Perhaps one of the most misremembered paper in macroeconomic theory is that which gave birth to the famous Phillips Curve. The original paper, which drew upon nearly a century's worth of UK data, demonstrated an empirical relationship between the rate of unemployment and the change in nominal wages (commonly referred to as wage inflation). The result - that money wages increase more rapidly with less unemployment - opens a few potential theoretical interpretations.
The Marxian interpretation would have it that larger pools of unemployed workers would make the overall workforce more disposable, hence reducing worker power to secure higher wages. This is largely the position that Phillips himself took although I don't believe he labelled it Marxian.
Another interpretation embraced by the mainstream of economics asserts that this relationship indicates the tendency for programs geared towards "artificially" reducing unemployment ultimately results in inflation. The underlying presumption here is that any sort of policy intervention to boost employment will necessarily be less productive than the private sector. Thus, additional workers employed at the going wage but producing less value will lead to more money chasing fewer goods. With a larger supply of money in the hands of workers but not as much to spend it on, these wage earners will simply bid up prices, leading to inflation.
Here, we have the recipe for the bastardization of the Phillips Curve by the likes of Ned Phelps and Milton Friedman. These two reinterpreted the Phillips Curve to be a relationship between unemployment and price inflation. This version of the model broke down soon after it was developed during the oil price shocks of the early 70s. Hence, a new breed of economists, led by the likes of Robert Lucas, Finn Kydland, and Edward Prescott, proclaimed that discretionary economic policy could not work because the model that economists had (over)developed failed to predict inflation due to structural changes. These economists attributed this to reactions to policy pronouncements by economic actors on the basis of their expectations. Hence, this later breed recommended policy rules that would set incentives for individuals to work around.
Anwar Shaikh has shown, however, that although the later versions of the Phillips curve don't hold, the original model still holds.