In The Righteous Mind, Haidt invokes Michael Tomasello’s notion of “joint intentionality,” calling it our evolutionary Rubicon; i.e., the critical trait the evolution of which made us irrevocably human and led inevitably to the development of a large number of our most distinctive human characteristics, especially our groupishness. Haidt writes:

When everyone in a group began to share a common understanding of how things were supposed to be done, and then felt a flash of negativity when any individual violated these expectations, the first moral matrix was born… That, I believe, was our Rubicon crossing. (239)

I think Haidt gets a little too carried away over joint intentionality. The purpose of this post is to explain why and to suggest a more sensible alternative proposed by anthropologist Joseph Henrich.

Tomasello is a noteworthy primatologist/psychologist/anthropologist. Joint intentionality is his Big Idea. I believe it lies at the core, explicitly or implicitly, of nearly all of his work, a large number of papers and books stretching back over 20 years. Certainly, it is the heart of his new (2016) book, A Natural History of Human Morality. Joint intentionality—and thus truly joint action, as opposed to merely acting in a group—requires that the agents share a representation of the action and its purpose, including each agent’s functional role. The example of carrying either end of a log illustrates this requirement beautifully, I think.

More specifically, the requirements for joint intentionality can be analyzed into four elements: (1) There must be a shared representation of the task as a joint task, implying therefore a joint agent, a “we.” (2) The task representation must specify a set of functional roles, to be occupied by the individual participants, such as the role of carrying one end of the log and the role of carrying the other end, or the role of driving the prey forward and the role of catching it in a waiting net. It is important that the roles are functionally specified and therefore (a) agent-independent and (b) normative. That is, any individual could potentially occupy any role, and the roles are conceived in terms of an ideal performance, which actual performance will meet to a greater or lesser degree. (3) There needs to be some form of commitment to the joint task, to the “we.” This is because engagement in joint action requires mutual trust. A joint action depends on the functional roles being filled, so there is no point in initiating a joint action if there cannot be an expectation that the roles will be and stay filled. (4) All this should be mutually understood. That is, the parties not only must each possess the global representation of the joint task and its requirements, but they must each know that the other parties possess it, know that the other parties know that they possess it, know that the other parties know they know the other parties possess it, ad infinitum. (And I kid you not, he cites Lewis’s Convention on this point. It is surprising to me how well-acquainted with the philosophical literature some non-philosophers are.)

These elements wouldn’t have popped into the human repertoire overnight, of course. They must have evolved over time. But—just as Haidt says (243)—by the time of Homo heidelbergensis, about 400,000 years ago, they will all have been well along in development, if not fully in place. (Tomasello speculates [44] that what drove this development were ecological changes beginning about 2 million years ago that made scavenging and then hunting large game a requirement for human survival.)

Tomasello thinks that joint intentionality is the fundamental human ability that separates us from the other great apes. That is, none of the other great apes—nor any other species, for that matter—exhibits joint intentionality, and joint intentionality is ultimately responsible for the bulk of important cognitive and behavioral abilities that distinguish us from the other great apes. With respect to morality, Tomasello argues that the logic of joint intentionality already implies the basic principles of the morality of fairness, which early humans (i.e., beginning with Homo heidelbergensis) accordingly began to work out. In particular, first, the ideality of functional roles implies a set of socially normative standards; i.e., of virtuous performance of those roles. Second, the impersonality of functional roles implies self–other equivalence. Since anyone might occupy any given role, there is a basis for putting oneself imaginatively in another’s place and considering the other as equivalent to oneself, apart from the role each happens to occupy. This implies that one must judge oneself impartially, not favoring oneself over others or privileging one’s own interests. Third, self–other equivalence implies mutual respect between partners and mutual deservingness; for instance, in sharing the proceeds of joint activities. Fourth, mutual respect and deservingness implies the standing to enter into (or not) joint commitments and to mutually regulate them (for example, to rebuke those who fail to honor the commitment or to perform to standard).

Tomasello claims that these points follow from the logic of collaborative activity and joint intentionality. They are a matter of “cooperative rationality” (79) (rationality for carrying either end of a log) as distinct from the personal-utility-maximizing rationality of nonjoint action (rationality for a herd of fast deer). These points that constitute cooperative rationality are not reducible to personal-utility-maximizing rationality. Further, recognition of them is recognition of reality, the reality of the logic of collaborative activity (81).

It is obvious that these points comprise all the essential elements of a morality of fairness (which Haidt would recognize as falling within his Fairness/Cheating moral foundation). That is what Tomasello intends: “The claim is not just that this [the moral psychology of cooperative rationality] was a modest first step on the way to a fully modern human morality but, rather, that this was the decisive moral step that bequeathed to modern human morality all of its most essential and distinctive elements” (78). It is also obvious that the conception of morality in play here is thoroughly Kantian. Notice that morality is being said to be fundamentally a matter of rationality. It is simply the rational recognition of the logic of collaborative action. This is no accident. Tomasello’s philosophical authorities on morality are Thomas Nagel, Christine Korsgaard, and Stephen Darwall. He cites them and speaks of their views as though they are simply true.

Cooperative rationality is not the end of Tomasello’s story. Cooperative rationality only gets us “second-personal morality,” which is limited to specific joint intentional agreements (to hunt bushpigs tomorrow, to carry a log today) between particular individuals. The scope of this second-personal morality is thus local and temporary. It does not draw on community-wide social norms or imply a larger moral community in any other way. To advance to today’s “‘objective’ morality” requires a second major step, a transition from joint intentionality to “collective intentionality” and from cooperative rationality to “cultural rationality.” This shift began with the emergence of Homo sapiens some 150,000 years ago and was well under way by 100,000 years ago. It was driven by population growth, as human groups grew from relatively isolated bands (such as chimps now live in, for example) to bands grouped into tribes, which is how hunter-gatherers are mostly organized today. Bands of 20–30 individuals belong to tribes of several hundred individuals who share a common language, culture, and religion. Bands do not compete within the tribe, but the tribe competes with other tribes for land, resources, and women.

Tribal-level competition created the problem of maintaining group cohesion and pro-group behavior among people who don’t all know each other very well, aren’t kin, and don’t share a context of knowledge and interests, as members of a single band would. The solution to this problem was to create a set of cultural norms, practices, and institutions that signal group identity and mandate pro-group behavior. Now instead of norms dictated by functional roles in a collaborative activity, there are norms dictated by cultural conventions. And rather than accepting norms as part of one’s agreement to participate in a collaborative activity, one accepts them because one identifies with one’s culture and so considers oneself to be a coauthor of those norms. (Thus, even 100,000 years ago, our ancestors were giving themselves the moral law.)

I won’t go into further details about step two “objective” morality (the scare quotes are Tomasello’s, by the way). It is clearly just a scaled up version of step one second-personal morality. It is scaled up in the sense that (a) the norms and commitments are now community-wide—so that they are universal, not local and temporary, and in that sense they are “objective”—and (b) they now encompass everybody in the community, not just the participants in a collaborative activity. Nevertheless, the same Kantian themes prevail: cultural practices, norms, and institutions specify agent-independent roles that anyone in principle could occupy; this implies a view-from-nowhere perspective in which roles might be occupied by any “rational person” (i.e., fellow tribe member); from this perspective, then, all tribe members are equivalent and thus equally deserving of respect, dignity, and concern; and the recognition of all this is simply recognition of reality, the reality of the logic of collective action; “objective” morality is thus a straightforward consequence of cultural rationality. And for Korsgaard fans, there’s also a great deal of discussion of “moral identity” and how it motivates moral behavior. But I said I wouldn’t go into details.

In case it is not clear, Tomasello’s theory, like Haidt’s, is supposed to be merely descriptive, not normative. Tomasello writes as a scientist. He is trying to understand how we have come as a species to make the moral judgments we do and also how we come as individuals to make them. He is not advocating or defending the correctness of any particular moral judgments over any others (although, disturbingly to a philosopher, he does simply take for granted the truth of one particular brand of moral philosophy). Also, although he presents his theory as a reconstruction of how moral judgments developed over the course of human evolution, it is not just his flight of fancy. He supports his claims with a wealth of data, especially from experimental studies conducted with chimpanzees and with 2- and 3-year-old humans, much of it conducted by himself and his research group. He is an excellent experimentalist, and his review of this data is easily the most interesting and useful part of the book (for my non-Kantian money, anyway).

Regarding Haidt, I can’t resist noting the irony in his enthusiasm for Tomasello’s joint intentionality idea, inasmuch as Tomasello’s theory of moral psychology is so very different from his. Haidt is a social intuitionist who thinks that moral judgments are mostly post hoc rationalizations of emotional and intuitive reactions; Tomasello is a rationalist who thinks that moral judgments recognize and express moral reality. Haidt thinks that moral reality, if it exists to be found at all, is elusive; Tomasello thinks that to discover moral reality is well within the mental capacity of H. heidelbergensis and Neanderthals. Haidt thinks that such moral reality as we will be able to find has an Aristotelian shape (leavened with plenty of Durkheim, of course), centered on human well-being; Tomasello thinks moral reality is Kantian, centered on the logic of collective rationality. Finally, Haidt thinks morality is richly multidimensional, deriving from at least six intuitive “moral foundations”; Tomasello thinks there are just two dimensions of morality, the morality of sympathy (i.e., compassion) and the morality of fairness. And where have we heard of this last before? Perhaps in the title of Part II of Haidt’s book: “There’s More to Morality than Harm and Fairness”.

Now, let me say why I do not think joint intentionality is the Rubicon—the essential threshold of our humanness—that Haidt suggests it is. For one thing, Tomasello has some trouble to establish that joint intentionality is really unique to humans at all. Chimps hunt in packs, for example. Tomasello argues that this behavior is not really collaborative. For instance, when chimps hunting a monkey surround it, this is not because they recognize surrounding as a joint activity in which each ape performs the role of plugging gaps in the circle; it is simply that each ape is individually trying to maximize his chances of getting the monkey first. Yes, but there are other examples. Chimp bands are highly territorial. Groups of males regularly patrol their borders, looking for strays from other bands. I believe they also sometimes deliberately raid other bands’ territories. Jane Goodall even reports observing a four-year “war” in which one band ultimately wiped out all the adult males in the other and took over its territory. This suggests to me that they may know what they’re doing in a more articulated way than Tomasello allows.

But let’s assume he’s right that joint intentionality is uniquely human. That does not make it our Rubicon. There are lots of uniquely human traits, such as control of fire, use of projectile weapons, personal adornment, burial of dead, use of tools for making other tools, and symbolic behavior, but being unique doesn’t make them Rubicons. To be the Rubicon, a trait should be the factor that is most fundamental in explaining the bulk of characteristics that uniquely distinguish our species. It is obvious how our acquisition of reason, or of language, could be claimed to be Rubicons in this sense. But it is not so obvious to me how the acquisition of joint intentionality can.

Certainly, Tomasello’s arguments that human morality derives from the logic of joint intentionality fall flat. For instance, perhaps the most important of these arguments is the claim that joint intentionality implies self–other equivalence through the shared representation of ideal roles. But this can hardly be true. What if the roles aren’t equal? Tom Brady has one role on the Patriots football team, the guy who squirts Gatorade into players’ mouths on the sideline has another. These are not interchangeable roles! Not the Gatorade squirter nor anybody else (except maybe the Gatorade squirter’s mother) thinks he could fill Tom Brady’s role. So there is no implied self–other equivalence here. Indeed, the opposite is implied. (Prehistoric example: fashioning the stone tools, quite a high-skill task, versus carrying the blanks from the quarry.) Roles may also be intrinsically unequal. Perhaps only I can perform the sacrifice at the religious rite, because only I am descended from the gods. Roles may also be socially unequal. Perhaps only I can have the prestige role of driving the prey forward and you have to wait with the net because I am son of the Chief. This brings out another point: not only do ideal roles per se not imply agent equivalence, but they are fully compatible with hierarchy and coercion. Just because A and B carry either end of a log doesn’t mean that A is not B’s master and B is not A’s slave. It might only mean that since A happens to have only the one slave, he must carry the other end of the log himself. Again, here is a collaborative activity: you polish my boots, and I wear them. It might seem that either of us might occupy these ideal roles, but not at all. For, boot polishing is a demeaning, menial task, and boot wearing is for elites. And, let us suppose, you are intrinsically of lower rank and I am higher. Nothing about joint intentionality per se contradicts this. Joint intentionality does not sponsor the “view from nowhere” any more than maintaining a mental map of the local terrain or a set of the general characteristics of bushpigs or any other abstract mental ability, so far as I can see.

Joint intentionality also does not imply norms any more than any other activity. Consider the art of cracking nuts open with a stone. It is not as easy or obvious as it might seem. It takes chimps months or even years to learn to do it by watching other chimps. It is clearly an activity with an ideal functional role, even though it is solo. It can be done better or worse; there is an ideal of performance which the individual seeks to achieve. So if the jointly represented role of carrying one end of a log implies a norm, so does the singly represented role of opening a nut with a stone.

An alternative idea is that norm psychology—the propensity to treat a certain way of doing something as the way it’s supposed to be done—developed as a genetically evolved adaptation that promotes cultural learning. This is what Henrich argues in his recent (2016) book, The Secret of Our Success. In Henrich’s view, the capacity for accumulated culture is what separates us from other animals. Think of it as an anthropologist’s take on one of the points Leonard Read was making in I, Pencil. None of us (or very few of us) knows the first thing about even so seemingly simple a thing as how to make a pencil. As individuals, we rely on and benefit from a sophisticated technological and social culture that has built up over years and centuries and even millennia, most of the details of which we do not understand.

For example, consider arrow-making in a group of hunter–gatherers. Some arrows will definitely be better than others—lighter, straighter, more stable in flight, sturdier, with sharper tips, a better fit to the bow, and so forth. Over many years, knowledge of the best tools, materials, and techniques for making arrows will gradually accumulate. Sometimes what makes certain tools, materials, or techniques superior may be fairly obvious or may be known for some reason, but in many cases it may not. Importantly, it need not be known as long as people learn to make their arrows by simply copying the procedures of the best arrow makers. And this is what people tend to do. Henrich provides abundant evidence that people naturally tend to model their behavior—not just in arrow making, but in many, many ways—on the behavior of people who exhibit signs of superior performance in the relevant domain, or else who exhibit signs of superior status, or else people model their behavior on that of the majority. Moreover, people tend to treat the modeled behavior as normative, not as, “this is how so-and-so did it, and it worked out pretty well,” but as “this is the way you do it.”

Henrich shows—convincingly, to me anyway—that this process can lead over time to an accumulation of cultural know-how that does not depend on individuals knowing all the whys and wherefores that underlie their techniques. This is fortunate, because often the whys and wherefores may involve scientific principles that are beyond the reach of the people in question. For example, certain Australian Aborigines eat the “seeds” (strictly speaking, sporocarps) of a certain fern called nardoo, which is inedible unless processed properly. The seeds must be ground into flour and the flour leached with water. The leaching helps wash out a thiamine (vitamin B1) destroying chemical in nardoo, ingesting which causes extreme fatigue, muscle wasting, and other ailments. The Aborigines then bake the resulting gruel in ash, which lowers its pH and also apparently helps reduce its thiamine destroying property. Of course, the Aborigines know nothing of thiamine or pH, but this has not stopped them from accumulating cultural know-how enabling them to exploit an important local food source that would otherwise be unavailable to them. It seems likely that this knowledge accumulated gradually, by lucky accident and copying the practices of the most successful, over time in the remote past. The Aborigines may or may not know the specific adverse consequences of failing to follow their customary procedures, but it seems unlikely that they need periodic reminders of the importance of following them by lapsing and thus inadvertently poisoning themselves. More likely they follow them without requiring elaborate reasons, as a matter of custom, because “this is the way you do it.”

An impressive line of evidence for the claim that cultural knowledge is mostly not the sort of thing that a single individual could develop but rather must accumulate painstakingly over many years is that cultural developments are often lost. We tend to think that once some cultural milestone has been reached—the invention of the wheel, for instance—we now “have” it, permanently, and are ready for the next development. But in fact, culture has often moved backward, as important achievements have been lost. For instance, over 100 years ago the anthropologist W. H. R. Rivers published an essay, “The Disappearance of Useful Arts,” that documented how various peoples of Oceania lost such vital arts as the making of canoes, pottery, and the bow and arrow. Henrich himself documents that when, about 12,000 years ago, due to the end of the last ice age and rising sea levels, Tasmania became separated from the Australian mainland and so its population became isolated from the rest of Australia, the Tasmanians lost bone tools, fishing, watercraft, fitted clothing, and even the ability to start a fire. Again, writing is another cultural attainment that has been lost repeatedly at various places and times. One way cultural attainments can get lost is if some catastrophe wipes out all the specialists in a culture who possess the relevant know-how. Henrich has other and more interesting ideas I won’t go into. The important point is that this evidence demonstrates conclusively how much of our know-how—even seemingly quite basic things—depends on cultural accumulation and is beyond the power of single individuals or even collaborative groups in a single generation to create. And cultural accumulation in turn is strongly aided by norm psychology, by our ready willingness to model our performance on that of those who seem adept—to take their performance as “correct,” without needing to understand all the whys and wherefores. If this is right, then norm psychology did not need joint intentionality to get started.

I suspect that Haidt is enchanted by the joint intentionality idea because of its emphasis on shared norms. It might thus seem to explain how we became ultrasocial: commitment to shared norms could readily come to include prosocial norms that bind groups together and promote their smooth internal functioning. But this is a non sequitur. A norm’s being shared hardly entails or even suggests that it be prosocial. (Here’s a possible shared norm: “Every man for himself!”) Therefore, a base of norm sharing is not required to explain the evolution of human cooperation and group-mindedness. Cultural learning and norm psychology alone are enough. Given the presence of group competition and group selection, groups in which prosocial norms took hold would outcompete groups in which they didn’t. And if such selection pressure prevailed for long enough, gene–culture coevolution would see to it that certain prosocial attitudes and traits became genetically fixed. This is the story Henrich tells (in his chapters 10 and 11), and it is really no different from Haidt’s, except that Henrich places our capacity for accumulation of culture at the center of his account—and thus explains a great deal more than just our groupishness—instead of joint intentionality—which, I have argued, does not have the explanatory power that is being claimed for it.