Just how seriously is virtue ethics under attack by moral psychology?

The reason that Virtue Ethicists do not consider the experimental results to be refutations is because they do not actually test virtue ethics. Virtue ethics does not predict that virtuous people will always behave virtuously, but that virtuous character habits increase the frequency of virtuous actions from the virtuous person.

That people are more likely to be helpful when smelling something nice -- does not change that the test subjects who actually are helpful in the experiments, will be more frequently those who acquaintances will describe by the character traits "helpful" and "kind". And while some circumstances will increase the frequency of lies that the test subjects tell, those people who acquaintances describe as "deceitful" and "liars" will feature predominantly among those who tell lies.

The above, that there are differential dispositional traits that we observe among people, IS experimental data. It is not the RECENT experimental data that is cited by Appiah, but that is actually the nature of faddism in philosophy. Studies generally pursue questions that could reinforce the researcher's predispostions, and there is a clear pattern of fads in published philosophy. Hence the recent trend of data supporting situationalism is reflective of little more than that recent PhD candidates lean toward situationalism.

It is a bit of a conundrum as to why Appiah, who favors a long view of philosophy, and realizes experiments to be intrinsic to most historical philosophiszing, was blinded by the recency fallacy to think that the historical Virtue ethics thinking was NOT based on experiment!

N.B. The question as written is presumptive and argumentative: it asserts that empirical psychology has seriously undercut rational moral philosophy, and then questions why moral philosophers haven't acknowledged that. As a rule, asking whether a group of people are 'in denial' is a poor start for any intellectual exercise; it speaks to bias.

That aside, The reason that virtue ethicists (not to mention deontologists and consequentialists) haven't had more of a reaction to empirical research in psychology is that they don't really need to. This kind of research calls into question some of the presumptions of virtue ethics, but doesn't really touch its core principles.

Virtue ethics is a old field, having roots all the way back in ancient Greece. Its core idea is that ethical behavior comes from virtuous character traits, such that a man who has such traits will act more ethically than a man who does not. We should understand, however, that this particular approach to ethics is steeped in a naïve classical psychology, in which character traits are malleable, easily modified by conscious applications of will and reason. For Plato or Aristotle, self-examination and philosophical introspection are the keys to creating virtue; anyone willing to put in the effort can improve their character and attain some measure of ethical standing. This was the prevailing worldview up until Freud and his introduction of unconscious mental activity. Most of the writings of Classical Liberalism (Locke, Smith, Rousseau, etc), shared this understanding of the human mind, and even today we can still find plenty of authors in fields like economics, political theory, and even theology who cling to 'rational actor' models of human behavior.

Of course — as the question highlights — modern psychology has provided a large body of experimental evidence that people do not always (or even generally) act with conscious rationality. People are subject to biases, misattributions, cognitive shortcuts, emotional reasoning, anti-statistical intuitions, and other non-rational influences. People do not (in the language of virtue ethics) demonstrate 'good character' on a regular basis.

But this is where we need to be careful. The fact that people do not consistently or frequently show 'good character' does not imply that people have a disposition towards 'bad character'. That tendency to think of character traits as dispositional, not contextual, is an example of the Fundamental Attribution Error. Any virtue ethicist who considers this psychological research will likely admit that the classical worldview is flawed, but hold that the general principle — that people can and should develop virtuous character traits — is still valid. All that's changed is the process by which those character traits are developed. Virtue cannot be achieved by the direct application of will to build character, but by a more subtle, indirect process of working through unconscious attitudes. One ends up at something like Jung's efforts to embrace and integrate the 'shadow' into one's overt, expressed nature.

Virtue ethicists might seem underwhelmed by psychological research, but only because psychological research isn't challenging the substance of their programme, only its implementations.

Looked a bit like one argument for moral anti-realism: there's a lot of disagreements about what to do, so maybe we don't "know" what is moral.

https://plato.stanford.edu/entries/moral-anti-realism/moral-realism-intuitive.html https://plato.stanford.edu/entries/moral-anti-realism/moral-realism-explain.html

may have some relevance.

Moral disagreement is widely held to pose a threat for metaethical realism and objectivity

https://link.springer.com/article/10.1007/s10892-008-9041-z

The fact that we're talking about psychological traits rather than knowledge, may be relevant, or it may not. It will suffice, for me, to point out that moral scepticism may need to respond to the idea of "progress" -- so what if the Aztecs practised human sacrifice -- in ways that the challenge against virtue ethics does not.

For me, if no-one is virtuous, then we can probably infer that virtue ethicists - 'virtue' is a long standing science - are not being very helpful, and that they may not even believe in virtue themselves. That would make it at best a thought experiment.

It is a sensible point to argue that valid moral theorizing is constrained by empirical findings, though to what extent I do not know. Surely normative claims hold against empirical shortcomings, but they become invalid upon demanding the impossible. Virtue ethics would not seem to be in denial, then, as the reported findings ('weak character', 'no character') fail to establish such a strong verdict. I'd agree that "the concept of virtuous character runs the risk of being part of a fictuous story", but reification strikes me as the more obvious response than abandonment. Yet again, I couldn't tell how to find the right balance there.