Carnap, Maturana, Bloom and Agrippa: four attempts to police a line

Tolerance was the real bequest, and "Two Dogmas" did less damage than the textbooks claim#

Logische Syntax der Sprache (1934; English 1937) carries the sentence that outlived everything else. Section 17 states the principle of tolerance: "In logic, there are no morals. Everyone is at liberty to build up his own logic, i.e., his own form of language, as he wishes." It converts disputes about what exists into disputes about which language to adopt. Metaphysics is not refuted on that view; it is reclassified as a proposal and asked what work it does.

Quine's "Two Dogmas of Empiricism" appeared in The Philosophical Review in 1951 and is generally credited with ending logical empiricism. The verdict is too tidy. Quine did not show the analytic–synthetic distinction incoherent; he showed that every attempt to define it ran through synonymy, interchangeability or semantic rule, each presupposing the notion under definition. Carnap replied in "Meaning Postulates" (1952) that in a constructed language the analytic sentences are those the rules stipulate: a decision, not a discovery. The real objection was to the decision: Quine denied there was any standpoint outside a framework from which to make it. Richard Creath's edition of their letters (1990) shows two men who understood each other and disagreed about what philosophy was for.

The last twenty years went on inductive logic. Logical Foundations of Probability (1950) distinguished probability as logical degree of confirmation from probability as relative frequency; The Continuum of Inductive Methods (1952) ranged the methods on one scale. Then came the embarrassment it never escaped: a universal law over an infinite domain gets confirmation zero, however much evidence accumulates.

Carnap asked what makes a string of symbols meaningful and replied that meaning lies in its logical relation to possible observation. A language model produces well-formed strings standing in no such relation, having made no observations. That looks decisive, and it is not. It establishes that well-formedness and fluency come apart from empirical grounding — Carnap's point against Heidegger, available unmodified against a transformer. It does not establish that the output is meaningless in any sense worth caring about, because the criterion failed on its own terms and he broke it himself: a line that could not be drawn sharply enough to exclude German metaphysics without also excluding theoretical physics will not now be drawn around a chatbot. Nor is the relation to observation absent: it is indirect and statistical, running through the writing of people who did observe things. Whether that is the right kind of relation is the whole question, and his apparatus cannot settle it — which is why the argument moved long ago to Searle's Chinese Room and what we talk about when we talk to these systems.

Maturana put the boundary of cognition at the boundary of the living#

Maturana read medicine in Santiago and took a Harvard doctorate in 1958. His first famous paper was a collaboration: "What the Frog's Eye Tells the Frog's Brain", with Jerome Lettvin, Warren McCulloch and Walter Pitts (Proceedings of the IRE, 1959), found that the frog's retina sends the brain no image but a handful of processed reports: small dark moving objects, edges, dimming.

The term autopoiesis followed around 1970. By his own account he coined it after a conversation about Don Quixote's choice between arms and letters, praxis and poiesis: he wanted a word with no history. An autopoietic system is one whose components continuously regenerate the network of processes that produces them, and thereby constitute it as a unity. De máquinas y seres vivos appeared in Santiago in 1972 with Francisco Varela; the English Autopoiesis and Cognition came from Reidel in 1980, carrying the 1970 essay in which the claim is put flatly:

Living systems are cognitive systems, and living as a process is a process of cognition. This statement is valid for all organisms, with or without a nervous system.

Cognition is therefore not representation of an external world; an observer who describes a system as tracking its environment reports his own vantage point, not the system's. The influence spread sideways: through Heinz von Foerster's laboratory at Illinois into second-order cybernetics, through Luhmann's Soziale Systeme (1984) into a sociology of communications rather than people, and through The Embodied Mind (1991) into enactivism, where Varela ended at some distance from his teacher.

On this account a language model is not cognising, whatever it emits, because it does not produce itself: it is built, maintained and powered by others, and nothing in its operation is directed at holding its organisation in being. That is a substantive position, not a definitional shrug, and it is contested. Ezequiel Di Paolo objects, from inside the tradition, that autopoiesis is all-or-nothing and gives no gradient of better and worse: nothing in the axioms, he writes, lets us derive that an autopoietic system "will attempt to improve a situation that leads to the future loss of autopoiesis". Donna Haraway rejects the premise outright in Staying with the Trouble (2016), proposing sympoiesis instead. Functionalists reject the framing altogether, self-production being for them an implementation detail — where the argument about substrate now sits. Maturana's line is defensible. It is not free.

Bloom's book is not the book that either side quotes#

Bloom took his first degree at Chicago at eighteen and a doctorate there in 1955, on Isocrates, under Leo Strauss. His literal Republic (1968), with its long interpretive essay, is still assigned; his Emile followed in 1979. He taught at Cornell from 1963 and resigned in 1970, with Walter Berns, over the administration's response to the armed occupation of Willard Straight Hall in April 1969 — the episode at the emotional centre of the book he wrote eighteen years on. Toronto kept him until 1979, then Chicago and its Committee on Social Thought.

The Closing of the American Mind, published by Simon & Schuster in March 1987 with a foreword by Saul Bellow, sold close to half a million hardbacks and led the New York Times non-fiction list for four months. Very little of it concerns curricula. Bloom's thesis is about moral psychology: that American students had absorbed a denatured German pessimism, the vocabulary of "values" arriving from Nietzsche and Weber stripped of the tragedy that gave it point, and that an openness understood as the refusal to rank anything removes the motive for enquiry. Rock music takes up as much of the book as the professoriate. Conservatives who read it as a great-books manifesto and liberals who read it as a demand for dead white men were both reporting a book Bloom had not written.

Martha Nussbaum's review, "Undemocratic Vistas", ran in the New York Review of Books on 5 November 1987, its title playing against Whitman. Her objections were technical, which is what made them hard to answer. Bloom's readings, she wrote, proceed with "almost never quotations or references to passages, and never an effort to discover the structure of an extended pattern of argument", and his Straussian Republic is "both bizarre and not accepted by any major non-Straussian interpreter". Most awkwardly for a man arguing from antiquity against feminism, he passes over the Hellenistic evidence against him: Musonius Rufus held that women should study philosophy. She added a point that has held up: "Bloom's indifference to the situation of all but a few highly atypical students should make us wonder about all general statements about 'students' and about American culture in this book." He was describing Chicago and Cornell undergraduates and calling them America. Nussbaum shared his premise that philosophy should alter a life; the quarrel was over the aristocratic conclusion he drew from it, and it is live wherever a syllabus becomes a legal matter.

Bloom died in Chicago on 7 October 1992, aged sixty-two. Bellow's Ravelstein (Viking, 2000) made him the title character and disclosed what had not been public: that Bloom was gay and had died of an AIDS-related illness. Bellow said Bloom had told him to tell it all. The man who spent a career on what belongs inside a serious education had the line around his private life erased by the one person best placed to know where he drew it.

Agrippa published the case against all knowledge, then the encyclopaedia of the occult#

The birth date is firmer than Renaissance birth dates usually are — 14 September 1486, at Cologne — and the death date is not: Grenoble, February 1535, or possibly early 1536. He soldiered for Maximilian I, lectured at Dole in 1509 on Reuchlin, and wrote for Margaret of Austria a treatise on the theological superiority of women. At Metz he defended a peasant woman accused of witchcraft against the inquisitor Nicolas Savin, and had to leave town.

De occulta philosophia libri tres was drafted by 1510 and sent to the abbot Trithemius; the first book was printed at Paris in 1531, the complete three at Cologne in 1533. It organises magic across three worlds — elemental, celestial, intellectual — as the practical part of natural philosophy, a discipline with rules. De incertitudine et vanitate scientiarum et artium, printed at Antwerp in 1530, takes the disciplines one at a time and demolishes each, magic included: reason proceeds by conjecture, and only faith is safe. It found readers as different as Montaigne, Descartes and Goethe, and is the more original of the two.

The reconciling reading treats them as one project in two movements, the sceptical book clearing away corrupted learning and the magical book rebuilding on theological ground. Coherent, and tidier than the texts. What the Vanitas does is enjoy itself, at length, at everybody's expense, the author of the Occulta philosophia included.

Three lines, and the man who would not draw one#

Carnap's boundary between the meaningful and the nonsensical was dismantled by Carnap himself, in instalments; what survives is the tolerance principle that replaced it. Maturana's line between cognisers and non-cognisers stands where he put it, at a price his successors will not pay: it settles by definition the question behavioural research is trying to settle by evidence. Bloom's boundary around the educated was drawn to defend philosophy from the culture and was at once conscripted by the culture, which is what happens to a line that can be used as a fence.

Agrippa is a control rather than a fourth case. He worked both sides of the boundary between the knowable and the unknowable, praised and buried the same disciplines within three years, and was in consequence itinerant and unemployable — which is one answer to what the boundary-drawing is for. A line makes a subject governable, and a man who draws none can be read for four centuries without being cited as authority for anything. Yesterday's column, on Midgley, Montaigne and Feuerbach, ended on the habit of mistaking a vocabulary for a discovery. Today's names its companion: mistaking a boundary for a fact, when it was only a decision about how to work.

Arthur Oncken Lovejoy was born in Berlin on 10 October 1873. He made the history of ideas a discipline, traced one idea from Plato to the Romantics in The Great Chain of Being, and helped write the rules of academic freedom he later qualified. The same date, in four Catholic countries, did not exist in 1582.

Arthur O. Lovejoy