The conditions never recur, because knowledge, technology and institutions accumulate

The argument

Those who hold that history does not repeat itself argue that the premise of recurrence fails where it matters. For an outcome to repeat, the conditions producing it must recur, and on this account they do not: the stock of knowledge is larger each time, the available technology is different, the institutions are different, and the participants know about the earlier episode. A famine in a society with grain reserves, forecasting and road transport is not a recurrence of a famine without them; an epidemic met with sanitation and vaccines is a different event from one met without; a financial panic in a system with a lender of last resort behaves differently from one without. Proponents argue that the superficial resemblance of outcomes conceals inputs that are never the same twice. The second argument is that the apparent pattern is supplied by the observer. Any period contains wars, financial crises, epidemics and leaders who overreach, so a pattern defined loosely enough to cover those will be found in every era by anyone looking. Proponents argue that the cases offered as repetitions are selected after the fact, that differences are discounted as detail while similarities are treated as structure, and that the same reasoning would establish any pattern one cared to propose. On this view the better description is that history rhymes without repeating: recurring human problems are met by changing means, and the outcomes diverge accordingly. Proponents argue that treating recurrence as a law is not only inaccurate but discouraging, since it implies that the accumulated knowledge which actually changes outcomes makes no difference.

Premises

Counter-arguments

Those who hold that history repeats answer that this confuses the form of an outcome with its substance. Reserves and forecasting change how a famine proceeds without changing why food was allocated away from those who needed it, and a lender of last resort alters the mechanics of a panic while leaving intact the overconfidence and leverage that produced it. They also argue that the charge of loose definition cuts the other way: insisting that conditions be identical before an outcome counts as a repetition sets a standard no historical claim could meet, and so rules the question out rather than answering it.

Rejecting the premises

[Rejecting P1] Requiring identical conditions before an outcome counts as a repetition sets a standard no historical claim could meet, and so rules the question out rather than answering it. [Rejecting P2] The examples change the mechanics of an outcome while leaving its cause intact: reserves alter how a famine proceeds without altering why food was allocated away from those who needed it. [Rejecting P3] Loose definition is a fault of careless versions of the claim, not of the claim itself, which can be stated as specific configurations with specific predicted consequences. [Rejecting C] The premises support the view that repetitions are never exact, which the opposing position already grants.