Defending My Moral Realism Article

Addressing a response article

Panth of naturalism next has written a response to my article arguing for moral realism. I do not think his responses are successful and here I’ll explain why. One frustrating feature of the article is that it failed to address many of the arguments I raised in the first article — often points I originally raised responded to points that he raised throughout the article. Overall though, the article is pretty good.

1 Intuitions: Good or Bad?

Intuitions and Phenomenal Conservatism 

Much of Matthew’s case rests on an appeal to intuition. A word of caution is in order about intuition. While I agree that intuition is a defeasible starting point for reasoning, I think we should be far more skeptical of it than many in the ‘pop philosophy’ spaces I occupy are. This is because our scientific picture of the world has consistently shown that many of our intuitive notions are false. Folk beliefs are notoriously unreliable. 

Let’s take one example from a domain I spent a lot of time doing research in during my undergraduate education: our perception is intuitively rich and filled with detail. We seem to perceive a world full of color across our visual horizon, and to have a detailed-rich picture of the world around us. Yet, research on phenomena such as inattentional blindness, change blindness, and crowding has shown that our visual world is far more limited than we think it is [1-5]. We frequently miss unexpected, obvious objects, we are extremely poor at detecting changes to scenes, and the way our periphery processes information is fundamentally limited. Specifically, it appears that we really just capture the general, rough ‘gist’ of the statistics of the information in our periphery, without the rich detail [6-7]. 

To see how shockingly poor our peripheral vision is, take a look at this video:

In this experiment, people in Virtual Reality viewed a series of scenes in which they were instructed to look around in whatever way they thought was most natural [8]. While in a critical scene, experimenters desaturated the periphery of their vision. Notice that very large proportions of people failed to notice this drastic alteration to the scene, even in the most extreme conditions. I propose to you that what you are looking at after the alterations in these scenes are made is not at all what intuition suggests our world looks like. Yet, psychology & neuroscience research strongly suggests that this is the actual way the world is represented in our brains (indeed, it’s even more limited than that, since in this video the periphery is detail-rich and in reality our peripheral vision lacks this detail!). Experiments looking at people’s metacognitive abilities (that is, their ability to be accurate about their own perceptual and cognitive abilities) demonstrate that people inaccurately think they do not have the visual limitations that research suggests they in fact have, pretty clearly demonstrating the unreliability of folk ideas in this area [9-10].

Notice that unlike the responses to phenomenal conservatism that Huemer talks about in Matthew’s article, this is an example of an intuition that we think about all the time in regular life. My goal here isn't necessarily to rebut phenomenal conservatism, but I do want to caution about putting too much weight on our intuitions. 

Several points are worth making.

1. While visual seemings are far from infallible as this shows, this does nothing to undermine the fact that things tend to be how they appear.

2. Pointing out a few cases of our intuitions being wrong doesn’t show that they aren’t generally justificatory.

3. Intellectual seemings can be evidence even if visual seemings aren’t.

4. The author seems to agree that seemings tend to be evidence, so the question is just whether seemings on balance decisively support moral realism.

Worth noting that none of the arguments for moral realism were addressed — I’ll quote them here.

1. Any argument against intuitions is one that we’d only accept if it seems true after reflection, which once again relies on seemings.  Thus, rejection of intuitions is self-defeating, because we wouldn’t accept it if its premises didn’t seem true.  

2. Any time we consider any view which has some arguments both for and against it, wecan only rely on our seemings to conclude which argument is stronger.  For example, when deciding whether or not god exists, most would be willing to grant that there is some evidence on both sides.  The probability of existence on theism is higher than on atheism, for example, because theism entails that something exists, while the probability of god being hidden is higher on atheism, because the probability of god revealing himself on atheism is zero.  Thus, there are arguments on both sides, so any time we evaluate whether theism is true, we must compare the strength of the evidence on both sides. This will require reliance on seemings.  The same broad principle is true for any issue we evaluate, be it religious, philosophical, or political.  

3. Consider a series of things we take to be true which we can’t verify.  Examples include the laws of logic would hold in a parallel universe, things can’t have a color without a shape, the laws of physics could have been different, implicit in any moral claim about x being bad there is a counterfactual claim that had x not occurred things would be better, and assuming space is not curved the shortest different between any two points is a straight line.  We can’t verify those claims directly, but we’re justified in believing them because they seem true--we can intuitively grasp that they are justified.  

4. The basic axioms of reasoning also offer an illustrative example.  We are justified in accepting induction, the reliability of the external world, the universality of the laws of logic, the axioms of mathematics, and the basic reliability of our memory, even if we haven’t worked out rigorous philosophical justifications for those things.  This is because they seem true. 

2 Is Moral Realism Intuitive 

Is Moral Realism Intuitive?

Now, how well does Matthew do in establishing that moral realism is, in fact, intuitive? His strategy in the first section of his post is to adduce a list of sentences that he thinks are obviously true, but that the anti-realist needs to deny, such as the following:

[When] dinosaurs experienced immense agony, having their throats ripped out by other dinosaurs…that was bad.

The holocaust is bad.

Torturing infants for fun is typically wrong. (Note, it’s funny that Matthew uses the word ‘typically’, since his utilitarian commitments might lead him to say this is sometimes morally acceptable. It seems just as obvious to me that ‘torturing infants is okay as long as it lets enough rats use heroin’ is false as the other moral statements Matthew comes up with, a point about intuitions I’m later going to bring up.). 

I’ve defended a similar version of this conclusion here.

Let’s just focus on one of these sentences for the sake of discussion, ‘The Holocaust is bad’. There are a number of problems with using the intuition favoring this sentence to argue for realism. I think they all rest on an equivocation: what the realist really need is the following sentence:

‘It is a moral fact that the Holocaust is stance-indendently wrong’. 

The realist treats ‘The holocaust is bad’ as if the only way of interpreting that sentence is in this manner. But, of course, many antirealists will be perfectly happy to affirm that sentence, just under an interpretation consistent with anti-realism. For instance, some non-cognitivists might say that ‘the holocaust is bad’ is equivalent to ‘boo the holocaust’ or ‘I disapprove of the holocaust’. Subjectivists might take ‘the holocaust is bad’ to mean ‘The holocaust is stance-dependently wrong’. Even an error theorist, who thinks that all moral assertions are false might assert ‘The Holocaust is bad’ in a fictionalized context. Richard Joyce has defended this position extensively, the idea is essentially that acting as if moral assertions are true can have prudential benefits: they affect our motivations in ways we might desire [11]. Of course, if you ask the error theorist to step back into the context of the philosophy classroom, they will agree that all moral assertions are false. But this does not stop them from using moral statements as a kind of useful fiction, and so uttering them under everyday contexts. And, of course, nothing is stopping the error theorist from saying ‘The holocaust was bad’ in a non stance-independently truth-evaluable way, where they mean something like ‘The holocaust was inconsistent with my subjective values’. So, really what the moral realist needs to do to make this argument from intuition is show that their particular interpretation of ‘the holocaust is bad’ is the intuitive interpretation. Once that interpretation is actually stated the Moorean force of their argument is significantly dented. And, the problem is that Matthew’s article is riddled with this implicit equivocation. He constantly opines on how obvious certain statements are when he is taking for granted his interpretation of those statements. Let’s alter some of his example sentences to explicitly state the interpretation needed for realism:

[When] dinosaurs experienced immense agony, having their throats ripped out by other dinosaurs…that was stance-independently immoral.

It is fact that the Holocaust is stance-independently wrong

Torturing infants for fun is a stance-independent moral wrong.

With these alterations, these sentences no longer appear to have significant Moorean force.

This is confused about the nature of the original argument. I don’t think that anti-realists always have to deny the statement torturing infants for fun is wrong. Rather, I went through and explained how, for each of the three schools of anti-realism, they have unacceptable implications. So denying that torturing infants for fun is wrong is only an entailment of error theory — but I raised problems for the other theories. On the subjectivist account, for example, the statement “torturing infants for fun would be wrong even if I thoughtA it were good” is false. However, that’s really implausible. All forms of anti-realism (with the possible exception of types of constructivism which fall prey to the shmagency objection) have to deny the following statement

Even if everyone thought it was good to cause maximum suffering on others, it would still be wrong to cause maximum suffering on others. Thus, Panth now mistates the view — as a result the objections have no force.

Now, I concede that some people will still find these sentences to be intuitive. However, this brings me to a second problem with these kinds of arguments. When Matthew says that these assertions are ‘obviously intuitive’, for whom is that meant to be true? If he intends to make an argument for moral realism that is persuasive, it cannot be just him. For, that he finds these sentences to be intuitive is an interesting psychological fact about him, but it does nothing to convince someone who does not find these sentences intuitive. So, he must be making a stronger claim: that these intuitions are widespread, and that the prevalence of these intuitions is some kind of evidence for realism. I have a few points to make about this. 

Firstly, as I pointed out earlier, folk intuition is a pretty bad guide to truth. But, secondly, Matthew provided no good evidence that these intuitions are widespread (a twitter poll is terrible evidence, for obvious reasons: it’s unsystematic, unrepresentative, etc.) . The relatively new field of experimental philosophy has spent a lot of time studying what the ‘folk’ consensus is on metaethics, and as far as I am aware those studies have very mixed results. There is no consensus that the way the ‘folk’ use moral statements is generally realist in nature; on the contrary, it seems that the ‘folk’ tend to be metaethical pluralists [12-14]. Perhaps instead Matthew intends to appeal to the consensus of academic philosophers. Approximately 62% of philosophers appear to be moral realists according to the latest philpapers survey, hardly an overwhelming consensus [15]. Ironically about 80% of philosophers are non-utilitarians according to the latest survey, showing that the very position Matthew endorses has a greater academic consensus against it than moral anti-realism! I don’t think either of these findings is significant evidence against the minority position, but if we are going to play that game Matthew is losing it. I think if the relevant experts agree on a position that is some evidence for the position, but at best it again establishes a defeasible presumption in favor of that position, not an overwhelming case. And, 62% is really not a very strong consensus. 

I didn’t make an appeal to the folk generally being moral realists — I’m not sure if they are. I think the folk are mostly very confused in their use of moral language. However, I think most folk would affirm the following statements

1. If everyone thought it was okay to torture infants for fun, it would still be wrong.

2. If everyone thought that the trans atlantic slave trade had been good, it would still have been bad.

3. If everyone thought the holocaust were good, it would still have been bad.

I don’t think one should always defer to the peers, though that should influence them a bit. My credence in hedonistic utilitarianism would be like 95% if it weren’t for smart people disagreeing, but it’s currently only about 66%.

3 Phenomenal introspection

The next section of Matthew’s article I would like to address is his response to evolutionary debunking arguments, which I generally regard as the best anti-realist arguments. When we tell the story of why we developed, say, our perceptual faculties, we will inevitably need to appeal to the reliability of those faculties: their ability to accurately reflect the world around us. However, research in evolutionary psychology and related fields can powerfully explain why we have the moral sensibilities we do without appealing to the truth of any moral assertions. Since we evolved our moral beliefs because they happened to be adaptive, it would be an enormous coincidence that these beliefs happen to align with the moral truth. The important part of this debunking argument is that in the case of morality, unlike other faculties, we can explain why we have the views we do without appealing to the truth of those judgements. Note that Matthew has very little to say about this core point as we look through his responses. 

First, I am going to truncate his first, second, third, and fourth points into one section since they are all fairly similar:

“First, as Sinhababu points out, we’d expect evolution to make us reliable judges of our conscious experience.  Belief in the badness of pain resists debunking because it’s formed through a mechanism that would evolve to be reliable.  Much like beliefs about vision aren’t debunkable, neither are beliefs about our mental states, given that beings who can form accurate beliefs about their mental states are more likely to survive…Third, there’s a problem of inverted qualia.  As Hewitt (2008) notes, it seems eminently possible to imagine a being who sees red as blue and blue as red, without having much of a functional change.  However, it seems like undesirability rigidly designates pain, such that you couldn’t have a being with an identical qualitative experience of pain, who seeks out and desires pain.  This means that the badness and correlated undesiredness of pain is a necessary feature, not subject to evolutionary change…Fourth, evolution can’t debunk the direct acquaintance we have with the badness of pain, any more than it could debunk the belief that we’re conscious.  Much like I have direct access to the fact that I’m conscious, I similarly have direct access to the badness of pain.  After I stub my toe my conviction is much greater that the pain was bad than it is in the external world.   [I put his first and third, and fourth points together here because I’d like to discuss them together] “

When Matthew talks in this article about having a ‘phenomenal introspection’ of the badness of pain I think he is seriously confused. First of all, I do not buy that our phenomenal introspection shows that there is a necessary link between pain and badness. There is significant pressure against the idea that pain is intrinsically bad from conditions such as pain asymbolia, wherein people experience the sensation of pain yet without any associated ‘badness’ or desire for the pain to go away, and from results in neuroscience showing that people under the influence of certain drugs can still recognize and identify being in pain yet without any associated negative valence (thanks to @ghostcoase for pointing out this latter research to me, and for the quote from the SEP below) [16]: 

“It has been well known that certain surgical procedures, some drugs and certain pathological conditions reduce or remove the unpleasantness of pain while preserving its sensory-discriminative aspect. These data typically come from patients who have undergone prefrontal lobotomy (Freeman et al. 1942; Freeman and Wattz 1946, 1950; Hardy et al. 1952; Barber 1959; Bouckoms 1994) or cingulotomy (Foltz and White 1962a, 1962b; White and Sweet 1969) as a last resort for their intractable chronic pain (as frequently involved in phantom limb pain, neuralgia, causalgia, severe psychogenic and cancer pains), from patients under the effects of hypnotic suggestion (Barber 1964; Rainville et al. 1997, 1999), nitrous oxide (laughing gas), and some opium derivatives like morphine (Barber 1959). These patients by and large claim that they are in pain, and they can recognize and identify their pain as such, but do not feel or seem bothered or distressed in ways characteristic to having pain experiences.”

The case of pain asymbolia I think doesn’t respond to the point — you can see a more detailed discussion from Rawlette in her book “the Feeling of Value.” But the basic idea is that it’s not pain itself that is technically bad; instead, there’s a basic quale of badness/unpleasantness/unenjoyableness. Pain asymbolia can make that no longer happen with traditionally painful experiences. However, this does change the experience to make it no longer feel bad — this is in much the same way that a drug to make you no longer disgusted by certain foods wouldn’t change the foods taste in one sense, but it would change the associated bad feeling.

Secondly, Matthew is conflating his phenomenal experience of pain with his belief that pain is stance-independently bad. It’s the latter thing that is going to ground realism, not the former thing. Even if it is your phenomenal experience of pain that leads you to form the beliefs you do about it, the second you form a belief you are now no longer simply ‘phenomenally introspecting’, you are implicitly using a particular conceptual framework (that of folk psychology) to interpret your mental states, and forming an internal representation. In the process of adopting that framework and forming that representation, you can obviously go wrong. To show that we have direct access to our beliefs that pain is wrong, you need to show that you somehow phenomenologically grasp your beliefs in a non theory-laden sense, something Matthew doesn’t even attempt and which I think is completely implausible. Therefore, this argument only works based on illicitly conflating what we (allegedly) have direct access to (our qualia) and our beliefs about our qualia. 

Two points are worth making.

1. If pain feels bad, that refutes error theory, even if it’s stance dependently bad.

2. I don’t know how the way that a feeling feels could even in principle depend on one’s attitudes towards it.

3. To quote my original article, where I respond to exactly the same point.

I agree that generally introspecting on experiences doesn’t inform us of their mind-independent goodness. But if we introspect on experiences that we don’t want but are pleasurable, they still feel good, showing that their goodness doesn’t depend on our desires.

I can remember when I was young I wanted to feel cold for some bizarre reason — small children have weird desires. Yet despite this, it still felt bad — this shows that it’s desire independent.

4 Debunking

I also don’t understand how “the badness and correlated undesiredness of pain is a necessary feature, not subject to evolutionary change” is actually a response to debunking arguments. Debunking arguments say that we can explain our moral beliefs only by appealing to why they are adaptive; we do not need to appeal to how they actually help us evaluate reality accurately like we do in the case of other domains (ie. vision). That pain would always be undesired is perfectly consistent with this: pain is consistently undesired because that is adaptive, not because it is a stance-independent moral fact that pain is bad! I think what Matthew means to argue is that there is a necessary connection between pain and the peculiar moral beliefs we have about pain. Even if this is true, it doesn’t seem to me that it undermines the evolutionary debunking argument so long as that connection is due to the fact that the connection is adaptive and not due to its truth. Irrespective, it definitely isn’t true. We are not bound to evaluate pain as morally bad for others, or even ourselves. Imagine a species who evolves in tribes where the head ‘Queen’ produces infertile offspring with high genetic relatedness to everyone in the tribe and especially to their queen, but very little genetic relatedness to other tribes. These mindless ‘drone warriors’ would benefit from sacrificing themselves for their leader and murdering and enslaving members of other tribes, and so easily could evolve beliefs about pain that lead them to disregard it when it happens to them or when they inflict it on others for the sake of their queen.

I’ll explain the argument because I don’t think that Panth’s responses are successful. If our belief that pain is bad could be debunked that would mean that we could have evolved with inverted qualia — having the type of responses we have to pain instead to pleasure and vice versa. But this doesn’t seem at all plausible — it doesn’t seem like beings could have evolved to feel exactly as they do when they’re in pain but to seek out that pain. The intrinsic bad feeling of misery (used more broadly than pain) seems like the type of thing that couldn’t be debunked.

Finally, I do not understand why Matthew thinks our beliefs about our basic conscious experience are not debunkable. As I illustrated at the beginning of this post, our basic beliefs about vision, the very process he said is not debunkable, turn out to be radically mistaken. All beliefs go through a basic representational process that involves adopting a particular conceptual framework, and so they can all be mistaken, in principle (and often are, in practice). And, the important point about our moral beliefs, is that we do not need any reference to them actually being true to explain why it was adaptive for us to hold them. As far as I can tell, Matthew did not attempt to respond to this basic fact about debunking arguments.

I do not think that our vision beliefs are “radically mistaken” — they’re certainly not infallible but that seems like a bit of an overstatement. However, discussing this in detail would require a bit of a digression and discussion of physicalism.

“Second, as Bramble (2017) points out, evolution just requires that pain isn’t desired, it doesn’t require the moral belief that the world would be better if you didn’t suffer.  Given this, there is no way to debunk normative beliefs about the badness of pain.”

First of all, this whole point is inconsistent with Matthew’s earlier insinuation that the necessary connection between pain and its undesirability was enough to respond to debunking arguments. If pain’s undesirability isn’t related to our moral beliefs about pain, then how could it possibly be relevant to realism or its debunking?

I don’t see the inconsistency. On my account pain feels bad and then by reflecting about it we can realize it is mind independently bad. However, evolution might explain why we don’t like pain but it wouldn’t explain why through reasoning we conclude that it is mind independently bad.

Regardless, this point, while partially true in that our moral beliefs about pain have no necessary connection to its undesirability, requires a naive view of the empirical evidence on the evolution of our moral beliefs. It isn’t just pain being undesired that is adaptive, rather, the core theories underlying these debunking arguments explain how our general moral emotions and jugements evolve. To illustrate, let’s take a quick detour into one of the most famous papers in applied evolutionary theory: the evolution of reciprocal altruism by Robert Trivers [17]. According to this theory, altruism can be beneficial for an organism if the incurred cost can later be reciprocated by a partner. However, this altruism needs to be discriminate: if we just helped others with no expectation of any assistance or response, then selection would favor the ‘cheater’ who takes help but never returns any help. It turns out that reciprocal altruism is the mathematically correct approach in repeated iterations of economic games like the Prisoner’s dilemma. How might this altruism play out evolutionarily? Well, as people became careful about who they were cooperating with, cheaters might want to develop more subtle methods of cheating, appearing as if they are cooperating without really doing so. This leads to an evolutionary back and forth where discriminant altruists will fine-tune their ‘cheater detection’ methods and cheaters will improve their subtle cheating. Anger protects people whose niceness leaves them vulnerable, guilt racks cheaters who are in danger of being caught, etc. We can see how it would be useful to propound and spread the idea that it is bad to inflict pain: the more this idea is spread, the less likely you or the people you care about are to be evolutionarily disadvantaged. Much more can be said about this topic, but suffice it to say that the theories underlying evolutionary debunking arguments have far more complex and sophisticated accounts of the origins of our moral beliefs, they do not merely appeal to pain being undesirable.

But we could evolve to just not like when we and others inflict pain. No part of this reasoning requires that we hold the judgment that it’s mind independently bad to inflict pain or that our pain is mind independently bad.

“Fifth, it’s plausible that beings couldn’t be radically deluded about the quality of their hedonic experiences, in much the same way they can’t be deluded about whether or not they’re conscious.  It seems hard to imagine an entity could have an experience of suffering but want more of it. “

Notice that Matthew here is not talking about our moral beliefs, he is talking about our desire not to suffer.

No — I’m talking about our belief that suffering is bad.

We might concede that it’s impossible for us to ‘continue to want to suffer and want more of it’ and use that in our evolutionary explanation of why we avoid suffering, and yet never reference the truth of any moral proposition. We might even use this to explain why we develop the idea that suffering is morally wrong, but, again, we don’t need to reference the truth of this belief to explain its evolutionary development. Maybe what Matthew wants to say is that we cannot be mistaken about the moral facts with regards to our hedonic experiences. I have already responded to this earlier, and Matthew provided no good evidence for that proposition here other than asserting it, again relying on conflating less controversial statements ‘one could never desire suffering’ with more controversial ones ‘one could never be wrong about the stance-independent badness of suffering’. 

I am directly acquainted with my mental states and through direct acquaintance I know that being in extreme agony is very bad.

The sixth response was not addressed, which I’ll quote here.

Sixth, there’s a problem of irreducible complexity.  Pain only serves an evolutionary advantage if it’s not desired when experienced.  Thus, the experience evolving by itself would do no good.  Similarly, a mutation that makes a being not want to be in pain would do no good, unless it already feels pain.  Both of those require the other one to be useful, so neither would be likely to emerge by themselves.  However, only the intrinsic badness of pain which beings have direct acquaintance with can explain these two emerging together.  

Panth next says

“Seventh, evolution gave us the ability to do abstract, careful reasoning. This reasoning leads us to form beliefs about moral facts, in much the same way it does for mathematical facts.”

This simply avoids the argument. According to the debunker, we need to reference the accuracy of our reasoning skills in the case of facts about things like math to explain why they would evolve. We do not need to reference moral truths to explain why our reasoning in the moral domain would play out as it does. 

This was an explanation of how evolution could allow us to arrive at moral truth. This does not, by itself, mean that this is the best explanation of our moral beliefs. However, we would then need to compare the theories and weigh up the considerations.

If our moral beliefs were the type of thing that wasn’t arrived at through reason and was just evolutionarily hypnotized into us we wouldn’t expect to be able to change our moral beliefs through careful reasoned reflection. Yet we plainly can.

Conclusion

I did not respond to everything Matthew said in this article, since a lot of it involved repetition of similar points, but I think I have said enough to deflate his core case. As I said at the beginning of this post, I am not a committed moral anti-realist, but I think realists need to do better than these intuitive dismissals of anti-realism in the future. 

I do not think that this deflates the core case. The two main arguments no addressed were

1. The argument from irrational desires (section 3).

2. The discovery argument (section 4).

Overall though, I think that Panth’s case was not successful. We do, in fact, have desire independent reasons to avoid unfathomable agony.