Philosophy, Film, Politics, Etc.

Wednesday, May 26, 2010

Sam Harris, Confused

The more I read Sam Harris on morality, the more annoyed I get. There has been a lot of good discussion in the blogosphere (1, 2, and 3, for example), and I've written about this twice already, but there is more to say. I want to make two basic points here. First, Harris fails to offer a cogent response to the main criticisms leveled against him. In one case, he ends up embracing a view which contradicts his own: While Harris claims that moral facts are scientific facts, he accepts that justification ultimately lies outside the boundaries of science (see point 7 below). This completely undermines his proposed view of morality. Second, while Harris sets moral relativism up as the bogeyman in the debate over science and religion, his vision of a scientific morality is nothing other than a variety of moral relativism (see point 4 below). In sum, Harris' position is a self-contradictory mess. If this wasn't clear from his TED talk, it is blatantly apparent in his latest attempt to defend his views.

1. First, Harris sets up moral relativism as the bad guy and fails to correctly identify the options on the table. He supposes the choice is between moral relativism, moral realism, or no morality at all. He writes:

the response to my TED talk proves that many smart people believe that something in the last few centuries of intellectual progress prevents us from making cross-cultural moral judgments -- or moral judgments at all. Thousands of highly educated men and women have now written to inform me that morality is a myth, that statements about human values are without truth conditions and, therefore, nonsensical, and that concepts like "well-being" and "misery" are so poorly defined, or so susceptible to personal whim and cultural influence, that it is impossible to know anything about them.

This is a jumble of ideas and does not clearly identify Harris' opposition. First, he suggests that some of his opponents believe that moral statements lack truth conditions--a position known as moral noncognitivism. Harris wrongly claims that this means that statements of human value are nonsensical. Noncognitivists do not suppose that moral claims are nonsensical. Moral judgments and expressions are meaningful and important; they just aren't factual assertions.

Harris believes that the only way to empower our moral faculties--the only way to justify our moral judgment of other people and other cultures--is to establish a scientific foundation for moral realism. I have already explained how Harris is presenting a false dichotomy between realism and relativism. There's no need to repeat myself here. What I will point out, however, is this: Of all the professional philosophers I have found who have responded to Harris, none have taken the moral relativism route. None have claimed that science cannot inform our moral judgments, or that moral judgments themselves are meaningless or inapplicable to people of other cultures or communities. (Sean Carroll, unfortunately, seems unclear on the distinction between moral relativism and moral noncognitivism. I think he really wants to argue for the latter--he claims that moral claims are neither true nor false, and he is quite explicit about his belief in the virtue of moral judgment--but he has not distanced himself from relativism. I recently pointed out this problem to him, but I don't know if he'll notice.)

2. Some of Harris' critics argue that moral disagreement is not based on questions of fact, and so cannot be resolved by appeal to facts alone. Moral disagreements are disagreements of goals. So, while we might be able to define morality in terms of "well-being, " this will not help us resolve moral disagreements, because "well-being" will be understood in fundamentally different ways. This doesn't mean that we cannot use science to understand well-being. It just means that the relevance of any particular scientific analysis of well-being is not a matter which could be determined scientificcally.

Harris responds:

Of course, goals and conceptual definitions matter. But this holds for all phenomena and for every method we use to study them. My father, for instance, has been dead for 25 years. What do I mean by "dead"? Do I mean "dead" with reference to specific goals? Well, if you must, yes -- goals like respiration, energy metabolism, responsiveness to stimuli, etc. The definition of "life" remains, to this day, difficult to pin down. Does this mean we can't study life scientifically? No. The science of biology thrives despite such ambiguities. The concept of "health" is looser still: it, too, must be defined with reference to specific goals -- not suffering chronic pain, not always vomiting, etc. -- and these goals are continually changing. Our notion of "health" may one day be defined by goals that we cannot currently entertain with a straight face (like the goal of spontaneously regenerating a lost limb). Does this mean we can't study health scientifically?

Harris' response is a non sequitur. He has not addressed the argument laid before him. Of course, as Harris says, we can study health scientifically, even though health is defined as a goal. But morality is not a goal. Rather, it is a process. The goal of morality is to establish agreement on what is and what is not justifiable behavior. Sure, morality, as a phenomenon, is just as open to scientific scrutiny as anything else. But there is a difference between making judgments about health and making judgments about morality. Science cannot tell us what we should do. It can only tell us what are the likely consequences of our actions.

3. Sean Carroll indicates that, if we are going to scientifically ground morality in terms of well-being, we must first agree on some methodological principles. We must have some way of identifying well-being in objective terms. Yet, not everybody agrees on what constitutes well-being. It is not a scientifically defined entity. Any attempt to define it scientifically will therefore likely exclude many plausibly legitimate views of what well-being is all about.

Harris responds with more confusion. He correctly points out that Carroll can make "a reasonably principled decision about whom to put on a panel at the next conference on Dark Matter without finding a neuroscientist from the year 2075 to scan every candidate's brain and assess it for neurophysiological competence in the relevant physics." This response is absurd, and shows a failure to grasp Carroll's point. There is relatively little disagreement about what counts as expertise in the subject of Dark Matter. Similarly, there is relatively little disagreement about what counts as expertise in the subject of morality. While intelligent people will surely disagree about who is the most important or relevant moral thinker, we can expect certain sorts of people to be invited to a conference on morality or metaethics. But this fact in no way undermines Carroll's point, which is that experts on morality and metaethics do not require a consensus on what constitutes well-being. And, if they do try to find a consensus, they very rarely treat it as a question of natural fact. The scientific study of morality does not require that people agree on what constitutes well-being. On the contrary, I think all evidence points to the opposite conclusion: morality is such a dynamic and often tumultuous process because people often do not approach well-being in the same way. While Harris has every right to say that people should approach well-being in some particular way (though he hasn't defined what that way is yet), his will be one voice out of many. Which is not to say that his voice doesn't count. It only means that, if he wants his views of well-being to be taken seriously, it doesn't help for him to pretend like alternative views are simply irrelevant.

4. Harris writes: "I would say that more or less everyone, myself included, is insufficiently interested in [universal well-being]. But we are seeking well-being in some form nonetheless, whatever we choose to call it and however narrowly we draw the circle of our moral concern." According to Harris, people all define well-being differently--they pursue it differently, according to their unique goals and circumstances. However, if facts about well-being make our moral judgments either true or false, and our own moral judgments are defined by the concerns of our own private moral spheres, then our moral judgments are really only true within that circle of concern. That, ironically, is the essence of moral relativism.

Harris has contradicted his claim that morality is an attempt to maximize the well-being of all conscious creatures. It turns out nobody is really interested in that. He has also contradicted his claim that a science of morality would overcome moral relativism. On the contrary, it looks like Harris' scientific approach has provided nothing more than a foundation for moral relativism.

5. Harris agrees with Carroll that what is important in our moral judgments is not just whether or not our actions produce certain neurological states, but how those states are produced. Yet, Harris offers a problematic response to the virtual reality example in which people every citizen was unknowingly hooked up to virtual reality machines which gave them orgasmic sensations all day, and made it unnecessary for them to eat or procreate. Carroll observes that people might have moral objections to this way of life, even though it would maximize pleasure. Harris' response is that we don't want to be delusional, or removed from real contact with other people. Yet, if those desires conflict with our ability to maximize our well-being, then--by Harris' own argument--those desires are irrelevant. Harris is cherry-picking, choosing exceptions to the rule when they suit his moral sensibility.

6. Harris claims that neuroscience should be able to give us a range of brain states which most qualified people would recognize as "good." However, this does not mean that any particular social norms would tend to maximize those states for all people. It is a well-known fact that people aren't all turned on by the same turns of events, or the same exercises of freedom. So the fact that we might define some brain states as obviously "good" (in the sense of being pleasurable) in no way helps us establish a foundation for morality.

7. There is a common view that scientific conclusions are not prescriptive. If we want to turn a scientific conclusion into a moral precept, we must move outside the boundaries of scientific discourse. It impossible for science alone to ground morality.

Harris' response to this argument is stunning, and shows just how incoherent his position is. He says, "we must smuggle in an 'unscientific prior' to justify any branch of science. If this isn't a problem for physics, why should it be a problem of a science of morality?"

Harris's claim is that science (and, presumably, any behavior) can only be ultimately justified by appealing to some unscientific principle. Thus, you cannot use science to justify science, or anything else.

Consider that for a moment. If you want to justify science, you are no longer working within the realm of science. That is what Harris said. And, of course, he's right! But this is exactly what his many detractors have been trying to tell him. If you are trying to justify something, you are doing morality, and this cannot rely on science alone. Moral judgments are not scientific conclusions, though scientific conclusions can be used in the formation of a moral judgment.

Harris accepts that, if you want to justify your move from scientific conclusions to moral precepts, you must "smuggle in" unscientific judgments. He has embraced the view that morality cannot be given a wholly scientific foundation--that attempts to justify our behavior, such as our scientific discovery itself--require unscientific judgments. That should be the end of the debate. Harris has rejected his position. Game over.

8. There is one more minor point to address (not as stunning as the previous one, unfortunately). Harris writes:

To say that morality is arbitrary (or culturally constructed, or merely personal), because we must first assume that the well-being of conscious creatures is good, is exactly like saying that science is arbitrary (or culturally constructed, or merely personal), because we must first assume that a rational understanding of the universe is good. We need not enter either of these philosophical cul-de-sacs.

If it is not clear already, this is a straw man. Harris' detractors (of whom I am aware) do not claim that morality is arbitrary or "merely personal." However, I do imagine that many probably do recognize the role of culture in the formation of our moral judgments. I would be surprised if Harris was going to deny that culture played a significant role there.

In sum . . .

Harris wants to open the public debate to a discussion of the role of science in morality. I don't think many, if any, of his detractors are against that proposal. Some, including Sean Carroll, have explicitly embraced it. Unfortunately, Harris' arguments are probably hurting more than helping such a discussion.

Wednesday, May 19, 2010

Sam Harris and the Moral Realism/Moral Relativism Myth

There is a popular misconception that, if you are not a moral realist, you are a moral relativist. Yet, most non-theistic professional philosophers may be neither. As I'll explain, I prefer a variety of anti-realism called moral noncognitivism. First, I want to discuss the moral realism/moral relativism myth.

Moral realists believe in objective truths about right and wrong, whereas moral relativists believe that moral truths are subjective, or limited by the beliefs and values of particular communities and cultures. Moral relativists are people who say, for example, that genital mutilation isn't absolutely or universally wrong--it's just wrong for some people. Moral realists respond, "No, genital mutilation really is [or, perhaps, really isn't] wrong, for all people and all times--all things being equal."

This is usually how the issue is framed in popular culture, particularly in debates over atheism and the role of religion in society. It is not uncommon to hear religious moralists claim that atheists have no foundation for morality, and that they must be moral relativists. Without religion, the religious moralist claims, science would produce as many holocausts as cures, as many massacres as breakthroughs. A wholly atheist society could not have any moral sense at all. In the debate over religion, moral relativism is the bogeyman.

Sam Harris recently put it thus:

But the consequences of moral relativism have been disastrous. And science's failure to address the most important questions in human life has made it seem like little more than an incubator for technology. It has also given faith-based religion -- that great engine of ignorance and bigotry -- a nearly uncontested claim to being the only source of moral wisdom.

Harris believes that, to overcome moral relativism without empowering theism, we must establish a scientific foundation for moral realism.

Some atheists accept the charge of moral relativism, but not the allegedly "disastrous" consequences. They say that the fact that there are only local moral truths is good enough--nay, it has to be good enough, because that's all there is. We're all just trying to get by, doing the best we can with what we've got. Nobody has an absolute foundation for condemning anything as immoral in any absolute sense. We only judge people according to how we want people to live. We don't have a choice, because this is just the way we're made.

Other atheists, like Sam Harris, embrace moral realism. Slightly more than half of the philosophy professors and PhDs who partook in a recent survey either accept or lean towards moral realism, while only 16 percent accept or lean towards theism. This means that, like Harris, a large number of serious philosophers are non-theistic moral realists. That said, I do not think a large percentage of those philosophers would agree with Harris' attempt to regard moral facts as scientific facts. For example, Jean Kazez, a philosophy professor and blogger, defends Sam Harris' moral realism: "In terms of addressing the religious moralist, what would [Harris's co-non-religionists] have him say? That there really aren't any facts about morality? That torturing babies for fun isn't really wrong? What a public relations disaster!" However, while Kazez thinks that the moral realist line is "strategically right," she disagrees with Harris' view that moral facts are scientific facts. According to one commenter on Kazez's blog, Harris does a disservice to atheism for even suggesting that moral truths require some kind of foundation. While atheism may not offer a foundation for moral facts, neither does theism. The problem of understanding moral facts is there for both the atheist and the theist. (The only difference, I suppose, is that the theist gives herself license to ignore the problem. That is the point of faith, isn't it?)

Interestingly, then, we have moral realists who are atheists, and who reject Harris' idea of grounding morality in science. Yet, for Harris to remain true to his naturalism, he has to think of moral facts as natural facts. So how could they not be scientifically discoverable? This is one problem with moral realism: if moral facts are not reducible to scientific facts, then what sort of fact are they? What place can normative facts have in nature?

As problematic as moral realism is (see the Stanford Encyclopedia of Philosophy entry on moral realism for an overview of the main problems), Harris is going out on a peculiar limb by claiming that moral facts are a particular variety of scientific fact. This wouldn't be so bad, were it not for the fact that his arguments are very weak. His upcoming book will not likely impress any religious moralists, or any non-religious philosophers who have spent any time on moral philosophy. Sean Carroll has eloquently pointed out some of the main problems with Harris' arguments. (See this post for my own analysis of where Harris goes wrong: science and morality). Harris' arguments aren't just weak; they're self-defeating. Harris' main point is that we shouldn't hesitate to condemn, for example, the obligatory use of burkas in some Muslim states. Yet, according to his view of moral claims as scientific claims, we should wait for strong scientific data before we can be sure we have discovered any moral facts. He suggests that, among other things, neuroscientific measurements of "well-being" must be involved. So, in effect, we cannot confidently make any moral judgments until we have generated data and published our results in peer-reviewed scientific journals. This cripples, rather than empowers, our moral faculties.

More important, and more dangerous, than the poverty of Harris' position is the fact that he is playing into the false dichotomy between moral realism and moral relativism. He suggests that moral realism is the only plausible alternative to moral relativism. This hurts the debate. Moral realism and moral relativism are not the only options on the table. In fact, it seems that moral realism and moral relativism are rejected by the majority of non-theistic philosophy professors and PhDs.

According to the PhilPapers survey, about 30 percent of philosophy professors and PhD's accept moral realism, while about 25 percent lean towards it. A 44 percent minority rejects moral realism. Before we interpret these results, we should correct for theism. About 16 percent of the same group either accept or lean towards theism, and this same 16 percent probably accepts or leans towards moral realism. Therefore, perhaps only 40 percent of professors and PhDs are both non-theists and moral realists, while 44 percent are non-theists and not moral realists. If we are to trust the results of this survey, then, the majority of non-theistic philosophers reject moral realism. According to the survey, of those who reject moral realism, about two thirds either accept or lean towards moral anti-realism. Moral anti-realists are usually either noncognitivists or error theorists. Though the survey is not helpful on these details, I think it is a fair bet that, of the remainder, very few accept or lean towards moral relativism. (Again, the Stanford Encyclopedia is helpful in explaining why moral relativism isn't so popular.)

As I said, I like moral noncognitivism. This is the view, widely respected and often supported, that there are no moral facts at all. Contrary to what Kazez says, this does not mean that torturing babies isn't really wrong. We can unhesitatingly say that it really is wrong; however, when we say that, we are not making a factual claim. It's not that our statement is not true. Rather, it is that our statement has no truth conditions. We are not saying something which could ever be either true or false. It's just not that kind of speech act. What we are doing when we make moral claims is subject to debate. The view I prefer is that we are directing human action, not by reporting a fact about the world, but by stigmatizing certain behaviors.

Sam Harris is right to draw our attention to the ways science informs our moral deliberations. A moral noncognitivist can appreciate the role of science in morality. After all, our desires and how we express them are both shaped by what we know and think about the world. Yet, there is a world of difference between forming a belief about the world and deciding how we want the world to be. The former is a matter of fact, the latter a matter of morals.

So, yes, I can say unflinchingly that torturing babies is "really wrong." That does not make me a moral realist. It just makes me a person who wants to live in a world where babies aren't tortured. Like the moral relativist, I can acknowledge the role of culture and circumstance in the formation of my moral judgments. But, unlike the moral relativist, I do not claim that those judgments are facts which are only "true for me." They're not facts at all. They're judgments of a different sort.

Harris says he thinks the world is worse off for philosophical discussions that involve words like "noncognitivism." On the contrary, a better appreciation for the philosophical issues is exactly what the world needs. Harris' refusal to even acknowledge the legitimacy of well-known and widely respected philosophical views is curious. It will make it difficult for serious philosophers to take his work seriously. More than that, it should make us all afraid of the effect his upcoming book will have on the public debate. I cannot see it helping the case for atheism.

Sunday, May 16, 2010

Science and Morality

In February, popular writer and atheist Sam Harris gave a talk about science and morality, which you can watch here:

He recently defended his views at The Huffington Post, where he spends a good deal of time responding to physicist and pop science writer, Sean Carroll. I'm not going to address the Carroll-Harris debate. I just want to look at some strengths and weaknesses of Harris' original argument.

The first point to mention is that Harris makes a serious error in his original presentation, and which he does not correct in his Huff Post entry: According to Harris, the only people (other than himself) who believe there are objective moral facts are religious demagogues. Yet, according to the preliminary results of a recent PhilPapers survey, slightly more than half of the professionals and PhD's surveyed accept or lean towards moral realism--they believe there are objective moral facts--while only sixteen percent accept or lean towards theism. Moral realism remains a heavily discussed and supported position, and clearly this is not out of any connection to theism. There is thus no reason for Harris to claim he is alone in his argument for a non-theistic objective morality. Either he is ignorant of much of the literature on the matter, or he just wants his arguments to appear more original and groundbreaking than they are. That said, I am not wholly persuaded by moral realism. As I will make clear, I think morality is objective, but not necessarily factual. I will explain, first by briefly presenting my view of morality, and then critiquing Harris'.

I take morality to be a process of establishing and revising the most justifiable prescriptions for human action. Moral precepts are judgments about what people should do based on the prevailing rational arguments, given various contextual constraints. The process of establishing these conclusions is a rational and objective one: It requires argument based on reason and evidence. However, that does not mean that such conclusions are scientific conclusions. It does not mean that we can, for example, define moral questions in testable terms which could then be answered with controlled experiments. However, it does mean that we can use scientific facts in the formation and refutation of moral arguments. Scientific evidence can be used to guide moral reasoning, but this does not mean moral reasoning is scientific reasoning.

The difference between science and morality is most easily (and most commonly) described by observing the different sorts of conclusions they produce. Scientific conclusions are essentially methodological: They describe methods for producing specific results, and (in the more formal sciences) they define the relationships between the methods and the results in mathematical terms. Scientists then use these relationships to predict how the world will behave. Moral conclusions are quite different. They are decisions about how we want people to behave, and not about what will probably happen in any particular circumstances.

This does not mean that moral precepts cannot be understood by science. Science can tell us why people make one moral judgment and not another. So the existence of moral precepts is not outside the purview of scientific methodology. Science can explain morality.

So, if Harris' point is only that science can help us understand morality, and can help us solve moral dilemmas, then I agree with him. However, if his point is that scientific experiments can provide evidence of the truth of moral precepts, then I think he is confused about what morality and science are.

His basic claim is that moral judgments are facts about the well-being of conscious creatures, and that these facts can be measured in much the same way that facts about health are measured. His view is that "morally right" is just another way of saying "healthy for the prosperity of conscious beings." While we can disagree about what consciousness, health, and prosperity are, this does not detract from the essentially scientific nature of the way we should approach the issue.

True, moral judgments strongly focus on conscious beings. But how do we establish a scientific concept of prosperity? Prosperity for whom? For the individual, the family, the community, the entire species? For all conscious life?

We might suppose that the object of moral concern is not important here. Some moral judgments are relevant to individuals, some to families, some to communities . . . Yet, what happens when the values of one community conflict with the values of another? Or when the values of the community conflict with the values of the individual, the species, or some set of non-human species?

The complexities involved in even the most common of moral disputes suggests that there may not be clear answers to them. We strive for the clearest arguments and judgments--at least, we should. But this does not exclude the possibility of better arguments which could undermine our judgments. We make moral judgments, not by observing a moral fact, but by making a decision about how we want to live. Mutually exclusive judgments may be equally supported by the facts. Thus, I believe that moral questions are not always, if ever, answerable questions of fact. (I should note that I have in the past tried to argue for at least some moral facts - I think there may be some moral facts about morality itself, for example - though I won't get into that here.)

In contrast, Harris argues that science can teach us moral facts. Unfortunately, he does not make a good argument. On the one hand, he has not shown how this could happen. It remains to be seen what a scientific demonstration of a moral fact would look like. But, even with that problem laid aside for the moment, Harris' argument seems problematic.

Harris claims that we should not hesitate to judge, for example, the way some Muslim women are forced to cover their entire bodies, or the way religious parents and teachers promote corporal punishment in public schools. The problem with his argument is that, even if we agreed that science could tell us whether or not women and children should be treated this way, he has not made the case that science has provided any clear conclusions on these or any other cultural matters.

Perhaps Harris' point is that we should be more willing to act on the (limited and tentative) conclusions of scientists when making moral judgments. Yet, it doesn't seem that people are ignoring any particular scientific conclusions here. It's not like we needed Sam Harris to tell us that we should respect the scientific evidence of harmful effects of various cultural practices. So, if Harris just wants to argue that we should be more aggressively against practices which are known to be harmful . . . well, then he doesn't need to make any deep philosophical arguments about science and morality. He just needs to make sound arguments about the dangers of those various practices. So why is he getting all philosophical?

I suppose it is because some religious people criticize atheists on the grounds that atheists have no moral foundations. To overcome this tired argument, atheists don't need to claim that moral judgments are scientific facts. They just need to explain why theism is incapable of providing a foundation for morality, and why no such foundation is needed. (See, for example, Atheism and Morality.) Harris' response to these misguided religious moralists is not helpful.

Consider what sort of intellectual commitment Harris is making for himself. If he wants to remain as scientific as possible, he should be open to the possibility that people who do many things he finds morally reprehensible might, in some ways, be more prosperous than people who do not. Maybe evolution favors burkas over bikinis. Maybe corporal punishment in public education is evolutionarily advantageous. Certainly no tests have suggested otherwise, so Harris is asking us to make moral judgments in the absence of strong scientific data.

Perhaps we could use computers to run simulations of civilizations which, for example, were more oppressive of women, and more violent towards children. What if we found that such civilizations had a small but significant survival advantage over those which were more egalitarian? Should we then say that the Muslims were right, and that all women should wear burkas? Should we make corporal punishment a mandatory part of public education?

No, I don't think anybody would draw those conclusions. While we might admit that burkas did confer some benefits to the societies which obligatorily enforced them, or that corporal punishment did have its benefits, we would find reasons to reject the conclusion that they were morally right. We might do this by claiming that the benefits given to society were outweighed by the harm done to individuals, families, or communities. Of course, such harm would have to be demonstrated scientifically. But how would we scientifically measure the relative harms here? What scientific test could tell us how much personal harm is required to reject a behavior which was clearly evolutionarily advantageous? There is no calculus for that, and there is no reason to think that there could ever be one, because it is not obviously a question of fact.

Of course, if respectable studies showed that burkas and corporal punishment did have significant advantages for the well-being of civilization, our moral discourse would change accordingly. The science would be relevant, no doubt about it. But the outcome of the scientific investigation is only a tool to be used in the argument for a moral judgment; it is not the moral judgment itself. That is the point which Harris seems to be missing.

I applaud Sam Harris for arguing for a rational approach to morality, and for rejecting the view that you need religion to justify your moral judgments. However, his desire to find scientific answers to moral questions is unfortunate. It is likely to give religious moralists more, not less, cause to criticize atheists.

Wednesday, May 12, 2010

Defending Hate

The infamous atheist, PZ Myers, recently posted a short bit about a legal matter in Poland on Pharyngula, his popular science blog. It turns out that Poland has a law against insults to religion. (A number of EU countries have such laws.) If you offend religious sensibilities in Poland by "publicly profaning an item of religious cult or a place intended for publicly performing religious rituals," then you are "subject to a fine, restriction of freedom or imprisonment up to 2 years." I have been told by students and practitioners of Polish law that it is very difficult to offend religious sensibilities in a legally recognized way. The law is intended to protect the freedom of religious expression and prevent a sort of hate speech. It is not used to silence objections to, or criticisms of, religion. Whether or not the law is used to infringe upon the freedom of speech is a question to be determined by the application of the law, and not its letter.

As it turns out, it seems that it is very hard to find examples of cases where this law has been applied at all. It was applied in 2003, and resulted in a probationary community service sentence against Dorota Nieznalska, an artist who had created an art exhibit which featured male genitalia on the cross. However, Nieznalska eventually won an appeal. Otherwise, it does not seem that this law has resulted in any convictions at all. This could be because Poles are not so sensitive to public offenses to religion, or it could be because Poles have a strong tendency to publicly respect religious objects and institutions. Either way, it is at least possible that this law could be used to infringe upon the freedom of speech. The only way to find out, I suppose, is to test it.

Enter Doda, a pop star who made a provocative comment about the authors of the Bible. Some Catholics in Poland are trying to bring a legal action against Doda. As Myers observes in his blog, what Doda said is not deserving of punishment. If she is taken to trial and loses, it will look bad for Poland's legal system. Fortunately, I hear there is no way Doda will be punished for what she said. So, if and when she is brought to trial, she will probably win, and it will prove nothing bad about Poland or Polish law--except, perhaps, that it is too easy to file law suits and get in front of a judge in Poland--but that problem is not specific to "religious insult" laws.

It might seem like I agree with Myers. Not quite. Myers titles his blog post, "Shame on Poland," as if the entire country, or even a significant portion of its population, had done anything wrong. All that has happened is that a conservative fringe has made a fuss. Myers says that the people who are responsible for this situation are "touchy little cowards." He is probably right--but he is using this as an opportunity to criticize the entire country, or at least a significant portion of its population. More, he implicates all of Europe, suggesting that all Europeans are hypocrites for criticizing the way Muslims respond to depictions of Mohammad. (According to some conservative Muslims, visual depictions of Mohammad are punishable by death, a view which has led to death threats and violence around the world, even against people who are not Muslim. Recently, for example, a cartoonist was attacked in Sweden.)

There are two things to observe about this comparison. On the one hand, what is so upsetting to many Europeans is not that some Muslims are offended by depictions of Mohammad. It is that they consider it a capital crime and are willing to threaten and execute anybody anywhere for breaking their rules. It is absurd to compare that to the fuss made in Poland over Doda. On the other hand, how can Myers suggest that Europe itself--not just Poland, and not just the European Union, but the entire continent--was somehow implicated in this Doda situation? This is also absurd. Not just absurd, but irresponsible and damaging. I think Myers has let his emotions get the better of him.

PZ Myers has very strong and no doubt well-earned feelings about religion and conservatism. I wouldn't criticize him for that, and not just because I share those same feelings. I do share them, or at least a lot of them, even if I don't always agree with the way he expresses them. I often enjoy his well-thought-out critiques and commentaries, and I often agree with his perspective. Unfortunately, he is not always careful; and considering his influence, that is very irresponsible.

According to the Penguin Dictionary of Psychology, hate is a "deep, enduring, intense emotion expressing animosity, anger, and hostility towards a person, group, or object." This is a healthy emotional response, and it serves many useful functions in our lives. I have no doubt that Myers hates many expressions of religion and conservatism. I hate them, too. For example, in another recent post, Myers talks about the brutal execution of a 13-year-old girl--a Muslim--whose only crime was to have been gang-raped. Myers says he does not want to share the world with people who could do that. I understand that sentiment. I can even agree with it. And I can think of no better word to describe its emotional foundation than "hate." I hate what happened to that 13-year-old girl, and I have no doubt that Myers does as well.

There is no reason to deny such feelings. The fact that you hate something does not preclude you from making a cogent argument against it. Emotions do not always interfere with our rational judgment. They also guide us. Consider: pro-life activists hate abortion; pro-choice activists hate unreasonable restrictions on a woman's right to control her own body. Does that mean that neither side has a legitimate argument to make? Of course not. The fact that an argument or perspective is motivated by hate is no reason at all to reject it.

If your adversaries in a debate criticize you for your hatred, they are not criticizing your arguments. They are trying to avoid your arguments by criticizing their source. This is called poisoning the well. It is not a valid counter-argument, and it should not be respected. The response is not to say, "oh, no, I don't hate!" The proper response is to say, "that is not a valid counter-argument; you are evading my argument." Don't pretend you don't hate. It is not convincing, and it does not further your cause. It only lets your opponents use your emotions against you. There's nothing wrong with strong emotions.

All of these issues came up in the comments section of Myers' post. I do not normally post on Pharyngula, and I do not normally read the comments section, so I went into the discussion without any prior knowledge of the participants. What I found was interesting. At first, a number of people were very supportive of my position. (I should mention that others had already pointed out some of the problems with Myers' post before I joined in.) However, it soon became apparent my use of the h-word was a problem. The Pharyngula community does not like to be accused of hating. It turns out that religious conservatives try to undermine their arguments by claiming they are expressions of hatred, so they have (perhaps tacitly) agreed never to claim to hate anything related to religion or conservatism. I find this deeply troubling.

I should also mention that I was not just concerned about the way PZ Myers misdirected his hatred. I was concerned with the way other people jumped on the bandwagon. Several people who responded in the comments section echoed sentiments of hostility and anger at Poland and the Polish people. One person claimed to have been happy about the tragic deaths of scores of Poland's military and civilian leaders last month. (This individual probably didn't bother to read about all of the people who had died, or he may have found out that they weren't all religious conservatives. For example, one was going to be the next Presidential candidate for the liberal party.) Another person said he was no longer going to be offended by Polish jokes. Yet another said the situation was disgusting, and that Poland was uncivilized.

As an American ex-pat living in Poland and married to a Polish woman, I felt a duty to defend my host country and its people against such unjust attacks. I was not defending religion or conservatism. I only criticized the misdirection of strong emotions. And, as should be clear by now, I was not criticizing those emotions themselves. I was only pointing out that (1) there were strong, negative emotions being misdirected, and (2) this made the posts not only absurd, but very irresponsible and sad.

As I said, some people agreed with and appreciated my comments. But then people started distancing themselves from my use of the words "hate" and "hatred." I was met with a little hostility and, finally, I was accused of being a "concern/tone troll." This apparently justifies ignoring me: Since I have been called a troll, nobody else has posted.

It is also worth noting the feeling of solidarity I sensed among the participants. Many felt very comfortable speaking about the emotions of the group. Perhaps they are all so connected to each other, they know that none of them hates anything. One wrote, "Hate requires energy which we don't have." Right. They have energy to post at length about the horrible effects of religious faith--but not the energy to have strong feelings about it?

Hopefully leading members of the atheist and liberal communities can make a change here. We shouldn't feel guilty for having deep-seated emotional responses to injustice. We shouldn't pretend that our arguments and actions are not motivated by our feelings. There's nothing wrong with hate, or expressions of hatred; just so long as those expressions are healthy and constructive, and not misdirected.

Thursday, May 6, 2010

Swamp Deviants, Part II

Peter Mandik sent me a draft of a paper in which he addresses one aspect of my criticism of Swamp Deviants. The issue, which I discussed here a couple of days ago (Deviant Phenomenal Knowledge) and elaborated upon earlier today (Can Swampkinds Be Deviants?), is this: Can a Swampkind have deviant phenomenal knowledge--that is, can they know what it is like to see red without first seeing, imagining, or hallucinating the color red?

A Swamp Deviant is a Swampkind who knows what it is like to see red without having ever seen, imagined, or hallucinated the color red. One objection to the idea of Swamp Deviants is the argument that Swampkinds cannot know anything at all, because they cannot have any intentional states. To have an intentional state which is about red, for example, you must have a causal history which connects your state to the color red. Knowledge supervenes over causal histories, and not simply over neurological states. Since Swampkinds have no causal histories at all, then a Swampkind cannot be born with knowledge of any sort. They could only learn what it is like to see red by having the right kind of experience. An experience that generates such knowledge is what I call an "earning experience."

Mandik argues that, if this objection to Swamp Deviants works, then it can work against the knowledge argument itself. According to the knowledge argument, Mary the super-scientist cannot learn what it is like to see red because she has no access to color experiences (other than experiences of black and white), even though she learns every scientific fact about colors from inside her monochromatic existence. Yet, according to the historicist, Mary's scientific knowledge of colors would have to be causally connected to redness. Thus, Mandik argues, we no longer have a reason to suppose that black-and-white Mary doesn't know what it is like to see red.

I think defenders of the knowledge argument have a response here. (Though I am no defender of the knowledge argument, as my previous posts make clear enough. See, for example, What Zombie Mary Knows.) They can claim that what the knowledge argument shows is that Mary must have the right kind of causal history before she can gain the right phenomenal knowledge. Yes, her scientific knowledge must be causally related to redness; but it is not causally related in the right way. I therefore do not think Swamp Deviants pose a serious threat to the knowledge argument.

In any case, I have another historicist argument which defenders of the knowledge argument could use against the possibility of Swamp Deviants. If phenomenal knowledge is propositional knowledge (as the knowledge argument requires), then it entails justified true beliefs. (This is generally accepted, even if justified true belief is not universally accepted as sufficient for propositional knowledge.) Thus, we should be able to talk about phenomenal beliefs. We could presumably also talk about true phenomenal beliefs which lack justification, and which therefore are not knowledge. What better than a Swampkind could exemplify this?

We could grant that a Swampkind is born with true, though unjustified, phenomenal beliefs. What supervenes over causal histories is not the belief itself, but only the justification for the belief. In this case, Swampkinds are not born with phenomenal knowledge. They are born with true phenomenal beliefs which await justification. There is an intuitively appealing idea that motivates this argument: The only thing that could justify a phenomenal belief is an earning experience. If this intuition is correct, then, by definition, a Swamp Deviant's phenomenal beliefs cannot be justified, and a Swamp Deviant cannot be a deviant at all. In short, if (1) phenomenal knowledge is propositional knowledge and (2) the only way to justify a phenomenal belief is with an earning experience, then (3) Swamp Deviants are logically impossible.

[Update, May 7, 2010: It occurs to me that the two premises in this last argument, if correct, establish the logical impossibility of deviants of any sort. But I am not convinced these premises are sound--more specifically, I am not convinced that phenomenal knowledge is propositional knowledge.]

Can Swampkinds Be Deviants?

In my last post, I challenged philosophical appeals to deviants. Deviants are beings who have phenomenal knowledge (they know what it is like to X), but who have not had any experiences which could "earn" that knowledge. For example, a deviant is somebody who knows what it is like to see red, though who has never actually seen, imagined, or hallucinated red. I previously argued that deviants are plausibly impossible--the very idea of a deviant is not obviously coherent--and that, if we do our best to make the notion coherent, the possibility of deviants becomes an empirical question. Here I want to further support the argument that deviants may be inconceivable.

The argument I will now develop is that Swampkinds cannot be deviants--or, rather, that any claim that Swampkinds are deviants will carry undesirable philosophical baggage. Swampkinds are beings who are physically and functionally identical to human beings, and who have phenomenal consciousness, but who are created spontaneously as the result of a random bolt of lightning.

In my last post, I mentioned the Davidson-Millikan view, which states that we cannot attribute intentional states or attitudes to Swampkinds, because they lack suitable histories. To see the basic rationale here, consider the following scenario: a bolt of lightning strikes a jug of paint, causing droplets of paint to splatter over a canvas. The result is what appears to be a realistic portrait of an actual tree which is standing nearby. The image may be as realistic as we like, so that anybody would be justified in supposing that it was the work of a very talented artist. Is it a representation of that tree? According to Davidson (and Kripke and Putnam), it is not a representation of anything. It is not a picture of a tree, though it looks just like one. The paint-on-canvas lacks a suitable causal history, and so cannot be said to be of anything at all. This does not prevent it from being used as if it were a picture of a tree, of course; but knowing what we do about its causal history, we are limited in what we can say about it. We cannot, for example, call it a painting, even though there is no observable difference between it and a painting. What licenses us to call something a painting (or a representation, or as something exhibiting intentionality) is the right sort of causal history; without that, we cannot make any attributions of intentionality at all. So it is with Swampkinds--they lack causal histories which could make their neurological states about what they are supposed to be about. So they cannot have the requisite phenomenal knowledge. They cannot be deviants.

To claim that Swampkinds can be deviants, then, we must suppose that there is something intrinsic about certain neurological states which gives them intentionality. This is a philosophical commitment which is not so easy to make.

Dennett might object here on the grounds that I am excluding an alternative hypothesis: Intentional states need not be attributed on the basis of causal histories, nor need they be attributed based on intrinsic properties of neurological states. Rather, we are justified in attributing intentionality just in those cases where such attributions lead to useful predictions about the behavior of a system. Since the behavior of Swampkinds can be accurately predicted on the basis of such attributions, we are justified in attributing intentional states to them. This is the gist of Dennett's "intentional stance" view of intentionality (Dennett, 1987).

The problem here is that, according to the intentional stance, we cannot attribute intentional states on the basis of how a system might behave. Rather, we can only attribute intentional states when the behavior of the system justifies it. In the case of a Swampkind, attributions are only justified when some behavior justifies them. That is, we cannot say that a Swampkind has the requisite phenomenal knowledge--we cannot know that it is a deviant--until it has exhibited some behavior which justifies the attribution of the right intentional states. Yet, any such behavior would have to entail an experience which earns that phenomenal knowledge. So, if we adopt Dennett's intentional stance, we are only able to claim that those Swampkinds which demonstrate the right phenomenal knowledge are deviants; yet, by virtue of such demonstrations, they cannot be deviants. Thus, the intentional stance would seem to preclude the possibility of Swampkinds being deviants.

In conclusion, the claim that Swampkinds are deviants relies upon a difficult philosophical commitment: We must regard intentional states as intrinsic properties which have no relation to causal histories or possible future behaviors. This makes the appeal to Swampkinds in discussions of deviants far less appealing.


Dennett, Daniel C. (1987) The Intentional Stance. MIT Press.

Tuesday, May 4, 2010

Deviant Phenomenal Knowledge

[Updated on May 5, 2010]

Phenomenal knowledge is normally defined as "knowing what it is like to experience [something]". This is usually further defined either in terms of, or in terms closely related to, the abilities to recognize, remember, and imagine experiences. Some recent philosophical discussions have focused on the issue of deviant phenomenal knowledge. Torin Alter (2008) explains:

Phenomenal knowledge is earned if the experience requirement is satisfied. For example, since I have seen ripe tomatoes, my knowledge of what it’s like to see red is earned . . . To access phenomenal knowledge is to exercise closely related abilities, such as the ability to imagine, recognize, or remember relevant experiences. I access my phenomenal knowledge when I visualize a ripe tomato, stop at a traffic light, or have an episodic memory of seeing oxygenated blood. Phenomenal knowledge that is unearned, inaccessible, or both is deviant.

In what follows, I use "deviant" to refer to anybody who has deviant phenomenal knowledge. If deviants are possible, then it is possible for a person to know what it is like to see the color red without ever having seen, imagined, or hallucinated the color red. Discussions of deviants have focused on whether or not they create a problem for the knowledge argument against physicalism (Alter, 2008; Dennett, 2007; Mandik, 2010). However, I do not wish to discuss the knowledge argument here, nor do I want to discuss whether or not deviants cause a problem for anti-physicalists. Rather, I want to argue that the notion of a deviant is not obviously coherent, and that deviants are plausibly impossible. If the notion of a deviant is coherent, then the possibility of deviants remains an empirical question. Furthermore, I argue that the burden is on the supporters of deviants to substantiate their claim that deviants are both conceivable and possible.

Under some definitions of "phenomenal knowledge," the possibility of a deviant may be rejected a priori. According to the ability hypothesis, for example, phenomenal knowledge is defined as the abilities to recognize, imagine, and remember an experience (Churchland, 1989; Lewis, 1988; Nemirow, 1990, 2006). Yet, deviants cannot have true memories of the experiences they are supposed to have knowledge of. A deviant may think they are accessing memories of past experiences, but they are wrong. They are merely imagining experiences and wrongly judging them to be memories. Thus, the deviant does not have all of the right abilities, and so does not have the requisite phenomenal knowledge. In short, the deviant is not a deviant at all. According to the ability hypothesis, then, the notion of deviant phenomenal knowledge is self-contradictory.

Of course, the ability hypothesis is not universally lauded. Those who reject the ability hypothesis might suppose that phenomenal knowledge does not require the possession of true memories. Some might suppose that no memories are required at all, while others might claim that a deviant could have implanted memories which, though false in the strict sense, might still satisfy the requirement of phenomenal knowledge. Though I do not find either of these alternatives compelling (I am a supporter of the ability hypothesis), I will--for the sake of argument--accept that true memories are not required for phenomenal knowledge, and on that basis consider whether or not deviants are possible.

One way of approaching this topic is to consider Swampkinds. Davidson (1987) introduces the idea:

Suppose lightning strikes a dead tree in a swamp; I am standing nearby. My body is reduced to its elements, while entirely by coincidence (and out of different molecules) the tree is turned into my physical replica. My replica, The Swampman.....moves into my house and seems to write articles on radical interpretation. No one can tell the difference.

Enter Swamp Mary, a Swamp-version of Mary the super-scientist. Mary, as you may recall, learns all about color vision while confined to a black-and-white existence. Then she steps out of her room and sees a red tomato for the first time. Now she knows what it is like to see red. She earned it. A moment later, lightning strikes and Swamp Mary is born. Swamp Mary has access to the same knowledge Mary had, and so she should know what it is like to see a red tomato. Yet, Swamp Mary hasn't earned that knowledge. Her history does not include an experience of phenomenal redness. Swamp Mary is thus said to be a deviant (Alter, 2008; Dennett, 2007; Mandik, 2010).

(In Dennett's version, Swamp Mary is born before Mary ever sees a red tomato: the lightning strikes Mary and reconfigures her brain so that she is in the exact state she would have been in had she experienced the sight of a red tomato. The result is the same: lightning produces a Swamp version of Mary who supposedly knows what it is like to see red without ever having seen red.)

Whether Swamp Mary is possible depends partly on what it would take to create her. Millikan (1996) observes that there are limitations on the ways a Swamp person can come into existence. Biological science tells us:

You can't in principle build large organic molecules such as hemoglobin, they say, just by throwing the right amino acid molecules together at the right angles with the right energies and having them stick. . . Molecules like that must be built up through long chains of enzymatic cascading reactions that proceed in just the right order and require support from just the right helping catalysts timed to come on stage at just the right moments.

Thus, one way for a Swamp person to come about is through a long and complex biological process. However, this is not how Swampkinds are usually imagined. They are supposed to be born out of a sudden jolt of creative energy. (Lightning is often involved, perhaps for dramatic flair.) Is there another, more spontaneous way for Swamp people to come into being?

Millikan says yes, there is a way. First, she reminds us that the laws of physics are time-symmetric. What can happen going forward in time can also happen in reverse. So, if a person can be disintegrated into tiny particles--blown to bits, as it were--then a Swamp person can be reintegrated in exactly the same manner. Yet, as Millikan notes, such a Swamp person would live backwards in time:

A Swampman so originated will himself be running backwards when he has been formed. His cells will be getting younger rather than older, coalescing rather than dividing, returning nutrients to his blood which will run backwards, returning the nutrients to his digestive system from whence they will eventually emerge as new good hamburgers getting fresher by the moment.

I'm skipping over some of the details of Millikan's analysis, but the point is clear enough. If Swamp Mary is going to happen in this world (or a world reasonably similar to this one), then it can only happen in conformity with the laws of physics, and this places constraints on how she can be created. It is most likely that Swamp Mary can only come about via a long and complex biological process, the full details of which are as yet unknown.

Supposing that Swamp Mary is possible, there is also the question of whether or not she really is a deviant. We cannot assume she is. For example, Millikan (1996) argues that we cannot attribute beliefs or other intentional states to Swampkinds, and this would preclude any knowledge attributions. She quotes a point made in her (1984):

Suppose that by some cosmic accident a collection of molecules formerly in random motion were to coalesce to form your exact physical double. ...that being would have no ideas, no beliefs, no intentions, no aspirations, no fears, and no hopes....This because the evolutionary history of the being would be wrong.

Davidson (1987) makes a similar point, though instead of appealing to the lack of evolutionary history, he appeals to the lack of individual history, writing: "I don't see how my replica can be said to mean anything by the sounds it makes, nor to have any thoughts."

The Davidson-Millikan objection is that, because Swampkinds lack the right histories (be they individual, evolutionary, or both), we cannot treat them like human beings. They cannot be deviants. I do not know how much weight is given to this argument in the literature, but it does appear to be a coherent objection. This makes the appeal to Swampkinds somewhat more difficult. [Update: see here for more on this line of thought].

In any case, all of the deviant eggs are not in the Swamp basket. Alter (2008) mentions other seemingly plausible ways of creating a deviant:

Unger imagines that scientists construct a man who is “cell‐part for cell‐part, cell for cell, nervous structure for nervous structure identical to” a man who both knows what it’s like to see red and has come upon this knowledge the ordinary way (Unger 1966: 50). Others (e.g., Lewis 1988, Alter 1998, Stoljar 2005) imagine cases in which surgeons operate on a person who has never seen red, creating brain structures similar to those found in the brains of people who have seen red.

In all of these cases, including Swamp Mary, the alleged deviant requires some long, elaborate, and neurologically complex process in order to produce their phenomenal knowledge. Yet, such processes cannot entail an experience of phenomenal redness--or any other experience which could earn phenomenal knowledge of redness, such as experiences which are simpler than experiences of phenomenal redness and out of which experiences of phenomenal redness might be composed.

(There may be disagreement about whether an experience of phenomenal redness is simple or complex. I will not argue one way or the other here. However, I think all sides agree that some experiences are composed of smaller or more primitive experiential components. Presumably, if one earns phenomenal knowledge of the more primitive components, one has earned phenomenal knowledge of the composites.)

Alter, Dennett, and Mandik are willing to assume that deviants could be constructed and function without ever experiencing even a hallucination of the color red. I find that remarkable, since there is no consensus about what sorts of physiological processes are necessary or sufficient for phenomenal consciousness. Any consensus can only be reached by appeal to the empirical sciences, and this makes the possibility of deviants an empirical question. I am inclined to think that deviants are impossible, though I admit that evidence could prove me wrong.

My thesis is thus: A person cannot be put into a functional state of knowing what it is like to see red without being put in experiential states which earn that knowledge. My intuition is that the elaborate process of constructing the right neurological configurations, or the very having of those configurations in a functionally operative state, necessarily entails perceptual experiences which are sufficient to earn all of their phenomenal knowledge. Of course, not being able to explain how such configurations can be constructed, I do not expect my intuition to be compelling to anybody. However, compelling or not, I do think this view is plausible. Furthermore, I think the burden of proof is on the people who disagree with me.

Here is my argument. We might define an "earning experience" as any local, biological process capable of producing the right synaptic configuration. In such a case, deviants are probably impossible, because deviants probably require some local, biological process to produce this configuration. Such a process would, by definition, make the deviant a non-deviant. So, if such a process is necessary, then deviants are necessarily impossible. On the other hand, if we claim that only some of the possible biological processes that create phenomenal knowledge count as "earning experiences," then we need some basis for making this distinction. Considering the obstacles in the way of correlating biological processes with primitive experiential states, it is possible that no empirical evidence can ever justify making such a distinction. Therefore, we cannot assume that such a distinction is coherent, let alone plausible. The burden is on the supporters of deviants to argue for a distinction here, and such a case must rely on empirical evidence--we need a reason to discount some biological processes, and not others. Thus, without further argument and evidence, we may suppose that deviants are plausibly impossible, and that the very notion of a deviant is possibly incoherent.

Having supported the main thesis under consideration, I will conclude by trying to motivate a narrower thesis: The mere having of the right functional brain states may entail the having of the right experiences. I begin with an appeal to Paul Churchland (1989), who presents some thoughts on what phenomenal knowledge entails. According to him, phenomenal knowledge is not discursive knowledge. It is manifest in abilities to access representational states, where representational states are "learned configurations of synapses" which apparently ground "the many abilities we expect from color-competent creatures: discrimination, recognition, imagination, and so on" (Churchland, 1989, 70). In sum, a person who knows what it is like to see colors is a person who has a neurological configuration which represents a full spectrum of primitive color-perceptual elements. In this light, the possibility of deviants depends on the answer to this question: Is it possible for such a configuration to function without instantiating experiences which earn the right phenomenal knowledge?

I submit that the answer is possibly, if not plausibly, no. Critics will point to the commonsense belief that we do not constantly have, for example, visual experiences of red, and that we do not thereby suffer spells of not knowing what it is like to see red. But is this so obvious? How do we know that we are not always activating our most basic color-perceptual configurations? Reportability is not the key here. The most basic elements of phenomenal experience are not necessarily reportable as such. We can distinguish between a reportable experience and experience simpliciter. The fact that we cannot always distinguish a "red" experience (or an experience which earns phenomenal knowledge of redness) does not mean that we are not always having one. So how do we know that we are not always experiencing some phenomenal qualities which earn our phenomenal knowledge of red? Even when we close our eyes, we see colors--even when we are seeing black, we are seeing colors.

While "black" is sometimes defined as "the absence of light," this is not an accurate description of the black we experience. It is more accurate to say that "darkness" is the absence of light, where even in darkness we see black. In terms of pigmentation, black is a color, and it can be created by combining more basic elements in the color spectrum. Seeing black is not the same as not seeing anything at all. There is an experiential difference between seeing a black object and having a blindspot. So, arguably, when we see black, we see colors--and, arguably, we see all of the most primitive elements in our color-perceptual toolkit. It is therefore plausible that any person with a functional representation of the elements of color vision will experience all of those elements, and will thereby earn their phenomenal knowledge.


  • Torin Alter (2008). "Phenomenal Knowledge Without Experience." In Edmond Wright (ed.), The case for qualia.
  • Daniel C. Dennett (2006). "What Robomary Knows." In Torin Alter & Sven Walter (eds.), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism. Oxford University Press.
  • David Lewis (1988). "What Experience Teaches." In Proceedings of the Russellian Society. Sydney: University of Sydney. Rpt. in Mind and Cognition: a Reader, W. Lycan, (ed.). Cambridge: Basil Blackwell, 1990, pp. 499‐518.
  • Millikan, R. (1984). Language, Thought and Other Biological Categories. Cambridge MA: MIT Press.
  • Laurence Nemirow (1990). "Physicalism and the Cognitive Role of Acquaintance." In William G. Lycan (ed.), Mind and Cognition. Blackwell.

  • Laurence Nemirow (2006). "So This is What It's Like: A Defense of the Ability Hypothesis." In Torin Alter & Sven Walter (eds.), Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism. Oxford University Press.
  • Peter K. Unger (1966). "On Experience and the Development of the Understanding." American Philosophical Quarterly 3 (January):48-56.

Saturday, May 1, 2010

What Zombie Mary Knows

[This post was revised and reposted on May 2, 2010.]

As I mentioned in my last post, Torin Alter alerted me to the fact that Victoria McGeer (2003) makes an argument about Zombie Mary which is similar to my own. I emailed McGeer and she was kind enough to send me a copy of her paper. It is not as similar to my argument as I was expecting, but we do cover some of the same ground. She makes strong use of Zombie Mary, though she does not construct an incompatability argument. Her aim is rather to show that (1) the knowledge argument relies on the very modal intuition it was supposed to establish, and (2) the knowledge argument is implausible. I will discuss her main arguments below, and I will also discuss some related issues.

First, McGeer makes a very important point about David Chalmers' position in this whole affair. Chalmers (1996) observes that the modal intuition--the intuition that zombies are conceivable--entails that zombies make judgments just like human beings do. This follows from the fact that judgments are behaviors--they are things we say and do. Yet, because zombies do not have phenomenal consciousness, their judgments about phenomenal consciousness are not true. Zombie Mary can say, "now I know what it is like to see red!" However, when she says it, she is wrong. This leads to a sort of paradox: What makes a phenomenal judgment true or false--what justifies it--cannot be a part of its causal history. As McGeer writes, "it seems that consciousness in the phenomenal sense is explanatorily and causally irrelevant to the making of phenomenal judgments, though not to their justification or truth-value" (McGeer, 2003, 386.) According to Chalmers, this paradox is "delightful and disturbing" (Chalmers, 1996, 181). Disturbing, indeed, for we must marvel at the fact that Zombie Mary makes judgments about phenomenal properties at all!

What I show in my defense of the incompatibility argument is that the paradox is wider than Chalmers may have realized. It is not simply a paradox of phenomenal judgment. It is a paradox of judgment, period. The knowledge and conceivability arguments cannot both work unless we completely detach knowledge from observable behavior. To save the knowledge argument, we must stipulate that knowledge cannot be observed, even when we observe clear cases of learning. Zombie Mary can learn how to identify colors, but learning is not sufficient for knowledge. Phenomenal experience is required for knowledge of any sort, but this experience has no causal or explanatory relation to the behaviors we regard as expressions of knowledge! It turns out Zombie Mary is not a super-scientist, though she looks and acts exactly like one in every observable detail.

The paradox of judgment does not just stem from the combination of the knowledge and conceivability arguments. It follows directly from the knowledge argument itself. McGeer notes that, if the knowledge argument is sound, then physicalism is false, and thus the modal intuition is correct. According to the incompatibility argument, if the modal intuition is correct, then the knowledge argument is not sound. Thus, the knowledge argument is self-defeating--unless we reject the incompatibility argument in favor of the paradox of judgment.

The paradox of judgment may also follow directly from the conceivability argument itself, if we accept Bertrand Russell's distinction between knowledge-by-acquaintance and knowledge-by-description. Russell (1912) maintained that all knowledge is composed of facts learned by direct acquaintance; even facts we learn by description are only known via facts we have learned by acquaintance. If this is true, then zombies cannot know anything at all, because they cannot be acquainted with anything. Thus, the modal intuition produces the paradox of judgment regardless of whether or not we combine it with the knowledge argument.

Chalmers might be willing to accept the paradox of judgment. He already accepts the paradox of phenomenal judgment, so why not also accept that all expressions of knowledge are unrelated to what makes them true or false? Indeed, it would be simpler to claim that whatever makes any judgment true or false is causally and explanatorily unrelated to the judgment itself than it would be to claim that phenomenal judgments are somehow distinct in this matter. Chalmers can say that this is just the way knowledge works, without having to make a special case for phenomenal knowledge.

The main reason to reject the paradox of judgment is that it makes it impossible for an anti-physicalist to explain zombie behavior. An explanation of zombie behavior in terms of evolutionary biology, for example, would be out of the question. Perhaps Chalmers is willing to purchase anti-physicalism at this price, but I cannot imagine how he could justify that move.

Furthermore, the paradox of judgment causes a problem for the knowledge argument, which stipulates that Mary can learn a completed science. If we stipulate that zombie behavior cannot be explained even in physical terms, then we are making a strong critique of the physical sciences. There would have to be physical facts which could not be deduced from the completed science Mary learns from inside her room, and the knowledge argument would no longer work against physicalism.

McGeer does not follow this line of thought exactly, but her approach is quite similar. She suggests a way for anti-physicalists to explain Zombie Mary's judgments: Zombie Mary experiences phenomenal illusions. Zombie Mary only thinks she is experiencing colors, as a result of physical transformations she undergoes when she sees a red tomato for the first time. However, as McGeer notes, this creates another problem for the knowledge argument. For, if Zombie Mary's physical transformation induces a phenomenal illusion in her, why wouldn't it induce such an illusion in the original Mary? Indeed, how do we know we are not experiencing phenomenal illusions when we think we are experiencing colors? The possibility of phenomenal illusions seems to undermine the knowledge argument, too.

I suspect supporters of the knowledge and conceivability arguments will reject the idea that Zombie Mary experiences a phenomenal illusion. How could a zombie experience an illusion, when they cannot experience anything at all?

Either way, the knowledge argument is in trouble. If anti-physicalists embrace the paradox of judgment and reject the claim that Zombie Mary experiences phenomenal illusions, they undermine the knowledge argument by claiming that not all physical facts can be accounted for with the physical sciences.

McGeer makes another argument which deserves mention. She notes that Zombie Mary seems to learn what it is like to see red, and even says, "Ah, so this is what it's like to see red!" Yet, the point of Zombie Mary is that she cannot know what Mary knows. So the fact that she seems to know what Mary knows is not enough to conclude that she knows what Mary knows. In the case of Zombie Mary, anti-physicalists ignore her behavior, because Zombie Mary does not really know what she claims she knows. Yet, in the case of the original Mary, anti-physicalists do not ignore her behavior, because Mary really knows what she says she knows. But then, anti-physicalists are licensing themselves to ignore evidence in one possible world, but not another. To justify this move, the anti-physicalist must first establish that the same evidence can be treated differently in different worlds, and that requires an argument for the modal intuition that zombies are conceivable. So anti-physicalists must first establish that zombies are conceivable before they can successfully argue for the knowledge argument. Thus, the knowledge argument relies on the very modal intuition it was supposed to establish.

Even if all of my previous arguments are flawed--even if the knowledge and conceivability arguments are compatible, and even if the knowledge argument is plausible--McGeer shows that the knowledge argument cannot be used to support the conceivability argument. The conceivability argument is thus significantly weaker, for the knowledge argument was supposed to establish the modal intuition which it requires. The knowledge argument is also significantly weaker, because it relies on the same modal intuition.

Before concluding, I will make one more observation about the poverty of the knowledge argument. Consider just what Mary is supposed to learn which is said to prove that there are non-physical facts. Frank Jackson (1986) puts it thus:

The trouble for physicalism is that, after Mary sees her first ripe tomato, . . . she will realize that there was, all the time she was carrying out her laborious investigations into the neurophysiologies of others and into the functional roles of their internal states, something about these people that she was quite unaware of. All along their experiences . . . had a feature conspicuous to them but until now hidden from her.

Anti-physicalists claim that Zombie Mary could not learn this fact. But how could the original Mary learn it? Her own experience of color vision cannot tell her anything about what other people experience. Anti-physicalists claim that phenomenal experience is absolutely subjective--nobody has access to anybody else's. So nobody's color-perceptual experiences can tell them anything about what other people experience, if other people experience anything at all. Indeed, how does Mary know that the rest of the world isn't populated by zombies? Mary cannot learn what Jackson says she can learn, so the knowledge argument's threat against physicalism is idle.

I have here presented several arguments which together suggest that (1) the knowledge argument is self-defeating and poses no threat to physicalism, (2) both the knowledge and conceivability arguments rely on the same modal intuition, and (3) the modal intuition entails a paradoxical view of knowledge and judgment. If anti-physicalists want to make a case against physicalism, they need a new argument for the modal intuition--preferably one which either justifies or overcomes the paradox of judgment.


Chalmers, David (1996) The Conscious Mind: In Search of a Fundamental Theory. New York: Oxford University Press.

Jackson, Frank (1986) “What Mary Didn’t Know,” Journal of Philosophy, 83, pp. 291–5.

McGeer, Victoria (2003) "The Trouble With Mary" Pacific Philosophical Quarterly, 84, 4, pp. 384-393.

Russell, Bertrand. (1912). The Problems of Philosophy. Oxford: Oxford University Press.