Clinical Ethics >
Volume 3, Number 4 >
Pp. 185-188
Papers |
Department of History and Philosophy of Science, University of Cambridge, Cambridge, UK
E-mail: abb24{at}cam.ac.uk
This paper argues that bioethics requires analysis, which is not explicitly ethical in character. The first part of the paper argues the general point, that ethical problems can arise not only on occasions when moral values make conflicting recommendations, but also in understanding the facts. I suggest that this is particularly so where the facts are provided by the biomedical sciences, since it is often not clear how to relate their conceptual framework to that in which we frame our value judgements. In the second part, I illustrate the argument by criticizing the moral conclusions drawn by the authors of a recent widely publicized study of paediatric obesity. There is a failure to translate properly these results into the language we use for moral evaluation. The case therefore illustrates exactly the analytic gap which I suggest bioethics might do more to fill.
 | Introduction |
|---|
 Top Introduction Bioethics and analysisA case study: paediatric...ConclusionReferences |
|---|
The specific point is an example to illustrate the general point. I take a recent study of paediatric obesity and argue that it fails to support the moral conclusions which its authors draw.
By ‘analysing the facts’ I mean working out how to apply our concepts to the facts. This will often require some conceptual analysis – by which I mean working out the proper extension of our concepts and the relations between them. These two activities often go together when we come up against facts whose classification is not obvious. I use ‘analysis’ as a general term for both these activities.
| Bioethics and analysis |
|---|
|
TopIntroductionBioethics and analysisA case study: paediatric...ConclusionReferences |
|---|
If a normative or evaluative principle is framed in terms of a predicate which has fuzzy edges (as nearly all predicates in practice have), then we are not going to be able to use the principle to decide cases on the borderline without doing some more normation or evaluation.1
Hare says it is ‘inescapably a moral decision’ whether a fetus is a human being.1 He recommends that instead we should ask whether there is anything about the fetus, or about the person it may turn into, which means we ought not to kill it. He thinks this question is more ‘direct’.1
Hare's preferred method, then, is to ask directly how we ought to treat a fetus. But that is a broader question than whether a fetus is a human being. In particular, as he recognizes, it includes considerations about what the fetus has the potential to develop into. Those considerations may also be relevant to the abortion debate, but they are not the same as the considerations about what the fetus is now. Likewise, the welfare of the mother is a consideration. These factors all have weight, along with many others; but I cannot see how conflating the relevant factors helps assess their weight.
Hare draws a parallel with a hypothetical example from jurisprudence (which appears to be a variation on the theme made famous by Hart2), in which an Act prohibits the use of wheeled vehicles in the park. Billy rides his roller skates into the park. Has he broken the law?
...no amount of cerebration, and no amount of inspection of roller skates, are going to settle for us the question of whether roller skates are wheeled vehicles ‘within the meaning of the Act’ if the Act has not specified whether they are; the judge has to decide whether they are to be counted as such.1
Hare is arguing that it is a legal question whether the extension of the concept ‘wheeled vehicle’ for the purposes of the Act includes roller skates. For this reason, the extension of the concept cannot be empirically determined.
In the legal case, judges as a matter of fact do not follow Hare's suggestion: they do not sidestep the question whether roller skates are wheeled vehicles for the purposes of the Act. The legal restrictions that the Act imposes on wheeled vehicles generate a legal obligation on judges to determine whether roller skates are wheeled vehicles. A person could hardly be convicted of using wheeled vehicles in the park under an Act that prohibited their use if the items in question did not count as wheeled vehicles for the purpose of the Act. Likewise, the moral implications attached to being a human being mean it is morally necessary that we decide whether a fetus is a human being if we are proposing to destroy it, even if (as Thomson argues convincingly) that decision is not sufficient to settle the moral question.3
Nor do judges ask Hare's substitute question: ‘Is there anything about the roller skates that should make us say that they ought not to be used in the park?’ In fact a judge would avoid any such question, precisely because it is plainly not equivalent to the question whether roller skates are vehicles for the purpose of the Act, and thus not a reliable guide to the duties imposed by that Act.
This is not to say that judges confine themselves to literal interpretation of statutes. A judge may apply the ‘Golden Rule’, eschewing or exceeding a literal construal so as to avoid absurdity;4 or considering the mischief the Act was designed to prevent,5 or the purpose more generally of Parliament in passing the Act.6 Clearly, such approaches include a normative element. Nevertheless, they remain rules for deciding whether roller skates are wheeled vehicles for the purpose of the Act; there remains a gulf between that question and Hare's question as to whether there is anything about roller skates which means we should prohibit them. The analogy with legal practice thus lends no support to Hare's argument. Lawyers find it helpful to do exactly what Hare says we should not do if we aspire to clear moral reasoning: that is, to settle matters of law by asking questions of a factual form and by thus breaking down the ‘direct’, all-things-considered question which Hare advocates.
Hare's approach was that adopted by the Warnock Report on Human Fertilisation and Embryology. This Report does not concern abortion directly, but does cover some areas where the human-ness or otherwise of an embryo might be thought relevant. The Report avoids such questions, however, and perhaps an advocate of Hare's ‘direct’ approach might cite the following passage in support:
Some people hold that if an embryo is human and alive, it follows that it should not be deprived of a chance for development, and therefore that it should not be used for research... Although the questions of when life or personhood begin appear to be questions of fact susceptible of straightforward answers, we hold that the answers to such questions are in fact complex amalgams of factual and moral judgements. Instead of trying to answer these questions directly we have gone straight to the question of how it is right to treat the human embryo.7
This strategy is in line with Hare's recommended approach. However it does not support Hare's approach for three reasons. First, the only apparent justification for the line taken by the report appears to be the pragmatic point that these questions (or rather, their answers) are ‘complex’. This pragmatic attitude may be appropriate on occasion, but that is not the reason to advocate it across the board. Just because an approach is appropriate in one case does not mean it will be appropriate in all. Second, the Warnock Report is a political document produced at Parliament's request. It may be, therefore, that in adopting this approach Warnock was deliberately seeking to avoid clarity. Clarity can be dangerous in politics for the same reason that it is desirable in academic debate: because it brings out differences of opinion. So, in contrast with Hare, it is plausible that Warnock chose this approach because it was less ‘direct’ with respect to the underlying questions; the political nature of the document thus renders it a shaky support for Hare's philosophical approach. Third, even with the caveat that this is a political document, the approach is not wholly satisfactory even in the context of this Report. After all, it is hard to imagine that the commissioning of the Report was not motivated in part by the worry whether abortion amounts to taking human life. It is, at the very least, legitimate to ask whether the Report ought to have shirked the question of whether an embryo is human. Again, this means that the Report cannot be held up as a successful application of Hare's approach without qualification.
More generally, the claim that concepts such as ‘vehicle’ and ‘human being’ may be of indeterminate extension does not tell against trying to settle their extensions, at least for a particular purpose. It tells only against any such effort which assumes that such an extension already exists – empirically discoverable or otherwise. But I am not advocating any such assumption. In the legal context, one of the judge's jobs is to sharpen up concepts that are not determined by legislation or case law, to fill in the gaps, in a way that seems harmonious with the legislation and the settled cases, and also coherent enough so as to offer some predictability regarding the outcome of future cases.
This analytic sort of activity in the ethical arena is what I advocate for bioethics. Hare is right that many of the concepts we use to frame ethical judgements have extensions that are not empirically discoverable or perhaps not antecedently determinate at all. But far from showing that we ought to neglect the question of the extension of these concepts, it shows that we need to undertake analytic work in order to apply them sensibly to new cases. In the case of bioethics, this means working out how to apply prescientific categories, such as ‘human being’, to scientific categories – categories either introduced by science or adopted and given a precise definition – such as ‘fetus’. What makes the matter all the more pressing in bioethics is that new results in the biological sciences often do have some apparent relevance to the classificatory questions upon which ethical judgements depend.
| A case study: paediatric obesity |
|---|
|
TopIntroductionBioethics and analysisA case study: paediatric...ConclusionReferences |
|---|
Wardle et al.8 set out ‘to quantify genetic and environmental influences on BMI [body mass index] and central adiposity [fat around the waist] in children growing up during a time of dramatic rises in paediatric obesity’. They achieved this objective with a twin study, which they used to estimate the heritability of obesity. The heritability of a trait is the proportion of differences in that trait between members of a given population that are attributable to genetic factors.9,10 Wardle et al.8 found that BMI and waist circumference were both 77% heritable.
Twin studies have been criticized for overestimating heritability.11 But rather than discuss general methodological issues, I want to focus on issues which are specific to Wardle's study and the moral conclusions it is supposed to support.
Wardle writes:
Discussions about the obesity epidemic almost invariably ascribe a key role to the family, but, in the present study ... siblings from the same family were only slightly more similar in adiposity [fattiness] than would be expected from their genetic similarity ... What is important is this finding means that ‘blaming’ parents is wrong ... Results from the present study highlight the fact that excessive weight gain in a child is unlikely to be the fault of the parents...8
I want to criticize two aspects of Wardle's attempt to draw morals from the results of the study. The first is that heritability is ill-suited to ground the sort of moral conclusions Wardle seeks to draw. Second, even if it were better suited to that purpose, Wardle's study does not license an answer to the question that we ask in the context of moral evaluation.
The first point to make about heritability estimates for obesity is that they vary greatly between studies in the region of 55–85%.8,12–15 It is, therefore, vitally important to distinguish between the confidence one might have in the results obtained for one data-set and the confidence one can have that those results will generalize. Care should be taken in generalizing Wardle's conclusions; and this goes equally for any moral claims based on those conclusions. Even if the parents in Wardle's study were blameless, we should be cautious to infer that parents in general are blameless. And it is clear, for example from the passage quoted, that Wardle does think that her results admit of some meaningful generalization; it is hard to see how else to read her claim that ‘what is important is this finding means that "blaming" parents is wrong’.
But even if Wardle's conclusions could be generalized, there is a fundamental difficulty with using heritability the way Wardle uses it. Heritability is a measure of the relative contribution of environmental and genetic effects, because it measures the proportion of a trait in a population attributable to genetic effects. As such, it is, by definition, partly determined by the environment. Parents generally control aspects of their children's environment. If parents control those aspects which play a role determining heritability, then parents themselves may help to determine how heritable a given trait is.10 Therefore, applying heritability crudely to questions about whether parents are responsible for defects in their children has the potential to be extremely misleading. Wardle's is an example of such a misleading application.
Wardle does not explain her reasoning. But presumably, the reason she thinks we ought not to blame parents for their children's obesity is that, being highly heritable, it is largely beyond parental control. She says:
This finding will ... come as no surprise to parents, who are well aware that their children come in different shapes and sizes despite having a similar upbringing.8
Suppose, perhaps fancifully, that the reason for the observed high heritability is that the studied parents generally let their children eat whatever they like, thereby allowing genetic predispositions full reign. The children in any given family would have had the ‘same upbringing’ in this sense. But the consequent high heritability would hardly show that the parents had behaved blamelessly. On the contrary, their behaviour would have directly contributed to the high heritability observed. For this reason, Wardle's moral conclusion, if the reasoning for it is as I have construed it, is fallacious. High heritability does not indicate low parental control; indeed, high heritability may be a direct result of parental behaviour.
Of course, I do not mean to suggest that, if parents have brought it about that obesity has a high heritability, they should automatically be blamed. Whether they are blameworthy depends on other factors, such as what they knew and ought to have known about nutrition, and the stresses and strains of their own lives. My point is not that parents are to be blamed after all: it is that heritability cannot by itself settle whether they are to be blamed.
The second point to make about Wardle's study is that, even if heritability were a good indicator of parental control over a trait, it would not automatically exculpate parents. As long as parents retain a degree of control, moral evaluation focuses on how they exercise that degree of control.
This point can be brought out clearly with an analogy. Some children are born blind for genetic reasons. Suppose that a particular group of parents brought them up as they treated their other children, they received the ‘same upbringing’. These blind children would not last long: they would fall down stairs, run into roads and suffer all sorts of mishaps.
Suppose Wardle then conducted a study of the heritability of some subset of these mishaps, such as falling down stairs. They would be highly heritable: they would be differences between those children and others in similar environments that were very largely attributable to genetic differences underlying their blindness. Would this exculpate the parents? Clearly not. To decide whether the parents of the blind children were to be blamed for the injuries, we would not ask about differences between those blind children and other, sighted children. We would ask about differences between those blind children and the blind children of parents – perhaps hypothetical ones – whom we considered to be morally upstanding. We would compare the upbringing the children had, not with the upbringing which other children had, but with the upbringing that they themselves ought to have received.
When we decide whether to place blame on an agent, we usually focus our attention on those factors that the agent can control. It is, therefore, of limited interest to point out that, in many cases, factors beyond our control are at work. What we want to know is whether the parents did the right thing with regard to those factors that they could control. A particularly direct way to satisfy ourselves on this point is to compare the upbringing the children received with the upbringing they ought to have received in those circumstances. Wardle's study clearly does not provide us with the material to frame this contrast, since it does not tell us what upbringing the children concerned ought to have received.
In sum, Wardle's paper makes strong moral assertions without supporting them. The failing is not one of scientific rigour: I have not argued, and do not believe, that the methodology of the paper is flawed. Rather, the failing is one of analysis: the results are mapped uncritically onto the conceptual framework in which we frame value judgements. High heritability is wrongly thought to suggest low degree of agential control, and hence to exculpate the agents in question – parents. And low degree of control is wrongly assumed to be decisive in exculpating agents from moral blame.
One might react to this example by asking whether it really matters what one study says. But it does matter: Wardle's paper received extensive publicity and is (at time of writing) the latest in a string of high-profile studies of the genetics of obesity. Another response might suggest that Wardle's moral remarks are akin to a slip of the tongue. That response would be incorrect. The remarks are emphasized in the discussion section; one might wonder whether the moral question of whether parents are to be blamed was a motivating factor in undertaking the study, especially as the policy recommendations are not very convincing (as the editors of the journal of publication pointed out16). Moreover, the moral dimension must have been reiterated in a press release, as the BBC quote Wardle as saying: ‘It is wrong to place all the blame for a child's excessive weight gain on the parents – it is more likely to be due to the child's genetic susceptibility’.17 It is easy to criticize the media for over-general and melodramatic claims. But it is too easy for academics, by blaming the media, to thereby shirk all responsibility: over-generalized and under-analysed moral claims are not all due to media amplification, but are also to be found at the source, in academic papers presenting the results of scientific research. In this paper, I have sought to highlight the need for better academic analysis of the results of biomedical studies, sensitive to the moral implications of those results yet prior to value judgements being attempted, or else as part of the process of arriving at an informed value judgement: the need, in short, for analytic bioethics.
| Conclusion |
|---|
|
TopIntroductionBioethics and analysisA case study: paediatric...ConclusionReferences |
|---|
| Acknowledgements |
|---|
| Footnotes |
|---|
| References |
|---|
|
TopIntroductionBioethics and analysisA case study: paediatric...ConclusionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. C Dunn, Z. Gurtin-Broadbent, J. R Wheeler, and J. Ives Jack of all trades, master of none? Challenges facing junior academic researchers in bioethics Clin Ethics, December 1, 2008; 3(4): 160 - 163. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
