Thought Experiment as Ideology: Rawls, Mills, and Ideal Theory 

“The question for political philosophy is not what we should do – but what we should think, even when what we should think makes no practical difference.” 

  - G.A. Cohen

“When reality is depicted, philosophy as an independent branch of knowledge loses its medium of existence.” 

                - Karl Marx 

Introduction

The debate around ideal and non-ideal theory is among the most critical in contemporary political philosophy. It is also incredibly polarizing, and the stakes are quite high: at issue is the extent to which political theorizing ought to be a practical endeavor, and the extent to which real-world facts and phenomena should be factored into the enterprise. John Rawls articulated the ideal/non-ideal distinction well in his seminal A Theory of Justice, which is now commonly seen as a classic example of the ideal theory approach. Here, Rawls develops an account of justice as fairness, specifying the principles which ought to govern the basic structures of an ideal society. To get there, he makes two key idealizing assumptions: one, that the chosen principles/demands of justice would be generally complied with by agents; two, that a society is in reasonably favorable historical and social conditions, such that all its citizens are able to cooperate politically without worrying about famine, war, and similar issues. Of course, Rawls recognizes that the society we live in currently is unjust. Ideal (strict-compliance) theory is meant to provide a systematic grasp of the various injustices we face, and provide some guidance for realizing a more perfect union. Non-ideal (partial-compliance) theory is in the business of finding the principles that govern how we are to deal with injustice here and now. 

The ideal-theoretical features of Rawls’ account have been subject to severe criticism, and from several directions. Certain critics say that Rawls’ account isn’t really ideal theory in the proper sense, since he isn’t in the business of identifying normative truths that are completely devoid of empirical fact (Cohen, pp. 114). Some believe that ideal theory has its merits, but that Rawls’ version is far too abstract or too demanding; indeed, it may not be necessary to know the principles of a perfectly just system in order to recognize the many injustices in our current ones and work on them (Sen, pp. 12). Taking it even further, others have argued that doing ideal theory is not just unnecessary, but rather actively detrimental to the project of rectifying the very real problems we see around us; using ideal standards of justice in the real world may, in fact, lead to more injustice (Mills, TRC, pp. 16). 

The Strong Critique 

That last kind of critique is more serious than the others, as it calls into question the very nature of the kind of normative political theorizing Rawls is up to. Charles Mills and others want to say that doing ideal theory misses the point of political philosophy in an important way, particularly if we want to be in the business of helping solve real-world problems. Idealizing veers away from what should be relevant in political philosophy, and it is therefore an incoherent approach – for, in this domain, relevance and coherence are directly related. 

Whether they acknowledge it explicitly or not, non-ideal theorists are committed to a view about what kind of thing political philosophy is. If they wish to say that non-ideal theory is still recognizably philosophy, the critic must say that the role of a political philosopher is to be directly involved in trying to make people’s lives better. Critics of ideal theory will therefore often draw comparisons between political and moral philosophy. If political theorizing is meant to tell us what to do – they might say – we can look to its closely-related neighbor, ethics, for inspiration. 

While there is plenty of truth to criticisms of ideal theory, non-ideal accounts often lack a precise statement of the exact philosophical convictions that make ideal theory philosophically moot. Such is the case, I would argue, for Charles Mills’ excellent work on the topic – The Racial Contract (along with the correspondence with Tommie Shelby). This shortcoming is the reason why a firm-minded ideal theorist can read Mills’ work and still respond as follows: “Yes, on the face of it, ideal theory does not address the social issues of the present moment; but idealizing seems like the best way to get a coherent grasp of the relevant concepts to strive for, and to orient ourselves to do good. It is certainly still a valid way of philosophizing.” Perhaps coming up with such devil’s advocate responses is the reason why readers like myself have come away from Mills’ eye-opening work wanting more.

In what follows, I hope to overcome these difficulties by presenting a case against Rawls-style ideal theory that explicitly addresses the problematic philosophical moves that underlie it. In particular, I look to the thought experiment as a vessel for the kinds of troublesome features of ideal theory that Mills is gesturing at, but is having trouble identifying as intrinsic to the very method. To this end, I will split the paper up into four sections, arguing for the following points: (1) that thought experimentation is a core component of the Rawlsian ideal-theory approach; (2) that thought experimentation is a core component of any ideal-theory approach in politics, and that it will end up appealing to moral intuitions as opposed to some distinctly political ones; (3) that reliance on thought experiments in political contexts is problematic, and therefore (4) that we should give up on ideal theory as it rests on shaky methodological and conceptual grounds because it relies on thought experiments (which lead its proponents to ignore real-world problems) – and not just because it ignores real-world problems. 

Thought Experiments are Central to Rawls’ Ideal Theory

Rawls, when discussing what makes his approach ideal, only mentions the idealizing assumptions mentioned above: “[In the book,] I consider primarily what I call strict compliance as opposed to partial compliance theory.” (Rawls, pp. 8) In Section 2, he outlines the distinction between ideal and non-ideal theory, and provides a brief justification for choosing the former over the latter: “The latter studies the principles that govern how we are to deal with injustice […] These are the things that we are faced with in everyday life. The reason for beginning with ideal theory is that it provides, I believe, the only basis for the systematic grasp of these more pressing problems.” (Ibid.) He adds that, after one gets to the principles of justice, “the principles define a perfectly just scheme; they belong to ideal theory and set up an aim to guide the course of social reform.” (Ibid., pp. 215) Finally, he states explicitly the lexical primacy of the ideal: “nonideal theory, the second part, is worked out after an ideal conception of justice has been chosen; only then do the parties ask which principles to adopt under less happy conditions.” (Ibid., pp. 216) 

Notice that Rawls justifies the use of ideal theory by appealing to its power as the provider of the only basis for the systemic grasp of our real-world problems. Then, he justifies its primacy over non-ideal theory by referencing the fact that it is able to guide the course of social reform – you derive the ideal conception of justice first, and then move on to real-world concerns. What Rawls fails to acknowledge, however, is the work the Original Position (OP, together with the Veil of Ignorance, VOI) is doing for both of these claims. Indeed, there is no way Rawls gets to his putative systemic grasp of real-world problems – and, indeed, no way he gets to the principles of justice – unless he brings a set of people, armed with basic facts about human life, behind the Veil of Ignorance and into the Original Position. In fact, the only basis for such a systemic grasp just is the combination of the VOI & OP. Because of their lexical relationship, there is also no way he justifies the primacy of ideal theory without appealing to the same methodological tool. This much is quite obvious, and it’s uncontroversial to say that the VOI & OP play a central role in the Theory of Justice. However, acknowledging the absolutely crucial role they play in Rawls’ justifications for using ideal theory instead of its non-ideal counterpart marks an important distinction for our purposes. The reason for this is because Rawls’ key tool is, in fact, a thought experiment. 

Recall that Rawls asks us to imagine ourselves in the position of free and equal individuals who are to jointly agree upon principles of social and political justice. To ensure that our judgements are impartial, we are envisioned as having no knowledge about our socio-economic or historical circumstances in the society (age, gender, race, economic standing, etc.) Since one won’t know where or how well-off they’re going to end up, one will vote on those decisions that advance one’s ends, i.e. accord with the principles of self-interest and mutual disinterestedness (Ibid., pp. 111). 

Clearly, a thought experiment is central to the Rawlsian ideal-theory approach. What about ideal theory approaches in general? The stronger claim is that all such attempts to do political philosophy will involve thought experimentation. The next chapter discusses this claim in detail. 

Thought Experimentation as Intrinsic to Ideal Theory

Part of the reason why non-ideal theories are appealing is because they may be better-equipped to answer the all-important question: what do we do to rectify injustice right now? This is why many, including Mills, are quick to jump on the comparison to normative ethics (Mills, ITI, pp. 166). In his paper, “‘Ideal Theory’ as Ideology,” Mills does a great job explaining the difference between the kind of idealizing that happens when we think about ethics, and the subversive kind that we see in political theory. If political philosophy is meant to help us get to a better state, he says, we must avoid the pitfalls that are unique to the employment of idealization in political theory.

In the article, Mills outlines the features that he sees as characteristic of ideal theory in the political context. These are meant to pick out the things that make this kind of ideal theorizing different from the kinds of idealization we do in considering cases in ethics. The list goes as follows: political ideal theory is prone to assuming (1) an idealized social ontology, (2) idealized capacities, (3) ideal social institutions, (4) an idealized cognitive sphere, (5) strict compliance. Notice that every single category he sees as picking out something unique about ideal theory in the political context is, in fact, picking out the common features of thought experimentation. Certainly, it can describe Rawls’ thought experiments; but we can also mix and match particular features that Mills brings up and compare them to thought experiments from the recent history of ideal theory – Nozick’s Wilt Chamberlain (2, 3, 4, 5; Nozick, pp. 161), Singer’s Drowning Child (2, 4, 5; Singer, pp. 14), etc. Of course, we have yet to justify whether any of these idealizations are bad or incoherent – this is the goal of the next section. The point, for now, is this: if Mills means to suggest that there is a way in which idealizing in ethics (permissible) is different from idealizing in politics (not permissible), and that these features mark all the difference between two forms of idealizing – he narrowly misses the point. They pick out thought experiments in political theory. 

It seems that we cannot think of an ideal political theory that doesn’t involve any of the elements Mills brings out, by his own definition. The next question to answer is whether we can think of a political thought experiment whatsoever that does not involve them. Rawls’ experiment is certainly not a good answer, because Mills created his list partly by reference to TOJ. For my part, I am unable to think of an example or come up with a new one that is devoid of these elements. There is no room here for an exhaustive list, but we may conclude, somewhat hesitantly, that it is impossible to do political ideal theory, in the sense Rawls means, without doing a political thought experiment. This may be an observation more than a strong conclusion – but it certainly seems difficult to conceive of some idealized political theory without idealized thought experimentation. Thought experimentation, therefore, seems intrinsic to ideal political theory. 

Rawls, like many other ideal theorists, assumes that thought experimentation is completely fine as a methodological tool. This is a controversial supposition in general, and Mills helpfully gives us a list of reasons why; these will be discussed in the next section. To provide a more fleshed-out account of the issues of relying on thought experiments in political theory, it is also worth discussing why ideal theorists might still use them, even if they acknowledge the aforementioned features.  

The reason ideal theorists do not doubt the import of the thought experiments they engage in might have to do with believing that there is an important difference between moral and political intuitions. Using Rawls as an example, the underlying idea is that the conclusions derived from the Original Position are different from those coming from a standard counterfactual moral thought experiment, because the OP appeals to a completely different notion. This would be important for Rawls in particular because he wants to say that any arbitrary humans (idealized in the way he describes) would reach the two principles of justice if put through this thought experiment. With the Veil of Ignorance, Rawls deletes all potential sources of interference, and wants to suggest that idealized humans would - using their blind political intuition - reach the same conclusion. 

To see why Rawls’ thought experiment does not produce a political intuition which is distinct from a moral intuition, we only need to unpack the kind of decision-making that would take place in his experiment. In the Original Position, the recourse to the normative realm is implicit, and it may not be immediately obvious. To an individual behind the veil of ignorance, a crucial concern is how they should act or make decisions. But, since you don’t know where or how well-off you’re going to end up, you should vote on those decisions that advance your ends, i.e. accord with the principles of self-interest and mutual disinterestedness. These ends are what is arrived at by moral intuition and deliberation. It is clear that said intuition is of the moral kind, as it is called upon in response to an inquiry about how one should act. In fact, Rawls can say that the principles of justice will be the result of the deliberative process with such conviction precisely because the original position collapses what an agent should and would do into one, such that rationally-behaving individuals are bound to come up with the principles of justice (Rawls, pp. 10).

Of course, Rawls may not think this at all, believing that the conclusion of the thought experiment is valid because it was predated by a leveling of the actors’ epistemic playing fields. I used the OP to make the main point clear by example: a putatively distinct political intuition is just moral intuition with extra steps. Looking to other thought experiments in political theory (like Nozick’s Wilt Chamberlain; Nozick, pp. 161), it will be clear that they must include an appeal to moral intuitions somewhere along the way. What this means is that concerns regarding regular moral intuitions, ones that are commonly charged against thought experimentation in ethics, also apply here. I explore these claims in the next section. 

Reliance on Thought Experiments in Political Contexts is Problematic

Relying on thought experiments in political philosophy – an approach we may now identify as ideal theory, in the Rawlsian sense of the word – is problematic for two sets of reasons. These are (1) the reasons why Mills believes that the ‘features of political ideal theory’ are problematic, and (2) the reasons why relying on moral intuitions is problematic in general. Since there is significant overlap between these two sets, they will be discussed in one fell swoop. 

One broad concern about thought experiments that veer into the normative realm goes as follows: Because a thought experiment allows us to transcend the limits of our comprehension of the real world, the use of hypothetical situations in normative and political theory brings up some serious concerns. Since the domain explicitly engages with the kind of things we ought to do, we want to apply the findings from our imaginary experiments into real life. But, the more we distance ourselves from the world, the more we will need to clarify the hypothetical scenario’s links to reality. Indeed, if our thought experiment is completely out of tune with the reality to which it is speaking, we must be able to discard it as useless. As Stephen Mulhall puts it, “what if the issue means what it does to us, has the moral significance it has for us, precisely because of its place in that complex web [of interrelated matters of fact within which we usually encounter and respond to it]?” (Mulhall, pp. 5). 

Mills echoes this point, but in a slightly different fashion. Broadly speaking, he says, what distinguishes ideal theory is the reliance on idealization to the exclusion or marginalization of the actual. This is not intrinsic to the operation of abstraction in general, as is clear from certain kinds of moral deliberation. We assimilate this to a concern about thought experimentation; the idealizing aspect often serves to work against inclusive, pragmatic concerns that ought to guide our action. This relates to Mills’ four points: 

(1) Idealized social ontology. Indeed, because moral and political theory deal with normative concerns, they cannot avoid some characterization of the human beings that make up the society. And this is cause for trouble; idealizing the social ontology will typically assume the “abstract and undifferentiated equal atomic individuals of classical liberalism,” and thus “abstract away from relations of structural domination, exploitation, coercion, and oppression, which in reality, of course, will profoundly shape the ontology of those same individuals, locating them in superior and inferior positions in social hierarchies of various kinds.” (Mills, ITI, pp. 168)

The points about (2) idealized capacities and (3) an idealized cognitive sphere are cogent here as well. Ideal theory often gives agents unrealistic capacities (Rawls’ strict compliance is a great example). Those who are theorizing will most likely not recognize the consequences of oppression for the social cognition of these agents; both the advantaged and the disadvantaged, will typically not be recognized, let alone theorized – “a general social transparency will be presumed.” (Ibid., pp. 169) 

Finally, (4) social institutions are often conceptualized into ideal models; things like the family, the economic structure, the legal system are imported into the idealized world with little or no sense of how their actual workings may “systematically disadvantage women, the poor, and racial minorities.” (Ibid.) 

A related concern has to do with the trust we place in our intuitions in general. Of course, the intuitive judgements that we form in response to political thought experiments provide credible initial evidence to support or refute certain political propositions (general principles, issues about legitimacy, political authority, etc.). But the discussion of the epistemically relevant features of the moral intuitions generated by political thought experiments is radically undermined by the kinds of criticisms that follow. Indeed, why would we treat our swift, spontaneous judgements as providing anything but evidence into how our mind works when it is faced with an unfamiliar counterfactual? 

It certainly seems like the reasons against trusting moral intuitions outweigh the reasons for doing so. First of all, there is a widespread disparity in the intuitive answers to particular moral thought experiments, and this certainly translates to ones in political theory. One would hope that an experiment is something one could count on, since you want to extract valuable information from its results. But in political thought experiments, more controversial examples will certainly produce equivocal results. Secondly, it seems that our intuitions are very sensitive to morally arbitrary features. What I mean by this, for one, is that the particular way a political thought experiment is phrased may produce dramatically different results – leaving certain terms vague, changing the order in which points are presented, rephrasing certain segments, etc., are all methods of manipulating or destabilizing moral intuitions. What’s more, how one feels at a particular time, whether they are hungry or tired, etc., are all arbitrary factors that can have a severe impact on the intuition itself. If you’re really aggravated one day, you would be more likely to say you’d push the large man off the bridge and into the path of the trolley than on a more normal day. Along similar lines, one’s biases, including those entrenched in us by our culture and upbringing, will inevitably be reflected in intuitions. Ultimately, there does not seem to be anything in the way intuitions feel that ought to give anyone confidence in their putative epistemic authority, especially in the face of these criticisms. 

Near the end of his discussion, Mills says: “if we start from what is presumably the uncontroversial premise that the ultimate point of ethics is to guide our actions and make ourselves better people and the world a better place, then the [five features] will not only be unhelpful, but will in certain respects be deeply antithetical to the proper goal of theoretical ethics as an enterprise.” (Ibid.) I concur that we may uncontroversially substitute “thought experimentation about political matters” for “the framework above”. In fact, we can do the substitution in order to appropriate the title of his article for our needs – “‘Ideal Theory’ as Ideology” becomes “Thought Experiment as Ideology.” Indeed – if the outcomes and presuppositions of thought experiments depend on who creates them, who engages with them, the background conditions they bring into the imagined world – the hermeneutical concerns Mills raises are made even more powerful, precisely because of the way they interact with the dubious epistemic nature of thought experimentation in general. 

Giving up On Ideal Theory – What Next? 

If what was said in the foregoing section is correct, it seems that we have found a way to bolster the kind of deep-cutting criticism of ideal theory that philosophers like Charles Mills have tried to push forward. This kind of criticism, for all its insightful assessments, does not attack ideal theory where it really hurts: as we saw with Mills’ criticism above, the focus is on the hermeneutics of ideal theory, or on the kinds of philosophers that tend to do it. For what it’s worth, one may think his critique provides sufficient reasons to give up ideal theory in its own right – Mills’ description of the very real consequences that ideal theorizing tends to have on marginalized communities, including its tendency to obscure or implement structurally oppressive structures into the ideal realm, are incredibly cogent. But ideal theorists might still think this criticism is not good enough reason to stop doing that kind of philosophy; indeed, they may heed Mills’ criticism and attempt to do ideal theory without creating such omissions. The central purpose of this essay was to deny the ideal theorist this possibility, and say that this method of philosophizing is intrinsically predisposed to such omissions. 

The way I tried to show this of ideal theory in general is by first utilizing Rawls as an example: his justification for ideal theory, as well as his definition thereof, relies centrally on a thought experiment – the Original Position along with the Veil of Ignorance. The centrality of thought experimentation to this particular ideal-theoretical approach was telling of a broader issue, and our broadening of the inquiry to ideal-theory approaches in general proved fruitful. This is where Charles Mills’ account came in – in “‘Ideal Theory’ as Ideology,” he outlined what he saw as the core features that distinguish ideal theorizing in politics from ethics. This was a misdiagnosis – what Mills was picking out was, in fact, a set of features that distinguished political thought experiments. This gives one enough reason to suspect that ideal theory, in the broad sense Rawls uses, is intrinsically dependent on thought experimentation. After this, it is easy to criticize the method known as ideal theory on the same grounds that we may criticize moral thought experiments on. A potential roadblock is the notion that there exists a political intuition that is distinct from a moral one; we briefly showed that political thought experiments must appeal to the normative realm, and therefore trade on moral intuitions. The list of reasons for suspecting such thought experiments is long, and integrates general concerns with those hermeneutical ones Mills describes. Ultimately, thought experimentation leads to dubious conclusions as it abstracts from the real world, relies on the limited epistemic value of intuitions, and is otherwise problematic. Taken holistically, I hope to have shown that ideal theory – because it relies on political thought experiments – is intrinsically on shaky grounds. This achieves the aim of bolstering Mills’ account, which might be pessimistically seen as saying that ideal theory is on shaky grounds for non-intrinsic reasons that may be overcome by the right kind of ideal theorizing. At worst, I will have shown that Rawls’ reliance on a thought experiment is dubious, and this is good grounds for arguing against his particular ideal theorizing. 

With that recap in mind, we move on to a final reflection on the state of political philosophy. Rawls says, at the end of part 2 to the Theory of Justice: “If ideal theory is worthy of study, it must be because, as I have conjectured, it is the fundamental part of the theory of justice and essential for the nonideal part as well.“ (Rawls, pp. 343) Against Rawls, if thought experimentation is the fundamental part of ideal theory – a claim which casts serious doubts onto the legitimacy of the very enterprise of ideal theorizing – where does that leave the desire for a theory of justice?

The biggest shortcoming to most critiques of ideal theory, I believe, is their lack of a positive account for what non-ideal theory is, how we ought to do it, and why. Mills provides a little bit of information in this light in the most concrete parts of the Racial Contract. However, he comes far from answering the aforementioned key points. What we do know, however, is that ideal theory is intrinsically dubious. So, if we do want to find a theory of justice, we may want to try and find a non-ideal theory of justice. In closing, I want to suggest that we do not need to find such a thing in order to do non-ideal theory, and to make a real difference. To practice political philosophy after ideal theory, we might say, it is insufficient to examine our ideas and fundamental concepts like justice, and see what they tell us. What is required is for philosophy to collaborate with other forms of inquiry and practical organization, whose chief aim is to question and transform oppressive social relations. This does not imply a pursuit of struggle for its own sake, without conceptual grounding; instead, we may look at philosophy as a field that is subordinate to the needs of social struggles. Recognizing ideal theory as fundamentally opposed to this goal is one thing, but noting that it is intrinsically incapable of achieving such goals because it relies on thought experiments is the stronger claim. Justice, as a thing in the world, requires much more than a conceptual explanation; it requires discovering its validity conditions practically, by challenging oppressive power structures all over, and with others’ help. Philosophy, seeing the shortcomings of the alternative, must adapt its intuitions to the real-world situations we find ourselves in. Political philosophy in particular ought to be rooted in a confrontation of idealization and abstraction; it thus becomes helpful, if not tantamount to, activism.