regions also inter-relate with one another and endogenous and exogenous forces construct regions as 'meeting places'

(Holmen 1995). What sorts of agencies bring regions into alignment for the purposes of innovation?

One recent strand of the literature has answered this question by referring to

Wenger's (1998) notion of 'communities of practice', which are characterized by: i) a mutual engagement among participants; ii) joint enterprise - formal/informal conventions of mutual accountability - and; iii) a shared repertoire of stories, discourses and so on.

Coe and Bunnel(2003) call attention to the debate between Arnin and Gertler over the efficacy with which communities of practice can transfer knowledge between sites. Arnin and Cohendet (1 999) assert that the organizational and relational proximity that define communities of practice substitute and even supersede physically proximate relations.

Gertler (2001 : 18), however, is sceptical of this view: "the idea that organizational or relational proximity is suficient to transcend the effects of distance (even when assisted by telecommunications and frequent travel) seems improbable."

Coe and Bunnel(2003: 446, emphasis added) adopt a sympathetic mid-range position that advocates a view of innovation systems as, "constituted by constellations of communities of practice." This perspective is dynamic and sensitive to the mechanisms that negotiate the boundaries between communities of practice. As they argue, "while such communities will originally almost certainly be local configurations, over time sustained and repeated interaction facilitated by 'boundary crossers' may create new spatially extensive constellations" (ibid).

These insights are complementary to Oinas and Malecki's (2002) notion of spatial innovation systems (SIS) which are comprised of overlapping and interlinked national, regional and sectoral systems of innovations. The SIS extends these frameworks by focusing on how technological systems (Carlsson and Stankiewicz 1991), which are sets of technologies in use in specific interlinked industries, evolve in space and time. The SIS gives emphasis to: i) the role of individual actors as intermediaries in global networks and; ii) the interrelations between former and emergent leading regions in the determination of particular technological trajectories. These aspects make the SIS particularly suited to model the geography of radical technical change via transectorial innovation. Oinas and Malecki (2002: 110) assert that, "spatial discontinuities may relate to specific phases of technological cycles". In addition, the SIS allows broad systemic processes, such as the emergence and ascendancy of TEPs via the 'difision of the engineering disciplines', to be reconciled with the practices of individuals who mobilize knowledge by putting it to work, within and between specific sites. Given this agency, through what channels is knowledge transferred?

2.3.3 Conduits of Knowledge Transfer

Various authors (Gertler 2001, Howells 2002) critically reflect on the three conduits by which spatial knowledge transfer is deemed to take place. These are: texts and other forms of codified knowledge, people, via their movement, association and other communicative exchanges and artifacts or products.

The first conduit, textual knowledge has received considerable attention in the regional science's local knowledge spillover literature (Jaffe et al. 1993, Audretsch and

Feldman 1996, for a review see, Breschi and Lissoni 2001). This research approach typically examines patent data, looking for patterns in codified information transfers.

Spatially localized concentrations in patent citation are considered to be indicative of knowledge spillovers. For a while these approaches seemed to have been translated into geography uncontested. Recently, they have met with strident criticism. For instance, the fluid metaphor 'spillover ' has been a lightning rod for criticism with various authors offering challenge to the assumption that knowledge can be said toflow (Howells 2002,

Doe1 and Rees 2003). Howells (2002: 876) takes issue with the conceptual inferences derived from this type of data analysis, arguing that patent citation is a metric that,

"impl[ies] the imparting of knowledge, but do[es] not actually measure it."18 Labour mobility, the second channel for spatial knowledge transfer, has been examined by several authors (Audretsch and Stephan 1996, Zucker and Darby 1996, Almedia and

Kogut 1997). These studies affirm that the mobility of scientists and engineers amongst laboratories and firms is an important vehicle of transferring ideas. Incidentally, that these studies also found that much of this mobility occurred locally serves to indirectly reinforce the local spillover authors. They also, however, share a drawback of the

l8 Doe1 and Rees (2003) similarly dismiss the spillovers literature for flaws in it assumptions, namely that patents 'represent' innovation or citations 'represent' knowledge transfers. For example they point out that the practice of citing prior art in patent documents is rarely undertaken by engineers (or patent applicants) themselves, but rather by a supplementary cast of patent agents, attorneys and examiners. Consequently, they argue citation analysis is a poor indicator that, for example, Engineer A learned fiom Engineer B no matter what their proximity may be. While I think that this contention requires testing, the general nature of their criticism suggests that patenting should be more thoroughly examined as practice. spillover studies in that they identify movements and linkages but do not actually demonstrate the process by which knowledge is displaced from one institution to another19. Finally, analysis of trade patterns has led to the proposition that technology transfer occurs through the movement of goods, including capital goods (Coe and Help man 1995). In the case of capital goods, and especially complex machinery, Gertler has repeatedly (1995,2001) called into question the limits to the transfer of these technological artifacts.

Given these criticisms, of what value is the 'texts, people and artifacts' typology for understanding processes of knowledge transfer? If, "by focusing on the medium through which interaction occurs rather than the nature of the process itself, this scheme is of limited use" (Gertler 2004: 354), then the task for research is to reveal the mechanisms that activate these channels in the formation of knowledge networks. A network ontology is useful here for it frames texts, people and artifacts as mutually constituting. Moreover, it directs attention at the agency, specifically the politics of persuasion, which lend these situations a resonance that extends beyond their local embeddedness. These concerns are addressed by referring to the science studies literature inspired by Bruno Latour.

2.4 Science in Action

The science studies approach popularized by Latour (1 987, 1999) offers a way forward in framing the geographies of knowledge transfers that are involved in technological change. More specifically it provides a highly nuanced interpretation of the way codified

19 Coe and Bunnell(2003) have added scope to our understanding of how labour migrations relate to knowledge exchanges. In particular they have drawn a linkage between the literatures on transnational and knowledge communities. and tacit knowledge interrelate in the innovative process. Science studies endeavors to shed light, "on the local, material, mundane sites where the sciences are practiced"

(Latour 1999: 309). More broadly, it develops a conceptual framework and associated vocabulary to interpret the circulation of scientists and engineers through society. In the last decade various authors have illustrated the utility of this perspective, which relates how embedded local practices connect with the networked space of flows, for the discipline of geography (Murdoch 1994, Bryson 2000, Barnes 2001,2004). Perhaps the commonest denominator in these accounts is the idea of 'translation'.

2.4.1 Translation

'Translation', according to Latour (1999: 88), "consists of combining two hitherto different interests . . . to form a single composite goal." Latour (Ibid: 3 11) employs the notion of 'translation' to discuss links in the chains of contingency, "through which actors modify, displace and translate their contradictory interests" in the formation of a network. Critical to this concept is the 'slippage' or 'drift' in meaning and objective that takes place as interests are displaced. As Latour (Ibid: 89) puts it, actors are forced to

"take detours through the goals of others". He writes of four interdependent circuits through which translations circulate in an actor-network (Figure 2.1).

The first loop concerns the 'instruments' actors use to 'mobilize the world'. For engineers and inventors, these instruments are the various inscriptions (circuit diagrams, textbooks, patents, etc.) used to 'map' their world by authorizing a territorial claim to technological space. Once codified, these objects take the form of 'immutable mobiles' that, because they maintain their form in different settings, are transportable and combinable and thereby allow for new translations. For example, a physics equation depicting voltage amplitude and periodicity could be applied to the musical scale to form a new relationship whereby one volt equals one octave in pitch. A subsequent 'digital' rendering of this 'instrument' makes a further translation by assigning a binary logic to the relationship.

The viability of these combinations is tested as the engineer seeks to enroll colleagues through the second circuit of 'autonomization'. This circuit strengthens in accordance with the degree to which a discipline defines its identity relative to a common problem. For example, an inventor organizes a session at the Audio Engineering Society

(AES) on the theme of 'Music and Electronics' and invites colleagues to participate. 3 Alliances

\ 5 ~inkband knots

4 2 Public Autonomization representation (colleagues)

Source: Reprinted with permission of the publisher from PANDORA'S HOPE: ESSAYS ON THE REALITY OF SCIENCE STUDIES by Bruno Latour, p. 100, Cambridge, Mass.: Harvard 1 University Press, Copyright O 1999 by the Mobilizations of the world President and Fellows of Harvard College. (instruments) Figure 2.1 The Circulatory System of Scientific Facts These participants seek to enroll more colleagues by publishing their contributions in the AES journal. In converting orallaural performances and accompanying manuscripts to published texts, editors displace, somewhat, the original intentions of the authors, while realizing new ones. This drift is likely acceptable since, a localized performance that took place in some hotel conference room is translated into an immutable mobile that can be disseminated broadly to subscribers around the world2'.

These inscriptions extend the loop by attracting new colleagues while deepening the autonomy of the fledgling connective.

In the third circuit, actors enlist the support, of 'allies' beyond their ken. "No instruments can be developed, no discipline can become autonomous, no new institution can be founded without the third loop", which Latour (1 999: 103) calls 'alliances'. For example an engineering section head must enroll the confidence of management to secure funding to develop an idea. The potential for a slippage in these translations is obvious2'.

For example, Katoh Tom, the president of Japanese synthesizer maker Korg, once weighed the fifteen technical features and $2000 (US) price projection for an advanced prototype instrument presented to him by his engineering department. Katoh's goal, on the other hand was to launch the first mass-produced synthesizer that would sell for under

$1500. "Katoh took a piece of chalk, and using it like a katana (a Japanese sword), slashed a line across the blackboard right after feature number seven. He hit the top part of the blackboard, which contained the first seven features and said, 'These you get.'

20 Important subscribers, for this author, include university and public libraries whose stores of these volumes enabled archival research. 2 1 I am reminded of one of the California scenes in Woody Allen's (1977) Annie Hall where the background conversation between script writers at a party of movers and shakers provides the memorable line, "They had a notion, which they wanted me to work into a concept, in the hope that it will eventually turn into an idea." Then he hit the bottom part with the remaining features and said, 'These you don't"' (Jim wightZ2as quoted in Anderton 1988: 56). The alliance forged between engineering and business interests authors its own instruments in the form of patents which codify technical developments in a proprietary fashion. These inscriptions enroll a further cast of allies comprised of patent attorneys, agents and examiners.

Latour's (1 999: 105-6) fourth circuit, 'public representation' furthers, "this massive socialization of novel objects", to the most critical realm. The three other loops utterly depend on translation in this public circuit. However, "information does not simplyflow from the three other loops to the fourth, it also makes up a lot of the presuppositions of scientists themselves about their objects of study." For example,

Japanese instrument maker Yamaha has long sought to enroll the public via its musical education programs. Today, Yamaha music schools are found all over the world.

Additionally, Yamaha trains dealers and retailers to properly demonstrate its products in music stores. In turn, engineers design functions for its beginner keyboard models to fit in with this educational philosophy. The instrument, in effect, trains the musician. More generally, Theberge (1997: Chapter 5) writes about the key role played by magazines

such as Keyboard and Computer Music, which are as much advertisements for the manufacturers as they are sources of information for consumers, in strengthening this public circuit.

These four circuits, instruments, colleagues, allies and the public are held together by a fifth circuit, which Latour calls '.knots and linkages'. Here, everything either comes together, or falls apart (Barnes 2004). Latour's (1 999: 106-107), "heart beating at the centre of a rich system of blood vessels" and "knot at the centre of a net" metaphors

22 An employee of Korg's US distributor, Unicord, who witnessed this episode.

59 capture how the function of this loop relates to its form. Science studies aims to explain the relationship between this fifth loop and the other four. Generally, the larger, more complex and interdependent these constituent circuits are, the larger and more robust are the knots that tie it together. This thesis endeavors to explain the chain of translations and circulation of science in music's own quantitative revolution - the transectorial adoption of digital electronics. It does so in a way that brings geography to the foreground. In crossing techno-economic paradigms the industrial trajectory of electronic musical instruments (EMI) has become pervasive and global in scale. However, increasingly it has also been centred in Hamamatsu, Japan, driven by the pulse of firms based there. To make this case, requires a 'local model' (Barnes 1996) of place and space that illustrates how these enterprises acted 'at a distance' (Latour 1987).

2.4.2 Topological Space

In drawing together these myriad sites of practice into a circulatory system, science studies presents a geographical vision that privileges topological space - a space of connections in which two far flung places may be far more tightly linked that neighboring places. Topological space is complex because networks overlap unevenly with regions.

Regions, both figuratively and literally, are principalities that possess their own internal

"topological rules about areal integrity and change" (Law 1999: 6). Yet regions themselves are 'constituted by networks' that are both internal and external. In this sense, argues Law (Ibid), "nation states are made by telephone systems, paperwork and geographical triangulation points." Conversely, to illustrate the process of exogenous constitution, we can imagine a product bearing the label 'Made in Japan' that is underwritten by fine print listing the relevant US, European and Japanese patent numbers that translate the meaning of the technology embodied in the good in terms of intellectual property regimes around the world. The manufacturer, in staking these claims, can 'act at a distance' to discourage the piracy of that technology.

To work through how these topological geographies unfold, in the following pages I highlight three instituted practices and one setting that proved important for the translation of digital logic into the musical instrument industry: patenting, tinkering, reverse engineering and trade shows. My intention is to conceptualize specific key practices relevant to understanding the endogenous and exogenous configurations of

Japanese innovation in networked space. The interrelations amongst these practices and situations resonate across geographical scales. For a hypothetical example, a Japanese engineer working on a technology that will be incorporated in a product sold in the US market, is likely to keep an accurate laboratory notebook that describes, in rigorous detail, all that takes place. The reason for this is that, since the US patent system works on the 'first-to-invent' principle, these notes may become critical evidence in the event of competing claims. In other words, the routines of the engineer reflect company policy which in turn must be attuned to their principal market's jurisdictional conventions governing intellectual property. This example, which views patenting as embedded practice, differs markedly from the knowledge spillover school in which patents are taken as given, ready-made science, and hence are recognized simply as proxies for innovation. Since much of my analysis examines the relations between patent statistics and patent practices, the following section elaborates on this theme. 2.4.3 Patents

Abraham Lincoln remarked that patents wed, "the fuel of interest to the fire of genius in the discovery and production of new things" (quoted in Basler 1953: 357). In other words, patents translate economic and scientific interests to produce an end, which happens to be a socially constructed version of novelty. Patent documents are particular types of inscriptions that codify a recipe for this novelty. Authored by inventors and assigned by institutions in what Latour (1 999) would refer to as an 'alliance', patent applications are rigorously examined by the state, and it is at this stage that the persuasiveness of their claims is tested. If the contribution is not obvious to a reasonably skilled person in the field and improves upon prior art - extant, 'authorized' inscriptions pertaining to a given field of practice - the application as described novelty is translated and re-inscribed as an intellectual property right. Patents' proprietary quality allows firms to deploy them as a means to mark off technological space from rivals, to the extent limited by the claims. However, this intellectual and technological space is highly congested, and though patents impose a degree of order to this unruliness, quite often patent claims partially overlap and any attempt to capitalize on proprietary knowledge will necessitate coalition or conflict between their holders. One solution is the cross- licensing of adjacent and ancillary claims held by others, a form of 'alliance' amongst competitors. This example illustrates another apparent property of patents as 'immutable mobiles' due to their transferability between parties. This claim will be critically evaluated shortly.

The ability to 'mobilize the world' in a patent is temporally a close-ended phenomenon - patent rights offer protection for a limited time. In the US this period is twenty years from the date of filing. Once sanctioned by the state, a patent does not necessarily allow its holder to use, make or sell a product derived from that knowledge.

Rather, as one of my informants pointed out, it merely entitles its owner to prevent other parties from using, making, or selling a product that intentionally or otherwise infringes on its claims. Patenting strategies are consequently regarded as being defensive as opposed to offensive in their orientation. Nevertheless, these strategies differ greatly between nations (Granstrand 1999), sectors23,and firms. Some companies patent every incremental improvement while others patent technologies they have no intention of using except as a deterrent to their competitors.

The quality of patents is highly varied, and this too affects the way they are used.

For example, some companies like to either fortify or obscure a particularly significant patent behind a wall or smokescreen of dependent, though insignificant improvement patents. Also, firms refer to their 'patent position' vis-a-vis competitors and this trope hints at the way they view the competitive environment. A different respondent in my study noted that patents are only as good as a firm's willingness to enforce them. This statement implies that firms have diverse commitments to enacting their rights from afar, which oftentimes involves enrolling allies to enforce these rights on their behalf.

Social scientists have long used patent statistics to discern patterns in innovation and economic development (Schrnookler 1966, Mensch 1979, VanDuijn 1983, Freeman et a1.1983). Indeed, much of the contention in long wave research hinges on competing interpretive frameworks of patent data. Most research in geography that uses patent data has been extensive in approach (for example Pred 1966, Ceh 2001). By comparison, until

23 I've heard that the figure for registered patents increased sharply following a decision by the USPTO to allow the patenting of software.

63 quite recently little was known about the social dimension of patenting. However, scholars working from a science studies perspective have begun to cast the problem in a different light that problematizes patenting as a socio-spatial practice (Winder 1995,

2001, deLaet 2000). deLaet (2000: 157; emphasis added) questions the 'immutability' of patents as 'mobilizations of the world'. Patents, by,

tying, subjects, objects and knowledge to one another in a particular but immutable way.. .create an order of property.. .A patent remains the same, no matter what its destination or its vehicle; patent documents are made to be tenacious and dependable representations both of rights and of knowledge, and patents perform with tenacity as they are part of a system that builds a proprietous order of things. This is all true. That is, it is true as long as the patent does not travel.

Patents are problematic because, though they are designed to 'act at a distance', there is an inevitable slippage once they are mobilized in various contexts. This slippage is produced because the social concepts of ownership and authorship are less mobile than the inscription itself. This contingency of place creates the potential for patents to

'unravel' when they travel (ibid: 163).

Another consideration is how the knowledge 'contained' in patents relates to other inscriptions, for instance papers published in engineering society journals. Though theorists have long recognized the distinction between patents and academic journals as being important (Machlup 1962, Freeman 1982), these lessons need to be re-interrogated, for they lend insight into the role of codified knowledge during episodes of technological change24. For instance, the logics of citation for patents versus other forms of codified knowledge are fundamentally different and this holds the two in tension. As Latour

24 These inscriptions in Latour's parlance are also 'nonhumans'. "This concept has meaning only in the difference between the pair "human-nonhuman" (Latour 1999: 308). The obvious implication is that there is a social mediation between these instruments. (1 987: 3 1) points out, the purpose of citation in scientific journals is to marshal authority to your argument in order to deter dissent - paraphrasing: 'if you want to refute me, you must also refute Jones et al. 1976 and so forth.' Doe1 and Rees (2003) further note that citation of prior art in patents is not usually undertaken by the inventor, but in more likely scenarios by the patent department of the firm or examiners. In occasional situations prior art searches are initiated by firms seeking re-e~amination~~.The motivations of these disparate interests are in conflict, since, for example, an examiner finding prior art, reduces the breadth of the claims that the invention covers. There must be a 'translation' to bring these interests into alignment. The examiner might say something like, 'claims one through five are valid, while claims six through eight have been covered previously.

Please revise the scope of your claims by referencing the appropriate literature.'

The relationship between patents and textbooks/technical reports is further held in tension in situations where jurisdictional conventions privilege the interests of those who are in a position to patent. An example that will be explored more fully in chapter seven concerns the domestic patent system in Japan that was in place until the early 1970s. A key feature of this system that has since been amended allowed large firms importing foreign products to patent the embodied technological knowledge 'as is' (Kakehashi

2002). This situation forced smaller firms to, for instance, publish ad-hoc textbooks in order to establish a claim to domestic prior art; to mobilize the world in a particularly public, as opposed to proprietary, fashion.

Having illustrated the various ways that science studies can be used to interrogate the processes of knowledge codification in the form of patents, I next consider the

25 Re-examinations are initiated at the behest of rival fmsor individuals who, if additional prior art can be found to refute the claims of the patent, seek to void its novelty. avenues by which these perspectives possess a similar utility for linking these insights with our understanding of the tacit dimension. To accomplish this, I consider two particular practices that have been important for the Japanese electronics industry in the cultivation of tacit knowledge through the reworking of foreign codified knowledge: tinkering and reverse engineering.

2.4.4 Tinkering

Immediately following the Second World War, Japanese industry found itself in a situation of acute material and financial scarcity, and the most significant problem was that much of its fixed capital needed to be rebuilt. Consumers were similarly affected and also had to make due with what they had or could scavenge on the black market.

Meanwhile the US office of general headquarters (GHQ) - the occupying administration

- regulated the rebuilding process and had as its broadest mandate the construction of a set of institutions designed to promote political democracy. One of the first policies it enacted in this regard was to lift the ban on the reception of foreign broadcasts while encouraging the production of radio receivers and associated components. At the same time, there were a large number of Japanese military technicians trained in the science of radio who took their skills to the civilian sector. An accompanying flood of inexpensive surplus radio components from the Japanese military found their way to urban 'flea markets'. One such market was Tokyo's Akihabara Electric Town, which to this day is the largest specialized retail district in the world for both low cost and leading edge electronic technologies, most of which are made by Japanese companies (Takahashi

2000). Out of this constellation of factors, argues Takahashi, emerged a culture of tinkering - defined as the repair of old damaged sets and the combination of spare components into working sets - which formed the basis for the unofficial sector of the

Japanese radio and electronics industry. Takahashi presents figures showing that until

1950, the ratio of unofficial to official radio production was almost three to one. The cost of these sets was frequently less than half that charged by the large firms which were still devoting most of their energies to rebuilding. Central to this boom in radio production was people's increasing technological literacy or familiarity with and understanding of electrical components. Amateur technical journals played a key role in promoting this literacy.

People living near Tokyo visited Akihabara frequently to buy components, and young people, carrying radio magazines for beginners, crowded there. In Osaka, the Nipponbashi [Den-Den Town] electric town played a similar role (Takahashi 2000: 468).

This quotation connotes the formation of an informal, if nonetheless effective, community of practice. It also illustrates the knot forged between the circuit of

'autonomization' and that of 'allies'. Indeed, the network of tinkerers defined itself through the consumption and use of these magazines - a position that necessitated a translation at some point that enlisted the support of magazine publishers. Further, the journals themselves take on the form of immutable mobiles that can be carried to the

'electric towns' to direct purchases and back home to direct the cultivation of practical tacit knowledge via tinkering

Takahashi also shows that a similar integrated 'circuit', that hinged on a relationship between amateur tinkerers, component manufacturers and hobby electronics journals, fueled a subsequent television boom in the 1960s. Later, even as the unofficial electronics sectors sector faded away in the 196O's, this culture of tinkering, "had a profound influence on the success of the Japanese electronics industry in export markets"

(ibid: 461). Many of these amateurs used their own stock of tacit knowledge as a source of human capital to gain entrance to engineering programs at universities, to gain employment at the emerging consumer electronics firms which are global household names today and even to found their own firms. In other words, the tacit knowledge developed through tinkering was readily transferred to other contexts. This process is less a 'diffusion of the engineering disciplines', than it is a translation of an amateur network into the corporate realm, a reformulation of a community of practice into institutionally differentiated constellations driven not by touchstones like hobby magazines, but by the interests of enterprise.

2.4.5 Reverse Engineering

The acquisition, effective adaptation, and improvement of technologies from abroad by Japanese industry served as the basis for Japan's rapid economic growth and international competitiveness in a wide variety of manufacturing industries ( [US] National Research Council 1997: 10).

Underlying this sequence of processes in Japan is the practice of reverse engineering, which involves the disassembly of competitors' products with a view to understanding and improving on the technology. In its narrowest form, reverse engineering consists of attempts to manufacture a product that is, "similar to one already available on the world market but without [hosting] direct foreign investment or [the] transfer of blueprints for product and process design (Freeman 1987:40). Moreover, it is a practice often associated with Japanese firms and comprises part of a larger set of scanning techniques that include analyzing patents and technical reports, attending conferences, visiting competitor's factories, hiring foreign experts, and so on. By broadening the definition of reverse engineering to include this entire suite of practices it is possible to view the process as an institutionalized version of tinkering that translates not individual products, but entire industry models and their underlying knowledge bases.

The Japanese verb, shuuttoku g@- which combines the kanji (Chinese character) for learning with the kanji for getting %#captures the essence of these instituted processes - 'learning by getting'26. The careful scrutiny paid to American cars, appliances and electronic components in the 1960's by Japanese engineers built up the stock of tacit knowledge that enabled the firms employing them to best their US rivals by the 1980s. It is a process with a long history and it is a practice that still continues to this day27.Reverse engineering as 'learning by getting', involves the collection of instruments

- 'mobilizations of the world' - from afar through 'cycles of accumulation' which are the iterative connections centres form with peripheries to enhance their conceptual and practical understanding of the latter (Latour 1987: 219).

Hayter (1997: 257-8) vividly illustrates this process via a discussion of the

Japanese chassis maker and Honda-supplier, F-Tech. In one room of their factory, the chassis of all their competitors from around the world are arranged on the walls. To paraphrase Latour's (1987: 224) paraphrasing of Kant, instead of the engineers' minds at

F-Tech revolving around the chasses of their competitors, through collection and arrangement the chasses are made to revolve around the engineers, literally. The walls of

F-Tech's 'focal' factory (Fruin 1997) bring to mind the collections of flora and fauna

2"n response to my research presentation entitled 'Gijyutsuhenka no katoki ni okeru gakushuu to kakushinn wo meguru chirigakuteki shitenn - Nihun no denshi kenbunn gakki sangyou nit suite no jirei given in Japanese during a weekly graduate seminar in the Department of Geography, Kyoto University, Professor Kinda Akihiro proposed shuutoku as a correction to my use of the verb gakushuu (?%) which connotes study, in an educational-institutional sense. Note that the second kanji in this word is the same one that begins shuutoku. '' Yamaha's founder Torakusu Yaniaha, a traveling medical equipment repairman, started his organ company in 1897 after repairing the imported organ in Hamamatsu's elementary school (Ohtsuka 1980). brought by explorers back to 'centres of calculation' like the Royal Society (Gregory

1994). An engineer at Yamaha told me that in the early 1960s' when engineers entered the then embryonic electronic organ division, they would spend the better part of two years taking apart American-made organs28.Synthesizer manufacturer Roland's former head of R&D flatly informed me that, "We always examine competitors products."29

Both of these musical instrument makers are engaged in roughly the same process. That they are also neighbors in Harnamatsu suggests that this city functions as a kind of

'centre of calculation', a core region which the global musical instrument industry is

made to revolve around.

Freeman (1987) suggests that technological learning via reverse engineering

produced a number of consequences for Japanese industry. Foremost among these was a tendency by engineers, managers and workers to view the production and innovation processes as being systematically integrated. It also produced the notion of using the

factory as a laboratory - 'a focal factory'. As can be seen from these examples, reverse

engineering literally opens-up technology's 'black box', and indeed disassembles it,

thereby challenging its 'ready-made' status. Reverse engineering further empowers

Japanese firms to 'act at a distance.' The difference between reverse engineering and

tinkering is more than one of degree. Reverse engineering, according to the conception of

translation outlined earlier, involves the enrolment of 'allies', for it takes the

orchestration of firm strategy to systematize the translation of foreign instruments into the

technological sphere of Japanese firms.

Nagahama, Yasuo, Professional AudioIDigital Musical Instruments Division, Section Head of Electone Division Personal interview, Hamamatsu, July 2002. 29 Kikumoto, Tadao, Senior Managing Director and Former Head of R&D, e-mail correspondence, October 2002. 2.4.6 Trade Shows

Beyond the laboratories and workshops dedicated to tinkering, reverse engineering and patenting, other more temporary sites ground the translation of technical knowledge. The specific example I wish to address here is trade shows, which I define broadly to incorporate both professional society conferences as well as conventions where manufacturers engage with distributors and retailers. The former facilitate the

'autonomization' of colleagues, the latter operate in the circuit of the Both of these events are carefully choreographed demonstrations of the state of the art.

Professional society conferences gather spatially disparate communities of practice into contact for a brief time. Barnes (2003) comments on the role these occasions and spaces played in the maintenance of ties in the network that wadis regional science. These sites, from convention centres to hotel bars, provide a forum for a range of practices. Foremost is the presentation of research, new ideas and reflections on the state of the field for public peer review. These events also enable the formation of coalitions to address common concerns, establish and regulate standards and serve as a market for job-hunting and head-hunting. Through conventions, a community of practice forges its identity by defining its autonomy.

In chapter three I consider the role of the Audio Engineering Society (AES), which especially during the late 1960s and early 1970s served as a crucible for the

'translation' of knowledge and 'diffusion of the engineering disciplines' from the electronic engineering and physics fields into the nascent field of electronic music. The

Meetings of the AES were remarkable for their diverse composition which was reflected

30 Even if this public circuit is one translation removed from the final consumer. The general public is excluded from attending annual National Association of Music Merchants (NAMM) Convention.

7 1 not only in the various sectors employing audio engineers - the military, hi-fi makers, musical instrument manufacturers -, but also for the diverse professions juxtaposed as presenters even in narrowly themed sessions: composers, educators, researchers, entrepreneurs and engineers employed by large firms all had something to say in some of the early sessions on electronic music. Institutional settings shape tacit knowledge, and so the reception apaper received in these sessions depended on the degree of overlap in underlying tacit knowledge between participants; a generally stable situation that was tested by radical ideas.

2.4.7 Interlinking Practices

Patenting and other means of inscribing knowledge, tinkering and reverse engineering are social practices that are institutionally embedded. Conversely, conventions such as the

AES are institutions governed by a suite of social practices. I suggest that in concert, these specific practices, and the locales in which they take place constitute the critical nodes in the formation of knowledge networks; links in a chain of translations; knots in the circulatory system of technological knowledge. Moreover, in a conceptual sense, I raise these practices because they animate, in a practical sense, the three conduits of knowledge transfer discussed earlier. In cases such as reverse engineering, it is the interaction between texts, people and artefacts that permits the cycling of tacit and codified knowledge and thereby the transfer of knowledge in space.

With this perspective in mind, we can return to the debate between Amin and

Cohendet (1 999) and Gertler (2001) over the efficacy with which organizational or relational proximities can be said to overcome distance. National institutional spaces, the scale that I think Gertler is particularly interested in (Gertler 1999, do indeed spawn national styles of industrial learning or national innovation systems (Freeman 1987).

These systems are firmly anchored in their regions of origin and there are powerful limits to convergence between these industrial models (Gertler 2001). However, specific national innovation systems such as Japan operate extremely well in the realm of topological space by producing constellations of practice in 'cores' that, through a 'cycle of accumulation', can develop a rich network of connections that enable them to 'act at a distance'.

This last point is illustrative of my broader conceptual aim which is to suggest ways in which science studies complement efforts by economic geographers to understand processes of knowledge transfer, especially those that are prone to arise at the boundaries between TEP. A science studies perspective bolsters this research trajectory by pointing out how industrial geographies are produced through practice. Practice is driven by the translation of inscriptions - the mobilization of the world. It is also shaped by the persuasive capacity of actors to enrol new actors into a network in which the conduits of knowledge transfer (people, texts artefacts) are relationally constituted.

2.5 Transectorial Innovation Systems, Engineering Biographies and the Formation of Core Regions as 'Meeting Places'

This brief, final part of the chapter draws evolutionary economics and science studies together by underscoring the contingent nature of what Rosenberg (2000) called the

'diffusion of the engineering disciplines'. The contingencies that I wish to highlight are first, the lives of engineers and inventors and second, the way these lives take on meaning via their relationship to industrial cores. Transectorial innovation necessitates disciplinary boundary crossing. This

'diffusion', however, is far from an even process. What Hughes (1992) terms

'technological salientsY3l,advance, hold firm and retreat in particular places and times and these outcomes clearly turn on social processes. A science studies perspective is not particularly interested in denoting these cases as successes or failures. Rather, it is concerned with understanding why particular connections in networks become strong or weak32.The identification and interpretation of these connections poses a practical concern for researchers interested in 'following [engineers] through society'. The temptation is to latch onto key accomplishments (or failures), or the 'Eureka!' moments where the genius of the engineer comes to the fore. However, even with the benefit of hindsight, these instances are merely Thomas Edison's 'one percent inspiration' as opposed to the 'ninety-nine percent perspiration' that goes into invention. Conversely, it is the 'traces', the 'chains of translations' that precede and follow these moments that need to be addressed. Only as a last resort should we look to 'special cognitive abilities'

(Latour 1987: 247) to explain technological 'progress'. As engineers circulate in networks of association with instruments, colleagues, allies and the public, they leave traces which are the accumulation of all those moments of slippage whedwhere heterogeneous interests are translated.

Bringing these 'traces' into relief is still problematic. Latour, after all, was afforded the opportunity to develop his insights as a participant-observer 'embedded'33 in

31 Salient comes from the Latin salire, meaning leap. 32 This perspective can inform our understanding of path dependent phenomena (David 2000), such as the QWERTY keyboard, which, though not the most eficient solution, nevertheless had the strongest connections when it mattered most, at the very beginning. 33 The highly contrived practice of 'embedding' reporters with the invading US Army in Iraq has, unfortunately, given this term a whole new set of connotations. the everyday world of science at the Salk Institute. His vantage point differs markedly from my own, for the only way I can follow musical engineers through society is to sift through patent archives, industry histories and engineering biographies, in other words, texts, authored by the observed themselves. As Barnes (200 1: 426) cautions, "if science studies are to use biographical accounts then their nature must be recognized."

Schoenberger (2000) and Barnes (2001) draw attention to the theorization of biography in economic geography. Both these authors, for instance, refer to the narrative tropes

(heroic, marital, filial) that colour these stories. Similarly, the construction of longitudinal profiles (Laulajainen 1982) from secondary sources like patent databases, annual reports and industry histories must also reflect on the supposedly 'inert' facts they contain.

The way forward is to set biographies against longitudinal profiles in order to evaluate what engineers accomplished versus what they have to say about these accomplishments. This juxtaposition looks critically at technological change in general and transectorial innovation in particular as contingent and contested associations between engineers and their inscriptions, their colleagues, their allies and the public. In this sense, 'the instruments' that gain recognition as radical innovations certainly make a name for their inventor as a 'hero' who overcomes a particularly significant threshold.

However, they also bear traces of all the translations that this very success obscures.

These slippages animate the lives of engineers. To put radically innovative products in perspective involves tracing, amongst other things: the career displacements and transectorial migrations of engineers who, eventually, arrived at 'The Idea'; the inscriptions that they accumulated and recombined along the way; the connections with colleagues and; the allies that either inspired or constrained their career, and so on. At this stage it is possible to re-insert the TEP model, which interprets long run patterns of technological change according to the succession of key technologies, leading industries and most important 'core' industrial regions. To make sense of these shifts as contingent and not deterministic involves highlighting the way that the lives of engineers relate to these cores. Science studies offers a way to through the connections between cores and peripheries and vice versa. Its view of topological space is further useful in making sense of the translation of interests between two cores and the ascendancy of one core over its predecessor. All that is required is to follow engineers through the networks that connect the two. Thus one can distinguish between engineers working from cores, versus those that work towards cores. In the empirical part of the thesis, I split the narrative between analysis of Japanese engineers who, acting at a distance, mobilized the technological world so that it revolved around firms based in Hamamatsu and, analysis of

American engineers whose interests were translated to the advantage of Hamamatsu.

From these quite different though interdependent vantage points I am in a position to highlight the twin features of Oinas and Malecki's (2002: 103) Spatial Innovation System

(SIS) framework, namely, "the external relations of actors and the variability of the relative weights of different places or regions as centre points of particular points in time."

Corporations proscribe limits to the mobility of engineers. Firms, as institutions, profoundly shape the everyday lives of these individuals through socialization. Moreover, through specific practices such as patenting, the 'alliance' between the engineer and the firm can become crucial for both the strategy of the former and the career of the latter.

This example raises two points. First, the patenting practices of engineers and inventors, when translated through the interests of firms, have profound implications for industrial dynamics. For instance, firms that assign highly significant 'keystone' patents can alter the configuration of topological space and hence industrial advantage, by shifting the centre of the network. The momentum they gain from these proprietary inscriptions modifies the hitherto flexible constitution of a community of practice because these inscriptions lock-in advantage in one particular constellation of interest while simultaneously locking-out rival constellations from whole swaths of technological space.

In this sense, the advantage of these 'cores' derives from their connectivity and indispensability to the maintenance of innovation systems.

The second point I wish to make concerns how the career movements of engineers reflect choices: 'With whom do I want to work?', 'How can I best guard my interests?' and so on. Career migrations are problematic with respect to patents, because, for instance, engineers who patent inventions on behalf of one firm can transport their tacit knowledge to a competitor in ways that challenge the ability of the first firm to act at a distance. In other words career migrations, as they relate to specific inscriptions, also have implications for regional development.

2.5.1 A Local Model of Transectorial Innovation Across Space

To conclude I present a transectorial model of innovation over space that is simultaneously a 'local model' (Barnes 1996) This model is partly inspired by Freeman's

(1 982,2000) view of industrial evolution. However, its content is structured around the theme of transectorial innovation and spatially with reference to the experience of

Japanese industrialization. This specific core is interpreted as a meeting place shaped by exogenous and endogenous processes. It presents an industrial trajectory that begins in a few centres but becomes increasingly centred on one region, Japan. Though it presents this evolution as a sequence, this is not a linear teleological model. Indeed, the numerous feedbacks between the various stages will become obvious once the empirical details are filled in. Nevertheless, the sequence is one of momentum.

The evolution of transectorial spatial innovation systems is a multi-stage process in which different scales of activity in topological space govern the process of knowledge formation and transfer (Table 2.2). The genesis of these systems occurs within localized communities of practice among a set of actors whose inspiration derives from their training in the carrier sector. Piatier (1988) called these subjects transectorial migrants.

Their experience and familiarity with the key technology and its underlying principles is a necessary but not sufficient condition for enrolling new actors into the network.

Consequently, their principal contribution is realized through inscriptions which translate knowledge in a way that can be more readily mobilized. Precisely because transectorial migrants are likely to move between enterprises as they cross sectors, firms do not play a significant role early on in structuring these systems.

Transregional inspiration captures the suite of practices which accomplish the reverse engineering of foreign technologies. Travel, especially factory visits, the translation of texts and tinkering are the principal practices that mobilize and re-embed knowledge. At the broadest scale these activities are the product of national innovation systems. Formal industrial policy sets overt directions for research activity by, for instance, emphasizing key sectors and funding overseas travel (Freeman 1987, Partner

1999). Another important way in which industrial policy structures the knowledge economy is through patenting regimes which may be biased to the interests of large firms. Beyond these formal mechanisms a social infrastructure governs the styles of practice. Takahashi's (2001) case study of the post-war network of tinkerers in Japan exemplifies how technological solutions are embedded in historical, social, and economic worlds. His study also suggests that, in the case of Japan at least, communities of amateurs are able for a time to persist in parallel with firms.

Whether the inspiration is transectorial or transregional, at their genesis, the structure of spatial innovation systems is not greatly affected by economy-of-scale effects. Indeed, a phase of entrepreneurial and intrapreneurial invention persists as long as barriers to entry remain low for the swarm of actors from diverse circumstances that flood into the nascent sector. Entrepreneurs in this case conform to Schumpeter's 'Mark

I' type of innovators and possess a dual agency as both inventors and founderslmanagers of small firms. Intrapreneurs are those inventors working within larger enterprises whose job is to implement the technological strategies of these firms. Depending on the firm, these subjects have varying degrees of autonomy for devising strategy. Beyond these two categories of agents are a class of independent inventors who work outside a corporate rable 2.2 The Evolution of a Transectorial patial Innovation System - The case of an industrial shift from the US to Japan Phase in the life Conduits/processes of Nature of knowledge formation and transfer (in Geography cycle of a knowledge formation and terms of tacitlcodified knowledge cycling) transectorial transfer technology system Transectorial Boundary-bridging texts 1. Translation via re-situation of codified knowledge Highly localized community of practice Inspiration : published enables formation of general disciplinary-level tacit form according to disciplinary protocols Formative Training in basic science knowledge. Impetus to shifting 'engineering common and institutions exogenous sense' Occupational migration National innovation systems set overt linkages 2. Formation of sectorally specific tacit knowledge as directions for research activity (i.e. Tinkering established through codified knowledge is worked out in new sectoral military industrial goals) and also transfer of basic setting govern style of practice in facilitative techno-scientific and prohibitive manners. knowledge 3. Pervasive applicability of key technology and underlying basic science allows for labour mobility across permeable sectoral boundaries Transreaional 1. Overseas factory, 1. Kengaku (learning be seeing)lShisatsuryoko Multi-scalar: Internationally, Inspiration Reverse laboratory and conference (inspection travel) constellations of practice are brought engineering (for visits 2. Translation via re-situation of foreign codified and into alignment. Nationally, formation of example from US 2, Import andlor translation embodied knowledge (N.B. subject to time lag tinkering culture. Locally, firm-specific to Japan) of texts and products imposed by cultural distance). resources govern tinkering style. Tinkering 3. Cultivation of thorough tacit understanding of state of National innovation systems set overt the art. Deconstruction permits insight towards path- directions for research activity (select breaking avenues of development sectors, fund travel) and also govern style of practice Entrepreneurial and 1. Small-scale 1. Embodiment of tacit and codified knowledge in Deepening and widening of regional Intrapreneurial demonstrations of early products as inventors seek to persuade management constellations of practice. Swarming of Invention prototypes. to support R&D trajectory actors from diverse circumstances 2. Publication of patents 2. Proprietary codified knowledge is transferred to public (withinlout firms, withinlout sector). embodying basic techno- domain National patenting regimes govern scientific knowledge. 3. Tacit knowledge is refined at frontier between definition of novelty; establish 3. Early adopters in the producer-consumer communities parameters for its appropriability. avant-garde. Limited Highly localized coalitions between commercialization but engineering and artistic communities of legitimization of practice not governed by business logic community of practice Phase in the life Conduits/processes of Nature of knowledge formation and transfer (in Geography cycle of a knowledge formation and terms of tacitlcodified knowledge cycling) transectorial transfer technology system Technical 1. Decisive demonstrations 1. Development of sectorally-specific radical 1. Intuitively, physical and cultural Crvstallization of technical feasibility technologies. Firm-specific tacit knowledge and vision proximity are expected to facilitate Formation and 2. Publication and shapes attitude of reception to these keystone positive reception to radical gestation of 'best appropriation of key technologies. technologies. (In practice the opposite practice' patents 2. This is the key instance when the codification of may occur making this a process that crosses a number of scales). 3. Commitment to adopt knowledge shapes the technological trajectory in a carrier technology highly proprietary manner 2. Technological space is defined in a manner that decisively favours one 4. Potential for highly- 3. Management must allocate significant resources to regional innovation system over selective migration of key the cultivation of firm-specific tacit knowledge for the another. inventors if there is a long-run mismatch with their 4. Tacit knowledge of key inventors follows transfer of 3. Deepening and widening of regional constellation of their inscribed knowledge locallregional production system and practice division of labour. 4. Trans-regional migration of key individuals - Commercial 1. Decisive demonstrations 1. Radical innovation sets the benchmark for rivals as a A single firm and its production system crvstallization of commercial feasibility bond is established between a single firm and an enjoy a phase of super-profits as as landmark product entire community of consumers. The growth of the landmark product spawns pervasive proves worth in the market market is fueled by rapid widening and deepening of impacts. tacit knowledge in artistic community as they prove the technology. Phase in the life Conduits/processes of Nature of knowledge formation and transfer (in Geography cycle of a knowledge formation and terms of tacitlcodified knowledge cycling) transectorial transfer technology system Explosive take-off Benchmark firm seeks to Benchmark firm demonstrates it boundary straddling 'A rising tide lifts all the boats' permitting and growth of establish feedback links to knowledge to carrier sector and forges links with a brief period when locational sector carrier sector institutions at the vanguard of basic research. advantages appear to be insignificant. Secondary swarming of New entrants bring their own tacit set of skills into In actuality, highly localized abilities to firms into expanding sector. effect internal and external economies of scale are selecting winners and market 'Follow the leader' process is constrained by patent losers. Attempts to position of benchmark firm. Some firms challenge emulatelunder-cut rival by testing its standing in technological space. Response to rival's patent positions necessitates search strategies that advantage of benchmark Codification of a common technological protocol (e.g. cross a number of scales firm MIDI). Coalition between rivals to Protocols define common technological establish product/process frame which brings various regional standards markets into alignment Consolidation via 1. Successful firms stalk 1. Broader-based processes of head-huntinglhiring of Regional production systems assert horizontal their rivals in order to key talent, as individuals are selected for their tacit their dominance - by integration absorb skilled labour and knowledge (which includes their ability to interpret wideningldeepening links with regional non-human codified knowledge contained in patents). constellations of practice that remain assets/infrastructure following dissolution of firms. Maturitv- 1. Expirylextension of key 1. Temporal limits to the proprietization of technological Innovation systems are rooted in region Reiuvenation patent positions space of control. Local rivals match spatial 2. Development of second, 2. Lead firms, losing their sectoral position of strategies through international third generation of dominance in technological space, cultivate ongoing divisions of labour. technologies feedback linkages to carrier sector Advantage persists through duration of techno-economic paradigm context, in situations as diverse as university labs or garages. The fruits of their labour may, however, be translated into the industrial realm through contract work. Regardless of whether innovation is of the Mark I, Mark I1 or independent variety, the pace of technological development is limited by developments outside the sector, specifically in advances in the key technology. For example, in the ICT paradigm, successive generations of developments in micro-processing power and memory defined the frontier of computational potential in sectors employing these fundamental technologies.

Entrepreneurs and intrapreneurs play a key role as gatekeepers who translate these advances to their own sectoral circumstances. To accomplish this they must first maintain a familiarity with these exogenous developments. Secondly, they must enlist a supporting cast of allies into the network by convincing these superiors or backers through small- scale demonstrations that hint at the potential of transectorial impacts. Independent inventors rely on the support of institutions such as patent offices to uphold claims to novelty.

At this stage the gulf between the practice of invention and the practicality of the inventions is still broad. Solutions are still being worked out in a context where advantage is dispersed. One implication is that in the patent arena, control of technological space is similarly diffuse in character. Neither entrepreneurs nor large firms enjoy an advantage. Meanwhile at the frontier with the market (the 'public' circuit), a small set of first users in the avant-garde, governed more by artistic logic that business logic, does little to produce an economic return for the technologies they endorse through use. Nonetheless, these actors lend a powerful legitimacy to the community of practice that weds technological and artistic spheres. Technical crystallization consists of those decisive demonstrations of technological feasibility that point to a best practice. Entrenched firm strategies and structures condition responses to these demonstrations. Rejection is a far more likely outcome than acceptance because the technical merits of the inventions are fully apparent only in hindsight - in Latour's terms, once the technology is 'black-boxed'. Nevertheless some firms have both the prescience to comprehend the value of the technology and the resources to devote to its development. This commitment locks-in advantage for that firm. At a tactical level, the enterprise seeks to internalize these technologies - remove them fiom the public domain - through various patenting practices. In particular, the exclusive licensing of a technology strands competitors at sea as the licensee secures a beach head to the massive swath of technological space opened up by the patent.

Commercial crystallization follows a lengthy phase of product development. At this stage the previous hunch to commit to the technology finally pays-off. Prior to this moment, the market has been slowing evolving in a manner that begins to favour economies of scale. However, until cost constraints and technical limitations are solved by the radical innovation, competing firm strategies reach towards the new paradigm but do not cross the threshold. The innovation proves sectorally radical by opening-up the market to a broad class of general users.

In the next stage, take-off, the success of the radical innovation produces a secondary phase of swarming by competitors who attempt to capture a share of the market opened by the benchmark firm. Against this onslaught, the benchmark firm engages in various tactics to protect its advantage, for instance by systematically publishing a battery of ancillary and dependent patents to colonize the technological space around the keystone technology. This tactic lengthens the lead time over rivals and forces the competing firms to navigate technological space by working around the patents; an expensive, time-consuming but necessary translation. By commanding the high ground, the benchmark enterprise is in a favorable position to exploit its advantage by deepening ties within the topological space of its innovation system. It may, for instance, forge relations with institutes at the vanguard of basic research who fund it with upcoming technologies. Economies of scale in innovation and production govern the landscape and favour those firms who can also gain a foothold in the rapidly growing market opened up by the radical innovation. This second wave of firms faces much higher barriers to entry than those that came into the sector during the phase of entrepreneurial and intrapreneurial invention. Entrepreneurial firm formation is likely to be limited. On the other hand, large firms from outside the sector can gain entry if they are able to develop complementarities that translate existing economies of scale into economies of scope via diversification (e.g. Casio). On the surface, a rapidly growing market enables marginal firms to survive. However, differences between broader regional patterns of rivalry and cooperation underlie the subsequent geographic shift in the industry's core that occurs during the phase of consolidation. An important precursor to the concentration in industrial structure is the establishment of a technological protoco!.

In the ICT paradigm, serial interface agreements such the internet protocol, reflect a confluence of technological and economic logics as firms realize that greater connectivity between their products fosters the growth of a market in which consumers are confidant that brand X can work with brand Y. Technological trajectories inevitably mature. As the sectors' rate of growth subsides to the level of normal returns, the weaknesses of marginal firms become apparent. Firms unable to maintain minimum efficient economies of scale are likely to fail. The financial assets of these firms may not be worth much on paper; however the tacit knowledge embodied in people, and codified in technical properties makes them attractive targets to larger firms. The latter seek to enhance their market share, and if possible absorb the talent employed by these 'sinking ships'. Transregional migrations and other substantial interregional connections in the network are likely to increase.

In the final phase of the life cycle, maturity/rejuvenation, firms in 'core' regions match spatial strategies through international divisions of labour. Basic research is performed close to home while applied research is conducted in various market regions.

Similarly, the most advanced production takes place at home in so-called focal factories

(Freeman 1987, Fruin 1997) while routine production conforms to the logic of Vernon's

(1966) product cycle model, taking place in low cost regions. Technically, feedbacks to the carrier sector extend the trajectory as transectorial impacts come full circle.

The sequence I have described is designed as an illustrative rather than exhaustive

'local model' of technological change used to interpret the ascendancy of one, particular

Japanese industry that happens to be spatially concentrated. I recognize that this sequence could be presented quite differently. This particular transectorial innovation systems challenges representation because it crosses not only continents and oceans but also techno-economic paradigms. It is an evolutionary process that unfolds through chains of contingency that are at times thoroughly embedded within enterprises, while at other times, impossible to contain within the firm due to the shifting opportunities and constraints imposed on individual talent. In these latter situations, the permeable boundaries of the enterprise facilitate situations where tacit knowledge is unmoored and mobilized to complement transfers of codified knowledge in ways that can profoundly affect the balance of regional industrial advantage. CHAPTER THREE: THE RISE AND FALL OF THE ELECTRONIC MUSICAL INSTRUMENT INDUSTRY IN THE USA FROM 1890-1980

3.1 Introduction

The previous chapter discussed transectorial innovation as an evolutionary and social process that unfolds as a 'diffusion of the engineering disciplines' (Rosenberg 2000).

From a neo-Schumpeterian perspective, this 'diffusion' from one disciplinary or industrial context to another, is uneven and punctuated by the, "bunching of technically related fmilies of innovations and inventions" (Clark et al. 1983: 76). Accompanied by the social and institutional momentum of techno-economic paradigms, transectorial innovations, like the cell phone discussed in chapter one, become pervasive in effect

(Freeman and Perez 1988). In turn, a science studies perspective endeavours to frame this process 'in action', as it unfolds through networks of relationally constituted actors

(Latour 1987, 1999). These networks feature certain centres where the connections between, or the translations of, heterogeneous interests are particularly strong (Latour

1987). Centres, which at one level could be seen as firms and, at a broader more collective scale, as 'core' industrial regions, anchor the linkages between instruments, colleagues, allies, and the public (Figure 2.1). In this conception the viability of firms and regions hinges on their capacity to maintain these networks, both locally and 'at a distance'. Through much of the past century, the USA dominated the world of electronic music (Chadabe 1997, Theberge 1997, Pinch and Trocco 2002). America was not so much the exclusive source of innovation in EMI, for key inventions were also developed by Russians and Germans, some of whom immigrated to the US. Rather, America was the polynucleated-centre of a much larger network of engineers and institutions whose interests combined to propel transectorial technological developments in musical instrument design. At the turn of the 2othcentury, Cahill's telharmonium was the first and perhaps most ambitious, if ill-starred, electronic musical instrument (EMI) project in the world (see chapter one). Had Cahill been able to enlist more subscribers, had he better aligned his interests with the New York phone company to eliminate interference, and had the instrument itself not been an immutable immobile, the network might have held together to realize Cahill's dreams to 'democratize' music (Weidenaar 1995). Half a century later, Hammond organs realized the mass commercial potential envisioned by

Cahill. Based in post-prohibition Chicago, Hammond succeeded commercially because the firm's organ's sold well to the church and jazz worlds. Moreover, unlike the telharrnonium, the Hammond was mobile. This feature made it the choice of the US

Army Chaplaincy during the Second World War (Vail 1997).

The chapter focuses on the period from 1950 to 1980 which witnessed a chain of developments that further mobilized the world around a constellation of key US institutions. Something akin to a 'quantitative revolution' occurred in the way that music was imagined, and this shift in engineering common sense brought instrument design into line with the constellation of forces represented by the ICT techno-economic paradigm. Latour (1987:224) would call this episode a Copernican revolution, for the scientist no longer revolves around the world; the world is made to revolve around the scientist. In the 1950s the nation's premier electronics laboratories (Bell Labs and RCA

Labs) conducted much of the basic research that drew the worlds of electronics and music together thereby establishing the foundation for the ICT techno-economic paradigm in music. In this manner, they acted as 'centres of calculation' (ibid) that mapped the frontier of the electronic music world, literally, in the circuit diagrams and computer algorithms that translated mathematical logic to music, and thereby drove the innovation system. By the mid-1 960s, the workshops of iconoclastic synthesizer makers such as Bob

Moog in places like Trumansburg in up-state New York rose to prominence by producing instruments that were complementary to, or, translated well with, the interests of progressive and experimental musicians. In the 1970s, California firms run by transectorial migrants from the state's aerospace sector commanded the heights of the industry. Over these three decades, umbrella organizations, especially the Audio

Engineering Society (AES) drew these distant interests together by providing a forum for enrolling sectorally disparate actors into a community of practice, the circuit Latour

(1 999) calls 'autonomization'. The point is that the post-war EM1 socio-technical trajectory in the US evolved amongst a dispersed set of sites. In some cases the connections between these spheres were direct, in others mediated through bodies such as the AES. In sum, these various actors worked individually and collectively to map the parameters of a problem and enlist other actors into the network. To all, the future looked bright, even if nofirm had yet broken through to the other side of the digital divide.

In the 1980s the US EM1 industry rapidly underwent a sudden demise, one that involved an almost wholesale shift of major production to Japan. Indeed, by the end of that decade, Japanese firms such as Yamaha and Roland had horizontally integrated much of their former US competition. This spatial shift 'coincided' with an important technological discontinuity at the point when EM1 entered the digital realm. From the perspective of the US, the demise of its industry is even all the more disconcerting because a select cohort of its engineers had generated the inventions that would subsequently radicalize the industry. Moreover, a number of American firms had been given the option to develop these technologies, only to decline. At this point, a network that had been centred in the US fell apart and the centre shifted to Japan.

The emphasis of this chapter is on the trajectory of the American EM1 industry because it served as the 'core' region in the early phases of this sector's technological trajectory. Especially in the period from 1950 to 1980 this trajectory corresponds to the

'transectorial inspiration' and 'entrepreneurial and intrapreneurial invention' phases of the transectorial SIS (Table 2.2). These stages of the narrative are crucial because they establish the context to understand why the US served as the key source of knowledge for the Japanese EM1 industry concentrated in Hamamatsu. Consequently, ensuing chapters focus on the rise of the Japanese industry and the processes through which American transectorial knowledge was translated into the orbit of Japanese firms. However, to understand these later trans-regional extensions of the innovation system requires an appreciation of their origins in the US.

After detailing the rise of various US centres of EM1 design and production, I then account for their downfall. Why was the US industry, after setting the course for this trajectory, unable to transform its foundation of knowledge into sustained industrial performance? I interpret the demise of the US according to Latour's ideas about the 'circulatory system of science' (see Figure 2.1). In particular I argue that, by the mid

1970s, the circuits in the network began to break-down as a community of entrepreneurial firms that was particularly rich in collegiality and well connected to the musical elite, failed because of its weak links with allies and the (general) public. The downfall of the

US industry occurred from this time for a number of reasons. Most generally, its collapse can be ascribed to the failure of US institutions to organize for radical innovation. While the US possessed the right engineering talent and institutional mix to undertake the very critical basic research and early invention that set the course of the industry's trajectory, these institutions were inappropriately configured to realize the ultimate potential of the digital age. Such a criticism is by no means new (Dertouzos et al. 1989). Nevertheless, the failure of US industry was a failure of its firms - as institutions failed to connect with the interests of their engineers, and with sources of capital. More specifically, the nature of enterprise failure hinges on mis-directions in corporate strategy, particularly in regard to technology. Admittedly the industry produced a number of highly influential engineers and inventors who still play key roles in the industry. However, their best efforts were realized only after they had left US firms and started working for the Japanese.

The demise of US firms played out differently depending on the size of enterprise.

Most critically, larger firms such as Harnmond were unable to recognize the moments of

'technological crystallization' (See Table 2.2), when they were presented with the opportunity of investing in the platform technologies of the digital era. Another structural feature similarly constrained the potential of smaller, highly inventive companies such as

Moog. The problem was conglomeration. Large corporations like CBS, and Norlin, heeding the prevailing wisdom of contemporary management discourse, absorbed these smaller enterprises during the conglomeration craze of the 1970s. These relationships, however, proved a poor institutional fit for the negotiation of radical technological change. While in one sense, conglomerations are often formed to be multi-sectoral, it does not follow that these structures are necessarily designed to successfully harness the technological synergies of transectorial innovation. In other words, a situation arises when the business interests of the conglomerate and the engineering interests of the acquired firm do not translate. There is a clash of cultures that arises from different ways of knowing and acting. Finally, US firms failed in their approach to the broader market.

Though successful at courting the attention of the musical avant-garde they neglected the average consumer who was unconcerned with programming algorithms and just wanted to make music.

The chapter is organized in the following manner. First, I situate the emergence of electronic musical instruments (EMI), particularly those that are controlled via a keyboard, by referring to their antecedents. Thus, the histories of the piano and organ industries are briefly outlined, paying particular attention to the shifting connections between the geographies of production and consumption. Next, the locational history of

EM1 is narrated as a diffusion of the engineering disciplines that passes through a number of key centres that successfully translated much of the foundational knowledge into the field and performed a good deal of the early practical problem solving. Technologically, the recurring characteristic of EM1 until the late1960s is, with notable exceptions such as the Hammond organ, how these instruments tended to be the brain-children of individual inventors working outside a corporate context and the darlings of the musical avant- garde. These characteristics insulated the EM1 trajectory fiom the popular market of amateur musicians. Finally, I focus on the particular period from the early- 1970s to early-

1980s when these instruments crossed the threshold to widespread popularity. In terms of the broader history, this period was marked by radical technological change, in which an entirely new market for flexible instruments based on digital synthesis was created. This is when things fall apart for the US and comes together for firms based in Hamamatsu as it is the latter who successfully mobilize the 'technical' and 'commercial crystallization' phases of the transectorial SIS.

Besides tracing the EM1 technological trajectory, I show how developments in this field bend back to relate to acoustic world, in particular to the class of keyboard instrurnents that are the antecedents of EMI. Indeed the popularization of the EM1 market proved concurrent with the de-segmentation of markets for pianos, organs and electronic keyboards, a trend which blurred the distinctions between these instrurnents. Not surprisingly it was Japanese firms that sought to capitalize on this substitutability by, for instance, developing both acoustic and digital pianos. I argue that this watershed emerged at the intersection of solutions to a constellation of technical problems which included achieving stable tuning, real time programmability and inter-instrument interface. In short, these solutions were all achieved through the application of binary digital logic as instrurnents started to take on many of the properties of computers. In order to understand why instrument and computers became intertwined, it is first necessary to chart the enduring legacy of keyboard instrurnents. 3.2 Instrumental Antecedents to EMI: Pianos and Organs

3.2.1 The Piano Industry

The piano is instrumental as a cultural artifact of the modern era (Weber 1978). It was the first instrument suited to forms of manufacture beyond craft production. Invented by the

Italian Bartolomeo Cristofori in 1709, and produced by hundreds of craftsmen

(principally cabinet-makers) on the continent, the pianoforte (soft-loud) was not manufactured on any substantial scale or dedicated manner until it met with the constellation of forces present in the crucible of the Industrial Revolution and the British firm of Broadwood, which turned out 400 instruments in 1800 (Loesser 1954: 234).

"With its manifold intricate structure - and especially its abundance of serially repeated parts - [it] seemed particularly suited to the new mechanical processes.. .The piano was the factory's natural prey; purely on the basis of its structure, it was an instrument of the time" (ibid 1954: 233).

In the New World, the piano soon became a significant force in America's industrialization. In 1853, the largest single industrial building in the country belonged to the Chickering Piano Company of Boston (ibid: 495). In 1900, the piano industry accounted for 111 20~of the entire US economy (Litterst and Malarnbri 1999). At that time, about one in six New Yorkers worked in some piano-related job (Cole 2003). "It is hard to picture now, but at one time the piano industry had an influence on the national economy of the US very similar to that of the auto industry today (Fandrich 200 1: 1).

Pianos are percussion instruments. Their technology of sound production is essentially mechanical. Consequently, innovation in acoustic piano design has proceeded in an incremental manner, propelled largely by advances in metallurgy. As Fandrich (2001 :1) points out, "by roughly 1875, the basic concept of the piano as we know it today had become fairly well stabilized.. .Many of the pianos in production today trace their designs back to the early 1900s.'' With a static design, the ability to make pianos, and consequently the geography of production has been dictated by market fundamentals such as levels of consumption and the ability to realize economies of scale in manufacturing. Notable industrial clusters involved in the production of pianos were centred around lead firms like Steinway (New York), Chickering (Boston) and Baldwin

(Cincinnati). Southern Ontario even had its own congregation of firms (see Kelly 1991)

Consumer markets for pianos reach saturation points. One informant portrayed this phenomenon by claiming that the most significant competitive threat facing the venerable USIGerman company Steinway is not Japanese brands like Yamaha or Kawai, but rather by the market created by the refurbishment of old Steinways. A similar phenomenon exists in Japan where one periodically sees television commercials for piano recycling services34.Moreover, there is a geography to this recycling, with many of these units being sent to South East Asia for sale. One of the principal markets for recycled pianos from Japan is Malaysia where firms such as the aptly named Cristofori employ

"technicians and tuners [who] were trained by experts from Japan" (New Straits Times

2003) to repair and recondition pianos for sale in the Subang Jaya district.

The dynamics in these factors have been responsible for an intriguingly progressive westward shift in the centre of production fiom Europe to the US (Table 3.1).

Thus the US sat at the top of the league tables in the early decades of the 20~century while Japan was similarly in ascendancy fiom the 1960s to 1980s. Not shown on the table

34 1 do not think I have ever seen a TV ad for pianos, let alone recycled ones in North America.

96 are subsequent developments. Since the 1990's East Asian countries, successively Korea and China, and S.E. Asian countries such as Malaysia and Indonesia have seen their

Table 3.1: Estimates of Piano Production 1870-1984 (1.000. , units) UK Franc Germany US Japan Russia S. Total e Korea 1870 25 2 1 15 24 85 1890 50 20 70 72 2 232 1910 75 25 120 370 2 592 1930 50 2 1 20 120 2 212 1960 19 2 26 160 48 88 243 1970 17 1 45 220 273 200 666 1980 59 218 393 166 81 964 1984 51 106 426 148 125 880 ~ource:Ripin et al. 1988 production market shares soar. These latter countries primarily produce pianos for

Japanese firms and local consumption has a long way to catch up with exports. The exception is China. Currently, the largest piano factory in the world belongs to the Pearl

River Piano company in Guangzhou, China, a joint venture between Yamaha and a local partner. Significantly, only Japanese firms have been able to successfully produce both pianos and EMI. They have done so by drawing these instrument classes together.

There will never be a replacement for the piano. However, statistics indicate that consumers are willing to purchase electronic substitutes. In 1999,92,10 1 acoustic pianos were sold in the In that same year consumers bought 86,086 digital pianos, many of which were manufactured by Japanese firms36(Music Trades 2000: 98). While the value of the latter market was considerably less - $158 million vs. $669 million - consumers appeared increasingly enchanted by digital models carrying average price tags under

$2000, in comparison to the $3,600 and $14,000 necessary to purchase acoustic vertical and grand pianos (ibid). The US International Trade Commission (1 999: 29) reported that digital pianos were especially considered "close substitutes for vertical acoustic pianos for beginner level consumers or those that lack a strong preference." One respondent to the commission, a piano dealer, revealed that, "children entering the store with their parents go directly to the digital pianos that they find more appealing than acoustic pianos" (ibid). While it is unlikely that the digital piano and its kind will usurp the position that the acoustic piano forte has enjoyed in concert halls around the world, it is the domain of the home where the digital revolution is being realized. Anecdotally, not

35 Specifically, this figure %,lo1 is comprised of 32,155 grand and 59,946 vertical pianos. (Music Trades 2000: 98) 36 In 1999, Japanese firms manufactured 249,000 digital pianos, in comparison to 134,000 acoustic units (Japan Music Trades 200 1). long ago, on a visit to my friend's house, I was pleased to learn that his 8 year old son was taking piano lessons. Eager to show me what he was learning, the child proudly ran through a series of scales and arpeggios on his $200 Casio electronic keyboard, which featured a virtually real piano tone function.

3.2.2 Organs

The concept of the pipe organ dates back to the 3rd Century BC when a Greek engineer named Ctesibius invented the hydraulis, a mechanical flute playing instrument with wind pressure regulated by means of water pressure. For the next millennium, the hydraulis spread throughout the Mediterranean and was played at banquets games and circuses. Its association with the church did not happen until the year 757, when Byzantine Emperor

Constantius sent an organ and other costly gifts as peace offerings to Charlemagne's father, Pepin the Short, King of the Franks. This act had remarkable consequences. In order that the gift might be copied, a Venetian monk was requested to teach organ building methods to students. From that time onward, the organ spread throughout

Europe and by the 1400s, the use of organs was well established in monastic churches and cathedrals throughout Europe. Beyond its ecclesiastical utility, however, the organ was not considered an instrument suitable for serious classical composition until the

Baroque era. In a letter to his father dated October 17h 1777, Mozart wrote that, "in my eyes and ears", the organ is, "the king of instruments" (Westfield Centre 2001). The

Dictionary of Music and Musicians by Sir George Grove noted in 1889 that,

The organ is, together with the clock, the most complex of all mechanical instruments developed before the Industrial Revolution. Among musical instruments its history is the most involved and wide-ranging, and its extant repertory the oldest and largest ...No other instrument has inspired such avowed respect as the organ (Ibid). Yet until the modern era, the organ's place continued to be the church, as its technology was inseparable from the architectural requirements of these spaces - churches and cathedrals were designed in anticipation of the installation of the pipe organ (Kakehashi

2002: 172).

Both organ technology and the instrument's role in music and society evolved rapidly following the industrial revolution. The Great Exhibition at Crystal Palace in

185 1 featured organs from several countries and this juxtaposition likely prompted the ensuing design improvements and technological branching points. The first of these was the reed organ or Harmonium (as it was known in France) which was invented in the

1850s as a cheaper alternative to the piano37.Waring (2002), in noting this instrument's popularity and inseparability from Victorian culture, points out that purchases of reed organs in the US between 1850 and 1910 exceeded that of pianos by almost two to one.

The reed organ was also the first keyboard instrument manufactured in Japan. A little less subtle an instrument was borne in 1855 when Joshua C. Stoddard of Worcester, England invented a steam-powered organ called the calliope. This instrument was apparently so expressive in its volume that Worchester's city council banned Stoddard from playing it within the city limits. That the calliope drew on the key technology of its TEP is not surprising and is illustrative of a recurring theme that carries over into the domain of electronic instruments (Ibid).

37 This instrument's technology involved a foot pump which activated a suction bellows for drawing air out from within the action of the organ. When a note is played by depressing a key, air is sucked into the action through a small chamber containing a free reed. The passing air causes the reed to vibrate, producing a tone. In America the dissemination of organs was propelled by patrons of the arts such as Andrew Carnegie who, beginning in 1873, donated over 8000 organs to churches, schools and other civic institutions. The next technological watershed occurred in 1896 when Robert Hope-Jones built an organ with an electric action embodying some radical tonal ideas for the cathedral in Worcester, Massachusetts. Hope Jones is regarded as the father of the modern organ, and he was soon hired by the Wurlitzer Co. of Cincinnati to design an instrument to accompany and complement another contemporary technology, the motion picture (Westfield Centre 2001).

Between 1910 and 1927 several thousand theatre organs were installed in movie houses to accompany films during the silent era. Known generically as The Mighty

Wurlitzer, this billing was certainly appropriate - the Wurlitzer installed in Radio City

Music Hall apparently contained more wiring than in all the rest of the Rockefeller

Center. Organ production continued apace with the number of units made annually in

America doubling to 2400 between 1909 and 1927. In this latter year, the largest organ in the world was installed in Wannamaker's Department store in Philadelphia. The release of the first talkie, A1 Jolson's The Jazz Singer in 1927 quickly brought an end to this period, and a further displacement of the organ. This event along with the Depression effectively decoupled the organ from the theatre and scores of organists had to seek employment in the broadcast booth, the sports arena and the nightclub (Kelzenberg

2002). The Hammond Organ, invented in 1935 would further redefine the technology and situation of the organ in American music.

As with the piano, the organ too is now fully digital. However, unlike the piano, the market for organs, particularly those destined not for churches but for homes, has experienced a rapid decline. In 1992 US consumers bought 17,660 home organs. In 1999 this figure had slid to 9,400 units (The Music Trades 2000). Even in the organ's traditional domain, the church, pipe organs have long given way to electronic and increasingly digital instruments3*.To make sense of this situation it is necessary to now explain the comparatively brief yet profound trajectory of electronic musical instruments

(EMI).

3.3 The Evolutionary Trajectory of Electronic Musical Instruments

Much has been written on the history and developmental trajectory of electronic musical instruments (Armbruster 1984, Mochida and Aoki 1994, Colbeck 1996, Chadabe 1997,

Theberge 1997, Pinch and Trocco 2002, Kakehashi 2002). Prior to the mid 1990s these accounts emphasized explanations of the technical properties of the instruments, biographical details of the pioneering inventors and provided cursory appraisals of the significance of these instruments for influencing both high and popular culture. However, according to Theberge (1 997: 43), these first generation histories frequently lacked any detail that adequately captured the,

context of invention, for example the accumulation of scientific knowledge and engineering expertise in a particular field, which often precedes the invention itself, and the musical social, economic and institutional forces that help or hinder it.

Moreover, Theberge also suggests that there is often a failure to separate the wheat from the chaff, or distinguish between the few innovations that mattered versus those inventions consigned to the tar-pits of history. The second generation accounts, especially

Chadabe (1997), Theberge (1997), Pinch and Trocco (2002), remedy many of these

38 Notable exceptions to this trend include Quebec's Casavant.

102 concerns and I draw substantially on their insights in shaping my analysis of the industry's socio-technical trajectory. In later chapters, I expand on their insights by focusing on the neglected Japanese dimension of the story.

My emphasis is first to place the history of electronic musical instrument design within the initial phases of the evolutionary economic framework detailed in Table 2.2 in a manner that highlights how these developments were embedded in specific geographic circumstances of a few key centres. Second, I trace the cumulative and derivative, yet ultimately contingent, trajectory of this evolution. Ideas regarding EM1 design proved contagious, and successive instruments often drew on the characteristics of their predecessors. More importantly, whole generational cohorts of instruments were frequently based on similar sets of scientific knowledge rather than radical departures, a point which is raised in Theberge (1997), but not examined in a systematic manner. The third point of emphasis is on the recurring transectorial character of invention and innovation which is manifested in biographical details of the actors involved. In other words I focus on whether new designs were made by solitary inventors, staff of research institutions like Bell Laboratories, or engineers employed by instrument manufacturers.

Transectorial effects are also reflected in the sources of knowledge: theories, papers and other texts that circulate between these actors, as scientific knowledge from outside was introduced to the music industry as a whole. 3.3.1 Spatio-temporal Clusters in the Innovation of EM1

A site on the internet entitled 120 Years of Electronic Music: Electronic Musical

Instruments 1870-1990j9 lists 109 instruments deemed to have had significance in the evolution of the industry's design trajectory. Undoubtedly this site's selection criteria is subjective, yet for an initial overview the inclusiveness of their approach is appropriate - notably, none of the highlight instruments that frequently appear in other accounts are omitted. To make sense of this list in its entirety, I plotted the frequency distribution of the listed 109 inventions over time and space (Figure 3.1).

Figure 3.1 The Distribution of EM1 Innovation over 120 Years

Other Japan USSR &I France Germany UK H USA

Source: data compiled from http:llwww.obsolete.coml120 years/ :Accessed 1016103.

Until 1925, the incidence of new EM1 is both sparse and sporadic, but beginning with this period a clear pattern is evident, with peaks around 1930 and 1980 - not coincidentally one Kondratieff cycle apart. Geographically, American inventions have always remained significant although regions of secondary importance have shifted from Europe, principally France and the USSR in 1930 to Japan in the 1980s. From this larger population, I have selected twelve key inventions that I judge to be important which appeared before 1970 for closer examination and the details of these are contained in

Table 3.2. This list is intended to be illustrative rather than exhaustive.

Each of the peaks in Figure 3.1 marks an acceleration or swarming phenomenon in the introduction of new instruments. Technologically, these episodes correspond with a set of individual circumstances which shaped the application of general scientific knowledge. For instance many of the new instruments that emerged around the 1930s

(e.g. the Hammond organ and the Orgel (Table 3.2)) drew on similar scientific principles

(e.g. the heterodyning principle40)and employed similar core technologies (e.g. vacuum tube oscillator^^^). Indeed, the vacuum tube oscillator remained the core technology in

EM1 until at least the early 1970s when transistorized integrated circuits (IC) took over as the basis of the second peak. In each case there appeared a swarming of interests seeking to exploit these fundamental technologies. These phenomena correspond to Schumpeter's ideas of technological clustering. An examination of individual biography captions summarized in Table 3.2 reveals how this general process of transectorial migration is composed of quite diverse individual circumstances.

Leon Termen a Russian physicist, for instance, enacted Lenin's dictum that 'socialism is proletarian dictatorship plus electrification' when he invented the Theremin in 1920 (see

Table 3.2). That year the Theremin was introduced to the public at the Moscow Industrial

Fair. Soon after, Lenin commissioned a private demonstration and requested lessons from the inventor. Lenin proved a valuable ally, ordering 600 models to be produced and

40 Heterodyning: when you mix two signals with two slightly different frequencies the resulting signal has a frequency equal to the difference of the two frequencies. (Armbruster 1984) 41 Until the invention and adoption of the transistor, electronic devices relied primarily on the triode or three electrode vacuum tube which was invented by Lee DeForest in 1907. toured around the USSR~~.Termen's ideas inspired later inventors, including Bob Moog, who entered the discipline by building and selling Theremin kits. The earliest 'trace' that

Termen's ideas traveled is found in pair of French inventions: the Ordes Martenot and the

Orgue des ~ndes~~.The latter, for example was born in 1929 out of the collaboration of

Armand Givelet, an engineer and physicist who worked the radio laboratory at the Eiffel

Tower and organ builder Eduoard Coupleaux. In this case the intersection of interests is quite clear.

42 Termen's story follows a number of interesting turns. For instance, some years after immigrating to New York, Termen was kidnapped by the NKVD (precursor top the KGB), transported back to Moscow and accused by Stalin himself of treason. His subsequent internment in the Magadan Gulag was a productive one, for Termen's tacit knowledge of electronic sound was translated to the task of designing 'bugs' for the espionage community (Source http://www.obsolete.com/l20 vears/machines/theremin/index.html; Accessed 25/06/04) 43 In English, Ondes translates as 'waves'. These instrument names would have been unthinkable before the discipline of physics introduced this metaphor to capture the form of sound. .e 3.2: Key Innovations in the 2othCentu Develo ment of Analo E mctronic Musical Instruments Invention Year Inventor(s) Technology of sound Comments generationlEmbodied ' Iknowledge Telharmonium T. Cahill (US) Helmholtz's' overtone The first electronic musical instrument. Powered by 6 dynamos, theory the telharmonium transmitted music to subscribers over Additive synthesis telephone wires. Tone wheel Theremin L. Termen Heterodyning principle Played by moving the hands around the metal loop for volume (USSR) Vacuum tube oscillator and around the antennae for pitch Lenin commissioned 600 models to be built and toured around the Soviet Union Heard on Beach Boys' Good Vibrations Ondes Martenot M. Martinot Heterodyning principle Based on Termen's ideas but incorporating a standard keyboard (France) Vacuum tube oscillator The first successful EM1 and the only one of its generation that is still used by orchestras today Orgue des Ondes A. Givelet & Vacuum tube oscillator Based on the same technology as the Theremin and Ondes- E. Coupleaux Additive synthesis Martenot but the 'Wave Organ" had an oscillator for each key (France) enabling polyphony Subtractive synthesis Trautonium F. Trautwein Neon tube oscillator Manufactured and marketed by Telefunken AG (Germany) Subtractive synthesis Consisted of a resistance wire stretched over a metal rail marked with a chromatic scale Hammond Organ L. Hammond Vacuum tube oscillator Most important (most well connected) innovation in keyboard (USA) Additive synthesis design before 1950. Tone wheel Warbo Formant Orgel H. Bode Formant filters Designed at Heinrich-Herst Institute in Berlin. Bode emigrated to (Germany) Geographical key the US in 1954 to work professionally for Bell Aerospace and on assignment an amateur basis on later designs of EMI. His inscriptions inspired Moog and Buchla. The RCA Synthesizer H. Olson & Vacuum tube oscillator First programmable EM1 H. Belar Digital control via punch- Charles Wuorinen's Time's Encomium a composition for the (USA), paper roll RCA synthesizer awarded the Pulitzer Prize in 1970 Bell Labs Music mk I-IV 1957 Max Pulse code modulation MUSIC series software went through stages of evolution tj Software Matthews (from telephony) following the development of the IBM computer which ended in * (USA) 1968 with MUSIC V written in FORTRAN and running on the -tj wavetable oscillator for 6 sound sampling IBM 360 machines ._ma I .s Siemens Synthesizer (studio for 1959 H. Klein and Vacuum tube oscillator Pierre Boulez and Karlheinz Stockhausen visited the studio and electronic music) W. Schaff wrote compositions to be performed on the technology 'w$ ~i~it~lcontrol via punch- --- (Germany) paper roll Buchla 1963 Donald Voltage controlled Collaboration with the San Francisco Tape Music Centre Buchla (USA) electronic instrumentation L ~~~ltime control P 2 Q Moog 1963 Bob Moog Voltage controlled Wendy Carlos' Switched on Bach performed on a Moog h modules q" (USA) becomes a best seller. - Sources: Vail 1993, data compiled from http://ww.obsolete.cod120 years/ :Accessed 10/6/03. 0 00 3.3.2 The Hammond Organ

Following these early inventions, the first major EM1 innovation was the Hammond organ. Compared to the inventors who preceded him, Laurens Hammond proved to be the only entrepreneur and it is this confluence of engineering and business interests which contributed to the emergence of a dominant design. In the 1930s' and for much of the following 50 years, the representative instrument was the Hammond organ (Table 3.2).

After the digital divide, Yamaha's DX-7 is credited with assuming this role. This latter instrument's story will be featured later in the thesis.

Originally intended as an electronic alternative to the typical church pipe organ, the Hammond organ represented an improvement over contemporary EMI. Both the

Theremin and Ondes-Martenot achieved popularity in orchestral and avant-garde circles yet failed to have much of an impact on popular music tastes. The Hammond advantage derives from transectorial links between electro-mechanics and music. Specifically,

Hammond's background of making clocks is reflected in his organ design. The series of tone wheels that comprise the Hammond's sound source are driven by a synchronous motor that was not unlike the motor of the electric clock - the first product of the

Harnmond Company. Based to a large degree on Cahill's Telhannonium, the

Hammond's use of vacuum tube amplifiers and silver dollar sized tone wheels effected a miniaturization of its predecessor. Upon its release in 1935 the Hammond cost about

$2600 in comparison to the price of an average pipe organ at $75,000 (Kakehashi 2002:

229) and this price difference certainly boosted its appeal to consumers. Concerned pipe organ makers took Hammond before the Federal Trade Commission (FTC) in 1936, arguing that the Hammond was not an organ at Hammond eventually prevailed in this battle and through successive models such as the Novachord, Solovox and the legendary B-3, defined the most widely recognized sounds in electronic music until the advent of commercial synthesizers in the late 1960s (Vail 1997).

The Hammond's success reflects superior intrinsic qualities (and especially its portability) and Latour's point about strong connections between heterogeneous interests.

The Hammond enrolled some formidable allies including many churches, the FTC, the

US Army Chaplaincy and even Henry Ford. Regarding the latter, an anecdote holds that

Henry Ford demanded to be the first Hamrnond organ owner (Vail 1997). Yet these sorts of connections cannot hold indefinitely, nor can their topology remain defined by the interest of one place. Later chapters frame Hammond's demise from the standpoint of

Japan, for it was with the Hamamatsu-based firm Roland that forged the strongest ties with Hamrnond on the threshold of the digital age. Since 1993, the Hammond brand of organs has been manufactured by Hamamatsu-based firm Suzuki.

3.3.3 Into the Laboratory

The next phase in the evolution of EM1 took place in the 1950s and early 1960s (see

Table 3.2). Large research laboratories, such as the RCA Lab at Princeton, Bell Labs at

Murray-Hill, NJ, and Siemens in Germany constituted the principal centres. A quantitative revolution took place in the field of music as fundamental research in math, physics, electrical engineering and the nascent field of computer science was applied to the problem of musical composition and instrumentation. RCA's research engineering

44 Similarly themed arguments were also voiced 50 years later by piano makers who took slight at the idea of the digital piano. team of Harold Olson and Herbert Belar, seeking to enact ~chillin~er's~'(1 949) "A

Mathematical Theory of Music", devised the first programmable synthesizer in which

pre-solid-state analog components for tone generation were controlled digitally by means

of punched paper rolls. Controversial, due to the claims of RCA's Chairman David

Sarnoff, that "it is not necessary that a composer be able to play a musical instrument, for

whatever effects he wants to create can be achieved by use of the synthesizer", RCA was

vilified by the American Federation of Musicians (Chadabe 1997: 108). In 1959, the Mk

I1 version was installed at Columbia University with the funding from a Rockefeller

Grant. The Colurnbia-Princeton Electronic Music Centre's reputation as the premier

institute of its kind through most of the 1950s and 1960s drew Milton Babbitt, Charles

Wuorinen and several other notable scientist-composers. At the same time in Germany,

composers such as Stockhausen and Varese were visiting Siemens to book time at that

firm's studio for electronic music. In 1957, Max Mathews at Bell Labs Acoustic

Research Department made the first computer generated sounds while conducting

experiments in telephony. In 1959 a recording called Music for Mathematics was produced. A copy sent to Aaron Copland elicited a reply stating that "the implications are

dizzying and, if I were twenty, I would really be concerned at the variety of responses

suggested" (quoted in Chadabe 1997: 109).

In practice, the instruments that emerged out of these institutional settings had

little immediate influence beyond these research communities and the avant-garde. Like

the computer mainframes of the time they occupied entire rooms and their sheer

complexity prohibited real-time performance. Babbitt characterized the experience of

45 Schillinger, a gifted physicist and violinist, emigrated from Russia to the United States in 1925. A long career at RCA began with a brief collaboration with Theremin. Schillinger's Mathematical Theory of Music held that music could be generated by a system of random probability (Armbruster 1984). working with these machines as being, "tedious, and demand that you learn so much before you can even make your first combination of sounds."(quoted in Armbruster 1984:

5 1). Once this skill was mastered the physical task of typing all the parameters onto punch cards took a few additional weeks. Nonetheless the Pulitzer awarded to

Wuorinen's piece Times Encomium affirmed the validity of electronic music as a legitimate, if limited field (Chadabe 1997). So eventually, the mundane world of the laboratory had been translated to the public sphere (or public circuit in Latour's terms), albeit in limited fashion.

In general the EM1 laboratories of this era reflect a different set of strengths via connection. Indeed, in the laboratory, the diffusion of engineering disciplines is managed and contained both organizationally and spatially. Moreover, similar to the telhannonium and unlike the Harnmond, the instruments produced in these contexts were immobile. As a result, heterogeneous interests were compelled to visits the labs to work with these instruments. That is, until the laboratories began mobilizing the world through inscriptions: matrices, algorithms, equations that represented the logic of the new synthesizers in a form that could be transported between contexts.

If instruments of this era 'failed' commercially, the ideas that lay behind them have had lasting legacies and it is their creator's performance as authors that been most influential. Olson's text, Music and Physics was published in 1952 and translated into

Japanese ten years later. Two generations of Japanese respondents who I interviewed cite this text as being foundational to their education and training in the field of EMI.

Similarly Matthews 1963 article in Science, "The Digital Computer as a Musical

Instrument inspired John Chowning at Stanford to develop FM Synthesis, the technology embodied in Yamaha's DX-7. These early texts codified much of the knowledge that would drive the problem solving trajectory in EM1 design for the next 40 years. In doing so they constituted the earliest frames of reference for the emerging community of practice in the field. Umbrella organizations, such as the Audio Engineering Society in the USA, also drew this community into focus.

3.3.4 The Audio Engineering Society (AES): A Community of Engineers Looks Forward to the Digital Age

From the mid-1960s through the early 1970s, the US-based Audio Engineering Society

(AES) was the key institutional forum for dialogue on the direction of electronic musical instrument design. In the pages of the society's journal and at the society's convention, engineers practicing their craft in diverse circumstances shared their interests with their peers. Analysis of the Journal of the AES (JAES) and convention programs from the years 1964 through 1972 furnished glimpses of this community's composition which included corporate engineers, entrepreneurs, lone inventors, researchers and academics

(see Table 3.3). These individuals were affiliated with a reasonably narrow list of leading institutions including: large corporate electronics laboratories (RCA, Bell), leading organ manufacturers (Conn, Harnrnond), upstart synthesizer makers (Moog, ARP, Buchla) and universities (Princeton, U. of Illinois, Stanford). Occasionally, some participants appeared out of 'left field' such as from the electronic firm Motorola and the aerospace company North American Rockwell - and it is these non-traditional, transectorial sources of knowledge which would later prove to be catalytic. For instance, a patent for an electronic organ developed by a Rockwell engineer is the most widely cited patent in the field of electronic music (as chapter five will prove). Table 3.3: Session Participant Affiliations at the AES Conventio~ 1964-1972 Session Presenter's Affiliation Location of Title Presenter's Affiliation Music and Bell Aerospace Co. N. Tonawanda, N.Y Electronics Electronic Music Studio, Brandeis U. Waltham, Mass. Research Lab of Electronics, MIT Cambridge, Mass. R.A. Moog Company Trumansburg, N.Y. School of Music, U. of Illinois (2) Urbana, Ill. National Research Council of Canada Ottawa, Ont.

Speech Bell Telephone Laboratories Inc. (2) Murray Hill, N.J. Processing Bolt Beranek and Newman, Inc. Cambridge, Mass. Philco, Corp. Blue Bell, Pa. Air Force Research Laboratories CRBS Bedford, Mass. RCA Laboratories Princeton, N.J. Nippon Electric Co. Tokyo, Japan IBM Yorktown Heights, NY Sperry Gyroscope Company Carle Place, N.Y. Music and Hammond Organ Company Chicago, Ill. Electronics Astrosonics Incorporated Syosset N.Y. R.A. Moog Co. Trumansburg, N.Y. Hofstra U. Hempstead, N.Y. Argonne National Laboratory (2) Argonne, Ill. Bell Aerospace Co. N. Tonawanda, N.Y. Contact Associates, Inc. New York, N.Y. Music and C.G. Conn (Organs), Ltd. Elkhart, Ind. Electronics Catgut Acoustical Society Montclair, N.J. Perma-Power Co. Chicago Ill. General Radio Company W. Concord, Mass. Motorola Semiconductors Phoenix Arizona R.A. Moog Co. Trumansburg, N.Y. School of Music, U. of Illinois (2) Urbana, Ill. Bell Telephone Laboratories Inc. Murray Hill, N.J. Princeton U. Princeton, N.J. Speech R.A. Moog Co. Trumansburg, N.Y. and Music Wayne State University Detroit, Mich. RCA Laboratories Princeton, N.J. National Research Council of Canada Ottawa, Ont. ldentitones Inc. New York, N.Y. Dept. of Computer Science, University of Toronto Toronto, Ont. Cunningham Dance Foundation New York, N.Y. The Wurlitzer (Organ) Company N. Tonawanda, N.Y. Date and Session Presenter's Affiliation Location of Location Title Presenter's Affiliation 511 968 Music and Bell Telephone Laboratories Inc. Murray Hill, N.J. ( Los Speech School of Music, U. of Illinois Urbana, Ill. Angeles) University of Southern California Los Angeles, CA Department of Music, Stanford University (2) Stanford, CA University of Massachusetts Amherst, Mass. UCLA Los Angeles, CA lndependent Washington, D.C. 1011969 Developm C.G. Conn (Organs), Ltd. (2) Elkhart, Ind. (New ents in Electronic Music Studio, University of Toronto Toronto, Ont. York) Electronic Music National Research Council of Canada Ottawa, Ont. System R.A. Moog Inc. Trumansburg, N.Y. Radatron Inc. Tonawanda, N.Y.

1011970 Electronic CBS Musical Instruments Fullerton, CA (New Music Hammond Organ Co. Chicago Ill. York) Dept. of Music, Queen's College New York, NY ARP - Division of Tonus Inc. Newton Heights, Dept. of Electrical Engineering, Ohio University Mass Alfred Mayer Ionic Corp. Athens, OH Norm Milliard Electronic Music Laboratories Cambridge, Mass R.A. Moog Inc. Hartford, Conn. Trumansburg, NY 1011971 Electronic ARP - Division of Tonus Inc. (2) Newton Heights, (New Music Hammond Organ Co. (2) Mass York) New England Conservatory of Music Chicago, Ill. Moog Musonics Co. Boston, Mass. Autonetics Division, North American Rockwell Inc. Williamsville, N.Y Anaheim, CA 511 972 Electronic Department of Music, Stanford U. (2) Stanford, CA ( Los Music lndependent New York, NY Angeles) Different Fur Trading Company San Francisco, CA Buchla and Associates Berkeley, CA National Research Council of Canada Ottawa, Ont. Moog Music Inc. Williamsville, N.Y. of the Audio lgineering Societv, various issues 1964- 1972. The AES data suggests something of the shifting engineering mind-set at the threshold of the digital age. With each successive year, practitioners engaged with the problem of electronic music joined together as session participants. How the community defined itself is revealed in the session titles, which early on shifted back and forth from 'Music and Electronics' to 'Speech and Music', even if the participants were often the same. By the 1970s, participants settled on the title of 'Electronic Music' (see Table 3.3).

Constants, such as the virtually assured participation of Bob Moog, served to maintain the community. Around this period, RCA's Harold Olson served as the society's president.

Geographically, participant lists point to a limited number of core centres - New

York and New Jersey, Upstate New York, Chicago and Greater Boston. Until the late

1960s, conventions were almost exclusively held in New York hotels and these ephemeral forums attracted participants only from a catchment area of limited range. The

1968 convention held in Los Angeles served to break the regional dominance of the north-east and mid-west as California participants took part. This westward shift was coincident with a cultural turn in the community's composition and outlook.

Both culturally and technologically a sense of transition pervaded professional discourse. A review of the 1969 conference proceedings noted that,

As serious and creative as before, the 1970 audio engineer doesn't look much like one any more. Beards, bell bottoms, "hair" and granny glasses made the scene in appropriate numbers.

This cultural change also disrupted the conventions of the convention. Don Buchla's participation in the 1969 convention in Los Angeles was one such occasion. "By refusing to rent a booth and instead playing a concert, and thus gaining lots of free publicity, [Buchla] broke most of the rules for such conventions"(Pinch and Trocco 2002: 5 1).

Apart from the contingent of young Turks such as Buchla, it is worth commenting on the older, first generation engineers, for their career paths indicate where this cohort learned their trade (see Table 3.4).

These career trajectories as longitudinal profiles reveal much about a community steeped in the aerospace and consumer electronics sectors. They also point to a pattern of transectorial boundary-crossing migration as the origin of this population's way of thinking. This thinking was quickly brought up to date, especially in instances when it was forced to confront the radically unfamiliar. Bob Moog's presentation in 1968 entitled

'Recent trends in electronic music studio design' featured a recording of Wendy Carlos playing Brandenburg No.3 from her soon to be released landmark album 'Switched on

Bach'- the first album to demonstrate that synthesized sound could be used to make music that the general public could appreciate. Its debut before the AES was also radical in effect. As Moog has recalled, "I could remember people's mouths dropping open. I swear I could see a couple of those old bastards starting to cry.. .those cynical, experienced engineers had their minds blown" (Pinch and Trocco 2002: 144); "here was something that was impeccably done and had obvious musical content and was totally innovative. The tape got a standing ovation." (Freff 1988: 55) Table 3.4: Sample career path profiles of authors publishing papers on electronic musical instruments in the Journal of the Audio Engineering Society (1969-72) I Engineer Career profile Daniel Martin Baldwin Piano MSc (1939) and PhD (1941) in physics from the University of and Organ Co. Illinois. Worked for RCA between 1941 and 1949 developing microphones and headsets before joining Baldwin in 1949 where he rose to the position of director of engineering and research. Victor J. Bong Conn Organs BSc 1956, MSc 1957 in electrical engineering from Notre Dame. Prior to completion of his degrees he worked in automotive engine research and as a project engineer in a jet engine division of Studebaker Corp. After graduation joined Conn, as research engineer in Electronic Organ Division. Later work fell in area of New Organ Model Department Since 1961 was the Department Section Head for organ section etc. Melville Clark BSc 1958 in physics from the Case Institute of Technology. PhD Associates - in physics from MIT in 1963. From 1963-64 research associate in Moog- Norlin Cooperative Computing Laboratory of MIT working on development of computer controlled printing methods. Currently consultant for Melville Clark. Melville Clark BSc from MIT in 1943, PhD Harvard 1949 in physics. Worked for Associates - MIT Radiation Laboratory on microwave radar, on the Manhattan Moog- Norlin Project at Brookhaven National Laboratory specializing in reactors, at Radiation Lab UC on atomic and hydrogen bombs, for Sylvania Electric Inc. on ionospheric propagation, auditory perception and speech research, for Avco on electrical propulsion, orbit calculations and gamma ray transport, and for NASA on plasma physics and solid state physics. For a number of years he has also had his own business concerned with the development of musical instruments and devices as well as research in musical acoustics. Was also faculty member of Dept of Nuclear Engineering at MIT. Rodgers Organs Graduate of Georgia lnstitute of Tech. in Electrical Engineering (1951). After graduation he specialized in aerospace flight control system design with the Martin-Marietta Corp. Was responsible for the Mace missile system and for various new space-craft altitude control systems. He later joined Radiation Inc. in Melbourne Fla. Worked on design and analysis of advanced digital data and communications systems. Also on technical staff at MITRE corp. Now organ designer with Rodgers. ARP Yale degrees in music and electrical engineering followed by graduate studies in industrial and management engineering Princeton, Before joining ARP in 1969 he worked for RCA labs in Princeton NJ Currently marketing manager for ARP. Harald Bode Bell Aerospace Graduated from Hamburg University and Heinrich Hertz Institute Moog of Berlin. Afterwards specialized in the field of electronic music, creating the Bode-Melochord and pioneering the industrial production of electronic organs (Polychord) in Germany. Came to US in 1954, held several positions before joining the Bell Aerospace Company in Niagara Falls NY where he specialized in the field of micro-electronics while remaining active in the field of instrument design. Articles he published influenced Bob Moog who later licensed a few Bode patents. Ray Hammond BSRE degree from Indiana Tech College 1949. Joined Sylvania Schrecongost TV and Radio Division, Buffalo, NY. Engaging in television design. In 1954 he was employed at Admiral Corp. Chicago doing design work on military and consumer television systems. Joined Hammond in 1960. Warren Brunsting Hammond Wilson College and Illinois Institute of Tech. Worked for seven years on the design of military communication equipment for ITT Kellogg and also was employed in recording and instrumentation section of Illinois Inst. Of Tech Research Institute. Joined Hammond in 1967 team leader of Phoenix Hammond's first MDD tone-generation instrument. [Currently Vice-president Manufacturing Hammond Suzuki Inc.] Source: Author profiles compiled from articles in the Journal of the Audio Engineering, Society, various issues 1969-72. Technologically, the implications of the ICT paradigm were beginning to be reflected in discourse. For instance, in 1969, David Martin, chief engineer at Baldwin, the large Cincinnati piano and organ manufacturer, wrote a piece for the JAES that earnestly prophesized the trajectory of his discipline. In an article titled "Electronic Music: Audio

1988', Martin (1969:387; emphasis added) presciently observed that,

In the circuitry, the adoption of IC is expected, but who knows what other discontinuities in the state of the art might be equally important or more so, not only for electronic organs but also for other instruments and for their tonal accessories.

Two years later, Richard Shaefer (1 971 : 570) of Rodgers Organ Co. also commented on the fusing of electronic and musical technologies:

In the last decade, with the potential for low-cost semiconductor devices beginning to be felt, some very significant advances have been made. The best of the modern electronic organs are a study in cost effective sophistication. Their development has been guided by more rigorous theoretical and analytic studies of the nature of pipe tones, supplemented with relatively comprehensive test programs and coordinated with the limitations imposed by practical semiconductor circuits.. .the problems facing the electronics engineer in this field are considerable.

By 1972, the trajectory had been refined yet again. In a session at the 42ndAES convention organized by Stanford University's John Chowning, Hammond Organ engineer Ray Schrecongost (1972: 602 emphasis added) proclaimed that,

Now the solid state era is here, with adaptations that greatly expand the whole art of musical instrument design. .. . The computer industry has been providing the impetus to this rapid development, and now it appears that other industries will be requiring answers in the linear applications as well as in the digital areas. With such universal interest in MOSFET LSI [Metal Oxide Semiconductor Field-Effect Transistor Large-Scale Integrated Circuit], the musical instrument industry, which is small compared to other industries can only benefit by adapting to this technology. Academics sensed the wider import of what was happening. Hubert Howe, Professor of

Music, Queens College, N.Y. put it this way during his presentation at the Audio

Society's 39thconference in October 1970:

The most important musical result of the revolution of the electronic method of sound generation has been a reevaluation of the fundamental ways in which music has been thought about.

The AES worked to fill out the connections of Latour's (1 999) 'colleagues' circuit. Peer reviewed articles in the society's journal acted as immutable mobiles while the conventions themselves further enabled the process of 'autonomization'. As the shifting session titles reveal, however, autonomy must be negotiated. What else was going on? Engineers were not only engaged in mapping the frontier of their field, via discrete inscriptions. Discursively they sought to mobilize the entire discipline, to make it work with the ongoing advances in the ICT paradigm's key technology. Between 1969 and 1972, conventional wisdom within the EM1 section of the AES proclaimed that radical technological change, derived from exogenous sources, was afoot. Yet amidst this transition, much uncertainty prevailed and despite the public pronouncement that the instrument industry would reap the benefits of the space age, hesitancy would characterize the response of American firms when they confronted the specific technologies that actually moved instrument design into the digital era. It is worth mentioning briefly, and in specific reference to Schrecongost's comments about the

MOSFET LSI, that in 1972, Yamaha was already producing its own application specific

LSI, a fact that was entirely overlooked by the US industry (Nakagawa 1984, Yamaha

1987). 3.3.5 Analog Days

Electronic music remained largely disconnected from the public 'circuit' until the late

1960s. The wave of popularity began with the Beach Boys' use of Theremin in their 1966 song Good Vibrations. A few years later, Wendy Carlos' Switched on Bach (1969), performed entirely on the new Moog modular synthesizer, translated electronic music to an audience of classical and pop fans. This era, which lasted until the mid 1970s, marked the high tide of the analog synthesizer, an instrument most characteristically identified with the sound of the emerging genre of art rock (e.g. Emerson, Lake and Palmer,

Tangerine Dream, and so on). It can also be regarded, aesthetically at least, as the high water mark of the US industry. During this time-frame, the cohort of young Turks led by

Bob Moog and Don Buchla eclipsed the influence of the large research laboratory as the domain of EM1 instrument design. The inventor-entrepreneurs of this period were a generation removed from the typical staff at RCA, and consequently embedded in a much more diverse set of linkages. Buchla shared close ties with the avant-garde, working with composers such as Morton Subotnick at the San Francisco Tape Music Centre. In his spare time Buchla supplied the synthesizer and served as resident technician for the series of moments, the so-called Acid Tests, orchestrated by Ken Kesey and his Merry

Pranksters around the San Francisco Bay Area in 1965-66 (Wolf 1969, Pinch and Trocco

2002). Similarly, early orders for Moog synthesizers were received from The Rolling

Stones, George Martin, and Jimmy Hendrix. Moog who claims that, "artist feedback drove all my development work" (Houston 2000), also had roots in the avant-garde, collaborated closely on product design with other key early adopters, such as Wendy

Carlos and Keith Emerson. Table 3.5 presents a chronology of the venture that Moog started. Moog's post- secondary education, which culminated in a PhD in engineering physics from Cornell in

1965, was funded by selling Theremin kits. Moog induced demand for these $50 kits by Table 3.5: A Chronology of Moog Music rnYear Firm Owner ~ocationl Comments RAMCO Moog Flushing Designs first Moog Theremin R.A.Moog Moog Flushing Produces vacuum tube Theremin kits. Publishes Co. article: 'The Theremin' in Radio and Television News R.A.Moog Moog lthaca Produces transistorized Theremin kits. Publishes Co. article: 'A Transistorized Theremin in Electronics World Trumansburg Moog moves to Trumansburg. Collaborates with Herb Deutsch on synthesizer development. Produces first synthesizer. Publishes article: 'Voltage Controlled Electronic Instruments' in Journal of the Audio Enaineerina Society Produces synthesizer modules and custom modular systems. Composers gather for concert at the factory. Establishes Modular System models 1, 2, and 3. Produces Demo Record: Moog 900 Series Electronic Music Systems. Begins quarterly publication of Electronic Music Review. R.A.Moog Moog Trumansburg Re-defines Modular System models 1,2, 3. Wendy Inc. Carlos' "Switched on Bach" album released. Modular system sales peak. Begin development of performance (portable) synthesizer (Minimoog). Produces Minimoog A,B, C prototypes. Releases Minimoog model D. Employee Dave VanKoevering initiates retail demonstrations. Moog Waytena Williamsville Sale of company, move to Williamsville in Musonics November. Last Minimoog shipped from Trumansburg serial number 1210. Moog Waytena Williamsville Release Sonic 6, Satellite, Modular system "IT, Music Inc. "921" Oscillators. Dave VanKoevering joins company as VP Marketing. Dave Luce joins company and begins development of polyphonic synthesizer. Moog Norlin Williamsville Sale of company. Minimoog production stops at sln Music 6920. Release Taurus and Polymoog Cheektowaga Move to Cheektowaga, Walden Ave. Bob Moog leaves the company. Re-start Minimoog production. Releases Multimoog, Polymoog Keyboard, Vocoder. Releases Liberation, Opus 3, Prodigy, Rogue, Taurus 11. Produces Gibson, SG, Cordovox, Maestro lines. Design and produce Lab Series amplifiers. Begins Contract (non-music) Manufacturing. Last Minimoog produced January, sln 13180. Release The Source and Memorymoog. Luce & Cheektowaga Sale of company to management. Emphasis on Chapman contract manufacturing. Attempt to enter into telecommunications market with a microprocessor- controlled feature telephone, initially named "Telesys 3",later "The Operator". Designs and manufactures "Song Producer". Develops "SL8" and builds prototype. Maintains repair operations for Moog products. Produces the "SSK Concertmate" synthesizer for toy manufacturer Tandy Corp. 1987 Moog EJE Cheektowaga Sale of company Move to French Road, Music Cheektowaga, NY. Primary focus is on contract manufacturing, no proprietary products. Continues service of Moog Music products. ETG Jamestown EJE owners form public company: ETG. Moog Music service work transferred to facility in Jamestown, NY. Sale of company. Moog Music ceases operation. Company assets warehoused, later sold at auction. I I Jotes: 1 : All locations in upstate 1 :w York [Source: Table by author, based on content provided by httv://www.moogarchives.com, obtained 1010 1/04] publishing an article that explained the principles behind and the way to construct a

Theremin in Electronics World in 1961. That year he sold 1000 units. Flush from this success, he opened up a shop near his university in Trumansburg, New York, and in collaboration with composer and Hofstra University Professor Herbert ~eutsch~~devised the first voltage controlled oscillator (VCO) synthesizer. This instrument was revolutionary because, unlike the RCA and Bell Labs synthesizers, it was operated by a keyboard interface and had a fixed, equally tempered tuning whereby one volt was equal to one octave. Moog introduced his prototype at the Audio Engineering Society's annual meeting in 1964 and published a paper in the Society's Journal the following year. The principal products develop by Moog until 1970 were large modular (multiple components linked by patch cords) synthesizers, many of which were customized and sold to an elite rank of professional musicians and universities. At their peak of modular production in

1969 Moog's 42 employees had an output of two or three per week (Freff 1988: 55).

Real-time live performance was possible if difficult with these instruments, and it was not until the introduction of the portable Minimoog in 1971 that sales figures really climbed as their price became reasonable for amateur musicians. According to Moog, "we'd been getting requests from studio musicians asking us to pack all this stuff into a nice package they could carry it to gigs with them, and we'd done that, but we had no way of selling it.

Music stores didn't carry synthesizers" (ibid: 56).

To compound matters, the US economy worsened during the 1970s, so the situation of backlogs during the modular phase (1 966-70) devolved rapidly into a glut of

46 Deutsch's Synthesis: An Introduction to Electronic Music was first published in 1974. Roland's founder Ikutaro Kakehashi translated and published this volume in Japanese in 1979. Deutsch went on to become Marketing and Product Consultant and a Clinician at Roland (Kakehashi 2002, htt~://www.hofstra.edu!academics/hclas/music/musicdeutsch.cfin; accessed 25/06/04). inventory. Scant sales, accumulating bills and no capital put the firm in a difficult situation, not just with respect to the short term, but also for its ability to invest in innovation for the future. Moog (quoted in Milano 1988:42) summarized how capital scarcity stunted innovation at Moog and a generation of US companies including ARP,

Sequential and Oberheim: "you're living from one NAMM [North American Music

Merchandise Association] show to the next. If you don't have a hit at one show, you'd better at the next or you're dead."

The bottom half of table 3.5 chronicles the demise of Moog. While the Moog brand identity remained in the public eye, behind the scenes the reality was one of corporate turmoil. In 1971, Bill Waytena, an individual who specialized in the buy-out of distressed firms, purchased Moog before flipping its control to Norlin in 1973. At the time Norlin was aspiring to become the Yarnaha-like one-stop-shop of the US industry - the Moog acquisition followed the purchase of Lowery Organs, Gibson Guitars and

Selmer Woodwinds. Norlin, a conglomerate previously known as the Ecuadorian

Corporation Limited (ECL) flew a Panamanian 'flag of convenience' and was registered in Delaware. Even with its late foray into the musical instrument industry, which began in

1969, Norlin's biggest source of profit came from a brewery in Panama. Former Moog employee and current Berklee College of Music professor Tom Rhea assessed the souring

Moog-Norlin relationship by stating,

Norlin's marketing people did not come from the music business. They were French's mustard, Gabriel shock absorbers, Beatrice foods, that sort of thing. They were all bright men, but they didn't have a handle on the music industry. Musicians accept new ideas very slowly. Moog was on the leading edge, and Norlin didn't understand that they had an education problem, not a distribution problem (Quoted in Vail 1993: 18). This quotation points to a fundamental flaw in Norlin's ability to enroll new actors (in this case, musicians - the 'public') in their network. Bob Moog lasted a few more years at this organization before leaving the firm in 1977. The company that remained sold products bearing the Moog label for a few more years until the dissolution of the company in 1983.~~

Similar to Moog, Buchla's attempt at the commercialization of his instruments depended upon the financial and production assistance of a major corporation, in this case

Columbia Broadcasting Studios (CBS), which itself had been rapidly acquiring and mismanaging a whole range of instrument manufacturers, including Steinway Pianos, and the Fender Guitar Company. As with Moog music, conglomeration did not help Buchla's fortunes. The deal soured in 1969 due to differences in vision. In particular, Buchla intended to remain at the leading edge of experimental sound, a commitment requiring large resources for R&D which CBS was unwilling to finance. This rejection likely prompted the narrowing of Buchla's focus towards a niche market of experimental musicians which was beyond popular tastes. To put these circumstances in perspective, around the time that CBS was vetoing Buchla's long term research plans, Yamaha was dispatching their engineers to the US to learn technologies that would define the next 20 years of music (Yamaha 1987).

The acquisition of Moog and Buchla by conglomerates at the turn of the 1970s needs to be contextualized as part of a much broader pattern of US managerial 'best practice' that was fixated on growth through the rampant consumption of unrelated

47 It is tempting to ascribe some part of Moog Music's decline to its relocation from Trumansburg to an old gelatin factory in suburban Buffalo, heart of the rust belt. Following later rounds of restructuring, the local press announced that "foreign competition...muffled the Moog synthesizer" (The Buflalo News May 30", 1987, C8) assets. Indeed, Bluestone and Harrison (1982: 41) point out that "during the 1970s, two out of every three new Fortune 500 manufacturing plants were actually not 'new' at all, but rather were acquired from other owners." These authors condemn the effect conglomeration had on the subsequent round of plant closures and layoffs. For the musical instrument industry, conglomeration produced a myopic attitude towards innovation; if not in the engineering departments, then in the boardrooms.

In comparing Moog and Buchla, we can say that, as entrepreneurs, operated in quite different networks than the generation of engineers who preceded them. Positively, by enrolling leading musicians they initiated the circuit with the public. Negatively, they fell into the orbits of the wrong kinds of allies, conglomerates. Moreover, both engineers had to rely on radical performances to win over colleagues at the AES. Geographically,

Moog and Buchla displaced the vanguard of practice from the institutional laboratory to the workshop. However, a technological branching point served to separate the fates of

Moog and Buchla. Pinch and Trocco (2002:309), suggest that Moog's early commercial success owes to his invention of and commitment to the volt-per-octave standard which,

"embedided] into his technology a piece of existing culture - the idea that music is about intervals. By defining octaves, the Moog preordained the keyboard as the controller of the synth." Buchla's design, on the other hand, had no such tuning standard nor was it operated by a keyboard, two features which consequently isolated Buchla's influence to only the vanguard of electronic music. Buchla's location on the West Coast did, however, prefigure the setting of the last wave of American EM1 manufacturers before the shift in industrial location to Japan in the 1980s. 3.3.6 Towards the Digital Era

By the mid 1970s the exponential increase in computer processing capability predicted by

Moore's ~aw~~had permeated the musical instrument industry. The typical actors at this stage were young, inventor-entrepreneurs, such as Dave Smith, founder of Sequential

Circuits in Sunnyvale, CA and Tom Oberheim in Los Angeles, who both started their firms out of garages. Smith's degree in computer science and electrical engineering from

UC Berkeley and experience in the Bay Area's aerospace (Lockheed, General Electric) and computer sectors (Standard Microsystems, Signetics, Diablo Systems), provided him with a solid basis in microprocessor technology. He was thus attuned to the recent availability of application specific synthesizer ICs produced by Solid State Music, another Silicon Valley start-up. The advent of microprocessors enabled the generation of a complete synthesizer voice at a fraction of the cost of vacuum tube oscillators. "The technology seemed obvious. I thought everybody, all the East Coast firms would be using this stuff so I was surprised to find out I was the only one" (Personal Communication,

1011 1/02). Sequential's Prophet 5 synthesizer, released in 1978 became the first completely programmable, polyphonic (multi-voice - capable of playing chords) keyboard. In 1979, Sequential had over 100 full-time employees and a 30,000 square foot production facility49. Sequential's generation of firms, which also included Linn

Electronics and Kurzweil, is notable in that it constituted the first set of American manufacturers to establish relations with their Japanese counterparts. In the following

48 Moore's Law: In 1965, in a paper published in the journal Electronics , the lead engineer at Fairchild Semiconductor, Gordon Moore predicted an exponential growth in the number of transistors per integrated circuit, prophesizing that chip density would double every eighteen months. Moore later became the president of Intel and his prediction has more or less held. (http://www.intel.corn/research/silicon/m ; accessed 26/06/04) 49 Prior to 1978, Sequential contracted out the production of synthesizers such as the Fugue to larger manufacturers including SEIL, an Italian firm which was eventually absorbed by the Japanese fm Roland. chapter I will relate the details of how many of these US firms, or at least the intellectual capital contained within these companies, was absorbed by the wave of acquisitions by

Japanese manufacturers in the mid-to-late 1980s. Prior to these events in the early 1980s,

American and Japanese firms collaborated to overcome a technological hurdle that, in the long run would prove pivotal in determining the institutional architecture of the industry.

3.3.7 MIDI [Musical Instrument Digital Interface]

The defining event in the collaboration between US and Japanese firms was the establishment of the Musical Instrument Digital Interface (MIDI) Protocol in 1983. MIDI was initially conceived as a hardwarelsoftware specification that would allow keyboards and other EM1 manufactured by different firms to be connected to one another. It also allowed computers to share data and interface with digital EMI. At the time, digital interface standards already existed in the computer market as diverse components, such as modems, printers, disks and cables of various brands, could all be used with each other, so the rationale held that if the EM1 market was to grow, then a similar type of standard was imperative.

MIDI'S genesis marked a critical threshold in defining the shape of the industry in the digital age. Theberge (1997) suggests this event is significant for several reasons.

First, the establishment of a protocol required a high degree of negotiation and cooperation amongst participants. Since MIDI was being introduced during a period of rapid technological change, for the protocol to work, competitors had to disclose proprietary technical details concerning present and projected products. Second, the non- proprietary nature of MIDI represented an important institutional innovation. "Without the initial joint decision that no one should profit directly from the development of MIDI, the necessary trust, cooperation, and good will between the participating companies could not have been guaranteed" (Theberge 1997: 86). Third, MIDI as a formal convention proscribed the form of the subsequent technological trajectory and thus served as a powerful selection device that favored adopters (particularly early ones) while punishing laggards and the non-compliant. This issue of compatibility was exceedingly important since at the time, the pace of technological change was perceived by all to be very rapid, and consumers in particular were concerned about product obsolescence. Until this point,

Loy (1985:20) characterizes the industry as being vertically integrated so that at the retail end, manufacturers often relied on, "sales of one item in their product line carrying a package of sale." With compatibility, MIDI facilitated company specialization and the horizontal integration of the industry. Indeed the latter half of the 1980s was characterized by a series of acquisitions. Finally, these moments of collaboration and discussion unfolded in quite specific places, in particular a succession of meetings such as the AES and NAMM conventions between 198 1 and 1983.

It is generally agreed that Roland and Sequential Circuits played the lead roles in orchestrating MIDI (Theberge 1997, Chadabe 1997). The idea of a protocol had been tossed about between Roland and the American firms Sequential Circuits and Oberheim.

In October 1981 a meeting was called between these three and three other Japanese producers: Yamaha, Kawai and Korg. The following month, Sequential's Dave Smith talked up the idea at an Audio Engineering Society Convention. This performance must have piqued some interest, for at the NAMM show in January 1982, 10-15 firms participated. Yet apart from the Japanese firms and Sequential, nobody had any interest in following up on the idea. According to Smith, No one in the States seemed to be interested anymore, and we lost interest trying to round everyone up, so we worked with the Japanese companies. At that point we started going back and forth. The Japanese made a lot of suggestion. I think that Roland did most of the work. They did most of the coordinating in Japan (quoted in Chadabe 1997: 195)

The US entrepreneur synthesizer maker Tom Oberheim, an early proponent, chose to

drop out at this stage due to a rift with the others over the degree of sophistication of the

interface. He wanted it to be a parallel interface, one more suited to the high-end professional market, whereas Roland's Kakehashi Ikutaro and the others opted for the

serial interface which was deemed fast enough for most consumers, either professional or

amateur. By January 1983 the remaining collaborators had a working specification, and the key moment of unveiling took place at that month's NAMM when Roland's JP-6

machine was hooked-up to a Sequential Prophet 600, and the two machines talked to

each other. As Smith recalls, "you could play the keyboard on one and the others would

play right along" (Ibid: 196). After a few more meetings that took place in Japan, a

version of MIDI was released to the public in August 1983. Chadabe (1 997) suggests that

the MIDI version 1.0 constituted a necessary compromise between cost, performance,

market preferences and the myriad of possible uses imagined by musicians. Nonetheless,

with the advent of MIDI a single performer could synchronize any number of devices

from different manufacturers. This newfound flexibility allowed musicians to build up

their own custom systems in an incremental fashion as their budget allowed and without

fear of the obsolescence of one component harming the ensemble's integrity.

A number of formal industry groups were born at this time, the first of which was

the Japan MIDI Society, which, according to Kakehashi functioned, "to exchange ideas,

collect information, form working groups, check future possibilities, keep the standard" (quoted in Chadabe 1997: 197). In the US, the International MIDI Association was created to perform communications and other publicity functions while the MIDI

Manufacturers Association was established to handle technical problems.

It is hard not to see the establishment of MIDI as one of the critical pivot points by which Japanese manufacturers, through cooperation, gained a competitive advantage over their US counterparts. The MIDI protocol was like the 'knots and linkages' at the centre of Latour's (1999) circuit diagram (Figure 2.1). It literally tied the industry together. Moreover, it translated the interests of manufacturers ('allies') with musicians

(the 'public'). It was therefore in the circuit of colleagues where things went amiss.

Indeed, Dave Smith sought to enroll his US peers by giving presentations at the NAMM and AES conventions. Their hesitancy points to an inability to connect the 'colleague' circuit and it is at that point when the network on the US side fell apart. In Japan, on the other hand the story was different. After all, the Japanese manufacturers, three of whom were neighbors and rivals in Harnamatsu had come together to cooperate (explained in

Chapter Four). Moreover they had initiated contact with the leading American firms and thus had a good idea of the state of the art in that country. On the other hand the US industry was still fragmented, secretive and continued to have a propensity to perceive their principal market as the professional musician. In MIDI and in EM1 technology in general, Japan was clearly setting the standard.

3.3.8 Legends of the Fall

In 1988, Keyboard magazine published an article by Dominic Milano (1988: 43) entitled

'American Synthesizer Builders: Triumphs and Crises for an Industry in Transition', that offered the following requiem for the rise and fall of the EM1 industry in the USA: America happens to be the land of the synthesizer - its birthplace divided between East Coast and West Coasts. Its fathers, Don Buchla and Bob Moog, were dreamers whose workbenches full of circuit boards turned the world music community on its collective ear and spawned a two-billion- dollar-a-year electronic music industry. That industry is now a quarter- century old. And many of Moog's and Buchla's offspring have come and gone, victims of the things dreamers tend to be victims of. Bad business sense. Bad timing. Bad luck. In one way or another, the money stopped and so did their companies.

This article literally wraps itself in the flag - the text is accompanied by an image depicting the knobs and keys of a synthesizer as the 'stars and stripes'50. Factoring in the patriotic, evolutionary and paternal tropes that pepper the above quotation, Milano is hardly suggesting that the demise of the US industry relates to bad genes. As he implies, it is not that the industry could not reproduce itself, but rather its inability to connect that mattered. The three 'bads' further allude to the highly contingent aspects of failure. 'Bad timing' and 'luck' are contingencies of fate, while 'bad business sense' is only factor tying agency and circumstance together. In short, like all business failures, 'the money stopped'. My point is that it 'stopped' in a variety of circumstances for largely similar reasons: specifically a 'short circuiting' of allies and customers.

In the article, Milano quotes various 'builders' as they pick over the remains of their industry. A recurring theme is the notions that under-financing killed the US industry.

Norlin did not spend enough money on product development. They did not have deep enough pockets (ibid: 44-5; Former Moog and Norlin employee Tom Rhea).

50 Theberge (1997: Chapter 5) discusses the way that the music magazine industry brought the interests of musicians and manufacturers together in a way that is sympathetic to Latour's (1999) concept of 'allies'.

135 We weren't capitalized enough. I didn't know business well enough. When I was engineering products myself, things went well, but as soon as I tried to get bigger we got into trouble (ibid: 45; Electronic drum designer Roger Linn).

We had never taken outside investments, so we never had the luxury of getting any breathing room (ibid: 45; Sequential Circuits founder Dave Smith).

We were under financed. I don't think foreign competition hurt us (ibid: 45 Oberheim founder Tom Oberheim).

Another theme, discussed earlier with respect to Norlin who, 'had an education problem', concerns the weak link with the 'public'. Scott Wedge, founder of synthesizer manufacturer EMUtalked about the failure of firms such as Moog as follows: "Company size is a red herring. The issue is the relationship between a company and its customers"

(ibid: 47).

3.4 Conclusion

By the early 1980s' the post-war developments in the US EM1 industry had run their course and were on the threshold of being eclipsed by Japanese competition. Independent of this threat, however, the community of practice engaged in the problem of instrument design was cursed with various problems that prevented it from sustaining its initial advantage. This advantage was won in the 1950s and early 1960s. During these decades, the cores of US R&D, Bell Labs and RCA, as the primary transectorial gateways, produced much of the basic research that would educate later generations of inventors.

Yet the US advantage was squandered in the 1970s by the mercenary interests of companies like Norlin. Conglomeration proved a poor business model for facilitating research in acquired firms. Finally, the failure for US firms to collaborate on MIDI proved their independent-mindedness, and subsequently their downfall, while the

Japanese, using a more collaborative business model realized the strategic advantages afford by connectivity.

Geographically, the US EM1 industry evolved in a manner that reflected broader economic patterns of development. In this regard, Bell Labs and RCA on the east coast of the USA were the fonts for a host of key technologies such as the transistor that would define the ICT paradigm. Similarly, the garage operations of the California entrepreneurs are in keeping with other Silicon Valley innovation stories like the Homebrew Computer

Club, which spawned Apple Corporation. At another level, the AES, served as an important national institution for the diffusion of transectorial knowledge. Within this body, engineering discourse spoke of a new age, even suggesting the technologies such as MOSFET which would bring the community to the 'promised land'. Yet repeatedly, firms, the institutions that possessed the power to enact, as opposed to facilitate, technological change failed in their strategies to realize the scope of the ICT paradigm.

Independent of the threat posed by Japanese competition, US industry could simply not maintain the network of heterogeneous interests that was so central t the early trajectory of the industry. CHAPTER FOUR: THE MUSICAL INSTRUMENT INDUSTRY IN HAMAMATSU JAPAN: CORPORATE STRATEGY, STRUCTURE AND RIVALRY 1880-2000

4.1 Introduction

This chapter sketches the development of the acoustic and electronic musical instrument industries in Hamamatsu, Japan over the last century. Specifically, it presents the technological strategies of Hamamatsu's three principal firms (Yamaha, Roland and

Kawai) as 'longitudinal profiles' (Laulajainen 1982). Though the chapter is primarily concerned with organizational evolutions, I narrate these trajectories by framing product development, and in particular the process of transectorial diversification, through the lens of corporate decision making. From this perspective, innovation is interpreted as a social process that involves persuasion and compromise between different levels and functional categories of the firm, as well as between the firm and various other actors.

The focus at this stage is to consider firm trajectories according to the agency of corporate leaders and entrepreneurs. Taken in concert, these profiles provide the initial point of reference from which I can begin to interpret the knowledge geography that underlies Hamamatsu's emergence as an industrial 'core' region. Later chapters, six and seven in particular, add depth to the narrative by providing a closer reading of the translation of foreign technology to Hamamatsu from the standpoint of American and

Japanese engineers. Geographically, the biggest difference between the last chapter and this one is that, while the former discussed many places, the present chapter is concerned with one place. Framed as networks, the US industry was poly-nucleated, a succession of shifting cores, whereas the Japanese musical instrument industry has been concentrated in

Hamamatsu for over a century. Moreover, as its firms have become world leaders, the world has come to revolve around Hamamatsu in the sense that a global network of heterogeneous interests come together in Hamamatsu. Production and knowledge networks, networks of domination and surveillance (Law and Hetherington 2000: 39) centre on Hamamatsu. A series of maps will illustrate, at the local scale, an industrial morphology that is highly clustered. Indeed a broad and deep social division of labour centred on the three leading firms suggests that the sheer sum of these parts is massive and significant. However to fully appreciate the nature of this specialization, Hamamatsu needs to be viewed as a kind of regional Gestalt, where the whole is greater than the sum of its constituent parts. This aspect is less easy to map, except by referring to the type of circulatory system of interlocking interests outlined by Latour (Figure 2.1).

The longitudinal profiles discussed in this chapter may also be viewed as forms of biography and this relates to the nature of my sources. These include stories told by the firms themselves as corporate histories (Yamaha 1987, Kawai 1997), various tables charts provided by these same firms which represent technological developments in an explicitly longitudinal manner, autobiographies of firm founders (Kakehashi 2002), books written by leading engineers on the state of their discipline (Mochida and Aoki

1994) and other interpretive accounts (Nakagawa 1994), including research papers published by geographers (Ohtsuka 1980, Takeo 1988). These secondary sources are almost all in Japanese and have been translated by the author, save for Kakehashi's autobiography which is in English. Following Barnes' (200 1: 426) reminder that, "if science studies are to use biographical accounts then their nature must be recognized", the various authorial positions need to be considered. Thus, for instance, firm histories frame the enterprise as a master of its environment. In doing so, the authors privilege certain voices and perspectives within the firm over others in order to impress a coherent narrative structure to a rather complex (and possibly contested) story. Similarly, autobiographies are often, "structured to present the person and his story and how the self is fashioned through this narrative approach" (Schoenberger 2001 : 277). As for the tables and diagrams that firms use to represent technological developments in schematic form, do their structures yield similar insight concerning the construction of history? The answer is yes, for tables impose an order on events and are designed to put everything in its place, to represent contingent relations by way of a grid. Finally, the maps by Japanese geographers construct geographies of phenomena such as 'spatial clustering' that demand further scrutiny to determine why proximate points on a map might be indicative of processes such as external economies.

The chapter proceeds as follows: First, the context for my discussion of the musical instrument industry by painting a general portrait of Hamamatsu as an industrial city. I then explain the origins of the musical instrument industry in Hamamatsu.

Industrial learning in this period is marked by the reverse engineering of Western instruments while the local structure of industry gradually evolves via the emergence of spin-offs like Kawai from their parent firms. Next, I discuss the set of strategies employed by Yarnaha and Kawai to engineer the mass production and consumption of pianos and organs in the post-war era. The early post-war period is also significant for its pattern of new firm formation by spin-offs. Third, I narrate the developmental trajectory of Yamaha's efforts to become proficient in the production of electronic musical instruments - a project capped by that firm's decision to manufacture its own microchips.

I next juxtapose Yamaha's pattern of development with those of its local rivals. It is at this stage where the story of Roland, a specialist in electronic instruments that now leads the world in this category largely by virtue of the efforts of its founder Kakehashi Ikutaro, is used to illustrate the considerable barriers to entry posed by Yamaha's sectoral leadership. Finally, I pick up where the last chapter ended by discussing the pattern of acquisitions of American firms by Hamamatsu's 'big three' in the period of the late

1980s after the establishment of the MIDI protocol.

4.1.1 Vignette: ICT, Yamaha, Roland and Hamamatsu

Each time I turn on my computer, just as the operating system finishes booting, 1 am greeted by the Microsoft anthem. The melody is performed by a sound chip embedded in the hardware of my computer. This device, a tiny synthesizer processing digital signals

into a polyphonic soundscape, functions as my computer's electronic larynx. The control panel menu indicates that the default sound and audio device is a Yamaha Audio chip.

The AC-XG WDM animates my Microsoft GS Wavetable Synthesis software package

that features a Roland Sound Canvas Sample Set. Every sound effect that the computer

makes draws from a palette of audio samples furnished by Roland that are reproduced via

a Yamaha sound chip.

Neighbors on my computer (Figure 4.1), Yamaha and Roland, the two largest

manufacturers of electronic musical instruments in world, are co-located in the city of Hamamatsu, Japan. If the industrial scope is broadened to include all musical instruments, Hamamatsu-based firms still sit 1-2-3 atop the global league tables (The

Music Trades 2000).

Figu

Cumaauk,mdena

The WdaudiomAFdd 6bMBHLI

Source: screen image captured by the author, by pre-permission Microsoft.com.

Kawai, Yamaha's long-standing rival in the piano market, occupies the number two position. There are few industries in the world that are this top-heavy in terms of their spatial concentration of control, and on this basis alone Harnamatsu instrument makers deserve attention for their economic geography. Each of these firms still retains a strong footing in their home region and it is this fierce local rivalry in Hamamatsu which fuels their global competition. In 1996, musical instrument manufacturing accounted for 9.7 percent 0290 billion - approx. $US 856 million") of Hamamatsu's economic output and

-- -

51 Assuming an exchange rate of 1 USD = 105 JPY Sea of Japan

Figure 4.2: Map of Japan showing the location of Hamamatsu (Source: Map drawn by author) employed 10.5 percent (832 1 people) of its labour force (Hamamatsu-shi 1997). This cluster could not have emerged and persisted had not firms like Yamaha and Roland so deftly negotiated the digital turn that proved fateful for former core regions like the US.

4.2 Hamamatsu: Industrial City

Hamamatsu, a city of 589'21 852,is located in western Shizuoka prefecture about halfway between Tokyo and Osaka (Figures 4.2'4.3). Some of this city's regional identity derives from its situation. Indeed, Hamamatsu's connectivity with Japan's two major

Metropolitan regions (Kanto and Kansai respectively) dates from the Edo era (1 603-

1867) when the Arai checkpoint, located just west of Lake Harnana, regulated traffic on the Tokaido Road.

52 Population as of January 1,2001 (Source: Hamamatsu City's world-wide web page: http://www.city.hamamatsu.shizuoka.jp/ham. Accessed: 26/06/04) Figure 4.3: Map of Hamamatsu City showing location of highways and railways and other features

Today, a clue to its relative position in the national urban hierarchy is the

frequency at which the Shinkansen bullet train stops at Hamamatsu station - roughly five

times an hour. That the Hikari, the second fastest grade of train pulls into Hamamatsu

station every half hour further alludes to a commercial/industria1 importance far greater

than its population would suggest. Yet despite this connectivity, it is almost as though

the influences of these two metropoles have cancelled each other out at the mid-point.

"Yaramaiku ", a phase in the local dialect that translates as "challenge everything" is

often cited anecdotally as typifying the local attitude vis-a-vis the received wisdoms of

elsewhere53.Yaramaiku hints at Hamamatsu's sense of place.

53 I thank Professor Takeuchi Atsuhiko for drawing this term to my attention.

145 I i 4.2.1 Entrepreneurial Foundations and Transectorial Legacies

Table 4.2: Hamamatsu's industrial pioneers

Entrepreneur Organization Principal Comments (Date of Sector(s) Incorporation) Torakusu Nippon Gakki Musical In 1884, Yamaha repaired an imported Yamaha (1889) > Yamaha lnstruments > organ in Hamamatsu school while working Motorcycles > LSI, as a traveling engineer who serviced (1851-1916) etc. medical equipment.

Koichi Kawai Kawai (1927) Musical Apprenticed to Torakusu Yamaha at age 11 lnstruments and became head of piano action section (1886-1 955) before founding his own firm.

Sakichi Toyoda Toyoda Power Textile Machinery Held 141 patents inventions of looming Looms (1896) > > Automobiles technology of successively sophisticated (1867-1 930) Toyota Motor Co. looms and contributed to the growth of textile and machine industries in the region. His son Kiichiro established Toyota Motor Co. in nearby eastern Aichi prefecture. - -- Suzuki Michio Suzuki Loomworks Textile Machinery Held 120 patents for looms and ancillary (1909) > Suzuki > Automobiles > machinery before steering Suzuki into (1887-1 982) Motor Co. Motorcycles vehicle manufacturing. Honda Soichiro Honda Motor Co. Motorcycles > The founder of Honda established Hamamatsu as "the ctty of motorcycles" (1906-1991) (1946) Automobiles Takaganaki Shizuoka Television A Professor of Engineering at Shizuoka Kenjiro University, Faculty University, Takayanagi invented the first of Engineering domestic TV set and the world's first all- (1899-1 990) (1946) electronic TV. Considered a 'father' of Hamamatsu's optical technological industry. ource: Hamamatsu L httu://hamamatsudaisuki.net/enalish~indus/to.htmlAccessed: 26/06/04

The Hamamatsu region boasts an impressive list of entrepreneurs who have played a major role in modern Japanese industrialization. Table 4.1 contains biographical and other details of six founding 'heroes' of local industry. Almost all of these entrepreneurs started companies that are household names around the world and, in doing so, set the standard and determined the latter-day composition of Hamamatsu's industry.

After Kanto (Tokyo, Yokohama), Nagoya and Osaka, the Hamamatsu region ranks as the fourth largest industrial region in terms of the value of manufactured shipments (Takeuchi 1996: 192). Hamamatsu's prefecture, Shizuoka, leads the nation in manufacturing output for a number of sectors (Table 4.2).

Table 4.2: Manufactured Goods Produced in Shizuoka Prefecture Ranking First in Japan Product Share of National Production I Pianos (2000) I 100% I I Canned Tuna (1999) I 88% I Electronic musical instruments (2000) 85.1% Plastic Models (2000) 78.1% I All musical instruments (1998) I 70% I I Compact Disks (1999) I 64% I I Motorbikes (2002) I 61.6% I Car Headlights (1999) 60% Green Tea (1999) 60% Cardboard (1999) 54% Lumber Machines (1999) 53% I Doll Parts and Accessories (1999) I 43% I I Light four wheel cars (2002) I 36.3% I I Mobile Telephones (1998) I 23% I I I I Sources: MET1 1998, 1999, 2000, 2002 "Industrial statistics table itemized version" Quoted in various Shizuoka Prefecture world wide web pages: http://~~~.pref.shizuoka.ip/-syoukou/svo-13Olindustrvlindustrvlindustrv04e.html http://www.pref.shizuoka.ip/kikakuki-20/englishl~ride/index.htm

Hamamatsu region has produced a set of enterprises with an almost singular ability to realize economies of scope via technological diversification. A number of authors have raised this point in various ways (Ohtsuka 1980, Takeo 1988, Yamashita 1991, Takeuchi

1996). Ohtsuka (1 980: 169) captures the evolutionary trajectory of industry in

Hamamatsu when he notes that,

these industries have never developed separately at all, but ...have developed in close relationship with one another,. ..the preceding industry in this region has functioned as the leading industry for the birth of a new one. 4.3 Hamamatsu's Musical Instrument Industry

4.3.1 The Origin and Pre-war History of Musical Instrument Manufacturing in Hamamatsu

The history of musical instrument manufacturing in Hamamatsu began in 1887 when

Yamaha Torakusu established the Yamaha Organ The founding of Yamaha in Hamamatsu choice was entirely contingent on what Lee (2000) calls 'practice and instance7.By this I mean that Yamaha Torakusu's practical repair skills, which had been honed in the medical instrument business, were translated to the task of fixing the organ at Hamamatsu's school. Until this point, Yamaha's career was literally on the move, and this moment forever settled him in ~amarnatsu'~.The company began in earnest the following year employing seven craftsmen. By 1889, 100 employees were turning out

122 organs, 78 of which were destined for export to Europe. Ten years later, the firm's capital had trebled to Y10,000 while organ production had climbed to 397 units per annum (Ohtsuka 1980). Ln 1900, Yamaha, which from 1897 was known as the Nippon

Gakki Company (Japan Musical Instrument Company), commenced production of pianos.

By 1932, Nippon Gakki employed 2050 persons, 1800 of whom were factory workers.

That year the company produced 4000 pianos, 20,000 organs, 240,000 mouth harmonicas

and 369,000 square metres of veneer. Contracts from the government were also awarded

for the manufacture of airplane propellers (Loesser 1954: 599).

Ohtsuka (1980) cites the local availability of capital and technique as key factors

which contributed to Yamaha's early success in Hamamatsu. Important sources of capital

54 ~t that time reed organs in Japan were known asfukin (RT)which is composed of the kanji for wind and koto, a Japanese zither. This term has largely disappeared from the lexicon, replaced by the katakana- ized orugan (&/L/?r'>). Counterfactually, there is no telling whether Yamaha might have been founded elsewhere had an organ r -. in that place broken down. included the tea and textile industries which had been leading industries since the Edo

period. In addition, a labour pool already skilled in woodworking could be drawn from

engravers located in Hamamatsu city and shipbuilders based 1Okm to the east in

Kakezuka, the old harbour of the Hamamatsu clan located at the mouth of the Tenryu

River. In Latour's (1987, 1999) terms, Yamaha enrolled these different interests towards

his project.

Following the sudden death of founder Yamaha Torakusu in 1916, the ensuing

decade unfolded as a succession of crises for Yamaha including, a fire, an earthquake and

a 105 day strike (Yamaha 1987). On the verge of bankruptcy, Kawakami Kaichi was

installed as president in May 1927. Kawakami received his scholastic training at the

University of Tokyo, finishing top of his class in physics. Not surprisingly, this

accomplishment landed him a job at the Sumitomo Corporation who further invested in

his training by dispatching him abroad for a two-year study tour of the US. When

Kawakami came on board at Yamaha, he brought not only a set of enterprise-specific

skills (Kioke and Inoki 1990, Patchell and Hayter 1995) and other accumulated tacit

knowledge honed at Sumitomo, but also the financial backing of Sumitomo. This

relationship with one of the large keiretsus6provided financial stability to Yamaha in

both the short and long run, in the latter by enabling the firm to enter new businesses.

Indeed, into the 1990s Sumitomo group firms held the largest block of shares in Yamaha

at 14% (Johnstone 1999: 223). In light of the almost unanimous cry of under-financing

coming from the American industry (chapter three), the stability of capital at Yamaha is

remarkable.

56 Keiretsu are the large Japanese enterprise groups that are usually headed by a bank and associated with a sogo shosha (general trading company). In theory the post-war reforms disbanded the zaibatsu in particular by wrestling control away from leading families; the keiretsu form is the result (Lazonick 1991)

149 1 E Besides the change in Yamaha's fortunes, 1927 proved a significant year for the musical instrument industry in Hamamatsu in another way. That August, Kawai Koichi, one of the original seven Yamaha employees, an engineer and head of the piano action section at Yamaha, established the Kawai Gakki Company. Kawai's ability to differentiate his nascent enterprise from Yamaha lay in a decision to inscribe his novel piano action in a patent document, enrolling the interests of the state to protect his own intellectual property. Stimulated by this competition, musical instrument manufacturing in the region took off. On the eve of World War 11, in addition to Yamaha and Kawai,

Hamamatsu and its environs were home to five other musical instrument manufacturers:

Fuji Gakki, Mitsuyo Gakki and Hamamatsu Gakki (which were offspring of Kawai), as well as the Taiyo Musical Instrument Manufacturing Company and Enshu Gakki

(Ohtsuka 1980).

During World War 11, Yamaha ceased production of instruments and began producing airplane propellers while Kawai also switched its output to aviation parts. The location of these strategic industries in Hamamatsu made it a target for American bombers and on June 18th 1945 much of the city was reduced to rubble including Kawai's

HQ factory. Yamaha's HQ factory, though damaged by the ensuing fires that raged through the city, remained operable and was able to revert to civilian industrial production of pianos soon after the war (Kawai 1997).

4.3.2 The Post-War Period

As a result of its wartime experience producing aircraft propellers, Yamaha was able to make the transition to the production of motorbikes in 1953. Since this time the Yamaha

Motor Corporation established itself as a separate corporate entity producing motorcycles, outboard motors and engines for automobiles. It is also likely that Yamaha's

success in these endeavours gave it confidence down the road in broadening its scope to

new fields such as integrated circuitry and electronid7.

In the immediate post-war period, overall capital scarcity in Japan meant that it

was not until 1950 that instrument manufacturing in Hamamatsu returned to pre-war

levels. From this point, Hamamatsu rapidly assumed its role as a regional production

system for musical instruments. Figure 4.4 shows the changing number of musical

instrument factories in Hamamatsu in the post-war period.

Until 1970, pianos, organs and harmonicas comprised the core of this output.

Thereafter, woodwind, brass, electronic instruments and a cluster of associated parts

makers were included in the tally. These factories relied on a highly specialized social

division of labour (see Table 4.3). By 1960 there were 26 different firms manufacturing

pianos in Hamamatsu. A third of these were subsidiaries of Yamaha and Kawai while

another third were independent spin-offs of these enterprises. Takeo (1 989) has shown

how these two groups of start-ups relied on a handful of key engineers trained at either

Yamaha or Kawai for the transfer of skills to the new ventures. In terms of output though,

piano production in Hamamatsu has been dominated by these two key firms. Yamaha and

Kawai accounted for just under 90 percent of Hamamatsu's piano output in 1965. By

1984, through the attrition of smaller makers, this figure had reached 98% (Table 4.4).

57 Amongst piano manufacturers, Yamaha was not alone in its war-time experience. Indeed it is hard to conceive of many consumer goods industries that were not forced into similar circumstances - a rapid switch from butter to guns - during the war. For instance, the US piano maker Baldwin produced wings, fuselage sections and other parts for fighter, bomber and other aircraft. Baldwin's website states that, "lessons learned in the construction of multiple-ply aircraft wings became the basis for Baldwin's 4 1 -ply maple piano pinblock." (Source: http:Nwww.baldwinviano.com/about/histo~l;accessed: 26/06/04) i Figure 4.4: Musical Instrument Factories (Final Assembly) in Hamamatsu 1948- 1977

Source: Translated from Ohtsuka: 1980: 161

Table 4.3: Establishment of musical instrument parts producers in Hamamatsu 1944-1977

Source: Translated from Ohtsuka 1980: 162

Yamaha, itself, has always accounted for around two thirds of Hamamatsu's piano production. The rapid growth in piano production capacity in Hamamatsu owes much to the orchestration of an increasingly sophisticated and spatially concentrated social division of labour. In 1960, the pattern of agglomeration shows three distinct clusters of parts makers located in close proximity to the assembly plants of Yamaha and

Kawai (Figure 4.5). Table 4.4: Share of piano production within Hamamatsu 1965-1984 I Production I Yamaha Kawai Other (units) 1965 65.80% 23.20% 11% 147,538

1980 ND ND ND 394,127 1981 67% 27% 6% 361,128 1982 67.40% 28.50% 4.10% 327,394 1983 71.90% 26.40% 1.70% 328,589 1984 72.20% 26.10% 1.70% 299,122 Source: Translated from Takeo 1989: 57

Figure 4.5: The agglomeration of musical instrument manufacturers in Hamamatsu

Employees Assembly Parts

10-99 o 100-499 0 0 500+ 0

K Kawai

(Source: based on Ohtsuka

By 1975, a number of new final assembly plants spread the pattern of agglomeration beyond the boundary's of Hamamatsu city to incorporate other centres of production in western Shizuoka prefecture (Figure 4.6). 4.3.3 Learning to Mass Produce, Enrolling the Consumer

A great deal of corporate learning went into creating the conditions under which piano production could achieve these economies of scale. As inferred in chapter three, the post- war US industry provided the inspirational model for this industrial learning. In the summer of 1953, Yarnaha's President Kawakami and other senior managers embarked on a 90-day shisatsu ryokou or inspection tour of piano factories in North America and

Europe. The purpose was to visit the factories of competitors and other companies in order to examine their methods of high quality and western style mass production. Thus began a 'cycle of accumulation' (Latour 1987), or a set of iterative connections Yamaha formed with centres in the US to enhance their conceptual and practical understanding of the latter. This process was inspired by, and reproduced the travels of Kawakami's father when he was with Sumitomo although in this case, it was an entire junket, rather than an individual who took part. Participants on the tour were shocked about the management practices58and degree of automation seen in factories59.They were additionally struck by the lifestyles of the consumers in these countries. Certainly the link between the latter two observations (mass production and consumption) was clear in their minds, for on their return a concerted effort was directed at ways of attaining high levels of each

(Yamaha 1987).

Yamaha and Kawai sought to encourage the mass consumption of pianos and other musical instruments through a number of strategies designed enroll consumers into the network. First, both firms established a branch network of music education centres throughout Japan. In 1954, the year after Kawakami's shisatsu ryokou, Yamaha inaugurated the Yamaha Music School Program while Kawai's own Music Classroom program began in 1956. Based on the instruction of keyboard skills to children in a classroom setting, rows of organs took the place of desks while innovative teaching techniques such as having mothers and their children take lessons together created a demand for instruments that would enable home practice. Financing innovations such as installment plans that could be initiated before the birth of a child and piano loan programs in which instrument makers teamed up with the large national banks like

Mitsubishi proved vital in facilitating mass consumption. Out of the education programs, formal music competitions and recitals, such as the Yamaha Popular Song Competition

58 They were struck, in particular by seeing 28 year old general managers overseeing 40 and 50 year olds $?maha 1987: 23). Comparing Wurlitzer's (Chicago) output of 2000 pianos a month, with their own capacity of 4- 5001month certainly established a challenge for their return (Ibid:24). and Junior Original Competition, were set up throughout the country. At the same time, each of these firms established a nationwide dealership network whereby retailers would exclusively offer either Yamaha or Kawai pianos. All of these programs which had proven successful in enrolling Japanese consumers by framing the piano as an educational instrument were eventually introduced into overseas markets.

The success of these strategies had two long-lasting implications. First, the retailing practice led to an almost complete bifurcation in the Japanese market that rarely deviated from either Yamaha or Kawai. Incidentally, this would create difficulties for rival firms such as Roland in gaining a foothold at home in the electronic era (Kakehashi

2002). Topologically, Roland was obliged to make overseas connections because its interests were difficult to translate into the domestic distribution system. Second, they created a level of domestic demand high enough so that profits from these activities could be directed towards new avenues such as electronic instruments, a luxury not afforded their American counterparts, whose research fimding was more immediately contingent of the short-term success of particular instruments (Milano 1988).

4.3.4 A National Institutional Basis for Mass Consumption of Musical Instruments

In stimulating demand amongst the Japanese populace, Yamaha and Kawai were really tapping into a long established pattern of state support for the promotion of Western music as an instrument of national education. Indeed the adoption of Western music was a central plank in the Meiji government reforms (1 868- 1912) (Malm 1971). For instance, the Tokyo Conservatory of Music, located in Ueno Park, was established in 1887 and was modeled on similar institutions in Britain and Germany. Significantly, native Japanese music and instrumentation, such as the shamisen and koto, were not taught there until 1936. Institutionally, this Westernization of music was most thoroughly embedded in the military and education systems, not as Malm (1 971 : 259) argues, "out of any special interest in the qualities per se, but rather as necessary parts of a Western-derived table of organization for the particular institution in question." Thus, the first concert performance of Western music was by a military band, whose leaders were likely emulating the discipline of rhythm and harmony seen in the band that accompanied US Commodore

Perry, the first US diplomat to visit Japan in 1853. In schools, Japanese children in the

Meiji period learned hybridized versions of Western songs, usually through singing, but gradually through the introduction of Western instruments, including the piano. Out of its cradle in these institutions, Western music gradually diffused into Japanese society.

Loesser (1 954: 597), noting how this pattern of adoption differed markedly from the mode of status emulation found in the Western context comments that,

Thus, without totally repudiating their ancestral musical habits, the Japanese became Westernized musically. This came about not because their upper classes affected Western music as a fancy foreign fad, but because the practice of it was implanted systematically and generally in the very young of all ranks, three quarters of a century ago.

So if public institutions drove the adoption of Western music in the Meiji period and established a basis of demand through which Yamaha and Kawai could prosper, in the postwar Showa period, it was the adoption and adaptation of Western marketing techniques by these same firms which enabled these enterprises to return to first, national and then international stature.

In the post-World War I1 period, the production of pianos grew rapidly to meet this demand (Figure 4.7). Until recently the domestic market absorbed the majority of this output and in 1965, it accounted for 89% of production. Twenty years later, 70% was still destined for the national market, although by 2000 this share had fallen to 35%. In terms of total output, Japanese manufacturers reached their peak around 1980, a time that coincided with the pronounced boom in electronic instruments (compare Figure 4.7 with

Figure 4.8).

Figure 4.7: Piano Production in Japan 1965-2000

1 Exports 1 1 Domestic 1

Source: Japan Music Trades 200 1

Figure 4.8: Sales of Keyboard Units by Japanese Manufacturers 1965-2000

rn Synthesize1 mn E. Keyboan E. Organ rn E. Piano s Piano

Source: Translated from the Japan Music Trades 200 1 4.4 Japanese Firms Enter the Electronic Musical Instrument Field

4.4.1 The Case of Yamaha

In 1952, at the same time as Yamaha initiated its program to learn methods for the mass production of pianos, President Kawakami decreed that the firm would enter the electronic organ field. Yamaha's strategy was once again guided by the process that I referred to in chapter two (Table 2.2) as transregional innovation, meaning that Yamaha sought to learn from and improve on the model presented by the US. From a science studies perspective (Latour 1987, 1999), Yamaha enhanced its 'cycle of accumulation' with the US. Samples of American organs, principally Hammonds and Baldwins were ordered from the US and thoroughly investigated. In contrast to the shisatsu ryoko which required travel, this process reflects a 'mobilization of the world' that hinged on

Yamaha's accumulation and examination of 'immutable mobiles'- blueprints, foreign musical instruments, and so on.

These very early efforts at reverse engineering could yield little, since at the time

Yamaha had virtually no staff possessing electrical engineering skills. In other words it possessed no disciplinary basis from which it could begin to make use of these foreign organs. Asheim's (1 999) notion of 'disembodied knowledge' an intermediate form of knowledge between the tacit and codified realms helps in conceptualizing this situation.

Outside of their context of origin, immutable 'instruments' such the American organs were indecipherable to engineers lacking the necessary tacit knowledge to make sense of them. To gain a practical mastery of these technologies required travel.

Consequently, the Ministry of Foreign Affairs sponsored two young Yamaha employees to receive English language training and supported the cost of dispatching them abroad for a two year sojourn to America. Graduate work at Yale was accompanied by visits (shisatsu ryokou) to organ manufacturers including Wurlitzer, Conn and

Baldwin, as well as to electronics research units of RCA, GE and Bell Labs (Yarnaha

1987). These endeavours mark the earliest practices in a sequence of transregional inspirations through which Japanese firms enrolled themselves in the spatial innovation system for EMI. At this stage, Yamaha's technological strategy was embedded within a national system of innovation, as the firm received support and guidance from the highest level.

In order to continue with the development of this technology, Yamaha had to enrol more allies. Between 1954 and 1957, Yamaha, in collaboration with Nippon

Electric, conducted a number of experiments in tone generation at the national broadcaster NHK's Tokyo Acoustic Laboratory. Like the shisatsu ryoko this relationship points to a very direct link between the private firm and the public institution, further evidence of the importance of national innovation system influences on the transectorial strategies of the firm. Conceptually (in reference to Table 2.2), though Yamaha continued to derive inspiration transregionally, its cultivation of a tacit practical knowledge of EM1 had entered the phase of intrapreneurial invention. A number of prototype organs emerged out of this work and in 1959, Yamaha released it first electronic musical instrument, the D- 1 Electone which employed a bank of 28 1 transistors to generate sound. The D-1 and its successors formed the basis for Yamaha's ascendancy in the field of EM1 (Table 4.5). Throughout the 1960s a number of transistor-based organs were conceived by the recently established electronics division. Until 1969, most of these products were projected for domestic sale, but thereafter the foreign market, in particular r(l!s~ayunnyoqol le spoqlaw uo!pnpo~dpue u6!sap Apnls 08 06 saaAoldwa eqeweA Year Models Features of Features of technological released instrument development development

1981 14 (03) Porta-tone mini-keyboard Production of instruments using FM tone series launched synthesis begins

1983 28 Digital synthesizers (especially DX-7) launched 1984 17 1985 21 Sampling porta-tone launched Models released in Jaoan (Source: translated from data provided by Yamaha. N.B.:the form of the table is largely unchanged from the original.) America, became of increasing importance. Table 4.5 is noteworthy because, apart from my translation from Japanese, the form and content of this diagram are unaltered. In other words it points to the way in which Yamaha represents its own technological development: as a succession of instruments, that evolve according to specific underlying technologies (e.g. transistors, integrated circuits) and production locations.

For large firms such as Yamaha, intrapreneurial invention required a human resource strategy that selected and empowered section-heads to fulfil technological objectives. In this regard, Kawakami made an important decision in 1957 when he hired

Mochida Yasunori away from the electronics firm Nippon Columbia to head the new electronics division at Yamaha. As with the employees who were dispatched on shisatsu ryoko, Mochida played a key role as a transectorial migrant whom Yamaha enlisted to translate the lessons of the emerging ICT paradigm to its advantage. Mochida has been characterized as possessing a restless curiosity which manifested itself in his love for tinkering with machines -taking them apart, putting them together, trying new combinations of parts etc. Mated with a quiet confidence in his own capabilities these traits have certainly contributed to his atypical career trajectory which has seen him switch firms a number of times (Nippon Columbia, Yamaha, Ricoh) and leave the corporate world for academia in the 1990s. Following his entry to Yamaha, Mochida's principal long-term goal was to develop a system for real-time tone synthesis (Johnstone

1999). 4.4.2 Yamaha's LSI Program

By the late 1960s it was apparent to Mochida that the transistor, as the core technology of the electronic organ, was not powerful enough to fulfill this aim. Indeed, in the general electronics field, the transistor was rapidly being eclipsed by the integrated circuit which contained thousands of transistors on one chip. Moreover, Mochida realized that generic chips would not work for the high quality tone generation he had in mind. Application specific integrated circuits (ASICs) were required. However, when Mochida approached suppliers such as NEC and Hitachi about making such chips, "they told us to stop thinking about something so difficult" (quoted in Johnstone 1994: 3). Against strenuous opposition from the company's board of directors, Mochida proposed to Kawakami that

Yamaha would probably have to spend S2 billion - an amount equal to Yamaha's entire capital at the time (Yamaha 1987: 52) - in order to become a chip maker in its own right.

Kawakami agreed, saying (according to Mochida), "if we can make the best musical instruments in the world, then no matter how difficult it is, no matter how much money it costs - we'll do it." (quoted in Johnstone 1994: 3) - further evidence of Hamamatsu's

Yaramaika philosophy. Mochida's ability to enlist the support of Yamaha's management to develop an internal competency in the production of not second-generation but leading-edge carrier technology reinforces the importance of intrapreneurial vision to its corporate strategy. In Latour's (1 999) terms, Mochida translated the interests of a key ally to his project. The strength of this connection endured and this first moment of persuasion by Mochida laid the foundation for Yamaha's later realization of

'technological crystallization' (Table 2.2).

Yamaha's quest to develop the capability to produce LSI and the central role that this decision played in the firm's subsequent evolution has been discussed by Nakagawa (1 984) who commented that such a grand endeavour made it imperative that Yamaha cultivate the knowledge and expertise necessary to make its own chips. Once again, in

1969 a team of young engineers was dispatched, this time to Tohoku University and the labs of Professor Nishizawa Jun'ichi. Today, Professor Nishizawa is known in Japanese engineering circles as the 'father of optical communication' and the 'DaVinci of semiconductors'. However in those days, he was one of the only people in the country in possession of the knowledge which Yamaha sought. Under the rigorous tutelage of

Nishizawa, these engineers acquired the fundamental and applied skills that would enable

Yamaha to voyage along a path, the trajectory of which was a radical departure from its basis in electro-mechanical engineering and instrument manufacture. Yamaha committed to LSI technology in a long-term manner whereby internal developments (location decisions, product/process innovations) kept pace with the external evolution of the key technology (Table 4.6).

In 1970, new land was bought for a factory, east of Hamamatsu, across the

Tenryu River in Toyooka. In 1971 Yamaha's Toyooka plant began producing MOSILSI

(Metal-oxide semiconductor/large scale integrated circuit) at a rate of 50,000 units per month. At the time, the Electone line was still located at its HQ factory near downtown

Hamamatsu, but from 1973, this production function was gradually shifted to Toyooka, a move that became complete in 1976. In 1976, spurred by government policies that encouraged the diffusion of production from core areas (such as the Pacific Industrial

Belt between the Kanto and Kansai regions) to peripheral locations, Yamaha established Table 4.6: Semiconductor technolorn develo~mentat Yamaha Year Semi-conductor technology Software Product related Industry's technology technology technical development trend First MOS-LSI users I969 In-house production decision 1970 Launch of Start Toyooka factory construction - note: it LOCOS semiconductor begins with a factory! production system - 1971 Completion of Toyooka factory; reliability check Circuit analysis SIT Q\ program QI SI Siristor development; ISOPLANAR Production management Ikd RAM system I PMOS 4bit micro- computer (4004) 1972 Start of MOS-LSI mass production; start of SIT Start of PAS 12L development (Passive analog PMOS 8bit micro- system) computer (8008) 1973 Establishment of Low Noise FET; Be oscillating board Introduction of nMOS 4bit full crystal development; ion pouring machine introduction CAD system microcomputer technology (IJCOM-4 1974 Static guidance 300W SIT mass production; 'One Master' tone SIT power amp; 4k RAM transistor LSI Beryllium oscillating nMOS 8 bit development board speaker; microcomputer GX-1 polyphonic (8080,6800) synthesizer Year ltem Semi-conductor technology Software Product related Industry's technology technology technical development trend 1975 Rhythm ABC-LSI LSI mass Kagoshima factory started; LSI for electronic Start of FM Electronic piano 280 production piano; Super Low Noise FET system; 2 Step Poly Si factory Semiconductor production management system II Digital Electone Arpeggio LSI; Complementary SIT; Ultra Low PASS (digital-analog era begins Noise FET hybrid Electone) series 4k16k EEROM RAM -- -- - I Dry etching (RIE) introduced; Projection aligner MC type cartridge introduced 1 8086 High speed LSI SlTL development; High frequency SIT; DBias simulator Piano tuner 64k RAM launch Electronic beam flash equipment introduced; development CCD 1M bit Circuit employing 'one chip-one board'; Low pressure CVD introduced Hubble 1M bit Start to go Stepper introduced; Low pressure AP Automatic layout beyond LSI introduced; High speed LSI tester introduced; program; High pressure resistance-high frequency SIT Process touch sensor; High pressure oxidation furnace simulator introduced Design rule TRX (prototype FM check program synthesizer) development Vector generator LSI; Image sensor YIS (home computer) 32bit

Digital product LSI for AHD use; FM operator LSI; Graphic Personal CAD maturation; start LSI; LSI for CD use; Signal processing LSI; system; LSI Year ltem Semi-conductor technology Software Product related Industry's technology technology technical development trend of sales to DAC; Blaze pressure sensor automatic design YGT (graphic terminal); outside system CD-X1 (CD player); MSX companies (home computer); Pro digital reverb 1984 Production ability MSX -VIDEO LSI magnification I - I iource: Translated from Nakagawa 1984: 194- 197 a second factory for LSI production in Kagoshima, a city in the southwest corner of

Kyushu, Japan's southernmost island6'.

Sales of LSI began in earnest in 1983 and Yamaha's mastery of the production of

ASIC chips capable of digital tone generation, and its patenting of this technology gave the firm a monopoly position in the new market for PC sound cards and other multimedia applications6'. In the process of acquiring the expertise to produce LSI, it was necessary for Yamaha to develop a set of fundamental skills in metallurgical technology and materials development (Figure 4.9). Figure 4.9 is remarkable, again, because it represents

Yamaha's own framing of their technological diversification as a longitudinal profile.

Along the top are listed all educational and promotional programs Yamaha has used to connect with the public. Below this is a family tree that illustrates all the diverse product areas that the firm has entered. The critical juncture in the table connects the D-1

Electone with all the electronics technologies that follow, including semiconductors, professional audio equipment and industrial robots.

The newly acquired core competency in electronics technology provided the platform for a series of incremental innovations in the field of EMI. Table 4.7 represents how two of Yamaha's engineers Mochida and Aoki (1994) fkamed this trajectory in their book on musical instruments and computers. With the benefit of hindsight this is how they map their world by connecting the fundamental knowledge they harnessed with the instruments they mobilized.

60 Policies such as the Relocation Promotion Act (1 972) which was authored under the auspices of the New Comprehensive National Development Plan (1 969) encouraged the investments by semiconductor manufacturers in peripheral locations such as the island of Kyushu (Murata 1980, Sargent 1980) In 1994, Yamaha's sound chips accounted for 95% of sales to of the $lbillion market for sound boards (Johnstone 1994). In later chapters I show how this process was secured through aggressive patenting

practices that laid claim to large swaths of intellectual space. In this regard, I will

mention two strategies that were of particular importance. First, was the licensing of key

patents in the field of digital synthesis and the subsequent process of improving on this

technological knowledge through a series of derivative and ancillary patents. Second, was

the exclusive license taken out on a patent developed at Stanford University detailing the

method of FM tone synthesis62.In combination, these sets of proprietary knowledge were

rigorously applied to organs, synthesizers, digital pianos and eventually video game and

computer sound cards.

There were however, limits to Yamaha's realization of economies of scope and

Yamaha's misguided venture into the computer business in 1985 is a case in point. Thus,

in trying to build a multimedia computer with built-in sound and graphics, Yamaha's

decision to take on development and production of the myriad of components necessary

for such a venture proved disastrous financially. Yet even in failure, important

technological lessons were learned and positive offshoots of this fiasco included a

contract with the game maker Sega to supply the sound and graphics processors for its

consoles as well as a brief role as a supplier of entire sound boards to IBM for that firm's

PSI2 model in 1986 (Johnstone 1999). As this example illustrates, the market often

imposes limits on a firm's ability to diversify transectorially.

62 FM (Frequency Modulation) tone synthesis was developed by Stanford's John Chowning. This b technology was licensed exclusively to Yamaha in 1973 (See chapter six) -'ablt 1.7: The h ory of ele :ronic musical in! ruments at ' amaha Year dethod of Electronic Organ Electronic Synthesizer Electronic Technology rone Models Piano Models Models Keyboard Models - ;eneration 1950 1952 Research Begins 1955 E-05 prototype 1959 D-I

1970 1974 SY-1 1970 EX-42 1974 GX-1 1977 CS- 1978 CP-80 80 1977 E-70 1980 1980 PS-1 1982 PC-100 1981 GS-1 1982 HS-4 1983 DX-7 1982 MP-1 I983 FX-1 1985 CVP-7 1984 PS6100 1988 B- 1986 SHS- 1987 HX-1 200 1OR 1990 1989 SY- 1988 PSS- 77 140 1991 CVP- 1991 EL-90 75 1993 CVP- 87 1993 VL1 1993 PSR- 1994 VP1 2700

Source: Translated f rorn Mochida and Aoki 1994: 17)

IC: Integrated Circuit LSI: Large Scale Integrated Circuit FM: Frequency Modulation AWM: Advanced Waveform Memory VA: Virtual Acoustic 1920 1930 1940 1950 1960 1970 1980 1990 ---- 1954 1964 1970 1987 I Education & I Yamaha Electone Junior Original Teens Music ------Popularization : Difiuiion- of : Music School concours Competition Festival !- -Y!=% - -1 I 1967 1982 1993 Pianos r ELhtpiano CF Concert Disklavier Silent Series Grand Piano Piano / Grand1902 piano Organs 1887 1935 1959 1975 1983 1991 Reed organ I Magna Organ D-1 Flectone Gx-1 Electone Clavinova EL Series I Electone I I Wind, String 1914 1942 I 1965 1995.1997 - Silent Brass & Percussion Harmonica Acoustic Trumpet, Drums I I Guitar : E. Guitar 1980, 1983 Silent Violin I I Digital EM1 II Portable Keys. 1990 I------1974 CSY-1 Synth DX-7 Digital Sequencers ~lectronics Synthesizer echnology Pro Audio I------1974 1980 1994 Mixing console Digital Signal Digital Mixing I I I Processor Console , 1 I Audio , 1922 1955 , 1974 1986 1995 I I Hand wound Yamaha hi-fi ', Hi-fi speaker Sound field Theater Sound I : phonograph I Processor System I I L I ------Electronic I 1971 1983 1993 I Devices I LSI Sales 4X speed I CD recorder Electronic ------1961 1971 1981 1992 Metals YFAP/CuT IC lead frames Etching grade lnvar material lead frame for shadow masks Industrial 1981 1995 Systems Industrial 2-I1 6 Industrial -. robots robot (eqewe~/(q pap!~o~d 1954 1960, 1968 1974 1987 1994 Sports and legaleur uo paseq '~oqlne6q am%y:amnos) - Leisure au!lam!$ pnpord s,ayamaA :6*pa.an%!d Motorcycle Powerboat Swimming Air Electro-hybrid YA-1 Snowmobile pool conditioners bicycle Where Yamaha has succeeded is in its ability to develop a range of EM1 that effectively leveraged its competency in producing tone generating ASIC (Application specific integrated circuits) that rely on FM and subsequent modes of synthesis. A short list of these accomplishments included its portable and professional synthesizers lines, its

Silent Series instruments (piano, violin, drums), and more recently its miniature synthesizer chips that could produce ring tones in cellular phones. The novelty of these ventures has not come without attendant unintended consequences. For instance, in response to Japan's Casio Corporation's aggressive entry into the high-volume, low- priced portable keyboard market in the early 1980s, Yamaha had to adopt a set of manufacturing and marketing practices that were more in line with a consumer electronics firm. An executive in Yamaha's product planning division commented that in many senses, the firm views products such as the Portatone, "not as musical instruments, but as consumer products" (Matsubayashi, K. Interview 8/27/01; translation by the author). I felt a sense from the tone of this statement that if he could turn back the clock, then perhaps Yamaha might have been better off ceding market share to Casio in this segment in order to focus its efforts on the higher margin professional keyboard market.

Another case concerns the development of the Silent Piano. The target for this type of instrument, which employs a complex set of technologies to enable practice with earphones, was thought to be the emerging markets of East Asia, such as Taiwan, Hong

Kong and China, where the premium on residential space necessitates quiet performance.

Contrary to these expectations, Yamaha found that instead of families wishing to muffle the sounds of their piano practicing children, they wanted their neighbors to know that they had 'made it' and could afford to pay for their child's lessons (Ibid). 4.4.3 Yamaha as a Mark I1 Innovator

To summarize, Yamaha's decision to make its own LSI semiconductor chips marks the defining event in the firm's entry into the field of electronic musical instruments.

Undertaken at a time when Yamaha's American rivals were only speaking of the virtues of LSI at venues such as the AES convention (see chapter 3), this decision has cast a long and favourable shadow over its subsequent firm structure and technological strategy

(Figure 4.9). It is difficult to find a branch of instruments to which Yamaha has not sought to apply its digital expertise. Given these accomplishments, it is still important to recognize Yamaha's roots as a piano and organ manufacturer, for it is from this basis that its transectorial innovation has evolved. After all, without the substantial revenues generated by the mass production of acoustic instruments in the 1960s, it is doubtful that

Yamaha would have had the steady stream of revenue with which to fund its ventures in the brave new world of electronics. With respect to this matter of funding, it is not clear the degree to which Yamaha's foray into EM1 has been underwritten by the relationship that the first of three Kawakarni CEOs forged with Sumitomo. Regardless of how it was funded, the LSI program has allowed Yamaha's technological strategy to work on a canvas sufficiently large to accommodate its 'big-picture' vision. In this sense it is better to highlight intrapreneurship, in particular the ability of Mochida to enlist the massive finances for the internal development of LSI, as the impetus for Yamaha's transectorial strategy. Against this competition provided by a large, capital rich, multidivisional firm, the efforts of Yamaha's local rival Roland are best contextualized. 4.5 Competition from an Entrepreneurial Upstart - the Case of Roland

Roland is the only producer in Hamamatsu who has, from the outset, specialized in EMI.

Without a history of making pianos and other acoustic instruments, the firm's set of competencies have never had to branch out from an established core into the electronic realm. In addition, relative to Yamaha and Kawai, its evolution and technological trajectory has been shaped by the entrepreneurial drive of its founder Ikutaro Kakehashi.

Indeed until very recently, Kakehashi's influence, particularly in the areas of product design has been considerable. In order to understand the personality that drove this process it is necessary to examine aspects of Kakehashi's autobiography in detail.

Consequently, much of the ensuing discussion is based on Kakehashi's memoirs: I

Believe in Music (2002).

4.5.1 The Origins of Kakehashi Ikutaro as a Transectorial Entrepreneur

Born in Osaka in 1930, Kakehashi was orphaned at the age of two and adopted by his grandparents. To escape the bombing of World War 11, the grandparents moved to

Kyushu while Kakehashi stayed on throughout the war to continuing with his schooling.

In 1946, Kakehashi also moved to Kyushu to spend the next four years there. His job upon arrival involved survey work to update the local land registry, but it was his hobbies, tinkering with radios and clocks, which caught his fancy and allowed him to gain a practical understanding of these technologies. When his survey contract expired,

Kakehashi apprenticed for two months with a local clock and watch repairman before starting his own business. His work on radios continued although post-war scarcity meant that there were few new components available, and these had to be ordered from Tokyo.

The solution was to purchase broken radios and spend long hours taking them apart and assembling working sets from the useful bits. This practice proved invaluable as it cultivated a life-long habit of hands-on technological learning.

When Kakehashi returned to Osaka, the profits from his watch repair business were earmarked to pay for his schooling, but a TB diagnosis meant that the next four years of his life were spent in the hospital. In this unlikely venue he continued his pursuit of a practical technical education, repairing broken watches and now radios of fellow patients and hospital staff. He even found the energy to make his own television set in time to tune into the nation's first NHK broadcast. As a new technology, there were no old models to scavenge, and so knowledge had to be gleaned from published articles available in Japan. To complement his skill at tacit learning, Kakehashi also became attuned to reading circuit diagrams and deciphering knowledge in codified form. These practices are a displaced version of those employed by Takahashi's (2001) network of tinkerers that formed around Akihabara and similar 'electric towns'.

Recovery from TB was bittersweet since the prolonged hospital stay left

Kakehashi too old to pursue a degree and too frail to suit the tastes of most employers.

Continuing a pattern of overcoming diversity, he struck out as an entrepreneur, opening

Kakehashi Musen, a small electric appliance store that assembled and repaired radios and

TVs. Soon the shop diversified and began also to engage in the retail of electric appliances such as rice cookers and washing machines. It was in this capacity that

Kakehashi first came across organs.

This was the time of 'transregional inspiration' (Figure 2.2) when organ technology in Japan had advanced from treadle (foot-pump) to motor-operated units and this shift caught the attention of the appliance manufacturers that supplied the motors. As these instruments came across his transom for repair, Kakehashi began to note their flaws. As he put it,

On the one hand, it seemed almost impossible for appliance makers to acquire the music expertise needed to produce satisfactory instruments. On the other hand, it was equally difficult for musical instrument manufacturers to master electronics technologies. The path to success seemed to lie in a niche somewhere between the two existing disciplines (Kakehashi 2002: 30, emphasis added)

From this perspective of Kakehashi's, technological change as a 'diffusion of the engineering disciplines' (Rosenberg 2000) is anything but unidirectional and deterministic. Indeed, Kakehashi appraised the problem solving frames of both instrument and electronics manufacturers and found them both incomplete. Though he drew on their lessons, he let his ear guide his own practice. After all, he had heard electric organs in churches sound wonderful, and he had become acquainted with the

Ondes Martenot (see chapter one) while helping tune his friend's instrument. His decision to enter the business of making his own electronic instrument was based on two factors. First, the better made instruments created sounds that were like 'candy to his ears', but perhaps more important was the "knowledge that the instrument was made of parts and materials that were relatively close at hand" (ibid: 32).

By 1960, Kakehashi Musen had morphed into Ace Electric Ltd., taking on 20 employees as it grew. Meanwhile, Kakehashi's latest hobby of building crude electric organs had also reached the threshold of experience sufficient to enter the field on a permanent basis. Kakehashi cultivated his entrepreneurial acumen in two ways. As we have seen, the first way was via learning by doing. To complement this tacit knowledge,

Kakehashi (2002: 34), "gained valuable knowledge by looking up information about the history of manufacturers such as Pioneer, Sansui, and Akai, all of whom supplied parts to the radio and TV manufacturers and who had originally operated on a scale that I could understand." By reflecting on these firms as business models, Kakehashi gave direction to his practical efforts to become a manufacturer in his own right.

4.5.2 Entrepreneurial Invention in Japan

A prototype one-manual keyboard with an electric motor was prepared. It was quickly apparent, however, that the venture lacked the capital to finance production on its own.

Consequently, Ace Electric entered a partnership with a company in Matsushita's business group whereby the former designed and produced an electronic organ, the

SX60 1, which was marketed under the latter's National brand name. The contract provided valuable experience in large-scale manufacturing while yielding a source of revenue which enabled Ace to develop it own products. This OEM (original equipment manufacture) partnership can be interpreted as a form of relation-specific skill (Asanuma

1989) in which Roland gained valuable experience 'serving the particular needs of a core firm'. Following this important contract, Kakehashi's entrepreneurial instincts sought further ways to cultivate relations with large firms, including those outside of Japan.

Perhaps it was all his experience with clocks, but the niche which Kakehashi sought to exploit lay in the range of accessories for 'combo' organs that included rhythm accompaniment units. Kakehashi sought to go beyond the then state of the art, which still employed vacuum tubes to create sound, so he devised the world's first transistorized non-automatic, percussion instrument, the Rhythm Ace R-I. 4.5.3 First Products Bypass the Home Market

Kakehashi was thrilled to market this product at the 1964 NAMM (National Association of Music Merchants) show in Chicago. Typically, firms build up domestic market sales prior to exporting (LeHeron 1986, McConnel 1980). However, for Ace Electric it was imperative to sell its products abroad from the start and this related back to the structure of instrument retailing in the Japanese home market. In other words, Ace Electric simply could not sell its products in the Japanese market as accessories to Yamaha and Kawai organs as these firms had their own exclusive stores. Faced with this situation, Roland had to acquire a 'relational market intelligence' (Reiffenstein et al. 2002) that was particularly tailored to serving the interests of US customers. This foray into the

American market took a number of years to nurture, but by 1967, the Rhythm Ace FR- 1 attracted the attention of Hammond who began importing these units for incorporation into their organs. Conversely the American organ giant was seeking to enter the Japanese market and secure a production source for its other international markets. Hamrnond

International Japan, the resulting joint venture allowed Ace Electronic collaborate on design with Harnmond and undertake a scale of production far greater than it could have orchestrated on its own.

As Ace 'translated' Hammond's interests to facilitate its own growth, the issue of industrial location came into the foreground. Kakehashi was forced to consider a shift in away from the increasingly crowded and costly Osaka area. Hamamatsu was the obvious choice due to its established status as an instrument producing region. Rather than set up shop on a greenfield site, Ace purchased the recently mothballed Zenon Corporation factory that had made reed organs until the pace of technological change in that sector forced it out of business. Changes in the corporate organization of Ace Electric were also taking place, as one of Kakehashi's partners sold their share in the venture to the

Sumitomo Chemical Company, a branch of the giant keiretsu with the same name. As the new majority owner of Ace Electric, Surnitomo was, to Kakehashi's mind, insensitive to the needs of the musical instrument industry and so in 1972, the firm's founder abandoned this connection to regain entrepreneurial autonomy by establishing the Roland

Corporation. This episode can be contrasted with the case of Yamaha, which, since the

1930s had been allied with Sumitomo.

Figure 4.10 longitudinally profiles the evolving corporate organization and technological pathways through which Roland asserted its place as the world's premier manufacturer of electronic instruments. From a foundation in synthesizers, the figure shows how Roland's pattern of technological diversification exploited Kakehashi's experience in producing accessories and efects for EMI. In particular, the Roland subsidiary company, Boss Corporation, which is headquartered in Hamamatsu but has its manufacturing base in Taiwan, specialized in producing effects pedals for electric guitars and basses. Along a similar vein, Roland's development of programmable bass and rhythm machines has had an immense impact on popular music. Thus the TB-303 Bass

Line and TR-808 drum machine were originally designed by Roland's head of R&D,

Kikumoto Tadao, to serve as accompaniment devices for guitarists. To even an unschooled ear, the Bass Line sounds nothing like a bass at all, and it failed to capture the imagination of its target audience. Yet when musicians working in the then nascent genres of 'house music' and 'hip-hop' purchased used TB-303 units in pawn shops, tweaked the knobs and pumped up the volume, 'acid-house' music was born. The TR-

808 met with similar fortunes of fate and served as the pulse for a generation of electronic 1960 1970 1980 1990 I960 Ace Electronics 1972 Roland Corporation 1968 Hammond- Intn'l- 1973 Boss Corporation (guitar effects pedals) Corporate 1977 Roland ED (guitar synthesizers, AV production equipment) Organization 1981 Roland DG (computer peripherals, scanners, plotters) 1987 Roland Tech (cabinetry) 1988 Rodgers Organs (digital pipelnon-pipe organs) 1994 Edirol Co. (AV editing equipment)

Education & Popularization 1984 1994 Roland Music School Roland Foundahon- Pianos 1986 b Digital Piano Organs 1960 1965 , 1988 b National Electronic Organ Ace Tone Rodgers (marketed by Matsushita) Combo Organ Organs Accessories 1964 1975 1977 1981 b & Effects RhyihmAce Percussion Guitar Amps Effect Programmable Bass, Accompaniment Machine Pedals Rhythm Machines

Synthesizer 1972 1976 1981 1987 b Keyboards SH-1000 Modular Polyphonic Digital Synthesizer Synthesizer Sythesizer Synthesizer Electronic 1995 1997 String & percussion Guitar Guitar System Digital Drums Instruments Synthesizer

Studio 1 1981 1988 1996 1999 Equipment Sequencer Desktop Music System Studio Digital Workstation Mixing System ProlAm AV 1990 1994 Electronics Roland Sounc Desktop Video 1982 1988 Canvass 1997 Computer Computer CNC Colour Peripherals Plotters Cutter Printer dance music in the 1980s and 1990s (Poschardt 1998). The immense popularity and unintended consequence of these episodes has been recognized by Wired Magazine

(2002), which counted both instruments amongst the top six most influential EMI.

Like Yamaha, Roland also sought to exploit various economies of scope from its basis in electronics (compare Figures 4.9 and 4. lo), albeit in a more focused manner than

Yamaha. Thus, Roland DG (digital group) was created in 1981 to diversify into complementary and even new spheres of electronics. From an initial specialization in making digital plotters, Roland DG has also ventured into computer numeric controlled

(CNC) machinery and, more recently, colour printers (see figure 4.10). Takeuchi (2002) cites this unit as an example of the type of firm - a spin-off that draws on the technological strengths of the parent firm to boldly enter new lines of business - which underpins the secret of Hamamatsu's industrial strength.

The case Roland contrasts markedly with that of its neighbor and rival Yamaha.

In particular, Roland's network of relations was linked to an entirely different set of interests. For instance, its business 'allies' were initially larger firms (Matsushita,

Hammond). In providing OEM services to Matsushita, Ace Electric cultivated 'relation- specific skills' that could be translated to other contexts. Additionally, Ace Electric first aligned itself with the EM1 sector by producing accessories, and this aspect of technical complementarity brought the firm into the sights of Hammond. Geographically, in order to establish its presence in the nascent EM1 sector, Roland was compelled to topologically 'act at a distance' since it could not immediately translate its products to the domestic retail context. Roland's developmental trajectory also highlights the contingent nature of unintended effects, for the initial failure of its TB-303 product line received a 'second wind' of support fiom a group of users who were never imagined by that product's designer. Finally, the personal history of Roland's founder Kakehashi, illustrates 'tinkering' as the basis of innovation.

4.6 The Case of Kawai

Kawai presents a different case from Yamaha and Roland. Yamaha has dominated the musical instrument industry for at least the last 20 years. Indeed, Yamaha's bold entry into electronic music signaled by a commitment in the late 1960s to develop the capability to make its own LSI, the core technology of the ICT paradigm, suggests that

Yamaha sagely 'read the tea leaves' of the emerging paradigm and took this and other decisive actions to ensure its own longevity in the field. Furthermore, it developed a number of key products, such as the DX-7, the world's first digital synthesizer, which have been widely described as being revolutionary (Colbeck 1996, Theberge 1997, Pinch and Trocco 2002). Finally, its almost singular realization of economies of scope, stemming largely from its original EM1 strategy completes the picture of Yamaha as a consummate learner. Roland's trajectory, on the other hand, very much bears the signature of its founder Ikutaro Kakehashi. Relative to its two neighbors, Kawai's role in the electronic age has been far more muted. Yet its firm size, structure and product range

(with the exception of motorcycles) at the start of the electronic age were similar to that of Yamaha and as such present an interesting point of contrast to its major rival (compare

Figures 4.9 and 4.1 1). co.E! E srn 4.6.1 EM1 Development within Kawai: the Lack of Local Linkages

Kawai initiated its own EM1 program in 1958 through a joint venture with Toshiba to produce electric organs under the brand name Sforzando. The latter provided the electric organs and assembly facilities. Soon after this, in 1962, Kawai purchased Meruhen, a manufacturer of electric organs located in Mori-cho, 20 km east of Hamamatsu. From this point, Kawai centred its EM1 strategy on the accumulation of assets from outside the firm. 1965 saw the purchase of Teisco, a firm located in Saitama Prefecture, north of

Tokyo, which made electric guitars, speakers and electric organs. For the next 20 years

Teisco was the core of Kawai's EM1 efforts. In 1975,50 engineers comprised the product development team for EM1 at Teisco in comparison to 10 at Kawai's HQ labs (Kondo, personal communication 24/7/02). This spatial division of labour between Kawai,

Meruhen and Teisco was unique in comparison to the more concentrated corporate arrangements pursued by its local competitors. In 1986, many of the R&D jobs at Teisco were moved to Hamamatsu. Until this point, the integration of R&D efforts within the

Kawai group was limited by the friction of distance between Saitama and Hamamatsu.

Indeed in one of my interviews with the firm, engineers who had been working at each of these labs remarked on the lack of overt coordination between sites63.

The relationship between branches in Kawai's technological trajectory is also much shallower when compared with its rival Yamaha (compare Figures 4.9 and 4.1 1). In particular, though it has entered many of the fields pioneered by its neighbour it has not

63 Group interview with Kondo Yoichi, Asst. Manager, R&D Electronic Musical Instrument Div. Kawai Corporation, Saito Tsutomo, Manager, R&D Electronic Musical Instrument Div. Kawai Corporation, Takaba Tsutomo, Chief Engineer, R&D Electronic Musical Instrument Div. Kawai corporation, Hiranabe, Yoshiaki Asst. Manager Marketing and Merchandising/Import, International Division, Roland Corporation, Hamamatsu. 22/06/02. invested in the fundamental learning that might better integrate these pursuits with its core business of instrument manufacturing. For instance, its precision metals and computer and software lines appear somewhat peripheral rather than integral to its overall direction. These endeavours are by no means frivolous. However, they do seem to have been undertaken to complement Kawai7score business rather than to radically extend the scope of the firm, technologically, into new fields.

In other words, Kawai's technological trajectory has been the most conservative of the three major EM1 firms in Hamarnatsu so it has never been a first mover in any category of EMI. On the other hand, it has never lacked the resolve to observe and then follow closely along the paths laid out by its neighbours. For instance, although Kawai did not play a lead role in the negotiations to establish the MIDI protocol (see chapter three), it sat at the table with its neighbours and rivals and reaped the benefits of this very important episodic collaboration. Indirectly, another linkage between Yamaha and Kawai is that the American, Ralph Deutsch, the most prolific patentee of EM1 and the inventor of some of the field's most cited technologies, who performed contract work for both firms. The implications of this indirect relationship will be examined in more detail in chapter six.

4.7 The Swarming of Japanese Firms into the EM1 Sector

In the previous chapter I discussed the Schumpeterian dynamics that underlay the swarming of American firms into the EM1 sector during the 1970s. Particularly after the introduction of path-breaking instruments such as Moog, scores of entrepreneurs (Dave

Smith, Tom Oberheim, and so on) eagerly leapt on the bandwagon, only to 'burn out and fade away'. The eventual demise of the American industry has been covered extensively (Milano 1993, Greenwald and Burger 1986, Greenwald 1987) and the sheer breadth of

'flash-in the pan' stories bears testimony to the Darwinian pattern of competition in this sector. On the other hand, as intimated earlier, there has been little attention devoted to an explanation of how Japanese firms fared during this same period. According to Takeo

(1 988), from the late 1950s onwards 18 Japanese firms entered the EM1 sector, half of whom were based in Hamamatsu. Remarkably, half of these companies currently persist today in this industry (Figure 4.12). As Harnamatsu asserted itself as the dominant centre for EM1 production, this new local landscape of instrument manufacturing developed atop the established social division of labour as layers on a palimpsest. Figure 4.13 shows a number of these entrants came from the consumer electronics sector, such as Casio,

Matsushita, JVC and Brother, the latter two performing OEM for beleaguered American organ makers Baldwin and Lowery. As suggested earlier, of these transectorial rivals, only Casio has provided any sort of competitive threat to the likes of Yamaha and

Roland. Figure 4.12: Electronic Musical Instrument Manufacturing in Japan: A Timeline of Industry EntranceIExit 1960 tam 1980 lOQb 2000

H: Harnamatsu D: E. Organ T: Tokyo : E. Piano 0:Osaka A : E. Keyboard N: Nagoya : Synthesizer Y: Yamagata M: Matsumoto

Source: Translated from Takeo, 1989, The Music Trades, December 2000: 80-104. I Figure 4.13: Hamamatsu EM1 Factory Location in 1986 (Source: translated fiom Takeo 1989:62)

4.7.1 The Casio Challenge

Casio took the model for its successful colonization of the digital watch sector in 1974 and applied it to the problem of musical instruments. Indeed, according to Casio's US national sales manager, Bob Larson, "the very first R&D that was done was to determine if the mathematical discipline could be applied to producing instrument sounds

(Armbruster 1983: 21 emphasis added)." Significantly, one of the firms founding brothers, Kashio Toshio, was instrumental in steering the family firm into the field of musical instruments (The Music Trades 1985). Indeed many of Casio's early patents in

EM1 were invented by Toshio. Casio keyboards were small, inexpensive, and unashamedly geared to the amateur in their toy-like configuration, yet they practically defined a whole new market segment, that of the mini-keyboard. The distribution of these portable, battery operated models in the American market proved particularly innovative as Casio broke with a longstanding industry convention which limited the sale of instruments only through music retailers. Boldly selling its keyboards in the US market in such unorthodox retail spaces as Kmart, Casio, more than any other Japanese firm, were responsible for the widespread global diffusion of keyboard instruments on a massive scale (compare Figure 4.14 with Figure 4.8).

Figure 4.14: Sales of Keyboard Units by Japanese Manufacturers - The Casio Effect

8000 7000 $ 6000 rn Synthesizer 3 .O 5000 E Keyboard C 4000 E Organ a 2 3000 E Rano I- 2000 1000

(Source: The Japan Music Trades 200 1)

In 1982 Bob Larson observed that,

We feel that the majority of people who bought Casio products in department stores had never before been in a music store and prior to purchasing a Casiotone, had never before had any exposure to music. (The Music Trades 1982: 42)

Though Yamaha and other Japanese instrument makers declined to follow Casio's lead on the distribution side, they grudgingly accepted the huge potential this market offered and soon introduced portable models of their own. So it came to pass that Casio's entry to the market, by deliberately targeting the lower end of the market through the formation of alliances with non-traditional retailers, had the effect of enrolling the public on a massive

scale into the Japanese EM1 network.

4.7.2 Post-MIDI Consolidation

Table 4.8: A History of Inter-Firm Relations, Mergers and Acquisitions in the EM1 Sector Ace (J) designslproduces organ for Matsushita (J) Ace (J) enters production joint venture with Hammond (US) Roland, Yamaha, Kawai Korg (J) and Sequential Circuits, Oberheim (US) collaborate on MIDI format JVC (J) designslproduces electronic keyboards for Lowery Organs (US) Brother (J) designslproduces electronic keyboards for Baldwin Piano Company (US) Suzuki (J), Oberheim's (US) Japanese distributor starts building keyboards under AEM contract that expands to include design functions

Drum machine maker Linn Electronics (US). . folds, Roger- Linn starts working for Akai (A Yamaha purchases 51% of Korg (J) Yamaha purchases Sequential Circuits (US), Sequential R&D staff start working for Yamaha Korg (J) purchases Yamaha's US R&D division - former Sequential staff now work for Korg Roland (J) purchases Farfisa Organs (ITL) Roland Purchases Rodgers Organs (US) Kawai purchases Lowery Organs (US)

I Suzuki (J) purchases Hammond Organs (US) (Sources: Kakc :hashi 2002, Milano 1993, Johnstone 1999, Dave Smith (former president of Sequential Circuits, Interview 11/10/02), www.synthmuseum)

The previous chapter discussed the introduction of the MIDI protocol, noting in particular

the potential this standard had for enabling horizontal integration in the industry.

Japanese manufacturers, especially the three Hamamatsu firms who participated from the

beginning, were in a prime position to reap the benefits of their collaboration. However,

the shakeout of the industry, understood in evolutionary terms as the inevitable

consolidation around a best practice following the selection phase, did not occur until the

mid-to-latel980s. It was at this point that US firms, relying on, at best, a few product launches per year, found themselves in especially vulnerable financial positions if their products were not immediate successes. Beginning with the demise of the drum machine pioneer, Linn Electronics, a series of US firms on the brink of bankruptcy were taken over by their Japanese counterparts. Table 4.8 shows that in some cases, as with Linn and Sequential Circuits, the product brands would be retired while the intellectual capital and tacit knowledge residing in the engineering and R&D departments would be absorbed by the acquiring Japanese firm. In the others, only the brand was purchased

(Harnmond) while in others the object of acquisition was another type of asset entirely.

One American respondent noted that Kawai "bought Lowery for its f~rniture"~~;the rationale being that in the digital era, the sound generating mechanism of organs had been miniaturized onto a few microchips embedded in a circuit board. As a result, most of the material value of these instruments was embodied in the wood surrounding the electronics. Roland's purchase of Rodgers Organs can be viewed in a similar light. One other important acquisition quietly took place in 1988. This was Yamaha's purchase of a controlling share in the Tokyo-based Korg Corporation. In chapter six I discuss this particular case in greater detail, especially regarding the way in which the intellectual resources of the former Sequential Circuits' team were transferred between these enterprises.

4.8 Conclusion

By the start of the 1990's Japanese firms dominated the global musical instrument industry, not only in the EM1 segment, but in virtually all instrument segments. This dominance was borne of endurance, and from the narrative presented thus far it appears

Interview with Ralph Deutsch, Inventor, Los Angeles 7/10/02.

192 that it is the stability of these firms that distinguishes them from the American industry.

Strong and sustained management - remember that Kawai and Casio are still family- firms; a succession of Kawakami's headed Yamaha from 1929-1989; the long-standing leadership of Kakehashi Ikutaro remains at Roland - have endowed these firms with a far- sighted approach to the pursuit of new markets, often consuming their rivals along the way (Figure 4.15). In this manner they have endured by championing the role of transectorial learning to their organizational strategy and structure.

Specifically, by applying digital technology to instruments other than synthesizers, these firms have challenged the definitional stability of instruments such as pianos, the design of which has remained static for at least the last 120 years. Yet in accordance with evolutionary theories of technological change, in particular the TEP model, the ultimate impact of radical technologies such as digital synthesizers has proven far more pervasive. Indeed, as noted in the vignette at the start of this chapter, computers and cellular phones routinely incorporate these electronic sound technologies, thereby closing the loop that began with the adoption of ICT by the music industry 40 years ago.

Hamamatsu-based firms, most notably Yamaha and Roland, though not the impetus of this trajectory, have been instrumental in its fulfilment. By looking to the longitudinal profiles of these firms, especially the vertical dimension illustrating the internal technological linkages between product groups, it is possible to obtain sense of how these circuits of technological learning evolved. However, industrial dynamics, in this case the wholesale shift of industry from the US to Japan, involved more than merely new product development. In Latour's (1987) terms, these product technologies were essentially

'black boxes' of 'ready made' science that obscure much of the 'messy practice' responsible for their success. In the following three chapters, I analyze respectively patenting statistics and patenting practices to show that long before Japanese firms became ascendant in the production of EMI, they cultivated an R&D infrastructure that was far more attuned to realizing the radical potential of EM1 within the context of the

ICT paradigm. CHAPTER FIVE: PATTERNS IN THE PATENT RECORD: GEOGRAPHICAL AND TECHNICAL SHIFTS

5.1 Introduction

This chapter interprets the geography of technological change in the musical instrument

industry from the standpoint of patent statistics. Specifically, it undertakes an analysis of

the 2,375 US patents in the field of electronic musical instruments (EMI) that were

registered between 1965 and 1995. Relative to the preceding and ensuing sections of the

thesis, this chapter stands out because it deals almost exclusively with numbers.

Nevertheless, this chapter is pivotal in terms of the thesis' overall objectives. Empirically,

it is important because it shows that regional patterns in the patent record illustrate a shift

in the volume of registered inventions from the US to Japan, and in particular to

Hamamatsu. Secondly, it looks at these same numbers at the enterprise level to show that

the dynamics in the geography of invention were driven by highly variable corporate

patenting propensities and ways of distributing invention within the firm. Thirdly, through a filtered analysis of significant, highly cited patents, it indicates the importance

of transectorial sources of innovation. Conceptually, the chapter draws on the

evolutionary economic (Freeman et al. 1982, Freeman 1987, Freeman and Perez 1987)

and, to a lesser degree, science studies (Latour 1987, 1999, deLaet 2001) literature to

frame its interpretation of the data. These perspectives, which share the view of patenting

as a social process, help to explain the broad shift in patenting to Japan as both a

reflection of the role that patenting plays within Japan's National Innovation System (Freeman 1987), as well as the outcome of highly contingent and relationally constituted individual and organizational geographies. From this standpoint, the patent analysis undertaken in this chapter hints at underlying social and institutional processes that are investigated more thoroughly in subsequent chapters.

The chapter begins by briefly reviewing how evolutionary economics and science studies approaches have informed geographers' views of patents. These two bodies of literature are drawn together by emphasizing the proprietary nature of patents. In this light, patterns in the patent record reflect varying underlying individual and organization approaches to mobilizing knowledge as property, more so than they relate directly to innovation per se. The discussion then provides an overview of conceptualizations of

Japanese approaches to patenting (Granstrand 1999). In the following empirical part of the chapter, I present and comment on the EM1 patent data as they are organized regionally, by firm and according to their significance for influencing later patents.

Following this is a discussion that compares the observed patterns in my study with other empirical literature. The chapter ends by engineering a segue that links the patent analysis to subsequent chapters' efforts to closely investigate patenting and other social practices of engineers.

5.2 Conceptualizing Patents

5.2.1 Patents as Proxy for Innovation (as a Social Process)

Patent statistics have long been used by evolutionary economists as proxies for innovation. Indeed much of the post-Schumpeterian research into the nature of long- waves incorporated patent analysis to assess relationships between technological change and economic development (Schmookler 1966, Baker 1976, Mensch 1979, Freeman et al. 1982, Clark et al. 1983). As Freeman et al. (1 982) point out, much of these debates turned

on the definitions and empirical elaborations of 'radical', 'basic', or 'fundamental'

inventions and innovations as well as their timings. One of the points of contention

between these authors concerned the (often subjective) determination of significant

inventions. In accordance with these studies, my analysis recognizes a distinction

between important and unimportant patents, but it derives this categorization objectively

by assigning a numerical threshold for significance.

Until recently, geographers have primarily followed the evolutionary economists

by using patents as proxy for innovation. In these studies, comparative patent

performance has been viewed as a measurement that sheds light on the innovative potential of regions (e.g. Pred 1966, Ceh 1996,2001). By drawing on patent statistics for

their research, these authors recognize the limitations of patents as data. For instance, one

of the major criticisms of patents is that they are only partial indices of novelty. Indeed,

firms employ many different strategies to make inventions proprietary. Some enterprises

might elect to patent a technology, while others prefer to maintain this knowledge as a trade secret. Moreover, even if an invention is patented, it may lie dormant and un-

commercialized - uncoupled from innovation. There are two reasons why this may be the

case. The first is that it may be too impractical to develop into a viable product. Or, this

dormancy may be the reflection of a deliberate strategy to block-off technological space

from rivals. Regardless of the reason why they may remain un-commercialized, this

aspect of patents leaves them open to criticism as a weak source of data. Ceh (2001 : 300)

addresses these concerns by pointing out that, Arguably, the more pertinent an invention, the more likely a company will acquire additional, supporting patents (which typically are never commercialized) to protect their new technology. As such, unfiltered patent data has a weighted element built into it, and researchers can consider it for data analysis rather than treat the information as a drawback.

Despite Ceh's argument supporting the utility of unfiltered patent data for research purposes, I suggest that it is still prudent to evaluate this claim of 'weightedness' by additionally developing tests to gauge significance. Indeed, I show how such a test for patents in the electronic musical instrument industry bears out the validity of Ceh's intuitive claim. The partiality of patent data must further be assessed with the recognition that different nations, sectors and firms exhibit varying propensities to patent.

5.2.2 Patents as Problematic Inscriptions

Recently, patents have been considered by a different geographical literature that works from standpoint of science studies (Winder 1995,2001, deLaet 2001, Doe1 and Rees

2003, see also Stirk 2001). These scholars begin from the assumption that patents are problematic and then investigate why this is so. This literature is less interested in patents as data, in the numerical sense, but rather in patents as texts, and the practice of patenting as a situated social process. From the standpoint of science studies (Latour 1987, 1999) patents are interpreted as inscriptions that translate the interests of inventors and their assigning firms ('allies') into the 'public' realm. In this sense they are akin to other instruments that actors use to 'mobilize the world' (Latour 1987: 223). Moreover, as deLaet (2000) suggests, patents can be considered to be 'immutable mobiles' (Latour

1987), or a particular type of instrument that maintains its form in various contexts. Indeed their immutability enables firms to 'act at a distance'. However, as she also notes, this conception is problematic, because the property rights associated with patents are

less mobile than the inscriptions themselves. Consequently within and among various

contexts, the 'tenacity' of patents may have as much to due with the intrinsic quality and

originality of the technical claims as it does to the strength of the network that upholds

these claims.

This chapter frames its analysis of patents by integrating the lessons of

evolutionary economics and science studies. In this regard I am sympathetic to the former

literature's recognition of the role that patents play in shaping the technological

trajectories of industries. Indeed, following Freeman et al. (1982: 66) who suggest

attaching, "less importance to the statistical clustering of basic innovations and much

more to their linking together in new technology systems", I am guarded about reading

too much into raw patent tallies without additional qualitative investigation. Patterns in

patent data, for instance temporal and spatial clustering, do not yield direct insight into

innovation as much they highlight regularities in approaches to inscribing novelty as

intellectual property. This view of patents as property connects to the broad aims of

science studies which seeks to work backwards and forwards from these regularities in

order to trace the networks of heterogeneous interests that come together to mobilize

knowledge in a proprietary manner.

5.2.3 Patents as Property

A patent is an officially inscribed right of ownership for a particular invention. As such it

constitutes a covenant between an inventor and the government whereby the latter

provides the former with an incentive to make new industrial knowledge publicly available through the provision of a legal framework that safeguards that intellectual property. Once novelty is proven through a rigorous examination process in which the claims of the inventor are measured against 'prior art', a patent is awarded, thereby offering protection for a limited time. In the US this period is 20 years from the date of filing. Once sanctioned by the state, a patent does not necessarily allow its holder to use, make or sell a product derived from that knowledge. Rather, as one of my informants pointed out, it merely entitles its owner to prevent other parties from using, making, or selling a product that intentionally or otherwise infringes on its claims6'. Patenting strategies are consequently regarded as being defensive as opposed to offensive in their orientation. As Blomley (2003: 121,2004) notes, property rights, intellectual or otherwise, and their associated spatializations are 'necessarily relational' and must be reproduced via 'enactment'. Thus quite often technological space is cluttered with partially overlapping claims to novelty and any attempt to capitalize on proprietary knowledge will also involve the cross-licensing of adjacent and ancillary claims held by others. In other words the maintenance of patenting regimes is an ongoing negotiation amongst various interests. In the following section, I discuss how patenting practices in

Japan have been institutionalized.

5.3 Patenting in Japan

5.3.1 The Patent Precedes the Territorialization of Industry

Almost a decade before the performance of Japanese automakers precipitated a trade war with the US and subsequent restructuring of the latter's auto industry in the 1980s,

Japanese firms inscribed their presence in technological space. Japanese inventors

65 Interview with Ralph Deutsch, Inventor, Los Angeles 7/10/02. registered twenty-four motor vehicle patents in the US in 1966. This figure was nine percent of the 275 patents published by US firms. By 1975, Japanese inventors had caught up with the US with each country publishing roughly 300 patents (Altshuter et a1

1985; quoted in Freeman 1987: 40). Freeman interprets the trend in these figures as signaling one facet of a broader shift in Japanese firms' approach to innovation that followed the decades of reverse engineering in the 1950s and 1960s. In this regard, the rapid increase in Japanese patenting in the early 1970s constitutes one dimension of the general process whereby,

Japanese management, engineers, and workers grew accustomed to thinking of the entire production process as a system and of thinking in an integrated way about product design and process design. (Freeman 1987:40).

Freeman (Ibid) further qualifies his analysis of Japanese patenting practices by claiming that,

whereas Japanese firms made few original radical product innovations, they did make many incremental innovations

This quotation implies that the Japanese refinement of radical technologies produced elsewhere, is literally underwritten by the systemic practice of patenting. By looking to the origins of this practice it is possible to get a sense of the importance of patents in

Japanese industrialization. Indeed, Takahashi Kunio, who was appointed the first

Director General of the Japan Patent Office in 1885 proclaimed that:

We have looked to see what nations are the greatest, so that we can be like them. We asked ourselves 'What is it that makes the United States such a great nation?' We investigated and found that it was patents, and we will have patents. (quoted in Granstrand 1999:137) A few decades after Lincoln remarked in 1859 that patents, "added the fuel of interest to the fire of genius in the discovery and production of new things" (quoted in

Basler 1953: 357), the Japanese recognized the value of this institution. However, in the case of Japan, the 'interest' in question has implicitly been collective in nature. Indeed, several studies have argued that the Japanese approach to patenting is predicated on a logic whereby the rewards for invention are measured in terms of social rather than individual benefits (Rosen and Usui 1994, Takenaka, 1994, Kotler and Hamilton 1995,

Granstrand 1999). To this day, it is a system designed to foster national industrial and technological development via the dissemination of patent information (Takenaka 1994), and as such may be considered a central, if often overlooked, facet distinguishing Japan's

National Innovation System. For example, the practice of publicly disclosing an application before it is granted and, until 1996, the right of third parties to challenge a patent's claims during the pending period, contrast markedly with conventions in the US

(JETRO 1999). So thoroughly has Japanese industry entwined its economic development with a propensity to patent, that today Japan files five times as many patents as any other country, with only 20% of these applications directed abroad (Kotler and Hamilton

1995). These foreign applications, which secure a position in export markets like the US, command the high ground in certain sectors including the electronic musical instrument industry. Indeed my analysis shows that even in the US market, Japanese firms had eclipsed their American counterparts by the mid 1970s in the registration of patents. I now examine the regional, corporate and individual geographies of patenting in the electronic musical instrument industry. 5.4 Patterns in Patents and the Spatial Dynamics of Inventive Activity

5.4.1 A Brief Note on Regions

In the analysis that follows, the regional aggregation of patenting activity in the EM1 sector employs a subjective definition of the different patent regions in the US and Japan.

Therefore, in the US, the 'East' is taken to mean all states that border on the Atlantic

Ocean; the 'West' includes those states that border the Pacific Ocean, as well as Arizona and Texas. The 'Mid-West', then, includes all the other states that lie between the East and the West. In practice, the Mid-West Patenting Region is far more concentrated in the states of Ohio, Illinois and Indiana, respectively the locations of the major US firms

Baldwin Organs, Hammond Organs and Kimball Organs.

In Japan, patenting regions are assigned as follows. The Kanto Region is narrowly defined to incorporate the Tokyo, Saitama and Kanagawa prefectures. The Kansai region is conventionally defined to include Osaka, Kyoto and Hyogo prefectures, although in practice, the Kanto patent region extends no further than Osaka City and its suburbs - the hearth of the Matsushita Electric Corporation, manufacturer of the National and

Panasonic Brands. This leaves the Hamamatsu region, which is centred on the city of

Hamarnatsu and its environs, but also includes the rest of western Shizuoka prefecture.

The few patents (less than two percent of the regional total) registered to inventors living in Nagano and Aichi prefectures are also included in the Hamamatsu region tally. 5.4.2 Regional Trajectories in Patenting Activity

Table 5.1 US Patents for Electronic Musical Instruments by Region of Assigning Com~anv US US J J J S. Mid-W West Kanto Hamamatsu Kansai Korea Total 138 24 4 9 3 0 242 149 40 13 177 12 0 464

1990-94 1 2 1 10 Total (A) 1 78 1 218 Source: USPTO (http://www.uspto.gov)

Table 5.1 shows the regional distribution of EM1 patents between 1965 and 1994.

Several patterns of concentration and dispersal are apparent. First, there is a pronounced spatial shift in the sponsoring of invention from the US to Japan. From 1965 to 1969, US companies in all three regions accounted for 86 percent of total patents (Table 5.1).

Between 1970 and 1974, however, the US share had fallen to 54 percent, while Japanese firms accounted for 44 percent. Thereafter, the gradual demise in patenting activity by

US based firms from 1970 onwards is coincident with an increase of Japanese assignees, such that from 1990 to 1994, the US share amounted to only 7 percent. Second, the locational dynamics of invention are manifested intra-nationally. In the case of the US, the regions that were dominant until 1984, such as the Mid-West where most of the large organ builders such as Baldwin (Cincinnati, OH), Lowery, Wurlitzer, Harnrnond

(Chicago, IL), and Kimball (Jasper, IN) were located, experienced the most significant absolute and relative declines from the mid-1980s. On the other hand, the contribution of the West in absolute terms remained fairly consistent over the 30-year period. In relative terms, the West had achieved parity with the East and Mid-West by the late 1980s and had surpassed them by the 1990s. In Japan, from 1970 onwards Hamamatsu, home to

Yamaha, Kawai and Roland, was overwhelmingly the dominant region although Kanto

(the greater Tokyo area) firms, in particular Casio, consistently increased their share from

Table 5.2: US Patents for Electronic Musical Instruments by Region of Inventor's Residence

1 990-94 4 10 3 27 109 355 6 7 521 Total (B) 103 233 407 306 253 999 67 7 2375 B-A from Table 5.1 25 15 -96 166 7 -1 06 -1 1 0 Source: USPTO (http://www.uspto.gov)

These same data are also classified according to the inventor's residence (Table

5.2). The broad similarities between the array of this data set and the array shown in

Table 5.1 suggest that in most cases inventors tend to live in the same region as their assignees and are likely their employees. A comparison of regional totals (Rows A in 1 and Row B in 2) reveals small but noticeable differences. At the bottom of Table 2, the row labeled B-A (from Table 5.1) shows that regions with positive values are net inventing regions, implying that there is a portion of inventing activity that is fulfilled for assignees elsewhere, while those with negative tallies are net assigning regions that to varying degrees rely on inventors from outside the region. Significantly, the core regions in each country, the Mid-West and Hamamatsu respectively, farm out a portion of their assignments to outside inventors. Indeed almost 115 of assignments by Mid-Western firms fall into this category. On the other hand, certain regions, the US West in particular constitute net inventing regions, indicating that inventors in places like California had an influence on technological development elsewhere, a fact that takes on greater weight once the significance of their contributions is considered.

5.4.3 National Systems of Innovation v. Corporate Strategies: Comparing the Patenting Performance of US and Japanese Firms.

I now organize these data according to assigning firms in order to provide an indirect vantage point of regional approaches to invention (Figures 5.1 and 5.2). Accordingly, the five most prolific US and Japanese assignees are compared. In the case of US firms, all five companies are based in the Mid-West. Within this sample, industry leadership in patenting shifts over the 20-year period until 1985. For instance, Baldwin is clearly dominant at the cusp between the 1960s and 1970s while The Chicago Instrument

Company, which was later acquired by the Norlin conglomerate (see chapter three), was dominant ten years later. In the case of Japanese firms, very few US patents were assigned prior to 1970, but thereafter, Yamaha consistently leads the industry in terms of its commitment to patenting with Casio becoming more important by the 1980s onwards.

There are significant differences between US and Japanese firms in levels of patenting by enterprise as well as the distribution of patents among inventors within the firm (Tables

5.3 and 5.4). In some companies a small core of very active inventors account for the majority of patents, while in others these efforts are more evenly distributed. It is important to keep in mind that Japanese firms only started patenting in any significant manner from 1970 onwards, while their US counterparts remained active only until the mid-1980s. Yamaha is discernibly the industry leader, with 856 patents from 180 different inventors. Consequently, relative to its competitors, patenting at Yamaha was also far less concentrated amongst its top group of inventors. In contrast, Kawai, the second most prolific Japanese firm, is notable for the concentration of patenting by its top three inventors. In particular, Kawai's top inventor, the former Rockwell employee, skews this figure and accounts for the comparably low percentage of local invention at 68 percent. Figure 5.1: A Comparison of the Top Five Japanese EM1 Manufacturers' Patent Outputs 1965-94

140

120

100

80 R AceRoland patents Matsushita 60 Casio Kawai Yamaha 40

20

0 1973 1977 1981 1985 year

(Source: USPTO www.uspto.gov) Figure 5.2: A Comparison of the Top Five US EM1 Manufacturers' Patenting Outputs 1965-1994

JasperIKimball patents !a Hammond w Wulitzer ChicagolNorlin

1973 1977 1981 1985 1989 1993 year

(Source: USPTO www.uspto.gov) Table 5.3: Patenting Characteristics within the Top Five Japanese Manufacturers 1965-94

Firm Patents Inventors Average TOP Top 3 Percentage of Per Inventors Inventors' Patents Inventor Share of Share of Assigned to Total Total Local lnventors Yamaha 856 180 4.8 5% 12% 97% Kawai 236 5 1 4.6 32% 44% 68% Casio 196 70 2.8 6% 14% 100% Matsushita 40 19 2.1 18% 38% 100% Roland 3 1 19 1.6 19% 39% 87% Source: USPTO(www.uspto.aov)

Table 5.4: Patenting Characteristics within the Top Five US Manufacturers 1965-94

Firm Patents Inventors Average TOP Top 3 Percentage of Per Inventors Inventors' Patents lnventor Share of Share of Total Assigned to Total Local lnventors Baldwin 115 22 5.2 22% 50% 94% Lowery 79 21 3.8 15% 44% 64% Wurlitzer 63 2 1 3 21% 46% 27% Hammond 56 23 2.4 32% 50% 88% Kimball 39 12 3 28% 51% 92% Source: USPTO (www.uspto.gov)

5.4.4 The Geography of Significant Patents

In order to ascertain which regions really drove the process of technological change, I looked within the larger data set at significant patents, here defined as those receiving 25 or more citations in subsequent filings (Table 5.5). If the average patent receives between

6 and 7 citations, then the distribution of this class of "super patents" suggests which regions were influential in steering the technological trajectory of the industry. Three points need to be highlighted. First, the years 1970 to 1974 mark a watershed in the registration of significant patents. Second, the US West, especially California, and particularly at this critical threshold, became a major contributor of key patents. Third, from 1980, Japanese inventors accounted for more than half of all significant patents. It is as though following an episode of radical invention in California in the early 1970s,

Japan and especially Hamamatsu asserts itself as the core region during the following decade.

Table 5.5: Significant Patents: US Patents for EM1 Receiving 25 or More Citations by Inventor's Residence

US US J J J S. Europe East Mid-W US West Kanto Hamamatsu Kansai Korea Total 1965-69 2 1 3 1970-74 3 3 12 3 22 1975-79 3 2 7 1 6 1 20 1980-84 2 1 6 1 9 19 1985-89 1 1 3 7 1 13 1990-94 4 6 6 16 Total 1 10 7 33 8 3 1 2 93 Source: USPTO (http://www.uspto.gov) Table 5.6: The Top 20 Most Cited US Patents for EM1 1965-94

Year Cites Assignee Assignee Assignee's Inventor Orgnztn Location Sector Location 1970 114 Rockwell CA Aerospace CA * 1971 111 Rockwell CA Aerospace CA 1 1974 1 91 1 Rockwell I CA Aerospace CA * 1991 80 Stanford U CA Education CA 1993 66 Pioneer Tokyo Electronics Tokyo 1975 62 Yamaha Hamamatsu M. Instruments Hamamatsu 1977 58 Stanford U CA Education CA 1973 52 Rockwell CA Aerospace CA 1972 50 Rockwell CA Aerospace CA 1966 50 Seeburg IL Jukebox IL 1983 49 Yamaha Hamamatsu M. Instruments Hamamatsu 1984 46 Mattel CA Toys CA 1978 44 Yamaha Hamamatsu M. Instruments CA 1978 42 Hitachi Tokyo Electronics Tokyo 1970 42 Motorola IL ITIElectronics AZ 1975 42 Pioneer Tokyo Electronics Tokyo 1992 42 Casio Tokyo Electronics Tokyo 1975 4 1 NIA NIA NIA PA 1978 41 Wurlitzer IL M. Instruments NY 1982 41 Norlin IL M. Instruments IL

A further gauge of significance is to weight each of these twenty patents according to the number of citations it receives - the top patent has a 114 weight, while the least cited is

4 1. If the sum of these citations is 1,119, then the 1970-1975 period scores a 605 rating, accounting for 54 percent of significance. A regional pattern of concentration largely overlaps the temporal cluster just mentioned. By region, California-based inventors also accounted for almost half this list (nine out of twenty) and so have a significance weighting of 64611 119 (58 percent). Finally, a single inventor, Ralph Deutsch, is responsible for a third of this top group. Deutsch, following a Master's Degree in applied math, employment with the US government on radar systems during the Second World War and his work with Rockwell, was hired by Yamaha and subsequently Kawai in the

early 1970s to advance those firms' understanding of digital systems. His role as a vector

of knowledge transfer to Japan will be discussed in greater detail in chapter six.

5.5 Discussion

These data were classified according to both the name of the inventor and the assigning

firm and were further organized at the national and (sub-national) regional scale. The first

level of analysis, which tallies all of the patents in the data set, implicitly considers all patents, significant and insignificant, to be of equal weight. Unfiltered, these data

nonetheless conformed to a pattern which clearly indicates a broad shift in patenting performance from the US to Japan. At finer scales of resolution and at the corporate

level, Hamamatsu-based firms and especially Yamaha asserted their dominance from the

early 1970s onwards. The second pass through the data compared the patenting

performance of the top five US and Japanese corporations in order to determine firm-

level statistics, such as the number of different inventors active within each enterprise,

the concentration of patents belonging to a 'top group' of prolific inventors, as well as the

degree to which firms assign patents to inventors located outside of their home region.

The third frame of analysis was concerned with identifying those significant patents that

exerted a greater influence on the sector's technological trajectory. This analysis,

undertaken by measuring the number of citations a patent received in subsequent patent

filings, revealed a different geography of invention that was far more concentrated in a

few firms and regions. Hamamatsu firms, and especially Yamaha once again proved

significant, but their record of improving performance was temporally staggered

according to this metric. Instead of rising to the fore in the early 1970s, we saw that their ascendancy was delayed until the late 1970s. Conversely, the US west, and specifically inventors based in California proved particularly important at the watershed of radical change in the early 1970s. Moreover, these data, when analyzed according to significance highlighted a transectorial dimension to innovation in the musical instrument industry.

Highly significant patents, including the most widely cited patent in the field, came from the aerospace firm Rockwell. The results of this investigation build on the case that

Yarnaha, in particular, practiced far more extensive approaches to the colonization of technological space than the rest of the field. How do these results relate to other literature that interprets patenting practices of Japanese corporations?

Granstrand (1999) reviews the various strategies employed by Japanese firm that help to explain their prodigious patenting activity across a range of industries. In general he argues that high rates of patent registration by Japanese firms in the US can not be dismissed as differences in the quality of these applications pointing out that, "Japan has upgraded the quality of her patents as she has upgraded the quality of her products"

(1 999: 15 1). This observation supports the earlier work of Narin (1 986) who shows how since the mid 1970s Japanese patents have outstripped the competition in their frequency of citation. Freeman (1987: 22) qualifies this aspect of Narin's study, by suggesting that higher rates of citation, while some rough proxy for quality, might be, "explained by commercial as well as by technical logic." Granstrand (1999), points to various strategies that employ high levels of patenting to maneuver in technological space. The first strategy involves gaining access to competitors' technological breakthroughs by

'surrounding' them with one's own patents, thereby engaging the competitor to seek cross-licensing agreements. The second strategy is to obscure an enterprise's own technological priorities by 'blanketing' the technological space around key patents with

minor derivative and ancillary patents (Granstrand 1999:220). Through the recognition of these practices it is possible to make sense of Japan's high patent propensity. Moreover, it

is important to keep these strategies in mind in the coming chapters which analyze patenting as a situated practice

5.6 Conclusion

Quite clearly firms in this industry display very different prerogatives when it comes to their patenting propensity. Some firms have opted for highly selective approaches to the

colonization of technological space while others have adopted a blanket approach. In this

latter regard, Yamaha's experience needs to be particularly highlighted, especially relative to specific patenting strategies that might explain this firm's performance.

Yamaha provides the benchmark against which all firms are measured - against

Yamaha's massive patent portfolio all its competitors appear in relief. Yet other firms

still exist, and it is Yamaha's Japanese competitors who provide both evidence of

different approaches to the negotiation of the patent arena, as well as commercial success

in remaining competitive, in spite of what would appear to be a massive advantage

enjoyed by Yamaha. To determine why this is so, the following chapters investigate

patenting as a situated practice that is embedded in particular geographies of the

individual inventors, firms, regions and nations. CHAPTER SIX: THREE WISE-MEN FROM THE EAST: VECTORS IN THE TRANSFER OF TECHNOLOGY FROM CALIFORNIA TO JAPAN

6.1 Introduction

This chapter examines the career biographies of three inventor/engineers who proved of critical importance in the transfer of technological knowledge from the US to Japan.

These are: Ralph Deutsch who, after the invention of the digital organ on behalf of

Rockwell Aerospace - a technology which was developed collaboratively with, and licensed to Allen Organ - worked for both Yamaha and Kawai; John Chowning, the

Stanford composer of computer music whose patent for FM synthesis was licensed exclusively to Yamaha, spawning a powerful relationship between the university and

Yamaha that exists to this day, and; Dave Smith, the entrepreneurial founder of

Sequential Circuits who played a lead role in the collaborative development of the MIDI

(Musical Instrument Digital Interface) protocol and who subsequently worked for both

Yamaha and Korg. All three of these engineers forged linkages with Japan from a base in

California and they each had problems enrolling the interests of US 'allies'. Apart from these contextual similarities, their stories highlight different themes. For instance,

Deutsch's story raises the problem of how personal conflicts reshape technological trajectories, Chowning's tale points to the way that major academic institutions like

Stanford University constrained and enhance the efforts of talented individuals, while

Smith was unique in his role as an entrepreneur. Conceptually, a focus on career trajectories allows for an investigation of how transectorial innovation, as the 'diffusion of the engineering disciplines' (Rosenberg

2000), works in practice. Chapter two concluded with a 'local model' of transectorial innovation across space that referred in general to the exogenous dimension of Japanese industrialization and specifically to the case of the nascent electronic musical instrument

(EMI) sector (Table 2.2). From an evolutionary standpoint, the key moments in the trajectory revolved around a small number of 'technical crystallizations', or decisive demonstrations of technical feasibility that, once appropriated by a particular firm or region, served to 'lock-in' industrial advantage. Geographically, these moments constituted the critical thresholds that enabled a shift in an industry's centre of gravity within an evolving 'spatial innovation system' (SIS) (Oinas and Malecki 2002). Each of the three engineers profiled in this chapter played the role of protagonist at these pivotal moments. Indeed, these career-defining accomplishments have been well documented

(Vail 1993, Markowitz 1989, Chadabe 1997, Theberge 1997, Johnstone 1999), even if the emphases in these narratives are on the technical importance of the instruments or techniques they invented, rather than the contingent circumstances in which they were embedded. The objective in this chapter is to frame these radical innovations through the lens of science studies (Latour 1987, 1999) in order to recover the biographical

(individual) and social (contextual) aspects of the story. My argument is that particular radical transectorial innovations derive their significance not from their intrinsic technical qualitiesper se, but rather from the strength of linkages that bind them to other instruments, colleagues, allies and the public (Latour 1999, see Figure 2.1). The narratives I present emphasize the 'chains of translation' that both gave rise to particular technical accomplishments of the US engineers, and served to unmoor them from their context of origin.

Topologically, these engineers, as 'vectors' of technology transfer, connect the demise of the US as an industrial core and the ascendancy of Japan. Indeed in all three cases, potential US allies balked at the opportunity to enrol the engineers in a domestic network. The 'slippage' that took place as the different parties 'translated' their heterogeneous interests proved too great, despite the geographical proximity of these allies. Moreover, this failure to connect recurred in each case and, in hindsight, is especially surprising since the inventors were working in institutional settings that are normally associated with California's post-war ascendancy as a centre of technoscience - settings that include the US aerospace industry and Stanford University. In contrast,

Japanese firms proved eager to wed their interests with these US engineers; to mobilize the latter's inventions. This chapter problematizes the inadequacy of these local linkages while offering an explanation of why Japan proved to be a 'better fit'.

In constructing these narratives I relied on my own interviews with Ralph Deutsch and Dave Smith in addition to secondary sources that include firm histories (Markowitz

1989, Yarnaha 1987), interviews and oral histories contained in trade publications (Vail

1993) and other interpretive accounts (Johnstone 1999). These multiple biographical sources serve as a counterpoint to reflect on the public accomplishments of these individuals (patents, instrument designs, etc.) while illustrating the social (as opposed to technical) dimension of innovation. With these sources in mind, it is worthwhile here to provide a brief theorization of biographies and how they can inform our understanding of the processes of technological change. Various authors have sought to problematize the role of biography in economic geography (Schoenberger 2001, Barnes 2001). Biographies, as Barnes (2001 : 41 5) notes,

"are a melange of fact and fiction - of events that happened, but also of rhetorical strategies, vested interests, and acts of interpretation, selective memory, and wishful thinking." Biographies of key individuals connect career trajectories with corporate interests in order to contextualize the harmony and dissonance that arise between the two.

Driven by social processes, the linkages within transectorial SIS are graced by synergies or fraught with conflict. The literature in economic geography has succeeded in highlighting the soft conflicts, such as discordant institutional structures (Gertler 1995), that prohibit technology transfer. However, accounts of hard conflicts, such as formal legal disputes, and personal animosities, have been largely omitted. Conflicts such as cases of patent infringement bring into relief the uneasy tension between people, texts and artefacts, the three conduits of spatial technology transfer (Howells 2002). Career migrations are also a source of conflict. For instance, in reference to Ralph Deutsch, the dust jacket of Markowitz's (1989: emphasis added) autobiography Triumphs and Trials of an Ornan Builder highlights, "How one of the key figures in the development of digital musical technology sold out to a leading Japanese company". One of Markowitz's chapter titles is: "Ralph Deutsch and the Dark Side". In basing my own narrative on conflicting sources, is it even possible to reconcile the cases where account of 'what happened' run counter to each other?

As Barnes (2001: 426) notes, "if science studies are to use biographical accounts then their nature must be recognized." He argues that the biographer must negotiate between the individual and their context (human agencylstructure). Moreover, there needs to be recognition that biography is the re-invention of lives in a manner that

Livingston (1 999, quoted in Barnes 2001 : 4 15) calls 'controlled fiction'. Third, the position and intention of the author must also be clear. In other words, the reader must recognize that the accounts that follow are a triangulation between: i) the 'facts of life', the texts and artefacts (patents, inventions, instruments) that mark the signposts of engineering lives as they were 'lived' (Barnes 2001); ii) the 'lives told' (ibid), which are the biographical material which reflects on these accomplishments, and; iii) my own interpretation and (re)construction of a narrative that fits with the themes of my project and aims to 'accurately' cover processes that took place before I was born.

Judged from the present, Deutsch, Chowning and Smith are pioneers of the digital era and their accounts are absolutely essential to understanding the process of transectorial technology transfer in the EM1 industry. Moreover, they represented the first wave of US engineers to forge a connection with the spatial innovation system that was to become centred on Hamamatsu. Geographically, their stories can be contextualized relative to past migrations of engineering talent to Japan.

6.2 Foreign Experts and Innovation in Japan: Historical Antecedents

The notion of foreign experts introducing technological knowledge to Japan is certainly not new. Even during sakoku - the phase of national isolation from 1639-1858 - foreign knowledge, for instance in the field of medicine, was permitted to enter the country via a handful of local ports, primarily Dejima in the city of Nagasaki in the north-west part of

Kyushu Island. Until this era drew to a close, the morphology of Japan's spatial innovation system hinged on the role these critical entrepots played in the regulation of the knowledge economy for the rest of the country. Apart from the principal connection between these places and the capital Edo, regional daimyos (feudal lords) had diverse geographies of access towards the foreign knowledge that moved through Nagasaki.

Proximity mattered, so clans located close to Nagasaki, such as the Shimazu of

Kagoshima, and the Hozokawa of Kumamoto, gained much of their power by exchanging their rice wealth for technology (Tamaki 2001). During the bakumatsu era which bridged the Tokugawa (1 603-1 868) and Meiji (1 868- 1912) periods, this power challenged the hegemony of Edo (Tamaki 200 1).

At the Meiji restoration, when Japan embarked upon its rapid modernization and attempt to 'catch-up' with the west, foreigners once again played key roles as vectors of knowledge transfer. From the Japanese perspective, these individuals' knowledge represented a mobile manifestation of specific national models that at the time were considered to be 'best practice'. Various authors have detailed the role of the some 3,000 foreign experts who were hired by the Meiji government to advise in the establishment of institutions ranging from the banking and postal systems to private commerce (Jones

1980, Westney 1987, Pedlar 1990). Very often these individuals were selected from specific countries which contained a particular institution from which Japan sought to learn. Thus major cases of organizational emulation included British mining and engineering, the military, postal and postal savings systems, the French primary school and legal systems, the German secondary school system and the American national bank system (Westney 1987: 13).

These foreigners, known collectively as yatoi, acted principally as catalysts in

Japan's rapid modern development. In many cases, once their utility to a particular project had been fulfilled they were dispatched back to their country, Japan having absorbed their tacit knowledge. From a science studies perspective (Latour1989), the waves of yatoi who brought their knowledge and talents to Japan, and the Japanese envoys such as Fukuzawa Yuichi (1835-1901) who returned to their home country with lessons from their travels, engaged Japan in a national 'cycle of accumulation'. These parallel ventures established the basis of Japan's national innovation system as a mediation between exogenous and endogenous forces (Freeman 1987).

After the Second World War, foreigners, American's in particular, were once again solicited to advise the reconstruction of Japan. Joseph Deming's introduction of US quality control practices to Japanese industry is the most famous case of this phenomenon

(Voehl 1995). In contrast to the Meiji period, the post-war pattern of institutional and technological knowledge transfer to Japan was initially a cooperative effort directed by

GHQ. The reconstruction of Japan's economy was deemed desirable as a strategic objective in the early years of the Cold War. Thereafter, in the 1960s and 1970s individual industries took the initiative in enhancing their own technological sophistication by hiring these foreign experts.

6.2.1 Technological Learning in Japan and the Role of Transectorial Yatoi

In its broadest sense, technological learning in Japan has been articulated through a spatial innovation system that has become increasingly refined in isolating the institutional scale from which it wants to absorb best practice. By this I mean that in the early Meiji period, Japanese bureaucrats selected particular nation states for their respective model institutions. At the time the state gave private businesses a hand in selecting more regionally - meaning sub-national - models for adoption. For instance, the

Kyoto prefectural government sponsored an envoy to learn French weaving in the city of Lyon, a textile centre in that country (Toyota 2001). In the post-war period, and especially in the episode that marks the ascendancy of the ICT paradigm, Japanese firms were attune to specific institutions (corporate laboratories, leading engineering schools), specific technologies (transistors, LSI) and especially key individuals who could translate this knowledge into its highest and best form. The difference between the Meiji (1 868-

1912)and late Showa (post WWII) 'economic miracles' is that in the early period, catch- up for technological autonomy was the goal; in the latter, catch-up for technological mastery, a process also known as 'leap-frogging', was the goal. The drive, long-term thinking and associated patience allowed Japanese firms to keep their sights on the leading edge, meaning that in many cases they were in a position to realize the potential of what they saw more deeply than the American firm performing the demonstration

(Partner 1999). This situation also meant that they were more inclined to be persuaded by radical new technologies and the individuals who conceived them.

As the chapter turns to a discussion of the career trajectories of Deutsch,

Chowning and Smith, especially vis-a-vis their relationships with the Yamaha

Corporation, it is worthwhile to consider in what ways their stories map out with the trajectories of the yatoi a century earlier. In the Meiji Period, yatoi were only yatoi if they were in Japan. In other words their identity was relationally constituted and hinged on their presence in a particular place. Conversely, as the following examples show, the geographies of Japanese - foreign-expert relations were more complex. Specifically, they indicate maturation in the way knowledge was accumulated and mobilized by Japanese interests. In particular, the relationships that Deutsch, Chowning and Smith shared with

Yamaha were mediated through inscriptions which ran the gamut of formality from patents to rough sketches. In other words their connection to Japanese 'allies7 turned on the way the latter mobilized the inscriptions that bore the former's name. Additionally the musical instruments spawned by these partnerships sit uneasily over top of the raft of inscriptions that constitute the sector's knowledge base - an issue that comes into relief when charges of infringement are levied. This chapter works through the cycling between tacit and codified knowledge by examining the inter-relations between people texts, and artefacts.

6.3 Ralph Deutsch

With one notable exception (Markowitz 1989), you will not find Ralph Deutsch's name in any of the books that chronicle and interpret the evolution of electronic musical instruments from an Anglo-American perspective (e.g. Chadabe 1997, Theberge, 1997,

Pinch and Trocco 2002). Deutsch's technical accomplishments are, however, evident through analysis of the patent data, that is through his inscriptions. Not only has he registered the most patents in the field of electronic musical instruments, he invented a number of highly influential patents, including the most widely cited patent in the field.

Something of his career trajectory is also apparent from the patent records (Table 6.1). Table 6.1: A synopsis of Ralph Deutsch's Patenting Record

1 Assigning Firm 1 years I Number of 1 Number of 1 Legal Representation 1 patents with co- (Patent AgentIAttorney) I inventor N.A. Rockwell 1970-74 9 Ia Haman & Hurnphries Deutsch 1974-79 17 "~owardSilber (Flarn & Research Labs Flarn); *' Ralph Deutsch Yarnaha Howard Silber (Flarn & 1 1974-77 1 2o 1 3v I Flam) Kawai 1977-89 94 236 Ralph Deutsch 4 firms 19 years 140 patents 32 co-invented 3 different legal representatives a: George Watson, employee of Autonetics Division Rockwell p: ~eslieDeutsch, son bf Ralph Deutsch y: Glen Grifith, Tustin, CA 6: Leslie Deutsch, son of Ralph Deutsch Source: USPTO (http://www.uspto.gov)

Rockwell, Yamaha, Kawai and his own concern, Deutsch Research Labs have taken turns assigning the technologies he invented. Collaboration with a succession of co- inventors, including his son Leslie Deutsch, accounts for just under one-quarter of his total output. One other feature, that will take on greater significance shortly, is that at varying points he has been legally represented by different patent attorneys and agents, most notably himself in the latter portion of the record. In other words, his interests as an inventor have been translated through an evolving set of relations with various 'allies'.

Indeed, the only constant in this restless career of generating intellectual properties is his address in Sherman Oaks, California and it is at that location, his home, where I interviewed Deutsch in October 2002.

The books lining the shelves in his office, inscriptions once again, add another level of depth to the story. One bookshelf contained volumes pertaining to aerospace including Deutsch's own book, The Orbital Dynamics of Space Vehicles (1963)~~,which was perched alongside its Russian translation. Another shelf contained books in Japanese, including language instructional texts. Finally, there was a wall devoted almost entirely to books about patent law. The patent records indicate that Deutsch represented himself as a patent agent, somewhat of a rarity in this field, or any other for that matter, and these volumes contained most of the knowledge he had had to absorb to engage in this practice.

Over the course of his career, Deutsch had accumulated a stock of knowledge that allowed him to mobilize the worlds of astrophysics, Japanese language and patent law.

Twenty-eight of Deutsch's patents have been co-invented with his son, Leslie Deutsch.

According to his father, Leslie is the musician of the family; competent at playing the tuba as well as the organ. Leslie Deutsch is currently the organist at CalTech University.

More importantly, Leslie Deutsch is as distinguished a scientist as his father. A PhD in physics has prepared him for his current occupation where he is the chief researcher at

NASA's Jet Propulsion Laboratory at CalTech. In this capacity he led the NASA team that rescued, remotely, the Galileo space craft on its mission to Saturn, a feat for which he received an audience with the

66 Deutsch, R. 1963. The Orbital Dynamics of %ace Vehicles Englewood Cliffs, NJ: Prentice Hall. 67 Leslie Deutch's biographical details obtained from the American Institute of Physics web site: htt~://www.ai~.org/aiv/comorate/1998/bios/deutsch.htm(obtained September, 2003) 6.2: Ralph Deutsch and the Story of the Rockwell Patent Event details Knowledge formationltransfer Ralph Deutsch's early training includes: Experience is gained in a carrier sector of the ICT paradigm MSc (Applied Math) Employed by US Government during WWll on RADAR development Post-war work in aerospace sector: (Sperry Rand, Hughes Aviation and Rockwell) Rockwell's seeks to develop commercial applications for its MOSILSI circuit Transectorial implications of digital knowledge initially go technology unrealized because organ makers own tacit knowledge is not Deutsch directs initiative to contact a number of US organ manufacturers in order suited to the interpretation of codified knowledge they are establish a joint venture to develop a new musical system based on digital shown technologies Allen saw, "the radically new technology associated with the All except for Allen Organ (Macungie, PA) decline. space program." Joint venture aareement sianed between Rockwell and Allen Organ Spatial and sectoral transfer of knowledge from West to East Rockwell unde;akes engineering design and manufacture of MOSILSI circuit coast (Pennsylvania) and from aerospace to instrument design devices Agreement hinges on the division of responsibilities and Allen gains right of first refusal to the new digital organ technology, in return for returns inherent in the codification process funding development through cash and securities transfers The significance of this knowledge makes it the Rosetta stone Rockwell patents resulting from this J.V. prove to be fundamental to the for the subsequent technological trajectory of the musical development of virtually all subsequent EM1 instrument industry US Patent #3,515, 792 becomes most widely cited EM1 patent Rockwell and Allen disagree on details of technology transfer but resolve to Aareement's fixedness makes it a barrier to bridge-. separate continue collaboration in order to bring technology to market. worlds of situated tacit knowledge. Allen's Computer Organ presented to public at National Association of Music Yamaha pursues access to radical knowledge Merchants Show Deutsch introduces Yamaha delegation to Allen Organ. Allen sues Rockwell over differences in their contract and litigation ensues Contested claims to intellectual space again shift the frame by Settlement is reached at the trial and Allen agrees not to exploit the matter in a which knowledge is put to use manner that would be detrimental to Rockwell Spatial transfer of knowledge Rockwell transmits the ownership of certain patents to Allen. Deutsch leaves Rockwell and enters a three year contract with Yamaha Spatial transfer of Deutsch's tacit knowledge to Yamaha Deutsch starts contract work for Kawai Potential knowledge transfer between firms in Hamamatsu Allen and Yamaha discuss possible conflict between 'Rockwell' patents and Conflicting claims to intellectual space Yamaha's development plans in the EM1 field Virtual licensing agreement reached in April '77 Yamaha serves Allen with motion asserting invalidity of patents in May. Allen Yamaha seeks to re-regulate key knowledge in its favour counter-sues. 0 Yamaha settles in November for $1-2 million and licenses Allen technology. Allen enters licensing agreements with 6 other Japanese manufacturers. Knowledge diffuses once more Sources: Ralph Deutsch, Interview, Sherman Oaks, CA 12/10/02, Markowitz 1989 The outline of Ralph Deutsch's story can be found in Table 6.2. Its details remain highly controversial - Deutsch's most cited patent (US # 35 15792) has been the subject of numerous lawsuits involving several different firms. Basically, Deutsch acted as a key conduit by which radical knowledge developed in the aerospace sector entered the musical instrument industry through Yamaha. That knowledge then became foundational in the ascendancy of other Japanese manufacturers, as the technologies he authored were licensed by these latter firms.

6.3.1 The Deutsch - Rockwell relationship

Deutsch's academic training and early career placements allowed him to cultivate a body of knowledge in applied mathematics that appeared perfectly suited to the organizational objectives of the aerospace firm North American ~ockwell~~,particularly the firm's electronics division Autonetics to which he was assigned. At the time Rockwell was one of the lead firms in the industrial consortium contracted by NASA to propel the US past the Soviets in the space race. Throughout the 1960s much of Rockwell's revenue was derived from contracts that included the guidance systems of the Minuteman missile and the electronic components of the Apollo landing craft. In 1964, Autonetics was

Rockwell's largest division, employing over 36,000 people. However, by the late 1960s' even while the firm was perfecting the micro-circuitry for the moon-shot, it was clear that the potential for sustained revenues from the military and NASA was limited.

Consequently the firm was hungry to extend its acquired knowledge to civilian applications and it was this task that constituted the re-focused divisional mandate of

Autonetics. One of Autonetics' first projects in this capacity was a $30 million contract it

68 North American Rockwell is the product of the merger between North American Aerospace and automotive parts maker Rockwell-Standard. signed with Hayakawa Electric (later to be renamed Sharp) to design and build LSI MOS circuits for the latter's calculator di~ision~~(~ohnstone1999). Within Autonetics, Deutsch was assigned to lead a team of engineers in a project to produce digital circuitry for the musical instrument industry

Deutsch filed a patent for the digital organ on August 8th1 96770. After a year-and- a-half of development which included the registration of derivative and co-dependent patents by Deutsch and fellow Rockwell engineer George watson7', Rockwell presented its idea around to the perceived beneficiaries of this technology, namely the large US organ manufacturers. Almost all of them, including Rodgers and Conn declined. In other words, even as the technology he demonstrated stayed the same - as an 'immutable mobile' (Latour 1987) - Deutsch failed to translate his (and Rockwell's) interests to the circumstances of these enterprises - the 'slippage' proved too great and these potential allies remained outside the network. Finally Allen Organ of Macungie, PA, agreed to pursue a relationship with Rockwell. Following a visit by its president, Jerome

Markowitz, to Rockwell's massive complex in Anaheim, California, the two firms signed a licensing agreement which outlined the responsibilities of each partner - a translation whereby Rockwell would undertake design and production of the MOSILSI~~devices while Allen would finance these endeavours in return for a right of first refusal to utilize the new technology.

Almost fiom the start, this agreement was mired in controversy that hinged on the contested claims of individuals and organizations to control technological space via the

69 This relationship between Sharp and Rockwell remains tight to this day - Sharp fax machines contain circuit boards mounting Rockwell chips (Partner 1999). 'O US # 3,515,792 71 US Patents: # 3,610,799; 800; 805 and 806 72 Metal oxide silicon large scale integrated circuit. manipulation of codified knowledge. These differences emerged precisely because each

side, Rockwell an aerospace firm, and Allen an organ manufacturer, perceived the technological horizon borne through their partnership in fundamentally different ways.

For example, from the standpoint of Allen's President, Jerome Markowitz (1989: 77) when drawing up the initial contract, "we expressed our requirements in the technical

language familiar to us and Deutsch 'translated' these requirements into the 'foreign'

language of the new technology." Markowitz (ibid: 70) colourfully refers to this language

as, "Anaheim tech-speak".

The outcome of this translation was inevitably a compromise, involving the

'slippage' of each others' interests. A the stage when Rockwell presented the first

engineering model to Allen, Markowitz was indeed impressed with the digital circuitry

but dissatisfied with the organ's tonal expression. Interpolating between the accounts

provided to me by Deutsch and that found in Markowitz's memoirs, it is clear that over

the course of the ensuing negotiation the personal differences between these two

heightened to the point that future relations beyond the letter of the contract became

highly unlikely. Nevertheless, the tenacity of the contract maintained the translation

between these increasingly fractious interests. Eventually Rockwell delivered Allen its

digital organ and, despite the early technical difficulties and escalating animosity between

the two principals, it proved an immediate sensation on its release in June 1971.

Coincidentally two Rockwell technologies shared the limelight as fundamental inputs to

the world's first digital consumer products: the Sharp calculator and the Allen Organ, the

latter winning accolades from Industrial Research as one of the hundred best products of

the year (Markowitz 1989). The technical advantages of the new organ were obvious, with Deutsch commenting that prior to the introduction of digital technology it took

Allen one week to test its analog organs in comparison to the one hour it took for the digital ones. More importantly Allen's ownership of the digital organ technology proved vital to the firm's long-run fortunes. So foundational is this technology that competitors have had no choice but to sign licenses with Allen, the royalties undoubtedly contributing to the fact that in 2004 Allen remains the only major organ company that is still

American-owned.

If English-speaking authors have overlooked the contribution of Deutsch in the development of EMI, the same cannot be said of Japanese commentators. Yamaha's corporate history (Yamaha 100 Nenshi 1987) and Nakagawa (1984) both ascribe importance to his catalytic role. Nakagawa, in his volume detailing the trajectory of

Yamaha's LSI program, provides a detailed description of the way in which Deutsch's knowledge was introduced to Yamaha's system. From these accounts it is possible to construct a synopsis of Deutsch's contribution to Yamaha's EM1 program.

6.3.2 The Deutsch - Yamaha relationship

Yamaha became aware of Allen's Digital Organ at the North American Music Merchants

(NAMM) trade convention in 1971. The significance of this innovation was registered by

Yamaha engineers whose reaction when translated to English was something akin to "My goodness, they've done it." (Yamaha 1987:54) In 1972, Yarnaha hired Deutsch with a three year contract and a broad mandate to come up with technologies for their Electronic

Organs (Electones). These inventive efforts, undertaken at his lab in California, were integrated periodically with those of Yamaha engineers in Hamamatsu. Topologically, in the absence of spatial proximity, the relationship was made to work. According to

Nakagawa (1 984: 85)'

From March of 1972, the HQ Electone division was provided with a new electronic tone research lab for the trial development and manufacture of Deutsch's NES (New Electone System). The core of the NES development consisted of Deutsch and his assistant [Glen Griffith] Deutsch also requested the team include several staff on the Yamaha side including Tomisawa morio] (translation by the author).

While the NES initiative was geared towards a digital solution to the problem of tone synthesis, a parallel program operating out of the newly constructed Toyooka factory focused its efforts on an analogldigital hybrid system known as PAS (passive analog system). These two teams developed a healthy rivalry whereby,

Deutsch's appearance stimulated a competitive spirit at Toyooka's PAS development centre and also raised the aggregate power for developing the logic circuit which had thus far proved to be a major hurdle.. .The PAS team had the spirit and determination that it was not acceptable to lose against Deutsch (Nakagawa 1984: 86-87) (translation by the author).

After a year and a half of this bifurcated development, an internal trial was staged in front of President Kawakami to select between the technologies. The PAS proved superior, although Tomisawa acknowledged that it still needed improvement (Nakagawa 1984:

87). This appraisal fits with Deutsch's viewpoint that Yamaha only partially incorporated the technologies he developed and patented on their behalf. Moreover, Deutsch's remuneration package did not include royalties accruing from these technologies.

6.3.3 The Deutsch - Kawai relationship

During his contract with Yamaha, Kawai approached Deutsch to undertake a similar function with regards to that firm's digital organ program. This turn of events, not surprisingly proved controversial, for when Yamaha caught wind of this development they threatened to sue Deutsch for breach of contract. Deutsch defended against these allegations by citing the non-exclusive nature if his contract. Moreover, as he claimed,

"there was nothing worthwhile [in their relationship] that was not covered by the patents"

Ralph Deutsch personal communication 10/12/02). In other words the relationship with

Yamaha was relationally constituted by the mobilization of texts. Later, when Deutsch's contract with Yamaha expired, indeed even before so, he had assumed the role of a contract inventor employed by Kawai's R&D department.

Deutsch believes that Kawai more fully incorporated his inventions into their products (Ralph Deutsch, Interview, Sherman Oaks, CA 12110102). Kawai assigned 94 of

Deutsch's patents over a ten year period. Deutsch made close to forty trips to Hamamatsu over the course of this relationship. During these face-to-face collaborations Deutsch supervised the construction, debugging and testing of engineering models. In contrast to the Yamaha relationship, he received royalties for the technologies developed for Kawai.

Clearly Deutsch's decision to cultivate the skills necessary to represent his own interests as a patent agent paid off and contributed to the longevity of the relationship. His study of

Japanese, a self-taught skill, also smoothed his contract negotiations with Kawai (ibid).

6.3.4 Ralph Deutsch: An Essential but Controversial Actor in the EM1 Transectorial SIS

There are a number of contentious issues wrapped up in Deutsch's decision to work for the Japanese firms, in particular concerns such as what sort of knowledge he was permitted to transfer to Yamaha. As the author of one of the most fundamental technologies in the field, could he separate his own tacit knowledge that formed the basis for the patent from the codified knowledge represented in that patent? Whether he could or not, Deutsch, R&D staff in the Electone division at Yamaha and R&D division at

Kawai soon published a battery of ancillary and dependent patents that laid claim to a large portion of intellectual space. Complicating matters, instruments that seemed dependent on the digital organ patent started appearing in the marketplace. Allen's objections to these actions resulted in initial and retaliatory lawsuits not only between

Allen and Yamaha, and Allen and Kawai, but also between Allen and other American organ makers who also appeared to be infringing. These lawsuits clearly underline the significance of the disputed knowledge. Different approaches to patenting, contesting patents and licensing in the US and Japan played a role in the far more fluid introduction of knowledge to the latter.

As an example, the lawsuit between Kimball International, Inc. and Allen Organ

Company lasted between 198 1 and 1988 - the latter accusing the former of infringement and the former contending the latter's patent invalid. Much of Kimball's claims of invalidity rested on the fact that Rockwell had 'demonstrated' its digital organ technology to Rodgers and Conn more than a year before it filed for patents #3,610,799 and

#3,610,806. The act of demonstration, Kimball argued, implied that this knowledge was already extant, or in the terms of patent law in 'public use', a condition which if proven would invalidate the patent. Eventually the jury ruled in favour of Kimball, but even that proved a pyrrhic victory for Kimball soon went out of business. Allen's president

Markowitz offered these comments on the industry's response to the knowledge embodied in the digital organ:

It is instructive to compare what I viewed as Kimball's aggressive litigation tactic with the approach taken by the Japanese companies. All of the Japanese companies we approached about licensing chose to limit the money they spent on arguing or litigating. Instead, they obtained licenses and concentrated their time and energy on product development. I believe this is one main reason why the Japanese presently dominate most of the musical instrument industry while most of the older American companies either have gone out of business, have been restructured, or have been purchased by foreign companies. (Markowitz 1987: 1945)

In this quotation, Markowitz is suggesting that though they acted independently, Japanese

firms nevertheless presented common negotiating tactics for obtaining US intellectual properties. This further indicates that the national context shapes approaches to knowledge formation.

Deutsch's keystone patent (# 35 15792) met with critical reappraisal when an anonymous competitor petitioned the US Patent and Trademark Office in 1985 to re-

examine its claims. The result of this scrutiny was that in 1987, one year prior to the

expiry of the patent, every one of its 46 claims were cancelled. Someone had done a lot of homework to signal to the examiner a host of records which in sum alluded to a far more extensive body of prior art than was contained in the original patent document.

Essentially, the originality of the invention was called into question. Articles published in the Journal of the Audio Engineering Society, technical reports published by Bell Labs'

Max Matthews and other documents were brought forward to suggest that Deutsch's digital step was not as radical as claimed. This result is surprising, and Deutsch admits in hindsight that there are conceptual similarities between his early work and that of Max

Matthews. Nevertheless, for close to 15 years, the musical instrument industry operated under the assumption that the Rockwell Patent defined the state of things to come. In this

sense the patent's 'tenacity' (deLaet 2001) proved radical even if the technology amounted to the most officially sanctioned articulation of the state of the art at the dawn

of the digital frontier. How are we to evaluate Deutsch's contribution to the economic geography of the

EM1 industry? His career path traces a time-geography through a number of key firms in both the US and Japan. The SIS could not have evolved without the high level of technological capability that existed in Japan. Yet it was Deutsch's movements that forged the links in a spatial innovation system that is both transectorial and transnational in nature. However, Deutsch's contribution goes far beyond these migrations. Patents bearing his name as the inventor (140 in total), especially key documents such as US

Patent # 35 15792, cast a long shadow over the trajectories of both his assigning firms and those companies forced to navigate the technological space demarcated by these documents. Moreover, his story captures the process that Nonaka and Takeuchi (1995) refer to as the 'cycling' between tacit and codified forms of knowledge. For instance in court, pertinent cases turned on the scrutiny of codified knowledge (specifically the claims in the patents) relative to Deutsch's actions (demonstration) which were surely derived from the tacit dimension. What emerges, from the standpoint of science studies is an uneasy tension between human and non-human actors in this network in which:

inventors initiate novel solutions that improve on prior art, individuals/firms inscribe and thereby mobilize this knowledge in patents, patents allow/preclude individuals/firms from making use of this knowledge products of competing firms infringe on other's patents, necessitating the enrollment of further interests (attorneys, courts, and so on)

To summarize, the contribution of Deutsch to Yamaha's and Kawai's technological trajectory is significant but indirect. Indeed much of the significance derives from the legacy of Deutsch's original work at Rockwell. That both Yamaha and

Kawai have endured legal battles, settled and purchased a license with Allen Organ, the 'owner' of the Rockwell patent, indicates a robustness and longevity to this original radical invention. This 'tenacity' has obliged several other Japanese firms to license the technology from Allen. Consequently, even if they do not directly infringe on its technological space, many of the patented technologies assigned by the latter relate in an indirect way to the radical invention that was conceived transectorially in the crucible of

Rockwell. This indirect aspect is subtle and it is difficult to distinguish exactly how this early, profound accomplishment influenced Deutsch's subsequent work for Yamaha and

Kawai. However, we can say that it was borne of the same tacit knowledge that was as intimately familiar with the technological fundamentals of the digital age, as it was wary of the motivations of others in the patenting arena. Deutsch's mid-career decision to extend his polyvalency and become his own patent agent is evidence in this regard.

Finally, this example demonstrates that, beyond the intrinsic technical qualities of the digital organ patent, this inscription is constituted by it connections with a shifting cast of allies.

Upon conclusion of our interview, Deutsch led me into another room which contained various organs, synthesizers, mixing and recording equipment. Retired from inventing, one of his current hobbies is to play these instruments and record, part-by-part, each of Beethoven's Symphonies. Using machines that harness the technologies he developed, synthesized strings, brass and percussion are performed in turn, mixed together and reproduced as a testament of his contribution to musical technology; a particularly novel and fitting legacy. Table 6.3: John Chowning and the Story of the FM synthesis patent I I Ilate Event- - - Details- - Knowledge formation/transfer 1957 Max Matthews at Bell Labs Acoustic Research Department makes the first Formation of tacit knowledge in carrier sector computer generated sounds while conducting experiments in telephony Realizes the potential for producing music. 1958-63 Mathews and John Pierce (Bell) develop Music I through IV, sound Accumulation of tacit knowledge generating computer programs. 1963 Matthews publishes "The Digital Computer as a Musical Instrument" in Knowledge codification in the public realm Science. 1963 John Chowning a graduate student of music at Stanford reads Matthews' Spatial transfer of codified knowledge from East to West Coast Science article and promptly takes a course in computer programming Acquisition of codified knowledge requires the accumulation of Chowning contacts Mathews and visits Bell Labs tacit knowledge to put into practice - codified to tacit knowledge cycling 1964-67 Chowning experiments with Music IV at Stanford's Comp Sci. department Tacit knowledge formation whilekeeping in touch with staff at Bell Labs 1968-71 While experimenting with Music IV, Chowning makes an "ear discovery" Tacit-Codified knowledge cycling Conceives the idea of producing tones to using FM synthesis and is Decision to seek patent changes course of how and whom this instructed by Pierce to, "Patent it!" knowledge is used 1971 Chowning takes idea to Stanford's Office of Technological Licensing American firms own tacit knowledge limits their ability to grasp "We then-approached a lot of American companies, none of which were significance of Chowning's discovery interested." Spatial transfer of knowledge to Japan Chowning contacts Yamaha, who send an engineer with a background in Right to bar others from using this technology. communications technology Yamaha initiates negotiations to exclusively license technology Over its 17 vear life cvcle this license becomes the Stanford's second highest generatorbf royalties 1973 Chowning publishes FM synthesis paper in the Journal of the Audio Knowledge is made publicly available but cannot be put to use, ~ngineeringSociety except by Yamaha. 1973-82 Yamaha develops FM technology and incorporates it in a number of Formation of tacit knowledge models with varying degrees of success Further codification of knowledge marks off yet more intellectual Builds wall of patents around Chowning's key patent further protecting their space. investment. 1983 Yamaha releases DX-7, the first mass-market all-digital synthesizer Radical innovation: the embodiment of fundamental codified incorporating Chowning's FM and Rockwell's patents knowledge and 10 years of development tacit knowledge ~amaha-sellsmore than 275,000 DX-7s 1983-85 Faced with the success of their rival's DX-7 Roland's Head of R&D, Competitor's response is to seek codified knowledge Kikumoto spends a full year researching-. old patents and articles trving- - to find prior art to challenge patent (see quotation on pg. 15) ources: J nstone 1999, Chadabe 1997, Vail 1993) 6.4 John Chowning, FM Synthesis and the Birth of the Stanford- Yamaha Relationship

The case of John Chowning and the FM synthesis patent is another story of transectorial and spatial knowledge transfer that travels the carrier wave from the font of ICT, through

California and on to Japan (Table 6.3). In 1963, Chowning was a graduate student of music composition at Stanford University. Inspired by a paper published by Bell Labs'

Max Mathews in Science, Chowning shifted his course of study to focus directly on computer music. The article, as an 'immutable mobile' was sufficient to enroll Chowning into a network of colleagues based around a few key 'centres of calculation' such as Bell

Labs. Course work undertaken in the computer science department followed by a visit to

Mathews' laboratory set Chowning on a 5-year path of largely independent study in which the bulk of his time was spent experimenting at Stanford's Artificial Intelligence

Laboratory with Mathews' Music IV program. One night in late 1967 while tinkering with a pair of oscillators, using one to control the pitch of the other, he makes an 'ear discovery'.

At a frequency of around 20Hz, he noticed that instead of an instantaneous change in pitch from one pure tone to another, a recognizable tone colour, one that was rich in harmonics, emerged from the machine. It was a discovery that an engineer would have been unlikely to make. What Chowning had stumbled upon, it later turned out, was frequency modulation - the same technique that radio and television broadcasters use to transmit noise-free signals. Of this, the composer was blissfully ignorant: All he wanted to do was make colourful sounds. Chowning began tweaking his algorithm and pretty soon, as he recalls, 'using only two oscillators, I was making bell tones and clarinet-like tones and bassoon-like tones, and I thought, you know, this is interesting.' (Johnstone 1994: 3) There were few researchers of Chowning's ilk at Stanford capable of confirming the

radical nature of his discovery, a factor which hurt Chowning's chances of getting tenure

in the department. Moreover, Chowning had yet to "connect his ear to the theory"

(Johnstone 1999: 2 17). As Johnstone (ibid) explains:

This conjunction would not occur until 1971, when Chowning remembered some synthetic trumpet tones that a Bell Labs researcher had played for him and wondered whether he could indeed achieve a similar effect using FM synthesis. It turned out that he could indeed produce some quite realistic brass tones. It was at this point that Chowning realized that his technique was a lot more than he had first thought.

In order to realize the potential of his invention, it was necessary for Chowning to enlist the support of 'colleagues'. When Chowning presented his discovery to the staff at Bell

Labs, its significance was immediately apparent, especially to Max Mathews' boss John

Pierce who instructed Chowning to "patent it!"

This turn of events deserves closer scrutiny, because invention at Bell Labs, a non-profit organization, was by convention not undertaken with pecuniary ends in mind - hence the public dissemination of knowledge evident in Mathews' Science article. Other

Bell Lab inventions found their way into the music industry along similar avenues to

Mathews' Science article. Mercer Stockwell, an engineer at quickly defunct New

England synthesizer maker MTI also consulted Bell when he,

realized that the technology had a ten year leap on other musical products of the time.. .so [we] wanted to get this technology out of Bell Labs. Now Bell Labs had no experience at all in selling anything. They didn't know how to do that and said it would take their legal staff a couple of years to research how to sell it to us. So the suggestion was made that [Bell] publish the thing and make it public domain, then we could use it, which was approved by the Laboratory, because they did that all the time. They had no built-in procedures for accepting money. Probably for a non-profit organization that's par excellence.(Vail 1993: 183). It this light it is very interesting that Bell's Pierce instructed Chowning to publish his

findings in 1970 with a view to remuneration, by patenting his idea. This necessitated the

recruitment of a different set of connections, the circuit of interests that Latour (1 999)

calls 'allies'.

6.4.1 US Firms: FM Falls on Deaf Ears

With echoes of the cold shoulder given to Rockwell and Deutsch by all of the US industry save Allen, in 1971 Chowning encountered a similar inability of US firms

(including Allen, Hammond and Lowery) to comprehend what was shown to them.

Despite Chowning's entreaties, ones that were fully backed by Stanford, a powerful ally, his interests failed to translate to the technical aspirations of any domestic firm.

Hammond Organ, for instance responded to Stanford's invitation by sending four engineers and a professional organist. While the latter was impressed with what he heard, the engineers failed to recognize that this technology heralded the dawn of a new frontier.

Here are Chowning's thoughts about that meeting,

They kept asking me about how many pins it would need. Well, I didn't know anything about pins, chips or analog circuitry at all so I couldn't answer them. I said, 'Look, it's an algorithm and here's the code [but] it was simply not part of their world (Johnstone 1999: 221; emphasis added).

Yamaha, on the other hand, sent an engineer, Ishimura Kazukiyo, with a background in communications technology. Ishimura, in Chowning's words, "took all of ten minutes", to grasp the technology and immediately initiated negotiations to take out an exclusive license on the patent. From the standpoint of Ishimura's boss, Mochida Yasunori, FM synthesis dovetailed perfectly with Yamaha's objectives. As an engineer, you are very lucky if you encounter a simple and elegant solution to a complex problem, FM was such a solution and it captured my imagination. The problems of implementing it were immense, but it was such a wonderful idea that I knew in my heart that it would eventually work (Johnstone 1994: 5).

In 1983, ten years after they started developing applications based on Chowning's invention, Yamaha released the DX7, the world's first, all digital programmable, polyphonic synthesizer. Pinch and Trocco (2002: 3 17) regard the DX7 as the

"breakthrough digital instrument, the first one to achieve commercial success", while

Colbeck (1995: 1) states that, "the DX7 opened manufacturers' eyes to the possibility that synthesizers could sell. From DX7 onwards, the synthesizer industry grew up." Whereas the signature synthesizer of the early 1970s, the Minimoog, achieved lifetime sales of

12,000, the DX7 sold over 200,000 units in three years. In the following two chapters I examine the response of Yamaha's competitors to the release of the DX7.

For Stanford, the institution which gave birth to Silicon Valley and which is synonymous with academic-industry collaboration, Chowning's invention has become the second highest generator of royalties for its Office of Technological Licensing

(Johnstone 1999). Stanford's relationship with Yarnaha has persisted to the mutual enhancement of both parties (Table 6.4). Most importantly, the royalties from

Chowning's FM patent funded the establishment of Stanford's Centre for Computer

Research in Music and Acoustics ('Karma'). CCRMA currently employs 14 research and teaching faculty including John Chowning and Max Mathews. Successive generations of key intermediate technologies include Julius Orion Smith's wave guide sampling73have

73 This technique which only became possible with the improvements in the memory capability of micro- chips allows actual samples of sound to be stored in memory for playback in real time. served to keep Stanford at the leading edge in forging university-industry liaisons.

Primarily these relations have been channelled through Yamaha.

Table 6.4: The deepening relationship between Stanford and Yamaha Date of Technology Details of Stanford - Yamaha relationship invention (of agreement) FM synthesis - Yamaha takes out an exclusive license on the FM invented by John technology. Chowning

Waveguide synthesis Waveguide synthesis is licensed in a non-exclusive sampling - invented fashion to Yamaha and California firms such as by Julius 0.Smith Ill Chromatic Research, Crystal Semiconductor and Seer Systems Sondius-XG Joint licensing agreement to cooperatively promote the development and use of their respective intellectual property portfolios (over 400 patents and applications) in the computer tone generation and sound synthesis areas. The Sondius program is Stanford's first effort to not only patent discoveries made on campus but to trademark them as well. ource: Johnstone I 99; htt~://www.sondiusx~.com/Dresrelease.htrnl;accessed 28/06/04. Date I Event - Details Knowledge formationltransfer

1968-72 a As an undergraduate Dave Smith plays bass and guitar in rock bands a Development of tacit knowledge through musical practice 1972 a Smith graduates from UC Berkeley with a degree in B.Sc. in computer science a Cultivates a basis of knowledge in the fundamental and electrical engineering sciences of the ICT paradigm Purchases a Minimoog and starts exploring possibilities of electronic sound Inventive tinkering with existing technologies modifies I manioulation approach to musical practice 1972-77 a Smith works for various aerospacelelectronics firms in Silicon Valley (Lockheed, Development of applied knowledge in carrier branches of G.E., Standard Microsystems, Signetics, and Diablo Systems) ICT paradigm a In his spare time builds a succession of analog and digital sequencers that a Application of scholastic and practical training to the interface with the Moog keyboard. In 1974 he trademarks "Sequential Circuits" problem of adaptinglimproving existing technology and starts sellina units he builds in his aDartment 1977 a Smith quits his regular job and directs his attention to full-time entrepreneurial Transfer of knowledge to the musical instrument industry management of Sequential Circuits a Transfer of codified knowledge through tacit agreement Agrees to license polyphonic keyboard technology from Emu systems1 of Santa with flexible, ill-defined terms. Cruz, CA. with understanding that royalty payments would be forthcoming 1978 a Sequential Circuits releases the Prophet-5, the first programmable polyphonic a Successful application of key technology to EM1 ensures synthesizer to incorporate a microprocessor. For a three-year window this product landmark status for Prophet-5 is the embodiment of best practice in instrument design 1980 1 a Sequential decides to stop paying royalties to Emu. A costly lawsuit ensues, Lack of a firm agreement between parties results in the eventually to be settled amicably contestation of shared technological space. 1981-83 Smith collaborates with Japanese and US manufacturers on developing the idea Japanese and American approaches to for a digital interface. Advocates idea at 1981 AES meeting and 1982 NAMM. instrumentlmarket development converge, with solidarity Facing reluctance from US firms to engage in further negotiations, Smith is sole marking one side and factionalism the other. American representative in the establishment of MIDI technology 1985 Sequential disastrously enters the home computer market at the time when sales Failed attempt to apply knowledge developed in branch in that segment are sagging sector back to carrier sector 1987 Sequential development team working on new platform technology called vector Formation of new incrementally significant knowledge svnthesis 1988 Yamaha purchases Sequential Circuits for $500,000. Product development team Knowledge is transferred from the US to Japan yet the under Smith is retained. (Yamaha purchases 51% of Korg) Yamaha interpretation of its codified embodiment is not favorably disposed to warrant further cultivation of Smith's tacit knowledae 1989 a Yamaha's contract for Smith's product development team is transferred to Korg Transfer of reservoir of knowledge from one Japanese firm to another lote: 1. Oberheim Electronics licensed this same technology for their four and eight voice synthe izers in 1975 (Vail 1993: 197) (Sources: Interview with Dave Smith, St. Helena, CA (1 1/10/02); Vail 1993) 6.5 Dave Smith

Dave Smith is both an inventor and an entrepreneur and it is in the performance of this latter role that he distinguishes himself from both Deutsch and Chowning. However, like these two he is a Californian whose career path and technological accomplishments intersected with the interests of Japanese musical instrument makers (Table 6.5). Smith's early career up until his founding of Sequential Circuits as well as his role in establishing the MIDI protocol has been covered in chapter three so these events will be referred to only in passing. The focus in this section is on two themes: Smith's approach to patenting and his work for the Japanese firms Yarnaha and Korg in the late 1980s.

6.5.1 Approaches to Patenting

In 1977, Dave Smith licensed a polyphonic keyboard patent from a firm based near Santa

Cruz, CA called E-Mu. Since the fledging Sequential Circuits had consumed most of its financial resources getting the Prophet-5 keyboard to market an agreement was made to postpone royalty payments. By 1979 the success of the Prophet-5, the world's first instrument to incorporate a micro-processor, had vaulted Sequential to the head of the industry and the royalty checks which Sequential could now afford to write represented a large transfer of funds. This cash flow in turn helped finance the R&D efforts of E-Mu.

However, in May of 1980, the staff at E-Mu received a letter from Smith informing them that he had decided to stop paying royalties on the Prophet-5 design. A lawsuit between the two ensued. According to E-Mu's founder Dave Rossurn, "E-Mu went from having a positive cash flow to having a fairly substantial negative cash-flow to pay the lawyers to get Sequential to pay us the money that we thought they owed us" (quoted in Vail

1993:199). The two parties eventually settled out of court in an amicable fashion.

This episode indicates that patents are honoured more in the breach than in the observance. Moreover, it suggests that over the course of the Sequential - E-Mu relationship, the latter's patent worked in a number of ways. At the outset, the patent sanctioned the transfer of codified knowledge in a manner that was only partially consistent with the conventions of licensing. Such a soft agreement between the two firms was likely the outcome of a shared entrepreneurial culture; one that was perhaps even embedded in a larger set of loosely defined regional codes of practice that characterized the formative days of Silicon Valley. A little later, after the patent's worth had proven itself, Sequential, for a time, recognized this value by conceding royalties to

E-Mu. The termination of royalty payments is more difficult to interpret except to say that Sequential viewed the worth of the E-Mu patent in limited terms. Finally, the lawsuit and resolution prove the immense power of this piece of codified knowledge to discipline the behaviour of the disputing parties in such a financially punitive manner that only extra-legal means and a touch of local culture could salvage the relationship from going the route of Jarndyce v. Jarndyce (Vail 1993). Nonetheless, this experience shaped

Smith's views on the value of patents thereafter.

Dave Smith holds no US patents for any of the technologies he has developed and he commented at length about his philosophy in this matter.

We never patented anything at Sequential. It was partially the cost, partially laziness and partially my own philosophy which was that in most cases if you are trying to patent something then it is only as good as your ability to defend it. Which I think maybe explains the difference between Roland vs. Yamaha. You know a lot of people, the only reason they get patents is for defence not for offence. If you have a lot of patents and somebody comes after you, chances are one of your patents will overlap enough with one of their patents that you can negotiate a deal so nobody gets hurt. Whereas if you don't have anything to offer and you have nothing in your stable of patents, then you're stuck. The nice thing about Yarnaha is that they do patent everything, but they rarely will go out and enforce it. Because they have a lot of fluff patents. They patent everything and most of them are just not worth the paper they are written on. And if they wanted to cause trouble they could go out and make a big deal about it, but in general they don't. With the FM thing they did get sticky a few times, because that was a big one for them. I think it is mainly just defence. If you have a lot of patents then you are pretty well covered and nobody is going to come after you (Dave Smith Personal communication, 1 111 0102).

Clearly Smith was attuned to the patenting practices of other enterprises. It is also likely that, given his early experience with contesting patent license agreements, he viewed them at best as a double edged sword. Also conceded, is that a firm without patents (such as Sequential), has no recourse for negotiation if its ideas are used by others.

6.5.2 A Brief Note on MIDI

I have already commented on Smith's role in the development of MIDI, but at this stage, it is important to reiterate one point that recalls the experiences of Deutsch and

Chowning. This is that American firms failed to recognize the value of yet another innovation that would prove fundamental in defining the ensuing industry-wide technological trajectory. The innovation in question here is both technological and institutional. Technologically the idea of an interface, even a digital one had been conceived by a number of firms, but only to link their own proprietary instruments. So it is really the institutional innovation, the adoption of a common serial interface that is particularly novel. The technological elitism that characterized the typical American firm's view of the market jaundiced their outlook to the potential of MIDI. Once again they failed to realize that compatibility and connectivity would be as fundamental to the diffusion of the ICT paradigm as any technology in and of itself.

6.5.3 The Demise of Sequential Circuits and the Collaboration with YamahaIKorg

Following the advent of MIDI, Sequential made the same mistake as Yamaha in the mid-

1980s by trying to break into the computer market with music applications.

We were way too early. We were self-financed and we never had any investment, because it was still too early, people weren't doing it and we were in a niche market. So at any rate, the business wasn't doing well because we tried to get into the computer market, spent a lot of money on it and it didn't work. By the time we tried to get back into our basic pro market, we were too late and we were faltering and getting ready to go under and trying to find buyers. As it turns out, Yamaha came in and knocked on the door one day. We hadn't approached them. They just came by and said, 'Well what's going on?' And we said, 'We're about to go under. If you want to buy us that would be cool. We can work out something, it will be cheap. But you have got to do it real quick'. And this is something that Japanese firms are usually not good at - but to their benefit, they pulled it off. In a very short period they worked it out with the banks and creditors and purchased the assets. Nobody at Sequential made anything. It prevented us from going into bankruptcy. And we became Yamaha DSD Inc (Dave Smith, Interview 11/10/02).

Even though Yamaha Japan organized the acquisition, DSD fell under the responsibility of Yamaha America, and so to Smith, they "fell in between the two worlds" with neither side knowing quite what to do with the organization formerly known as Sequential.

The first step was they stopped production and we ended up just being an R&D group for Yamaha. Then we were working on an instrument for them for a few months and then they pulled the plug on that. They just came over one day and said, 'We're closing everything down. We're giving up.' Even though we had an instrument that was fairly far along in design they just could not decide what to do. Then we are in the process of shutting everything down, everybody is starting to scatter in the R&D group and then Korg comes in a few weeks later and says, "We want to start an R&D group." And I'm going, "Once again if you want to do that, you better do it quickly because everybody is splitting. So Korg set up a new group (ibid). The relatively rapid transfer of the Sequential assets from Yamaha to Korg occurred at a

time when Yamaha was taking a controlling interest in the latter. These factors made the

relationship between the various parties, in terms of knowledge assets and product

development particularly interesting. When they were bought by Yamaha, the Sequential

R&D staff had a platform technology - vector synthesis - in its pipeline and the story of the knowledge embodied in these latent designs took a number of tricky turns.

One of the things I did when I [first] came over to [Yamaha] was I brought a two-page list of products that could be built - brief descriptions, a paragraph on each one. Some of them were follow-ups of products that we had already done and others were new ideas. I was describing one of our ideas for a new product and so I was drawing some stuff on the board and it ended up that it was a product that we were working on for them. And we ended up actually finishing the design at Korg and Korg ended up selling it - it was called the Wavestation. And so basically, what I drew on their [Yamaha's] chalkboard was the original concept, describing this stuff, saying how it's going to work, it's going to have these things etc. And the net result of the meeting was their engineers saying, 'Well we don't think that's going to work.' And I said, 'Well, sure it's going to work'.

We ended up doing the product, not for Yamaha, but for Korg. We got part of the way through it and Yamaha closed down the R&D group and while we were developing it for Korg - it was right around when we were going to start shipping it, or just after - I was looking through a trade magazine - the AES Journal always prints a summary of patents in the back, a paragraph on each one - and I was browsing through it just to see what was going on. I was reading one from Yamaha, because, of course, they always have a bunch of them in there. And I thought, 'wait a minute', that sounds familiar. So I went back and re-read it and it was basically a patent of what we were doing for that synthesizer and it was based on those drawings that I did when I was describing it to them. So what they had done was they first told us the concept wouldn't work. Meanwhile they probably had an engineer writing all the stuff down and they went out and patented it without ever telling us. The funny thing is, we worked for them. So they would have owned the patent. It was their technology. We worked for them and came up with the idea, so it was theirs. But for some reason they did it on their own. When the Wavestation finally came out at a trade~how'~,I was there and some people were saying, 'How does it feel to have two new products at this show?' I said, 'Two? We only have one, the Korg Wavestation.' They said, 'Well haven't you seen this Yamaha SY-22?' I said, 'No, I don't know anything about it.' So I went and looked at it and it was an instrument that was based very specifically on one of the things on that sheet of paper that I took to them. At the time, I said, 'Here's an idea for an instrument.' So they went away and they did the instrument without ever acknowledging it or asking about it or talking about it, even when we were still employees of Yarnaha. I just found it very strange that they would be so secretive about it. It was like, here is an idea, 'We're not going to tell you we're going to use it, we're not going to ask your opinion about it. We're just going to go ahead and do it.' It was very strange, but I have also heard from other Japanese companies that Yamaha is a strange company. (ibid)

The episodes described in this lengthy quotation point to some of the facets that

technology transfer manifests when the boundaries between firms are permeable. First,

resistance to new ideas emerges at quite specific moments. Prior to this particular

engagement, Yamaha seemed at a loss regarding how it should best use the physical and

human assets of Sequential. So a face-to-face meeting was organized. Even when Smith

explained his plans, and more importantly, scribbled this information on a Yamaha

blackboard, his ideas still might have seemed foreign and unworkable - that is, if his

appraisal of their response is correct. Second, codified knowledge is only a recipe for

industrial learning. Why Yamaha did not recognize its import at the time is unclear.

Suffice to say, either the inscription was not immediately persuasive or yamaha was not swayed by Smith. However, when Smith says, "it's going to work",his own tacit knowledge fortifies his convictions vis-a-vis the idea on the board. At this moment, however, in the terms of Latour's (1 987) science studies, Yamaha had not yet been successfully enrolled into the network that revolved around the vector synthesis platform.

74 The 1990 NAMM trade show in Anaheim, CA Codified knowledge does not, however, go anywhere when its author leaves the room (or firm), especially if someone has copied these inscriptions down - which suggests another facet: that translations can take time - or they can be as rapid as

Ishimura's ten minute recognition of the importance of Chowning's FM idea. At some point after the meeting with Smith, these ideas took on a different light and Yamaha did what seemed natural - patent the technology and set to work applying its lessons.

Meanwhile, over at Korg, the mobilization of this very same product design idea took root in an organization with an infrastructure favourably disposed to allowing Smith and his team to steward this translation through to fulfilment in the Wavestation. That these ideas first echoed to Smith while he browsed the AES Journal also points out that such inscriptions, once translated into patent form, perform in a variety of contexts. In the case of the AES Journal they communicated to a particular community of practice both what

Yamaha was doing and, in theory at least, what Yamaha could proscribe others from doing. In the end, Yamaha's controlling interest in Korg mitigated against the chance of what otherwise would appear to be an obvious case of infringement when the SY-22 and

Wavestation emerged as ready-made science at the same NAMM show.

Smith worked for Korg for a few years under varying contractual arrangements that varied from a consultant to employee of their multimedia division. Eventually he quit and decided to put on his entrepreneurial hat, collaborating with several other local engineers in the establishment of SEER Systems, a small firm working in the MIDI enabled cottage industry to provide software synthesizer packages that could be used as easily with both computers and keyboards (Dave Smith, Interview, 1 111 0102). When Smith left Korg, about half of the original, "seven or so", engineers that were part of the Sequential team remained. An appraisal of both Smith's and their career paths suggests something about the articulation of the spatial innovation system that they constructed between Japan and California. Drawing on the transectorial knowledge that percolated out of Silicon Valley, these individuals were steeped in both the science and culture of the region. Wisely the Japanese firms that absorbed this human capital sought to cultivate their tacit knowledge as it formed in situ.

Smith's story should be further highlighted because it sets the mould for a subsequent pattern of US engineers who became vectors of knowledge transfer to Japan, once their firms failed (Table 6.6). The eponymous firms of Roger Linn and Tom

Oberheim failed, as Sequential did. Thereafter Linn did contract work for Akai, while

Oberheim turned his talents to problem solving for Roland. Alberto Kniepkamp, once

Lowery's head of engineering who subsequently worked for Norlin before it too dissolved, gained employment with Roland as, amongst other things, a patent consultant. Table 6.6 Career trajectories of transectorial yatoi to Japanese EM1 manufacturers Name Employment prior to Japan Employment for Japanese firms Ralph Deutscha North American Rockwell 1965- Contract invention and consulting 1973 for Yamaha 1973-1975, Kawai 1975-1986 John chowningb Stanford University Department of Licensed FM synthesis patent to Music 1966mesent Yamaha, consulting- 1975 Dave SmithC Founder and President of Headed Yamaha DSD 1988-89, Seauential Circuits 1974-1988 Kora R&D US 1989-1994 John ow en^ Moog Music 1973-76, Sequential Yamaha DSD 1988-89, Korg R&D Circuits 1982-87 US 1989-98 Chris ~e~er~Sequential Circuits 1984-87, Roland R&D 1990-1997 (tasks Digidesign 1988-1989, Marion included analysis of competitors Systems 1990 intellectual ~ro~erty). . -. Roger Linne Founder and President of Linn Contract invention for Akai 1987-94 Electronics 1978-86 Tom Oberheimf Founder and President Oberheim Contract invention for Roland 1988- Electronics 1970-1985, Marion 1990 Systems 1987-present Al berto Head of R&D for Lowery Organs Patent and product design Kniepkampg 1965-69, Norlin 1969-78 consulting for Roland 1989-present Herb ~eutsch~ Hofstra University and Moog Music Contract consulting for Roland 1964-1977. Norlin Carlo Lucarellig Technical manager Farfisa (Itl), President of Roland Europe 1988- Founder president of SlEL (Itl.) Dennis Houlihang Lowery Organs, author of 'Owner's President of Roland US Manual', VP Marketing- Anthony ~illias~ Professional Musician, Music Retail Product Specialist at Korg, Director Shop Floor Demonstrator of Marketing for Technology Products Yamaha John ~emkuhl~ Music retail Synthesizer voicing Korg 1988- Nick ~owes~ Degree in Astrophysics, Yamaha R&D (UK) 1992- Professional Musician, Music Retail (UK) Mark Moffat' Studio Engineer EM1 (Aus) 1972- Contract consulting for Roland 1980, Record Producer, Nashville 1978- 1982- I I I Sources: a - personal communication, 12110102; b - Johnstone 1999; c - personal communication 11/10/02; d - Sonik Matter website (http://sonikmatter.com) accessed 01/04/04; e - Roger Linn Design website (http://www.rlinndesian.com) accessed 01/04/04; f - Sound on Sound website

455209cfcl c8bOc61 edOdd6b190eOa) accessed 01/04/04; g - (Kakehashi 2002: 149-15 1); h - Herb Deutsch website (http://www.hofstra.edu~academics/hclas/music/musicdeutsch.cfm)accessed 0 1/04/04; i - Roland User's Group on-line newsletter (www.rolandsus.com/community/ru~/Fall0 1 /currents.asp) accessed: 021 12/02 6.6 Conclusion

A comparison of the Deutsch, Chowning and Smith career trajectories is instructive in a number of ways. First, in their own way, each of these biographies illustrates that knowledge formation at the cusp of the musical and electronic worlds was the result of grounded practice. For example, Chowning's 'Eureka! ' moment, his 'ear discovery' of

FM synthesis, was clearly not the product of rational calculation. As Johnstone (1 999:

2 16) infers, it was a discovery that "an engineer would have been unlikely to make."

Nevertheless, Chowning's musical training, his later interest in the learning the practicalities of computer science and his connection to Bell Lab contributed to his stock of tacit knowledge. From a practical standpoint, these diverse knowledge bases grounded his experimentations with sound. In this sense, his 'discovery' reflected a 'diffusion of the engineering disciplines', although the nature of this diffusion was contingent on the social and geographical contingencies of Chowning's career. Similar 'practical masteries of technology' (Storper and Walker 1989) infuse the defining moments of Smith's and

Deutsch's careers.

Second, the juxtaposition of their biographies further points to how diverse sets of institutional contexts and individual practices shape the approaches to and outcomes of knowledge codification. For Ralph Deutsch, a lone inventor more comfortable working under his own terms or with his son, patenting every technology he developed became a practice that would ensure his recognition as an individual. It was his way of making sure that he, or more accurately his inscriptions, as 'instruments' (Latour 1999), became indispensable to an evolving network of interests that, with time, shifted towards the core of instrument manufacturers in Hamamatsu, Japan. In the case of FM synthesis, Chowning could not persuade American organ manufacturers of its value. Yet before FM synthesis was communicated to the industry, Chowning presented his findings to the team at Bell Labs who first inspired his entry into the fold of electronic music. Upon advice from Bell Labs' John Pierce, who himself worked in a context which eschewed both patenting and licensing, Chowning patented this idea. This patent cemented a relationship between Stanford and Yamaha, the legacy of which is CCRMA, an institution which has fostered the development of a number of fundamental technologies that have directly benefited both the university and Yamaha. In other words, the moment of 'technical crystallization' remained latent until it was translated to the interests of

Yamaha. Also notable in this sequence is that soon after Yamaha communicated their interest in the FM technology, Chowning's tenure was reinstated. Finally, for Smith, the entrepreneur, the inscription of knowledge in patent form was a practice more suitable for those with the resources and energy to defend that knowledge. Smith's approach to codified knowledge, whether in respect to E-Mu's patent, or to the designs he jotted down on Yamaha's blackboard, appears far more laissez faire than Deutsch. This might be a personal or generational matter; it also might indicate a difference in the industrial cultures of the Los Angeles aerospace industry versus that of the milieu which blossomed at the intersection of the Bay Area music scene and the Silicon Valley community.

As an aside, it is worthwhile to briefly contemplate the nature of the San

Francisco Bay Area as a creative space, and how this milieu contributed to the career paths of inventors located there. Dave Smith, for instance received his formal engineering training at UC Berkeley and at various aerospace firms. However, it was his passion as an amateurlsemi-professional musician that inspired his inventions. After our interview, Smith could not wait to show me his home studio where he demonstrated his latest invention, an instrument called '~volver'~~. Like Deutsch, he too is a 'tinkerer' at heart and his hobby and his occupation have dovetailed seamlessly. Chowning's early career trajectory was also embedded in academic as well as artistic networks - Johnstone (1 999) writes about Chowning's informal collaborations with Phil Lesh, the bass player for the

Grateful ~ead~~.As these examples show, Smith and Chowning had strong connections with the local music community that, in Latour's (1999) terms, fulfilled the role of both

'allies' and the 'public'. However, it was outside of these overtly 'cultural' connections where their network in the US started to unravel.

Each of the three inventors participated to varying degrees in the 'community of practice' (Wenger 1998) organized around the Audio Engineering Society. Fr instance,

Chowning gave papers in 1968 and 1972, Deutsch presented in 1971, while Smith took his turn, in 1981. As organizers and chairs of conference sessions, Chowning and Smith appear to have been more consciously interested in enrolling 'colleagues' into the network. However, it appears as though these collegial ties were poorly linked with the circuit of domestic 'allies'. Perhaps these three inventors were 'ahead of their time'. This is where the (ir)reconciliation between the individual and their context comes into play, opening up the door for the types of connections that rely more on networks in topological space than on geographical proximity.

Indeed, the way that these individuals related to their Japanese partners is particularly instructive in what it says about the geographies of embedding, dis- embedding and re-embedding of knowledge from one setting to another. At the outset,

75 For an explanation of this instrument, please see: http://www.davesmithinstruments.com~# 76 This author rejoiced in discovering this particular connection.

257 these engineers cultivated a body of knowledge positioned at the cusp between two

'communities of practice' (Wenger 1998): electronics and music. In the case of Deutsch and Smith this setting was the high-tech sector in California at a time when military technologies were been applied to meet consumer ends. For Chowning, the network he forged with Bell Labs and the training he undertook in Stanford's computer science department would appear to have provided a sufficient basis for further translations.

However, in attempting to translate these quite radical sets of knowledge into the music industry, each of these individuals faced resistance from prevailing habits of thought in the circuit of allies: the digital organ idea was rejected by most US firms; Chowning's

FM idea was not part of Hammond's 'world'. Even Smith, who at the time was the bright young star of the US synthesizer industry, could not persuade his US colleagues of the importance of MIDI. It is also surprising, at least in the Chowning and Smith cases, that

Silicon Valley, a region that serves as the global standard for industrial learning failed to adopt these radical ideas. Smith, for instance was unable to secure venture capital from outside sources to finance Sequential.

In each case, Hamamatsu firms, and especially Yamaha were prescient enough to offer these individuals the best deal for bring their ideas into fulfilment. Even at this stage, the transfer of knowledge was far from straightforward and in these spatial transfers, translations had to be performed in both the literal and figurative sense.

Deutsch took it upon himself to learn Japanese. The only way for Smith to communicate was via a blackboard. In their relations with Yamaha, both Deutsch and Smith found that their ideas were not automatically celebrated and their roles privileged. Deutsch's NES lost out to the PAS in the Yamaha trials of 1974, while Smith left his meeting at Yamaha with the belief that they were not going to use his idea for vector synthesis. Yet in the final analysis, each of these contributions played a role in transferring radical knowledge to Japan and Yamaha did 'use' their inputs in key ways. Deutsch served as a vector in the implementation of digital sound, Chowning transferred the method by which digital synthesis could be realized, while Smith' collaboration on the MIDI protocol enabled the entire horizontal integration of the industry that particularly swung in the Japanese favour in the latter half of the 1980s. In comparison to the scores of yutoi who counselled the uptake of Western institutional practice in the Meiji period, the trajectories of these three individuals continue this pattern of knowledge transfer in the general sense. However, they also extend and refine the mode of transfer to a scale of practice that could effectively harness the leading edge at a time when this technological space was highly uncertain. By commanding the technological high ground at such a pivotal time, Japanese firms were able to leapfrog the very regions that spawned these ideas. As Smith comments,

You know I don't subscribe to the idea that the Japanese just copy and don't innovate. I think that was true in the past, but it's not true now. If there's one thing they're good at, it is learning. A lot of these companies will send people overseas for a few years and then bring them back, just so they get more of an international experience. (Vail 1993: 25)

The following chapter examines Japanese perspectives on how they came to absorb and improve on the radical technologies transferred fiom the US. Indeed, it would have been impossible for Yamaha and other Japanese firms to 'use' the knowledge of Deutsch,

Chowning and Smith unless they were already technologically advanced in tehri own right. CHAPTER SEVEN: JAPANESE ENGINEERING PERSPECTIVES ON THE TRANSFER OF KNOWLEDGE AND THE COLONIZATION OF TECHNOLOGICAL SPACE

7.1 Introduction

This chapter addresses the manner by which engineers at Japanese firms transferred and acquired a practical understanding of electronic musical instrument (EMI) technologies that originated in the US. It presents the process of technology transfer as a multifaceted exercise that involved foreign travel as well as the mobilization of texts and products. As in chapter six, I am especially eager to point out the ways that engineers dealt with codified knowledge both as interpreters and authors of texts, as well as the firm strategies that underlay these practices. The narrative focuses on two periods of time. The first of these is the initial phase of 'transregional inspiration' (see Table 2.2) in the 1960s and early 1970s when Japanese firms first started accumulating foreign knowledge. The second phase covers a few key episodes in the early 1980s when Yamaha's competitors responded to that firm's 'commercial crystallization' (Table 2.2) of digital EMI. In both of these cases, engineering practice relied on the mobilization of codified knowledge, whether this was to translate or thwart the interests of the competition. The chapter makes a further distinction between the contextual circumstances of engineers working in smaller firms versus those at larger enterprises. Evidence for these accounts comes from primary data gathered through open-ended interviews conducted in Japanese with engineers at Yamaha, Roland and Kawai, as well as secondary accounts in biographies

(Kakehashi 2002) and firm histories (Yamaha 1987).

Once again, the theorization of these processes draws on science studies and evolutionary economics. As the methods by which Japanese engineers accumulated and introduced foreign knowledge became more sophisticated and systematic through the

1960s, they relied increasingly on the translation of documents such as textbooks, technical reports and blueprints to guide their efforts. Latour (1987, 1999) calls these types of codified knowledge that maintain their form as they travel 'immutable mobiles'.

With these texts in hand, the engineers had a script with which they could mobilize other forms of knowledge, for instance the lessons gathered through: the 'cycles of accumulation' (Latour 1987) - shisatsu ryoko (inspection tours) of the US (see chapter four); 'tinkering' and the reverse engineering of foreign instruments (chapter four), and; foreign experts (chapter six). These processes, when enacted amongst a set of competitors who were co-located in Hamamatsu transformed the geography of the spatial innovation system. Collectively, the efforts of these engineers propelled Hamamatsu from its peripheral status relative to the US to that of an industrial core that functioned as a

'centre of calculation' (Latour 1987). Essentially, engineers based in Hamamatsu had tipped the scale of advantage "by forcing the world to come to the centres - at least on paper" (Latour 1987: 233). As these engineers systematically mobilized these foreign texts to their own interests, they enhanced their ability to remap the contours of technological space by re-inscribing this knowledge in the battery of significant and insignificant patents that were discussed in chapter five. The perspectives recorded here are unique, for in the literature on geographies of innovation in Japan, the voice of engineers is largely absent. Indeed in economic geography more generally, the primary subject of 'corporate' interviews has been the manager (for an exception, see Schoenberger 19971~~.However, science studies (Latour

1987, 1999) gives a much more significant role to the engineer, for their everyday practices dissolve the distinction between scientific and social worlds. Moreover, for evolutionary economists such as Rosenberg (2000), the process of technological change needs to be understood as, 'the diffusion of the engineering disciplines'. This chapter continues to investigate the spatial dimensions of this diffusion in the nascent field of electronic music.

As both a discipline and occupational category, engineering is the sine qua non of applied techno-science. Engineers are trained to be problem solvers not only in an abstract sense, but in a concrete practical sense as well. Veblen, for instance viewed engineers as the heroes of capitalism and put great stock in their 'transformative capacities' for wrestling control away from the pecuniary interests of business owners

(Tillman 1993: xix). The engineer's contribution to technological change, and by extension economic development, is reflected in Galbraith's (1967: 8) observation that,

''the enemy of the market is not ideology but the engineer." An extension of Galbraith's institutional approach, however, subsumes the agency of engineers within a techno- structure whose goal is planning. In contrast, this chapter seeks to situate individual practices within the context of various organizational structures. In other words it

- 77 Undoubtedly this type of research has recorded the insights of engineers, although I suspect that these accounts have more often than not been framed in such a way as to accentuate the subject's role in the fm as 'managers'. constructs a biography of engineers that negotiates the "knife edge between social context

(structure) and personal creativity (agency)" (Barnes 200 1: 4 15)

In this regard, the contextual differences between R&D workers in the US and

Japan could not be more stark. For instance, relative to their American counterparts,

Japanese engineers working for medium and large corporation are far more likely to have remained with the same company over the course of their careers (Dore 1973, Abegglen and Stalk 1985, Shapira 1995). This career stability needs to be viewed alongside the fact that CEOs in Japan are additionally prone to come from an engineering background, while in the US, it is the ranks of finance that tend to beget CEOs (Tachibanaki 1998).

Together, these marked differences enable both individual engineers and the firms that employ them to think in a long-term, practical fashion.

A brief discussion about my primary sources is appropriate here, for the goal of the interviews was to hear first hand accounts of the digital revolution in EM1 from engineers whose careers spanned back to the 1960s. Thus it was not surprising that at the time of the interviews in 2001 and 2002, many of my subjects occupied senior management positions at their respective firms. Of the fifteen engineers who I spoke with at four firms (Yamaha, Roland, Kawai and Casio), eleven had joined their respective enterprise right out of university, two had initially worked for a brief time for larger electronics firms prior to their recruitment by the musical instrument industry and only one had transferred from another musical instrument firm. These stable career trajectories of the Japanese cohort stand in contrast to the varied paths traversed by most US engineers in this industry (chapter six). Finally, in discussing the competitive dynamics amongst the various

Hamamatsu-based firms, I draw an important distinction between the relative contexts of engineers working within large firms versus those individuals operating outside of the corporate context or within smaller enterprises. Conceptually, the difference between

Schumpeterian Mark I (entrepreneurial driven) and Mark I1 (large firm endogenous science) perspectives on innovation (Phillips 1971, Freeman et al. 1982) serves to explain the different engineering contexts found in Roland versus Yamaha. A science studies perspective helps to frame these different circumstances, by highlighting the different networks of interests through which small and large Japanese firms mobilized American knowledge and interacted with each other. However, at the very beginning of this evolutionary process, back in the late 1950s and early 1960s, formal institutional context mattered little, as both sets of engineers approached the set of knowledge that would constitute the basis of the digital age from the standpoint of amateurs.

7.2 Origins: From Tinkering Amateurs to an Innovation System

In the late 1950s and 1960s, the field of electronic music was small in the US, and even more marginal in Japan. At the time, Japanese firms relied on three strategies of knowledge accumulation. The first involved the tacit hands-on practice of tinkering, a second, that worked from codified sources such as textbooks and the third, which hinged on overseas travel or the shisatsu ryoko trips to see American technology in action. These three modes of learning frequently overlapped in the case of larger firms. However, at this time there were also many industrious amateurs working outside of these institutional circumstances, like Roland's founder Kakehashi who tinkered away with whatever old technology he could get his hands on. I will present these two vantage points in turn because they indicate a confluence of the two types of fundamental innovation noted by

Schumpeter: Mark I Innovation, which was the domain of the entrepreneur and; Mark 11, which relied on the resources of large firms. In these formative years, however, both sides were starting virtually from scratch. The discipline, if it could be recognized as such, was quite small and only the most established foreign ideas were known in Japan.

One American whose accomplishments stood out in determining the subsequent direction of the technology in Japan was RCA's Harold Olson. Nagahama Yasuo an engineer at

Yamaha, put it this way:

HF Olson is the father of the electronic organ as a musical instrument. He published a very original book in 1952 78 .. .and developed the RCA synthesizer in, I think 1956.. ..Because [Olson] was regarded as the father of the field, it was the intention of President Kawakarni to direct his engineers to the source of this new style of music.. .The first lessons that were learned came from the engineers who were sent overseas, and especially the Olson book dominated our discussions around the time the D-1 came out. Even until maybe 1985, the Olson book greatly contributed to our thinking about technology, not in a narrow sense, but in a general way (Nagahama Yasuo [Yamaha], interview 21/7/02, translation by the author)

Yamaha's entry into the electronic age necessitated the cultivation of a whole new set of skills that were radically different from the engineering common sense found in

Yamaha's piano division. This established way of thinking was based to a large degree on a craftsmanship ethos.

For a long time the piano has been the field of craftsmen - it is the product of experience education and mastery. Yes, engineers draft blueprints when the piano is made for mass production, but if you want quality at the scale of mass production, the craftsman is in charge. For instance in developing a new piano prototype the designer will always go on the recommendation of the craftsman. Even until the 1970s, this craftsman way of thinking influenced our approach to the organ (ibid).

78 Olson's Music and Physics was not translated into Japanese until 1962 (Source: author's fieldwork). Tinkering, formalized in Yamaha as reverse engineering, allowed Japanese engineers, amateur and professional, to be the original craftsmen of the electronic age.

This was the time when amateurs were reading magazines to figure out how to make black and white televisions. Before that it was amateurs, not Matsushita [a major electronics manufacturer based in Osaka] who were making good radio sets from spare parts. So it was also the time when amateurs were making organs for themselves. Since the level of technology in Japan was increasing rapidly but was still quite low after the war, the way we made instruments was also like an amateur ... Around these times, when engineers entered the company they would train for two years taking apart American organs in order to study the parts (ibid).

While the conventional wisdom holds that such strategies amounted to little more than mimicry, I would counter this argument by pointing out the value this practice serves in maintaining a culture of learning that keeps the sights trained on the state of the art.

Moreover, from the standpoint of science studies, the practice of collecting and deconstructing foreign 'instruments' allowed these engineers to 'mobilize the world' around them (Latour 1999), and thereby 'know at a distance' (Law and Hetherington

7.2.1 Enrolling Colleagues

It was, however, unreasonable to transfer human resources from the piano division to the upstart organ unit within Yarnaha. An entirely new labour force needed to be hired and trained, and so began a new phase of development within Yamaha - a Japanese version of

Rosenberg's (2000) 'diffusion of the engineering disciplines'.

Since we were only really making pianos and spinet organs back then, there were no electrical engineers. However, it was President Kawakami's intention to make electronic instruments so money was earmarked to attract engineers - for instance by university student scholarships (ibid). In this novel mode of recruitment, which operated as a Yamaha ROTC~~,university students would receive scholarships to learn the appropriate disciplines in exchange for a commitment to engage in apprenticeship sessions once a year at Yamaha, with guaranteed employment to follow. This effort was costly and it took a long time to reap the benefits. However, in the long run it paid off, as Yamaha had accumulated an engineering team that benefited from academic training and was socialized within the firm to acquire as set of enterprise-specific skills (Kioke and Inoki 1990). Essentially

Yamaha had institutionalized an 'autonomization' of 'colleagues' within the firm (Latour

1999; see Figure 2.1)

As this labour force took shape, the organization of R&D became more formalized with a view to linking the tacit and codified domains of knowledge.

This way of doing things changed into a system, especially for electronic instruments where we started to work from diagrams.. ..President Kawakami wanted Yamaha organs to be the best in the world and this conviction was at the front of his mind. So he gave instructions to do work based on documents [textbook schematics, assembly modules for US organs etc.]. He called up Mr. Kondo, head of the laboratory and told him to, 'work from the documents', because much of what was learned in the US [during the shisatsu ryoko] also came from documents. In fact there was a system for learning from documents - getting the specifications by inspection - here is the circuit diagram for the unit - here is the assembly figure for the unit.. .Many general things can be learned, if the structure of a transistor is known (ibid; emphasis added).

This distinction between the ways that piano design and EM1 design are viewed within

Yamaha has persisted.

Even today with CAD, piano design does not go smoothly unless there is a craftsman involved (ibid).

79 Reserve Officer Training Corp: A US Army institution whereby the Army bears the cost of university education, in exchange for a commitment to reserve military service. Of course, Yamaha did not arrive at success over night. In fact, for close to 20 years

Yamaha engineers worked diligently with few triumphs, and consequently they flew under the radar of their US competitors who took a patronizing view of Yamaha's efforts.

Upon the introduction of the first Electone D-1 in 1959, the engineers were certainly presented with a series of challenges.

The D-1's sound was unreliable and we couldn't stabilize the pitch frequency. It was like other Japanese products of the time: low quality. It was not built in a factory but in the laboratory. Work at the factory started the following year when a new section was established. For the D-1's development there were seven people in total. Five people worked in the electrical shop, and the other two worked on relating the other mechanisms such as cabinets, keyboards, switches and design.

To give you an episode which really speaks about those times, once the D- l moved to the factory, development took place off in this comer (katasumi) from which the most deplorable sounds came. It was like these impure sounds really grated on the ears of those in the piano section. Yet we had been given a mandate by Kawakami that allowed us to work this way. When 5:OOpm came, the whole factory would go pitch black except for the katasumi. This situation is very symbolic. We produced many failures because we set our sights high. Until 1970, we were really laying the groundwork for the whole Electone line (ibid).

In this quotation, when Nagahama describes the working environment as being off in a

corner, the term katasumi (Kffi)is entirely appropriate and symbolic, because in

Japanese the connotations of katasumi can mean both a corner nook or a state of

existence that is marginal from the status quo. Toiling away in a marginal space removed

from Yamaha's bread-and-butter piano division, the organ engineers persisted, confident

in their work which had been given a long term boost of support from none other than the

mercurial president of the firm. Gradually the ranks of engineers were augmented and a more complex division of labour also took shape, particularly from the late 1960's after Mochida Yasunori,

Yamaha's head of electronics R&D convinced President Kawakami that Yamaha needed to develop its own chips. This was a particularly pivotal stage in Yamaha's evolution, as a number of projects were engaged in rapid succession.

7.2.2 Linking (Domestic) Colleagues and Allies

Johnstone (1 999) argues that it is important that Yamaha had access to a generation of top engineers who were schooled at domestic universities for this enabled them to appreciate the value of home-grown R&D. Mochida dispatched a team to learn under

Professor Nishizawa in the Department of Electrical Engineering, Tohoku University all six of these engineers were trained in Japan. The collaboration between Yamaha and

Nishizawa to design and produce semiconductors was funded by the Research and

Development Corporation of Japan (JRDC) - an offshoot of the Science and Technology

Agency, and supposedly a rival body to MITI (Johnstone 1999).

The fact that Yamaha had had no prior experience with semiconductors was one reason that the relationship between Nishizawa and Mochida's people had gone so well. Engineers who had been educated - "contaminated"- in the United States would argue with Nishizawa based on what they had learned at MIT or Stanford or wherever. But to Yamaha's young engineers, who knew nothing else, Nishizawa sensei was the be-all and end-all of semiconductor technology. (Johnstone 1999: 3 10)

The collaboration between Yamaha and Professor Nishizawa to produce semiconductors fulfilled two 'circuits' within the network (Figure 2.2). Nishizawa played the role of 'colleague' par excellence. Moreover, by forging connections with a domestic

'centre of calculation' the act of 'autonomization' was doubly inflected - it broadened and deepened the network of engineers , and it did so in a way that advanced the autonomy of Japanese institutions.

In the early 1970s key foreign technologies were also sought. In the previous chapter, I noted how Ralph Deutsch's arrival in Japan at Yamaha also led to the establishment of two teams, one at Yamaha's HQ in Hamamatsu, the other 15km away at the Toyooka focal factory, who were working on competing parallel technologies. On the heels of Deutsch's arrival, the Chowning patent was also licensed. Finally, after accessing the two fundamental sources of radical transectorial knowledge embodied in the Rockwell (digital organ) and Stanford (FM synthesis) patents and licensing these technologies, it sought to shield these Ur texts behind a wall of derivative patents which, in part, explains its dominance in the patent data (chapter five). In the early 1970s patenting activity at Yamaha increased dramatically and has remained high ever since.

To the present, Yamaha's EM1 R&D division maintains a systematic approach to transectorial innovation. Individual engineers translate knowledge from diverse sources.

As another Yamaha engineer, Aoki Eiichiro pointed out:

Individuals inventors introduce knowledge from fields other than electronic musical instruments to Yamaha. However, what these individuals learn is passed on to others at group study meetings [guru-pa benkyokai] (Aoki Eiichiro Interview, 15/01/03; translation by the author).

As this quotation illustrates, the ongoing task of translating transectorial knowledge to the firm involves the continual cycling of tacit and codified knowledge within the circuit of

'colleagues'.

Manager Technology Development Group Product Development Dept. Pro Audio & Digital Musical Instrument Division. 7.3 Roland

Thus far I have presented only a partial view of early EM1 developments in Japan, that of

Yamaha. In the following section I contrast this large firm perspective with that of

Roland, an entrepreneurial upstart.

From the late 1950s and through much of the 1960s those Japanese engineers seeking to access American knowledge but lacking the corporate resources to travel relied primarily on the import of literature and products. In other words, engineers and inventors in these circumstances, lacking the ability to 'act at a distance' could only engage in surveillance -knowing at a distance' (Law and Hetherington 2000). Moreover, contrary to the present time, where knowledge in codified forms is assumed to be readily

'cosmopolitanized', to use Storper's (1997) term, access to these materials proved very difficult. Roland's founder, Kakehashi Ikutaro, offers this appraisal of his efforts in the

days before Kakehashi Musen morphed into Ace Electronic,

When I learned of a relevant book having been published, I would immediately place an order and then would have to wait as long as five months for delivery. It was only in 1962 that I was able to obtain Electrical Musical Instruments, which was published in the US in 1958. To say there was a technology and information 'gap' between Japan and the US would be a major understatement.. .. In those early days the only [other] way we could pursue our development efforts was by examining imported instruments and adapting their technology (Kakehashi 2002 159- 160).

Kakehashi took every opportunity available to examine foreign instruments which to

purchase were quite expensive in Japanese currency in the 1950s and 1960s.

Occasionally, organs, typically cheap American and Italian brands such as Organo, were

brought in for repair to his shop Kakehashi Musen in Osaka. Following the path of

Yamaha Torakusu, it was during these repairs that he hatched the next stage of his career as an entrepreneur. Beyond these repair jobs, and similar to the electronic enthusiasts who flocked to specific places such as Tokyo's electronics district Akihabara (Takehashi

2000), Kakehashi gravitated to places where pipe or electronic organs could be found.

When he learned that there were two rare brand American Minshall organs in Tokyo, he mad the trip from Osaka to Tokyo, even going as far as to visit the home of the owner of one these instruments (Kakehashi 2002: 160).

When I opened the back of the instrument I could see that the design of the Minshall represented an intriguing advance in technology. All electronic instruments of that era utilized the 12AX7 (twin triode) vacuum tube. However in order to produce the same number of keyboard notes, the Minshall needed barely half as many tubes as the Organo instrument, which had pioneered the field. It was a clever design that could generate two notes an octave apart and with very few components and utilizing only a single triode tube. The Organo was still a more stable instrument, but the engineering creativity inherent in the Minshall was an exciting step forward.

There is a Japanese verb kengaku (Ry)which combines the kanji for looking or seeing

(E)and that for learning (y).Translated as 'study by observation', its common usage is

as verb-form of 'fieldtrip'. What else is Kakehashi doing with the Minshall but learning

by looking? There is some light tinkering going on, but it would not endear him to his

host if, in this case, he actually took the organ apart. Similar instances, repeated over a

career of opening up the technological black boxes of other firms to learn from them has

taught Kakehashi much about what works and what does not work. For example, during

Ace Electronics' collaboration with Harnmond Organ (see chapter four), Kakehashi

figured out why Hammond organs performed so well in Japan relative to species like the

Organo's that often required some repair. The answer lay in the materials used for the key

contacts. Most organ makers cut costs by using gold and silver plating on their electrical contact materials. Since the initial signal of the Hammond tone wheel was so subtle it had to have more reliable contacts made of the more expensive platinum or palladium metals.

The Japanese summer can be unbearably humid (mushi atsui) and those organs using cheaper contacts (gold and silver) routinely failed in these conditions due to increased oxidation. It was fortunate that Kakehashi was cooperating with the firm which, by a fluke of local climate, proved to be the most reliable supplier to Japan. Of course, the corollary is that virtually all organs constructed in Japan were forced to use platinum and palladium, a feature which in turn established their reliability in the US.

Despite these early efforts to access foreign technology, structural constraints in the market prevented Kakehashi from accessing the most advanced technologies. This

was nowhere more evident than in the patent system.

The patent system then in use in Japan was unreasonable in that importers of new products could apply for and be granted patents, even if the technology was public knowledge in other countries. The law held that claims could not be refuted in the absence of supporting domestic documentation. Even if particular technologies were known in other fields inside Japan, patents would be granted if the claims were limited to the field of musical instruments. Circuitry was patented "as-is", and we could do nothing to prevent the practice because such claims were legal at the time.. ..The situation put independent engineers at a great disadvantage compared to parties who had sufficient capital to import foreign instruments. The only way to limit the damage of the situation was for amateurs and small companies in Japan to immediately publish all the information they had. By establishing the existence of domestic documentation, patent claims could be refuted (Kakehashi 2002: 162).

Clearly Kakehashi had no ally in patent office. As a result, in 1966, Kakehashi, two other

co-authors and many other independent developers published a volume, which if

translated, is titled Everything About Electronic Instruments. Six of the seventeen authors

went on to work at Roland. This would not be the last time that Kakehashi leveraged the

I collective learning of his engineering colleagues outside the firm. Indeed, echoes of this approach to innovation were subsequently resounded in the (co-)lead role that Roland played in working out the MIDI protocol (see chapter three).

Later on, Roland's attitude to patenting was reflected in the way that Kakehashi hired his chief engineer, Kikumoto Tadao, who had gained much of his experience working for one of Japan's large electronic firms. My interview with Kikumoto Tadao in part dwelt upon the rather unique way in which he joined Roland.

My back-ground was in audio, wireless communications and microprocessor applications. I developed Japan's first microcomputer mainframe-based Intel 8008 in 1972. I wanted to apply the microprocessor to music synthesizers which were becoming popular. I invented Synthesizer Guitar to control the microprocessor-based polyphonic synthesizer and applied for a US patent. I quit the [other] company and created a prototype as a venture business. When I finished, Roland launched Guitar Synthesizer which was just the inverse of mine. (Kikumoto Tadao, Interview 15/12/02)

While Roland's version was monophonic (capable of producing only one tone at a time) and based on analog technology, Kikurnoto's prototype was polyphonic and controlled digitally through a computer. It was thus very expensive to develop and consequently

Kikumoto, "spent all the money I had" (ibid). Once developed though, it had the prospect of a clear cost advantage over the Roland model.

I wanted to sell my patent to Roland which was the most aggressive venture company in the industry at the moment. Mr. Kakehashi bought me instead of my patent (ibid).

Unlike Yamaha's approach, which at this stage was geared towards the acquisition of rival technologies via licensing, Roland sought to absorb not only the codified knowledge embodied in Kikumoto's patent, but also the tacit (and transectorial) knowledge responsible for its conception. In doing so they brought a key 'colleague' into the fold. Upon joining Roland, Kikumoto had to acclimatize himself to the music industry, a world which he knew little about.

During the training period, I learned what the musical instruments, the industry and market were all about. I cannot play any instruments at all. Digital Signal Processing was totally different from the micro processor application skill I had. I had a hard but exciting study (ibid).

In conceiving and bringing to market such legendary products as the TR-808 and

TB-303, Kikumoto learned much about the workings of that market. He could hardly fail to notice the strategies of Roland's neighbor Yamaha, particularly when confronting the massive tracts of technological space they had effectively colonized through their patenting practices.

I noticed three thousand patents [domestic and foreign] already had been applied by Yamaha. Besides many easy ideas and trivial techniques, a lot of digital signal processing techniques had been applied (ibid).

From these documents he surmised something about the way in which Yamaha

'mobilized the world' of 'instruments' in the development of various technologies.

Most [Yamaha] patents might be derived from US research like Stanford University and I respect their effort to realize the FM technique. In fact, Chowning's patent is fundamental but very primitive in order to make products for music. Yamaha engineers invented many useful techniques to integrate the system. At the same time, [they] already had started semiconductor production to apply for their music instruments. So, dedicated digital semiconductor [technology] was the key and heart for the digital musical instruments business (ibid).

7.3.1 Responding to the Commercial Crystallization of Competitors

It is instructive to view how Kikumoto, and by extension Roland, reacted to the commercial success of the DX-7: by seeking sources of codified knowledge in order to challenge the patent. Kikumoto, recalls the shock and response which the DX-7 elicited noticing that,

They launched DX-7 in 1983. It was a shock for all the competitors. After finishing the MIDI protocol design and agreement I was studying Yamaha patents, reading all the articles. It was a full year of patient work. I learned where we were and where we would go. I noticed that the easiest and most useful techniques were applied in order to protect their business from new comers. In fact, Panasonic, Hitachi, Toshiba, Sony were trying to join the market for EMI. The [Chowning] patent first seemed like such a big barrier to break without infringing. However, I noticed that there were much better techniques, if we contrived to avoid the infringement. The harder to avoid, the better the technique. I took advantage of the barrier. I did not have enough time to apply the patents of the technical idea. If the techniques were copied by others, it would be very hard to prove and protest. Digital technology in the LSI is a kind of black box. I spent our effort on developing not on protecting. I spent my time on research of many US institute papers and articles. I went to the Computer Music conference every year. I visited many research institutes not only in the US but also Europe in order to find prior art to protest Yamaha's patent. But finally we had to develop our own techniques. Patents are a double- edged sword. (ibid: emphasis added)

This series of quotations is all about the study of inscriptions. It suggests little of the rigors of tinkering, and stands in contrast to the favorite practice of Kikumoto's boss

Kakehashi. This is not to say that Kikumoto was unfamiliar with this facet of learning.

On the contrary, he acknowledged that, "We always study our competitor's products"

(ibid). My point is that by the 1980s the shape of the spatial innovation system - the range of technological options open to pursue - was greatly circumscribed by the galaxy of patents which attended the accelerating accumulation of knowledge during the 1970s.

Navigating this space involved careful study. Moreover, the practice of kenguku had become far more spatially expansive in its search mechanisms. Expediency necessitated that it also be a much more active process. Compare Kakehashi waiting two years for the arrival of books fiom the US, with Kikumoto scouring the world for textual evidence of prior art to contest the FM patent.

This shift in the scale of surveillance, 'knowing at a distance' (Law and

Hetherington 2000) was apparent to outsiders. For example, Nashville music producer

Mark Moffat, who has worked with Roland in matters relating to product design has commented on the R&D process at Roland. Referring in particular to Mr. Kakehashi and

Mr. Kikumoto, he stated, "They were both so far ahead. They were always reading

[engineering] papers from all around the world." (Roland User's Group 2001: 1).

Relative to other EM1 manufacturers, Roland spent the greatest percentage of its income on R&D (Colbeck 1996:99). In 1987, Roland's R&D staff of 240 engineers in

Japan accounted for almost half of the firm's labour force of 500 employees (Doershuk and Milano 1987, Roland Corporation 1999). Roland's Kikumoto explained the organization of R&D at Roland as follows,

The Hamamatsu R&D center focuses on fundamental research creating core technologies. All the design divisions including overseas ones develop the products that apply the core technology. Twenty core engineers work for fundamental research. Engineers fiom the design division join the project sometimes with rotation." (Kikumoto Tadao, Interview 15/12/02)

Fundamental research cycles at Roland unfold within a "two-to-five year" time fi-arne

(ibid). Since the early 1990s, Roland has organized its R&D as a 'producer' system, with independent teams developing their own products and preparing them for production.

This structure forces engineers to view innovation and production systems as one, in a manner consistent with the Japanese style of industrial learning (Freeman 1997).

Moreover, the parallel development aspect echoes the pattern seen at Yarnaha a decade earlier, as that firm staged trials between its own passive analog (PASS) system and

Deutsch's new Electone (NES) system. Colbeck (1 996:99) has suggested that,

while this system of 'competing' teams keeps spirits high and products focused and also keeps an absolute avalanche of products on the starting blocks, it also throws up the odd anomaly wherein sequential Roland products seem to overlap or have blurred technologies because they were designed and produced by different teams.

Whether consumers perceive an overlap of products or not, from an engineering standpoint, the most difficult stages in bringing ideas based on a 'core technology' to market are consistently, "programming and debugging" (Kikumoto, T. Personal communication 15/12/02). Clearly, the practice of engineering in the digital age had fully come into alignment with norms of other ICT.

7.4 The View from Kawai

Kawai has never been at the forefront of developments in EMI. Nevertheless, it has also not been very far behind in following the trajectories established by its neighboring rivals. Ralph Deutsch's migration from Yamaha to Kawai in the mid 1970s is illustrative of this shadowing behaviour. As an engineer at Kawai recalls,

Ralph Deutsch came to our labs in 1975 and taught us sine synthesis techniques and about the digital organ. I thought it was amazing. I don't know of any other contact [our firm had] with Americans. [Saito Tsutomu, Manager, R&D Electronic Musical Instrument Division, interview (2216102); translation by the author].

From the previous chapter, we know that Deutsch was commissioned to start inventive work for Kawai while he was still under contract to Yamaha, and that this move greatly angered the latter. Moreover, it was Kawai who initiated this relationship. I was not furnished with an account of how Deutsch came into Kawai's orbit although it is reasonable to speculate that his reputation in the patenting arena raised his profile. Going one step further, it is also likely that Kawai picked up on some of the 'local buzz', precipitated by the arrival of Deutsch at Yamaha's labs in Hamamatsu.

Even before the arrival of Deutsch, engineers at Kawai had their sights trained on developments in the US. When speaking of the atmosphere of technological learning in the late 1960s and early 1970s respondents at Kawai raised a similar theme to their neighbouring counterparts.

In those days it was anyone's guess what would come next, although most Japanese felt certain that it would come from the US. As a result, Japanese were all looking to America. Magazines and other reference materials were consulted. [Saito Tsutomu, Manager, R&D Electronic Musical Instrument Division, Kawai Corporation, Interview (2217102); translation by the author].

As long as the critical developments were perceived to be exogenous to the region,

Kawai was at no particular disadvantage. However, once the leading edge was seen to pass from overseas to domestic rivals, engineers at the firm were made acutely aware of their relative position. Surprisingly, it was not Kawai's immediate neighbours who brought this intensification of competition into focus.

In the case of Kawai, we were watching Yarnaha and Roland. But it was not until Casio entered electronic music in 1980 that that we became conscious of all the developments in the field. [Takaba Tsutomu, Chief Engineer R&D Electronic Musical Instrument Division, Kawai Corporation, Interview (2217102); translation by the author].

So how did the Kawai Corporation learn from its domestic rivals?

From 1975 until 1990 we could make conjectures about developments based on the [domestic] patent record. We could keep pace with these developments until 1990, but after that year things were different. Yamaha started to go large scale into [applications] for cell phone and karaoke. Kawai was still a pure electronic musical instrument [company]. [Saito Tsutomu, Manager, R&D Electronic Musical Instrument Division, Interview (2217102)l.

Apart from these efforts at parsing through the patents and other written records of rivaling firms, there is scant evidence that direct face-to-face encounters with other

Japanese engineers played any role in stimulating learning at Kawai. This feature stands in marked contrast with the way that Kawai engineers mined their contacts with

Americans. The following series of quotations from a group interview with these respondents drives this point home.

Kondo : Occasionally I went to the US for trade shows and had quite open discussions with American engineers [on matters of a technical nature]. Around here it is different. Japanese don't have these things. Here, Yarnaha engineers don't have many things to say"

Takaba: There is no talk about our jobs.

Saito: At occasions such as nomikai [After-work drinking festivities of an informallformal kind], we do nothing but talk about the past.

7.5 Discussion

The reflections and recollections of Japanese EM1 engineers shed light on the nature of learning practices within the Yarnaha, Roland and Kawai corporations. Most generally, these practices unfold in a relational space that in chapter two I have called a

Transectorial Spatial Innovation System. Conceptually, the viability of these various relations hinged on the ability of firms to translate the interests of 'instruments' (texts and artefacts), 'colleagues' and 'allies'. At a corporate level, the Japanese end of this system is interpreted by referring to Schurnpeter's two modes of innovation. Specifically,

Yarnaha, in this sense, appears to conform to the Mark I1 variant. This is evident in the way that it augmented its internal division of labour to systematize the translation of externally derived knowledge. It is also apparent in the manner in which it exploited internal resources to organize trials of competing internally developed technologies.

Finally, if these practices are viewed in tandem with Yamaha's strategies in the inscription of codified knowledge in the patent arena, then we obtain a sense of how it further leveraged internal economies to shape the landscape of technological space outside of the firm.

Running against the grain of these practices is their origin in the practical yet unsophisticated culture of tinkering that prevailed both outside and within large firms.

When faced with technological discontinuities, Yamaha did not initially behave like a

Mark I1 innovator. Rather, its foray into the field was tentative; space for the new division was allocated 'off in a corner' (katasumi), not enjoying pride of place. However, in the 1970s Yamaha accelerated its transectorial evolution by creating a set of places, the

HQ laboratory and the Toyooka focal factory, to work systematically through the development of instruments. Exogenously it committed to the licensing of key technologies, the hiring of foreign experts and most importantly, the development of an internal economy of scope in the production of LSI. All these activities necessitated an extension and deepening in the scale of its innovation system to access talent situated in distant constellations of practice (from US inventors and universities in California to university research labs in Tohoku in northern Japan). Endogenously, the adoption of a

'systematic' approach to reverse engineering prevailed in the 1960s. By the 1970s this system shifted gears. Rigorous patenting, particularly in relation to key technologies, reflected a transformation in emphasis towards capturing the rents of Yamaha's substantial practical familiarity of unexploited gaps in technological space. In this manner, Yamaha's laboratories became what Latour (1987) calls 'centres of calculation'.

In summary, the accounts of engineers at Yamaha point to a reorganization of corporate strategy that gave direction and eventually momentum to a trajectory. Strongly demarcated proprietary boundaries locked in advantage in a way that enabled Yamaha to be first to harness the moment of technical crystallization. In these respects Yamaha bears the signature of a Mark I1 innovator due to the scale and strength of linkages in the constellation of interests in which it occupied the centre.

Independent of, yet in response to the practices of its larger rival, the engineering principals at Roland, Kakehashi and Kikumoto, speak from a vastly different point of view, one that appears representative of the Mark I mode of innovation. This is especially the case back in the early days of the industry in the 1960s and early 1970s. Kakehashi, in particular, was instrumental in rallying similarly positioned and like-minded amateur

'colleagues' to stake their claim in the emerging knowledge economy that was initially structured in the interests of larger firms. The publication of Everything about Electronic

Instruments is testament to both the way in which the entrepreneurial innovator exploits external economies amongst its network of tinkerers, as well as to the importance, again, of inscription, as a situated practice. This instance contextualizes the role Roland later played in instituting the MIDI protocol, which drew various firms into cooperation. The difference with this later collectively-oriented endeavor is that Roland was able to enlist

Yamaha, as an ally, into the project. In the meantime both firms had evolved. By the time Kikumoto was reacting to the 'shock' caused by the Yamaha DX-7, Roland was at the stage in its evolution that is best classified as a medium-sized threshold firm (Hayter 1997, ch.10). As such it had the resources to expand the scale of infrastructure devoted to searching the technological landscape. At this stage is it still fair to characterize Roland as a Mark I innovator? To answer this question it is necessary to turn to the next chapter to investigate how Roland and other enterprises behaved during the take-off and maturity of the innovation system and specifically, how they maintained linkages in the existing network and fostered linkages with the 'public' circuit. CHAPTER EIGHT: BEYOND THE GREAT DIGITAL DIVIDE: THE TAKE OFF, CONSOLIDATION AND MATURITY OF THE EM1 INDUSTRY - INCREASING RETURNS FOR HAMAMATSU

8.1 Introduction

To this point, the thesis has examined the evolution of a transectorial spatial innovation system up until the phase of 'commercial crystallization' (Table 2.2) in the early 1980s.

This chapter discusses the geography of what followed. Specifically, it investigates the mechanisms by which firms in Hamamatsu both concentrated their control over the global industry, via corporate consolidation and renovations to interface protocols such as

MIDI, and extended their scope of innovation and manufacturing activities within increasingly global divisions of labour. As such, it provides analysis of the changing morphology of EM1 production within Hamamatsu and around the world, and in doing so characterizes the nature of Hamamatsu as an industrial core. This evolutionary sequence, which traces the industry's development through the phases of 'take-off, 'consolidation' and 'maturity' (Table 2.2), is interpreted, in part, through the lens of science studies. In particular, Latour's (1999) Circulatory System of Scientific Facts (Figure 2.1) heuristic is drawn upon to explain the centralization of 'linkages' that connect Hamamatsu with a spatial innovation system. Since the discussion begins from the point of the industry's take-off, I focus especially on the formation of connections with the 'public' circuit, which has been covered only briefly to this point. Prior to the stage of 'commercial crystallization', which was ushered in by

Yamaha's release of the DX7 synthesizer and the establishment of the MIDI protocol, transectorial impacts had yet to be fully 'translated' to the public. Moreover, the rest of the industry, beyond Yamaha, had yet to 'get on the bandwagon'. This is because the rents from radical innovation were localized within Yamaha, the particular firm that locked-in the advantage of technical crystallization. It is unlikely, however, that its advantage would remain static for long. Indeed, the previous chapter discussed the tactics

Yamaha's competitors used to respond to the release of the DX7. A classic case of

Schumpeterian 'swarming' ensued. Like the cluster of four-minute miles that followed

Bannister's benchmark run at Vancouver's Empire Stadium on July 1, 1954 (Loasby

1999: quoted in Maskell 2001), once a commercial breakthrough has been achieved, competitors swarm to make the breakthrough themselves. The sector then takes-off. More generally, radical changes spawned by technical and commercial crystallization drive a phase of explosive growth. An expanding market, though initially kind to all competitors, ultimately rewards economies of scale in innovation and production. At some point this take-off stabilizes, the technological trajectory matures and the sector confronts the limits to expansion. Theories of path dependence (David 2000) suggest that initial advantages are cumulative in their impact. Consequently, the firm who authored the technological crystallization and those that respond most effectively are best positioned to sustain themselves through the industry's maturity.

Maturity is signalled in two ways. The first signal is industrial concentration

(Table 2.2), a readjustment of a sector's structure that reflects the motivations of firms to enhance their share of a finite market. Concentration occurs through the failure of marginal firms as well as through mergers and acquisitions initiated by stronger corporations. The second signal is a restructuring in the spatial division of labour in production. By shifting routinized production processes overseas, firms often obtain significant cost savings.

Innovation systems undergo a similar restructuring as they mature. This restructuring is necessitated by the institutional realities of concentration as well as by technological developments. With sectoral concentration, the surviving population of firms can tap a pool of talent, comprised of engineers who were formerly engaged by their now defunct competitors. These individuals are sought for their tacit knowledge, as well as the ideas in their heads that can be codified (See the case of Dave Smith; chapter six). Technologically, ongoing incremental innovations in the carrier sector are translated to the receiving sector. In the ICT paradigm, improvements in memory and processing capacity in the US drove incremental transectorial innovations in the Japanese musical instrument industry. In other words, firms in Hamamatsu translated these general advancements to their firm-specific circumstances.

The chapter is organized in the following manner. The discussion begins by illustrating the magnitude of change that followed the moment of commercial breakthrough. A longitudinal profile compares the characteristics of particular instrument production runs of the 1980s relative to instruments of previous decades. This analysis brings to light the pattern of responses to Yamaha's bench-mark DX7 that I touched on fiom an engineering standpoint in the previous chapter. In this context I discuss the advent of sampling technologies as the significant incremental innovation that promoted the second generation of digital EM1 in the late 1980s. After foregrounding processes of technical and institutional adjustment, represented here by sampling and consolidation, I then discuss the geographical restructuring of the production and innovation systems of the dominant firms during phase of maturity. Both of these systems evolved locally and globally and I tackle issues such as the reorganization of divisions of labour within

Hamamatsu, as well as the globalization of production. I continue by reflecting on the contemporary articulation of the innovation system, as a global hub and spoke that continues to be centred on Hamamatsu. This necessitates a discussion of Hamamatsu as a core. Takeuchi's (1 989) local model of industrial linkages in Hmamatsu contextualizes developments in the EM1 sector. In this regard manifestations of institutional thickness hinge on inter-industry complementarities within and among firms. These transectorial linkages resound in other parts of the milieu, and Hamamatsu's almost singular success as a Technopolis is presented as an example. Embedded within a particular type of learning region (Morgan 1997), the EM1 sector in Hamamatsu has evolved through a set of spatial relations that reflect the competitive strategies of its two principal firms,

Yamaha and Roland. To illustrate the nature of this local/global competition, I comment on the architecture of interconnection that is embodied in subsequent generations of the

MIDI protocol. Intended as an open-access non-proprietary platform, the key decisions regarding MIDI have increasingly evolved as a dialogue between Yamaha and Roland.

8.2 Take-off

The advent of programmable digital synthesizers, of which the Yamaha DX7 was the first, revolutionized the EM1 market. Pinch and Trocco's (2002:3 17) description of the

DX7 as "the breakthrough digital instrument, the first one to achieve commercial success", corresponds well with definitions of radical innovation (e.g. Freeman 1987, Freeman and Perez 1988). In comparison with the analog synthesizers of the 1970s (e.g. the Minimoog) which, at best, sold in excess of 10,000 units, digital instruments reaped enormous returns with production runs of 100,000 to 200,000 units. Data compiled from

Colbeck (1996) puts this digital frontier in perspective (Table 8.1). Not since the legendary Harnmond B-3, have electronic keyboards sold to such a mass market.

Moreover, while the 275,000 Harnmond B-3s sold over a twenty-year period, Roland took only two years to sell 200,000 D50s (Table 8.1).

This latter instrument's story, from a science studies perspective, begins where it ends, with 200,000 'black boxes'. This engineering and commercial feat "is made invisible by its own success" (Latour 1999: 304). The D50 was the outcome of Roland's four year strategic evolution to engineer a product that could compete with the Yarnaha's

DX7. Roland, like everybody else, was 'shocked' by their competitor's innovation when the DX7 was first released in 1983. Technologically, Roland's tone synthesis platform remained analog well into the mid-1980s (as in models Jupiter 8 through Juno 66 - Table

8.1). Ironically, Roland's lag in becoming digital until the mid 1980s has not hampered the aesthetic reputation awarded this generation of instruments by contemporary m~sicians*~.Indeed, it is not the sonic qualities of analog versus digital which need emphasis in situating this radical shift in instrument design, but the fact that the latter instruments are compatible with the ICT paradigm. Roland, one of the principal architects of MIDI was certainly attune to the information intensification of modem music production systems. On the other hand, it needed the prompting of Yamaha to embark on its own strategy to respond to the market potential of digital instruments. To do this, the firm had either to challenge the DX7 directly, or work around it. Roland tried both strategies. Initially the firm's chief engineer, Kikumoto Tadao, spent a full year searching the technical literature and scouring the world to find prior art to protest Chowning's FM patent which lay at the core of the DX7 (see chapter seven). This did not work, so Roland conceptualized and developed its own means of digital tone synthesis that did not rely on the FM technology. Their solution: Linear Arithmetic Synthesis (LAS), which hinges on

8 1 I still seehear Roland's analog generation keyboards gracing many a stage. Another example is the TR- 808 drum machine, the backbeat of house music.

290 a design logic that is referred to as 'sample+synthesis' (Colbeck 1996:97)~~.In lay terms, the D50s tone production system was "based entirely in software that went on to grace a generation of Roland products (ibid: 98, emphasis added)". Not long after the D50,

Yamaha and Korg, which was 50 percent owned by Yamaha, introduced their own

'wave-table' sampling83software-based synthesizers.

8.2.1 Sampling and the Emergence of a Division of Labour to Exploit Software Synthesis

The shift to sampling/software-based synthesis and instruments that recalled stored tones from memory proved yet another step in the transectorial revolution of the industry, drawing musical and ICT spheres even closer together. At this stage it is worth recalling

Theberge's (1 997) characterization of electronic musicians as 'consumers of technology'.

With the advent of sampled tones and software, musicians who used digital EM1 (which were increasingly likely to be manufactured in Japan) consumed a palette of sounds which were prepared as software, often by a third party. Instruments that followed the

DX7 were marketed as being programmable. In practice, however, few consumers could be bothered to learn the applied math necessary to program their own tones. This fact impressed itself on firms ranging from Sequential Circuits to Roland, who discovered that fewer than 10 percent of their customers elected to alter the factory pre-sets

(Theberge 1997) 84. In comparison to the EM1 pioneers from the previous generation (e.g.

RCA's Olson) who required advanced scientific and engineering degrees to produce synthesized sound, the musician of the 1980s could approach their practice which much

82 Colbeck (1997:97) suggests LAS is, "essentially, subtractive analog-style processing of sampled waves stored in a large bank of ROM - a phase near impossible to paraphrase. 83 A technology which derived from a patent by Stanford's Julius Orion Smith 111 (Johnstone 1999). 84 Factory pre-sets are comprised of the bank of 20-200 programmed voices that come with the instrument. of the messy science already programmed into their instrument for them; a ready-made black box.

The dawn of sampling and software synthesis also give rise to new industrial spaces that could be exploited by entrepreneurs. Through the early 1980s entrepreneurs faced insurmountable barriers to entry in the hardware sector. Sampling technology facilitated the birth of a cottage industry comprised of professionals (and to a lesser degree hobbyists) who programmed original and emulative patched sounds for each successive hardware model. The high information-intensity of their product allowed these entrepreneurs to market these 'sound libraries' and ancillary support systems directly to consumers, for instance through advertisements in magazines such as Keyboard and

Electronic Musician. For instance an ad by Hollywood, CA firm Club Midi Sofmare

(Electronic Musician July 1988:71) touted the merits of Prolib, 'the ultimate librarian for

IBM PC' as giving 'extensive patch data and storage for multiple instruments.'

The hardware producers caught wind of this trend and sought to internalize an external economy. In Latour's (1999) terms these potential 'allies' needed to be enrolled in the network. For example, there is the case of John Lemkuhl, a retail assistant whose tacit knowledge of EM1 earned him a job as a product support specialist with Korg in

1988.

First and foremost, I learned everything I could about all of Korg's products and then went to music stores to share my knowledge and enthusiasm with the sales people and customers by doing music clinics and in-store concerts. I also created demo sequences and product support that we then left at the stores to help them more effectively sell our gear (Cowell 2001a: 1). At this point, Korg was keen to enlist 'allies' such as Lemkuhl in order to foster linkages with the public 'circuit'. Soon thereafter, Lemkuhl's responsibilities evolved beyond service functions.

When I joined Korg, I already had a fairly large number of cool MI patches I had created while in Seattle (I sold over 30 Mls in the first month the M1 was shipping alone). I sent these sounds to my boss in New York and he sent them off to Japan. Next thing I know, I'm flying to New York to work on the MI Drum & Percussion card as well as to work on the Voicing for the T-series. Next I was asked to work on the M3R and the 23 FM Guitar synth and was on my way to work in Tokyo Japan. From about that time on, I was more involved in the programming/development side of Korg products rather than the product specialist/support side. MIDI Patch-Boys (MPB for short) is the self-given nickname for the team that does the voicing work on most Korg instruments (ibid: 1).

Lernkuhl's story is not unique. In 1990 Yamaha enrolled Nick Howes, a British astrophysics major and part-time EM1 retail assistant, in a similar manner.

Two reps from Korg and Yamaha came in basically on the same day telling my then boss that I had to go to Frankfurt [site of the European MusikMesse trade show] as I would be the "only person" who would understand some new "top secret" toys being developed (these were the VLl and the Wavedrum). ...This is because I had studied wave mechanics, and also some of the work by Julius Smith (on waveguide). .. the Yamaha rep mentioned that they were looking for good support staff at the Yamaha UK main office, and that I should apply. (Cowell 2001b: 1)

Soon after joining Yamaha, Howes also moved from a service to a creative position.

They told the boss to talk to me, about possibly getting involved with some synth voice development. I had that year been working for Ultravox (the pop group), on an album and world tour as programmer and keyboard tech, so I had some credentials in this area.

Dan gave me a prototype keyboard to do 32 voices85for (the W7 as it turned out) of which 31 were accepted and made it into the voice ROM (apparently a very high percentage for a new guy!). I then took the job full

85 Voices are the various types of tone produced by a synthesizer. Examples include trumpet, cello, slap bass, and rotor organ. The last of these is Roland's approximation of the Hamrnond B-3 organ sound.

293 time at Yamaha Kemble [Milton Keynes] doing tech support and tech sales, but always hankered for the job at R&D London.. ..After about 18 months at Yamaha Kemble, I received a call from the head of the European Music Software section at the R&D center. He was asking if I wanted to work for him developing voice data, and also building the new Yamaha.co.uk website (something I was getting into at the time was web development), and also work on a new soundcard project. (Ibid)

These two stories illustrate a number of characteristics of the evolving global innovation system for EM1 that centred on Hamamatsu. First, Japanese firms, unlike the US conglomerate Norlin (see chapter three), were quick to recognize that creating a market for new technologies involved educating the customer and thereby enrolling the 'public'.

Indeed, since the 1960s Japanese firms had given a strong emphasis to educational and service functions in their organizational structure (e.g. Yamaha Music Schools, and

Keyboard Clinics). In the digital era, these functions were extended to serve major overseas markets hence the hiring of Lemkuhl and Howes as product support specialists.

Second, Japanese firms sought to stay abreast of the sampling phenomenon by employing musician's whose practical knowledge of the hardware-software interface could be translated to the role of 'voicing' specialists. In this manner, Japanese firms extended their spatial innovation system to enroll a second generation of foreign talent. Third, the similarity in Lemkuhl's and Howes' career trajectories, not to mention the same day visit of Korg and Yamaha to the latter, is suggestive of more than coincidental human resource practices at the two Japanese firms. Yamaha's investment in Korg enabled the former to enact parallel technological strategies via complex inter-firm and international relations.

Finally, Howes' later job descriptions point to the elaboration of transectorial innovation strategies that emphasized the complementarities between hardware, software and information systems. 8.3 Consolidation

The first signal pointing to the maturation of the EM1 sector was the phase of consolidation that began in the late 1980s. US firms that could not compete in an environment that demanded economies of scale in production and innovation folded or were purchased by Hamamatsu-based firms. Table 4.8 catalogues some of these acquisitions. After the purchases, Japanese firms were interested in taking-on few assets other than engineering talent (e.g. YarnahaKorg's purchase of Sequential Circuits,

Akai's hiring of Roger Linn) and brands (Roland's purchase of Rogers, Suzuki's purchase of Hammond). One exception that I mentioned at the end of chapter four concerns the organ sub-industry, which followed a different logic. Organs are electronics encased in wooden cabinetry. When Japanese firms purchased US organ makers (e.g.

Kawai's purchase of Lowery, Chicago and Roland's purchase of Rodger's Hillsborough,

OR), according to Ralph Deutsch (personal communication, 12110102), the Japanese firms kept the US factories open to keep these 'furniture' production systems in situ. The legacy of these trans-Pacific acquisitions was that Japanese firms absorbed US engineering talent. These relations took two forms. In some cases, companies like

Yamaha, put these individuals to work in R&D satellite labs in regions like California. In other cases, US engineers performed contract R&D and consulting tasks for Japanese firms (Table 6.6). I address these reorganizations of the innovation system shortly.

The second sign of maturity in the EM1 sector concerns the reorganization of the production system. Causal factors, of both a general and industry-specific nature, prompted Japanese EM1 makers to reorganize their production geographies. Generally, the locational strategies of firms in the EM1 sector correspond to broader patterns of overseas corporate investment by Japanese industry that took shape from the mid-1980~.

Endaka, the appreciation of the yen following the Plaza Accord in 1985, was a principal catalyst, along with trade friction, for these responses. Within this climate the geographic pattern of investments conformed to sectorally specific logics. The auto sector, for example invested in the major market regions such as North America and Europe.

Conversely, for the leading firms in the electric/electronics sectors, the western pacific (in particular east and southeast Asia) received almost half of the investment in new manufacturing operations following 1985 (Edgington 1993).

8.3.1 The Evolving Spatial Divisions of Labour: The Production System

The shifting pattern in imports of electronic keyboard musical instruments (EKMI) to the

US points to an evolving geography of production in the sector (see Figure 8.1). In 1989, imports from Japan comprised 87% of total imports to the US. However, from 1990, imports from countries other than Japan have taken on an increasing share.

Figure 8.1: Imports of Electronic Keyboard Musical Instruments to the US 1989- 2001

L Source: United Nations Statistics Division (http://www.unstats.org) Surpassing Japan in 1995, these countries, including Malaysia, Indonesia and China, accounted for 63% of imports in 2001 (see Figure 8.2). Nonetheless, in considering this trend in the data, it is important to recognize that production of EKMI is dominated by

Japanese firms, no matter where it is located. Therefore the increase in imports of instruments from sources other than Japan is a largely an extension of these firms' investment strategies.

In 1990 the only significant imports, other than Japanese, came from South Korea and Italy. Imports from South Korea are accounted for by Kurzweil (a division of the piano maker Young Chang) and Samick. Five years later, the second and third leading source nations for EKMI to the US were China and Mexico. By 2000, the league table of source nations has shifted once again; following Japan were China, Malaysia, Mexico,

South Korea and Indonesia (Figure 8.2).

Figure 8.2: Imports of Electronic Keyboard Musical Instruments to the US 1989- 2001 (excluding Japan)

El UK Germany rn Indonesia Malaysia Thailand EI Mexico China DI Korea rn Italy

Source: United Nations Statistics Division (http://www.unstats.org) Table 8.2: Yamaha's geography of factory/lab establishments since 1959 Year I FactoryILab in I FactoryILab in Japan FactorylLab outside of Hamamatsu IJapan 1959 Miyatake Factory(pian0 Yarnaha de Mexico, SA de CV parts), HQ ~ab~(EMI) 1963 Nishiyarna Factory (upright

pianos) lwata Factory (metal

Taiwan Yarnaha M.I. (piano, electone) 1970 ToyookaFactoryLr Saitarna Factory (brasslwind Kaohsiung Yamaha Co. (electones, LSI) instruments) Taiwan (guitar) 1971 Kernble & Co. UK (piano) 1974 Toyohashi Factory (wood Yarnaha Musical Products, MI, processing) USA (woodwind, drums) 1975 P.T. Yarnaha lndonesia (piano) 1976 Kagoshirna Factory (LSI) Yarnaha Music Manufacturina.". GA, USA (piano) Yarnaha DSD Inc.'. . CA.. USA (EM1 R&D) Tianjin Yarnaha Electronic, China (EMI) P.T. Yarnaha Music. lndonesia (guitar, drums) Yarnaha Electronics Alsace, France (AV parts) Yarnaha Electronics Manufacturing- Malaysia-. (AV CD-ROM) Guanazhou Yarnaha Pearl ~iverbianoInc. China (pianos) 1997 Xiashan Yarnaha M.I. China (piano parts, wind instruments) P.T. Yarnaha Musical Products lndonesia (wind instrument parts) P.T. Yarnaha Music Manufacturing Asia, lndonesia (EMI) 1998 P.T. Yarnaha Electronics Manufacturing lndonesia (AV s~eakers Jotes: L: dedicated laboratory. LF: facility that contains laboratory functions (e.g. focal factory) (Source: data poLided by Yamaha 26/8/01, translation by the author) Table 8.3: Roland's geography of factoryllab establishments - Year FactoryILab in FactorylLab in Rest of FactoryILab Overseas - Hamamatsu Region Japan Ace Electronics, Osaka - Hammond lntl Japan LF - (e. organs) - *Roland HQ, Osaka, Roland's Hamamatsu - ~actory~~(EMI) Takaoka ~actory~~(EMI) - (100 employees) Boss Mfg.s,Taiwan (EMI, - effects pedals) Hosoe actor^^^ (EMI) - (500 employees) Matsumoto Factory (EM11 Rodgers Organ Mfc~.~OR, IT peripherals) (80 USA (organs) employees Roland Europe San - Benedetto5, (EMI) Hamamatsu LaboratoryL (Basic Research) (30 employees) Roland Tech. (30 employees Roland1 80 - employees BOSS) Miyakoda Factory (EMI) - (250 employees) Roland Electronics Suzhou6, China (EMI) - I I Nlotes: L: dedicated laboratory LF: facility that contains laboratory functions (e.g. focal factory) 1) Hammond International, a 50150 joint venture between Ace and the US firm Hammond, initiated production in an old Zenon organ factory in Hamamatsu 2) Nominal Headquarters. 3) Boss Mfg. was established as a 50150 joint venture between Roland and a local Taiwanese Partner 4) Roland purchased Rodgers organs, but maintained production at the latter's factory 5) Roland purchased 65% of Italian organ and EM1 maker S.I.E.L. SPA, maintaining production at the latter's factory 6) Roland Electronics Suzhou is wholly owned by Roland (Source: data provided by Roland 28/8/O 1, translated by the author)

The connection between import patterns and factory location strategies becomes

that much more concrete when specific firms are analyzed (see Tables 8.2, 8.3). Until the late 1980s EM1 production at Yamaha took place almost exclusively at the firm's

Toyooka factory on the outskirts of Hamarnatsu. From separate factory visits undertaken in August 2001 and July 2002, I discerned that this facility still plays a role as a kind of focal factory for the latest EM1 production runs. However, Yamaha's most significant investments in new plants for the production EM1 and other products have taken place in

China, Indonesia and Malaysia from the start of the 1990s (Table 8.2). The firm has also allocated some EM1 production to its long-established plant in Mexico. In contrast,

Roland has continually maintained a strong production base for EM1 in Hamamatsu, employing 880 workers in the Hamamatsu region. Its own recent investment in China suggests that it too has used location strategy to lower production costs (Table 8.3). This trend towards the globalization of production can only be understood by looking at the restructuring within the core.

8.3.2 The Local Production System

Table 8.4: The social division of labour in Hamamatsu's musical instrument

(Source: All Japan Commercial Handbook 200 1: 352-358) Table 8.5: Yamaha's social division of labour within its kyolyokokai (supplier cooperative association) in 2001 Tier Total 0-24 25-49 50-99 100-499 500+ Firm Employees emplys emplys emplys emplys emplys Totals One 923 7 4 5 1 0 17 Two 587 2 9 4 0 0 15 Three 1880 3 5 5 5 1 19 Four 736 7 8 3 1 0 19 Total 4126 19 26 17 7 1 70 (Source: Data provided by Yamaha, translated by the author)

Yarnaha's local production system is illustrative of the general morphology of core-firm supplier relations (Table 8.5). Yarnaha orchestrates its production system through a cooperative association (kyoryokokai) comprised of 70 firms. This system has multiple tiers that correspond to the hierarchical model of subcontracting (Hayter

1997:322). The number of firms in each tier is relatively similar, although the third tier accounts for 46 percent of employment. By establishment size, just under two thirds of

Yamaha's suppliers have fewer than 50 employees while a small number of medium- large sized companies (>loo) account for a significant share of employment.

Assuming increasing transactional distance with successive tiers down the sub- contracting hierarchy (ibid), the pattern indicates that if Yamaha anchors the kyoryokokai, then a small group of medium-large sized companies in the third tier constitute the backbone of the supply association.

The structure of Yarnaha's sub-contracting is also organized by function (Table

8.6). Piano suppliers account for 24 percent of firms but only 15 percent of employment.

Conversely, electronics suppliers have only a 6 percent enterprise share, but provide 16 percent of employment. If EM1 suppliers are added to this last tally, the two sectors account for one quarter employment in the local kyoryokokai. As the largest musical instrument maker in the world, let alone in Hamamatsu, the scale of Yamaha's production systems is by no means typical. Nevertheless, in the specific case of EMI, the structure of its production systems is morphologically representative of the overall pattern for Harnamatsu.

Table: 8.6 Yamaha's functional division of labour within its kyoryokokai (supplier cooperative association) in 2001

(Source: Data provided by Yamaha, translated by the author)

In comparison to 1986, Hamamatsu's social division of labour in the EM1 sector has contracted in both the number of producers and suppliers. Table 8.7 shows that the total number of producers have become both fewer and larger to include only the four core firms mentioned previously. Moreover, 93 percent of the parts suppliers employ fewer than 100 employees. There is no information indicating the degree to which the trend towards globalized production has modified the social division of labour at home.

Similarly, we do not know the share of foreign parts that are assembled at Hamamatsu factories. On several tours of EM1 factories belonging to various firms in Hamarnatsu, I

saw parts boxes that had been shipped from Res Ipsa Loquitur.

Table 8.7: A comparison of Hamamatsu's social division of labour for EM1 in 1986 and 2001

(Source: Takeo 1989: 62, All Japan Commercial Handbook 2001: 352-358, translated by the author)

8.4 The EM1 Sector as Part of Hamamatsu's Industrial Mosaic

Ultimately, it is necessary to view the EM1 industry as a component in the local industrial

mosaic. This polysectoral basis makes Hamamatsu a particular type of 'learning region'

(Morgan 1997). According to Takeuchi (1 996), Hamamatsu's industrial evolution has

unfolded via synergistic development. This means that its lead firms have leveraged the

technological complementarities between diverse sectors (Figure 8.3). Yamaha represents just one example of this phenomenon, for its longitudinal profile (see Figure 4.9)

indicates the cultivation of successive competencies to makes not only pianos,

motorcycles, and synthesizers, but the robots that assist in the manufacture of these

products. Another example that illustrates how the local learning style facilitates the

transectorial evolution of firms is Roland's subsidiary Roland DG, which has extended its

scope to produce computer plotters. Beyond EMI, the local style of industrial learning,

86 At another firm's piano factory, industrial hollowing-out (kudoka) had advanced that much further as the production line appeared to be relocating process by process to a greenfield site in South East Asia. This factory had closed by 2003.

303 when viewed collectively, appears as a geo-historical palimpsest featuring successive layers of sectoral competency that build with rather than on, what came before.

Figure 8.3: The Synergistic Development amongst Manufacturing Industries in Hamamatsu

Shizuoka University

Motorcycles . Automobiles

New Materials

Machinery Processing Hamamatsu Region Technopolis

(Source: Translated and modified from Takeuchi 1996: 194)

A further and more recent illustration of Hamamatsu's industrial strength has been the success of its Technopolis. At the national level, the overall success of the

Technopolis program as regional development policy has been mixed (Glasmeier 1988,

Masser 1990, Tsukahara 1994, Sternberg 1995). Hamamatsu, however, remains an exception, exhibiting the best performance of the 36 designated regions (Table 8.8). On the one hand, Hamamatsu's situation, offers significant advantages over more peripheral locations. Its central position along the Pacific industrial belt grants connectivity.

Hamamatsu is also sufficiently removed from the congestion of the Tokyo, Osaka, and

Nagoya metropoli. On the other hand, the existing industrial structure has also played an equally powerful role in reproducing success in start-ups. JETRO (1 998: 8) reports that, (6Historically, Hamamatsu is an area where industries with special technologies have been born. Accordingly, the area has a business climate in which large well-established and prospering companies support and nurture new business ventures." More research needs to be done to understand the way the Hamamatsu's Technopolis contributes to local industry (although see Masser 1990). Table 8.8: The Ranking of Japan's Technopoli According to Several Criteria

Category 1St Rank 2ndRank 3rdRank 5thRank Number of factories with attached Hamamatsu Shinanogawa Do-ou Ehime Yamagatal research function in the technopolis' (Niigata) (Hokkaido) Asama(Nagan0) Number of research institutes North Sendai Hamamatsu Kagawa Hiroshima1 situated in the technopolis2 Kumamoto Number of joint university-business Hamamatsu Akita Utsunomiya Kokubu-ltayata Kurume-Tosu research programs in technopolis3 (Tochigi) (Kagoshima) (Nagasaki) Number of enterprises affiliated Hamamatsu Shinanogawa Kokubu-ltayata Akita Kibi-Takahara with technopolis4 Amount of capital assets belonging Hamamatsu Shinanogawa Kokubu-ltayata Akita Kibi-Takahara 0 to enterprises in the technopolis5 Value of factory shipments from Nishi-Harima Utsunomiya Hamamatsu Ehime Toyama enterprises in the technopolis6 (H~ogo) Value added from enterprises Nishi-Harima Hamamatsu Shinanogawa Utsunomiya Toyama affiliated with technopolis7 7 Notes: Data compiled from 1980-96 Data compiled from 1980-96 Data compiled from 1983-95 Data compiled from 1983-95 Data compiled from 1983-95 Data fiom 1994 ' Data from 1994

(Source: On to hikari to iro no mirai toshi: Hamamatsu chiiki tekunoporis, Shizuoka-ken 1997 pgs. 9,lO) 8.5 The Maturation of ICT: Implications for the Yamaha - Roland Relationship and by Extension, the Health of the EM1 Industry in Hamamatsu

Takeuchi's (1 989) model indicates that ICT technologies have become critical lynchpins

in Hamamatsu's industrial mosaic (Figure 8.3). Generally, the pervasiveness of ICT

technologies hinges on their compatibility. Machines of various kinds have to be able to

work together to transfer and manipulate bits and bytes of data. Similarly, in the case of

EMI, the MIDI protocol is the format for connecting instruments from various

manufacturers. With Yamaha and Roland now dominating the global market for EMI,

their position relative to MIDI has become increasingly central to all the technologies and

industries that employ MIDI. I now discuss how Yamaha and Roland shaped this

architecture of interconnection in order to shed light on the contemporary relationship

between Hamamatsu's two main EM1 makers.

In 1982, the first version of MIDI resulted from the collaboration of four Japanese

firms and Sequential Circuits (see chapter three). By the late 1980s, MIDI was the lingua fianca of electronic musicians. The protocol was revised in 1991 to the General MIDI

(GM) format, which solved some of the teething problems of the earlier versiong7.This

new tone generation format though not a formal standard, was again recognized by the

Japan Association of Musical Electronics Industry (AMEI) and the MIDI Manufacturers

Association (MMA) as "Recommended Practice" (Yamaha 2001). The cooperative

aspect of these public dealings obscures the motivations of private interests in the

87 The problem with the first MIDI (1982-91) was that the only types of standard information that could be transferred between instruments were note-onfnote-off information and pitch values. The banks of pre-set voices had no standard catalogues. Roland's Al might have been a piano, while Yamaha's Al was a snare drum (Vail 1993). background. Colbeck (1996: 99) points out that Roland seized advantage with the release of GM which,

was another standard that Roland not only helped promote but practically helped itself to, since the first GM product (the Sound Canvass) was actually in production as GM was still in the process of being ratified.

This timing allowed Roland to derive large profits from the Sound Canvass technology which became, "the de facto standard in the emerging multimedia race" (ibid). Seven years later in 1998, the protocol was revised once again to General MIDI 2 (GM2).

Quoting from a Yarnaha (2001) press release,

The GM format only specified the minimum number of available instrument sounds, simultaneous note polyphony and so on. The need soon emerged to expand the format so it could be applied to a wide diversity of musical genres and be used to create data displaying a richer range of musical expression. Accordingly, both Yamaha and Roland moved to develop their own proprietary formats, XG and GS respectively, and each company endeavored to popularize its respective format to enable more advanced musical expression.

The existence of two different formats, XG and GS, has however created inconveniences for the MIDI instrument industry and for software vendors and users alike. And this state of affairs was further complicated in 1998 when GM went through an upgrade to GM2: GM2 is compatible with neither the XG nor the GS format, so its inception created yet another format and increased the total in use to three.

In 2000, Yamaha and Roland got together to resolve these difficulties.

The solution eventually arrived at was that both companies would actively support the GM2 format as the global standard while at the same time allowing open access to their respective XG and GS Formats in order to develop hardware and content for both formats.

Interestingly this joint press release from the US NAMM trade show on January 19th

2001, described Yamaha as being headquartered in Hamamatsu, and Roland's HQ as

"based in Kita-Ku, Osaka". No mention was made that these firms, for all intents and purposes are neighbors and that most of the latter's manufacturing and intellectual capital is tied up in Hamamatsu. The computer I am using now, a Toshiba, came with Yamaha

XG and Roland GS as its factory sound defaults. The agreement of 2001 ensured the co- existence of these firms in technological space; an industry's architecture of interconnection determined via dialogue.

The nature of competition (and collaboration) between Yamaha and Roland has matured along with the EM1 sector. On MIDI, they are obliged to work together. With respect to everything else, they compete. In the digital EM1 synthesizer segment, the competition is head to head. However, in the business lines that feed back towards the

ICT carrier industries (e.g. LSI, digital plotters) the two firms do not venture onto the others' turf. Their spatial strategies with respect to both production and innovation systems unfold in a landscape of countervailing power that conforms roughly to the locational overlap model (Vernon 1971, Hayter 1997). Roland's continual commitment to situate the balance of its EM1 production in Harnamatsu is the one aspect that deviates from this form, while distinguishing Roland from its neighbors.

The explanation for Roland's resiliency against the forces of hollowing-out

(kudoka) lies in the way it organizes its manufacturing as a 'cell' production system as opposed to an assembly line (Shizuoku Keizai Shinbun 2002). This flexible form of

'human centred' production is emerging as a new 'best-practice' within Japanese manufacturing (Isa and Tsuru 2002: 550).

A cell production system is a production system in which a single worker or small team of production workers (two to five members) perform multiple production jobs (multitasking) in short segment lines. Cells are, with few exceptions, arranged in U- shaped lines in which unfinished components enter at a point adjacent to the point that they leave as finished products. The cell design places a wide array of tools and equipment in close proximity to the workers, enabling them not only to perform a wide range of production tasks but to customize the products as well.

Cell production is ideal for small-lot production runs and allows the firm to respond rapidly to changes in demand. Another implication of cell production is that the morphology of Roland factories is very different from Yamaha. The latter's Toyooka factory was still producing synthesizers and Electones on linear production lines in 2002.

Consequently Yamaha and Roland's production systems differ primarily at their core.

On the other hand, the innovation systems of both firms are remarkably similar.

Basic research takes place at home, which for both firms is Hamarnatsu, despite Roland's nominal HQ in Osaka. In Roland's case at least, the R&D core feeds on satellite R&D facilities and contract inventors located in largely similar locales around the world

(Kikumoto Tadao, Interview 15/ 12/02).

Institutional incompatibilities or 'mismatching' (Freeman l982), in part covered in chapter three, prevented the US industry from crossing the great divide to the digital frontierg8.In comparison, the absence of institutional conflict is an enduring distinctive characteristic of the competition among Hamamatsu's leading EM1 manufacturers.

Operating in close quarters, these firms co-exist as 'sparring partners'. One of the defining features of Yamaha's innovation system during the late 1960s and early 1970s, the cusp of EM1 'technical crystallization', was the practice of splitting the engineering division of labour to undertake parallel prototype development (e.g. the trials between the

PASS and NES systems at Yamaha; see chapter six). The collocation of firms in

Hamamatsu is a larger regional version of that practice, with firms as parallel models.

88 Three examples: TechnicalIPersonal conflicts between Deutsch and Allen Organ, Informal Loose agreements that went awry (Sequential's royalty conflict with EMU), Moog's fate under the conglomerate Norlin. Yamaha and Roland, two lead models, and Kawai and Hammond-Suzuki, their 'loyal opposition' (Hayter 1997), lent a tremendous diversity to the local industrial ecology.

On the surface, the internal architecture of the Hamamatsu cluster is straightforward. Except when aligning themselves around formalized institutions like

MIDI, there are next to no horizontal relations between these firms except as competitors.

Yet at a deeper level, these firms know very well what each other are up to, in a general sense. In technological space, they are acutely aware of each other's patent positions. In the market, their products overlap and compete for space. They 'chase each other around the globe' investing in new factories. On the other hand, their local production systems are mostly mutually exclusive of each other. Formal and informal R&D alliances are similarly absent in the innovation system. 'Untraded interdependencies' (Dosi 1988,

Storper 1997) are everywhere but invisible. Competition, on the other hand, is entirely visible. For example, at the entrancelexit to the Shinkansen platform at Hamamatsu station, Yamaha and Kawai have installed booths on either side of the concourse to demonstrate their instrument technologies. Given these characteristics, how does one explain the emergence and persistence of Hamamatsu as the core of the global EM1 industry? With Alfred Marshall's famous aphorism: There's something in the air ... ?

Well, not quite. As noted earlier in chapter three, it is difficult for even Japanese outside of Hamamatsu to translate 'yaramaika ', a word peculiar to the local dialect. The closest English approximation is 'challenge everything'. Whatever it means, 'yaramaika ' captures the je ne c 'est quoi of industrial learning in Hamamatsu. Indeed, there is something of an X-factor to Hamamatsu's success that eluded this researcher.

Nevertheless, its dominance as an industrial core is inescapable. The global musical instrument industry, including the EM1 sector literally all comes together in Hamamatsu.

All four circuits of Latour's vascular system (Figure 2. 1)' 'instruments' (a particularly apropos term in this case study), 'colleagues', 'allies' and the 'public' are bound together and pulsate according to the beating heart of Hamamatsu. CHAPTER 9: CONCLUSION

Economic geography has largely overlooked analysis of radical change, at the heart of which is transectorial innovation. To address this lacuna, this thesis framed its explanation of the rise of Hamamatsu as the world's leading centre of EM1 from the perspective of inter-sectoral knowledge transfers and innovation. This concluding chapter offers three sets of summary reflections. The first relates the lessons of Hamamatsu's experience for economic geography's analysis of innovation. The second part notes contributions of this analysis for Japanese industrial geography. The last part of the chapter presents several directions for future research.

9.1 Hamamatsu's Changing Place in the World

Hamamatsu (@&) translates as 'coastallbeach pine'. Ten years before he seized power and rooted the bakufu system89 in Edo (Tokyo) in 1600, Tokugawa Ieyusu bestowed this toponym on a place that had formerly been known as Hikuma (51 ,R),which means

'pulling the reins'. Harnamatsu is situated at the mid-point along the nation's principal artery between the Kansai and Kanto regions, the famous Tokaido highway. Tokugawa

he Bakufu or shogunate was Japan's centralized military government Fat ruled from 1603 until 1868. Sakoku or national isolation defined Japan's relation to the rest of the world, whil; the internal 'relational space' (Smith 1986) was governed by the Sankin Kotei, or alternating attendance system. did not want a place of such strategic importance to be saddled with connotations of retreat so he selected a luckier, more stable and rooted appellationg0.

Since the Meiji Restoration, Hamamatsu's place in the world has changed considerably. Nevertheless its toponym endures as a fitting metaphor for the region's modem industrial evolution, a trajectory given impetus by Yamaha Torakusu. Strong root systems anchor coastal pines against the elements. Similarly, the musical instrument industry provided the basis from which Hamamatsu's industrial mosaic evolved into the information age. Yamaha developed its capability to produce microchips because it wanted to make a better sounding organ (Nakagawa 1984, Yamaha 1987, Johnstone

1999).

'Radical' derives from the Latin radicalis - 'from the roots'g1.With these etymological origins in mind, a geographical understanding of radical innovation involves tracing the roots of evolutionary processes to connect origins and outcomes, whether these are at a distance or not. The origins of the trajectory I studied, namely the post-war 'quantitative revolution' in instrument design, lie in the US and they have been well documented (Chadabe 1997, Pinch and Trocco 2002) as have the implications of radical technological change for musical practice (Theberge 1997). Missing, are accounts that foreground the geographical dimensions of this (r)evolution especially the ascendancy of Hamamatsu as the industrial core. Once place and space are brought into the picture, explanations of technical change in EM1 can begin to highlight the agency of

Japanese firms. It is these enterprises who translated and tied together the various actors

90 Thenceforth, the Tokugawa Shogunate installed only its closest allies in Hamamatsu and reciprocated loyalty by thereafter appointing many of these lords to positions of prominence in the central government. For this reason, Harnamatsu castle is known as Shusse-jo or 'Castle of Promotion'. The Concise OED (Fifth Edition) 1964. that collectively maintain the network of interests associated with EMI. These include: patents, machines and other 'instruments', engineers, inventors and other 'colleagues', managers, financiers and other 'allies' and finally the 'public' - the galaxy of musicians, professional and amateur who use these technologies to produce music, but also the

'general public' who employ digital EM1 every time they turn on their computer or answer their cell-phone (Latour 1999; see figure 2.1). In concert the 'heterogeneous engineering' of these varied interests has propelled the musical instrument industry into alignment with the pervasive impacts of the ICT techno-economic paradigm

Hamamatsu's EM1 industry evolved via a succession of ex-regional linkages which have complemented its local roots. These exogenous and endogenous factors are interdependent and hinge on the cycling between codified and tacit knowledge.

Hamamatsu's firms, most notably Yamaha and Roland, have developed quite distinct strategies in forging non-local and local links. Texts (books, patents), artefacts (products) and people (jatoi) represent the conduits through which these firms mobilized foreign knowledge in particular ways to enact the general philosophy of shuutoku or 'learning by getting'. In Latour's (1987) terms, shuutoku, might be considered as the process by which

Hamamatsu's firms have 'mobilized the world', so that it converges on the centre. On the other hand, evidence of overt local horizontal linkages is hard to come by. Nevertheless, I have characterized the nature of competitive and cooperative strategies between neighbours as being akin to the parallel development trials that take place within firms. 9.2 Contribution of the Study to the Theory of Innovation and Formation of Knowledge

This thesis frames the processes of transectorial innovation, radical technology transfer and the re-embedding of knowledge within Hamamatsu's firms as a 'diffusion of the engineering disciplines' (Rosenberg 2000) which unfolds as a 'chain of translations'

(Latour 1999), the sum of which constitutes a 'spatial innovation system' Oinas and

Malecki 2002). The general empirical details of this trajectory are contained in Table 9.1. Table 9.1: The Evolution of the EM1 Transectorial Innovation Svstem Phase in Life- Characteristics Representative Examples Key Locations Cycle Inspiration - from Localized transectorial filtering of - Schillinger's "A Mathematical Theory of - Bell Labs carrier sector (1949- knowledge Musicn(l949) - Aerospace Sector (US) 63) - Olson's Music and Physics (1952) - The Engineer-Composer Transrenional -National Innovation System - Kakehashi's clock-electronics repair Inside Technologies Black Box Inspiration - Japan -Reverse engineering philosophy - Translation of Olson's Music ancj Physics (1966) - Kakehashi's TB Ward at the (general 1867- -Corporate and amateur tinkerers Osaka Hospital present) (EM1 specific -Access to foreign technology is slow - Tucked away in a corner 1959-present) and structured in interests of larger (katasumg Yamaha Piano firms Factory Entre~reneurialand - Diffusion of texts US - Upstate-NY lntrepreneurial - Diffuse small-scale development of - Chowning reads Matthews article in Science - San Francisco Bay Area Invention - (US) pre- prototypes and initial patents (1963) -AES radical incremental - Small scale interactions between - Moog synthesizer (1966) innovation instrument makers and the musical - AES sessions on EM1 (1966-72) (1966-80) avant-garde Switched on Bach Japan The R&D Laboratory -Yamaha Electone D-I (1959) - Hamamatsu (Yamaha) - Kakehashi's organ for Matsushita (1960) - Tokyo (Korg) - Kakehashi (Ed.) "Everything about Electronic - Saitama+ Hamamatsu (Kawai) Musical Instruments" (1966) - Osaka 3 Hamamatsu (Ace - Hammond International Japan : 50% Ace Electronics1 Roland) Electronics (1968) Technical - Codification of key patents - Digital Organ: Rockwell-Deutsch-Allen Organ (1967 -Technological Space Crvstallization - - Protocols invented) (1970 codified- patent) (1973 Deutsch to - California: Stanford, Rockwell Concentration of - Selective absorption of foreign talent Yamaha) - Hamamatsu Technical Leadership -FM synthesis (Chowning-Stanford-Yamaha) (1968 (1969-82) invented) (1973 codified AES) (1977 codified patent) - MIDI: Roland and Sequential Circuits - leaders ; Kawai, Yamaha, Korg and Casio followers (1982) Commercial - Localized firm-specific monopoly - Yamaha DX7 Hamamatsu - World Crvstallization (1983) returns, pervasive impacts Take-off (I980-87) EmulationlAjustment of best-practice by - Casiotone: department store models (1980) Japan firms ~a~ableof economies of scale - Roland D50: sampling (1987) Phase in Life- Characteristics Representative Examples Key Locations Cycle Consolidation - Acquisitions, firm failures, absorption of - Yamaha buys Sequential and invests in Korg Horizontal Integration talent - Roland buys Rodgers, SElL California: Stanford (1987-90) - Kawai buvs Lowerv I - Suzuki b&s ~amiond'Brand' MaturitvIReiuvenation Incremental Innovation ~~p I I - Successive MIDI generations (1 990-?) Spatial division of labour, - EM1 in cell phones, PCs connective Architecture (MIDI) Transectorial feedbacks [Hamamatsu). Software (US) The research makes two important contributions in the study of knowledge creation. The first involves a specific spatial dimension. Until recently, economic geography's conceptualization of technological learning has manifested a bias towards place-based perspectives that are concerned with understanding the ways in which the formation of local tacit knowledge fixes territorial advantage. Against this grain, a number of scholars have called attention to the role that non-local linkages play in the formation of innovation systems (Amin and Cohendet 1999, Gertler 200 1, Bunnel and

Coe 2001, Oinas and Malecki 2002, Coe and Bunnel2003). With few exceptions, this latter research trajectory has sought to unravel the emerging industrial geographies of a

'globalizing' world. Its focus is very much on the present. In contrast, there has been little attempt to harness these perspectives to the interpretation of historical processes of inter- regional network formation.

Secondly, through the analysis of historical processes, my research challenges the assumption that codified knowledge constitutes a 'ubiquity' (e.g. Maskell and Malmberg

2001). Though texts can be mobilized to facilitate technology transfer, I have shown that this is a far from straightforward process. Knowledge is codified in particular ways in specific circumstances by a range of agents who subsequently mobilize that knowledge with varying motives and efficacies. In short, knowledge must be 'translated' between various interests, and this inevitably involves some degree of 'slippage' (Latour 1987).

To this end, I have presented considerable empirical evidence which details the problematic ways in which tacit and codified knowledge inter-relate. In this sense, I agree with Arnin and Cohendet (199997) who question the, "separability of the two forms of knowledge by suggesting that business networks largely dependent on local tacit knowledge and incremental learning may prove to be inadaptable in the face of radical shifts in markets and technologies." Yet these same comments imply the primacy of tacit knowledge in understanding radical change and local competitive advantage. In sympathetic contrast, this thesis offers an enriched understanding of tacit and codified knowledge by revealing the vital importance of the latter to transectorial innovation.

In particular, I emphasize that during episodes of radical change, the codification of knowledge - one form of 'mobilizing the world' in 'centres of calculation' (Latour

1987) that accompanies other situated practices such as tinkering and reverse engineering

- provides the mechanism for emerging regions to displace the entrenched advantages of cores. For this reason, the thesis highlighted patenting as a key situated process for inscribing technological space in a proprietary manner. By wedding the 'fuel of interest to the fire of genius', patents connect human agencies with social structures. These agencies are as diverse as people's personalities: Ralph Deutsch patented everything, while Dave Smith patented nothing (see chapter six). The 'interests' here are also manifold. Corporations with a large patent portfolio claim technological space in a way that thoroughly interlinks the firm's property to a host of other interests that include previous inventions ('prior art'), other firms, patent agents, courts of law, and even the consuming public.

Chapter five demonstrated that Yamaha stood head and shoulders above the rest of the field in its commitment to patenting. In pattern, Yamaha's behaviour is significant

- the US patent office awarded 856 patents to Yamaha between 1965 and 1994, far more than any other company. Nevertheless, this story turns on specific events and particular translations. Twice, at critical junctures that not coincidentally both took place in 1971 in California, Yamaha made decisions about very particular technologies (the Deutsch

'digital organ patent' and the Chowning 'FM synthesis patent') that, once developed and combined in the DX-7, proved radical in their impact.

On the one hand, the DX7 could perform a range of functions that were qualitatively different from previous instruments: it was programmable, polyphonic and it could be linked to other EM1 via its MIDI ports. In this respect, the DX7's technical characteristics were indeed 'radical' as Pinch and Trocco (2002) infer. However, from this standpoint, its 'radicalness' is still 'black boxed', to use both Rosenberg's (1 982) and

Latour's (1 987) terms. On the other hand, from a geographical standpoint, the DX7 is radical in a quite different way. Specifically, the knowledge 'traces' that lie behind it and stretch ahead of it, connect a set of interests on opposite sides of the Pacific Ocean.

Indeed, the Stanford University - Yamaha relationship persists to this day, with funds from various Stanford's licenses, to Yamaha in particular, underwriting the former's computer music centre. These complementarities serve to maintain the network that translates Yamaha's interests as a manufacturer with Stanford's interests as an institution that generates knowledge. This connection hinges on texts. Clearly, knowledge codification has tremendous implications for influencing the topology of connections in space as well as the ways in which firms, and by implication regions are locked-in or locked-out of technological progress.

Finally, this thesis also offered an interpretation of Schurnpeter's Mark I and

Mark I1 forms of innovation in the Japanese context. It developed this perspective through the analyses of longitudinal profiles of Roland and Yamaha's technological strategies as well as qualitative accounts of the manners by which these two firms internalized accumulated knowledge and jockeyed for position in technological space.

Based on this evidence, my research contributes to the industrial geography of Japan

literature in a number of ways.

9.3 Contributions to the Industrial Geography of Japan Literature

First, the thesis illustrated how at the start of the ICT techno-economic paradigm

in Japan both Mark I (entrepreneurial) and Mark I1 (large firm) innovators relied

substantially on the practice of tinkering with off-the shelf and usually foreign technology. In doing so, they prised open technology's black box, learning from but also

challenging everything about its 'ready-made' status. Gradually, management disciplined

these cultures of tinkering into innovation systems with formalized practices for the translation of non-local knowledge. Yamaha's President Kawakami instructed the R&D

head to 'work from the documents'. Later, trials between competing technologies

constituted another systemic practice at that firm. In Roland's case, that firm's founder

Kakehashi Ikutaro proved to be the consummate tinkerer, whose origins lie in watch and

electronics repair - in highly unique circumstances such as the TB ward at Osaka

hospital. Contingencies such as these, no matter how idiosyncratic, matter.

Second, my analysis of Roland as a Mark I 'entrepreneurial' innovator served as a

case study to interrogate how Asanurna's (1 989) relation-specific skill, operates in

dynamic spatial and temporal contexts to shape the horizontal strategies of firms as they

evolved in size and orientation from small suppliers into core finns whose stature was

that which Hayter (1997) terms 'big firms locally'. From the outset, Roland has had to

operate in a technological space that, in areas like patenting, was biased towards the

interests of Mark I1 innovators such as Yamaha. On top of this, the distribution systems of the large firms locked Roland out of the domestic market, forcing it to cultivate foreign

'relational market intelligence' (Reiffenstein et a1.2002) early in its evolution. Given this situation Roland could only have begun by drawing on external economies, and by being a supplier to other firms which allowed it to accumulate a stock of both relation-specific

(Asanuma 1989) and enterprise specific skills (Patchell and Hayter 1995) by supplying

Matsushita and Hammond. Later, Roland's structure further evolved when Kakehashi devolved power and relinquished the oversight of innovation when he hired Kikumoto

Tadao from the electronics sector to head the R&D division. The account of how

Kikumoto sought to challenge Yamaha's DX7, and by implication the FM patent by going back into the literature, suggested that particular inscriptions, no matter how scrutinized, deconstructed and contested, can persist to claim space in decisive ways. In short they manifest the quality that deLaet (2000) terms 'tenaciousness'.

Third, the juxtaposition of Roland and Yamaha with a focus on their interdependencies enabled us to view these enterprises as the lead firms in a highly unique learning region centred on Hamamatsu, a place whose signature is industrial development based on synergistic diversification or transectorial innovation. Several authors have characterized Hamamatsu in this manner (Ohtsuka 1980, Takeuchi, A. 1996,

Takeuchi, H. 2002); although our understanding is limited as to how inter-firm dynamics in the horizontal dimension of the cluster contributed to this regional style of learning. In this sense my analysis of the evolution of Roland and Yamaha as 'sparring partners' or local 'parallel models' of development lent an institutional perspective to these accounts that was centred on the firm yet was nuanced enough to identify the agencies of individuals within and amongst these contexts. In particular, my discussion in chapter six of the renaissance in foreign vectors who contributed to the formation of knowledge in

Yamaha and Roland extends the geographical literature on Hamamatsu by directing attention at the ex-regional linkages of firms.

9.4 Agenda for Future Research

This thesis has provided substantial insight into the geographies of musical instrument manufacturing, in particular by noting how the contingencies of knowledge formation and information transfer amongst agents contributed to the spatial dynamics in industrial advantage. In doing so it has drawn the connections between a technological discontinuity, a spatial discontinuity and the subsequent spatial concentration of industry in a particular place. Beyond the resolution of these original objectives, this research opens the door for further investigation of a number of topics grouped according to two themes. The first theme of subsequent inquiry involves further understanding

Hamamatsu's place in the world. The second themes concerns elaborations on the concept of transectorial innovation

One broad avenue that needs to be better addressed concerns the labour dimension to the manufacturing geography of EM1 in Hamamatsu. In particular, little is known about the social economy of the labour process in production, and how this might be shaped by the characteristics of the labour force. For instance photos of Yamaha's

Miyatake piano parts factory in the early 1960 show that most of the production labour force were young women, many of whom, the accompanying text notes, were migrants from depopulating rural regions like Hokkaido(Yamaha 1987: 30). We do not know the degree to which this gendered division of labour, which provided Japanese firms with

'nimble fingers' (Partner 1999: 207) at a low cost, persists today. Similarly, though studies have documented the social implications of Hamarnatsu's increasing foreign population (Yamanaka 2002)~~,the focus is on the world outside of the workplace.

However, there is recognition that foreign labourers such as Brazilian Nikkei are playing an increasing role in the manufacturing industries of Shizuoka and eastern Aichi prefectures (Belson 2003), not necessarily within core firms but in firms at lower tiers in the supplier hierarchy. At the level of the firm, strategies to employ foreign labour need to be understood in the context of the evolving international internal and social divisions of labour. In doing so, Hamamatsu's place in the world could be interpreted by drawing connections between the spatial implications of Kudoka, or the hollowing out of manufacturing, and the converse practice of bringing the world to Hamamatsu, in the form of foreign labour.

A second avenue for research is to address, in general terms, the institutional characteristics of the Japanese patent system. Anecdotally, Roland's founder Kakehashi emphasized the way that this system was biased to the interests of larger firms in the

1960s. Since that time Roland has devoted far less attention to patenting its own technologies, even as it has been compelled to remain abreast of the advances of its rival

Yamaha. Clearly much more research must be done to appraise how the domestic patent system has regulated the knowledge economy in Japan, both in terms of how it institutionalizes the relations between Japanese firms, as well as how it establishes a boundary to the literal and figurative translation of foreign knowledge. This research program would enable a slightly different framing of Japan's and Hamamatsu's place in

92 Hamamatsu's foreigner population increased 600% to 10,000 over the 1990s following the Revised Immigration Law which de-regulated employment for immigrants of Japanese descent. Brazilian Nikkei now account for two-thirds of Hamamatsu's foreign population (Yamanaka 2002). the world, this time by addressing the social dimensions to the policing of inter-national knowledge frontiers.

Additional research is needed to understand how instrument manufacturers link themselves with the cultural industries of musical production. Scott (1 999) has written about the organizational geographies of music recording firms in two such nodes in the

US: Los Angeles and Nashville. An appreciation of backwards linkages with producer goods (instruments and recording technologies) firms would serve to enhance that research by broadening the scale of analysis. We know for instance that both Yamaha and

Roland have recruited music producers and retail specialists for contract work. Moreover, like athletic shoe companies, instrument manufacturers covet professional endorsements as a further way of shaping market tastes. Additionally, Theberge (1 997) devotes attention to the ancillary though crucial role played by publications geared towards both producers (The Music Trades) and users (Keyboard, Computer Music Journal). These various strands of analysis need to be woven together to explain the geographies at the producer/consumer interface, especially the degree to which these connections are shaped by central forces (i.e. from Hamamatsu) or local circumstances.

The second broad theme for future research, involves further investigation of transectorial innovation as a general process. For instance, the widely cited research on the British motor-sport industry (Pinch and Henry 1999, Henry and Pinch 2000), illustrates how personnel and technology transfers from the aerospace industry contributed to the formation of that milieu. Transectorial innovation is a largely implicit theme in those reports, but its importance could be made more explicit for comparison purposes with other case studies. For instance, intuitively, the digital camera sector, which is largely a population of Japanese camera and electronic firmsg3,presents an obvious case to elaborate on the geographical impacts of transectorial innovation.

More generally, patent analysis similar to that found in chapter five would reveal possible transectorial linkages in the formation and evolution of a range of different industries. As an example, an anticipated result of my own filtering of US patent data in the field of EM1 confirmed that the geographical shift in patent activity from the US to

Japan prefigured the ensuing upheaval in the industrial ecology and the shift in competitive advantage to the latter. An entirely unanticipated consequence of that analysis was my discovery of the importance of transectorial sources of knowledge: principally the aerospace sector and the agency of individuals like Ralph Deutsch, whose

'mobilizations of the world' which include The Orbital Dynamics of Space Vehicles

(1 963) and US Patent #35 15792 (1970), mark an individual's transectorial migration.

Who knows what sorts of stories might come out of analyses of this nature?

9.5 A Final Word

Naturally, since hatching the idea for this topic, I have listened closely to the way EM1 colour my experience of music. For example I recently attended a Santana concert where, beyond the soulful guitar pyrotechnics of the band's eponymous leader, my attention was tuned into the accompaniment provided by veteran keyboard player Chester Thompson whose rig consisted of three instruments: a vintage Hammond B-3 organ, a Roland VS 2 synthesizer, and a Yamaha Motif 7 instrument. Different songs demanded different textures and this configuration lent Thompson the necessary flexibility to match the mood

93 Even Kodak's digital cameras feature a label which reads that the products are designed in Japan and manufactured in the Philippines for the Eastrnan-Kodak company of Rochester, New York. set by his bandleader. The juxtaposition of the Hammond, a 'classic' instrument first produced in 1954, and the state-of-the-art Yamaha illustrates, quite clearly I think, that the latter has in no way replaced the former in contemporary musical practice. Indeed, these two instruments, which bookend an entire Kondratiev wave worth of technological change in instrument design, complement each other.

Similarly, despite the profound regional differences over time between the US and

Japan in their ability to implement the transectorial innovation of EMI, an

interconnection between these places endures. It is the way that the two regions have been made to work together that matters and constitutes the principal lesson of this research. In light of recent calls by geographers to gain a better understanding of trans- regional innovation networks, my research demonstrates that these interconnections need to be contextualized in an evolutionary manner. Concepts such as Freeman and Perez's

(1987) notion of techno-economic paradigms (TEP) serve to frame these evolutionary

processes. However the contingent and nuanced geographies of translation that arise in

particular places reminds us that space-based perspectives still need to be grounded in

regional geographies, in which places like Hamamatsu are conceived as 'meeting places'

(Holmen 1995) for endogenous and exogenous forces. BIBLIOGRAPHY

Abegglen, J. and Stalk, G. 1985. Kaisha: The Japanese Corporation New York: Basic Books Acs, Z. Audretsch, D. and Feldman, M. 1994. R&D Spillovers and Recipient Firm Size. Review of Economics and Statistics 76:336-40. Almedia, P. and Kogut, B. 1997. The Exploration of Technological Diversity and the Geographic Localization of Innovation. Small Business Economics 9:2 1-31. Amin, A. 2000. Industrial Districts.In E. Sheppard and T. Barnes (eds.) A Companion to Economic Geography Oxford: Blackwell (pp. 149- 168). Amin, A. and Cohendet, P. 1999. Learning and Adaptation in Decentralized Business Networks. Environment and Planning D: Society and Space 17: 87-104. Anderton, C. 1988. Twenty Great Achievements in Musical Electronics 1968-1988. Electronic Musician (July): 28-98. Angel, D. and Savage, L. 1997. Globalization of R&D in the Electronics Industry: The Recent Experience of Japan. In R. Lee and J. Wills (eds.) Geomaphies of Economies London: Edward-Arnold (209-2 18). Aoki, M., Gustafsson, B. and Williamson, 0. (eds.) 1990. The Firm as a Nexus of Treaties. London: Sage. Armbrsuter, G. 1983. Portable Keyboards: Thanks to Microchips and Mass Marketing They're Popping Up Everywhere. Keyboard June: 20-3 1. Armbruster, G. 1984. The Art of Electronic Music New York: Quill. Asanuma, B. 1989. Manufacturer-Supplier Relations in Japan and the Concept of the Relation-Specific Skill. Journal of the Japanese and International Economies 3: 1- 30. Asheim, B. 1999. Interactive Learning and Localized Knowledge in Globalizing Learning Economies. Geoiournal49: 345-52. Audretsch, D. and Feldman, M. 1996. R&D Spillovers and the Geography of Innovation and Production. American Economic Review 86: 630-640. Audretsch, D. and Stephan, P. 1996. Company Scientist Locational Links: The Case of Biotechnology. American Economic Review 86: 64 1-652. Baker, R. 1976 New and Improved - Inventors and Inventions that have Changed the Modern World London: British Museum Publications. Barnes, T. 1996. Logics of Dislocation: Models, Metaphors and Meanings of Economic Space New York: Guildford Press. Barnes, T. 1999. Lives Lived and Lives Told: Biographies of Geography's Quantitative Revolution. Environment and Planning D: Society and Space 19(4): 409-29.

Barnes, T. 2003. "What's Wrong with Regional Science?" - Invited Plenary Presentation, Canadian Regional Science Association, Victoria, BC, May 30th, 2003. Barnes, T. 2004. Culture: Economy. In P. Cloke and R. Johnston (Eds.) Spaces of Geographical Thought: Deconstructinn Geomaphy's Binaries London: Sage (pp. 6 1-80) Bathelt, H., Malmberg, A. and Maskell, P. 2003. Clusters and Knowledge: Local Buzz, Global Pipelines and the Process of Knowledge Creation. Paper prepared for the 99' Annual Meeting of the Association of American Geographers, 4-8 March 2003, New Orleans, LA. Belson, K. 2003. Rethinking the Town that Toyota Built; as Car-maker Moves Plant Overseas, its Japanese Base Fears its Future. The New York Times 10121 : Cl Best, M. 1990 The New Competition Cambridge, MA: Harvard University Press. Blomley, N. 2003. Law, Property and the Geography of Violence: The Frontier, the Survey and the Grid. Annals of the Association of American Geographers 93:121-141. Blomley, N. 2004. Unsettling the City: Urban Land and the Politics of Property New York: Routledge. Bluestone, B. and Harrison, B. 1982. The Deindustrialization of America New York: Basic Books. Braczyk, H., Cooke, P. and Heidenreich, M. (eds) 1998. Regional Innovation Systems: The Role of Governances in a Globalized World. London: UCL Press. Breschi, S. and Lissoni, F. 2001. Localized Knowledge Spillovers vs. Innovative Milieux: Knowledge "Tacitness" Reconsidered. Papers in Regional Science 80(3): 255-274. Breshnaham, T. and Trajtenberg, M. 1995 General Purpose Technologies: "Engines of Growth." Journal of Econometrics 65:83- 108. Britton, J. and Gilmour, J. 1978. The Weakest Link: A Technological Perspective on Canadian Industrial Underdevelopment Ottawa: Science Council of Canada. Brown, L. 1981. Innovation Diffusion London: Methuen. Bryson, J., P. Daniels, N. Henry and J. Pollard (eds.) 2000. Knowledge, Space, Economy. London: Routledge. Bunnel, T. and Coe, N. 2001. Spaces and Scales of Innovation. Progress in Human Geography 25(4): 569-589. Carlsson, B. and Stankiewicz, R. 1991. On the Nature, Function and Composition of Technological Systems. Journal of Evolutionary Economics 1: 93- 118. Ceh, B. 200 1. Regional Innovation in the United States: Evidence of Spatial Transformation. Papers in Regional Science 80(3): 297-3 17. Ceh, B. 1996. Temporal and Spatial Variability of Canadian Inventive Companies and Their Inventions: An Issue of Ownership. The Canadian Geographer 40: 3 19-37. Chadabe, J. 1997. Electric Sound: The Past and Promise of Electronic Music Upper Saddle River, NJ: Prentice Hall. Christopherson, S. 1999. Rules as Resources: How Market Governance Regimes Influence Firm Networks. In T.Barnes and M. Gertler (Eds) The New Industrial Geography: Regions, Regulations and Institutions London: Routledge (pp. 155- 175). Clark, J. Freeman, C. and Soete, L. 1983. Long Waves, Inventions and Innovations. In C. Freeman (Ed.) Long Waves in the World Economy London: Buttenvorths (pgs 63-77) Coe, D. and Helpman, E. 1995. International R&D Spillovers. European Economic Review 39: 859-887. Coe, N. and Bunnell, T. 2003 'Spatializing' Knowledge Communities: Towards a Conceptualization of Transnational Innovation Networks. Global Networks 3(4): 437-456. Colbeck, J. 1995. Take Ten Synths.. .,A Journey Through Synthesizer History. Sound on Sound (Dec.). [Obtained during December 2003, on-line at: http://www.soundonsound.com/sos/1995~articles/dec9511Ogreatsynths.html?sessi on=95455209cfc 1c8bOc6 1edOdd6b 190eOa ] Colbeck, J. 1996. Keyfax Omnibus Emeryville, CA: Mix Books. Cole, D. 2003. Fanfare for the Uncommon Piano. The New York Times (June 6": E29). Cowell, B. 2001a. Interview with John Lehmkuhl. Sonikmatter: Music +Technologv+Mind http://interview.sonikmatter.com/data/j ohn_lehmkuhl.php. Accessed 1411 0102. Cowell, B. 2001b. Interview with Nick Howes. Sonilunatter: Music +Technology+Mind http:l/interview.sonikmatter.com/data~nick~howes.php.Accessed 14110102 David, P. 2000. Path Dependence and Varieties of Learning in the Evolution of Technological Practice. In John Ziman (Ed.) Technological Innovation as an Evolutionary Process New York: Cambridge University Press (1 89-203). de Laet, M. 2000. Patents, Travel, Space: Ethnographic Encounters with Objects in Transit. Environment and Planning C: Society and Space 18: 149-168. DeBresson, C. 1989. Breeding Innovative Clusters: A Source of Dynamic Development. World Development 17(1): 1-16. Dertouzos, M., Lester, R., Solow, R. and The MIT Cornrnision in Industrial Productivity 1989. Made in America: Regaining the Productive Edge Cambridge, MA: The MIT Press. Dicken, P. 1992. Global Shift: Transforming the World Economy London: Paul Chapman Publishing. Dodson, S. 2003. Where the Mobile is Killing the PC. The Guardian (April 17'4. Doel, M. and Rees, K. 2003. Negotiating Culture and Economy in High Technology Industries: the Governance of Knowledge in New Economic Spaces. Paper presented at the IGU Commission on the Dynamics of Industrial Space Annual Residential Conference, Vancouver, BC. Doershuk, B. and Milano, D. 1987. Inside Roland: Visionaries in Action. Keyboard (August): 14-16. Dore, R. 1973. British Factory, Japanese Factory: the Origins of National Diversity in Industrial Relations London: Allen & Unwin. Dosi, G. 1982. Technological Paradigms and Technological Trajectories: A Suggested Interpretation of the Determinants and Directions of Technical Change. Research Policy 1 1: 147-63. Dosi, G. 1983. Technological Paradigms and Technological Trajectories: The Determinants and Directions of Technical Change and the Transformation of the Economy. In C. Freeman (Ed.), Lon~Waves in the World Economy London: Buttenvorths (pp. 78-1 01). Dosi, G. 1988. Technical Change and Economic Theory London: Pinter Publishers. Edgington, D. 1993. The Globalization of Japanese Manufacturing Corporations. Growth and Change 24: 87-106. Edgington, D. 1999. Firms, Governments and Innovation in the Chukyo Region of Japan. Urban Studies 36(2): 305-339. Edgington, D. and Hayter, R., 2000. Foreign Direct Investment and the Flying Geese Model: Japanese Electronics Firms in Asia Pacific, Environment and Planning A 32(2): 281-304. Fandrich, D. 2001. A Piano Designers Notebook. Paper Presented at the First Annual BC Tone Wood Conference, Wells, BC, October 1,2000. Fransman, M. 1998. Visions of Innovation: the Firm and Japan. Oxford: Clarendon Press. Freeman C. and Perez, C. 1988. Structural Crisis of Adjustment, Business Cycles and Investment Behaviour. In Dosi et al. Technical Change and Economic Theory London: Pinter (38-66). Freeman, C. 1982. The Economics of Industrial Innovation Harmondsworth; Baltimore: Penguin. Freeman, C. 1987 Technology Policy and Economic Performance: Lessons from Japan London: Pinter Publishers Freeman, C. 1990. Technical Innovation in the World Chemical Industry and Changes of Techno-economic Paradigm. In C. Freeman and L. Soete (Eds.), New Explorations in the Economics of Technical Change London: Frances Pinter (pp. 116-134) Freeman, C. Clark, J. and Soete, L. 1982. Unemployment and Technical Innovation: a Study of Long Waves and Economic Development London: Frances Pinter. Freff 1988. Shuffle off to Buffalo: The Rise and Fall of Moog Music. Keyboard (May: 44-65. Fruin, M. 1997. Knowledge Works: Managing Intellectual Capital at Toshiba Oxford: OUP. Galbraith, J. 1967. The New Industrial State London: Hamish Hamilton. Gertler, M. 1995. "Being There": Proximity, Organization and Culture in the Development and Adoption of Advanced Manufacturing Technologies. Economic Geography 7 1: 1-26. Gertler, M. 2001. Best Practice? Geography, Learning and the Institutional Limits to Strong Convergence. Journal of Economic Geography 1: 5-26. Glasmeier, A. 2000. Manufacturing Time: Global Competition in the Watch Industry, 1795-2000 New York: The Guilford Press. Glasmeier, A. 1988. The Japanese Technopolis Programme: High Tech Development Policy or Industrial Policy in Disguise? International Journal of Urban and Regional Research 12: 268-84. Granovetter, M. 1985. Economic Action and Social Structure: The Problem of EOmbeddedness. American Journal of Sociology 9 1: 48 1-5 10. Granstrand, 0. 1999. The Economics and Management of Intellectual Property Cheltenham: Edward Elgar. Greenwald, T. 1987. It Came from the Music Industry Part 11: Mutants and Missing Links. Keyboard January: 82-98. Greenwald, T. and Burger, J. 1986. It Came from the Music Industry Part I: Dredging the Tar Pits of Technology." Keyboard December: 42-48. Gregory, D. 2000. "Diffusion" (Entry in R. Johnston, D. Gregory, G. Pratt and M. Watts eds.) The Dictionary of Human Geography Oxford: Blackwell (1 75- 178). Hagerstrand, T. 1968. Innovation Diffusion as a Spatial Process (translated by A. Pred) Chicago: University of Chicago Press. Hall, P. and Preston, P. 1988. The Carrier Wave: New Information Technology and the Geography of Innovation. 1846-2003 London: Unwin Hyman. Hamamatsu-shi 1997. Hamamatsu-Shi no Kyoshi Keikaku. [Hamamatsu's City Plan]. Mimeo. Harrington, J.W. 1985. Corporate Strategy, Business Strategy and Activity Location. Geoforum 16: 349-356. Harris, C. 2004. How Did Colonialism Dispossess? Comments from an Edge of Empire. Annals of the Association of American Geographers 94(1): 165-182. Harrison, B. Industrial Districts: Old Wine in New Bottles? Regional Studies 26(5): 469- 483. Hayter, R. 1976. Corporate Strategies and Industrial Change in the Canadian Forest Product Industries. Geographical Review 66: 209-28. Hayter, R. 1997 The Dynamics of Industrial Location: The Factory, the Firm and the Production System. Chichester:Wiley. Hayter, R. 2004. Economic Geography as Dissenting Institutionalism: the Embeddedness, Evolution and Differentiation of Regions. Geografiska Annaler 86B: 1-20. Hayter, R. and Edgington, D. 2004. Flying Geese in Asia: the Impacts of Japanese MNCs as a Source of Industrial Learning. Tiidschrifi voor Economic en Sociale Geografie 95: 3-26. Hayter, R. and LeHeron, R. 2002. Industrialization, Techno-Economic Paradigms and the Environment. In R. Hayter and R. LeHeron (Eds) Knowledge. Industry and Environment: institutions and innovation in territorial perspective Aldershot: Ashgate. Hayter, R. and Watts, H. 1983. The Geography of Enterprise: a Reappraisal. Progress in Human Geography 7: 157-8 1. Henry, N. and Pinch, S. 2000a. (The) Industrial Agglomeration (of Motor Sport Valley): A Knowledge, Space, Economy Approach. In Bryson, J., P. Daniels, N. Henry and J. Pollard (eds.) 2000. Knowledge, Space, Economy. London: Routledge (pp. 120-141). Henry, N. and Pinch, S. 2000b. Spatializing Knowledge: Placing the Knowledge Community of Motor-Sport Valley. Geoforum 3 l(2): 191-208. Hodgson, G. 2001. How Economics Forgot History London: Routledge. Holmen, H. 1995. What's New and What's Regional in the "New Regional Geography"? Geografiska Annaler 77B: 47-63. Howells, J. 2002. Tacit Knowledge, Innovation and Economic Geography. Urban Studies 39(5-6): 871-884. Hudson, R. 1999. The Learning Economy, the Learning Firm and the Learning Region: a Sympathetic Critique of the Limits to Learning. European Urban and Regional Studies 6:59-72. Hughes, T. 1992. The Dynamics of Technological Change: Salients, Critical Problems and Industrial Revolutions. In G. Dosi, R. Giannetti and P. Toninelli (eds.) Technology and Enterprise in Historical Perspective Clarendon Press: Oxford. Isa, K. and Tsuru, T. 2002. Cell Production and Workplace Innovation in Japan: Towards a New Model for Japanese Manufacturing. Industrial Relations 41(1): 548-78. Jaffe, A. 1989. Real Effects of Academic Research. American Economic Review 79: 957-970. Jaffe, A., Trajtenberg, M. and Henderson, R. 1993. Geographical Localization of Knowledge Spillovers as Evidenced by Patent Citations. The Quarterly Journal of Economics 108: 577-86. JETRO 1998. Local Business Climate Nurtures High-tech Ventures in Hamamatsu. Focus Japan July-August:8-9. JETRO 1999. Industrial Property Rights in Japan. Tokyo: JETRO Marketing Series. Jewkes, J., Sawers, D. and Stillerman, R. 1969. The Sources of Innovation New York: MacMillan. Johnson, C. 1982. MITI and the Japanese Miracle: The Growth of Industrial Policy, 1925-1975 Stanford, CA: Stanford University Press. Johnstone, B. 1994. Wave of the Future. Wired 2(3): 1-7. http://www.wired.codwiredlarchive/2.03/wave~ides.html?pg=2&to~ic Johnstone, B. 1999. We were Burning: Japanese Entrepreneurs and the Forging of the Electronic Age. New York: Basic Books. Jones, H. 1980. Live Machines: Hired Foreigners and Meiii Japan Vancouver: UBC Press. Kakehashi, I. 2002. I Believe in Music:Life Experiences and Thoughts on the Future of Electronic Music By the Founder of the Roland Corporation Milwaukee: Hal Leonard. Kawai. 1997. Sekui Ichi no Piano Tsukuri Mezashite [Aiming to make the world's best piano]. Hamamatsu: The Kawai Corporation. Keirstead, B. 1948. The Theory of Economic Change Toronto: MacMillan. Kelly, W. 1999. Downright upright: A history of the Canadian piano industry. Toronto: Natural Heritagematural History. Kelzenberg, D. 2002. 'American Theatre Organ Society: The Story of the Theatre Organ' http://www.atos.org/Pages/Story/innovations.htmlAccessed 6/7/03. Kenney, M. and Florida, R. 1988. Beyond Mass Production: Production and the Labor Process in Japan. Politics and Society 16(1): 121 - 158. Kioke, K. and Inoki, T. 1990. Skill formation in Japan and Southeast Asia Tokyo: University of Tokyo Press. Knox, P. and Agnew, J. 1989. The Geomaphy of the World Economy Harlow, Essex: Longman Scientific Technical. Kodama, F. 1995. Emerging Patterns of Innovation: Sources of Japan's Technological Edge Boston, MA: Harvard University Press. Kotler, M. and Hamilton, G. 1995. A Guide to Japan's Patent System Washington: Japan Information Access Project. Kuhn, T. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Latour, 1999b On Recalling ANT. J. Law and J. Hassard (eds.) Actor Network Theory and After Oxford: Blackwell (pp. 1- 14). Latour, B. 1987. Science in Action: How to Follow Scientists and Engineers Through Society. Cambridge, MA: Harvard University Press. Latour, B. 1999. Pandora's Hope: Essays on the Reality of Science Studies Cambridge, MA: Harvard University Press. Laulajainen, R. 1982. Temporal Hierarchy of the Corporate Space. GeoJournal6: 399- 408. Law, J. 1999. After ANT: Complexity, Naming and Topology. In J. Law and J. Hassard (eds.) Actor Network Theory and After Oxford: Blackwell (pp. 15-25). Law, J. and Hetherington, K. 2000. Materialities, Spatialities, Globalities. In Bryson, J., P. Daniels, N. Henry and J. Pollard (eds.) 2000. Knowledge, Space, Economy. London: Routledge (pp. 34-49). Lawson, C. and Lorenz, E. 1999. Collective Learning, Tacit Knowledge and Regional Innovation Capacity. Regional Studies 33: 305- 17. Lazonick, W. 1991. Business Organization and the Myth of the Market Economy New York: Cambridge University Press. Lee, R. 2000. Growth Pole. Entry in R. Johnston, D. Gregory, G. Pratt and M. Watts (eds.)The Dictionary of Human Geography Oxford: Blackwell. 326-327. Lee, R. 2002. 'Nice Maps, Shame about the Theory'? Thinking Geographically about the Economic. Progress in Human Geography 26(3): 333-355. LeHeron, R. 1980. Exports and Linkage Development in Manufacturing Firms: the Example of Export Promotion in New Zealand. Economic Geomaphv 56: 28 1-99. Liker, J., Fruin, M. and Adler, S. (eds.)1999. Remade in America: Transplanting and Transforming Japanese Management Systems Oxford: OUP. Lipsey, R. Bekar, C. and Carlaw, K. 1998. What Requires Explanation? In E. Helpman (Ed.) General Purpose Technoloaies and Economic Growth Cambridge, MA: The MIT Press (pgs. 15-54) Litterst, G. and Malambri, D. 1999. The Social Instrument. Piano and Keyboard NovIDec: 9-26. Loasby, B. 1999. Industrial Districts as Knowledge Communities. In Bellet, M. and L'Harmet, C. (eds.) Industry, Space and Competition: The Contribution of Economists of the Past. Cheltenham: Edward Elgar, (pp. 70-85). Loesser, A. 1954. Men, Women and Pianos: A Social History. New York: simon and Schuster. Loy, G. 1985 Musicians Make a Standard: The MIDI Phenomenon. Computer Muisc Journal 9(4): 8-26. Lundvall, B-A. 1992. National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning London: Pinter Publishers. Lundvall, B-A. 1998. The Learning Economy: Challenges to Economic Policy and Theory. In Neilsen, K. and Johnson, B. (eds) ~nstitutionsand ~conomicChange: New Perspectives on Markets, Firms and Technology Aldershot: Edward Elgar (pp. 33-55). Machlup, F. 1962. The Production and Diffusion of Knowledge in the United States. Princeton: Princeton University Press. Mair, A. 1997 'Strategic Localization: The Myth of the Postnational Enterprise'. In K R Cox (ed), Spaces of Globalization: Reasserting the Power of the Local, New York: The Guilford Press (pp. 64-88). Mair, A. Florida, R and Kenney, R. 1988. The New Geography of Automobile Production: Japanese Transplants in North America. Economic Geography 64(4): 352-73. Malecki, E. 1997. Technolo~yand Economic Development: The Dynamics of Local Regional and National Competitiveness Harlow, Essex: Longman. Malm, W. 1971. The Modern Music of Meiji Japan. In D. Shively (Ed.) Tradition and Modernization in Japanese Culture Princeton: Princeton University Press (pp. 96- 117). Malmberg, A. and Maskell, P. 2002. The Elusive Concept of Localization Economies: Towards a Knowledge-Based Theory of Spatial Clustering. Environment and Planning A 34: 429-49. Mansfield, E. 1988. Industrial Innovation in Japan and the United States. Science (30 September): 1769-1774. Markowitz, J. 1989. Triumphs & Trials of an Organ Builder Macungie, PA: Allen Organ Company Markusen, A. 1996. Sticky Places in Slippery Space: A Typology of Industrial Districts. Economic Geography 72: 293-3 13. Martin, D. 1969. Electronic Music: Audio 1988. JAES 17(4): 386-389. Maskell, P. 2001a. Growth and the Territorial Configuration of Economic Activity. Paper presented at the DRUID Summer Conference. http:llwww.druid.dkjconferences/nw/paperl/maskell.pdf; accessed l4/lO/O2. Maskell, P. 2001b. Towards a Knowledge-Based Theory of the Geographical Cluster. Industrial and Corporate Change 1O(4): 92 1-943. Maskell, P. and Malmberg, A. 1999. The Competitiveness of Firms and Regions. 'Ubiquification' and the Importance of Localized Learning. European Urban and Regional Studies 6(1): 9-25. Masser, I. 1990. Technology and Regional Development Policy: A Review of Japan's Technopolis Programme. Regional Studies 24(1): 4 1-53. May, W., Mason, C. and Pinch, S. 2001. Explaining Industrial Agglomeration: the Case of the British High Fidelity Industry. Geoforum 32: 363-376. McConnel, J. 1979. The Export Decision: an Empirical Study of Firm Behaviour. Economic Geography 55: 171-83 Mensch, G. 1979. Stalemate in Technology: Innovation Overcame the Depression Cambridge, MA: Ballinger. Milano, D. 1993. American Synthesizer Builders: Triumph and Crises for an Industry in Transition. In Vail, M. Keyboard Presents Vintage Synthesizers: Groundbreakinp Intsruments and Pioneering Designers of Electronic Music Synthesizers San Francisco: Miller Freeman Books (pg. 17-28) Ministry of Public Management, Home Affairs, Posts, and Telecommunications. 200 1. Joho Tsushin Hakusho (White Paper on Information and Communications in Japan) Mimeo. Mizuno, M. 2001. Kigyoukan Netowakku kara Umareru Inobessyon to Kyori: Jidousya Sangyou o Jirei to sum Tokkyo Detta no Chiriteki Bunseki [Technological Innovations, Inter-firm Networks and Distance: a Geographical Analysis of Patent Data in the Japanese Automobile Industry] (In Japanese). Jimbun Chiri 53(1): 18- 35. Mochida, Y, and Aoki, G. 1994. Gakki to Konpyu-ta [Musical Instruments and Computers] {In Japanese) Tokyo: Kyoritsu Shuppan. Morgan, K. 1997. The Learning Region: Institutions, Innovation and Regional Renewal. Regional Studies (3 1): 491 -503. Murata, K. 1980. The Role of Industrial Location Policy. In K. Murata and I. Ota (eds.) An Industrial Geography of Japan London: Bell & Hyman (1 77- 188). Murdoch, J. 1994. Sustainable Knowledge. Geoforum 25(2): 115- 132. Nakagawa, Y. 1984. Nippon Gakki no LSI Kaihatsu Senrvaku [Yamaha's LSI Development Strategy] (In Japanese). Tokyo : Diamond. National Research Council 1997 Maximizing US Interests in Science and Technology Relations with Japan (Committee on Japan Framework Statement and Report on Competitiveness Task Force Committee on Japan Office of Japan Affairs Office of International Affairs) Washington: National Academy Press. Nelson, R. and Winter, S. 1982. An Evolutionary Theory of Economic Change Cambridge, Mass: Belknap Press of Harvard University Press. Nelson, R., ed. 1993. National Innovation Systems: A Comparative Analysis New York: OUP. Nonaka, I. and Takeuchi, H. 1995. The Knowledge Creating Company: How Japanese Companies Create the Dynamics of Innovation New York: OUP. North, D. 1990. Institutions, Institutional Change and Economic Performance Cambridge: Cambridge University Press. Oda, H. 1992. Hamamatsu-Shi ni Okeru Kiki Kinzoku Kougyou no Ricchi Doutai. [Locational Dynamics of the Machinery Industries in the Hamamatsu Area, Shizuoka Prefecture] {In Japanese) Chirigaku Hyouron 65(11): 824-846. Odagiri ,A. and Goto, H. 1993 Innovation in Japan Oxford: Clarendon Press. Ohtsuka, M. 1980. Hamamatsu Chiiki ni Okeru Gakki Kougyou no Shuuseki. [Musical Instrument Industry in Hamamatsu] {In Japanese)Chirinaku Hyouron 53: 3 157- 70. Oinas, P. and Malecki, E. 2002. The Evolution of Technologies in Time and Space: From National and Regional to Spatial Innovation Systems. International Regional Science Review 25(1): 102-13 1. Paine, A. 1912. Mark Twain. a Biography: the Personal and Literary Life of Samuel Langhorne Clemens: with Letters, Comments and Incidental Writings Hitherto Unpublished: also New Episodes, Anecdotes, etc. New York: Harper. Park, SO. and Markusen, A. 1995. Generalizing New Industrial Districts: a Theoretical Agenda and an Application from a Non-Western Economy. Environment and Planning A 27: 81-104. Partner, S. 1999. Assembled in Japan: Electrical Goods and the Making of the Japanese Consumer Berkeley: University of California Press. Patchell, J. 1992. Shinchintaisha: Japanese Small Business Revitalization. Business: The Contemporary World 4: 50-6 1. Patchell, J. 1993. From Production Systems to Learning Systems: Lessons from Japan. Environment and Planning A 25(7):797-8 15. Patchell, J. 1996 Kaleidoscope Economies: the Processes of Cooperation, Competition and Control. Annals of the Association of American Geomapers 86: 48 1-506. Patchell, J. and Hayter, R. 1995. Skill Formation and Japanese Production Systems. Tiidschrift voor Economische en Sociale Geografie 86(4): 339-356. Pavitt, K. 1986. Chips and Trajectories: how does the Semiconductor Influence the Sources and Directions of Technological Change? In R. Macleod (Ed.), Technology and the Human Prospect London: Frances Pinter. Pedlar, N. 1990. The Imported Pioneers:Westerners who Helped Build Modem Japan. Folkstone, Kent: Japan LibraryLtd. Perez, C. 1983. Structural Change and Assimilation of New Technologies in the Economic and Social Systems. Futures (Oct: 357-375). Perez, C. 1985. Microelectronics, Long Waves and World Structural Change. World Development 13 : 44 1-63. Perroux, F. 1956. Travail et Civilization. Esprit 243:48 1-497. Phillips, A. 1971. Technology and Market Structure: a Study of the Air-Craft Industry. Lexington, MA: Heath Lexington. Piatier, A. 1988. Transectorial Innovations and the Transformation of Firms. The information society 5(4): 205-3 1. Pinch, S. and Henry, N. 1999. Discursive Aspects of Technological Innovation: The Case of the British Motor-Sport Industry. Environment and Planning A 3 l(4): 665-682. Pinch, T. and Trocco, F. 2002. Analog Days: The Invention and Impact of the Moog Synthesizer Cambridge, MA: Harvard University Press. Polanyi, M. 1967 The Tacit Dimension New York: Anchor Books. Porter, M. 1990. The Competitive Advantage of Nations New York: Free Press. Poschardt, U. 1998. DJ Culture (Translated by Sham Whiteside) London: Quartet Books. Pred, A. 1966. The Spatial Implications of US Industrial Growth Massachusetts: MIT Press. Reiffenstein, T., R. Hayter and D.W. Edgington. 2002. Crossing Cultures, Learning to Export: Making Houses in British Columbia for Consumption in Japan," Economic Geography 78 (2): 195-2 19. Rigby, D. 2000. Geography and Technological Change. In E. Sheppard and T. Barnes (eds.) A Companion to Economic Geography Oxford: Blackwell. (pp. 202-224). Ripin, E., Caldwell, J. and O'Brian, G. 1988 Piano (The New Grove Musical Instrument Series) New York: Grove Publications. Roland Corporation. 1999. Roland Corporate Guide. Hamamatsu Rosen, D. and Usui, C. 1994. The Social Structure of Japanese Intellectual Property Law. UCLA Pacific Basin Law Journal 13(1): 32-69. Rosenberg, N. 1982. Inside the Black Box: Technologv and Economics Cambridge: Cambridge University Press. Rosenberg, N. 1976. Perspectives on Technologv Cambridge: Cambridge University Press. Rosenberg, N. 1992. Science and Technology in the Twentieth Century. In G. Dosi, R. Giannetti and P.A. Toninelli (Eds) Technology and Enterprise in a Historical Perspective Oxford: Clarendon Press (pp. 63-96). Rosenberg, N. 2000. Schurnpeter and the Endogeneitv of Technology: Some American Perspectives London: Routledge. Rosenberg, N. Ince, P., Skog, K. and Platinga, A. 1994. Understanding the Adoption of New Technology in the Forest Products Industry. In N. Rosenberg (ed.) Exploring the Black Box: Technology, Economics and Histow Cambridge: Cambridge University Press (pp.232-249). Sargent. J. 1980. Industrial Location in Japan since 1945. Geoiournal4(3): 205-2 14. Saxenian. A-L. 1994. Regional Advantage: Culture and Competition in Silicon Valley and Route 128 Cambridge, Mass: Harvard University Press. Schmookler, J. 1966. Invention and Economic Growth Harvard: Harvard University Press. Schoenberger, E. 1997. The Cultural Crisis of the Firm Cambridge, Mass: Blackwell. Schoenberger, E. 2001. Corporate Autobiographies: the Narrative Strategies of Corporate Strategists. Journal of Economic Geography l(3): 277-298. Scott, A. 1999. The U.S. Recorded Music Industry: On the Relations Between Organization, Location and Creativity in the Cultural Economy. Environment and Planning A 3l(ll): 1965-1984. Shaefer, R. 1971. New Techniques for Electronic Organ Tone Generation. Jounal of the Audio Engineering Society 19(7): 570-575 Shapira 1995. The R&D Workers: Managing Innovation in Britain, Germav, Japan and the United States Westport, CN: Quorum Books. Sheard, P. 1983. Auto-Production Systems in Japan: Organization and Locational Features. Australian Geographical Studies 21 :49-68. Shizuoka Keizai Shinbun 2002. Rorando Hosoe Koujyo: Seru Houshiki de Shou Ryou Tahinshu: IT ka to Ginou Kojo Ryourinni [Roland's Hosoe Factory: the Small Quantity, High Diversity Cell Production Method: Improvements Based on the Twin Wheels of Information Technology and (Human) Technique] (2616102: pg 37). Steed, G. 1982. Threshold Firms: back in^ Canada's winners Science Council of Canada, Background Study 48, Ottawa: Minister of Supply and Services. Sternberg, R. 1995. Supporting Peripheral Economies or Industrial Policy in Favour of National Growth? An Empirical Analysis of Goal Achievement of the Japanese Technopolis Program. Environment and Planning C: Government and Policy 13: 425-39.) Stirk, N. 200 1. Intellectual Property and the Role of Manufacturers: Definitions from the Late Eighteenth Century. Journal of Historical Geography 27(4): 475-492. Storper, M. 1997. The Regional World: Territorial Development in a Global Economy New York: The Guildford Press. Storper, M. and Scott A. 1992. Pathways to Industrialization and Regional Development London: Routledge. Storper, M. and Walker, R. 1989. The Capitalist Imperative: Territory, Technology and Industrial Growth New York: Blackwell. Tachibanaki, T. 1998. Who Runs Japanese Business? Management and Motivation in the Cheltenham: Elgar. Takahashi, Y. 2000. A Network of Tinkerers: The Advent of the Radio and Television Receiver Industry in Japan. Technolonv and Culture 41 : 460-484. Takenaka, T. 1994. The Role of the Japanese Patent System in Japanese Industry. UCLA Pacific Basin Law Journal l3(l): 25-3 1. Takeo, H. 1988. Hamamatsu ni Okeru Gakki Kougyou no Henyou [Metamorphosis of Musical Instrument Industry in Hamamatsu] (InJapanese) : Chirigaku Houkoku 67: 55-66. Takeuchi, A. 1992. Activities of Small Scale Industries in Japan Through Inter-Enterprise Cooperation. Report of Research - Nippon Institute of Technology 21 : 283-92. Takeuchi, A. 1996. Koumou Chiiki no Hendou [The Dynamics of Industrial Regions] (InJapanese) Tokyo: Taimeidou Hakkou. Takeuchi, H. 2002. Hamamatsu Kigyou Tsuyousa no Himitsu [The Secret of Hamamatsu's Industrial Strength] (InJapanese). Tokyo: Toyou Keizai Shipannsya. Tamaki, N. 2001. Yukischi Fukuzawa, 1835-1901 : The Spirit of Enterprise in Modern Japan. Basingstroke, Hampshire: Palgrave. Taylor, M. and Thrift, N. 1983. Business Organization, Segmentation and Location. Regional Studies 17: 445-65. The Music Trades 1982. 'Portable Electronic Keyboards: The New Key to Expanding the Market'. September Edition (pp.38-44) The Music Trades 1985. 'Casio Unveils Synthesizer Line'. December Edition (pp.60-61). The Music Trades 2000. 'The Global 225: The World's Largest Music Products Suppliers'. December Edition (pp.80- 102). Theberge, P. 1997. Any Sound You Can Imagine: Making Music/Consuming Technology Hanover, NH: Wesleyan University Press. Thrift, N. 2000. "Actor-network Theory" (Entry in R. Johnston, D. Gregory, G. Pratt and M. Watts eds.) The Dictionary of Human Geography Oxford: Blackwell, pp.4-6) Tillman, R. 1993. A Veblen Treasury: From Leisure Class to War, Peace and Capitalism. Aromon, NY: M.E. Sharpe. Toyota 2001. Toyota Commemorative Museum of Industry and Technology Nagoya: The Toyota Corporation. Tsukahara,Y. 1994 Tekunoporis Seisaku Hyouka. [An Evaluation of Technopolis Policy] {In Japanese) Keizai Chirinaku Nenpo 40(3): 56-64. Ullman, E. 1980. Geography as Spatial Interaction Seattle: University of Washington Press. US International Trade Commission 1999. 'Pianos: Economic and Competitive Conditions Affecting US Industry' Investigation No.332-401 .Washington, DC. - available online at: ftp://ftp.usitc.gov/pub/reports/studies/UB3 196.PDF Vail, M. 1993. Vintage Synthesizers: Groundbreaking Instruments and Pioneering Designers of Electronic Music synthesizers San Francisco: GPI Books. Vail, M. 1997. The Hammond Organ: Beauty in the B. San Francisco: Miller Freeman. Van Duijn, J. 1983. Fluctuations in Innovations over Time. In C. Freeman (Ed.) Long Waves in the World Economy London: Buttenvorths (pgs 19-31) Veblen, T. 1904. The Theory of the Business Enternrise New York: Charles Scribner's Sons. Vernon, R. 1971. The Product Cycle Hypothesis in a New International Environment. Oxford Bulletin of Economics and Statistics 255-67. Voehl, F. 1995. Deminn: The Way We Knew Him Delray Beach, F1: St. Lucie Press. Von Hippel, E. 1988. The Sources of Innovation Oxford: OUP. Waring, D. 2002. Manufacturing the Muse: Estey Organs & Consumer Culture in Victorian England Hanover: Wesleyan University Press. Weber, M. 1978. Selections in Translation W. Runciman (Ed.), translated by E. Matthews. Cambridge: Cambridge University Press. Weidenaar, R. 1995. Magic Music from the Telharmonium Metuchen, NJ: The Scarecrow Press Inc. Wenger, E. 1998. Communities of Practice: Learning. Meaning and Identity. Cambridge: Cambridge University Press. Westfield Center, 2001. Curious Facts from the Organ's History. http://www.westfield.org/curious.htmAccessed 7/6/03. Westney, D. 1987. Imitation and Innovation: The Transfer of Western Organizational Patterns to Meiii Japan Cambridge, MA: Harvard University Press. Winder, G. 1995. Before the Corporation and Mass Production: The Licensing Regime in the Manufacture of North American Harvesting Machinery, 1830-19 10. Annals of the Association of American Geographers 85:52 1-52 Winder, G. 2001. Building Trust and Managing Business over a Distance: A Geography of Reaper Manufacturer D.S. Morgan's Correspondence, 1867. Economic Geography 77(2): 95-121. Yamaha 1987. Yamaha 100 Nenshi Hamamatsu: Yarnaha. Yamaha, 2001. Yamaha and Roland Agree to Cooperate to Improve MIDI Data Compatibility. Press Release, Buena Park California January 19th. Yamanaka, K. 2002. Anna Bortz's Lawsuit and Minority Rights in Japan. Japan Policy Research Institute: Working Paper (No.88 - September). Available on-line at: http:llwww.i pri.ordpublications/workin~ml.Accessed 22/05/04. Yamashita, H. 1990. Hamamatsu-Suzuka ni okeru Nidousha Kougyou no Chiikiteki hatten. [The Regional Development of the Motorcycle Industry in the Hamamatsu and Suzuka area] {In Japanese) Chirigaku Houkoku 7 1: 9- 19. Yeung, H. 2000. Organizing 'The Firm' in Industrial Geography I: Networks, Institutions and Regional Development. Progress in Human Geography 24(2): 30 1-315. Yeung, H. 2001. Regulating 'The Firm' and Sociocultural Practices in Industrial Geography 11. Progress in Human Geography 25(2): 293-302. Yeung, H. 2002. Producing 'The Firm' in Industrial Geography 111: Industrial Restructuring and Labour Markets. Pro~essin Human Geography 26(3): 366- 378. Zucker, L. abnd Darby, M. 1996. Star Scientists and Institutional Transformation: Patterns of Invention and Innovation in the Formation of the Biotechnology Industry. Proceedings of the National Academy of Science, 93: 12709-127 16. APPENDICES

Appendix A: List Of Interviews

Title Firm Location1 Date - - Matsubayashi Koji Asst. General Manager Yamaha Hamamatsu, Planning Division Corporation 4ugust 27,2001 Musical Instruments Group rKobayashi Kouichi Section Manager Yamaha Hamamatsu, IE Section Corporation 4ugust 27,2001 Product Engineering Group Furusawa Masahiro Supervisor Yamaha Hamamatsu, kPersonnel Division Corporation August 27,2001 Miyazaki Kenji Deputy Division Manager Yamaha Hamamatsu, Musical Instrument Business Corporation 4ugust 27,2001 Unit ISubcontract Control Division Komori Michihiko Human Resources Yamaha Hamamatsu, Development Section Corporation July 21, 2002, Section Head January 15,2003 Gompei Masashi Human Resources Yamaha Hamamatsu, Development Section Corporation July 21, 2002, January 15,2003 rNagahama Yasuo PA-DM1 (Professional Audio- Yamaha Hamamatsu, Digital Musical Instruments) Corporation July 21, 2002 Section Electone Division Head - PA-DM1 (Professional Audio- Yamaha Hamamatsu, Digital Musical Instruments) Corporation January 15,2003 Section Product Development Department Technology Development Group Music Software Development Manager - I Koga lppei International Division Kawai Hamamatsu, Product Planning and Import Musical August 28,2001 Corporation Name Title I Firm Location1 Date Manager Kamino Kennichi Personnel Division Kawai Hamamatsu, Public Relations Group Corporation August 28,2001 Manager ltoh Hiroshi Personnel Division Kawai Hamamatsu, Public Relations Group Corporation August 28,2001 Section Head Saitoh Yasuhiro Piano Manufacturing Kawai Hamamatsu, Department Corporation August 28,2001 Hiranabe Yoshiaki International Division Kawai Hamamatsu, Marketing and Corporation July 22, 2002 Merchandisingllmport Asst. Manager Kondo Yoichi R&D. Electronic Musical Kawai Hamamatsu, lnstruments Division Corporation July 22, 2002 Assistant Manager Saito Tsutomu R&D. Electronic Musical Kawai Hamamatsu, lnstruments Division Corporation July 22, 2002 Manager ~- Takaba Tsutomu R&D. Electronic Musical Kawai Harnamatsu, lnstruments Division Corporation July 22, 2002 Chief Engineer Nakata Haruaki Adviser Roland Hamamatsu Corporation August 29,2001 - -- Kikumoto Tadao Senior Managing Director Roland Email Interview Corporation December 15,2002 - --- Hoshiai Atsushi R&D Engineer Roland Email Interview Corporation January 17,2003 - Sakashita Akira Global Marketing Headquarters Casio Tokyo Consumer and Educational Computer January 20,2002 Products Department Corporation EMP Marketing Section Yuzawa Kiminori Public Relations Department Casio Tokyo Computer January 20,2002 Corporation Dave Smith President Dave Smith St. Helena, CA Music Inc. October 11,2002 Ralph Deutsch Inventor, Retired Sherman Oaks, CA October 12, 2002 Appendix B: Ethics Approval

SIMON FRASER UNIVERSITY

OFFICE OF RESEARCH ETHICS RURNARY, RRmSH COLUMBIA ROOM 2105 SIRAND HALL CANADA VSA IS6 Telephone 604-291 -3Mi FAX W-268-h7R5

Mr. Tim Reiffenstein Graduate Student Department of Geography Simon Fraser University

Dear Mr. Reiffenstein:

Re: Transectorial innovation, location dynamics and knowledge formation in the Japanese electronic musical instrument industry Title Change

Your application on July 22,2004, for a title change from, The Japanese piano/kcyboard production system: forms of learning and the echo of modernity, has been categorized as Minimal Risk and approved by the Director, Office of Research Ethics, on behalf of the Research Ethics Board in accordance with REB policy.

Sincerely,

Dr. Hal weinbe\rg, Director Office of Research Ethics

c: Dr. Roger Hayter, Supervisor