Of Minds and Language (56 page)

Consider a very similar case: bystander, loop, man, but now the man on the looped track is irrelevant because he's too thin to stop the trolley. However, in front of this man on the loop is a weight which is heavy enough to stop the trolley. This case can be interpreted as killing as a foreseen side-effect. Aiming for the man on the looped track makes no sense as he can't stop the trolley. Aiming for the weight makes sense as it can stop the trolley. Here, 76 percent of subjects say that it is morally permissible for the bystander to flip the switch, which is significantly greater than in the previous loop case. Importantly, these two cases involve impersonal harm, the trolley is redirected, there is only one man on the looped tracks, and the greater good is five saved in both. One of the potentially significant differences is between intended vs. foreseen harmful consequence. That is, using the man as a means to save five as opposed to foreseeing the man's death to save five.

These cases are just the beginning of the story, a flavor of how we have begun to move by thinking about principles. But let me flag something crucial about the notion of a principle. My use of this term is completely different, and ultimately wrong, relative to the level of abstraction of principles that people in linguistics have moved toward. In the case of morality, this is merely a starting point. The intuition is that as we move deeper into this problem, the abstractness of the problem will surface, and the relationship between actions, intentions, and consequences will be as complex and nuanced as are the relationships between the conceptual-intentional system and the syntactic operations that provide structure and, downstream, variegated meaning. So, when I say “principle,” think of it in this looser sense, at least for now.

Let me describe three principles, with the first mapping to a distinction I just called upon: the
Intention Principle
. It is basically the principle that Thomas Aquinas invoked as the “doctrine of double effect”: harm intended as a means to a goal is morally worse than equivalent harm foreseen as the side effect of the goal. Second, the
Action Principle
: harm caused by action is perceived as morally worse than equivalent harm caused by omission; and lastly, the
Contact Principle
: harm caused by physical contact is morally worse than equivalent harm caused by non-contact. To explore these principles, we developed a large set of moral dilemmas (we now have some 300–400 different moral dilemmas). For each principle, we presented a set of paired
dilemmas that only differed in terms of the crucial psychological dimension captured by the principle. Subjects provided judgments on a scale from 1 to 7, with 1 mapping to forbidden, 4 to permissible, and 7 to obligatory. For paired cases in which subjects noticed a difference, we both evaluated this difference statistically and also asked subjects to justify their responses.

Results showed that for both the Intention and Action principles, six out of six scenario pairings revealed support for the operative force of the principle, whereas for five out of six scenarios in the Contact principle, this was also the case. Thus, subjects judged intended harms as morally worse than foreseen harms, actions as worse than omissions, and contact harm as worse than non-contact harm. The next critical question, very much analogous to questions in linguistics, was: are these principles not only operative in that they influence people's judgments, but can they be expressed? Are they recoverable, are they used consciously in deliberations in creating these moral judgments?

For the Action principle, subjects recovered a sufficient justification 80 percent of the time, appealing to comments such as “actions are worse.” For the Contact principle, subjects appealed to “contactful” harm as worse than non-contact about 50 to 60 percent of the time. Quite consistently, however, they denied the moral relevance of contact, saying such things as “Well, if you physically touch somebody and it hurts them, that is worse than if you don't touch them … Nah, that can't be relevant.” So they rejected the principle, and often invented assumptions to explain what was driving their judgment. But perhaps most interesting of all, very few people recovered the Intention principle. People who failed would say things such as “I don't know” or our favorite, “Shit happens!” So here we have a distinction between principles that in each case are clearly operative in that they are driving the nature of the judgment, but only in some cases are they recoverable in that they seem to be expressed in people's distinctions between Case 1 and Case 2. That suggests that some principles seem to be having effects as intuitions and that maybe these intuitions are absolutely not recoverable or are inaccessible, in the same way that linguistic principles that have been discussed here in this conference are inaccessible.

One of the big questions, then, coming back to some of the themes in the conference, is the extent to which we see these kinds of principles as universally in play. To begin addressing this question, we can pinpoint different variables that have classically been invoked as causally relevant to cross-cultural variation and explore the extent to which they influence the patterns of judgment. One of our first stabs has been in terms of religious background. As a first cut, we simply contrasted all subjects indicating some kind of religious background with those marking “atheists.” For this initial analysis, we didn't concern ourselves with the specific kind of religious background,
but rather with its presence or absence. The clear result thus far is “no.” There was not a shred of evidence that people who claim to be religious showed different patterns of moral judgment or moral justification (except that we did of course see people who were religious invoking more, “Well, God must have done something”). Furthermore, we found no differences between people who expressed different degrees of faith or religiosity: individuals who said that they were not very religious showed the same patterns of judgment as those who said that they were very religious; and within the limited sample that we collected, there were no consistent differences among the types of religions (Hauser 2006).

Let me digress for a moment to relate this finding to a recent experience I had in a class at Harvard, and in particular, during the presentation of this material. During my presentation, I could see that some students were getting extremely anxious. I therefore stopped the lecture and said, “You all seem a bit antsy. If you have concerns or questions, please pipe up and let me know.” Upon finishing the last syllable of my sentence, one student exploded and said, “Look, I know where you are going with this. This is one of those biological, Darwinian explanations, but there is a clear alternative explanation: simply, God created all the universals.” These are tough moments for a teacher. On the one hand, you want to respect the variety of views that people can have, and on the other hand, you want to explicate the positions, and show that issues of faith and science are entirely different ways of knowing or understanding. I responded, “We may be at an impasse here. I can either capitulate because I can't call up any evidence to show that your position is wrong, or we can take the following path together. If it is true that most, if not all religions take as inspiration some divine power, and divine power provides the intuitions that create religious doctrine, then I think you have a problem. Since religious doctrine can't explain the pattern of judgments we observe, but you want to argue that God or some divine power provides the universals, then you have to say that religion rejects divine inspiration when it comes to these moral judgments. This just strikes me as very problematic for the religious position, at least if you think that there is an empirical issue, as opposed to an issue that strictly relies on faith.”

The other point I would make is that of course everyone taking the moral sense test logs on to the Internet, and thus our sample is very skewed. In fact you could say, “Even if you are not religious, you've been exposed to Christianity at some level, so of course that is why you are finding the pattern that you have.” To address this problem, we have begun to present the same kinds of moral dilemmas to small-scale hunter-gatherer societies that have no explicit religious system – this doesn't mean that they lack beliefs, but rather, that their system of beliefs is not made explicit in the form of religious doctrine or accounts. And they certainly haven't been exposed to Christianity. So if we find similarities
I think it argues even more strongly for certain patterns of universality driven by some biologically set-up system. An interesting example comes from the Kuna Indians, a very small-scale hunter-gatherer/subsistence society in Panama that has had little contact with the outside world. We have given one village community various kinds of moral dilemmas, cases that in important ways mimic the trolley problems. Now here is an intriguing, albeit preliminary result. If you give them an example that is like the bystander case, a canoe going down a river which can displace crocodiles away from five onto one, most people said that it is permissible for the bystander to redirect the crocodiles. If you give them a version of the fat-man case, in which a person can push a fat man out of a tree in front of a herd of stampeding boar, saving the five, but killing the fat man, only about half say that pushing is permissible.

Now, here is the important lesson, I think. This culture, as well as others that we have been able to look at such as a Mayan community in the Chiapas area of Mexico, see the difference between intended and foreseen harms. In the case of the Kuna, however, the difference between means-based and foreseen harms seems to be less than it is in the Western and developed societies that we have tested on the internet. This is very preliminary and could be driven by all sorts of confounds, but for the moment, let us assume the pattern is real. We can explain the increased permissibility of intended harms in the Kuna by looking at their recent history of infanticide. That is, one sees almost no physical deformities in this society, primarily because those with such deformities are killed early in life. So intended killing is part of the society. What we think is happening, perhaps as a form of parametric variation, is that all societies will show the Intention principle, but each society can tune the degree of difference between means-based and foreseen harms – but not eliminate it.

I hope this provides a flavor of the argument and the work that lies ahead. I think the principles here are nowhere near where they need to be. I think in some sense we need to go back to some of the work that was started a long time ago in the philosophy of action, laying out in greater detail the nature of computation in action perception that may provide some of the primitives to our moral judgments. Even with all the empirical holes, however, I think we now have a new and important set of questions, with answers forthcoming.

In this final empirical section, I will focus on music. Again, there is limitless variation in music as there appears to be for language and morality; the question is whether there are some primitives that are both driving and constraining the
variation. What I don't want to spend too much time debating is a definition of music, as this could lead us down a never-ending path that would be quite fruitless. Here, however, is a quote that I like because it captures at least some of the functionality of music: “The purpose of music is to sober and quiet the mind, thus making us susceptible to divine influences.” What I actually like about this quote is that it begins to capture two important aspects of music that I want to explore here, which is the interface between some kind of perceptual pattern recognition and our emotions. John Cage, of course, who really started the minimalist move in music, made exactly this kind of argument. Here is one of my favorite quotes by Cage: “I can't understand people who say I am frightened of new ideas, I am frightened of the old ones.” This certainly seems to capture a flavor of the minimalist program in linguistics. There is a little piece from John Cage that relies on only three notes, over and over again, but with crucial changes in tempo and the intensity or attack. This tradition continued, for some people to their great horror, including Cage's famous “ 4'33”” – a piece of simple silence, lasting for four minutes and 33 seconds, precisely the length of time of typical “canned music.” Taking liberties with the views espoused by the minimalist movement, I think that minimalist music focuses on breaking up the intentionality created by music, emphasizing the silences, randomness, slowness, and tempo in particular. As such, it attempts to strip music to its core, its skeletal features, and assess how such structures mediate perception. Even if this rather loose interpretation is too loose, I think it is a wonderfully ambitious and exciting project that sits at the interface of the arts and sciences.

Now here is what I want to do to show again how comparative work can bear on questions of music structure and perception. It is true that every single culture that we know about has music as part of its system, and the question is, are there invariants? Two invariants that appear to emerge, cross-culturally, are that consonant intervals are perceived as more pleasant than dissonant intervals, and that lullabies have virtually identical structures, simple, repetitive elements, slow tempos, and a restricted range of frequencies. Assuming these are invariants, part of our species' biological endowment, we can next ask: how did these perceptual–emotional biases evolve and are they uniquely human?

To address this question, we turn to studies of non-human primates. In particular, my students and I wanted to understand not only what primates perceive, but whether they spontaneously discriminate certain musical styles, and especially, like some more than others. Thus, our goal was to explore how potentially ancient perceptual mechanisms interface with the emotions to generate distinctive musical preferences.

To explore this problem, my recent graduate student Josh McDermott worked with me to design a very simple experimental approach using a V-shaped maze (McDermott and Hauser 2005). We released an animal, either a cotton-top tamarin or a common marmoset, into this maze, and while they were on one branch of the V-maze, a hidden speaker played a particular sound; as soon as they crossed over to the other branch, a different sound played. What this method provides is a kind of listening station where the animal gets to choose its musical selections, at least within the options of a session. They don't receive any physical rewards for choosing, simply the exposure to different sounds. Before we explored some of the more interesting musical contrasts, we first wanted to establish that the method would work and thus contrasted loud white noise with soft white noise. We found consistent preferences in both tamarins and marmosets: a strong preference to spend time on the soft white-noise side, as opposed to the loud. Similarly, if you present tamarins with a choice between their own, species-specific food chirps (associated with food) and their submissive screams (associated with fear), they spend more time on the chirp side than the scream side. These results reveal that the method works, providing a tool to explore spontaneous preferences for particular sounds.