Machine Dreams: Economics Becomes a Cyborg Science

by Philip Mirowski

This is the first cross-over book in the history of science written by an historian of economics, combining a number of disciplinary and stylistic orientations. In it Philip Mirowshki shows how what is conventionally thought to be "history of technology" can be integrated with the history of economic ideas. His analysis combines Cold War history with the history of the postwar economics profession in America and later elsewhere, revealing that the Pax Americana had much to do with the content of such abstruse and formal doctrines such as linear programming and game theory. He links the literature on "cyborg science" found in science studies to economics, an element missing in the literature to date. Mirowski further calls into question the idea that economics has been immune to postmodern currents found in the larger culture, arguing that neoclassical economics has surreptitiously participated in the desconstruction of the integral "Self." Finally, he argues for a different style of economics, an alliance of computational and institutional themes, and challenges the widespread impression that there is nothing else besides American neoclassical economic theory left standing after the demise of Marxism. Philip Mirowski is Carl Koch Professor of Economics and the History and Philosophy of Science, University of Notre Dame. He teaches in both the economics and science studies communities and has written frequently for academic journals. He is also the author of More Heat than Light (Cambridge, 1992) and editor of Natural Images in Economics (Cambridge, 1994) and Science Bought and Sold (University of Chicago, 2001).

Genres: Economics, History, Nonfiction, Politics, Philosophy, Science, History Of Science

672 pages, Paperback

First published December 3, 2001

About the author

Philip Mirowski (born 21 August 1951, Jackson, Michigan) is a historian and philosopher of economic thought at the University of Notre Dame (Carl E. Koch Professor of Economics and Policy Studies and the History and Philosophy of Science). He received a PhD in Economics from the University of Michigan in 1979, and is a Director of the Reilly Center for Science, Technology, and Values.

Reviews

[Aaron Arnold · Rating 5 out of 5]

The hidden history of postwar economics, with special attention paid to game theory, linear programming, thermodynamics, and information theory/cryptography. He explores a ton of fascinating linkages between these subjects with a scathing, almost Nietzschean denunciation of how the flaws and assumptions of past economists have turned modern economics into what it is today. However, it's also a celebration of the ideas behind economics, and the ending part where he shows an example of an auction market actually simulating another market is awesome. The amount of research he must have dug through is staggering.

[Humam Fauzi · Rating 4 out of 5]

This is seems like mix of history of post war science and economies. Some of the section is too dense that it require more time to digest. Contains many interesting information about many organization in cold war. Not everyone cup of tea.

[Martin]

Machine Dreams is a text of dizzying ambition and resolution. A particular intellectual history of neoclassical economic (largely micro) theory within the 20th century, with particular emphasis on the relationship it had to war studies, computation, complexity and automata theories, the formalist program, and related topics. But this analysis spreads across domains as much as it does across time, if not more. It has fungal qualities, in the way that Mirowski burrows under the surface to explain some of the intellectual, discursive dark matter connecting various developments in neoclassical theories of price, decision-making, and equilibrium. Godel and the demise of formalist ambitions in the ream of metamathmatics; Shannon and entropy and the concept of information itself and its struggle with major figures shaping the discipline; the western-centric fetish with the Self as primary atomic element of social analysis - and clones; some of the first spreadsheets being fathered by RAND for weapons system cost analyses simulations. None of this is off limits to Mirowski as he tries to contextualize the development of a (I would say) ill-fated, strange machine that is neoclassical economic theory in its various voodoo doll incarnations.

One of the major accomplishments of this book is to be part of the case advocating for re evaluating assumptions of rationality in economic theory. To argue, or open room at least, for the possibility of a heterogenous range of possible rationalities beyond the "Nashionality" of game theory, as he calls it, which is confined to a particular set of rules, from a particular kind of personality. Most undergraduates glaze over assumptions of rational actors - it is obviously silly, but beside the point. Mirowski's geneology herein helps make it painfully obvious through a variety of lines of argumentation specific to this book's context of why this is so. Among these arguements is one made on computational grounds: the neoclassical school was in denial of its image as reproduced (or not) by the encroaching cyborg prostheses it was taking on, like a transplanted organ rejecting it's host rather than the other way around. Various folks wanted to incorporate the computer into economic theory and practice but deny its critiques.

Mirowski uses game theory as a vector of both attack on the foundations of neoclassical thought, and as a first class example of the cyborgian march on the institutions of economic thought. He exhaustively explains the origins of operations research, its various strains, and how those filtered through different institutions - MIT, RAND, Cowles, Chicago - which ultimately became feeding grounds for game theory and related projects. Ultimately the largest throughline of the book is von Neumann's intellectual arc with economics being a potential path of salvation for a properly integrated cyborg science - one which recognizes its own flaws, inclusive of those within game theory itself, and moves forward through properly recognizing computation's critique.

The above are just tidbits of some notes, there's so much in here it's hard to summarize. His final thoughts, which echo earlier and blistering critiques of Hayek and the ontology and nature of markets, are worth pondering and I'm not sure how much subsequent work has been done here, though I think the time seems ripe. Mirowski asks, in the spirit of bringing computation into the fold of economic theory properly: how do we view markets as information processors, but not in the romantic quasi-reactionary way of Hayek? What does it mean to understand markets as a class of heterogenous mechanisms or algorithms - and once we see them that way, to understand their computational complexity, and ask questions of how we should evaluate them? Given current problems and contexts around things like online marketplaces, carbon pricing, power markets, etc - these are interesting questions! But I'm not sure they've been tackled through the von Neumannian automata theory lens that Mirowski was gesturing at.

This book is at times dry (some institutional details are simply lengthy), but never devoid of polemical spunk. The rest of the time it is vertigo inducing but ocassionally deeply enlightening in terms of seeing through a particular way in which neoclassical theory has stumbled forward continuously. Understanding it all - thoroughly - is really only for the select. I'd be very curious to see Mirowski's updated thoughts on the last chapter given the pass of time from then to now, and the enormous increase in the importance of computers - to the point of universality - even since the time this was written.

[Avery · Rating 5 out of 5]

Finished this 5/13/21.
This is my favorite Mirowski book so far. He managed to weave together many disparate subjects in a cohesive manner. I learned a lot about RAND, Cowles Commission, Maxwell's Demon, Computer Science, and many other topics from engaging with this work. I would highly recommend it to anyone interested in these subjects.