This essay originally appeared in ‘Knowledge and Information – Perspectives from Engelsberg Seminar, 2018’, Bokförlaget Stolpe, in collaboration with the Axel and Margaret Ax:son Johnson Foundation.

The evolution of human language was driven by the need for co-operation. In each child, language development springs from the urge to develop the individual’s social footprint. Mastering language is to master a rules-based system that reflects the mind and is built on a mixture of explicit and implicit rules plus a myriad of exceptions. This is computationally a great challenge and it is a wonder that the slow machinery in the brain can handle language in real time. Speaking the same language contributes substantially to the capacity for social co-operation and thereby to the contextually bound social capital. As our social context is moved to the virtual level, local social negotiations decrease in importance. We therefore run the risk that the socialisation process, the process that shapes the responsible, predictable and co-operating adult, will contain far less friction than hitherto and therefore not prepare the individual for adult life.

Our individual social footprint is based on communication where language is the key. Our rules for social interaction are nature’s own illustration of game theory, ie. the modelling of conflict and co-operation between intelligent, seemingly rational decision-makers. Social interaction is a complex dynamic process with multiple factors that influence the quality and quantity of the phenomenon. The thesis that there is a unitary rationality based on self-interest underlying social interaction is flawed: several experimental studies have shown deviation from such rationality, expressed in a preference for long-term stability and trust over maximal immediate reward. This lack of single-dimension egotism has been demonstrated in the Ultimatum Game. In that game, a bidder is given a fixed one-off amount of money and instructed to share a portion of it, the sum to be determined by the bidder, with the other player. If the bidder’soffer is rejected, neither the bidder nor the other player gets to keep what is proposed, but if the bid is accepted both keep their respective cut. If the proposed split is 50:50, the rejection rate from the second player is generally 0 per cent, whereas the rejection rate tends to increase when the bid-fraction dips below 30 per cent. It should be noted that to reject a bid is costly for the second player, and since the game is a one-off game only, the short and long-term rational response is to always accept the offer — €2 is after all always more than €0, even if the bidder does go away with €98. Thus, there seems to be a general willingness to punish others at one’s own expense in order to maintain and develop stable social norms on reciprocity.

Many psychological experiments have demonstrated the confabulatory tendency in seeking to explain one’s own irrational behaviour as rational. We demonstrated that it was possible to pharmacologically manipulate and increase the acceptance rate to unfair offers in spite of the fact that the description of the rating of the experienced unfairness was the same. Thus, the immediate ‘fast’ reaction (delineated under System 1 in Daniel Kahneman’s Thinking, Fast and Slow) seems in this case isolated from the slow thinking (under Kahneman’s System 2) that language tends to reflect. The rapid rejection response arguably reflects the evolutionary value in upholding norms of reciprocity in the social context, especially when resources are scarce.

This ‘hierarchy of needs‘ — as defined by the American psychologist Abraham Maslow — reflects the cognitive organisation of the brain in the order of its evolution. Immediate responses to strong and emotionally anchored stimuli stem from the phylogenetically older regions (the limbic system), whereas language and other higher order functions are more related to the upper levels in the hierarchy. When there is a threat to ‘basic needs’ (including a threat to social norms), corrections are prioritised above and beyond all others. Studies have clearly demonstrated that social stability and the upholding of norms for social interaction also belong to the area of basic needs as they were anchored in phylogenetically ancient brain systems that maintain essential survival mechanisms. Thus, reciprocity as a basis for survival is anchored not only in higher order moral concepts, but also in early evolutionary mechanisms, with representation apparent in the older anatomical regions of the brain. This conclusion predicts that the upholding of norms on reciprocity can be found also in social co-operating animals.

The Ultimatum Game yields similar results across the world, including in pre-literate cultures. In such cultures, all norms are upheld locally and are of core importance for the survival of the group. Several of these described norms, for example sharing, are very similar between different cultures almost to the point of being universals. In his book, The Blank Slate, Steven Pinker effectively argued that innate mechanisms lay behind such universals and hypothesised that human behaviour is substantially shaped by evolutionary psychological adaptations. The alternative concept of the newborn baby’s mind as a ‘blank slate’ would pose too great a risk for survival, implying that the principles for human interaction would have to be worked out for each relation. Language is part of the transfer of knowledge and behaviour within social groups and also across generational boundaries, as Pinker posited in The Language Instinct. It also shows structural similarities right across the world. Pinker argues that language acquisition can almost be equated with an instinct, driven by an individual’s ability to control and understand his or her immediate environment. A further argument is that children spontaneously invent a consistent speech even if they grow up among a mixed-culture population speaking an informal mix of languages with no consistent rules. These signs suggest that, rather than being a cultural invention, language is an innate human ability. Pinker thereby supports Noam Chomsky’s concept of a universal grammar, a meta-grammar into which all human languages fit. A universal grammar is instantiated in specific structures in the human brain that recognise the general rules of other people’s speech, such as whether the local language places adjectives before or after nouns, and begin a specialised and very rapid learning process not explicable as reasoning from first principles or pure logic. This learning machinery is most effective during a specific critical period of childhood and later loses its capacity. The brain’s social systems drive this development, and when the social drive is low — as it is for instance among individuals with an autism spectrum disorder — a marked slowing of language development is often observed. An important conclusion is that language and human interaction are intertwined and have a powerful role in shaping an individual’s social circumstances. The decreased plasticity that comes with age that is observed in the language system is true also for other brain systems including the social norm system.

The quantum leap of the Gutenberg printing revolution generated a huge increase in the language-based spread of cultural patterns. The pattern of communication was maintained in that only a fraction of society produced the material for the masses. General knowledge was made accessible by means of books, and the development of cultural values was firmly established by that technological development. Technology has evolved, and half a millennium after Gutenberg the structure of information availability has changed radically. The last thirty years have empowered everybody to both publish and access information about everything. Nowadays, with ubiquitous computing power and much of the global population being online, more or less continuously, what had been a scarcity of information has turned into an overflow of information. This overflow is having a radical impact on our social and cultural patterns as well as posing challenges to our cognition and health. The important effects of the change in the availability of information about every thing on individuals and on their cultural patterns are ever more apparent.

The stability of cultures, including social patterns for co-operation, is in the midst of an upheaval as a result of the extension of the social foot-print of each individual. The need to nurture one’s immediate surroundings by acting as a socially polished agent becomes less important as the self-reported narrative only has to resemble the truth in order to make new social connections. A decade after the introduction of mass tools for the extension of an individual’s social reach, the first trans-generational effects are emerging. Pinker’s argument that peer-to-peer influence out-weighs parental influence in the shaping of the individual has been amplified in the era of social media. Parental influence is in decline, in spite of the fact that children are leaving home at a later age than they once did.

The social arena has changed. Howard Rheingold’s seminal book of 2007 on the social impact of modern technology — Smart Mobs: the next social revolution — predicted the social effects of breaking up hierarchical information structures within a generation. He defined a ‘smart mob’ as a group whose coordination and communication abilities have been empowered by digital communication technologies. Smart mobs are particularly known for their ability to mobilise quickly and without the use of pre-internet democratic structures. Essentially such mobs could be described as democracy on steroids and be perceived as high impact destabilisers of established hierarchical models for power distribution.

Nearly all of the predictions in Rheingold’s book have materialised, which is remarkable as they were essentially made before the impact of Facebook, Twitter and Instagram was evident and the only mass-market technologies that existed were internet web pages and SMS text messaging. With the advent of ubiquitous computing and multi-peer chatrooms, the local social realm has lost its boundaries and thereby also important social control mechanisms.

*

How do you build a reputation in order to become a successful individual when social interaction has moved online? It is virtually impossible to verify any statement or effectively retaliate against the breaking of norms when, at worst, such transgression leads to the blocking of communication channels with an individual? The ability to shape the norms for co-operation is rather limited for an individual. The online use of very strong language, that would be unacceptable in other contexts, involving racism, sexually explicit language, brutal accusations and blatant lies, is a direct effect of the absence of immediate reciprocity across the web. Growing up in and being exposed to such a social landscape will clearly affect the next generation.

An important concept that Rheingold introduced was the ability to co-operate over long distances and without any social connection beyond what the internet enables. This makes the mob less stable when it is evolving and also renders it prone to manipulation as negotiation for reciprocity is diminished. Hence, such mobs are expected to be short-lived and will often fade quickly away.

Trust is the hard currency in any social relation and it is built on sameness, reputation and experience, as Francis Fukuyama suggested in Trust: the social virtues and the creation of prosperity. Any evolving social relation is based on continuous evaluation of the trust at the heart of it. Moving from local to remote relation-building makes this evaluation far more difficult. In spite of this, explicit negotiations in combination with third parties that guarantee the identity of the negotiators mean that large-scale social and economic interactions between individuals who didn’t know each other previously continue to develop. People share their houses and cars with complete strangers purely on the reputation of the internet-persona they have met online and whom they know only from that stranger’s self-presentation. This is in line with Robert Putnam’s concept of how to build and maintain social capital, as outlined in his book, Bowling Alone: the collapse and revival of American community. Putnam’s concept has three components: moral obligations and norms, social values (especially trust) and social networks (especially voluntary associations). The sharing economy seems to follow suit, with the introduction of explicit rules for obligations, communicated values and recommended behaviours — and concluding with the results of each transaction being communicated in the form of a two-way review. Thus, even reputation-building is made explicit in the digital arena.

Putnam’s central (pre-internet) thesis is that if a region has a well-functioning economic system and a high level of political integration, these are the result of that region’s successful accumulation of social capital. As social networks move to the internet, the dimension of locality diminishes in importance and it becomes difficult to tie social capital to any one place. As local reputation-building becomes less important, altruistic activities such as caring for the local community or participating voluntarily in local democracy are at risk of receiving less interest.

*

Much has happened since Rheingold’s book was published in 2007. The effects of the broken information monopolies have eroded institutions, radically changed professions and influenced politics. The president of the most powerful nation in the world seeks daily social approval via Twitter. With dominant parts of the population being online constantly, social patterns for learning and knowledge-finding have changed. The world is now available at the touch of a finger. The health effects of constant availability and self-propelled demand for split attention have started to take their toll. It is now possible for an individual to display their entire social persona to the world around the clock, while the separation of work time from private life is no longer possible for many. This taxes several cognitive processes simultaneously. The most vulnerable cognitive functions are the first to suffer. It has become a disease with many names and with common symptoms. Practically all cognitive functions are affected.

First and foremost in the line of fire is the ability to focus — and to maintain that focus and concentration over time. A simple thought that leads to a conclusion and insight takes a matter of several minutes, especially if it also includes information gathering by reading. Constant interruptions lower cognitive efficacy quite markedly. This in turn affects learning, specifically the consolidation of factual and procedural memories. Our brain is comparable to our social world — a complex dynamic system in constant change, based on its interaction with the environment. Today, most people do not allow themselves to be ‘undisturbed by themselves’ and a substantial fraction of the population is clearly feeling the health effects of this. The short-term reward mechanisms provided by the social mechanisms in the brain challenge the focus and endurance (‘grit’ or ‘resilience’) needed to conquer the unknown. The decline of academic performance in schools in the industrial world can most probably be attributed, at least partly, to this emerging phenomenon.

Everybody has experienced the effects of too much information and many realise that the rationing of social media or mobile phone use would increase focus and productivity, yet most tend not to do so. The key explanation is the manipulation of relevance that is built into the information. Tabloid newspapers have used this method for years. Choosing to use a word like ‘fuming’ instead of ‘angry’, or ‘dying’ instead of ‘ill’, is designed to grab attention and increase the likelihood that the content of the publication will be consumed. We are just as likely to look for the good stuff, because it is cute, as we are to look for the gory stuff in order to experience the relief of not being affected by it. Both yield the hard currency of reward — dopamine and endorphins. Social media has an extra advantage in terms of grabbing our attention because, as a result of evolutionary mechanisms, we find it unbearable to be excluded from a social community and will do practically anything to be part of the in-group. Conversely, participation in social communities drives the reward system. In that sense social media suffocate an important evolutionary system for the building of stable social connections by smothering it with pro-forma shadow connections with shallow reach.

*

The ability of software to communicate based on the ability to collect and classify knowledge, learn from it and recode itself, is growing quickly and challenges the idea that the human mind contains the most advanced intelligence on our planet. Threatening that pecking order today are chatbots which use language and reasoning in ways that make them indistinguishable from human beings. Another invention that challenges our trust is ‘deepfake’ technology, so far confined mostly to movies, by which an apparently ‘real’ actor is merely a composite fake. The scary thing here is that certain media which could previously be seen as broadly speaking trustworthy may nowadays be completely false, thereby inserting distrust into yet another channel of communication.

Adaptive computing, with endless processing power, will be able to track the communication habits and preferences of any individual. When logging on to Facebook, for instance, a user’s entire digital profile — all behaviours, preferences, patterns of search and contacts — is instantly identified. The profiling is extended if a mobile phone number is supplied, giving away travel patterns, social connections, social and commercial habits. The adaptive profiling of individuals is made possible only by the value that an individual perceives it to have, and the hard currency that is negotiated is the enlargement of the individual’s social footprint. The same mechanism is hidden in Gmail, Twitter, Instagram, Tumblr and LinkedIn, to name just a few. Make no mistake, information leaks between these applications as well. Most mass-market technological advances that have spread rapidly are more or less based on this offer of a larger social footprint — be it the telephone, the fax machine, SMS, WhatsApp or social media in general. The mass application of standard deep (machine) learning on social data makes it possible that, in the very near future, individually assigned ‘omnibots’ will be constructed with the capacity to serve every individual for the ‘small’ cost of leaving all information about yourself — including all those personal quirks and deviances — to transnational data-collecting giants.

*

Four trends arising from the revolution in global society’s information structures will affect the social and cognitive systems in our brains. All derive from the change of social structures around an individual as technology has evolved. The drivers for these developments are generally based on the evolutionary mechanisms that advance social interaction. The four trends are:

1. A decrease of local social capital in favour of boundless peer-to-peer networking in social media.

2. Stress from the difficulties of determining reciprocity in social connections. Stress-related disease and behavioural influence from the sheer overload of information.

3. Very different adaptive formation of the social mechanisms in the brain in the next generation as training in reciprocity will take place in circumstances where the ability to form lasting social norms based on reciprocity will be less evident than before.

4.The now emerging AI-based adaptive tools will flood our brains with even more information. Such tools will learn all about the individual and have an almost endless power to manipulate on a grand scale.

We need a clearer characterisation of these phenomena in order to understand the societal rules around them. Otherwise, the dystopic tendencies that have been touched upon in this chapter may self-amplify.