In this post, I outline a few notes on Scott’s Two Cheers for Anarchism ahead of a discussion with Ryan and Jesse.

What is Anarchism About?

Scott endorses what he calls a “anarchist squint” as being useful in various aspects of politics and political analysis. The idea is that aspects of anarchism are both useful and more common than they are often given credit for. We gain something by looking at the world through anarchist glasses.

A few anarchist principles that Scott highlights:

  • An emphasis on “cooperation without hierarchy or state rule”
  • A “tolerance for the confusion and improvization that accompanies social learning”, and “confidence in spontaneous cooperation and reciprocity” (xvii)
  • A defense of politics / conflict / debate (and their attendant uncertainty) against de-politicized administrative insitituions & systems based on “objective” science and modes of assessment.

Interestingly, Scott believes that “abolition of the state is not an option” (xvi); the state is necessary as a tool to have any chance of ensuring the “relative equality” that is necessary for mutuality or anything like democracy (though this is not explored deeply in the book).

Still, as discussed further below, Scott is very pessimistic about modern institutional democracy, and believes that – as a historical point – major episodes of structural change have primarily occurred in moments where there was substantial non-institutionalized threat to the order of things: widespread disorganized protest, looting/property damage, and other forms of disruption. In the historical narrative, these examples get accommmodated as cases of the democratic system “working”; however, it is really the rising threat of extra-institutional disorder that is driving change within the democratic system. Scott further argues that social movements paradoxically tend to have the most institutional success when they are at their “most disruptive, most confrontational, least organized and least hierarchical” (xviii) – i.e. when they pose the greatest threat to institutional order. In short: disorder plays an important political role in social change.

On the Logic of Extra-Institutional Action

As discussed above, Scott is very pessimistic about the ability of institutional democracy to achieve what it promises. Here is a relevant quote (Preface, page xvii):

the very insititutions designed to avoid popular tumults and make peaceful, orderly legislative change possible have generally failed to deliver. […] Episodes of structural change, therefore, tend to occur only when massive noninstitutionalized disruption in the form of riots, attacks on property, unruly demonstrations, theft, arson, and open defiance threatens established institutions.

One thing that Scott’s pessimism clarified for me is how the rational calculus of extra-institutional political action (protest, organizing, disruption, mutual aid etc.) has to do with two distinct but related assessments about the political environment around a particular issue:

  1. (An assessment of) how radically I would like to change the status quo policy.
  2. (An assessment of) how likely institutional action is to be successful in addressing my desired policy change.

Of course, these two assessments are related. If I am very radical (far from the “status quo”) in my policy preferences on an issue (1), it’s presumably rational for me to be more pessimistic about the system’s likelihood of accommodating my changes (2); hence, on both accounts, extra-institutional action becomes more appealing.

However, what Scott’s discussion clarified for me is that radicalism (1) is far from necessary to rationally justify extra-institutional action if one is sufficiently pessimistic about (2). In other words, even if I don’t want to change the status quo very much (i.e. I am not all that “radical” in some objective sense), if I am very pessimistic about the workings of the modern insititutional political system (e.g. because of inequality, the Electoral College, campaign finance rules, gerrymandering, or whatever), then extra-institutional action quickly becomes far more appealing. This logic extends further: even if I am, say, 50% confident that we can solve whatever problem (e.g. climate change) within institutional bounds, then this should still probably orient me more positively toward extra-institutional action; after all, there is still a 50% chance the system won’t be able to handle it…

The Idea of Infrapolitics

An interesting theme in the book is Scott’s discussion of “infrapolitics”:

“By infrapolitics I have in mind such acts as foot-dragging, poaching, pilfering, dissimulation, sabotage, desertion, absenteeism, squatting and flight” (xx)

Scott believes that these forms of infrapolitics are ubiquitous (especially among groups that otherwise lack political power), consequential, and “embody [a] mutuality without hierarchy” that resonates with anarchist perspectives:

“The accumulation of thousands or even millions of such petty acts can have massive effects on welfare, land rights, taxes, and property relations. The large mesh net political scientists and most historians use to troll for political activity utterly misses the fact that most subordinate classes have historically not had the luxury of open political organization. That has not prevented them from working microscropically, cooperatively, complicitly, and massively at political change from below” (xxi)

To summarize Scott’s argument about infrapolitics:

  1. “Infrapolitical” actions like foot-dragging, poaching, desertion etc. are political (or can be), even though they are often not coded as such.
  2. Infrapolitical activity of this form is historically ubiquitous (especially among groups with limited formal political power) and consequential.
  3. Despite this, infrapolitical action is rarely canonized/narrativized as consequential in history books.
  4. This result (3) is not incidental but rather intentional on behalf of the political actors engaged in these behaviors.
  5. Political actors engaged in these behaviors are often doing so precisely because they lack the political power / luxury of open rebellion or protest.

I can think of various examples of how this type of infrapolitics emerge in digital environments as forms of resistance to dominant platforms. For example, I think of gig workers that form groups to exchange tips about the best ways to game platform incentive systems. Roughly: various ways to “follow the rules” on paper without really following them. The recent (so-called) epidemic of “quiet quitting” is another example. Interestingly, contra Scott’s argument, quiet quitting was widely baked into high-level narratives about what was happening in the labor market at the time; see, for example, this New York Times article.

On Flexible Use and Bottom-up Design

A big focus in the middle of the book is on “top-down”, rigorously-organized, forms of design and planning vs. “bottom-up” modes of design. For example, Scott discusses the case of institutionalized forestry (also coverd in Seeing Like a State), which wound up being highly brittle despite being rigorously-planned and visually-ordered. Scott also discusses the design of “open” style playgrounds that assume little about the interests and creative pursuits of users vs ones that assume a specific mode of play.

On the topic of bottom-up system/institutional design, he offers a normative standard:

“Almost any human institution can be evaluated in these terms. How open is it to the purposes and talents of those who inhabit it?” (60)

A similar standard could be applied to the design of technological artifacts: how much is this artifact open to the purposes and needs of those who will use it? How much does this artifact’s design assume about how it will be used?

What is Scott’s basis for preferring bottom-up design? It is a mix of descriptive and normative points. On the descriptive front, Scott seems to be arguing that bottom-up systems/technologies are:

  1. More robust, adaptable, and flexible in the face of change. (less brittle)
  2. Better able to take advantage of or respond to the unique skills / talents / needs of those who use them (and hence: more effective).

From a more normative perspective, bottom-up systems are preferrable because:

  1. They enact a desirable humility about the needs/goals/desires of users.
  2. They grant users more control over their engagement with the system/artifact.

I want to think more about these themes. There are clear connections to philosophy of technology ideas like the designer fallacy; these themes also connect to conversations that Ryan and I have had about the value of modularity in technological design.

On Audit Society & Quantitative Measures

In chapter five, Scott expresses thoroughgoing skepticism about the use of “quantitative measures of quality” in a range of policy contexts (e.g. the SAT, standardized testing, SSCI). These points resonate with (and rehash) a lot of other STS work, including Porter (who he quotes), and Scott’s own Seeing Like a State.

What are Scott’s main points about these quantitative standards? First, two obvious points:

  1. Quantitative standards are often inaccurate, limited, and biased/political in the first place. Inevitably, any quantiative measure cannot capture all aspects of the broader qualitative concept like “merit” that it purports to represent. The choice about what aspects to “count” (and how) is a political one.
  2. Even if a quantitative standard is not inaccurate to begin with, it will become so once used as a regulatory device due to the endogenous response of those being regulated – i.e. “gaming” the system, teaching to the test etc. Essentially, Goodhart’s law.

More deeply, Scott emphasizes that quantitative assessment systems are “de-politicizing” machines. They are tools for taking sets of issues off the democratic stage of political discourse/deliberation/disagreeement– and instead delegating their management to expert technocrats and/or their technical tools (like tests or audits). According to Scott, the presence of a seemingly neutral, “objective” standard like the SAT neuters the impetus for democratic political debate about the meaning of qualitative concepts like “intelligence” or “merit”.

Scott’s points about quantitative measures are well-articulated, but have been made before by both Scott and other STS scholars (and can also be applied more broadly to many types of technical “knowledge”). What is novel about Scott’s articulation of these points is the connection with anarchism. What is the nature of this connection? A few thoughts:

  1. Quantitative assessment systems are often deployed hierarchically–they are most preferred by managers or states that are furthest from the reality of what’s happening “on the ground”. They are fundamentally tools of administration, management, hierarchy in some sense (a point I also discussed here).
  2. There is a connection with the broader anarchist preference for “messy” localism, deliberation etc. over tidy, top-down organization.

I wish Scott engaged more with the (non-managerial) benefits of clear quantitative standards. Clear quantitative standards for judgment–even if imperfect–do have an aspect of making decisions “more fair”: they remove arbitrary power & discretion (which might be influenced by all sorts of things, including racial or gender bias) from decision-makers and put it back in the hands of those being assessed. To this point, Scott admits that e.g. excising the SSCI from tenure debates leaves more space for other sorts of bias and power dynamics to enter decision-making; however, he seems to view this as a necessary evil.

For my part, it’s not clear that we ought to do away with all systems of quantitative assessment; instead, combining fairness considerations with Scott’s arguments about de-politicization point towards a middle path: recognizing the value of these systems while also keeping them open for democratic debate; not naturalizing (say) the SAT as the be-all-end-all measure of merit, but holding it loosely in a pragmatist sort of sense, recognizing it is useful but also has limits, and is always up for re-assessment.