Maria Nieves and Juan Carlos Copes are passionate about Argentine tango. They have been dancing together for 40 years and are among the best-known dance pairs in the world. Copes was once heard to say that if he had not found Nieves--someone to whom he is remarkably attuned--he would need four different partners to fully explore the tango's expressive spectrum. Anyone who has ever done pairs dancing will understand just how difficult it is to forge a merger out of differing styles and capacities, while coordinating movements with near perfection in space and time.

These coordination problems fascinate researchers who study cooperation. Whereas Copes and Nieves have learned to harmonize their movements to an extraordinary extent, it is clear that ordinary people are constantly attuning to one another, even during the most mundane daily activities. We set the table together, carry a large package or navigate between other drivers in heavy traffic. As Harvard University social psychologist Floyd Henry Allport recognized more than 80 years ago, such daily acts of cooperation are anything but trivial. In the end, two or more persons must coordinate and fine-tune their thoughts and actions. And, unlike dancers, we often have no opportunity to rehearse this choreography.

Prior knowledge and cultural conventions help sometimes. Let us assume that you are told to meet a stranger in Paris tomorrow. Except for the date "tomorrow in Paris," you know absolutely nothing. In such cases, people tend to orient themselves to the salient points in space and time. To maximize the chances that your paths will cross, you would be more likely to head toward the Eiffel Tower than to some nameless alley. In addition, you would be more likely to go there at noon than at 3:50 A.M.--on the assumption that the other person would do likewise.

Deciding on Turquoise
The basis for cooperation--the common ground, so to speak--frequently gets built up only during an interaction itself. Cognitive psychologist Simon Garrod of the University of Glasgow and Martin Pickering of the University of Edinburgh in Scotland have studied the role of speech as an instrument for coordination.

The researchers discovered that people tend to agree quickly and involuntarily on common concepts. For example, if two individuals are talking about a turquoise-colored tie and one party says at the outset that it is green, the other party will tend to identify it as green as well. This implicit agreement greatly simplifies understanding because, as turquoise is somewhere between blue and green on the color spectrum, it could be called any number of things, which can easily lead to misunderstanding. Often people even attune their sentence structure and dialect to the other person so that the conversation can go off without a hitch.

But what happens when speech is simply not the appropriate mode? Spoken instructions do not work for activities that require rapid-fire coordination, such as the tango. Words take too long to convey an idea. Over the past several years, researchers have been studying the role nonverbal cues play in the coordination of action and have discovered several mechanisms, most of which occur unconsciously.

For example, people are able to recognize in a flash where another person's attention is focused and to then redirect their gaze to the same object or event. This capacity for "joint attention" develops quite early. Long before their first birthday, babies will follow the gaze of another person. When babies are 12 to 18 months old, they already understand that when someone is looking at the same object as they are, the other person sees the same object at that moment.

Psychologist Herbert H. Clark of Stanford University and Meredyth Krych of Montclair State University have demonstrated just how important joint attention is for cooperative action. The researchers gave pairs of people instructions for building a model out of Lego blocks. One of the two test subjects played the "director," who read out instructions from a prepared manual. It turned out that the pairs worked faster and made fewer errors when the supervisor and the builder not only talked to each other but could see each other and the blocks as well. In contrast, when they were separated by a partition, coordination became considerably more difficult.

A brief glance is also often more than enough to allow us to recognize what another person is doing--and what that action will accomplish. For example, when we see someone picking up a glass of water in a restaurant, we can pretty much predict that she will raise it to her mouth and drink. A number of neuropsychological and brain-imaging studies indicate a close connection between the perception of what others are doing and our own planning and control mechanisms.

A direct connection between observation and the execution of actions was first observed in macaques. The "mirror neurons" in the premotor and parietal cortices of the brain fire not only whenever the monkey executes an action but also when the animal observes other monkeys performing that action. Since these initial findings, numerous studies have shown that the same areas of the human brain are activated when we act and when we observe others [see "A Revealing Reflection," by David Dobbs; Scientific American Mind, April/May 2006].

For example, brain researcher Beatriz Calvo-Merino and her colleagues at University College London showed dance films to ballet and capoeira dancers. In all participants, activity was seen in brain regions that are otherwise activated only when the subjects themselves dance. Interestingly, brain activity increased when the test subjects observed their own type of dancing, whereas when a ballet dancer, for example, watched a performance of capoeira, her gray matter cells had a weaker response. The more similar the actions of the observer and the observed, the greater the resonance in the brain's motor system.

A close connection between perception and action may also explain why we occasionally mimic actions, body language and facial expressions. For example, when two friends drink a glass of wine in a restaurant, they may raise their glasses at the same time. Likewise, people often cross their legs when someone they are talking to crosses theirs. This phenomenon of involuntary impersonation may arise from the activation of the brain's action programs with sufficient strength to trigger movements spontaneously.

Our talent for mimicry may serve an important purpose. Some studies imply that spontaneous imitation acts as a "social glue," promoting feelings of friendliness and a sense of togetherness. Studies conducted by Tanya L. Chartrand and her colleagues at Duke University, for example, have shown that people tend to assess those who frequently ape their movements during a conversation as friendlier; people who do not echo actions are viewed as less agreeable. Chartrand also showed that people who feel socially excluded from group activities tend to mimic others more often--presumably to get back in their good graces. When our actions resemble those that we observe in others, it is taken as a sign of unity.

This tendency for coordination with others is so deeply ingrained that people will follow it even when it hampers their ability to perform an assigned task, as suggested by a series of studies that I conducted with my Rutgers University colleague Guenther Knoblich and Wolfgang Prinz of the Max Planck Institute for Human Cognitive and Brain Sciences in Munich. On a computer screen, a test subject viewed a succession of images of a hand, each of which pointed to the left or to the right. Each index finger wore either a red or a green ring. The subject was told to hit a key on her right with her right hand whenever a green ring appeared, to hit the key on her left side with her left hand for a red ring, and, in the process, to ignore the direction of the virtual finger. Nevertheless, most test subjects were influenced by finger direction. If, for example, the hand with the green ring pointed to the left, they tended to hesitate before responding with the key on the right.

"You Are (Always) on My Mind"
We repeated this experiment with two persons sitting next to each other looking at the same monitor. Each test subject could press only one key. The person on the right was instructed to respond only to green rings by hitting her key; the one on the left was responsible for red rings.

The direction in which the index finger pointed affected the test subjects in this experiment as well. They reacted faster when the finger pointed in their direction. Yet when the subjects carried out exactly the same task by themselves--reacting only to red or green--the pointing direction of the finger did not affect their reaction time.

So two people, each of whom assumes part of a task, behave as a single person who is responsible for both parts of that task. Apparently, people pay attention not only to their own instructions but also to the potential actions of the other person. Whenever the hand points to my partner, a mental representation is automatically activated in my brain that relates to the potential action of my partner. There is a lag time before I realize that although the hand is pointing to the other person, I need to react to my color. That is why I will hesitate for a moment before hitting my key. In further studies on healthy test subjects, we showed that the "integration effect" described above is generalized well beyond tasks involving rings, hands and computer monitors.

We wondered whether this impulse to cooperate might be absent in some people as a result of a neurological disorder. People with autism, for example, are often unable to draw conclusions about what others are thinking or feeling in a particular situation--they have trouble attributing mental states to other people. The question was whether this stands in the way of their ability to integrate the actions of others with their own behavior. In collaboration with Luitgard Stumpf of the Integration Center for Autistic Persons in Munich, we used the ring experiment to determine whether autistic adults of normal intelligence take other people into account just as nonautistic persons do, even when their own task does not actually require it.

To our surprise, subjects with autism behaved exactly like the other test subjects. The same task led to different patterns in reaction time, depending on whether the task was done alone or in concert with a second person. We concluded that the basic connections between perception and action, which support social interactions, may be completely intact in autistic persons--even though these individuals have difficulty intuiting the thoughts of others. It will fall to other experiments to show whether people with autism perform as well in collaborative situations that require a higher level of coordination. It may be that the automatic connection between perception and action developed very early in our evolution--even before human beings were able to infer what others were thinking.

Suppressing Imitation
It is interesting to note that some forms of joint action require special effort to ensure that people do not react when it is the other person's turn. The ability to take turns plays a crucial role in conversation, paddling a canoe in unison, or playing a piano duet. Looked at another way, it seems that the strong instinct to mimic must be suppressed to make coordination possible. In our experiments, we were actually able to measure this suppression in the electrical activity of the brain using electroencephalography. A specific electrophysiological component--called Nogo P3--shows the magnitude of inhibition processes needed to suppress an action, such as when the red-ring test subject in a pair sees a green ring and is therefore not supposed to hit his key. Sure enough, more inhibition was recorded when subjects were working in pairs than when they worked alone.

Knowing that a partner will join us in a task can even alter our perception of situations and objects. It has long been known that we perceive things differently depending on our intentions and the resources at our disposal. For example, psychologist Dennis R. Proffitt of the University of Virginia demonstrated that a hill appears steeper when we are carrying a heavy backpack--something any hiker will confirm.

Similarly, someone's estimate of a box's weight will differ, depending on whether the individual has to lift it alone or with someone else. This phenomenon was demonstrated by an experiment conducted by Maggie Shiffrar and me at Rutgers, in which test subjects were asked to estimate the weight of transparent boxes that were filled with different quantities of potatoes. If the subject believed that another participant was going to help lift the box, her weight estimates were actually lower than if she thought she would have to hoist it by herself. Persons in groups underestimate weight across the board, even though their estimates may be on the mark when they do so alone.

These results suggest that we may see the world not only through our own eyes but also through the eyes of the groups we form. We plan our actions guided partly by what we think we can achieve with others. Our tendency to take others into account may have its roots in our evolutionary history. Those who were able to coordinate their actions with others may have had many advantages. One of the first forms of cooperative action might have been activities where two or more people performed the same action at the same time, for example, when they pushed a heavy rock in front of a cave to protect its entrance. Later, individuals may have started to engage in complementary actions, such as when one person chased an animal so that others could catch it.

Research from our lab and others suggests that the challenges posed by these different forms of joint action shaped our perception-action system and our unconscious cognitive processes. In many cases, then, cooperation is not just an exercise of social duty. Rather we simply cannot do otherwise.