Political theorists developed a constructivist approach to political representation around 2010; Saward (2010) seems to be the canonical citation for this approach, but I think M. B. Brown (2009) is doing something quite similar. Prior to 2010, work by political theorists on representation focused on either (a) formal procedures of elections and voting or (b) deliberative practices, either in civil society (that is, outside of government) or in boundary institutions that connected public deliberation to policymakers. Recently several philosophers of science have picked up ideas from the latter (citations in a previous post).
Constructivists challenge an implicit assumption of these approaches, namely, the there is a constituency or public that exists prior to representation and has more-or-less well-defined interests or preferences. This assumption might (might) make sense when we’re talking about legislative districts: the State of California existed before Diane Feinstein was first elected Senator in 1992. But it doesn’t work as well for less formal types of representation, for example, Martin Luther King, Jr., or Greta Thunberg as representatives of social movements.
Saward (2010) apparently (I haven’t actually read the book) draws on Gricean philosophy of language that will be familiar to most philosophers today. Specifically, Saward (2010) examines representative claims as speech acts, claims that the speaker represents a certain public or constituency to an audience in a certain context and that this constituency believes or wants or does something. These acts are successful insofar as they are taken up by the audience. In other words, to be a representative for a constituency is nothing more or less than to be taken as a representative for that constituency.
Importantly, the constituency does not need to exist for a representative claim to be successful. One practical effect of the speech act might be that the constituency comes into existence as members of the audience identify themselves as members of that constituency. For example, what we now call the “MAGA movement” or “MAGA Republicans” didn’t exist until after Trump presented himself as a representative for that then-unnamed constituency in 2015. Drawing on Hobbes, M. B. Brown (2009) calls this virtual representation1, and he uses it to develop an account of (natural) scientists as representatives of nature in policy contexts. Even if it makes sense to talk about “nature” as a single coherent entity, that entity can’t do something like authorize a scientist as its representative through a formal election or have beliefs or interests that are accurately or inaccurately represented by a scientist’s representative claims.
Urbinati (2018) discusses this constructivist approach to political representation. I found her writing somewhat difficult to follow, so I’m not going to try to fully reconstruct her views. But she does seem to present three objections to constructivism:
- It neglects inequalities in who will make representative claims and whose representative claims will be taken up.
- The representative is accountable to the audience rather than the constituency.
- There’s no reason to think that successful representative claims (as constructivists understand that success) will have any effects on policy. (Urbinati 2018, 74ff)
The first objection echoes Young (2000)’s critique of deliberative democracy, which (prior to Young’s critique) tended to focus on idealized deliberative scenarios and norms of deliberation that favored highly educated white men. The third objection seems to provide the best support for Urbinati (2018)’s “diarchic” approach, which emphasizes the relationship between formal institutions of political power and the open-ended deliberation of civil society. (I suspect some of the later contributions to this collection will do this as well.)
On the “pre-constructivist” models of representation, representatives are primarily accountable to the constituency that they represent. This might be done formally — the representative needs to stand for (re)election — or it might happen informally — as constituents call representatives to give an account of their actions qua representatives.2 Urbinati’s point seems to be that, by basing legitimacy on audience uptake, constructivists lose the demand for democratic accountability. Audience uptake matters, and representative and constituency co-construct each other (again, consider Trump), but good representatives must still represent their constituents accurately.
For M. B. Brown (2009), a constructivist account of scientific representation (scientists as representatives of nature) is attractive because democratic accountability to “nature” seems to be nonsensical.3 So if scientists are to be held accountable to anyone besides themselves, the audience seems to be the only option.
In the context of Brown’s account, we might understand Urbinati’s second concern as a critique of “wishful thinking” or “anything-goes” relativism: nothing seems to be stopping the audience from rejecting scientists whose claims are inconvenient or unfavorable (no matter how well-supported), and accepting scientists whose claims are convenient and favorable (no matter how flimsy) (Sarewitz 2004; Steel 2017; Lichtenstein 2022).
One constructivist response to this critique might be to reject the notion of accuracy, as in the formulation “good representatives must still represent their constituents accurately.” Such a constructivist might argue that “accuracy” assumes a pre-existing constituency; but representative claims are prior to constituencies, etc. On the science side, this leads to the interminable and fatally abstract debates over “realism” of the 1970s and ’80s.
An alternative approach might draw on Dewey’s account of inquiry, specifically as articulated by M. Brown (2020). On this account, inquiry starts with a problematic situation — things aren’t working — and inquiry is successful insofar as it resolves the situation — develops alternative things that do work. Importantly, “working” here has both “realist” and “constructivist” aspects. It’s realist insofar as the agent can incorrectly believe that a certain alternative will work when “in fact” it will not. But it’s also constructivist insofar as one tactic to resolve the situation is for the agent to change their understanding of the activity and its goals, and resolving the situation does not require a “true” representation of the world (in a classical correspondence sense).4
I suggest we can view political representation as “inquiry” using this account. The problematic situation is that some people5 have some felt dissatisfaction with the status quo. A representative claim is then a hypothesis that there is a group (the constituency) whose interests are not being served by the policy status quo (the diagnosis), and that adopting a different policy will resolve the problematic situation (the remedy).
On this model, a representative claim will have locutionary success criteria, that is, it needs to be taken up by the audience, especially people with the power to enact the proposed policy. And the constituency can be constructed by the representative claim, insofar as the “some people” subsequently decide that they are indeed members of the group and their interests are not being served by the status quo. But the representative claim also has pragmatic success criteria: would the remedy actually resolve the problematic situation as characterized by the diagnosis? This is unlikely to be the case if, for example, there is good evidence that the felt dissatisfaction experienced by group members is not caused by the factors identified in the diagnosis.
Consider how this model might work for Trump’s 2016 campaign. We can reconstruct the central representative claim of Trump’s campaign as something like the hypothesis that rural poverty is caused by mass undocumented immigration and so strong immigration restrictions will solve this problem. This claim was taken up by the media, who did numerous profiles of “Trump voters” (the newly-constructed constituency) in impoverished rural areas and their attitudes towards immigration. So it was evidently a locutionary success. But its pragmatic success is tenuous at best; agricultural consolidation and demographic shifts (depopulation and aging as young adults go to college and move to urban areas) are probably much more important causes of rural poverty than immigration, undocumented or otherwise.
References
Footnotes
Or maybe “fictional representation”? Need to check.↩︎
Young cite here?↩︎
Yesterday I was preparing Simpson (2014) as a reading on Indigenous epistemology and land relations for the Environmental Philosophy course that I’m teaching this semester. From the perspective of Indigenous land relations, accountability to “nature” — or rather, nonhuman agents — does make sense. Indeed, as presented by various Indigenous authors (Simpson 2014; Liboiron 2021) a key idea in Indigenous knowledge-making practices is consent and collaboration with nonhuman agents. I’m going to bracket this thought for the purposes of this post.↩︎
Sorry Matt Brown, nobody else thinks Dewey was a correspondence theorist.↩︎
This should be understood as a Boolosean plural quantifier: “there are some \(x\)s,” rather than “there is a set \(X\).”↩︎