carnes@gargoyle.UUCP (Richard Carnes) (12/05/85)
I want to take a break from the controversies currently raging in this newsgroup and post something which I think will be of general interest. The following is an excerpt from an unpublished lecture by John E. Roemer, an economist at UC Davis. A mathematically rigorous version of the argument is published in *Econometrica* 53 (1985). _______________ 1. INTRODUCTION. Revolutions have been thought of by social scientists, for the most part, as inexplicable or irrational events, events exogenous to the jurisdiction of the models we build to explain economic and social behavior. (Historians, being less enamored of models, have placed revolutions more centrally in the purview of their discipline.) The hurdle for economists and political scientists in explaining revolutions is the collective action problem: why should any individual join a revolutionary movement, when the costs to him of participation are potentially very high, and the benefits, if the revolution succeeds, will be enjoyed by him even without his participation? Rosa Luxembourg was not unaware of the problem of explaining mass action (she wrote on the psychology of the mass strike); I suppose, in our modern jargon, she could be translated as saying that in certain situations, the preference structure of the Prisoner's Dilemma game does not characterize how people think about participating or not participating in the mass movement; they have, instead, what are called Assurance Game preferences, in which each person derives more utility from participating if others do than from defecting or scabbing if others participate. Saying this doesn't make it so, of course, and the hyperrationality which characterizes economic theory today pushes us to ask: what, then, explains the switch in preferences from those which characterize the "individualistic" Prisoner's Dilemma, to those of the "cooperative" Assurance Game, facilitating possibility of mass action? In this talk, I sidestep this central question. Too often, I think, in economics, we fail to study phenomena because they don't make sense given our way of thinking about things. Or, worse, we deny a phenomenon exists because we can't explain it. The attitude of many economists towards mass unemployment is an example of this last casualty of our scientific method. And while no one, so far as I know, denies the existence of revolutions, they tend not to be studied, not because of their unimportance, but because our present models of individual behavior do not enable us to get to first base in explaining them. It is not a big step to conclude, as a consequence of this failure of theory, that revolutions are aberrations, or exogenous like earthquakes. I have avoided calling upon the *deus ex machina* of "paradigm myopia" to explain the non-study of revolutions, because I do not wish to suggest that the "other" so-called paradigm, Marxism, has all the answers regarding them. Certainly for Marxism revolution occupies a central place, but I do not think Marxists have answered very well questions of collective action and revolutionary dynamics. [John Roemer] [to be continued] -- Richard Carnes, ihnp4!gargoyle!carnes
laura@l5.uucp (Laura Creighton) (12/07/85)
For any given revolution-type action there will always be people who will understand that *nothing is going to happen unless somebody does it*. They are not particularily interested in how fair it will be that some people will enjoy benefits which they have not earned by joining the revolutionary movement -- they are interested in the fact that if they don't do something they will not get the benefits that they want. -- Laura Creighton sun!l5!laura (that is ell-five, not fifteen) l5!laura@lll-crg.arpa
carnes@gargoyle.UUCP (Richard Carnes) (02/11/86)
While this article is formally a reply to Wasilewsky, it will also serve as a further reply to Fencsik's question about the nature of contemporary academic Marxism. Jan writes, concerning "A Tale of Lenin and the Tsar": >I realize these are two fictitious characters. However, the fact that >they behave completely out of character for their historical >namesakes *does* reflect, to me, on the value of the analysis. The >author says: here is how two forces behave, let us understand why. >The answer: because you invented them this way - leaves his >conclusions without foundation. > >The fictitious "Lenin" has, as his main priority, the overthrow of >the tsarist regime - which, however, treats him with relative >leniency, and we are invited to understand that. Having this >priority, he calls for "progressive income redistribution" to obtain >the support of the people. All this, for historical Lenin and the >tsarist government, is complete hogwash. Jan here misses the point of Roemer's article, which may not have been very clear although I thought it was. The name "Lenin" was chosen for one of the agents in the tale simply because Lenin is the most famous revolutionary of our time. "The Tsar" is the obvious name for "Lenin's" opponent in a two-person game. Roemer did not state or imply that the model (or rather models) he proposed had any particularly close correspondence to the historical reality of the Russian Revolution, nor do his conclusions depend on a close correspondence. The principal question Roemer is addressing is this: Is it possible to understand revolutionary ideology in terms of individual rationality? Revolutions need not be black boxes; we can (Roemer suggests) understand how they work. As a step towards this understanding, Roemer constructs a model in the form of a two-person game. The purpose of constructing models is to enable us to substitute precision for handwaving; by starting with a set of precisely defined assumptions, we may be able to derive interesting results that shed some light on the real world. It goes without saying that the real world is more complex than the model; it also goes without saying that if the model's assumptions are too wildly unrealistic (e.g., "people seek to minimize their income"), then the model is of little use. Jan appears to believe that the model's assumptions are unrealistic: >As for "income redistribution" - the error is even worse here. >This was not *anywhere* on Lenin's, or the Bolsheviks', list of >priorities. (Redistributionism and "petty-bourgeois egalitarian- >ism" are among the swear words of Bolshevik theoretical jargon). >Their agrarian program was very characteristic: the land of the >landlords (about half of the land in use) was to be expropriated, but >*not* parcelled to the peasants. Rather, it was to be developed >by the state ("bourgeois-democratic" state) with hired labor, because >large-scale production units were more progressive. The Mensheviks' >program differed in that the land was to be municipal, not central >government, property. There was a party that wanted all the land to >go to the peasants - the Socialist-Revolutionaries (SR) and some >other parties with more moderate redistribution agenda. In 1917, >after their coup, Bolsheviks suddenly scuttled their own program and >adopted the SR program - for the good reason that it had the >overwhelming support of the peasants. Such lightning changes of >policy were quite characteristic of Lenin. But of course that had >nothing to do with defeating the Tsar who had been deposed long >since. Jan's historical account is interesting but beside the point. It shows that the Russian Revolution did not follow exactly the course of Roemer's model. Of course -- no one has claimed that it did. But Jan's account also shows that progressive redistribution (from rich to poor) was very much on the minds of the revolutionaries, even if they wished to redistribute wealth in the form of collective rather than private property. And indeed there is a large class of historical situations (including e.g., Central America and South Africa today) in which we may presume that the desire to improve one's material situation is one of the principal motives for engaging in revolutionary activity. So I think it is not grossly unrealistic for the Lenin of the fable to propose a progressive redistribution of income. In fact, one of Roemer's main results is that, given the other assumptions of the model(s) (which of course I have not yet specified), Lenin MUST propose a progressive redistribution if he wishes the revolution to succeed. This is what Roemer means by "rationalizing revolutionary ideology"; under the terms of the model(s), proposing a progressive redistribution can be viewed as a rational strategy for winning the two-person game with the Tsar. If I have time, I will post some more excerpts from the article to clarify these points. For now let me post these general methodological remarks from another article by John Roemer. Reprinted in defiance of bourgeois copyright laws. __________________ Marxism's method, it is often claimed, is the dialectical analysis of social forces; Marx's key insight was to see the individual was determined by his social milieu, indeed his relation to the means of production; methodological individualism ignores this insight by placing the individual in the center of the story. Classes must be the atoms of the Marxian system, not strategic individuals [or so it is claimed]. The power of the Marxian argument must come, however, not by postulating that classes are the atoms of the system, but by proving that as a *theorem*. One has no choice but to take individuals as given and separate. We observe that individuals frequently act in class units, the problem being to explain this in a deductive way, based on postulates of individual behavior. A huge deficiency of bourgeois history and social science is its inability to understand class behavior and solidarity: as well as observing the importance of such behavior, Marxists are obliged to explain it. Marxism consists of at least these claims: (1) that classes act as units in certain historical situations; (2) that class position is best defined with reference to a person's relation to the means of production; (3) that class struggle is the mechanism of historical change; (4) that historical change has a certain pattern (feudalism, capitalism, socialism); (5) that capitalism is ethically bad (exploitation). To argue these points, Marxism requires at least three types of work -- history, social science and philosophy. I think the role of history is to formulate inductive laws, with reference to (1)-(4) above. The role of social science is partly inductive, to formulate laws on a somewhat more micro level than history does, but also deductive, to formulate axiomatic theory that can explain the inductive laws. For example, I have referred to proving a theorem that individuals will act as classes in some situations as an example of how (1) [classes sometimes act as units] can be given such han axiomatic basis. An example of (2) [class position being defined by relation to means of production] is the work of [E.O.] Wright, who shows by analysis of data that defining class by reference to the relations to the means of production gives more cogent explanations of certain phenomena than defining class with reference to income or status as non-Marxist sociology does. Game theory can be an important tool ... for claim (1) at least. I have argued above it is also useful in (3) [class struggle as the mechanism of historical change]. It is the role of Marxist philosophy to argue for (5) [capitalism is ethically bad], although that argument must depend on historical work. It is also interesting to note that game theory can be useful for (5), in clarifying ethical positions.... On this account, what distinguished Marxist from non-Marxist social science is not the postulate of methodological individualism, but the level of aggregation deduced as applicable in social analysis. In the non-Marxist view history is determined originally by man's struggle against nature, and no further level of aggregation is fruitful. In the Marxist view history may be determined originally by man's struggle against nature, but the struggle against nature leads to the formation of classes that struggle against each other. Indeed, class interests can then affect the struggle against nature. It is in this sense that Marxism performs an aggregation that bourgeois thought does not. The most efficacious lens for analysis may not be the one with the highest magnifying power, and by resolving always to the level of the individual one may lose the pattern. ....If we had a good axiomatic theory of class formation and development of class capacity under capitalism, perhaps a generalization would suggest itself applicable to socialism. I think the inadequacy of the Marxist understanding of developments in twentieth century socialism in large part results from the lack of formulation of a general deductive theory of class, inspired by the inductive observations that constitute historical materialism. There is a secondary purpose for using deductive models in Marxism: to win over non-Marxists by showing them that Marxism can be formulated as a consistent system, requiring that the choice for or against it must be made on the basis of which world-view better conforms to historical evidence, not on the basis of internal consistency. I do not mean to imply these ideological decisions are taken in the pristine vacuum of the library; but if we believe academic Marxism has any role, we must acknowledge the power and importance of rigorous argument in ideological debate. Both world-views, it appears, can be made internally consistent by modern standards of rigor. But I think this pedagogic purpose is secondary. The axiomatic method, and methodological individualism in particular, is useful not mainly as a language to convince skeptics, but as a tool to teach ourselves. [John Roemer] -- Richard Carnes, ihnp4!gargoyle!carnes