We are all a collective brain, so what's new?
Cesar Hidalgo 1 Contributor
March 15, 20165 Versions
Not Featured in any Journals


@article{Weareallacol2016, title={We are all a collective brain, so what's new?}, author={Cesar Hidalgo}, year={2016}, note={version: 57a22b715a4037b8e01a2a95}, publisher={PubPub}, }


Cesar Hidalgo. (2016). We are all a collective brain, so what's new?. PubPub, [https://www.pubpub.org/pub/networked-intelligence] version: 57a22b715a4037b8e01a2a95


Cesar Hidalgo. "We are all a collective brain, so what's new?". PubPub, (2016). [https://www.pubpub.org/pub/networked-intelligence] version: 57a22b715a4037b8e01a2a95


Cesar Hidalgo. "We are all a collective brain, so what's new?". PubPub, (2016). [https://www.pubpub.org/pub/networked-intelligence] version: 57a22b715a4037b8e01a2a95
There is this meme or idea that is prevalent in society, which is that we are all a collective brain. We can call this collective intelligence, decentralized mind, global computer, extended intelligence, socially augmented cognition, global nervous system, or global brain, but no matter what name we use, we will be speaking basically about the same thing. But this idea is not new. In fact, it is an idea that can be traced back to several branches of literature, many of which do not deal with technology. On the one hand, the idea of a collective mind can be seen as an extension of the idea of the market, which is our decentralized ability to coordinate productive activities among strangers. On the other hand, the idea of collective minds can be traced back to the vestigial social structures and institutions that anthropologists study. The social structures that created the social room that allow markets to exist.
These vestigial social structures matter because collective intelligence does not happen in a vacuum, where self-interested individuals spontaneously generate cooperation. Collective or extended intelligence originated thousands of years ago in pockets, where individuals who shared genes and developed traditions evolved the rules they needed to cooperate. These vestigial institutions, the ones that created room for markets to exist, are well explained in David Graeber's 2011 book: Debt:The last 5,000 years [1]. In Debt Graeber, an anthropologist and professor at LSE, explains that markets and currencies did not emerge in the context of strangers looking to barter different goods (as in an Adam Smith proverbial village). What the anthropological record show is that, prior to having physical currencies, people had concepts of virtual currencies and debt, since they lived in small social groups where tabs did not need to be settled simultaneously. In these small tribal societies, people kept mental tallies of who owned something to whom, and where relationships would eventually balance over time. In these societies, Graeber argues, the double coincidence problem of bartering that economic textbooks use to motivate currency did not exist. Therefore, currency did not emerge as a solution to the double coincidence problem, but as an extension of the social debts that originate in small groups. In fact, what is observed in pre-modern societies are people without currencies, but with the concept of debt. It is this indebtedness to one another, and the social mechanisms that enforce the settling of these tabs over extended period of times, what provides the social glue that these groups need to exist, and of course, to develop their collective intelligence.
But eventually we transcend the scale of hunter-gatherer groups. This is where technology, such as physical currency, comes into play. The way that physical currency was introduced in many places was—once again, not to solve a double coincidence problem—but as a government innovation designed to resolve the logistic problems needed to support an army. The mechanism by which governments introduce currency to finance armies is quite simple, clever, and elegant. The idea is that early governments (those existing in a hostile, fragmented, pre-modern world) survived only if they could organize an army to protect or expand their borders. But providing food, weapons, and shelter for armies is expensive. So what governments began doing, instead of providing those things to their armies directly, was to give soldiers money to buy food, shelter, or even weapons. But citizens did not want money (they had no use for it because they had other ways of settling debts). So what governments had to do was to tax people (in many cases using the army to enforce the collection taxes). Now people needed money to pay taxes. Boom! By taxing people on the one end, and giving money to the army on the other, circulation of physical currency was turned on. In turn, physical money increased our ability to interact with strangers (and develop larger collective brains), but it did not historically emerge from that necessity.
But there are also other mechanisms that have helped create the pockets of humans where our collective intelligence can reside. According to Curtis W. Marean, an archaeologist at Arizona State, a key moment in the evolution of our species was triggered by inhomogeneities in the productivity of food patches, such as fishing holes in Africa, tens of thousands of years ago [2]. The theory (but there is also evidence) is that great variations in the productivity of shellfish made some locations extremely attractive to early human societies. In the best places, which were few and scattered, humans could extract well over their caloric need in a single day with little effort, and therefore, could grow and thrive if they were able to secure access to these places. But since many groups competed for these places, the inhomogenity of food patches created a strong selection pressure for groups with individuals with strong feelings of group loyalty to win the best patches (as they would work coordinatedly to protect them). So our motivations to act collectively emerged in the context of selection pressure caused by inhomogenous resources. Here, it was still the environment what provided us with the triggers needed to develop the feelings and institutions that allowed us to extend our intelligence beyond individuals and to act as groups, but by creating the idea of a group, humans began putting together their brains.
Now if we fast-forward dozens of millennia to the twentieth century, we can find ideas of collective minds in Hayek's famous use of knowledge in society [5]. This 1945 paper—which has had huge political implications that I will not discuss here—formalized the notion of market prices as an information revealing mechanism. This is the idea that what prices do is help reveal information about the supply and demand of goods, and therefore, the main role of markets is to process information in a decentralized manner (achieving a computational capacity that is beyond that of central planning efforts). In the case of Hayek, however, we are talking about a society that is much different from the ones described by Graeber: a society with global communication and transportation technologies, scalable manufacturing, and a highly developed trade and financial system. In this society, prices are the highly compressed pieces of information that summarize the information needed for large groups of strangers to coordinate their wants and needs. A collective computer powered by a very tenuous but pervasive stream of price data. Once again, the collective computer idea comes to mind.
But the previous literature on collective intelligence does not only describe the mechanisms that gave rise to the coordination of groups, but also, focus on the pairwise interactions that help us create the links by which we extend our intelligence. Take the work of the cultural anthropologists who have studied the biases determining who we are more likely to learn from. The idea here is that, since there are too many people in the world, choosing who we learn from is important [3]. And social learning is crucial for humans because our ability to adapt to an environment depends on it. So whom do we learn from? Dozens of experiments with children have shown that people are naturally biased to learn from those who show skill, accomplishment, and prestige, or that are similar to them in age, gender, and ethnicity. This rules help determine how information is passed on from generation to generation, and what information we consider valuable. While in principle, the veracity or value of a piece of information should be independent of who is providing it, in reality, it is not. That's why Einstein quotes, even when they are fake Internet memes, sound like life lessons. Of course, valuable information does often come from those who are skilled, accomplished, and prestigious (like Einstein), but the opposite is not always true (not all information provided by those who are skilled, accomplished, or prestigious is valuable). Like all natural mechanisms, the ones from social learning are also ones that backfire.
But sometimes these pairwise interactions are aided by something other than mutual feelings of indebtedness, social pressure, currency, or social learning biases. They are aided by a social-glue that has a component of delusion. As the Israeli historian Yuval Harari argued in Sapiens, our ability to coordinate activities with strangers is facilitated when we share the same fictions. When we believe in the same gods, currency, or legal persons (technically companies), than other people. Sharing delusions allow us to agree on values, procedures, goals, and moral judgments. In fact, the "gluing" power of these shared fictions is such, that much of our history has been dedicated to hone these fictions. I am pretty sure that there isn't a book with more versions than the book of the Abrahamic religions (the book that has been edited and appended to create the Torah, the Bible, and the Koran).
So why do we need to develop fictions, create mental accounts, and decide who we should learn from? The reason is simple. It is because the computational capacities of individuals are finite, and the only way in which we can accumulate increasing computational capacity is by processing information collectively. To become a collective brain, we need to first become collective.
If you are reading this, is because you descend from a long lineage of individuals who transcended their personal level limitations by developing the institutions and technologies they needed to work collectively (those who did not, were selected out). In a nutshell, this is basically the point of Why Information Grows [4], a book where I explain that the capacities of economies to process information are constrained by two things: the social hardware we need to create to accumulate computational capacity (our social networks), and the knowledge and knowhow that we accumulate through cultural learning process and that allow us to build what we imagine.
But the idea that our ability to compute collectively depends on the size of the networks we can form, and their ability to accumulate knowledge, is also supported by observations. Consider the disappearance of the knowledge and knowhow that Tasmanians used to construct tools and weapons once the land bridge between Tasmania and Australia disappeared. When the population became small and isolated, the knowledge they had diminished. Another example is the strong correlation between the diversification and sophistication of a country's export mix—its economic complexity—and the mix of products that a country will make next.
So we can say that the idea of a collective brain, an extended intelligence, a human hive, a decentralized computer, a universal mind, or a global Village, is nothing new. It has been around for long, and it will be around for long.
So what is new this time around? What might be different this time is the technology. Is the opportunities we now have to download computation to devices and outsource tasks of sensing, classifying, and driving, that we could not outsource to machine intermediaries before. What is different is the potential we have to coordinate, gather, and analyze data in ways that could complement, expand, and extend the coordinating capacities of the price system, a reliable but tenuous and imperfect channel that still mediates much global coordination. Yet figuring out how to connect the world's data, and how to create the algorithms needed for that data to begin operating on itself is still an unsolved and unprecedented technical and conceptual challenge. What is also different this time around is that much of the communication that will be forthcoming will not be embodied in objects, like the screens of our phones or the ink in papers, but that it will probably happen directly between nervous systems, maybe through non-invasive wireless systems, or through implanted devices.
So what is different this time around is what has always been different when it comes to collective minds. It is not the idea of collective intelligence, but the way in which we continue to build that eternal dream.
Add Comment
Publish Reply
Hi Cesar, may be this is more a personal reflexion. Just in case it could be useful: Cooperation can be un-intentional and an example is the functioning of impersonal markets, explained through the metaphor of invisible hand by A. Smith. I have always understood the invisible hand metaphor as a coordination mechanism, leading to particular institutions that are product of social and historical evolution. Cooperation can also be intentional, used as a strategy to reach a common end. Its use depends on cultural factors and leads to community actions. An example is co-management as an alternative to public regulation or privatization for common pool resources (Ostrom). Your example of fishing holes in Africa also refers to common pool resources leading to the tragedy of the commons and having collective action as a possible solution. We also learn in a collective way because we are able to identify common goals and the group realise about an improvement in terms of efficiency if they act collectively after having identified those goals. Moreover, “creating the idea of group” depends on culture. There are societies where collective action and the idea of group have been weakened.
The book ‘Economics: a very brief introduction’, by P. Dasgupta explains in a very nice way the main economic and institutional factors for a society, by describing and comparing the world of two children, one of them living in United States and the other one from Ethiopia. The importance of personal/impersonal linkages, markets/community exchanges, trust and social capital...are highlighted in the book. María
Some parts of your text remind me to Scott (Seeing like a State). Thank you so much for the recommendation.
Another term used for Hayek’s explanations is the Walrasian Auctioneer.
“But sometimes these pairwise interactions are aided by something other than mutual feelings of indebtedness, social pressure, currency, or social learning biases”... can also be understood as symbolic public goods? “One way of thinking about how culture and civic identity affect the capacity for collective action is by thinking about the formation of what Rao (2008) calls “symbolic public goods.” Rao builds on the work of Chwe (1999, 2001), who demonstrates how collective action needs to distinguish between structure and strategy. Chwe’s basic argument goes as follows. Most models of collective action assume, implicitly, some preexisting “common knowledge.” When a group of individuals plays a collective action game, whether static or dynamic, it is assumed that individual A knows the payoffs, information sets, costs, incentives, possible moves, and so forth faced by individual B. Individual B, in turn, knows all of this about individual A and knows that individual A knows everything about individual B. Individual A, in turn, knows that individual B knows that individual A knows, and so on. This common knowledge assumption permits games of strategy to be played with a common understanding of the rules of the game: everyone knows how everyone else is playing. In order to understand collective action, therefore, it is crucial to understand its social context through the symbolic public goods that facilitate it. Yet symbolic public goods are themselves the product of strategy and contestation. They can take a variety of forms, including intangible processes of identity formation such as nationalism; physical entities, such as mosques and temples; and periodic ritual events, such as festivals. All of these forms share characteristics of public goods, in the sense that they can be simultaneously “nonrival” (consumption by one person does not reduce the ability of others to consume the same good) and sometimes “nonexcludable” (it is not possible to deny anyone access to the good).” (Mansuri, G. and Rao, V., 2013, Localizing development. Does participation work?. Washington: The World Bank. P. 69-70.)
"he use of knowledge in society" -> "The use of knowledge in society"
Thanks. Fixed reference.
There's a big difference between 1) a society maintaining a set of common stories, understood by each of its members, with which to coordinate itself, and 2) a 'society' that emerges from countless interactions of its members with each other - each member perceiving no common story but instead just a random-seeming set of local stimulus/response events. In the first case the individual is a human, harmonized with his/her community through story. In the second he/she is some kind of 'societal neuron', whose personal understanding of his/her society (and whose personal anxiety and confusion) is apparently irrelevant. This second situation has resulted from technology and is, in fact, new. It is also a very pressing and increasing problem for enormous numbers of people, and so far the only proposed solutions seem to involve religious dogma, machine guns, slogan-based politics and perhaps some kind of vague techno-utopianism. We are very much past due for a secular, analytical and humane response to this 2nd situation - a response that optimizes for the experienced lives of human beings rather than for the cognitive powers of some kind of societal 'super-organism'.
When you say "new," what do you mean. Within the last 100 years, or since the cognitive revolution of 100,000 years ago.
Essentially the last 50 years - I’m loosely talking here about Lyotard’s ‘post-modern condition’, which was probably present to varying degrees much earlier but only really became widespread in the 50’s/60’s/70’s in the wake of new mass communications technologies. This is the notion that the social world is less and less a contest between competing narratives and more and more a contest between narrative and the absence of narrative. Given the primary role of the narrative function by human beings in sense-making, this loss of society's ‘common stories’ is likely to have profound effects - Moisés Naím's book ‘The End Of Power’ perhaps describes some. The daily news perhaps describes others.
A more specific example: the difference between a market that can be understood by the members of a society through a shared ‘market story’ (currency, trading of goods & services, etc), and a ‘market’ for which the members of a society have no common stories with which to understand it (abstract financial instruments in the global financial market). I’m suggesting that these cases differ based on the ability of societies’ members to form common stories with which to understand them, that the second case is socially unstable, and that this ‘neuron-level’ limitation may also be a limitation on how far an emergent societal collective intelligence can develop.
ability to coordinate productive activities among strangers
Milton Friedman’s Pencil might be a nice example
That's a classic example, although it is an essay by Leonard Read, not Milton Friedman.
collective intelligence, decentralized mind, global computer, extended intelligence, socially augmented cognition, or global nervous system
or also global brain, cf Heylighen
data to begin operating on itself
I'm not sure I understand, could an example of that be the coordination of Linux developpers thus relying on the network and meta-data like their commit of logs while Linux itself is used to run the hardware that runs that underlying physical network?
The idea here is to think of AI governments, where data from the census, taxes, and other sources, is processed algorithmically to make public investment decisions. People vote on the algorithms, rather than for Trump or Bernie 😃
"People vote on the algorithms" so things like DAOs or Ethereum or other concensus based system, using the Blockchain or not