“Considering the Computing・Algorithmic Environments of our Lives and Society - from the viewpoints of history and philosophy of computing”


私たちの⽣活と社会を考える ―― Computing の科学・技術史学と科学・技術哲学による視点から

WORKSHOP Co-Organized by

The Japan Association for Philosophy of Science (科学基礎論学会)

The History of Science Society of Japan (日本科学史学会)

HaPoC (IUHPST/DHST-DLMPST InterDivision Commission for History and Philosophy of Computing)


関連する6月14日サテライトWS url:

Presentators / 提題者

[HaPoC] Henri Stephanou (Université Paris 1 Panthéon-Sorbonne, philosophy of computing)

[ HSSJ(日本科学史学会) ] Mai Sugimoto :(杉本舞) (Kansai University, History of Computing)

Discussants / 特定質問者

[JAPS (科学基礎論学会)] Yuko Murakami (村上祐子) (Rikkyo University, Information Philosophy and Philosophy of AI)

[JAPS (科学基礎論学会)] Kiyotaka Naoe(直江清隆) (Tohoku University, Philosophy of Technology)

Workshop Organizers

JAPS 科学基礎論学会 (serving as the Discussion Co-Coordinators, as well) , Mitsuhiro Okada 岡田光弘 (Keio University) Jun Otsuka 大塚淳(Kyoto University)

HSSJ 日本科学史学会: Mai Sugimoto 杉本舞 (Kansai University)

HaPoC(*): Alberto Naibo (Université Paris 1 Panthéon-Sorbonne)

(*)HaPoC: InterDivision Commission for History and Philosophy of Computing of IUHPST/DHST and IUHPST/DLMPST(国際科学史技術史科学基礎論連合(IUHPST)/科学史技術史部門(DHST)・科学基礎論部門(DLMPST)共同設置によるコンピューティングの歴史と哲学)大塚淳

June 16th Main Workshop (6月16日WSの最新情報告知ページのurl):

Poster of the June 16th Workshop 6月16日ポスター

Note [Satellite Workshop June 14th]

A satellite workshop on this theme will be held on Friday 14 June at the Toyama Campus of Waseda University, The details will be posted soon at url:


[サテライトワークショップ6月14日] 本テーマに関するサテライトワークショップが大会前日6月14日金曜日に早稲田大学戸山キャンパスで開催されます。HaPoCからのStephanou氏のほか、日本科学史学会及び科学基礎論学会から登壇者を迎えて討論いたします。サテライトワークショップはオンライン参加も可能となる予定です:上のURLで情報を近日中に掲載します。

The Background and the Purpose of this Workshop

 Our lives and society are surrounded by the rapidly changing computing and algorithmic environments. The purpose of this workshop is to discuss how the evolving changes of the computing and algorithmic environments affect our lives society from the perspectives of history and philosophy of computing.
 The Japan Association for Philosophy of Science and IUHPST/DLMPST co-organized a symposium to discuss some social issues regarding fairness, integrity and transparency of computing and algorithmic systems, which resulted in fruitfil results. This workshop is designed to discuss with perspectives from the history of computing together with the philosophy of computing. The organizers of the Japan Association for Philosophy of Science has prepared this workshop in cooperation with the History of Science Society of Japan and the DHST-DLMPST InterDivision Commission for History and Philosophy of Computing. The workshop was co-hosted by the three organisations together with the Japanese Society for the History of Science and the IUHPST/DHST-DLMPST InterDivision Commission for History and Philosophy of Computing. In this Workshop, “Automation” in the computing-algprithmic environments and its influence to our society will be discussed, among some others. (Some of the other historical and philosophical questions, e.g. AI and philosophy, will be discussed in Satellite Workshop on June 14th.)
 The abstracts of the two main presentations are enclosed bellow. (Mitsuhiro Okada and Jun Otsuka)

Purpose and Bachground in Japanese

 2021年度大会シンポジウムではIUHPST/DLMPST(科学史技術史・科学基礎論連合/科学基礎論部門)との共同主催の形式で、computing・algorithms環境における公平性・公正性・透明性の問題を議論しました。コンピューティングの歴史学と哲学の視点に重点を置く今回のワークショップでは、日本科学史学会及び、DHST-DLMPST InterDivision Commission for History and Philosophy of Computingの協力を得て、これらの両組織とともに3組織共同主催形式で開催するに至りました。本ワークショップでは上で触れましたようにコンピューティング・アルゴリズムズ環境がもたらすAutomationと社会との関係について特に検討されます。(その他の歴史的・哲学的諸問のいくつか(例えばAIと哲学の問題)については、サテライトワークショップで議論されます。) 本ワークショップ使用言語は英語です。二人の提題者のAbstractsは以下の通りです。


1. “Automation v Augmentation: the two forms of technology at work”

Henri Stephanou (Université Paris Panthéon-Sorbonne, IHPST)

 In the past ten years, the rapid progress of AI has reignited the debate on the potential of technology to automate full spans of business activities and its dire consequences on employment. This possibility has been demonstrated in many physical places, such as data centers, warehouses, and factories. However, experts are more nuanced. Full automation has its own, daunting, challenges, and some rather claim that the future of "smart industry" lies in increased agility enabled by a new alliance between human and machine, sealed by AI. This debate raises the question of what is automation in the first place, a term which is notoriously difficult to define.
 In this talk, we will consider a few examples of automation over time, starting with Jacquard's loom in the early 19th century to fully automated warehouses at the end of the 20th century, in order to show that automation should not be defined relatively to the human tasks which they replace, but relatively to the problems they solve. Automation is a specific way to solve a problem through the operation of a machine. It requires the problem to be solved /systematically/, a term which we will explicate in detail.
 Automation has however strong conditions, as it requires the problematic situation itself to behave systematically. This is why it works best in closed, carefully designed environments such as a factory. Open situations, such as a lawsuit, are until now the domain of traditional human problem solving. In that case, the human agent keeps the central stage in solving the problem at stake, and technology is only here to facilitate her performance of some tasks, -- for example a legal database search -- i.e. to /augment/ her. /Automation/ and /augmentation/ are the two basic modes by which machinery can be put to work, because they are based on two fundamentally different approaches to problem solving.
 However what we believed were open situations have been regularly redefined to become systematic ones. How can this happen? Our thesis is that, whereas typical machines operate through rigid mechanisms that can only solve a narrow set of problems, computers have mechanisms which are easily reconfigurable ("programmable"). The progress of computability theory and software design has made machines increasingly /flexible/ to solve ever larger classes of problems -- from the rigid mechanisms of the industrial revolution to algorithmic behaviour, then to concurrent interaction, and now to adaptive, inductive "learning".
 Is there a limit to this trend? AI is about the automation of problem-solving itself ,and therefore it could in theory expand vastly the scope of automation. However, AI is today mostly presented as an augmentation technology (a "co-pilot"), and not an automation one. We suggest three reasons for this limitation : First, the need of AI to be complemented by traditional computational tools (such as compilers), which will still face the same traditional difficulties. Second, the data collection cost for achieving systematicity, which will make it in many cases unprofitable. Third, the fact the human meanings evolve over time. Language is a chaotic system which evolution cannot be predicted, because it reflects the evolving and intersecting conversations, actions and practices of communities of people.

2.“Low-Grade Human Labor” and Computers: Discourses on Technological Unemployment in the Mid-20th Century

Mai Sugimoto (Kansai University)

 Algorithms became available in a variety of environments with the advent of high-speed digital computers in the 1940s and the commercialization of programmable computers in the 1950s.
 When digital computers were developed in the 1940s, there was much discussion among mathematicians and logicians familiar with them about what the new digital computers could do and how they would affect the work of mathematicians. This was connected to the fact that digital computers were built to perform calculations and computations that were too large and tedious for humans to handle. For example, Alan Turing stated in his “Lecture to the London Mathematical Society Feb. 20 1947” that good “mathematicians will be needed in order to do the preliminary research on the problems, putting them into a form for computation”, while “stereotyped” tasks, such as card punching, would be easily replaced by machines.
 Discussions about “automation” began in the United States in the 1950s. This was triggered by the incorporation of control devices and later computers into production lines, i.e., the beginning of general-purpose use of computers, including commercial use, beyond the scope of numerical computation work. The term “automation” became widely known with John Diebold's article in 1952.
 The argument of factory automation and labor, and thus technological unemployment, in the 1950s is well-known for Norbert Wiener’s book The Human Use of Human Beings. In several works, including this book, Wiener discussed how the introduction of computers and feedback control devices into various locations could transform labor and how this could affect human economic life and culture. About mathematics, Wiener noted in his unpublished paper that “mathematicians who can make adequate use of [machines’] great capacities” would perform “marvelous tasks” using computers, and that in the future “logical computing laboratories” will develop, where experts will “use the machine as an extension of their brains”. He added, “[t]he machine as an adjuvant and supplement to human activity will unquestionably be of great value and importance in the future in all ranges of human activity”.  According to Wiener, “low-grade human labor” was the “cement” that kept the human society together, but once it was gone, “other motives, other employment of leisure, and other human valuations” would be needed as the new cement.
 This Wiener’s argument could be understood in relation to the issues addressed by John Maynard Keynes in “Economic Possibilities for Our Grandchildren” (1930) and partially discussed in Bertrand Russell’s “In Praise of Idleness” (1932). Weiner was taught by Russell in the 1910s, and Russel referred to Weiner’s book in a 1952 essay, in which he went further in his conclusions about the consequences for society of technological unemployment and the reduction of labor. Russell said the development of “mechanical brains” would take the place of factory labor and human decision-making, and the labor movement would inevitably end as working hours got limited. Russell stated that “our whole ethical system” based on the idea that “people ought to be useful and should show their usefulness by work” would collapse.
 More than 80 years later, has the working environment concerning computing changed dramatically? In the 20th century, it might be said that the substitution of labor as described by Turing, Wiener, and Russell was only partially realized due to the limitations of computer performance. On the other hand, however, “low-grade” computational labor has not necessarily been replaced by machines in the 21st century. Instead, previous studies have shown that still it has been carried out by people in weaker positions, for example, women and people in developing countries. The process of mechanization and automation over computation and control over the past century needs to be re-examined historically.