Moral judgments often affect scientific judgments in real-world contexts, but Knobe's examples in the target article do not capture this phenomenon.

We have found that moral considerations interact with belief ascription in determining intentionality judgment. We attribute this finding to a differential availability of plausible counterfactual alternatives that undo the negative side-effect of an action. We conclude that Knobe's thesis does not account for processes by which counterfactuals are generated and how these processes affect moral evaluations.

Moral judgment – even the type discussed by Knobe – necessarily relies on substantial information about an agent's mental states, especially regarding beliefs and attitudes. Moreover, the effects described by Knobe can be attributed to norm violations in general, rather than moral concerns in particular. Consequently, Knobe's account overstates the influence of moral judgment on assessments of mental states and causality.

In criminal law, foresight betrays a guilty mind as much as intent does: both reveal that the agent is not properly motivated to avoid an illegal state of affairs. This commonality warrants our judgment that the state is brought about intentionally, even when unintended. In contrast to Knobe, I thus retain the idea that acting intentionally is acting with a certain frame of mind.

Scientists are generally more moral, and moralists more scientific, than Knobe suggests. His own experiments show that people, rather than making unscientific judgements about the moral intentions of others, are behaving as good Bayesians who take account of prior knowledge.

In this commentary we explore Knobe's ideas of moral judgments leading to moral intuitions in the context of the moral thought and moral action debate. We suggest that Knobe's primary moral judgment and the setting of a continuum with a default point is in essence a form of cultural capital, different from moral action, which is more akin to social capital.

For Knobe, observers evaluate mental states by comparing agents' statements with “defaults,” the attitudes they are expected to hold. In our analysis, Knobe's model relies primarily on what agents should think, and little on expectancies of what they would think. We show the importance and complexity of including descriptive and prescriptive norms if one is to take expectancies seriously.

Children learn from people and about people simultaneously; that is, children consider evidentiary qualities of human actions which cross traditional domain boundaries. We propose that Knobe's moral asymmetries are a natural consequence of this learning process: the way “child scientists” gather evidence for causation, intention, and morality through early social experiences.

If Knobe is right that ordinary judgments are normatively suffused, how do scientists free themselves from these influences? I suggest that because science is distributed and externalized, its claims can be manipulated in ways that allow normative influences to be hived off. This allows scientists to deploy concepts which are not normatively suffused. I suggest that there are good reasons to identify these normatively neutral concepts with the folk concepts.

Knobe considers two explanations for the influence of moral considerations on “non-moral” cognitive systems: the “person as moralist” position, and the “person as [biased] scientist” position. We suggest that this dichotomy conflates questions at computational and algorithmic levels, and suggest that distinguishing the issues at these levels reveals a third, viable option, which we call the “rational scientist” position.

We propose Knobe's explanation of his cases encounters a dilemma: Either his explanation works and, counterintuitively, morality is not at the heart of these effects; or morality is at the heart of the effects and Knobe's explanation does not succeed. This dilemma is then used to temper the use of the Knobe paradigm for discovering moral norms.

I agree with Knobe's claim in his “Person as Scientist, Person as Moralist” article that moral considerations are integral to the workings of people's competence in making causal judgments. However, I disagree with the particular explanation he gives of the way in which moral considerations influence causal judgments. I critically scrutinize his explanation and outline a better one.

We are neither scientists nor moralists. Our mental capacities (such as attributing intentionality) are neither akin to the scientist's exact reasoning, nor are they “suffused through and through with moral considerations” (Knobe's target article, sect. 2.2, last para.). They are more similar to all those simple capacities that humans and animals are equally capable of, but with enhanced sensitivity to counterfactual situations: of what could have been.

Knobe reports that subjects' judgments of whether an agent did something intentionally vary depending on whether the outcome in question was seen by them as good or as bad. He concludes that subjects' moral views affect their judgments about intentional action. This conclusion appears to follow only if different meanings of “intention” are overlooked.

Knobe cites both relevant alternatives and defaults on a continuum to explain how moral judgments influence intuitions about certain apparently non-moral notions. I ask (1) how these two accounts are related, (2) whether they exclude or supplement supposedly competing theories, and (3) how to get positive evidence that people consider relevant alternatives when applying such notions.

Knobe relies on unhelpful analogies in stating his main thesis about the mind. It isn't clear what saying the mind works, or doesn't work, “like a modern university” or “a scientific investigation” means. We suggest he should say that some think that human cognition respects a ban on fallacies of relevance, where considerations actually irrelevant to truth are taken as evidence. His research shows that no such ban is respected.

Knobe contends that in making judgments about a wide range of matters, moral considerations and scientific considerations are “jumbled together” and thus that “we are moralizing creatures through and through.” We argue that his own account of the mechanism underlying these judgments does not support this radical conclusion.

Depressive mood can be involved in the moral judgments made by people with depression. Here, we focus on the negative judgments depressed patients have of themselves and the world. Possibly, the alterations in moral judgment in subjects with depression can be understood by taking into account the neural basis of depression.

Although we agree with the spirit of Knobe's competence model, our aim in this commentary is to argue that the default position should be made more precise. Our quibble with Knobe's model is that we find it hard to ascribe a coherent view to some experimental subjects if the default position is not clearly defined.

The commentators offer helpful suggestions at three levels: (1) explanations for the particular effects discussed in the target article; (2) implications of those effects for our understanding of the role of moral judgment in human cognition; and (3) more theoretical questions about the overall relationship between ordinary cognition and systematic science. The present response takes up these three issues in turn.