COMMUNICATIONS IN THE 21ST CENTURY The Mobile Information Society
	NEW PERSPECTIVES ON 21ST-CENTURY COMMUNICATIONS Conference, May 24-25, 2002 Hungarian Academy of Sciences

The electronic revolution of the last decade of the twentieth century can genuinely be said to have revolutionised our way of life. Things can happen faster, further away and on a wider scale than anyone could ever have dreamed possible even in the preceding decade. In the end, the extent to which we can exploit the opportunities offered by the new technologies depends only on our ability to take advantage of them.

In this lecture, I want to pursue two themes in this respect. One is to illustrate ways in which we can find new ways to exploit these technologies for purposes which were wholly unintended. The other is that, despite our ingenuity in this respect, our ability to exploit these new technologies may nonetheless be limited. Our cognitive machinery - the design of our brains - did not evolve in a world characterised by twenty-first century electronic technology, and this may impose real limitations on what we can in fact do with it. But let me begin with a brief introduction to the evolution of the social brain.
 

The Evolution of the Social Brain

Primates have unusually large brains for body size compared to all other vertebrate groups. This is primarily a consequence of the fact that primates have unusually large neocortices (essentially the thinking part of the brain). In primates, the neocortex represents in excess of 50% of total brain volume (ranging up to 80% in humans), whereas it never exceeds 50% in other non-primate mammals (and can be as low as 10%). The growing concensus is that the pressure to evolve large neocortices (and hence large brains) came from the fact that primates developed an intensely social style of life. So important was the careful management of the relationships that held their social groups together that primates had to evolve the computing capacity needed to deal with the information processing required to keep track of the ever-changing nature of social relationships and to use that information in managing a way through the social complexities of life in a primate group.

Known as the Machiavellian Intelligence (or Social Brain) hypothesis, this view has received considerable support from analyses of the co-variation between neocortex volume of primates and various aspects of primate social behaviour, including social group size, grooming clique size, the use of subtle social strategies such as alliances and tactical deception and the levels of social play. There is, for example, a linear relationship between relative neocortex volume and social group size in primates. In a nutshell, if a primate species wants to live in a larger social group in order to be able to solve its ecological problems more effectively, then it first has to evolve a sufficiently large neocortex to be able to support the computations needed to allow it to do so.

Humans seem to fit rather neatly onto the end of the primate distribution in this respect. The regression line for primates suggests that, with a neocortex the size that humans have, we should live in groups of about 150. Surprisingly, this turns out to be well supported by the evidence. Although humans can obviously cope with very large urban environments and even nation-states, the number of people within those large population units with whom one can say that one has a direct personal relationship is very much smaller. Censuses of the population units of hunter-gatherers, the size of scientific sub-disciplines, the number of people to whom one sends Christmas cards and the number of people of whom one can ask a favour all turn out to be about 150 in number.

There is, however, considerable evidence from the study of primates to suggest that what is important here is not simply the number of individuals whom you can recognise. Rather the problem is more concerned with the kinds of relationships you can have with other individuals. It seems that how well a relationship works depends on the level of trust that exists between the two individuals, and that this in turn is related to the frequency of their interaction. Frequent affiliative social interactions build a level of emotional bonding between two individuals that allows the relationship to be the basis for commitment to future social support. Among primates, individuals who groom together support each other in future conflicts with a third party. They are more likely to come to each other's aid when attacked by a predator or a member of another group. That much is also surely familiar from our everyday experience of humans: we do not easily rush to the aid of strangers - when we do, it is apt to arouse comment and, indeed, approval in a way that would surely not be necessary if such behaviour were a prominent natural feature of human nature. Indeed, we know from our studies of human social networks that the intensity of a relationship is related to the frequency of contact involved (or, at least, that has been involved in the past).

We do not at this point really understand why there appears to be limits to the number of individuals we can hold in a particular relationship to ourselves. It seems likely that these are cognitive limitations on the number of individuals that can be held in a relationship of a given degree of intensity. There is some longstanding evidence, for example, that the number of individuals we can have a particularly close bond with is limited to around 12-15, and that within this there may be an inner circle of about 5 individuals with whom this relationship is especially strong. There is, in addition, evidence to suggest that there may in fact be a series of layers, with boundaries at around 35 and 80-100, each associated with a declining level of emotional intensity and closeness. It is as though each of us sits in the centre of a series of expanding circles at 5, 15, 35, 80 and 150 individuals.
 

Cognitive Constraints on Network Size

The social brain hypothesis has important implications for several aspects of the way our social relationships are structured. The evidence shows that there is a simple linear relationship between neocortex volume and group size at various levels. We seem only to be able to hold a certain number of individuals in a given degree of intensity of relationship. The intensity of that relationship, in any given case, reflects a combination of factors in the past history of those two individuals.

One important component is the knowledge that they share about each other (the way they behave, their likes and dislikes), but also aspects of their shared social history (how they are related to each other in genetic or affinal terms, shared worldviews and culture, shared experiences). This shared knowledge base provides each with the basis off which to interact in an effective way. The wheels of the relationship are oiled by this knowledge. They know just how much they can trust each other, how far they can push each other both in teasing and in making demands on each other.

In addition to this knowledge base, however, it is clear that there is a more fundamental emotional stream to that relationship. In the end, of course, shared knowledge and shared experiences come from having spent a lot of time together. You cannot simply be told that X holds certain relationship to you and then naturally behave in certain way towards them. We can use such knowledge as a stepping-stone to establishing a working relationship with someone, but ultimately the effectiveness of that relationship - and in particular the level of trust that we are prepared to put on that individual - comes from interacting with them. The relationships that we hold dear are those that involve individuals we have spent a lot of time with. Building a relationship requires a lot of time. It requires time sharing activities that give us pleasure; it requires time spent mining the mental world of that individual, building up a virtual simulation of that individual in the mind's eye as it were.

This combination of understanding and trust built on time spent interacting lies at the base of primate societies. For primates, the glue that binds individuals together is social grooming. Although in origin a mechanism for removing debris and ectoparasites from fur, the amount of time that many monkeys and apes devote to grooming each other far exceeds that required for hygiene. A gorilla, with the largest area of body fur to keep clean, copes quite adequately in this respect with just 1-2% of its day devoted to social grooming. Yet some species, such as baboons and macaques, can spend up to 20% of their day engaged in this activity. It hardly seems plausible to suggest that macaques with a quarter of the body surface of a gorilla should require ten times as much grooming to achieve an acceptable level of cleanliness. In fact, grooming has been coopted by the social primates as a mechanism for social bonding. It works because it is particularly effective at releasing endorphins (the brain's natural painkillers). The sense of relaxation, euphoria and contentedness that this induces creates a bond between those who groom each other regularly. The consequence of that bond is that those individuals are more willing to support each other against common enemies than individuals that do not groom. It is this that provides the social glue that welds primate societies together.

We humans are no less bound by these mechanisms. We too engage in this kind of grooming in a very primate-like way. We lack most of the fur that provides the excuse for grooming in primates, but we engage in the activity every bit as much as they do. In our case, the actions are reduced to forms of patting, rubbing, hugging and touching that produce the same mildly euphoric effects. But such pseudo-grooming occurs only in the context of our more intimate relationships. Physical stimulation of this kind is regarded as unacceptable among less closely acquainted individuals. Yet, we need to engage in that bonding process with all those who come within our social circle. For humans, that level of bonding is serviced by language - or, more precisely, by conversation mediated by language.

Among nonhuman primates, there is simple more or less linear relationship between time devoted to grooming and social group size. This is principally because the bigger the group, the better one's friendships have to work to provide defences against harassment by other group members. Inevitably, of course, time is a limiting factor. For animals that have to earn their living in the real world, the amount of time that can be spared for social interaction is constrained by the amount of time it takes them to find and extract food from their environment. In the limit, primates cannot afford to spend more than 20% of their total day time in social interaction. This sets an upper limit of group size at about 80-100 individuals. Since human groups are 150 in size, some mechanism is needed to bridge the gap between what is possible to bond by grooming and the group size that the species requires to survive successfully.

Language seems to have evolved to meet that need. Studies of time budgets in a wide range of modern humans suggests that, on average, we spend about 20% of our time is social interaction (principally, of course, conversation). This is exactly the limiting value seen in nonhuman primates. But language allows us to use that time more efficiently. It allows us to interact with more individuals at once (we can talk to several individuals simultaneously, but grooming - even in the form that we humans have it - is very much a one-on-one activity. Language also allows us to seek and exchange information about our social network. A nonhuman primate's knowledge about the state of its social network depends entirely on what it has seen for itself; in contrast, we can find out about what has happened in our absence, thereby allowing ourselves to keep track of the ever-changing social world even when we are not there in person.

However, language lacks one key feature that grooming provides - the opioid kick that seems to play so essential a role in social bonding. That gap appears to be filled by laughter. Laughter has the ability to produce the same kinds of opioid effects as grooming does: we feel relaxed, euphoric, at peace with the world and well disposed towards those with whom we laugh. It appears to play a very fundamental role in human social interaction.

These findings would seem to have a number of implications for the structure of social networks even in modern post-industrial environments. One is that no matter how good the technology, electronic media that are used to service individual's social networks need never be significantly larger than this value. Even the most social individual is unlikely to be contacting more than about 200 other individuals at any one time. A second implication is that the kinds of relationships we have are severely constrained. The fact that we can contact very large numbers of people does not mean that we can get to know them better. Our expanding circles of relationships remain where they have always been, and continue to be dependent on the same old-fashioned kinds of emotional intensity that can, in the final analysis, only be generated by direct intimate contact.

One obvious difference that electronic technology can make, however, is to allow us to extend the circle of relationships over a much greater geographical area than has previously been possible. The intimacy and intensity of a relationship needs to be constantly serviced - reinforced by frequent interaction at the appropriate level. Failure to reinforce the relationship causes it to fade away. The individual that once occupied one of the boxes in our circle disappears and is replaced by a new face whom we see more often, perhaps someone who occupied a box in a more distant circle of acquaintanceship, perhaps someone who has newly arrived in our social world. That constant turnover of identities was, in the past at least, terribly dependent on the physical proximity of the players. Someone who moved elsewhere faded gradually unless they were emotionally especially important. Letter-writing helped to slow the rate of fading, but could not prevent it altogether because letters travelled slowly and replies often arrived back months, perhaps years, after the initial letter had been sent.

Modern telecommunications technology allows that exchange to happen faster, so the rate with which old friends fade in our memories may be reduced if we so wish. But still, it cannot abolish the process altogether. The reason friends cease to be inmates and become mere acquaintances with separation is that we are unable to update our mental picture of them. They change with time, and we change with time, and gradually our old similarities, our once shared knowledge of each other and the social world in which we live, diverge too far to allow us to pick up the pieces when we meet again. We become strangers who need to start the process of getting to know each other all over again.

One example of how the design of our minds can impose limitations on the size of our social groups comes from the rather mundane example of conversations. In principle, we can get very large numbers of individuals into a conversation, as evidenced by what we manage to achieve during lectures or sermons. However, to do that, we have to impose very strict social rules on those present. Everyone must agree to remain silent and listen to the single speaker. Indeed, everyone must agree to give up the opportunity to intervene or ask questions. When an opportunity to make comments is provided - as at the end of a lecture - the process has to be stage-managed very carefully by a chairman who dictates who can ask questions and how long they can hold the floor for. If we do not act in this way, chaos ensues very quickly.

The reason for this is very simple. It seems that there is a natural limit to the number of people who can be involved in a free-flowing conversation, and this limit is set at four individuals - one speaker and three listeners. If more than four people join a conversation group, the group will split into two or more separate conversations within at most a minute. This effect is very robust and can be observed during any cocktail party or similar social gathering. There are two likely reasons for this. One is strictly psycho-acoustical: our ability to discriminate speech sounds decreases rapidly as the distance between speaker and hearer increases, and, once the group exceeds four individuals, the distance across the circle becomes too great. The other is social: since we only ever allow one person to hold the floor at a time during conversations (otherwise chaos intervenes), each individual's opportunity to contribute to a conversation declines rapidly with group size. With five people in a group, each can spend at most 20% of the time talking if everyone is to get an equal turn. Very quickly the group ceases to become an interaction and instead becomes a lecture. This constraint probably sets the limit on the number of people that can be involved in a multiway communication system.

At the same time, it is clear that modern telecommunications has unforeseen dangers for us. These are already familiar phenomena - flaming (the tendency to over-react on email exchanges), stranger danger (our tendency to respond too intimately on emails). Most of these seem to arise because electronic communications currently lack the immediacy of face-to-face interaction. In everyday life, nonverbal signals play a crucially important role in allowing us to interpret the meaning and intentions of those with whom we interact. In email - and even telephone - exchanges, we lack these cues and thus fail to interpret messages correctly. In face to face encounters, we make tentative forays into intimacy and watch very carefully to see how they are received; the feedback we get encourages or inhibits what we do next. The absence of these crucial cues in electronic communication (combined by the slowness of the response in email) encourages us to go one step further than we would normally do. Premature declarations of eternal love are notorious in internet chat rooms for this reason. At the same time, we have few cues as to the honesty of the person we are interacting with. In real life, facial cues play a very prominent role in our assessment of individuals' suitability as partners or friends. When we lack those cues, we are apt to make assumptions about a person's honesty that aren't justified.
 

The Freerider Problem

I have suggested that relationships are based in large part on trust. Let me turn now to one of the central issues that underpins this, namely the problem of freeriders. Primate societies (and that includes human societies) are essentially cooperative solutions to the problems of survival and reproduction. In effect, individuals accept an implicit bargain whereby they forego some of their immediate objectives in order to obtain a greater longterm benefit. The problem that all such social systems face is that they are highly susceptible to freeriders - individuals who take the benefits that do not pay all the costs. This problem is ubiquitous in human societies. We find it in those who don't quite pay all the taxes they should, who park in the no parking zones (because it saves them a few moments of time); we find it on the grand scale in the so-called "Tragedy of the Commons", the over-fished oceans and the denuded landscapes once clothed in tropical forest. The problem is that, whenever there is a social bargain of this kind, it always some individuals to cheat - to allow others to bear the full cost of the communal benefit.

Human societies are especially susceptible to these kinds of problems because of their size. As social group size increases, so freeriders have more places to hide, more naïve individuals willing to allow them to join even after their anti-social behaviour has been uncovered elsewhere. Part of the problem here is the fact that freeriders can move faster than the information about their misdemeanours. They can always stay one step ahead of discovery.

Because the freerider problem is so intrusive in human societies, we have evolved a number of mechanisms for dealing with it. These include a particular sensitivity to social cheats, the use of language to exchange information about freeriders, verbal reprimands to make backsliders toe the social line and the use of dialects to identify members of our communities.

Language provides us with several important mechanisms whereby we can keep track of and police freeriders. One is the simple fact that we can use it to exchange information about what's been happening within our social circle while we have been engaged elsewhere. In one sense, this was probably the original function for which language evolved - a mechanism for monitoring the constantly changing state of a very dynamic system. Even today, 60-70% of conversation time in natural situations is devoted to social topics ("gossip" in its broadest sense). This is a phenomenal investment in time. Given that we devote 20% of our waking day to conversation, this means that we spend something like 15% of the waking day discussing the state of our social network. That's a very significant investment of our time. But ensuring that we know who is "in" and who is now "out", who has formed a partnership with who, and who has now broken up with who is crucial if we are to be able to negotiate our way through the complexities of everyday social life.

But language allows us to do more than that. It allows us to police the behaviour of our friends and colleagues - not just to catch up with their social misdemeanours but also to remonstrate with them when they offend against the communal code. There is nice experimental evidence of this from studies of cooperative games. In computerised investment markets where cooperation leads to significantly higher returns but lack of cooperation leads to very poor returns, the temptation to cheat is overwhelming: the problem is that if you don't cheat, you run a very high risk of doing very badly because someone else has cheated. So, in experiments in which individuals can observe the state of the market but not communicate directly with each other or see which individual is cheating, payouts can be as low as 20% of what full cooperation would yield, even when it is quite clear to everyone that cooperation would be a better strategy. However, allowing subjects to take a coffee break together mid way through an experiment can result in a three- to four-fold increase in payouts, mainly because it provides the group with an opportunity to remonstrate with those who fail to adhere to the cooperative strategy. The effect works even if the subjects do not know which individual is responsible, but merely state their frustrations and irritations with whoever is responsible. Allowing the group to impose sanctions on the still anonymous freeriders improves payouts even further.

One other interesting feature of human languages is the speed with which they diversify to produce, first, local dialects and, later, distinct languages that are so different as to be mutually incomprehensible. We have modelled this issue as a problem in social badging. In other words, we suppose that dialects are really designed to allow you to identify people that come from your own very small local community, and in particular the community that you grew up in. The implication of the last point is that you share with them a significant level of genetic relatedness. Relatedness is important in the context of freeriders because it identifies individuals with whom you share sufficient common genetic interest that both of you will be willing to adhere to the communal agreement. More importantly, perhaps, genetic relatedness has one important consequence: it makes it less problematic is your partner cheats on the system because you will still benefit from their good fortune because you have a stake in that person's genetic success. This is one reason why alliances and agreements with relatives tend to be more open-ended and less demanding of strict reciprocation than similar contracts with unrelated individuals.
 

Innovative Exploitation

Finally, let me turn to the first of the two consequences of the social brain that I mentioned at the start of this talk. We humans are perhaps the most inventive species there has ever been, at least on Planet Earth. Our very survival as a species, our ability to colonise every continent with all their different habitats, even our less laudable ability to drive all other species to extinction - all these have been the consequence of our inventive minds. Religion, science and enterprise all share a common origin in our remarkable capacity to visualise new opportunities.

The example I want to offer here concerns uses to which electronic technology might be put which have absolutely no relationship with their intended function. The technology in question is cell phones, the use is as a mate-attraction device.

The origins of this particular study go back to some discussions I was having with Hewlett Packard's research staff while involved in a small project with them. They had noticed that teenage males (in particular) were fascinated with handsets during focus group discussions. They wanted to have them, even though they were in fact dummies. At the same time, we noticed a newspaper report from South America which said that one of the chic things to do on the dance floors of local clubs was to hold your mobile in your hand. However, the number of calls on these had become so irritating to everyone that some clubs had started to insist that clients hand in their mobiles at the door, rather in the way gunslingers were required to hand over their guns in the old Wild West saloons. The puzzle was that many people didn't bother to pick their phones up at the end of the evening: sometimes as many as a third of them were left behind on the conciege's shelves. Eventually, the clubs had to do something about them, and at that point they discovered that many of the phones were in fact dummies. The cell phone as fashion accessory had arrived.

These observations reminded me of something that I had noticed, but not paid much attention to. I spent a lot of time travelling on trains at that time, and I had become subconsciously aware of something odd: men almost always put their cell phones on the table beside them - often after taking them out of their briefcase or a pocket - but women didn't. Women clearly had phones, because every so often one would ring and the owner would search for it in her handbag or briefcase, take the call, and then carefully put it back again.

What could be going on, we wondered?

One obvious possibility suggested by this contrast in the behaviour of the two sexes is that it has something to do with sexual advertising. Men were using cell phones to display something that women were not especially interested in displaying for themselves. The obvious answer is that phones are cues of mate quality. In the study of human mate choice, it has become clear that a man's status or wealth (the two are often much the same thing) are an important consideration in women's evaluation of potential mates. It is not, of course, the only consideration, but it remains an important one nonetheless. A man's ability to provide the resources needed for successfully rearing children is just as important in modern society as it is in traditional societies. In more traditional societies, it was a man's ability to provide the land for agriculture or the products of his labours that made it possible for a woman to successfully feed her children. The only difference today is that the emphasis has now shifted into how wealth affects children's education and placement in a technological society. Rearing children successfully is now even more expensive than it has ever been, because there is so much more available both in terms of opportunity (careers in computing or communications that requires expensive training and/or equipment) and in terms of display (having the right kinds of trainers or clothes with the right labels).

To explore this in more detail, we set up a small study of what people do with their cell phones. We chose a pub in the centre of Liverpool, close to the Law Courts, where the clientele was likely to be up-market (lawyers, business executives) and then sampled various aspects of the behaviour of those who came into the pub during the early evening immediately after work (5-7pm) on 23 separate days over a period of several months. Thirteen tables that were clearly visible from an upstairs balcony were sampled at regularly intervals, recording the number of cell phones on view, the frequencies with which phones were used (both to answer calls as by being toyed with) and the number of sex of the people at each table.

As we had anticipated, men were significantly more likely to have their cell phones in view (for example, on the table) than women were, even when we take into account the fact that men owned slightly more cell phones than women did. (At the time of the study, around 59% of registered owners with one major UK provider were men.) More interestingly, however, men were more likely to expose their cell phones as both the number of men in the group round the table increased and as the sex ratio increased (more men per woman in the group). When there were more than four other men in the group, around 80-90% of all cell phones known to be owned by the men in the group were on display, compared to around 50-60% when there were four or fewer men present. More intriguingly, the proportion of known phones on view increased more or less linearly as the number of males per female at the table increased, from about 10% when the sex ratio was equal to around 80% when the sex ratio reached three men for every woman.

It seemed that, at least at the time of the study, men were using cell phones as signals of quality. The level of competition they faced from rivals over access to women apparently spurred them on to ever greater efforts of display. In this respect, cell phones were being used in exactly the same way that Rolex watches and Armani suits are used as signals of wealth. The message is: that I can afford to splash out on such a relatively useless item shows how wealthy I am. The important thing about such signals is that they are expensive and therefore rare. A general principle in evolutionary biology known as Zahavi's Handicap Principle reminds us that honest signals are costly: if cell phones (or Rolex watches) were cheap and abundant, they would lose their value as a signal.

We have not revisited this issue since we completed our study, but now that cell phones are both cheap and commonplace in the UK (their price was just starting to fall as we did our study), we can predict that their signal value will have been eroded and men no longer use them so conspicuously. Instead, they will have been replaced by another newer piece of technology or fashion accessory that carries the same message because it is new, expensive and rare. WAP phones or digital cameras are obvious examples, perhaps. This looks suspiciously like the phenomenon known in evolutionary biology as the Red Queen Effect: like the Red Queen in the children's book Alice in Wonderland, organisms have to keep running (in terms of changing their appearance) in order to stand still in the game of life. Animals that fail to do so are swept backwards to extinction by the tide of competition. Similarly, we humans are forced to seek for constant novelty in our displays in order to stay one step ahead of the opposition in the mating game.
 

Conclusions

I have tried to show how and why our ability to exploit the new technology of the e-world may ultimately be constrained by some very basic features of the way the human mind has evolved. Our task, in one sense, is not to try to exploit this technology to do the same tasks as we do with conversation but on a bigger scale, but to find innovative ways to exploit that technology to do things we cannot do naturally. I offered one example of this kind of unexpected use - the use of cell phones in mate advertising. Our task in the future is, perhaps, to understand how the human mind limits natural interaction, and to find ways to circumvent those limitations in ways that do not contravene the basic principles on which the human mind is built.