Deontic Facts, Agents, and Hayek

If you read the supplemental material to which I link with the diligence I expect and require, dear reader (tongue is firmly in cheek here!), you will have read this paper I referenced that examines proposed symmetry breakers between the modal ontological argument for the existence of god and the reverse modal ontological argument against the existence of god. One of the symmetry breakers, and the response to it, reminded me of something F. A. Hayek said when evaluating the concept of “social justice.” 

The symmetry breaker in question is the deontic symmetry breaker, which deals with deontic properties. Deontic properties are properties related to what ought to be the case, “properties of obligation and permission (e.g., rightness, wrongness, oughtness, etc.)”, which are distinct from evaluative properties that deal with “properties of value and disvalue (e.g., goodness, badness, etc.)” The deontic symmetry breaker goes as follows (with citation removed):

Therefore, according to this proposed symmetry breaker, we have reason to prefer premise 1 of the modal ontological argument over premise 1 of the reverse modal ontological argument. 

One objection to this comes from William Vallicella, who argues that deontic properties can’t sensibly be applied to non-agential contexts. That is, it doesn’t make sense to speak of what ought or ought not be the case in situations that are not under the control of any agent: 

This notion of the inapplicability of deontic properties to non-agential contexts reminded me of Hayek’s criticism of social justice, an idea idea he maintained “does not belong to the category of error but to that of nonsense, like the term ‘a moral stone.’” To Hayek, the reason “social justice” was nonsense is because the outcomes of social processes are non-agential. There are no agents with sufficient knowledge and power to bring about or prevent specific end results of social processes.

As Hayek put it in The Mirage of Social Justice, the second volume of Law, Legislation, and Liberty: “If we apply the terms to a state of affairs, they have meaning only in so far as we hold someone responsible for bringing it about or allowing it to come about…Since only situations which have been created by human will can be called just or unjust, the particulars of a spontaneous order cannot be just or unjust.” And the inability of agents to control the outcomes of social processes isn’t exactly an idea that’s only held by those on the political right – Friedrich Engels likewise said “What each individual wills is obstructed by everyone else, and what emerges is something that no one willed.” So you can be on the left, even the very far left, and still acknowledge that the outcomes of social processes are beyond anyone’s control.

To use an analogy, suppose there is a father who deliberately favors some of his children over others. He deliberately showers his favored child with love, attention, and resources, while outright neglecting and ignoring his other children. This, Hayek would say, is unjust, because the outcomes experienced by the children are entirely agential. But the outcomes of vast and complicated social processes are non-agential, and to speak of those outcomes as just or unjust, as if they were analogous to the hypothetical father above, is nonsensical. 

But not everyone shares Hayek’s take that the outcomes of social processes can’t be controlled in a reliably agential way. Jeffrey Friedman wrote extensively of people who hold to a “simple-society ontology” and who believed that certain actors (politicians, technocrats, etc.) can reliably control social outcomes in a way that is analogous to the hypothetical father’s ability to control the way he treats his own children. Thus, the more one holds to a simple-society ontology, the more likely they are to embrace “social justice” and find it a meaningful project, because they believe social outcomes are in fact under reliable agential control. Friedman described how such people expressed themselves in political polling data:

These voters believed that “the reason social problems persist is that elected officials have ‘the ability but not the will to take care of the nation’s problems.’ The ability was, for them, the easy part, or so it seems; the hard part was the will.” But if you think that politicians and technocrats haven’t solved social problems because they simply don’t know how to do so, then you lose the ability to meaningfully ascribe deontic properties. This doesn’t mean one can’t still ascribe evaluative properties to certain outcomes, and speak of the goodness or badness of such outcomes. If a landslide that nobody created and nobody could have prevented wipes out a village tomorrow, I can ascribe evaluative properties to that event (“it’s a tragedy this happened”) even though it makes no sense to ascribe deontic properties to that event (“all those rocks and mud ought not to have overrun that village.”) But people who harbor a simple-society ontology can lose sight of the distinction between evaluative claims and deontic claims – leading them to believe that an outcome that is evaluatively bad is therefore deontically unjust. But this is a mistake, and we should resist falling into it.