On Drawing Conclusions from the Empirical Data
In the following, I consider the following concerns that one might have with my use of the empirical data to ground discussion of conceptions of moral appraisal:
The experiments might fail standards of ecological validity; the experimental context might be so distant from natural contexts as to preclude generalisations to the ‘real world’
The experiments may be conceptually irrelevant; the conceptions operationalized in the empirical work may not correspond to the related conceptions figuring [in the theories discussed] (Doris and Stich 2005: 122–123)
One might think that the experimental conditions in which the motivational tendencies of the subjects were assessed are unlike the conditions under which individuals are subject to moral appraisal. For example, one might object that in the experimental conditions, subjects were encouraged to engage in certain behaviours with the expectation of task contingent rewards. But such rewards would not ordinarily be expected for such behaviours, and it was clear that after the experiments, no such rewards could be expected.
This appears to be a point of crucial disanalogy that prevents generalisation from the empirical data to the theoretical models of moral appraisal. Contexts of moral evaluation are ongoing, and expectations of moral evaluation may persist even in the absence of an appraiser.
This concern does not mean that the empirical data has no import for ‘real world’ scenarios. Firstly, concerns about generalisations to the real world should be allayed by attention to the breadth of contexts in which such empirical data were collected (which go beyond laboratory settings, and include studies of the motivation of residential patients to adhere to health care plans, motivational tendencies of students to complete homework projects, motivation of workers to take on overtime, and so on).
What these concerns do indicate, however, is that whilst we might suppose subjects of moral appraisal to suffer decreased motivational tendency to comply with moral norms were the evaluative contexts to be removed, we in fact need not concern ourselves with this: the evaluative context is omnipresent, when it comes to moral appraisal. As such we need not concern ourselves with the potential diminishment of motivation to act well. This would affect my claims at premise (2) of my argument against (CA) – we should not expect there to be fewer moral actions if agents are motivated to act morally solely in contexts where rewards and sanctions are present. This is because all contexts are such contexts.
By way of response, I offer the following: if by ‘context of evaluation’ the objector means that the subject will be subject to evaluation for her action, clearly all contexts are not contexts of evaluation. Not all good acts are praised, and not all moral wrongs are blamed. Alternatively, the objector might mean that contexts of evaluation are contexts in which the subject expects to be evaluated for her action, or (weaker) believes there is a possibility that she will be so evaluated.14 I accept that if ‘context of evaluation’ is understood in this way, then there may well be a wide range of circumstances in which we need not concern ourselves with subsequent decreases in motivation: the evaluative contexts will remain, and motivation will be sustained. Even so, there presumably will still remain a perhaps small number of contexts that do not count as evaluative contexts – contexts in which the subject can be sure that there will be no evaluation made (perhaps because no one else will ever know of the action performed).
These are the contexts in which we should expect subsequent decreased motivation. Should one accept premise (1) of the anti-consequentialist argument, then the presence of such scenarios is sufficient for the argument to run. An objector is committed to the claim that there are never any contexts in which an agent does not expect to be evaluated, by others or by herself, and this claim does not seem to be a plausible one.
Having defended against this concern, however, I do concede that this objection might lead to the conclusion that we rarely need to worry about the potential diminishment of subjects’ motivation. This does not, however, detract from my claim that the possibility of such diminishment, in these albeit rare circumstances, is not one that should arise in an adequate conception of moral appraisal.
A more pressing concern is whether the empirical data can be generalised to the realm of moral appraisal and moral motivation. The concepts used in the inquiries might be subtly, but crucially, different from the concepts at use by theorists of moral responsibility. If one believes that moral motivation is sui generis, then one will not accept that conclusions can be drawn from these empirical studies which do not incorporate data on specifically moral motivation (and one might be concerned about the ethics involved in experiments that potentially decrease subjects’ moral motivational tendencies).
I cannot here address the complex issues surrounding the nature of moral motivation, and whether or not such motives share the relevant features with our non-moral motives. I do, however, sketch a line of thought that an objector might propose: that there is a distinction between conventional and moral motivation, and the experiments are concerned with the former. As a result generalisations to claims about moral motivation are unwarranted.
Such an objection might take the following form: the experimental contexts concern only motivation to avoid sanction for conventional transgressions (e.g. disobedience to the commands, or failure to meet the expectations of, the authoritative figure involved in the inquiry). There is reason to believe that subjects distinguish between conventional and moral norms (Nichols 2002, 2004). There might be reason to believe, then, that whilst individuals experience diminished subsequent motivation to comply with conventional norms (no conventional authority; no compliance) this need not be the case when it comes to motivation to adhere to moral norms. More importantly, if all the data we have is concerned with adherence to conventional norms, no conclusions can be drawn out with regards motivation to adhere to moral norms.
A fully adequate response to this concern would need to be grounded in further inquiry concerning the relevant similarities and differences between motivation in moral and conventional realms. I cannot address such issues here. Hopefully the following partial response might allay some concerns.
We should reject the supposition that the empirical data concerns solely motivation to adhere to conventional norms. Some of the experimental contexts can plausibly be understood as being governed by conventional norms – for example, contexts in which experimenters gave task-contingent rewards for completing maths puzzles – the convention in play being one of compliance with the demands of authority figures.
However, in other contexts it is less plausible to cast the norms governing the subjects’ behaviour as solely conventional. Inquiries into motivation to adhere to health care plans, for example, seem to address motivational tendencies to act on the basis of norms that go beyond norms of compliance with authority figures. This is not to say that there are moral norms in play (although Kantian duties of self-perfection might cast adherence to such norms as a moral matter) – it is merely to suggest that the empirical data can extend beyond the motivational tendencies of subjects to comply with conventional norms; perhaps to warrant generalisation to the realm of moral motivation, and to inform our models of moral appraisal.