
The cycle's engagement with systemic effects—recommendation algorithms reshaping what a generation finds normal, AI tools restructuring the texture of knowledge work, personalized media dissolving shared factual ground—gains moral precision from Eliot's web. The dominant frame for thinking about a technology's effects is causal and local: we build a tool to do X, ask whether it does X well, attend to its direct outputs. Eliot's web is a standing refutation of that frame. Her insight is that the significant effects of an action are rarely its intended local ones; they travel through the social medium and emerge somewhere the actor was not looking. [YOU] on AI documents the personal-scale encounter with the AI transition; the web of consequence is the frame for understanding how that encounter propagates into shared life.
The web also explains why good intentions are insufficient. Dorothea Brooke's tragedy—real sympathy, genuine idealism, and naive effect—is the cycle's standing warning to builders who invoke their mission to benefit humanity. Eliot's answer is that benevolence was never enough; that intention loosed into a web you have not troubled to understand is its own kind of egoism, the assumption that good will guarantees a good result. The web does not care about intentions. It only transmits consequences.
Eliot's prelude to Middlemarch announces the ambition to unravel certain human lots and see how they were woven and interwoven. The novel makes good on it through a formal architecture that includes four principal narrative threads—Dorothea and Casaubon, Lydgate and Rosamond, Bulstrode and his past, Fred Vincy and Mary Garth—each intersecting the others at points no character fully understands. The final page gives the web its canonical statement: the growing good of the world is partly dependent on unhistoric acts, those of people who lived faithfully a hidden life and rest in unvisited tombs. The good diffuses; the diffusion is not the accidental byproduct of action but its moral architecture.
The concept is prepared in Eliot's early essays on realism, particularly the 1856 “Natural History of German Life,” where she insists that the novelist's obligation is to the whole web of social life rather than to the sentimental individual—that accurate rendering of provincial society requires tracing the interdependencies that actors within it cannot see. The essay anticipates the novel's form: particularity in the rendering of each life, systematicity in the account of how the lives entangle.
Consequence is not diluted by diffusion. A recurring temptation is to treat the propagation of consequence through many hands as distributing and therefore diminishing responsibility. Eliot's moral physics runs the other way: to act into a web is to assume a duty proportional to the reach, however uncomfortable the math. Bulstrode's tragedy is the tragedy of a man who believed his private sin was sealed off. The novel's machinery exists to prove there is no such sealing.
The web is reflexive. The people the action enters change their behavior in response, and the change alters the medium through which the next effect must travel. A piece of information about Bulstrode does not simply land; it is taken up, transformed by the moral character of each person it reaches, amplified by gossip, softened by charity, until the social consequence bears the mark of the whole community and not just the original act. A language model deployed into a human population enters a medium of the same reflexive kind—its outputs gamed, resisted, repurposed, so that the effects cannot be read off the design.
The unhistoric acts. The web's most important implication is Eliot's most radical conviction: that the real moral weight of human life is borne by actions that leave no measurable trace. The faithful ordinary act that slightly improves the good of the world diffuses through the web in ways no metric can register. Any system of measurement or optimization is, by construction, blind to most of what matters—blind to the unvisited tomb, blind to the fellow-feeling extended in private, blind to the roar that no statistic captures.
The central dispute concerns whether the web of consequence can be adequately modeled by the kinds of causal and systems analysis that engineering disciplines deploy. Optimists argue that complex network effects, second-order market dynamics, and multi-agent simulation can capture what Eliot described qualitatively—that her literary intuition has been mathematized. Eliot's framework pushes back on a deeper register: the web is not simply high-dimensional causality but a medium of meaning-making agents who respond not just to the direct effects of an action but to their moral interpretation of it, and that interpretation depends on everything they know about the actor's character and motives. No simulation of causal propagation includes the variable of Mrs. Cadwallader's opinion of Bulstrode. The web, for Eliot, is not a system to be modeled but a reality to be attended to—which is why the sympathetic imagination, not the causal diagram, is the tool she reaches for. Judea Pearl's framework for distinguishing observational from interventional reasoning is a partial technical complement: acting into the web changes it in ways that observing it cannot predict, which is why the do-operator must sever a variable from its ordinary causes. Eliot makes the same point in moral rather than mathematical language.