Responsibility-Based Collectivist Arguments from the Discursive Dilemma

It is one thing to establish that in certain situations collectives face discursive dilemmas, it is another to use discursive dilemmas to show that collectives could in some cases be collectively responsible.

21.5.1 The Simple Argument

Imagine that a premise-based decision procedure is in place. Imagine further that the decision taken is morally objectionable because it is unfair or harmful. Finally, imagine that the distribu­tion of individual judgments or votes is exactly as in Table 21.1 above.¹⁷ In such a case, who can be held responsible or blameworthy for the morally objectionable decision? If you have a com­plaint against the decision whom can you remonstrate with? Each member can point out that they were against the decision, and further, that none of them could foresee what decision their votes would lead to (see esp. List & Pettit 2011, 216). Therefore, the only appropriate addressee seems to be the collective as a whole.

Note that the underlying motivation for attributing responsibility to the group is similar to other responsibility-based arguments for collective responsibility surveyed above in section 21.3.2. The decision is objectionable because it is unfair or harmful. However, no individual seems to be at fault. So, unless we ascribe responsibility to the group as a whole, there will be a “deficit in the accounting books” (Pettit 2007a: 194).

On closer scrutiny, several objections arise, however. Granted, those finding themselves at the losing end of the decision may feel bad about the outcome. However, first, it is question­able whether the decision is morally wrong in the sense that it involves the violation of a moral norm or obligation.

To take an example where the moral implications are more dramatic than in the case of a tenure decision, consider a panel of three deciding whether to grant parole. Let us stipulate that rules for granting parole determine three (and only three) conditions of eligibility for parole: the felon’s criminal history and behavior while in prison must be satisfactory, the risk of granting parole must be low (e.g., no strong indicators of a tendency to reoffend), and suitable accommodation upon the prisoner’s release must be assured. Let us further stipulate that the panel uses the premise-based procedure. The distribution of votes is as follows:

Table 21.2 The Parole Board

The prisoner is released because the premise-based procedure was used. And now suppose that he commits another crime while on parole.

However, it is unclear that releasing the felon was morally wrong — despite the unfortunate outcome of the decision. Parole decisions always carry a significant risk of doing harm: either to the felon, in cases where she is kept locked up needlessly, or to the felon’s victims, in cases of recidivism. So, the mere fact that the decision turned out to have regrettable consequences does not mean it was not justified as long as the risks were assessed with sufficient care. After all, many of our justified beliefs turn out to be false, sometimes with disastrous consequences.

Of course, it is possible that the risks involved have not been assessed by the individual members with sufficient care. In that case, these individual judgments should have been different and releasing the prisoner was indeed the wrong thing to do. So, in that case, the individual members could certainly be at fault. However, that conclusion is obviously not the one the col­lectivist is out to establish.¹⁹

By contrast, if individual judgments were as careful as they could be, then the only way the decision can be unjustified is if there is something wrong with the decision procedure used to aggregate them. And it is true that in this particular case the conclusion-based procedure would have been in some sense preferable, at least ex post. If the board had used the conclusion­based procedure, parole would have been denied and the criminal would not have committed another crime.

However, it does not follow from the fact that the premise-base procedure yielded a harmful outcome that its use was unjustified, let alone morally wrong.²⁰ As noted, the inevitable risk of recidivism upon releasing a felon must be balanced against risk of keeping someone locked up who deserves to be set free.

Collectivists may at this point grant that there was no wrongdoing in Parole Board. Nevertheless, collectivists could still insist that while the board is not collectively blameworthy, it is still collectively responsible for the decision. So, if the victims seek some recognition of the harm caused to them — say, by way of an apology or compensation — they should turn to the board as a whole.

Responsibility-individualists are likely to remain unconvinced. Even if they agree that victims may be entitled to some recognition for the harms suffered they will not accept that the nor­mative basis for the victims’ claims will be the collective’s responsibility. The reason is that, argu­ably, if individual members of the collective are not responsible for the outcome, then whoever is responsible for instituting the premise-based procedure is not responsible for the outcome either. Consider Parole Board again. The premise-based procedure may have been instituted by an individual, by the board itself, or it may have been adopted spontaneously as a “natural” way of rendering the decision without any explicit reflection by anyone. It may seem that in the second and perhaps the third case the collective is responsible for the relevant outcome.

This conclusion, however, does not follow. According to the collectivist, the reason why no individual member of the board is responsible for the outcome is because each actually opposed that decision and none of them could foresee that their actual contributions would result in a decision which yielded the specific outcome. Lack of knowledge is of course a standard and familiar responsibility-undermining condition, so if this condition applies to the group members, then indeed most will grant that the individual group members are not responsible for the decision.

It is important to see that these findings have little to do with the specifics of the example analyzed above (e.g., the legal or institutional context in which parole boards or tenure committees operate). The general problem with the Simple Argument is that it fails to show that the collective decisions in question constitute the violation of a moral norm or obligation by the collective (or anyone). As long as the individual contributions to the collective decision are justified, the collective decision will be justified as well, and so the collective will not be at fault. Moreover, the Simple Argument even fails to show that the collective (or anyone) is responsible for the fact that the use of premise-based procedure has led to a potentially harmful outcome.

21.5.2 The Contingency Argument

A different kind of responsibility-based collectivist argument from the discursive dilemma avoids some of the problems just mentioned. The basis for this argument is that the collective’s decision is contingent on the decision procedure (Copp 2007; Szigeti 2014a) in the sense that which decision will carry the day depends only on whether a conclusion-based or premise-based aggregation rule is used. One could argue that such contingency could be unfair or harmful in certain situations. This argument is made more plausible by the fact that in such situations the decision seems to be vulnerable to challenges by those who were negatively affected by it.

Imagine, for example, that in a variant of Tenure Committee if the candidate for tenure at a department is successful, then an adjunct, Dr. Secondbest, will have to be let go due to financial constraints.²³ If Dr. Borderline is refused tenure, Dr. Secondbest can stay. Now, if the premise­based procedure is used by the committee, then Dr. Secondbest can raise what seems to be a legitimate complaint — after all, no member was in favor of giving Dr. Borderline tenure. If, on the other hand, the conclusion-based procedure is used, then Dr. Borderline can protest — after all, the university’s by-laws for tenure clearly state that excellence in the three areas of research, teaching, and service is sufficient for tenure.

So, the Contingency Argument differs from the Simple Argument in a crucial respect. The Simple Argument is based on the analysis of cases such as the Parole Board where it seemed that the use of a specific decision procedure, namely that of the premise-based procedure, led to a morally objectionable outcome. By contrast, the Contingency Argument seeks to locate the basis for the moral objection in the potentially worrisome contingency of the decision on the decision procedure used. Unlike in the case of the Simple Argument, this moral objection could apply whether a premise-based or a conclusion-based procedure is used.

Collectivists of course will also have to show that the collective as a whole should be the addressee of the relevant moral criticism. Their suggestion once again is that the members cannot be reproached for anything since they acted in accordance with the procedure in place and voted fairly and after careful consideration of the applicant’s merits.²⁴ Consequently, if the complaint is justified, then the only plausible addressee of Dr. Borderline’s or Dr. Secondbest’s complaint will be the committee itself. Unless the committee is held collectively responsible for the unfairness of the decision, we will have a responsibility gap, i.e., a “deficit in the accounting books.”

As in the case of Simple Argument, the individualist response will once again consist in, first, questioning whether the situation involves any wrongdoing, and second, whether the collective bears any responsibility with or without wrongdoing. As regards the first issue, there is a good case to be made that the Contingency Argument fails to identify any violation of a moral norm or obligation. Both the premise-based and the conclusion-based procedures are reasonable, impartial, and procedurally fair.²⁵ If so, then there is no room for challenges on the basis that some alternative decision procedure would have produced a different result.

Moreover, note that if it is true that the same kind and degree of unfairness results whichever decision procedure is used (just the victims will be different), then it follows that all options open to the committee were equally objectionable, and therefore, equally bad (or equally good). It is worth recalling here (see section 21.4) that the paradox of collective judgment aggrega­tion, as demonstrated in the pertaining impossibility theorems, consists in the fact that it cannot be guaranteed that for any distribution ofjudgments held by rational individual members of a group the respective outcomes of the premise-based and the conclusion-based procedures will be the same. This is why there is no procedure that could guarantee outcomes against which nobody could raise a legitimate complaint. In that case, however, it is justified to opt for either decision procedure, and so it does not constitute a violation of a moral norm or obligation to choose one or the other.

Perhaps we do not have to take this to mean that there is no basis at all to, say, Dr. Borderline’s (or Dr. Secondbest’s) complaints. Arguably, there is something morally objectionable about the fact that someone’s tenure (or dismissal) depended solely on whether the committee members’ views are aggregated via the premises or via the conclusion. Thus, perhaps the decision could be unjust even ifjustified.²⁶ If so, then the demand for some recognition of the injustice or harm caused by the contingency may be appropriate, and the collectivist could argue that responsi­bility for the injustice caused by the contingency must be laid at the doorstep of the collective as a whole.²⁷

It is, however, not clear why the committee should be singled out as responsible for the injustice (if any) suffered by the party at the losing end of the committee’s decision. The con­tingency is nobody’s fault. As we have seen, it is generated by the paradox ofjudgment aggrega­tion exposed by the discursive dilemma. It may be true that nevertheless the losing party’s bad luck should be compensated. In Tenure Committee, for example, one could consider repeating the procedure by adding a fourth criterion, or one could create a position for Dr. Secondbest, and so on. However, the crucial point is that the normative basis (if any) for undertaking some compensatory or remedial action of this kind is not the collective’s responsibility.

21.5.3 The Collective Irrationality Argument

Collectives can be criticized for being irrational. Collectivists maintain that such a failure can itself constitute the ground for collective responsibility. Specifically, it is argued that certain groups can be collectively responsible for failing to meet fundamental demands of rationality such as coherence and conversability (List 2006: 365f).

First, in the case of many collectives we require that their current decisions cohere with past decisions (Tollefsen 2004; Pettit 2007a; Pettit 2007b; List & Pettit 2011). Imagine, for example, that in Tenure Committee (as in the version of the case discussed in Copp 2007) votes are taken on the three premises, the results are made public, but at a later stage a vote is taken on the conclusion as well (see also Pettit 2003). In such a case, the collective can be charged with inco­herence because its position on the four propositions will be inconsistent.

Incoherence due to inconsistency is not just an epistemic failure, but can have morally prob­lematic consequences. For example, in his analysis of Tenure Committee, Copp (2007: 380) argues that it would be unfair to deny tenure to Dr. Borderline who, having learnt about there being majority support for her within the committee for each criterion of tenure, had a legitimate expectation or even a right to receive tenure. In general, Pettit warns us that unless a collective takes steps to ensure group-level coherence, it will not be regarded as “an effective or credible promoter of its assumed purpose,” and will not be taken seriously as an institution or organiza­tion (Pettit 2003: 177).²8 If the existence of the institution or organization serves an important moral purpose, then once again the collective’s epistemic failure can also be morally wrong. But since the epistemic failure and the resulting moral wrong is the collective’s fault, and not that of individual members (none of the members have voted inconsistently or could foresee the result of their individual contributions), the collective as a whole could be morally blameworthy and responsible for it.

Second, collectives can be criticized for failing to be conversable. In general, to be conversable means to be able and willing to provide reasons for one’s views and attitudes (List 2006: 365). It is said that we expect many collectives to justify their decisions, policies, and operation by pro­viding reasons for them, not least because this will also allow those affected by these decisions to challenge or contest them (Pettit 2001). Perhaps not all collectives must live up to such high standards, but those that are can be held responsible for failing to do so.²⁹

Individualists are likely to complain that the Collective Irrationality Argument involves a non sequitur. Most of them will be quite happy to accept that groups can be irrational, and even that group-level irrationality can emerge even when members are fully rational.³⁰ Furthermore, they can also accept that some groups are expected to be coherent and conversable. In Tenure Committee, for example, the collective might be required to take a straw vote to avoid group- level inconsistency. Perhaps the straw vote is taken because members are familiar with the lit­erature on the discursive dilemma. As a result, the committee may discover that it judges that p, and q, and r since there is a majority support for each of these propositions. However, it may also discover that it judges that -(p&q&r) because no individual member is in favor of awarding tenure. If this is the case, then the committee could change its “mind” and come around to embrace the judgment that (p&q&r) to resolve the inconsistency.

Individualists will reject the claim, however, that the collective is to be held responsible for the failure to meet the demands of rationality to which collectives are also subject to. They will argue that if inconsistency threatens the group’s rationality because the group is confronted with a discursive dilemma (or for some other reason), then the threat of inconsistency will generate reasons for individual members to revise their attitudes regarding the propositions to be decided upon. Failure to consider these reasons will be an individual failure not a collective one. For example, strongly convinced that - p&q&r) some individuals may refuse to allow the group to endorse (p&q&r). Or they may not notice that there is a problem, or not care about it. In any case, as a result, the collective could fail to meet the relevant requirements of rationality, which in turn may or may not have morally problematic consequences. But the crucial point is that when this happens it will be the responsibility of the individuals who failed to provide for the rationality of the collective because they were not sufficiently attentive or were not in a pos­ition to do so. And conversely too, it is individuals, not the group as a whole, who ensure that the collective remains coherent and conversable by supplying reasons for voting for or against a decision.

21.6