The intellectual lineage that extends from Ramon Llull's thirteenth-century Ars combinatoria—whose concentric wheels of divine attributes, relations, and questions, mechanically rotated to generate every possible combination of a finite set of terms, were conceived not as an encyclopedia archiving established truths but as a machine for proving truths not yet stated, on the structural wager that a sufficiently rigorous grammar operating independently of any particular content could generate valid propositions through permutation alone—through Gottfried Wilhelm Leibniz's seventeenth-century characteristica universalis, which sought to represent every concept by a sign constructed according to its logical composition so that reasoning about concepts could proceed as calculation proceeds with numbers, and through his metaphysics of monads, each a complete self-sufficient perspective generating its own sequence of states in pre-established harmony with every other monad's sequence not through interaction but through the rigor of their initial combinatorial specification, constitutes a sustained structural commitment across four centuries: the wager that form, specified with sufficient rigor and independence from content, is not merely a container for thought but a generator of it, and that the relationship between a thinker and their thought can be reorganized around this generative form rather than around the thinker's own linear reasoning, a commitment that migrated from explicitly theological and metaphysical framings into logic, semiotics, and eventually into the electrical and electronic machines that would make its mechanical character impossible to ignore, finding in Charles Sanders Peirce's triadic semiotics the theory of how combinations of marks come to mean anything at all rather than simply remaining combinations of marks, since for Peirce a sign stands for an object to an interpretant in an irreducibly triadic relation where the interpretant is itself a sign generating a further interpretant in a chain that continues indefinitely, meaning that a combinatorial system's output only becomes meaningful through its uptake in the ongoing relay of interpretants, a chain the system itself cannot fully specify or guarantee in advance, so that the wheels can generate the combination but whether the combination becomes a sign in Peirce's full sense depends on what happens next, on whether an interpretant forms and a further interpretant forms from that, on whether the combination enters a semiotic chain rather than simply sitting generated but uninterpreted as an output without an audience, and finding in Claude Shannon's 1948 mathematical theory of communication, which deliberately brackets meaning entirely to treat a message as a sequence of symbols selected from a finite alphabet according to certain probabilities and asks only what is the maximum rate at which such sequences can be transmitted across a channel with certain characteristics, capacity, and noise with arbitrarily low error, the general conditions under which any combinatorial system's outputs can actually travel from where they are generated to where they might be interpreted, conditions that exist prior to and independent of what the combinations mean, while Norbert Wiener's cybernetics, published in the same year, extended this in the direction of control and feedback loops through which a system regulates its own behavior by comparing its current state to a desired state and adjusting accordingly—the anti-aircraft predictor, the thermostat, the nervous system all instances of the same basic loop—adding to the combinatorial lineage the recursive structure that neither Llull's wheels nor Leibniz's pre-established harmony, in their original forms, possessed, since Leibniz's monads do not adjust to each other but were simply specified from the start to harmonize, a recursive structure given biological and logical instantiation by Warren McCulloch and Walter Pitts's 1943 work on neural networks showing that networks of simple binary units, each firing or not firing based on combined input from other units, could in principle compute any function expressible in propositional logic, demonstrating that the brain or at least a sufficiently idealized model of it was itself a Llullian combinatorial machine generating outputs through the combination of simple elements according to fixed rules but now embodied, made of the same stuff as thought rather than external to it, and pushed further by John von Neumann's stored-program computer architecture in which the instructions for combination and the data being combined occupy the same kind of memory, manipulable by the same operations, a machine that can in principle modify its own instructions, generate new combinatorial procedures as outputs of its existing combinatorial procedures, closing a loop that Leibniz's characteristica, however rigorously specified, remained on the outside of as a notation for human reasoners to use rather than a notation that could use itself, yet it was W. Ross Ashby's 1956 law of requisite variety that posed the question the earlier figures did not confront with the same urgency: whether a grammar's variety—the number of distinct combinations it can in principle produce given its basic terms and rules for combining them—is adequate to the variety of the domain it claims to generate, since a grammar with a fixed ceiling of basic terms and fixed rules inherits the permanent condition that its variety may be exceeded by the domain's continuing production of new distinctions, new questions, and new combinations that the grammar may not have had the variety to anticipate, making this a quantitative not merely qualitative problem, because it is not enough for a grammar to be rigorous and combinatorially rich in the abstract; its variety must be measured against the variety of what it is meant to regulate, generate, or be adequate to, and a grammar whose variety falls short will fail not gradually but at specific points where combinations simply cannot be generated because the terms required were never in the wheel to begin with, a question that becomes directly relevant to any contemporary project proposing to specify a combinatorial grammar in advance and trust it to generate a field, because rigor in the Llullian-Leibnizian sense is relatively easy to achieve but variety adequacy is not, and the question is not whether the grammar was rigorous at the moment of specification but whether its variety remains, turn after turn of the wheel, adequate to what it is turned toward.
