OPR VI.17. Arguments from Abtraction and the Claims of Agency

Summary of OPR.VI.17

Chapter VI begins by reminding us of an important conclusion from the previous chapter, namely, that the Members of the Public (MoP) will be confronted with a large set of rules of social morality, and that with respect to these rules, the MoP (as a group) is indifferent (they prefer any member of the set to no rule at all, but do not converge on any particular member of that set).

The goal of this chapter is to advance to two partial solutions to this ‘problem of indeterminacy.’  They both concern individual rights, specifically, those rights commonly known (from Benjamin Constant’s famous essay) as the ‘liberty of the moderns.’  These solutions are only ‘partial’ because they serve only to narrow somewhat the set of eligible rules of social morality, but do not pick out any particular rules.

Section 17 presents the first of these two partial solutions.  In this section, drawing on the work of Benn, Gewirth, and Rawls, Gaus employs an ‘argument from abstraction’ to show that all reasonable Members of the Public would be committed to endorsing, at least in an abstract form, certain fundamental individual rights (the ‘liberty of the moderns’), as such rights are essential for effective agency.

Gaus begins the section by reminding us of the ‘Kantian-Rawlsian two-step procedure’ for arriving at justified principles under circumstances of reasonable pluralism (diversity of ends and values among the reasonable MoP). Roughly, this procedure involves ‘bracketing’ our disagreements, adopting a shared perspective, and reasoning on the basis of this shared perspective (the perspective of pure practical reason for Kant, the perspective of the original position for Rawls).

Gaus advances his own ‘argument from abstraction’ in this section in order to show that the MoP would support certain individual rights for all persons.  However, specific interpretations of individual rights, that is, specific rules, acceptable to all MoP in accordance with the ‘deliberative model,’ will need to be formulated at a later stage.  Nonetheless, showing that all reasonable MoP endorse such rights can serve to narrow the set of eligible rules of social morality (rules that deny such rights to some persons or deny them altogether are ruled out).

Gaus claims that the success of any argument from abstraction (whether Rawls’s original position argument, or the argument that Gaus advances in this section) depends on three claims:(a) the successful identification of a shared perspective (the original position for Rawls; the perspective of abstract agency for Gaus); (b) the importance or weightiness of the evaluative standards identified by the shared perspective (why the conclusions of the shared perspective should be taken seriously by the MoP for the purposes of evaluating rules of social morality); and (c) the ability of the conclusions generated via the shared perspective to survive the return of the Members of the Public to their ‘full set of evaluative standards’ (i.e., the ability of the conclusions of ‘pro tanto justification’ to survive ‘full justification’).

Gaus asserts that it was a concern with (c), the compatibility of the conclusions of the shared perspective (the conception of ‘justice as fairness’ endorsed by the parties in the original position) with reasonable persons’ various ‘comprehensive doctrines,’ that prompted Rawls’s move to political liberalism.  Rawls’s commitment to the original position device as the appropriate perspective for ascertaining principles of political justice remains constant from A Theory of Justice to Political Liberalism (p. 336).  As we’ll see, Gaus thinks that while the first principle of justice as fairness (the basic liberties principle), or some version of it, survives (c), the difference principle cannot.

The “second abstraction characteristic of Rawls’s original position,” Gaus explains, is that it focuses on the justification of abstract principles rather than rules.  Gaus restates his claim (from 14.3) that “principles are too vague and too subjective to interpretive controversy to provide an effective framework for cooperation” (p. 337).  Nonetheless, identifying principles shared by the MoP can be useful, since such stably shared principles would at least eliminate many proposed rules for social morality.

Gaus then identifies the ‘perspective of agency.’  In a pluralist society we cannot help but think of ourselves as agents — as persons who form intentions in light of our beliefs and values, and who act on those intentions.  Even people who endorse ‘traditional’ ways of life cannot help but think of themselves as choosing to do so (and acting accordingly), given their knowledge of the alternative ways of life available in their society.

The idea of agency is explicated by Gaus by drawing on S. I. Benn’s account of ‘autarchy’ (from his classic A Theory of Freedom).  Most normal adults are ‘autarchic’ agents – they are ‘self-ruling’ and ‘self-directing,’ in the sense that they understand themselves as determining their own actions on the basis of their intentions, and they understand themselves as forming their own intentions.  Human beings who lack autarchy (say, schizophrenics or those who suffer from compulsive obsessive disorders) are not fully ‘persons.’

(I should note here that Gaus, for most part, uses the term ‘autarky’ instead of ‘autarchy.’  This may be somewhat confusing.  Benn himself consistently used the term ‘autarchy’ throughout A Theory of Freedom. The term ‘autarky’ typically is used to refer to an economically independent or self-sufficient society, not a ‘self-ruler.’  I assume that this simply is a typo.)

Autarchy is not to be confused with an ideal of ‘personal autonomy,’ understood as a form of reflective, critical ‘self-authorship.’  The latter is a controversial ideal that not all MoP share.

From the perspective of agency, Gaus claims that the MoP would endorse ‘The Presumption of Liberty’ (PoL): “(1) agents are under no standing moral obligation (in social morality) to justify their choices to others; (2) it is wrong to exercise one’s liberty so as to interfere with, block, or thwart the agency of another without justification” (p.341).  Gaus then considers some possible objections to the PoL (341-45).

Gaus then argues that, from the perspective of agency, the MoP would endorse a right not to be coerced, understood at this stage as a right “not to have threats made against one’s natural person” (p. 352).  This is because such coercion prevents or constrains the exercise of agency.  (Coercion at this stage of Gaus’s discussion cannot be understood as involving threats against one’s property, since property rights cannot be assumed to exist yet.  Such rules concerning persons’ relations to external objects need to be justified. Eventually, though, as the MoP recognize more expansive rights for persons, including some form of property rights, the scope for coercion also increases.  Such more expansive rights are not defended in this section.)

The MoP would also endorse a right against deception, as deceiving a person subverts their autarchy, and a right against being harmed (i.e., an assurance that an agent’s “basic welfare interests – bodily integrity, health, the absence of severe pain, absence of psychological torture and distress, reasonable security of necessary resources – are not set back severely by the agency of others” (p. 357)).

The perspective of agency also leads the MoP to endorse freedom of thought/conscience.  In order to be effective agents, persons must be free to arrive at their own beliefs and values (the ‘internal’ dimension of freedom of thought), and to act on the basis of those beliefs and values (the ‘external’ dimension of freedom of thought).  (Obviously the latter right is defeasible.)  The importance of freedom of thought for Gaus cannot be overstated.  “Freedom of conscience is a core liberal right as it sums up the main grounds of free agency, uniting free deliberation and free action.  It has a basic place in the order for agents.  Appropriately, it is the core right that Rawls’ parties endorse in his deliberative original position.” (p. 354)

With respect to what other (abstract) rights might be justified from the perspective of agency, Gaus considers Gewirth’s claim that all agents would endorse a ‘strong’ duty to assist other agents in need (when doing so would not impose on them any comparable costs).  Gaus argues that while a ‘weak’ right to assistance could be justified to all reasonable MoP (he outlines such a ‘duty of rescue’ briefly on p. 367), any stronger right to assistance could not be justified.

Gaus’s argument for this claim is that once the MoP leave the abstract ‘perspective of agency,’ and evaluate the various rights tentatively endorsed from that perspective in light of their full set of beliefs, commitments and values (or, in Rawlsian terms, their ‘comprehensive doctrines’), many will find themselves unable to still endorse a strong right to assistance (although the other rights survive intact, at least in abstract form).  A strong right to assistance cannot survive claim (c) of the Rawlsian argument from abstraction (as outlined at the beginning of this section and in 3.1).  In other words, the ‘pro tanto justification’ for the right to assistance fails once we reach ‘full justification.’   As Gaus puts it, a “strong duty of assistance is not stable under full justification” (p. 363).

There are many reasons why the MoP might object to a strong right to (and corresponding duty of) assistance in light of their respective full sets of evaluative standards.  Some will object because they are committed to a principle of desert.  Others will have concerns about the problem of ‘moral hazard’ that such a right to (and duty of) assistance might cause.  ‘Luck egalitarian’ justifications for such a duty fall afoul of Hayekian views concerning the proper operation of markets.  And so forth.

Interestingly, Gaus interprets Rawls’s difference principle as a strong right to assistance.  Also noteworthy is Gaus’s claim that his argument here does not rule out a welfare (or even a more egalitarian) state.  Rather, Gaus simply claims that the justification for such a state cannot be made from the foundational ‘perspective of agency’ (of which the original position is a species).

The section concludes with Gaus noting certain limitations to the argument that he has advanced.  First, the principles agreed upon by the MoP still need to be translated into specific moral rules, and such translations must be publicly justified to all MoP as free and equal persons.  Given the ‘deliberative model,’ though, it is unlikely that a unique translation will be generated.  Instead, the MoP will face “a set of socially optimal eligible interpretations of an abstract justified right” (p. 370).  Nonetheless, Gaus claims that progress has been made, as we can see that the MoP would endorse certain basic individual rights (the ‘liberty of the moderns’) as part of social morality.

Comments on OPR.VI.17

1. The Relation between ‘Pro Tanto Justification’ and ‘Full Justification’

Gaus writes: “Full knowledge of evaluative standards may change what is validated from the perspectives of citizens. Hence Rawls’ claim that under ‘full’ justification the normative importance of the pro tanto argument from abstraction is preserved (i.e., condition iii is met). Let us, then, call this third requirement the stability of abstract justification under full justification. This requirement is immensely important: unless the conclusion of the argument from abstraction can be affirmed in light of a rational and reflective free and equal moral person’s full set of evaluative criteria, the abstract justification will be defeated by these other elements of her evaluative set.” (p. 336)

This account of the relation between principles of political justice selected on the basis of a shared perspective (in Rawls’s case, the original position device) and citizens’ various ‘full sets of evaluative criteria’ (comprehensive doctrines) seems to presuppose, I think, an overly ‘static’ or ‘fixed’ picture of citizens’ respective ‘full sets of evaluative criteria’ (comprehensive doctrines).  That is, Gaus’s argument appears to assume that the comprehensive doctrines of citizens are unalterable by reasonable persons in light of the political principles that they come to see as justified from the (abstracted) shared perspective.  This may be Gaus’s view of the relation between the rules of social morality and the full set of evaluative criteria of the MoP (this is how I’ve understood his position thus far).  However, I don’t think that this is Rawls’s position.

Instead, very roughly, I understand the Rawlsian argument to proceed as follows:
(a)    Reasonable persons are committed to the criterion of reciprocity and acknowledge the burdens of judgment.
(b)    Drawing on the ideas found in the public political culture of democratic society, especially the ideas of fair social cooperation and citizens as free and equal (ideas to which citizens are committed prior to their deliberations about political justice), reasonable persons consider the argument for justice as fairness via the original position device (the original position device is a way to realize, in a rigourous manner, the criterion of reciprocity).
(c)    Once reasonable persons appreciate the (‘pro tanto’) argument in favour of justice as fairness, they see whether they can endorse the two principles of justice as fairness in light of their comprehensive doctrines via reflective equilibrium (‘full justification’).
(d)    Importantly, the process of arriving at a reflective equilibrium with respect to the principles of justice as fairness can involve revising some of reasonable persons’ pre-reflective values, beliefs, and commitments (aspects of their respective ‘full sets of evaluative criteria’ or comprehensive doctrines).

As far as I can tell, Gaus’s presentation of Rawls’s ‘argument from abstraction’ does not consider (d). (Perhaps I’m reading Gaus uncharitably?)

2. The Original Position Justification for the Basic Liberties versus the Difference Principle

Gaus writes: “Our question here is whether an extensive right to assistance is justified at, as it were, the ‘ground level’ of basic claims of agency. Rawls thought so. Although in his principles that were agreed to in his original position deliberative model, agency freedom had priority, a strong principle of economic redistribution, requiring that individuals be secured a bundle of resources that satisfy their basic needs, was also justified as part of the fundamental claims of free and equal agents. But we know that free and equal agents deeply disagree on questions of distributive justice… To assert that this is a matter of consensus for all reasonable and rational free and equal people under full justification is either simply implausible, or stipulative, deeming a wide range of individuals in Western societies to be unreasonable. These are complex and contentious issues; a claim that they have been resolved because there really is no reasonable dispute about them and that the level of consensus of reasonable agents on them is akin to that on freedom of speech, is simply not credible.” (Pp. 367-8)

I think that Gaus misrepresents Rawls’s own assessment of the relative strength of the justifications available for the two principles of justice.  That Rawls thinks that the difference principle is less secure (the arguments for it are less decisive) than the first principle (the basic liberties principle) can be seen in two aspects of his work.

(a) In Justice as Fairness: A Restatement, Rawls explicitly concedes that his argument for difference principle (by means of the ‘second fundamental comparison’ in the original position) is less conclusive than his argument for the basic liberties and their priority (in the ‘first fundamental comparison’).

(b) In his final writings on political liberalism, Rawls explicitly acknowledges that there exist a family of different ‘reasonable’ liberal conceptions of justice.  All members of this family satisfy three criteria: roughly, they include a set of basic liberties; assign to those liberties a ‘special priority’; and ensure for all citizens adequate primary goods to make effective use of their basic liberties over the course of their lives (see The Law of Peoples, pp. 14, 141). Thus there is greater legitimate variation among reasonable conceptions of political justice with respect to distributive justice than there is with respect to the basic liberties.

Thus I think, contra Gaus, that Rawls holds that the level of consensus among reasonable agents on the basic liberties (the ‘liberty of the moderns’) is much greater than the level of consensus of reasonable agents on the difference principle.  The gulf between Rawls and Gaus on this matter does not seem to be as great as Gaus asserts.

3. The Difference Principle as a Right/Duty of Assistance

The characterization of the difference principle provided in the quote above — Gaus’s presentation of the difference principle as a ‘right of assistance’ — strikes me as problematic (and perhaps ironically so, given what Gaus says earlier in this section concerning coercion and property, viz., that coercive threats involving one’s property presuppose that one has a right to that property, that it is in fact ‘property’ and not merely ‘possession,’ and that such a right to property cannot be presupposed from the perspective of pure, abstract agency).

The difference principle, as I understand it, is not a principle meant to establish ‘rights to assistance’ of some citizens (those who happen to find themselves among the least advantaged) on others (those who are more advantaged).  Rather, the difference principle concerns the overall structure of a society’s system of entitlements and property rights.  That system should be organized so as to benefit the least advantaged over time more than any alternative system.  It is not a system that ‘assists’ particular persons who find themselves among the least advantage by transferring to them the (justly held) property and wealth of the more advantaged.

4. Abstract Principles versus Social Rules?

Gaus writes: “A second abstraction characteristic of Rawls’ original position is that we focus on the justification of abstract principles rather than rules. I argued in section 14.3 that principles are too vague and too subject to interpretive controversy to provide an effective framework for cooperation; they are not the ultimate aim of public justification. However, they can help us delimit our justificatory problem: if Members of the Public could agree on some abstract principles, and if these were stable under full justification, they would at least have narrowed their problem to the justification of rules that are adequate interpretations of the already justified principles.” [337]

Is this really a deep difference between Gaus and Rawls?  Rawls doesn’t think that the principles of justice selected in the original position are sufficient for determining the terms of social cooperation in actual societies.  Rather, the principles selected in the original position provide the basis for the selection of more specific rules (constitutions, laws, and the application and interpretation of those laws) by persons with additional information of the society in question.

Recall Rawls’s ‘four-stage’ procedure adopting and applying the principles of justice as fairness.  Only in the first stage, that of the original position, are principles chosen.  The second stage, the ‘constitutional convention,’ involves the selection or a particular constitutional structure in light of the representatives’ knowledge of the society in question (Iceland will require a different constitutional structure in order to realize the principles of justice as fairness than, say, India).  The next stage is the legislative stage, in which laws concerning ‘basic justice’ are determined (including the difference principle).  The final stage is that “in which the rules are applied by administrators and followed by citizens generally and the constitution and laws are interpreted by members of the judiciary” (Justice as Fairness, p. 48).

In short, I don’t think that the difference between Gaus and Rawls concerning the relation between principles and rules (social rules for Gaus, laws for Rawls) is that great.  Indeed, Gaus’s overall argument in this section strikes me as quite Rawlsian in nature.

This entry was posted in Books, Discussion, Posts, Reading Group and tagged , , , , , , , , . Bookmark the permalink.

3 Responses to OPR VI.17. Arguments from Abtraction and the Claims of Agency

  1. My apologies for not posting this comment on Monday. I regret any inconvenience caused by my tardiness.

  2. Hi Blain,

    I think you raise good points about the relationship between pro tanto and full justification (the Rawlsian idea that the abstraction stage of argument can lead to revisions in people’s overall evaluative standards, which Gaus doesn’t seem to acknowledge), and about the difference principle not being a duty of assistance (though it may be still to early in the book to form any definite conclusions about principles of institutional assessment, since at the moment we are just talking about moral rights). I want to raise a question about the structure of Gaus’s argument, in moving from “abstraction” to “full justification,” specifically about what I think is a discrepancy between the way Gaus treats the right to assistance and the right to free speech.

    The abstraction strategy is to bracket disagreement and look for a shared perspective that provides a common basis for evaluating moral rules. The results of this process of abstraction and reasoning from a common basis must be stable under full justification, i.e. they must be endorsed by all Members of the Public when they consider their full set of evaluative standards. In the abstraction argument, “we focus on the justification of abstract principles rather than rules” (337). This focus on principles makes sense because we are looking for a (common) basis for evaluating proposed moral rules. The shared perspective Gaus identifies is that of agency. He argues that from this shared perspective (not yet considering full justification relative to people’s complete evaluative standards), we should endorse a right not to be coerced, a right to freedom of thought, and “some fairly strong welfare rights” (duties of assistance to those in need) (358-9). However, “when the abstraction is lifted, and the deliberators are aware of the full range of their evaluative standards” only the first two rights are endorsed, not the strong right of assistance that seemed reasonable considering the shared perspective of agency alone. For once we bring in the full diversity of reasonable evaluative standards, we have to recognize that differences of opinion about desert, risk, responsibility, and so on lead some reasonable persons with reasonable points of view to reject the strong right of assistance.

    Gaus recognizes – in fact, he insists – that the rights justified on the grounds of agency and endorsed under full justification are abstract (350). Thus in the section on the limits on the argument from abstraction (17.6), Gaus imagines a debate between three Members of the Public who agree on the abstract right to freedom of speech, but disagree about how exactly it should be specified, and so disagree about how far the right should protect explicit sexual content, for example. Gaus remarks of this case that “we again confront the problem of a maximal set with no agreed-upon optimal element: they [the three Members of the Public in question] see reason to select from the maximal set [one of the three specifications of the right], but the choice from within it is, from the perspective of the Members of the Public, indeterminate” (369-70).

    There seems to be an important difference between Gaus’s treatment of the right of assistance and his treatment of freedom of speech. In the right of assistance case, he classes competing specifications as weaker or stronger, recognizes reasonable disagreement about which is optimal, and concludes that we must default to the weaker specification. In the freedom of speech case, he treats competing specifications as being on a par, and concludes that our task is to find a way of selecting from this set. Why the difference between the two cases?

    One answer would be to claim that it is reasonable to reject the strong right of assistance as against the alternative of not having any right of assistance at all, whereas none of the three specifications of the right of free speech is reasonably rejectable as against not having any such right. But I don’t see that Gaus actually makes the case for this claim on pp.362-368. He shows that it is reasonable to deny that the strong right of assistance is best; I don’t think he shows that it is reasonable to prefer no right of assistance at all, as against the alternative of the strong right. Of course, there must some extravagantly strong specification of the right to assistance that is reasonably rejectable as opposed to none at all, but it’s not clear to me that Gewirth’s fits that bill, and even if it does, all we would need to do is to weaken the right somewhat in order to arrive at a set of specifications of the right to assistance, which would then constitute part of the eligible set.

    A second answer would be for Gaus to offer some criterion for ranking specifications in terms of strength, and a principle that in the absence of unanimous agreement on a stronger specification we default to the weaker specification, which we might think of as the incremental application of the unanimous acceptability criterion. But it’s not clear what that criterion would be, and whether any such principle would itself have to be (and actually be) unanimously acceptable.

    Sorry to go on at such length – hopefully this issue will ring a bell with someone other than me.

  3. Hi Andrew,

    Although I don’t have much to add to what you’ve written, I thought that I would mention that I also found Gaus’s treatment of the right to assistance somewhat puzzling. The default to the weakest interpretation seems far too quick. Presumably there would be a set of acceptable interpretations of that duty, most of which would be stronger than the most minimal one, but that reasonable MoPs nonetheless would prefer over *no* such duty at all.

    I haven’t read much further in the book at this time, but based on Gaus’s “Classical Tilt” article (Social Phil & Policy 2010), I suspect that he would argue that there would exist greater overlap among the MoP on ‘weaker’ interpretations of the duty of assistance than there would exist on ‘stronger’ interpretations (and, at a certain point, some of the MoP would prefer no duty at all to a very strong one). But that kind of argument is quite different from claiming that, given the existence of objections to strong(er) interpretations of the duty, we must default to the weakest conception available.

    In short, the issue did “ring a bell” with me, for what that’s worth.

Leave a Reply