Power, Knowledge, and Anarchism

ABSTRACT

While Jeffrey Friedman’s Power Without Knowledge offers a welcome corrective to the technocratic statism that dominates modern politics, Wittgenstein’s view of language suggests that the problem of ideational heterogeneity is less worrisome than Friedman maintains. In addition, Friedman’s “exitocracy” is as epistemically demanding as ordinary technocracy and thus cannot provide an alternative to it. Anarchism, however, might provide a more consistent alternative to technocracy.

Keywords:

• anarchism
• anarchy
• positivism
• technocracy
• Wittgenstein

Notes

1 This amounts to saying that, in Friedman’s terms, we might have good reasons (not reducible to the question of technocratic knowledge) to grasp the “second horn of the [technocratic] dilemma” (316) and embrace people’s right to govern their own affairs despite their potential ignorance. We might, that is, decide that they are entitled to “power without knowledge.” This conviction is consistent, however, with the idea that we should seek forms of political decision making that are maximally responsive to the reality of pervasive ignorance and uncertainty.

2 Though I will not explore this further here, a puzzling aspect of Friedman’s position on ideational heterogeneity is his claim that, because belief formation is deterministic, idiosyncratic ideas are “not random” and thus can’t be dealt with through “probability theory” (144). I found this assertion confusing. Calling variation “random” for purposes of statistical analysis implies only that the variation in question is not systematically correlated with the variable(s) of interest, not that the variation is uncaused. The fact that ideational variation has deterministic causes, then, does not entail that it cannot be dealt with through statistical methods. I would have liked to see this point elaborated in the text, as it seems important to Friedman’s argument. If idiosyncratic variation can be dealt with through statistical methods (as I suspect it often can), this would offer an additional reason for optimism about behavioral prediction.

3 Rorty here associates this view with Quinean holism. Given that the “web of belief” idea on which Friedman draws is also of Quinean origin, Friedman might have reason to be committed to a similar view.

4 This is the kind of thing that philosophers of science have explored through the notion of a “paradigm” (Kuhn), “thought collective” (Fleck), or “sociologic” (Latour). Friedman himself explores similar ideas in his discussion of “spirals of conviction,” but does not consider whether such dynamics might apply to ordinary linguistic communities as well. Latour (1987, ch. 5) explicitly discusses scientific paradigms as paralleling ordinary linguistic paradigms.

5 Friedman holds that while first-order knowledge can be drawn on to determine, “case by case,” whether particular homogeneities exist, it cannot provide evidence that heterogeneity in general is not worryingly pervasive. Thus, on Friedman’s view, no amount of successful behavioral prediction can alleviate the worry posed by the heterogeneity thesis (which is a purely second-order consideration).

6 This seems to be an important (and perhaps unjustified) shift in Friedman’s argument. While earlier he claimed not to know whether ideational heterogeneity would be consequential for behavior, at this point in the argument he seems to be claiming such knowledge. It’s not clear how the fact that no two webs of belief are identical establishes behavioral heterogeneity as a “rule.” It’s hard to see how Friedman’s modest heterogeneity thesis—i.e., that no two webs of belief are identical—could give rise to such a conclusion. Whether such ideational heterogeneity actually leads to behavioral heterogeneity would seem to be unknowable at the second order.

7 A problem here might be that in these passages Friedman only considers experimental inquiry in a rigidly positivist vein (i.e., as focused exclusively on detecting correlations, with no theoretical underpinnings whatsoever). It is hard to imagine that real experiments are ever carried out in this way. What could possibly motivate one to carry out a field or policy experiment absent a theory guiding the specific choices of treatment administered and outcomes measured? Friedman (48, 208) himself makes similar observations. While his criticisms of a purely “positivist” experimentalism are certainly persuasive, I find it hard to imagine that this extreme positivist position is pervasive among social scientists using experimental methods. The text he cites in defense of this characterization (Green et al. 2010) considers reasons why definite causal identification is difficult in experimental social science. It claims that dealing with complex questions of “causal mediation” through experimentation is often very challenging and that strong causal claims should be scaled back. Admitting that experimental evidence cannot directly prove a specific causal theory, however, is not equivalent to saying that no theory should be guiding the generation of hypotheses (nor that consideration of people’s ideas should not be part of the process of hypothesis formation). Friedman’s criticism seems to conflate these positions.

8 For a general analysis of wealth taxes that highlights some of these issues, see OECD 2018.

9 Though Friedman's endorsement of competition is not explicitly Hayekian and he is in fact quite critical of Hayek's epistemology, Hayek's observations about the role of technocratic knowledge in determining where competition can be “made to function” still seem germane.

10 This seems especially appealing in light of Friedman’s admission that even what we consider to be “natural” arenas of exit, “such as consumer goods and labor markets,” are in fact dependent on technocratic judgment (342). Many anarchists have held that the specific understandings of property and law dominant in the modern world are skewed towards the interests of the wealthy (which the state tends to privilege). Given the reality of disagreement concerning even these basic components of “exitocracy,” it seems that maximally consensual arrangements regarding these institutions would be preferable to their technocratic imposition.

11 Some anarchists also believe that consensual property regimes would restrict the amount (or type) of property that one could justifiably accumulate, but there is no consensus on this issue.