A little while ago I looked at Parfit's treatment of the "mere means" principle that he identifies with the second half of Kant's Formula of Humanity as Parfit treats it in Climbing the Mountain. In this posting I'm going to look at the ways Parfit revises this discussion in the pre-publication version of On What Matters that was reached towards the close of 2008. The discussion parallel to that of Climbing the Mountain takes place in Chapter 9 of the pre-publication version of On What Matters.
The opening of Chapter 9 closely parallels the beginning discussion in Climbing the Mountain distinguishing again between treating someone as a means and treating them merely as a means (which latter involves relating to them as a mere instrument or tool). In response to the latter notion Parfit mentions the objection that a rich person could respond to the duty to give to the poor by giving extremely little and thereby still fail to follow their duty. Similarly, we might only partially fail to treat people merely as means (though it is rather less clear how this would go). This leads, as in Climbing the Mountain, to a revision of the mere means principle which now says it would be wrong to treat someone as a means or doing something that comes close to that. Now, after mentioning this revision of the mere means principle Parfit next refers to two ways in which our conduct can be said to be such that it definitely does not involve treating someone merely as a means.
These two ways are if we either treat this person in a way that is governed by a relevant moral belief or concern or we are or would relevantly choose to bear some great burden for this person's sake. However, given that relevance and importance are themselves matters of degree it is often unclear whether we have met these conditions. Further, we might well be prepared to meet the condition of bearing a great burden for someone whilst normally treating them merely as a means.
Parfit distinguishes further between the kinds of consideration appropriate to saying that we have treated someone as a means as opposed to treating them merely as a means. In the first case Parfit takes it that reference to "intentions" is appropriate whereas in the second it is more apt to refer instead to "underlying attitudes or policies" (hence to view these on the lines that would fit many characterisations of the Kantian notion of a "maxim"). However, despite having introduced this complexity, Parfit is keener on a different distinction, namely that between regarding someone as a mere tool and acting towards them in a way that reflects such a view of them. Whilst he takes the former to be wrong he is less convinced that the latter is always so. Part of his case for this view is strikingly weak as when he claims that a gangster treats a coffee seller merely as a means because they purchase the coffee rather than stealing simply because this is simpler. However surely this is a case of instead only treating someone as a means and doing so in a way that lacks moral worth without thereby having become "wrong"?
Rather than drawing this conclusion, which would seem the obvious one, Parfit allows a further emendation of the mere means principle so that it now claims that we don't treat someone merely as a means if our acts will not harm this person (adding that we know that this is so). The introduction of this criteria of "harm" was, however, unnecessary in relation to the case and the example that Parfit introduces to cause trouble for this emendation is not one difficult to meet. Parfit asks us to imagine the case of someone marrying an elderly person merely for their money without acting so as to harm them, a form of conduct which surely involves treating them merely as a means. It is clear that this example undercuts the new formulation since it shows actual harm is not necessary for conduct to be such that it involves wrongfully treating someone as a mere means. However, it is also the case that this example illustrates wrong conduct on the earlier formulations of the mere means principle and thus demonstrates further the otiose character of the emendation provided.
Parfit does at this point call in the notion of acts that lack "moral worth" though the case just introduced deserves a stronger response than this. However, rather than drop the otiose reference to harm, he retains it in a further formula of the mere means principle that describes wrong-doing as consisting in treating someone as a means or coming close to that if our act "will also be likely" to harm them. The principle seems to be insufficiently differentiated from the previous version and no more necessary than it.
Parfit next turns to the task of combining the mere means principle with the consent principle that he argued was the root of the first part of the Formula of Humanity. At this point his discussion repeats the invocation of trolly examples that was given earlier in Climbing the Mountain. Subsequent to this Parfit reverts to what he calls the "standard view" of the mere means principle that claims that if we harm people, without their consent, as a means of achieving some aim, then we are treating them merely as a means. The invocation of "harm" is, again, quickly shown to add nothing relevant to the picture since it is possible to harm someone without treating them as a mere means or doing anything wrong as Parfit himself confesses. Similarly, treating someone as a means has already been distinguished from treating them merely as a means so it remains unclear what Parfit has here added to his previous account.
The account of harm distorts Parfit's subsequent treatment of examples since he keeps coming back to cases involving it as supposedly definitive of treating someone merely as a means and then showing that this is insufficient for the action to be viewed as wrong. But if treating someone merely as a means has no essential link to "harm" then these cases only prove the point that the introduction of the reference to "harm" has not proved helpful.
A different element of Parfit's discussion is his response to the claims of Christine Korsgaard and Onora O'Neill that coercion and deception are essentially connected to treating others merely as means. The negative reaction Parfit has to these suggestions turns again on cases where reduction of harm would ensue from deceiving others so that the question of harm is taken to be an obvious sign of wrong actions, a point that Parfit simply assumes and does not argue for.
Parfit also returns to the Bad Samaritan case used in Climbing the Mountain to show that not regarding someone as worthy of being taken as an end of action can be as bad as treating them merely as a means. However it is not evident what relevance this is supposed to have for assessing the right way of viewing the mere means principle. Whilst the mere means principle may be insufficient on its own as a ground for moral appraisal this does not show that it doesn't pick out an important wrong-making characteristic of some actions.
As in Climbing the Mountain Parfit's overall aim is to make the claim that regarding people as mere means is different to and worse than acting in ways that treat them as mere means. But it is unclear that his discussions do any more than conflate questions of treating others merely as means with discussions about "harm" and effectively treat questions about harm as definitive for the view we should have of the "mere means" principle. For these reasons it is unclear to me that the argument of Chapter 9 of the pre-publication version of On What Matters marks any real progress on the account in Climbing the Mountain.