Document Title
Introduction
W. Brian Arthur is an economist and complexity scientist, serving as an external research faculty member at the Santa Fe Institute, renowned for its focus on complexity studies. In 1990, he won the Schumpeter Prize for his groundbreaking paper on positive feedback mechanisms in the economy, a recognition given for exceptional contributions to research in innovation.
In The Nature of Technology: What It Is and How It Evolves, Arthur continues a Schumpeterian exploration of innovation. Although published in 2009 and somewhat dated, the book remains relevant because innovation is still not fully understood by policymakers. Moreover, Arthur's insights hold enduring value, as evidenced by Giovanni Dosi, the 2024 Schumpeter Prize winner, who extensively references Arthur’s work in his recent book, The Foundations of Complex Evolving Economies. Dosi uses Arthur’s theory of combinatorial evolution and historical examples of innovation to enrich his analysis.
Arthur identifies three perspectives on technology: first, as a means to fulfill human purposes; second, as an assemblage of practices and components, representing technologies in the individual sense; and third, as a collective system encompassing all technologies.
Arthur’s primary goal in this book is to provide a theory of how technology evolves, tracing all technologies to earlier ones and detailing the mechanisms underlying this progression. Previous thinkers, such as economist Joseph Schumpeter, historian Abbott Payson Usher, and sociologist S. Colum Gilfillan, hinted at the combinatorial nature of technological origins. While they suggested that novel technologies arise by combining existing ones, they stopped short of offering a comprehensive framework for understanding how the entire body of technology builds upon itself.
Arthur also emphasises the significance of the existing stock of technologies as the basis for further combinations—a concept that traces back to sociologist William Fielding Ogburn. However, Ogburn did not develop this into a full theory of technological evolution. Arthur unites these ideas, asserting that novel technologies arise from existing ones, and that technologies continuously give rise to further innovations.
“… [O]ver time, many technologies form from an initial few, and more complex ones form using simpler ones as components. The overall collection of technologies bootstraps itself upward from the few to the many and from the simple to the complex. We can say technology creates itself out of itself.”
The theory Arthur presents is compelling, and his detailed explanation of its logic, along with his exploration of the general features of technology and the innovation process (chapters 2 to 8), is particularly valuable for policymakers.
The book is broadly divided into three parts. In the first half, Arthur lays the groundwork for his theory by discussing the fundamental elements of technology. He begins with its definition, the shared principles underlying all technologies, the essence of technology itself, the natural emergence of technological groupings or domains, and how these groups resemble languages. These foundational topics are addressed in chapters 2, 3, and 4. With the basics established, he shifts focus in chapters 5 through 8 to the process of innovation. These chapters cover critical topics such as the role of engineering in problem-solving, the origins of technologies, the typical paths through which technologies improve, the nature of technological revolutions, the emergence of new technological domains, and their effects on existing technologies and organisations. Finally, in chapter 9, Arthur brings all these components together to present his theory of how technologies evolve. In chapter 10, he discusses the impact on the economy as technologies evolve and he concludes in chapter 11.
The book’s structure provides clarity and progression, culminating in an "aha" moment when the reader recognizes how its arrangement mirrors the characteristics of scientific theories— purposeful constructs made up of interdependent components which he discusses in the book. Arthur’s deliberate approach, beginning with fundamental elements and processes before introducing his theory of technological evolution, reflects his commitment to practicing what he preaches. While some may feel he takes too long to reach the main theory, the book’s organisation effectively demonstrates how each step builds toward the central idea.
Notable Concepts and Ideas
Opening The Black Box of Technology
What is interesting about the book is the novelty of the ideas you encounter about technology especially how it contrasts with the ideas of mainstream economics. Mainstream economics mainly treats technology as a “black box”. Although it assumes technological change happens and studies its products like productivity or economic growth it does not dive into the internal process of technological development. Mainstream economics often treat it as an exogenous variable, focuses on aggregate measures such as TFP without looking at the processes through which they are developed, and pays little attention to the processes happening at the micro-level. This book will provide much of the needed remedy for those who are only familiar with the mainstream approach with its detailed descriptions of the essence of technology and the innovation process. People who are familiar with evolutionary economics and have read some papers of Nathan Rosenberg would find the discussions familiar. Although you would still benefit the author’s book owing to the the new theory, the detailed discussions on the innovation process, and the myriad of historical examples he provides to build his theory.
The Three Principles of Technology and Science’s Reliance on Technology
Even if you have only a vague understanding of technology, his discussion of its essence is intriguing. Technologies are combinations, recursive in nature, and exploit natural phenomena. They arise from the combination of existing technologies, forming a hierarchy where each level comprises smaller components. This structure means that improving a main assembly (the part that carries out the base function) often requires changes to subassemblies (parts that support the main assembly), as initial improvements can disrupt the system's balance. Adjustments must be made across all levels to resolve this imbalance.
The reliance on natural phenomena also highlights the role of modern science. As phenomena become less apparent, science is needed to uncover them, and technology is required to probe and harness them. This reciprocal relationship shows that scientific discovery depends on technology, just as technological development relies on science. Thus, technology is not merely applied science but a self-creating (autopoietic) process, with humans as intermediaries.
Technology Share Traits with Other Purposed Systems
Technologies also share similarities with social organisations, scientific explanations, mathematics, and languages. As “purposed systems,” technologies resemble social structures like businesses, legal systems, or monetary systems, though conventional technologies rely on physical effects, while social systems depend on human behavior. Similarly, scientific explanations and mathematical systems are constructed from simpler parts that combine according to established rules, reflecting the compositional nature of technology.
One interesting idea was his comparison of technological design to an expression within a language. Technologies can be grouped into domains because they share common features, making such clustering logical. A new device is assembled from components within a domain, with the domain acting as a specific "language" and its components as the vocabulary. This analogy makes technology a form of composition, similar to language, with varying degrees of articulateness, appropriateness, conciseness, and complexity. Like language, technology can be simple or intricate, akin to an entire book with a central theme and supporting subthemes. Additionally, like language, technology operates within rules—its own "grammar"—that dictate allowable combinations.
Notable Ideas Useful for Policymaking
Apart from offering a unique perspective on technology, the book provides deeply insightful ideas that extend beyond the theory of combinatorial evolution, which itself is valuable for researchers and policymakers. The real strength lies in the foundational discussion of the building blocks of this theory, offering a basic analytical framework that can inform innovation and industrial policy.
Understanding the fundamental elements of technology and the processes through which it develops allows us to better identify gaps in the prerequisites needed to create our own technologies. The components and theories of technological change outlined in the book clarify how proposed policies may influence the development of desired technologies across different levels. They also provide insights into whether these policies create an environment conducive to technological innovation.
The Role of Standard Engineering and Mental Association
The book also explores the process of innovation, emphasising the importance of creating favourable conditions for it. One key process is standard engineering, described as problemsolving that typically starts with design—a conceptual structure to achieve a purpose. This process progresses through three stages: conceptual design, detailed design of assemblies and parts, and manufacturing with necessary feedback loops.
Invention, on the other hand, requires familiarity with base principles, or the "methods of the thing," which exploit natural phenomena for human purposes. Inventors often identify these principles through mental association, linking achievable actions and deliverable effects to arrive at solutions. While the process may appear brilliant, it fundamentally resembles problem-solving in everyday life. For example, finding alternative transportation when a car is in the repair shop involves selecting and combining everyday functionalities. Similarly, inventors navigate familiar territory within their expertise to resolve problems.
These processes underscore the importance of sustaining active engineering practices across technological domains within a country. Without ongoing problem-solving and invention, engineers cannot gain the experience needed to form strong mental associations between functionalities and solutions. Exposure to diverse engineering fields and technological domains is also crucial, as it fosters familiarity with knowledge bases and facilitates the borrowing of base principles from other domains. Furthermore, translating scientific discoveries into practical engineering applications requires robust engineering faculties and enhanced collaboration between academia and industry, ensuring industries can quickly leverage scientific advancements.
Different Types of Innovation
Arthur also explores various types of innovation, which can inform policy design to encourage their development. After a technology is invented, it often evolves to overcome performance limitations. One method is internal replacement, where components limiting performance are replaced with better alternatives. Another is structural deepening, which addresses bottlenecks by adding components or assemblies to work around constraints, making the technology more complex. These approaches also apply to lower levels of the technological hierarchy, as rebalancing may require modifications in subassemblies, which could face their own limitations.
Radical innovation, or the creation of novel technologies, involves applying new base principles to meet specific needs. These needs often exist for some time, with practitioners aware of them but unable to find evident solutions using existing technologies. When new base principles emerge—often through mental association and borrowing from existing concepts—they are embodied in physical form and developed for commercial use. Such innovations may eventually enter the economy, transforming it. Since base principles cannot be invented from nothing, crossdomain interaction is essential.
A fourth type of innovation occurs when entire bodies of technology emerge, forming a new domain. These domains often originate from established fields and evolve over time, transforming industries that adopt them. For instance, the banking sector’s encounter with computation led to digitised accounting, blending bookkeeping and computational functionalities to create new capabilities.
The prerequisites for these types of innovation differ. Internal replacement and structural deepening rely heavily on standard engineering and the iterative practices of engineering-based industries. These industries develop capabilities over time through feedback in design and manufacturing. In contrast, radical innovation and the emergence of new technological domains depend on diversified industries. Cross-domain interaction among engineers fosters the adoption of base principles from different fields, while diverse industries enable the creation of new functionalities through technology adoption.
These insights underscore the need for policies that sustain industrial growth and prevent deindustrialisation. A strong and varied industrial base supports all forms of innovation and ensures a fertile ground for technological advancement.
Technology Maturity and Technology Cycle
Chapter 7 of Arthur’s work introduces the concept of technology maturity. Technologies, after successive rounds of internal replacement and structural deepening, eventually reach a performance limit, signifying maturity. This concept challenges assumptions about the suitability of short-cycle technologies for latecomer countries. The cycle time refers to the speed by which technologies change or become obsolete over time, and the speed and frequency at which new technologies emerge. Keun Lee (a 2014 Schumpeter Prize winner) suggests that technologies with short cycle times may offer lower entry barriers due to rapid technological obsolescence, reduced royalty payments, and opportunities for first-mover advantages. However, technologies nearing maturity, even with short cycles, may face dead ends in progress, reducing their potential for catch-up industries. For instance, as Moore’s Law approaches its limits in semiconductors, it raises questions about semiconductors as an industry for latecomer catch-up even though at present, it is still considered a short-cycle technology. Developing countries must, therefore, evaluate whether entering such industries for economic catching up is worthwhile and consider the timeline before technological maturity sets in.
Informing the Concept of Technology Cycle Time
Keun Lee’s insight, however, is not very clear regarding factors that drive the cycle time of technology. Some comments he provides is that inter- and intra-industry investments can increase chances of getting involved in short-cycle technologies. He also states that the cycle time can be endogenous if the firm introduces and improves the technology itself or exogenous if some critical component that firms depend on that may reset the industry is provided by an external firm. The case of Google introducing the Android Operating System (OS) can be seen from both the endogenous and exogenous perspectives. The OS which superseded the self-developed OS by Nokia and BlackBerry is endogenous from the perspective of Google since they developed it. From the perspective of mobile phone makers that started using Google’s OS, the technology is external. Thus, Google does each cycle of improvement, making the cycle time exogenous from the perspective of mobile phone makers. However, Lee does not reveal much about the influence of competition. Arthur’s perspective enriches this discussion, emphasising the role of economics and competition in shaping the pace of technological progress. When competitive pressure is intense, advancements accelerate; when absent, progress slows. Even when a technical path is clear, firms may delay improvements until it is economically viable. This nuanced view explains how economic and competitive factors influence a technology’s cycle time.
Critical Reflections
Questioning the Role of “Deep Craft”
Arthur attributes the sustained technological leadership of regions to “deep craft”—a blend of knowledge, cultural practices, and local experimentation. He argues that deep craft is challenging to transfer and underpins long-term technological dominance. However, this perspective can be questioned. The post-WTO era has seen developed countries restrict policy tools previously used by latecomers, a phenomenon often referred to as “kicking away the ladder”. Intellectual property regimes and constrained policy spaces may contribute more to incumbency than deep craft.
There are also examples from Taiwan and South Korea that illustrate latecomers can surpass incumbents which hints at the insufficiency of deep craft to sustain their position. These countries were able to catch up because they “industrialised by learning”. They prioritised continuous learning and adaptation, introduced strategic government intervention, and built national champions. They ensured that firms had absorptive capacity to learn and adapt foreign technologies, invested in R&D and technological development, and built strong linkages between firms, universities and research institutions. These countries have emerged as leaders in chip fabrication and memory technologies, industries once dominated by Western firms. This challenges the notion that deep craft inherently prevents catch-up.
Basic Science vs. Applied Science for Catch-Up Economies
Arthur advocates for developing countries to invest in basic science to expand their innovative capabilities. However, the experiences of successful catch-up economies like Korea and Taiwan suggest otherwise. Rather than prioritising basic science, these nations focused on technological development within private industrial sectors. Domestic firms invested in R&D after assimilating foreign technologies, have increased demand for applied science and has enabled the coevolution of the industrial sector with the science sector.
Studies by Kim and Lee indicate that technological knowledge (e.g., patents) is more directly linked to economic growth in latecomer countries than scientific knowledge (e.g., academic publications). In contrast, in regions like Latin America, where industrial sectors are weak, research priorities often follow international scientific trends, which is to prioritise basic science, with limited alignment to local industrial needs. This results in an R&D system favoring scientific advancements over technological achievements, hindering support for industrial growth and technological development, contrary to Arthur’s claim that investment in basic science would lead to innovative capabilities in developing countries.
Summary of The Theory
Arthur’s theory of technology is built on three key propositions: all technologies are combinations of components; these components are themselves technologies; and all technologies harness phenomena for specific purposes. The third proposition emphasizes that technology fundamentally programs nature by orchestrating phenomena to achieve human goals. Each individual technology integrates multiple phenomena, directing them toward a specific purpose.
Once new technologies emerge, they become building blocks for further innovations, leading to what Arthur describes as combinatorial evolution. Unlike Darwinian evolution, this process relies on combining existing technologies to create novel ones, forming a self-sustaining cycle. Progress in technology depends on the ability to capture and harness new phenomena, which, in turn, requires existing technologies. This cyclical relationship means that technology evolves out of itself, advancing from simple components to increasingly complex systems.
Arthur identifies the central mechanism of technological evolution as the linking of needs with effects that can fulfill them. This process begins with envisioning a concept—an idea of how specific effects can work together—and finding a combination of components to realize it. The process is recursive: solving one problem often generates new subproblems, requiring iterative refinement at multiple levels. At the core of this evolution is combination—assembling suitable parts or functionalities to form solutions. However, the process is also driven by needs, which frequently arise from the limitations or problems of existing technologies rather than direct human desires. Consequently, technological solutions often create new needs, fostering further innovations. This dynamic interplay makes combinatorial evolution as much about the emergence of needs as it is about their resolution.
Technological evolution is neither uniform nor smooth. At any moment, technologies are being introduced, modified, and discarded, creating opportunities for new innovations and revealing previously unexplored phenomena. Individual technologies evolve through internal refinements, such as upgrading components or adding subassemblies, to enhance performance. Meanwhile, bodies of technology evolve collectively, altering their "vocabularies" and integrating into industries. This process of constant transformation generates perpetual novelty, with technologies interacting like a metabolic system, building on existing entities to create new ones and generating further needs. The process is organic, with new layers forming atop old ones, and creations overlapping in time.
The economy plays a critical role in driving and mediating this evolution. It identifies needs, tests ideas for commercial viability, and demands improvements. However, the economy is not merely a passive recipient of technological upgrades—it is an expression of technology. Its structure consists of interdependent arrangements, such as businesses, production systems, and institutions, all of which can be considered technologies in a broad sense. These arrangements create opportunities for further innovations, shaping structural changes within the economy. As a result, the economy inherits the dynamic qualities of its technologies: it is open-ended, historydependent, hierarchical, and ever-changing. Like technology, it evolves continuously, teeming with change and novelty over the long term.
Conclusion
To conclude, Arthur’s The Nature of Technology is an engaging and insightful read for anyone interested in understanding technology, particularly policymakers. Its in-depth exploration of the fundamental elements of technology and the innovation process offers a valuable framework for analyzing how policies shape technological development. Despite some debatable views, such as the role of deep science for developing countries and the persistence of incumbents, the book remains an exceptional contribution to the theory of technological evolution and a benchmark in the field.
References
- Andreoni, Antonio, Ha-Joon Chang, and Isabel Estevez. 2019. “4. New Global Rules, Policy Space, and Quality of Growth in Africa.” In The Quality of Growth in Africa, edited by Akbar Noman, Joseph E. Stiglitz, and Ravi Kanbur, 111–45. Columbia University Press. https://doi.org/10.7312/kanb19476-006.
- Arthur, W. Brian. 2009. The Nature of Technology: What It Is and How It Evolves. New York: Simon and Schuster.
- Chang, Ha-Joon. 2003. Kicking Away the Ladder: Development Strategy in Historical Perspective. London: Anthem Press.
- Dosi, Giovanni. 2023. The Foundations of Complex Evolving Economies: Part One: Innovation, Organization, and Industrial Dynamics. Oxford University Press. https://books.google.com/books?hl.
- Kim, Yee Kyoung, and Keun Lee. 2015. “Different Impacts of Scientific and Technological Knowledge on Economic Growth: Contrasting Science and Technology Policy in East Asia and Latin America.” Asian Economic Policy Review 10 (1):43–66. https://doi.org/10.1111/aepr.12081.
- Lee, Keun. 2013. Schumpeterian Analysis of Economic Catch-up: Knowledge, Path-Creation, and the Middle-Income Trap. Cambridge University Press. https://books.google.com/books?hl.
- Mathews, John A., and Dong-Sung Cho. 2000. Tiger Technology: The Creation of a Semiconductor Industry in East Asia. Vol. 389. Cambridge University Press Cambridge.
- Reinert, Erik S. 2008. How Rich Countries Got Rich ... and Why Poor Countries Stay Poor. PublicAffairs.
