The Ego Tunnel: The Science of the Mind and the Myth of the Self - Thomas Metzinger (2009)

Part II. IDEAS AND DISCOVERIES

Chapter 6. THE EMPATHIC EGO

Have you ever watched a child who has just learned to walk run toward a desired object much too quickly and then trip and fall on his face? The child lifts his head, turns, and searches for his mother. He does so with a completely empty facial expression, showing no kind of emotional response. He looks into his mother’s face to find out what has happened. How bad was it, really? Should I cry or should I laugh?

Toddlers do not yet have an autonomous self-model (though probably none of us has a self-model that is truly independent from others). In such small children, we observe an important fact about the nature of our own phenomenal Ego: It has social correlates as well as neural correlates. The toddler does not yet know how he should feel; therefore he looks at his mother’s face in order to define the emotional content of his own conscious self-experience. His self-model does not yet have a stable emotional layer to which he could attend and, as it were, register the severity of what just happened. The fascinating point is that here are two biological organisms that just a few months ago, before being physically separated at birth, were one. Their Egos, their phenomenal self-models, are still intimately coupled on the functional level. When the toddler gazes at his mother and starts to smile in relief, there is a sudden transition in his PSM. Suddenly, he discovers that he didn’t hurt himself at all, that the only thing that happened to him was a big surprise. An ambiguity is resolved: Now he knows how he feels.

There are kinds of self-experience that an isolated being could never have. Many layers of our self-model require social correlates; more than that, they are frequently created by some sort of social interaction. It is plausible to assume that if a child does not learn to activate the corresponding parts of his emotional Ego during a certain crucial period of his psychological development, he will not be able to have those feelings as an adult. We can enter certain regions in our phenomenal-state space only with the help of other human beings. In a more general sense, certain types of subjective experiences—interpersonal connectedness, trust, friendship, self-confidence—may be more or less available to each of us. The degree to which individuals have access to their emotional states varies. The same is true of their capacity for empathy and the ease with which they can read the minds of other human beings. Ego Tunnels develop in a social environment, and the nature of this environment determines to what extent one Ego Tunnel can resonate with other Ego Tunnels.

So far, we have been concerned only with how the world and the self appear in the tunnel created by the brain. But what about other selves? How can other agents with other goals, other thinkers of thoughts, other feeling selves, become parts of one’s own inner reality? We can also express this question in philosophical terms. At the beginning of this book, we asked how a first-person perspective can emerge in the brain. The answer was that it does so through the creation of the Ego Tunnel. Now we can ask, What about the second-person perspective? Or the “we,” the first-person-plural perspective? How does the conscious brain manage to get from the “I” to the “you” and the “we”? The thoughts, goals, feelings, and needs of other living beings in our environment constitute part of our own reality; therefore, it is vital to understand how our brains were able to represent and create not just the inward perspective of the Ego Tunnel but also a world containing multiple Egos and multiple perspectives. Perhaps we will discover that large parts of the first-person perspective did not simply emerge in the brain but were in part causally enabled by the social context we all found ourselves in from the very beginning.

The self-model theory holds that certain new layers of consciousness, unique to the self-model of Homo sapiens, made the transition from biological to cultural evolution possible. This process started on an unconscious, automatic level in our brains, and its roots reach far down into the animal kingdom. There is an evolutionary continuity to such high-level social phenomena as the unique human capacity for consciously acknowledging others as rational subjects and moral persons. In chapter 2, I pointed out that in the history of ideas, the concept of “consciousness” was intimately related to possession of a “conscience”—the higher-order ability to assess the moral value of your lower-order mental states or your behavior. What kind of self-model do you need in order to become such a moral agent? The answer could have to do with the progression from a mental representation of the first-person-singular perspective to that of the first-person plural, along with the ability to represent mentally what the benefits (or risks) of a particular action would be for the collective as a whole. You become a moral agent by taking the coherence and stability of your group into account. In this way, the evolution of morals may have had a lot to do with an organism’s ability to distance itself mentally from a representation of its individual interests and consciously and explicitly to represent principles of group selection, even if this involved self-damaging behavior. Recall that the beautiful early philosophical theories of consciousness-as-conscience rested on installing an ideal observer in your mind. I believe the human self-model was successful because it installed your social group as an ideal observer in your mind, and to a much stronger degree than was the case in any other primate brain. This created a dense causal linkage between global group-control and global self-control—a new kind of ownership, as it were.

Investigators of these phenomena will have to look at chimpanzees and macaques, at swarms of fish and flocks of birds, and maybe even at ant colonies. They will also have to look at the way infants imitate their parents’ facial expressions. Intersubjectivity started deep down in the realms of biological behavior coordination, in the motor regions of the brain and the unconscious layers of the Ego. Intersubjectivity is anchored in intercorporality.

SOCIAL NEUROSCIENCE: CANONICAL NEURONS AND MIRROR NEURONS

Sociological and biological approaches to human consciousness have traditionally been treated as antagonistic to each other, or at least mutually exclusive. But today, in the new discipline of social neuroscience, the assumption is that a multilevel integrative analysis may be required and that a common scientific language, grounded in the structure and function of the brain, can contribute to it. The self-model theory is an attempt to develop exactly this type of language.

It has been known since the 1980s that there is a particularly interesting class of neurons in an area called F5 in the ventral premotor region of the monkey brain. These neurons are part of the unconscious self-model; they code body movements in a highly abstract way. Giacomo Rizzolatti, a professor of human physiology at the University of Parma and a pioneer in this exciting field of research, uses the concept of a “motor vocabulary” that consists of complex inner images of actions as a whole. Words in the monkey’s motor vocabulary might be “reach,” “grasp,” “tear,” or “hold.” The interesting aspect of this discovery is that there is a specific part of the brain that describes the monkey’s—and our own—actions in a holistic manner. This description includes the goals of the actions and the temporal pattern in which the actions unfold. The actions are portrayed as relations between an agent and the target object (a piece of fruit, say) of his action.1

Now we know that human beings, too, possess something similar. From a neurocomputational perspective, this system in our brains makes sense: By developing an inner vocabulary for possible actions, we reduce the immense space of possibilities to a small number of stereo-typical body movements. This allows us, for instance, to perform the same grasping movement in widely differing situations (recall the Alien Hand syndrome of chapter 4).

One of the most fascinating features of these so-called canonical neurons is that they also respond to the visual perception of objects in our environment. Our brain does not simply register a chair, a teacup, an apple; it immediately represents the seen object as what I could do with it—as an affordance, a set of possible behaviors. This is something I could sit on, this is something I could hold in my hands, this is something I could throw. While we’re seeing an object, we are also unconsciously swimming in a sea of possible behaviors. As it turns out, the traditional philosophical distinction between perception and action is an artificial one. In reality, our brains employ a common coding: Everything we perceive is automatically portrayed as a factor in a possible interaction between ourselves and the world. A new medium is created, blending action and perception into a novel, unified representational format. The second fascinating discovery about canonical neurons is that you also use them for self-representation. The motor vocabulary is part of the unconscious self-model, because it describes the goal-directed movements of one’s body. The unconscious precursors of the phenomenal Ego in our brain thus play an essential and central role in our perception of the world around us.

In the 1990s, researchers discovered another group of neurons. Also a part of area F5, they fire not just when monkeys perform object-directed actions, such as grasping a peanut, but also when they observe others performing the same type of action. Because these neurons respond to actions performed by others, they are termed mirror neurons. They are activated when another agent is observed using objects in a purposeful way. Thus, we are matching the bodily behaviors we observe in others with our own internal motor vocabulary. This action /observation matching system helps us understand something we could never understand using our sensory organs alone—that other beings in our environment pursue goals. We use our own unconscious self-model to put ourselves in the shoes of others, as it were. We use our own “motor ideas” to understand someone else’s actions by directly mapping them onto our own inner repertoire, by automatically triggering an inner image of what our goal would be if our body also moved that way.2 The conscious experience of understanding another human being, the subjective feeling that pops up in the Ego Tunnel when we intuitively grasp what others’ goals are and what is going on in their minds, is the direct result of these unconscious processes.3

The conscious self is thus not only a window into the internal workings of one’s own Ego but also a window into the social world. It is a two-way window: It elevates to the level of global availability the unconscious and automatic processes that organisms constantly use to represent one another’s behavior. This is how these processes become part of the Ego Tunnel, an element of our subjective reality. They lead to an enormous expansion and enrichment of our inner simulation of the world. As soon as our brains are able to represent not only events but also actions—that is, goal-directed events caused by other beings—we are not alone anymore. Others exist, with minds of their own. The fact that more than one Ego Tunnel might exist in the world is now reflected in our own tunnel. We can develop our conscious-action ontology, and we can put it to use by sharing it with others.4

A considerable body of evidence using a variety of neuroimaging techniques shows that the mirror-neuron system exists not just in monkeys but in humans as well. However, it appears that the system in humans is much more generalized and does not depend on concrete effector-object interactions; consequently, it can represent a much greater variety of actions than it does in monkeys. In particular, researchers have now discovered mirror-neuron systems that seem to achieve similar effects for emotions and for pain and other bodily sensations. When human test subjects are shown pictures of sad faces, for example, they subsequently tend to rate themselves as sadder than they were before—and after being shown happy faces they tend to rate themselves as happier. Converging empirical data show that when we observe other human beings expressing emotions, we simulate them with the help of the same neural networks that are active when we feel or express these emotions ourselves.5 For instance, certain regions in the insular cortex are activated when subjects are exposed to a disgusting smell, and the same regions are active when we see an expression of disgust on another person’s face. A common representation of the emotional state of disgust is activated in our brains whether we experience it ourselves or observe it in another individual. Parallel observations in the amygdala have been made for fear.6 It is interesting to note that our ability to recognize a particular feeling in another human being can be weakened or switched off by blocking the relevant parts of the mirror-neuron system. It is believed, for example, that certain areas in the ventral striatum of the basal ganglia are necessary in recognizing anger; patients with damage to this area show impairment in identifying aggression signals emitted by others. If these areas are blocked pharmacologically (by interfering with dopamine metabolism), subjects can recognize other emotions but can no longer recognize anger.7 Similar observations have been made for pain. Recent fMRI (functional magnetic resonance imaging) experiments show that areas in the anterior cingulate cortex and the interior insular cortex are active when we experience pain but also when we observe someone else experiencing pain.8Interestingly, only the emotional part of the pain system is activated; the part associated with the purely sensory aspect of pain is not. This makes perfect sense, because the sensory aspect is exactly what we cannot share with anyone else: We cannot share the cutting, throbbing, or burning sensory quality of pain, but we can feel empathy with regard to the emotions it causes.

Other neuroimaging experiments have demonstrated that a similar principle exists for other bodily sensations. Certain higher levels of the somatosensory cortex are activated both when subjects observe others being touched and when they are touched themselves. Again, the immediate sensory quality associated with the activation of the primary somatosensory cortex cannot be shared, but a higher level in the bodily Ego is active regardless of whether we are being touched or just observing someone being touched. There seems to be an underlying principle uniting these new empirical discoveries: Certain layers of our self-model function as a bridge to the social domain, because they can directly map abstract inner descriptions of what is going on in ourselves onto those of what goes on in other people.

Of course, intersubjectivity is not only about the body and emotions. Thinking plays a role as well. Reason-based forms of empathy appear to involve yet other parts of the brain—specifically, the ventromedial prefrontal cortex. Still, the discovery of mirror neurons helps us to understand that empathy is a natural phenomenon, acquired step by step in the course of our biological evolution. First, we developed the self-model, because we had to integrate our sensory perceptions with our bodily behavior. Then this self-model became conscious, and the phenomenal self-model was born into the Ego Tunnel, allowing us to achieve global control of our bodies in a much more selective and flexible manner. This was the step from being an embodied natural system that has and uses an internal image of itself as a whole to a system that, in addition, consciously experiences this fact.9 The next evolutionary step was what Vittorio Gallese, Rizzolatti’s colleague at Parma and one of the leading researchers in the field, has called embodied simulation,.10 In order to understand the feelings and goals of other human beings, we use our own body-model in the brain to simulate them.

As recent neuroscientific findings show, this process also cuts across the border between the unconscious and the conscious. A considerable part of this constant mirroring activity happens outside the Ego Tunnel, and thus we have no subjective experience of it. But from time to time, when we deliberately attend to other people or analyze social situations, the conscious self-model is involved as well; in particular, as noted, we can somehow directly comprehend, almost perceive, what somebody else is up to. Often, we “just know” what the purpose of the other person’s action is and what his likely emotional state is.11 We use the same internal resources that make us aware of our own goal states to discover automatically that others are goal-directed entities themselves and not just other moving objects. We can experience them as Egos because we experience ourselves as Egos. Whenever successful social understanding and empathy are achieved, we share a common representation: of one and the same goal state in two different Ego Tunnels. Social cognition has now become tractable to empirical neuroscience on the level of single-cell recordings—showing us not only how Ego Tunnels started to resonate with each other but also how complex cooperation and communication between self-conscious organisms were able to evolve and lay the foundations for cultural evolution.

My idea is that social cognition rests on what is sometimes called an exaptation. Adaptation led to an integrated body-model in the brain and to the phenomenal self-model. Then the existing neural circuitry was “exapted” for another form of intelligence: It suddenly proved useful in tackling a different set of problems. This process began with low-order motor resonance; then, second- and third-order embodiment12 led to embodied simulation as a brand-new tool in developing social intelligence. Like everything else in evolution, this process was driven by chance. There was no purpose behind it, but it eventually led us where we are today—to the formation of intelligent, scientific communities peopled by conscious agents trying to understand this very process itself.

The new emerging general picture is inspiring: We are all constantly swimming in an unconscious sea of intercorporality, permanently mirroring one another with the aid of various unconscious components and precursors of the phenomenal Ego. Long before conscious, high-level social understanding arrived on the scene, and long before language evolved and philosophers developed complicated theories about what it takes for one human being to acknowledge another as a person and a rational individual, we were already bathed in the waters of implicit, bodily intersubjectivity. Few great social philosophers of the past would have thought that social understanding had anything to do with the premotor cortex, and that “motor ideas” would play such a central role in the emergence of social understanding. Who could have expected that shared thought would depend upon shared “motor representations”? Or that the functional aspects of the human self-model that are necessary for the development of social consciousness are nonconceptual, prerational, and pretheoretical? The first inklings of these ideas came at the end of the nineteenth and the first half of the twentieth century, when there were numerous attempts in experimental psychology to better understand so-called ideomotor phenomena.13 Philosopher Theodor Lipps wrote about Einfühlung (empathy) in 1903—that is, the ability, as he put it, to “feel yourself in an object.” He had already spoken of “inner imitation” and of “organic feelings.” For him, objects of empathy could be not only the movements or postures we perceive in other human beings but also objects of art, architecture, and even visual illusions. He held that aesthetic pleasure was “objectified”—that is, “the object is ego and thereby the ego object.” 14 Social psychologists began talking about concepts such as “virtual body movements” and “motor mimicry” or “motor infection” decades ago.

From a philosophical perspective, the discovery of mirror neurons is exciting because it gave us an idea of how motor primitives could have been used as semantic primitives: that is, how meaning could be communicated between agents. Thanks to our mirror neurons, we can consciously experience another human being’s movements as meaningful. Perhaps the evolutionary precursor of language was not animal calls but gestural communication.15 The transmission of meaning may initially have grown out of the unconscious bodily self-model and out of motor agency, based, in our primate ancestors, on elementary gesturing. Sounds may only later have been associated with gestures, perhaps with facial gestures—such as scowling, wincing, or grinning—that already carried meaning. Still today, the silent observation of another human being grasping an object is immediately understood, because, without symbols or thought in between, it evokes the same motor representation in the parieto-frontal mirror system of our own brain. As Professor Rizzolatti and Dr. Maddalena Fabbri Destro from the Department of Neuroscience at the University of Parma put it: “[T]he mirror mechanism solved, at an initial stage of language evolution, two fundamental communication problems: parity and direct comprehension. Thanks to the mirror neurons, what counted for the sender of the message also counted for the receiver. No arbitrary symbols were required. The comprehension was inherent in the neural organization of the two individuals.”16

Such ideas give a new and rich meaning not only to the concepts of “grasping” and “mentally grasping the intention of another human being,” but, more important, also to the concept of grasping a concept—the essence of human thought itself. It may have to do with simulating hand movements in your mind but in a much more abstract manner. Humankind has apparently known this for centuries, intuitively: “Concept” comes from the Latin conceptum,meaning “a thing conceived,” which, like our modern “to conceive of something,” is rooted in the Latin verb concipere, “to take in and hold.” As early as 1340, a second meaning of the term had appeared: “taking into your mind.” Surprisingly, there is a representation of the human hand in Broca’s area, a section of the human brain involved in language processing, speech or sign production, and comprehension. A number of studies have shown that hand/arm gestures and movements of the mouth are linked through a common neural substrate. For example, grasping movements influence pronunciation—and not only when they are executed but also when they are observed. It has also been demonstrated that hand gestures and mouth gestures are directly linked in humans, and the oro-laryngeal movement patterns we create in order to produce speech are a part of this link.

Broca’s area is also a marker for the development of language in human evolution, so it is intriguing to see that it also contains a motor representation of hand movements; here may be a part of the bridge that led from the “body semantics” of gestures and the bodily self-model to linguistic semantics, associated with sounds, speech production, and abstract meaning expressed in our cognitive self-model, the thinking self. Broca’s area is present in fossils of Homo habilis, whereas the presumed precursors of these early hominids lacked it. Thus the mirror mechanism is conceivably the basic mechanism from which language evolved. By providing motor copies of observed actions, it allowed us to extract the action goals from the minds of other human beings—and later to send abstract meaning from one Ego Tunnel to the next.

The mirror-neuron story is attractive not only because it bridges neuroscience and the humanities but also because it illuminates a host of simpler social phenomena. Have you ever observed how infectious a yawn is? Have you ever caught yourself starting to laugh out loud with others, even though you didn’t really understand the joke? The mirror-neuron story gives us an idea of how groups of animals—fish schools, flocks of birds—can coordinate their behavior with great speed and accuracy; they are linked through something one might call a low-level resonance mechanism. Mirror neurons can help us understand why parents spontaneously open their mouths while feeding their babies, what happens during a mass panic, and why it is sometimes hard to break away from the herd and be a hero. Neuroscience contributes to the image of humankind: We are all connected in an intersubjective space of meaning—what Vittorio Gallese calls a “shared manifold.”17

CHAPTER SIX APPENDIX THE SHARED MANIFOLD: A CONVERSATION WITH VITTORIO GALLE SE

01

Vittorio Gallese is professor of human physiology in the Department of Neurosciences of the University of Parma, Italy. As a cognitive neuroscientist, he focuses his research interests on the relationship between the sensory-motor system and cognition in primates, both human and nonhuman, using a variety of neurophysiological and neuroimaging techniques. Among his major contributions is the discovery, with his colleagues in Parma, of mirror neurons and the elaboration of a theoretical model of the basic aspects of social cognition. He is developing an interdisciplinary approach to the understanding of intersubjectivity and social cognition, in collaboration with psychologists, psycholinguists, and philosophers. In 2002, he was the George Miller visiting professor at the University of California at Berkeley. In 2007 he received the Grawemeyer Award for Psychology for the discovery of mirror neurons. He has published more than seventy papers in international journals and is coeditor (with Maxim I. Stamenov) of Mirror Neurons and the Evolution of Brain and Language (2002).

Metzinger: Vittorio, what exactly do you mean by the shared manifold hypothesis. What is a shared manifold?

Gallese: The question I started out with is the following: How can we explain the ease with which we normally understand what is at stake when we interact with other people?

I used this term to characterize what happens when we witness the actions of others, or their overt behavior expressing the sensations and emotions they experience. Basically, it describes our capacity for direct and implicit access to the experiential world of the other. I think the concept of empathy should be extended in order to accommodate and account for all different aspects of expressive behavior enabling us to establish a meaningful link with others. This enlarged notion of empathy is captured by the term “shared manifold.” It opens up the possibility of giving a unified account of important aspects and possible levels of description of intersubjectivity. I tried on purpose not to employ the term “empathy,” because it systematically induces misunderstandings, mainly because of its different connotations in different contexts. The shared manifold can be described at three different levels: a phenomenological level, a functional level, and a subpersonal level.

The phenomenological level is the one responsible for the sense of similarity—of being part of a larger social community of persons like us—that we experience anytime we encounter others. When confronting the intentional behavior of others, we experience a specific phenomenal state of intentional attunement. This phenomenal state generates the peculiar quality of familiarity with other individuals, produced by the collapse of the others’ intentions into those of the observer. This seems to be one important component of what being empathic is about.

The functional level can be characterized in terms of embodied simulations of the actions we see or of the emotions and sensations whose expression we observe in others.

The subpersonal level is instantiated as the activity of a series of mirroring neural circuits. The activity of these mirror neural circuits is, in turn, tightly coupled with multilevel changes within body-states. We have seen that mirror neurons instantiate a multimodal shared space for actions and intentions. Recent data show that analogous neural networks are at work to generate multimodal emotional and sensitive “we-centric” shared spaces. To put it in simpler words, every time we relate to other people, we automatically inhabit a we-centric space, within which we exploit a series of implicit certainties about the other. This implicit knowledge enables us to understand in a direct way what the other person is doing, why he or she is doing it, and how he or she feels about a specific situation.

Metzinger: You also speak of “embodied simulation.” What exactly does that mean? Is there also something like “disembodied simulation”?

Gallese: The notion of simulation is employed in many different domains, often with different, not necessarily overlapping, meanings. Simulation is a functional process that possesses a certain representational content, typically focusing on possible states of its target object. In philosophy of mind, the notion of simulation has been used by the proponents of the “Simulation Theory of Mind-Reading” to characterize the pretend state adopted by the attributer in order to understand another person’s behavior. Basically, we use our mind to put ourselves into the mental shoes of other human beings.

I qualify simulation as embodied in order to characterize it as a mandatory, automatic, nonconscious, prerational, nonintrospectionist process. A direct form of experiential understanding of others, intentional attunement, is achieved by the activation of shared neural systems underpinning what others do and feel and what we do and feel. This modeling mechanism is embodied simulation. Parallel to the detached sensory description of the observed social stimuli, internal representations of the body-states associated with actions, emotions, and sensations are evoked in the observer, as if he or she were performing a similar action or experiencing a similar emotion or sensation. Mirror-neuron systems are likely the neural correlate of this mechanism. By means of a shared neural state realized in two different physical bodies, the “objectual other” becomes another self. Defective intentional attunement, caused by a lack of embodied simulation, might explain some of the social impairments of autistic individuals.

I should add that—in contrast to what many cognitive scientists think—social cognition is not only social metacognition, that is, explicitly thinking about the contents of someone else’s mind by means of abstract representations. We can certainly explain the behavior of others by using our complex and sophisticated mentalizing ability. My point is that most of the time in our daily social interactions, we do not need to do this. We have a much more direct access to the experiential world of the other. This dimension of social cognition is embodied, in that it mediates between our multimodal experiential knowledge of our own lived body and the way we experience others. I therefore call simulation “embodied”—not only because it is realized in the brain but also because it uses a preexisting body-model in the brain and therefore involves a nonpropositional form of self-representation that also allows us to experience what others are experiencing.

Metzinger: Vittorio, according to our best current theories, what exactly is the difference between social cognition in monkeys or chimps and social cognition in human beings?

Gallese: The traditional view in the cognitive sciences holds that humans are able to understand the behavior of others in terms of their own mental states—intentions, beliefs, and desires—by exploiting what is commonly called folk psychology. The capacity for attributing mental states to others has been defined as “theory of mind.” A common trend on this issue has been to emphasize that nonhuman primates, apes included, do not rely on mentally based accounts of one another’s behavior.

This view prefigures a sharp distinction between all nonhuman species, which are confined to behavior reading, and our species, which makes use of a different level of explanation—mind-reading. However, it is by no means obvious that behavior-reading and mind-reading constitute two autonomous realms. As I said before, in our social transactions we seldom engage in explicit interpretive acts. Most of the time, our understanding of social situations is immediate, automatic, almost reflex-like. Therefore, I think it is preposterous to claim that our capacity for reflecting on the real intentions determining the behavior of others is all there is to social cognition. It is even less obvious that in understanding the intentions of others, we employ a cognitive strategy totally unrelated to predicting the consequences of their observed behavior.

The use of the belief/desire propositional attitudes of folk psychology in social transactions is probably overstated. As emphasized by Jerry [Jerome S.] Bruner, “When things are as they should be, the narratives of Folk Psychology are unnecessary.”18

Furthermore, recent evidence shows that fifteen-month-old infants recognize false beliefs. These results suggest that typical aspects of mind-reading, such as the attribution of false beliefs to others, can be explained on the basis of low-level mechanisms that develop well before full-blown linguistic competence.

The all-or-nothing approach to social cognition of mainstream cognitive science—its search for a mental Rubicon, the wider the better—is strongly arguable. When trying to understand our social-cognitive abilities, we should not forget that they are the result of a long evolutionary process. It is therefore possible that apparently different cognitive strategies are underpinned by similar functional mechanisms, which in the course of evolution acquire increasing complexity and are exapted to sustain cognitive skills newly emerged out of the pressure exerted by changed social and/or environmental constraints. Before drawing any firm conclusion about the mentalizing abilities of nonhuman species, methodological issues related to species-specific spontaneous abilities and environments should be carefully scrutinized.

A fruitful alternative strategy I fully endorse is that of framing the issue of the investigation of the neural bases of social cognition within an evolutionary perspective. The evolution of this cognitive trait seems to be related to the necessity of dealing with social complexities that arose when group-living individuals had to compete for scarce and patchily distributed resources.

Cognitive neuroscience has started to unveil, both in monkeys and in humans, the neural mechanisms at the basis of anticipating and understanding the actions of others and the basic intentions promoting them—the mirror-neuron system for action. The results of this ongoing research can shed light on the evolution of social cognition. The empirical data on mirror neurons in monkeys and on mirroring circuits in the human brain suggest that some of the typically human, sophisticated mentalizing skills—such as ascribing intentions to others—might be the outcome of a continuous evolutionary process, whose antecedent stages can be traced to the mirror mapping system of macaque monkeys.

Thus, as you are asking, what makes humans different? Language certainly plays a key role. But in a sense this answer begs the question, because then we must explain why we have language and other animals do not. At present, we can only make hypotheses about the relevant neural mechanisms underpinning the mentalizing abilities of humans, which are still poorly understood from a functional point of view.

One distinctive feature of our mentalizing abilities is our capacity for entertaining potentially infinite orders of intentionality: “I know that you know that I know . . .” and so on. One important difference between humans and monkeys could be the higher level of recursion attained—among other neural systems—by the mirror-neuron system for actions in our species. A similar proposal has recently been put forward in relation to the faculty of language, another human faculty characterized by recursion and generativity. Our species is capable of mastering hierarchically complex phrase-structure grammars, while nonhuman primates are confined to the use of much simpler finite-state grammars. A quantitative difference in computational power and degree of recursion could produce a qualitative leap forward in social cognition.

Metzinger: Can you speculate about the role of mirror neurons in the transition from biological to cultural evolution?

Gallese: A possibility is that mirror neurons and the embodied simulation mechanisms they underpin might be crucial for learning how to use the cognitive tools of folk psychology. This typically occurs when children are repeatedly exposed to the narration of stories. In fact, embodied simulation is certainly at play during language processing. But certainly the aspect of human culture that is more likely to benefit from mirror neurons is the domain of imitation, the domain of our incredibly pervasive mimetic skills. If it is true that ours is basically a mimetic culture, then mirror neurons, which are deeply involved in imitation and imitative learning, certainly are one important and basic ingredient of this crucial cultural transition. And indeed there is plenty of evidence that when we imitate simple motor acts, such as lifting a finger, or learn complex motor sequences, as when learning to play chords on a guitar, we do this by employing our mirror neurons. But instead of drawing a line between species like ours, who are fully competent in imitation, and other species, where this capacity is at best only emerging—again, we are dealing here with the anthropocentric dichotomies so appealing to many of my colleagues—we should concentrate on understanding why mimetic skills are so important for the cultural evolution of our species. And to answer this question, we must place the issue of mimesis in the larger context of our peculiar social cognition, in which the period of parental care is much longer than in all other species. There is a clear-cut relationship between the prolonged dependency of infants on their parents and the learning processes that this dependency promotes. The longer the period of infantile dependency, the greater the opportunities to develop complex emotional and cognitive strategies of communication. Increased communication in turn fosters cultural evolution. Given the central role that mirror neurons seem to play in establishing meaningful bonds among individuals, their connection to cultural evolution seems very plausible.

For most of history, the culture of our species has been an oral culture, where the transmission of knowledge from one generation to the next had to rely on direct personal contact between the transmitter of cultural content and the recipient of the cultural transmission. As pointed out by scholars like [Walter J.] Ong and [Eric A.] Havelock, for millennia cultural transmission had to rely on the same cognitive apparatus we still exploit in our interpersonal transactions—that is, our ability to identify and empathize with others. Again, I think that if we look at cultural evolution from this particular perspective, the role of mirror neurons appears to be central. At present, we are witnessing a cultural paradigm shift. The impact of new technologies, such as cinema, television, and more recently the Internet, with its massive introduction of multimediality, is drastically changing the way in which we communicate knowledge. The mediated, objective status of culture as transported by written texts like books is progressively being supplemented with a more direct access to the same contents by means of the new media of cultural fruition. This media revolution will most likely introduce cognitive changes, and I suspect that mirror neurons will again be involved.

Metzinger: In the field of social cognitive neuroscience, what do you consider to be the most burning and urgent questions for the future, and in which direction is the field moving?

Gallese: The first point I would like to make is a methodological one. I think we should definitely try to focus more strongly on the nature of the subjects of our investigations. Most of what we know about the neural aspects of social cognition—with few exceptions pertaining to the study of language—derives from brain-imaging studies carried out on Western-world psychology students! Even with present technologies, we could do a lot better than this. It is an open question whether cognitive traits and the neural mechanisms underpinning them are universal or, at least to a degree, the product of a particular social environment and cultural education. To answer this question, we need an ethno-neuroscience.

Second, even within the average sample of subjects normally studied by social cognitive neuroscientists, we do not know—or at best know very little—to what extent the results correlate with specific personality traits, gender, professional expertise, and the like. In sum, we should move from the characterization of an unrealistic “average social brain” to a much more fine-grained characterization.

A third issue I would like to see addressed more specifically in the near future is the role played by embodied mechanisms in semantic and syntactical aspects of language. Let me be clear about this. Even though I spent a considerable part of my scientific career investigating prelinguistic mechanisms in social cognition, I do not think you can avoid language if the ultimate goal is to understand what social cognition really is. All our folk psychology is language-based. How does this square with the embodied approach to social cognition? To me, this is a burning question.

A fourth important point pertains to the phenomenological aspects of social cognition. I think we should try to design studies in which a correlation can be drawn between particular patterns of brain activation and specific qualitative subjective experiences. Single case studies are now possible with high-resolution brain imaging. I am fully aware that dealing with subjective states is a tricky issue, from which empirical science so far has tried to stay clear, for many good reasons. But in principle it should be possible to carefully design well-suited and well-controlled experimental paradigms to crack the boundaries of subjective phenomenal states.

Metzinger: Vittorio, you have repeatedly cornered me with pressing questions about Edmund Husserl, Maurice Merleau-Ponty, and Edith Stein. Why are you so interested in philosophy, and what kind of philosophy would you like to see in the future? What relevant contributions from the humanities are you expecting?

Gallese: Scientists who believe that their discipline will progressively eliminate all philosophical problems are simply fooling themselves. What science can contribute to is the elimination of false philosophical problems. But this is a totally different issue.

If our scientific goal is to understand what it means to be human, we need philosophy to clarify what issues are at stake, what problems need to be solved, what is epistemologically sound and what is not. Cognitive neuroscience and philosophy of mind deal with the same problems but use different approaches and different levels of descriptions. Very often, we use different words to speak about the same things. I think all cognitive neuroscientists should take classes in philosophy. Similarly, philosophers—at least, philosophers of mind—should learn a lot more about the brain and how it works. We need to talk to one another much more than we are doing now. How can you possibly investigate social cognition without knowing what an intention is, or without understanding the concept of second-order intentionality? Similarly, how can you possibly stick to a philosophical theory of cognition if it is patently falsified by the available empirical evidence? There is another aspect for which I think philosophy may be helpful. Our scientific bravado sometimes makes us think we are the first to have thought about something. Most of the time, this is not true!

As I said, philosophy should listen more carefully to the results of cognitive neuroscience. But things are changing rapidly. The current situation is much better than it was ten years ago. There are more and more chances for multidisciplinary exchanges between our disciplines. One of my PhD students, currently involved in neurophysiological experiments, has a degree in philosophy.

Broadening these considerations to the overall field of the humanities, I think incredibly fruitful contributions can result from a dialogue with anthropology, aesthetics, and literary and film studies. As I said before, a mature social cognitive neuroscience can’t limit itself to scanning brains in a lab. It must be open to the contributions from all these disciplines. I am rather optimistic. I see a future of ever-growing and stimulating dialogue between cognitive neuroscience and the humanities.