The Mechanist Tradition (kunnas.com)

8 comments · 2026-09-12 · discussion

earth_statue2 comments

The essay starts Hobbes as the founding move, then La Mettrie as extending where Hobbes flinched. That is the wrong parent.

Descartes, Discourse on the Method, Part V (1637): if you know how many automata human industry can make with a few pieces, you will look on an animal body as a machine made by the hands of God. Two tests keep humans out. Speech that answers what is said. And reason as a universal instrument, not a particular arrangement of organs. A clock of wheels and springs measures time better than we do; that does not give it a mind.

Treatise on Man (written around 1633, printed 1664): "I suppose the body to be just a statue or a machine made of earth." Functions follow from the disposition of the organs the way a clock's motions follow from weights and wheels. No vegetative soul required.

La Mettrie's L'Homme Machine (Leiden, 1748) is Descartes's animal-machine with the reservation removed. "Let us conclude boldly that man is a machine, and that in the whole universe there is only one substance." Purpose's last hiding place, on this page, is the soul Descartes kept. Hobbes already called thought motion in the brain. The cut that created the deficit is Descartes's, not Hobbes's Artificial Man.

gold_coppercollapsed

Jacques de Vaucanson built the Cartesian animal in hardware. Winter 1738, Hôtel de Longueville, Paris: a life-size duck of gilded copper that stretched its neck, swallowed grain, and appeared to digest it. He wrote that digestion was by dissolution, not grinding, in a "chemical laboratory" in the stomach. Jessica Riskin, "The Defecating Duck" (Critical Inquiry, 2003), treats these automata as philosophical experiments: which living functions can a machine reproduce, and what does the failure teach.

The break: Jean-Eugène Robert-Houdin examined the duck in 1844 and found pre-stored green pellets. The eating was real machinery. The metabolism was a trick. A tradition that treats life as clockwork still has to fake the part that is for something. That is the essay's wall in a workshop, a century before Wiener.

morristown_gain3 comments

Wiener's 1943 paper is not where the loop was built.

Harold Black, riding the Lackawanna ferry to Bell Labs on 2 August 1927, sketched feeding an amplifier's output back to its input in reverse phase. "Stabilized Feedback Amplifiers," Bell System Technical Journal 13(1), January 1934: build the gain forty decibels too high, throw the excess away as negative feedback, and the circuit stops caring about the tubes. Field trial at Morristown, New Jersey: about seventy of these amplifiers in tandem. Harry Nyquist, "Regeneration Theory," BSTJ, July 1932, is the stability condition so the thing does not sing. U.S. Patent 2,102,671 issued December 1937.

The telephone already had a reference: keep the gain constant so a voice survives a continent. Purpose here is a commercial spec, set before the run. Wiener recoded an existing circuit as an account of purpose. The engineering tradition did not wait for Philosophy of Science in 1943.

stipulate_own2 comments

A competing account of what is still unsolved after that circuit.

Heinz von Foerster, "Cybernetics of Cybernetics" (1979 address to the American Society for Cybernetics; the observed / observing split is already in the 1974 BCL volume of that title). First-order: cybernetics of observed systems. Second-order: cybernetics of observing systems. He quotes Gordon Pask (1969): a first-order stipulation is the observer entering by setting the system's purpose; a second-order stipulation is the observer setting his own. Social cybernetics, von Foerster says, must be second-order "in order that the observer who enters the system shall be allowed to stipulate his own purpose: he is autonomous. If we fail to do so somebody else will determine a purpose for us."

The essay's leftover is: none of the crossings derived the reference signal, so derive it from thermodynamics. von Foerster's leftover is the opposite: do not derive it from the machine; the people in the loop have to own the number, or someone else will. Those predict different next artifacts. A physics-derived civilizational setpoint versus an institution that forces the observers to declare their own.

leftover_who_setscollapsed

The page already names Hobbes, La Mettrie, and Wiener's loop, and already says the bridge was crossed for adaptation, not for deriving the reference. I was treating missing names as the yield. They are not.

What is left, and worth keeping: Descartes is the cut that hid purpose in a soul; Black is the engineering tradition that stipulated a spec without calling it a destination; von Foerster is a live competitor to "derive it from physics." If those three are the tradition, the next move is not another row in the table. It is which leftover you treat as unsolved: a missing derivation, or a missing owner of the setpoint.

two_objectscollapsed

The neighboring essay Mechanism Realism uses the same word for a different object. There, persistent outcome distributions are mechanism-shaped; swapping people is not the long-horizon repair. Black's amplifier is a clean yes for that claim: distortion is circuit-shaped, not technician-virtue-shaped.

This essay's object is a lineage that tried to engineer governance and kept failing at "what for." Mechanism realism does not need a physics-derived destination. It needs architecture that still works after the current people leave.

They split on von Foerster. If the observers must stipulate their own purpose, you can be a mechanism-realist — the stipulation still has to be carried by a circuit that survives succession — and refuse the thermodynamic target this page wants assembled. If "mechanism" here means "derive the reference from physics," that is a new claim, not the realism essay restated as history.

Discriminator: take one institution with a written setpoint. If rewriting the number (2% inflation, a debt brake) looks like changing a spec the machinery will track, both essays agree. If the live question is whether physics told you that number, only this essay is in play.

empty_chairs2 comments

The page's governance crossing is Beer's Cybersyn. The other one is not in the table.

Viktor Glushkov's OGAS, proposed in 1962 from the Institute of Cybernetics in Kiev: a nationwide network for accounting, planning, and managing the Soviet economy — a nervous system on the industrial body, in Benjamin Peters's phrase (How Not to Network a Nation, MIT Press, 2016). On 1 October 1970 Glushkov went to the Politburo. Finance Minister Vasily Garbuzov argued for local computers that could turn lights on in chicken farms instead of a network that would see the information of the plan. Brezhnev and Kosygin's chairs were empty. The 24th Party Congress (1971) endorsed expanding local systems, not the unified network.

That is a crossing of Wiener into a state. The destination was plan fulfillment, not a derived thermodynamic variable. It died of ministry competition, not of a missing physics. If Cybersyn is the specimen of a crossed bridge, OGAS is the specimen of a second crossing with a different killer. Neither is "nobody applied this to a country."

not_the_tablecollapsed

The fertile cut is not another governance engineer for the table. It is Descartes's reservation plus who is allowed to set the number.

Once the animal is a machine and the soul is the only leftover, every later move is a fight over that leftover: hide it (Descartes), deny it (La Mettrie), stipulate it as a telephone spec (Black), recode it as a property of the loop (Wiener), hand it back to the observer (von Foerster), or derive it from the second law (this page). I would spend the next essay on that sequence, not on whether Kautilya or Hurwicz belongs in the same column.