I am pleased to have the opportunity to respond to the thoughtful and insightful comments of 11 of my colleagues. I have had only the briefest time to consider their comments; no doubt more prolonged reflection will yield more insights. Already, however, the comments have assisted my thought. Some of the comments reflect a misunderstanding of my arguments on the part of commentators, but these misunderstandings are useful because they point to a lack of clarity on my part (sometimes a lack of clarity in my expression and sometimes in my thought). Most of the comments focus in one way or another on methodological issues in normative and applied ethics; these comments have been especially useful to me, because they have helped me to become clearer on just what role I think scientific data can and should play in moral enquiry and philosophy more generally.Several of the commentators (e.g., Boddington 2011, Braude 2011, and Rommelfanger and Boshears 2011) take me to be committed to a kind of positivism about science and philosophy—perhaps even a view on which neuroscience should replace philosophical enquiry. I do not advocate any such view. In the target article, and in my book (Levy 2007) I emphasized the interdisciplinary nature of neuroethics. Neuroethics is, or ought to be, a conversation between philosophers (of various kinds), neuroscientists, social and cognitive psychologists, and researchers in other fields. I envisage the tools of neuroscience and the other sciences of mind as supplementing and testing the tools used in philosophy, not simply replacing them. The reductionism that Braude sees in my article, according to which neuroethics “seeks to validate only those intuitions whose rational foundations can be demonstrated through scientific measurement,” is foreign to me. Rather, I advocate a highly interdisciplinary and multilevel approach to philosophical issues.It is important to stress the multilevel nature of the methodology urged here. Several commentators argue that cognitive science cannot contribute to philosophical debate unless and until substantive philosophical questions are settled. Bruni (2011) argues this most forcefully; Carbera (2011) makes a similar claim, though she limits its scope to the appeal to Greene's neuroimaging data for the resolution of ethical questions. Both commentators claim that the appeal to scientific data begs the question against rival views: For instance, if we argue on the basis of Greene's data that deontological intuitions are irrational because they are driven by affect, we beg the question against cognitivist views of emotion. But the begging-the-question charge is forceful in this context only when directed against an argument that uses this data to provide a premise for a simple deductive argument. That is not how intellectual inquiry should proceed; it is not even how scientific inquiry proceeds. Rather, we should seek multiple overlapping lines of evidence. We ought not to expect a single experiment to settle substantive debates; rather, we should bring to bear the best evidence from neuroscience, psychology, philosophy, the social sciences, and the humanities. Rather than thinking that we must first settle substantive philosophical issues before we appeal to data, we ought to be working on the ways in which philosophical views and scientific data illuminate one another. Here, as elsewhere, we should seek reflective equilibrium.Similar remarks apply to Rommelfanger and Boshear's (2011) claim that I overinterpret neuroscientific data in taking them as evidence for psychological states. Though the view that the BOLD signal does not provide any evidence of personal-level psychological states has its distinguished defenders among scientists (Coltheart 2006), it is very much a minority view and it is one that I reject. By itself, a single neuroimaging study is indeed merely correlational, as Rommelfanger says, but we ought to interpret such studies in the light of a great deal of other data: neuroimaging data, behavioral data, and our best psychological and philosophical theories (Shea and Bayne 2010). Neuroscience is a young science, as several commentators emphasize. But this is not a reason for thinking that it cannot inform our theories. Taken together with one hundred years of work in scientific psychology, as well as with an enormous amount of high-quality work in the other natural and social sciences, as well as the humanities, it can provide us with evidence that is already altering our conception of ourselves and our minds. Thus, while I endorse Giordano's (2011) call for a neuroethics that is alive to our social and biological natures, I do not take this as a reason for endorsing his pessimism about the powers of science to help us settle difficult philosophical questions. Rather, I hold that a multilevel methodology is more likely to give us defensible answers to questions about the nature of mind and its relationship to its physical realizers.This holistic and multilevel methodology, I claim, is not only the appropriate method to utilize; it is in fact the method that characterizes successful intellectual inquiry including the natural and social sciences today. Recognition of this fact explains its success, and also explains how other disciplines—including philosophy—are capable of generating important findings. It is for this reason that I think that Epstein's (2011) charges against philosophy miss their mark. Epstein thinks that moral philosophy cannot play any role in distinguishing between permissible and impermissible actions because its principles are too broad to yield determinate answers; if we supplement these principles with further assumptions to yield sufficient determinacy, however, we beg the question against certain views. But this broad-brush rejection of the value of philosophy neglects the fact that all intellectual inquiry—including the sociology that Epstein advocates oughttoreplace philosophy—makes and assesses claims against the background of many other claims. No proposition entails anything by itself. No researchers can hope to investigate the whole range of claims that their work presupposes; that is why intellectual inquiry is necessarily socially distributed. Progress on any aspect of inquiry depends upon the work of researchers in many other fields. Thisis true inphilosophy, and in the natural and social sciences as well. I heartily endorse Epstein's call for sociological inquiry into the presuppositionsofneuroethics. But this work should not be regarded as a replacement for work in neuroscience, psychology, philosophy, and so on: rather, it should be seen as one more—important—voice in the ongoing interdisciplinary conversation.Similar remarks apply to Boddington's (2011) praise of dialogue as a way of testing moral claims, and to Braude's (2011) call for the integration of first-person experience with third-person neuroscience. We should not deprive ourselves of possible sources of evidence. However, we should also not shirk the concurrent task of investigating whether and when these methods misfire. Take the use of first-person report. Some researchers think that introspection is of limited reliability: Peter Carruthers (2009), for instance, has argued that we cannot even introspect the content of our own attitudes. Other researchers are more optimistic about our introspective powers, but have mounted a powerful case that introspection is valuable only when it is managed properly, utilizing skilled interviewers and special equipment (Hurlburt 2011). Appeals to first-person data should proceed hand-in-hand with careful consideration of this work. We ought not, for instance, acquiescein Duggan's claim that “only the agent … can be entirely certain of his or her intention” without careful probing of whether intentions are the kind of things that can be reliably introspected (a claim that Carruthers, among others, would deny), and if they can be, whether we can reliably introspect these states without special training and perhaps special equipment (a claim that Hurlburt would deny).Let me turn now from broad methodological issues to more specific questions. A number of commentators misinterpret my claims about the reliability of intuitions. Levin, for instance, accuses me of confusing the context of discovery with the context of justification. That is, she takes me to overlook the fact that even if intuitions are caused by processes that are unreliable, we may nevertheless come to see that there are after all good reasons for the judgment toward which the intuition disposed us. I agree completely, and have no quarrel with the heuristic use of intuitions. If you have good arguments for a claim, I don't care what motivated you to investigate the claim. I am concerned with the evidential value of intuitions independently of supporting arguments. Philosophers routinely invoke intuitions in this kind of way (Pust 2000; Goldman 2007); that is, they claim that because a claim is intuitive we ought to give it some (perhaps defeasible) evidential weight. It is because I am concerned with intuitions invoked as evidence that the focus on thought experiments and the responses they generate, for which Boddington takes me to task, is in fact appropriate. It is precisely intuitions elicited in this kind of way that are in question here.A further remark is worth making in this context. Levin (2011), like Braude and Bruni, thinks that I aim to test which intuitions are reliable. Bruni in particular takes me to task on this score. I agree with Bruni that showing that some intuitions are reliable requires showing a range of other things as well—perhaps even settling some substantive philosophical debates. ButIdonot aimtoshow that some intuitions are reliable. Rather, I aim to show that some intuitions are unreliable. Showing that requires weaker evidence than showing that they are reliable; to pass as not unreliable, there simply has to be no evidence of their unreliability. I do not therefore have to offer a way of distinguishing reliable from unreliable intuitions, as Bruni claims: rather, I only have to offer sufficient (rather than necessary) conditions for unreliability. Bruni's request for further elucidation is well-taken, however. Here is my working criterion for unreliability: An intuition is unreliable if it is generated in response to features of thought experiments or actual situations that neither are the features that are genuinely up for debate in a particular dialectical situation nor reliably track features that are genuinely up for debate in a particular situation. If our intuitions about whether an action was intended or merely foreseen, the feature that is up for debate in the context of the doctrine of double effect, are driven by our prior moral views, then these intuitions are unreliable, I claim.The specific methodological issue of the evidential status of intuitions should not be divorced from the broader methodological question. Multilevel and overlapping lines of evidence should be brought to bear on our intuitions too. Zarpentine (2011) notes that sometimes the same intuition is generated by different causal routes in different subjects. As Zarpentine says, this fact complicates the project I have set myself, but it does not render it impossible. Rather, it necessitates the broad methodology advocated here: Because no single experiment can provide evidence about the many ways in which intuitions are actually generated, we shall need to consider a broad range of experiments, and other data too. Of course, if there are too many causal routes whereby subjects come to have intuitions (where intuitions are identified by their contents, not their etiology), then the project will become intractably complex. On the available evidence, however, there seem to be tractably many processes to investigate. The task does not require, as Zarpentine thinks, that we show that all possible routes to the generation of an intuition are suspect. We are concerned with actual human beings and their intuitions.Let me finish with a brief discussion of the doctrine of double effect, upon which both Duggan (2011) and Michael (2011) take me to task. Duggan thinks that I have “a very peculiar understanding of DDE;” I take it to be concerned with whether kinds of actions are permissible or impermissible, whereas the doctrine was actually developed to help agents themselvesto decide how to act. I know relatively little about the history of the doctrine, so I do not wish to take issue with Duggan's claims about its development. Howeveritis absolutely clear that the doctrineisstandardly used today in the way I claimed. The first sentence of the “Doctrine of Double Effect” article in the Stanford Encyclopedia of Philosophy (McIntyre 2009), the closest thing that contemporary philosophy has to a standard reference work, makes this explicit:The doctrine (or principle) of double effect is often invoked to explain the permissibility of an action that causes a serious harm, such as the death of a human being, as a side effect of promoting some good end.If Duggan wishes to restrict the doctrine to first-person action guidance, it is true that some of the considerations I advanced would be avoided. But this would be so only on the supposition that agents have better insight into what they intend in the cases at issue than do third parties. This is not a supposition that is obviously true, as I have already noted.John Michael (2011), finally, accuses me of overlooking the distinction between judging the goodness or badness of a side-effect and the judgment about the moral permissibility of an action. Subjects in Knobe's experiments, Michael suggests, generate an intentionality judgment as a consequence of a prior judgment about the moral significance of the side effect; the judgment about the overall permissibility of the action is subsequent to this. I agree, but I do not see this as a problem for my view. My claim is that insofar as the judgment about overall permissibility is influenced by the prior judgment about intentionality, mediated by an implicit or explicit invocation of the doctrine of double effect, the judgment is suspect.There is, of course, a great deal more to be said. I have not been able to respond to every point made by every commentator. Moreover, I do not take myself to have shown that those who advocate different methodological approaches to ethics in general or to the studies I have cited are clearly wrong. Rather, I hope to have shown that the approach I advocate is worth taking seriously and deserves further exploration and development. I thank the commentators for providing me with the opportunity to make a start on that further exploration by challenging me to refine my thinking.