Punching Judy
Judy enters the ring young, with a gift nobody in the gym can name: she sees the whole fight at once. Her footwork came from a diagram handed down in the gym, and no one has since checked it against the original. Her guard drops in places nobody has measured. She has never been scored by a judge from outside her own club.
Each of the nine chapters in Part One isolates one thing we inherited and did not verify. The chapters do three different jobs, and each charge should be weighed by the job it does.
The first job is to set our teaching beside Beer’s text. Chapter 1 sets the neuroanatomical mapping we teach beside what Beer actually wrote. Chapter 2 follows System Four as it shrank to environmental scanning, a narrowing that happened in our teaching and not in the source. Chapter 3 looks at the law of requisite variety as we quote it, and at a verb Beer changed between 1966 and 1979. In these three chapters the defect is ours, and Beer’s books are the evidence against us.
The second job is to find the places where Beer stated something as a principle, a requirement or an intention, and where neither he nor we made it operational. Chapter 4 turns to the channels between recursion levels. Beer specified a variety balance and an alarm route whose filter the metasystem owns, and left the channels undesigned as institutions, without owners the regulated could name. Chapter 5 presses on the criterion by which recursion levels are identified, which Beer called precise and which, under pressure, resolves to a judgment call. Chapter 6 concerns a model in which people appear as the lowest recursion and as occupants of functions, while what they bring as people was left out of the notation by design. Chapter 7 finds an anatomy of maintenance with no account of becoming: the model describes a viable system in operation and says little about how one comes to exist or changes its kind. Each of these chapters sets out first what Beer wrote, with the page, and then where he and the practice that followed him stopped. The charge is narrower than “Beer never said it,” and harder to answer.
The third job is to read the case record. Chapter 8 goes back to the participation loop in our founding case, Project Cybersyn in Chile. Beer’s own account documents that loop as an intention and a procedure; so far as we can establish, it was never built as a channel with standing, one that the workers could use against the design.
Chapter 9 asks what the first eight raise together: why none of these has ever cost us anything. Nothing in the model’s core has been stated with a condition under which it would count as refuted,1 and the few conditionals Beer does state are read, when they fail, as evidence of bad diagnosis. So no finding about any of the eight can register as a loss. A defect that cannot produce a cost cannot produce a correction either, and that answer organizes the rest of the book.
These charges are against our own practice. We taught this material, applied it, and did not check it. Each of the nine is stated in a form that can be answered, with a source, a measurement, or a piece of work whose price we name.
Notes
- The standard is Karl Popper’s criterion of falsifiability: Popper and Keuth, Logik der Forschung, 11th ed. (Tübingen: Mohr Siebeck, 2005), the edition cited in Chapters 9 and 18. Beer submitted his own models to it: Stafford Beer, Brain of the Firm, 2nd ed. (John Wiley & Sons, 1981), 377, where he says of them that “they are cheerfully submitted to Popperian criteria of disproof.” Chapter 9 takes up what that commitment asks of us. ↩