Terry Anderson and I have written a fair bit about the different social forms that apply in (at least) an educational context. We reckon that they fall fairly neatly into physically overlapping but conceptually distinct categories of groups, nets and sets. In the past, we used the term 'collectives' instead of 'sets' but we have come to realise that collectives are something else entirely.This post starts with an overview of the distinctions and then drifts into vaguer territory in an attempt to uncover what it might be like for something to have meaning for a social entity. That's a rather bizarre concept at first glance: is there any sense at all in which a collection of people, not the people within that collection but the collection itself, can feel or think anything and, if not, how can anything be said to have meaning to it? And yet, oddly, we do ascribe human attributes to collections of people all the time in our everyday speech - 'Apple is a creative company', 'Canada got another gold medal', 'We came top of the league', 'the crowd is angry', 'this is the most enthusiastic class I've ever taught', 'Google beat Oracle in the court case', 'Athabasca University is committed to open learning' and so on. While this is often just a shorthand notation for something else or a poetic metaphor, the ubiquity of such language makes it worth examining further.

Groups, nets, sets and collectives 

Groups are the stuff of conventional teaching and learning: they are distinct and intentional entities that people join and know that they are members. You are in a group or out of it: you might be more or less engaged, but there is no real in-between state. Groups are generally characterised by things like purposes, collaboration, hierarchies, roles, exclusion. We know a lot about groups and their effects on learning, and the whole field of social constructivist models of teaching and learning is based on them.

Networks are more tenuous entities. To join a network you connect with one or more of its nodes. You might intentionally wish to make connections with particular people or kinds of people, but a network has no formal constitution, no innate roles and hierarchies, no innate exclusion: it's about individuals and their connections with one another. It is composed of nothing but connections and ties and has no formal boundaries. Networks are traversable and offer ways of linking and connecting to others and their knowledge. Learning in networks tends to be informal, connected and undirected by any individual. Networks are great for on-demand and serendipitous learning, combining social ties with unbounded knowledge.

Sets are about categories and topics. Set-based learning is about finding people and knowledge based on shared characteristics, typically a topic about one wishes to learn. Wikipedia, YouTube, and Google Search epitomise the nature and value of sets in learning, with ascending social interest sites like Pinterest or Quora beginning to enter the fray. However, libraries and bookshops are also primarily set-oriented, so this is nothing new. Unlike networks, there may be no direct connection with others and certainly no expectation of sustained interaction (though it may occur and develop into other social forms). Unlike groups, there is no formal constitution of a collection of individuals. It is just a bunch of people joined (in a set-theory sense) by a shared interest.

When social forms act together as a single entity, they become collectives - not a social form, as such, but the result of social forms and the interactions of individuals within them. A collective may be the result of direct or indirect interactions of individual autonomous agents, such as may be found in natural social forms like ant or termite nests, herds, flocks or shoals or, in human systems, in the operations of money markets, mobs, stock exchanges, group-think and forest path formation. The 'invisible hand' is a collective in action, the result of myriad local interactions rather than a deliberate global plan. The environment plays a strong role in this: things like the availability of resources, sight-lines, weather patterns, topology and more play a role in determining how such dynamics play out.

In computer-based systems, the combination that leads to a collective is not just a result of the emergent results of individual agents but may be effected and consequently notably affected by a machine: Amazon recommendations, Google Search, PayPal reputations and so on are all combining intelligent and independent actions of humans using algorithms in a machine in order to affect human action. The computer system extends what is possible through direct/indirect interaction alone, but it is still powered by individual intelligent beings making intelligent choices. It leads to a cyborg entity where collective emergence is part-human, part-machine. This makes such systems very powerful and flexible as a means to create collective intelligence that is directed to some end, rather than being simply an emergent feature of a complex system that happens to have value. Not only does the environment itself play a role in shaping behaviour, as in 'natural' systems, but it actually creates some of the rules of interaction. In effect, it bends and sometimes creates the rules of social physics.

Values in collections of people

In some sense, groups, sets, nets are all identifiable entities in the world that achieve some kind of action or purpose that is distinct from the individual actions or purposes of the people of which they are comprised. Clay Shirky talked of them as first class objects - things in themselves. But are these entities, these first class objects, anything like people? Are there values we can ascribe to them? Do they have intentions and purposes that are analogous to those of individuals? Do they have attitudes that are separable or different from the attitudes of those that comprise them? This is a problem that my student Eric Von Stackelberg has been exploring in his masters thesis and he has made some very interesting progress on this by using categories, that are used in psychology to describe individual values, as a means of describing group values ('group' used here in the generic sense of a collection of people of some identifiable sort). I've been challenging him to clarify what it would mean for that to be true. Can a bunch of people (not the individuals, the bunch itself) be kind, or hedonistic, or happy, or avaricious, or whatever in a manner that is meaningfully different from saying that the individuals themselves, or even a majority of them, have those attitudes? It seems that a corollary of that implies we might ascribe to them something akin to emotion. Could a bunch of people (the bunch, not the people in the bunch) feel happiness, amusement, tiredness, anger, pain, hate or love? I find this a difficult concept to get my head around. And yet...

It seems intuitively obvious that there is something organism-like in a social cluster. It is certainly normal to speak of organizational values, national values, group beliefs, group norms and so on. Athabasca University, for example treats itself as a unified entity in its mission statement that talks of values, purposes and intentions as though it were (almost) a human being. Corporations are treated in the law of some countries almost exactly like people (albeit odd ones, given that all would be diagnosed as having, on analysis, serious psychopathic disorders). Nations are very similar - we can talk of America invading Afghanistan without batting an eyelid, even though it is very clearly not something that is literally or physically the case in the way it would be were, say, a bully to pick on someone in a playground. A similar but far more worrisome phrase like 'the French have always despised the English' sounds like it plays on a similar notion but suggests something rather different. When we say that a country has invaded another we are talking about a group activity, something organized and intentional, whereas when we suggest that a whole population of people thinks in a certain way we are talking about a set: people with the shared attribute of nationality (the same applies to race, or gender, or physical attribute, etc - that way bigotry lies). There are interesting hybrids: it is normal to say 'we won' when a hockey team wins even though 'we' had negligible input or nothing to do with it at all. We identify at a set level (we, the supporters of the team) in a manner that encompasses the team (a distinct group). It is harder to find examples of networks being treated in quite the same way, though the flow of memes that is so easily facilitated through social networking sites may be an example of values of a sort being a feature of networks. However, the innately diffuse nature of a network means it is significantly less likely to have values of its own. It may be predicated on individuals' values (e.g a network of religious believers) but a network itself does not seem to have any, at least at first glance. Networks are primarily about individuals and their connections to other individuals, each seeing their part of the network from their own unique perspective. This is not promising territory to find anything apart from emergent patterns of value.

There are natural parallels though, that suggest an alternative view. It makes no sense to think of an ant colony as just a load of autonomous ants - the colony itself is undoubtedly a super-organism and an ant from such a colony is, on its own, not a meaningful entity: it is constituted only in its relation to others, as part of a single network. We can use telological language about the colony, and even ascribe to it wants, desires and intentions. It is also absolutely reasonable to think of an organism like a human being as a group/network/set of tightly coupled cells that are behaving, together, as a single unified entity that is not dissimilar to an ant colony in its complexity and interdependence. An individual cell may live on its own, but its meaning only becomes apparent in the presence of others. Even at a cellular level, our cells are a community of different symbiotic organisms. The vast majority of the cells in our bodies don't even have human DNA (that still staggers me - what are we?) but we still cannot think of ourselves as anything other than individuals that have values, intentions, meaning and - well - an autonomous life of their own. Are social forms so very different? It seems that at least one contained network that constitutes an entity may well have values because, well, we have values and we can be viewed as networks. In fact, we can also be thought of as sets and, in some senses, as groups.

While chatting about this kind of thing, a friend recently remarked that perhaps the most crucial value that we can ascribe to an individual is the value of survival: the will to survive. An arbitrary collection of entities does not have this. If we are thinking in terms of organisms, then I guess we might more properly think of it in evolutionary terms as a bunch of genes seeking to survive, but that's a layer of abstraction higher than needed here.

At the individual organism level it is the organism that tries to survive. This is one obvious reason that it is logical to think of an ant, termite or bee colony as a single organism: individuals will readily sacrifice themselves for the colony exactly as the cells in our own bodies continuously sacrifice themselves in order to protect and sustain the entity that we recognise as a person. We can easily see this survival imperative in intentionally created groups, from small departments to sewing circles, from gangs and teams to companies to countries (groups). If a group exists, it will typically try to preserve itself, and individual members may often be seen as expendable in meeting that need: thing of countries at war, political parties, hockey teams and so on. We can also see it in less rigidly defined entities such as cultures (sets/nets) and institutions (sets/groups). Even though individuals may have no formal connections with one another with, at most, tenuous networks and no unifying constitution, the simple fact of observable similarities and shared features leads to a self-reinforcing crowd effect that leads to survival. Often, intentional groups will be formed to support these but the interesting thing is that they are not groups defending their own 'lives' but a kind of collective antibody formed to protect the broader, sometimes barely tangible, set. People who form organizations to defend society against some challenge to what they see as being its central cultural, aesthetic, ethical or social values are doing just that. The set of which they feel a part is somehow greater than the group that they form to protect it.

It is harder to see this in human networks. Although there do appear to be emergent and dynamically stable features in many networks, that's just it: they are emergent features like a solonic wave in a river, the rhythmic dripping of a tap, or a whorl of clouds in a storm. It makes far less sense to talk of a cloud formation as trying to survive than it does of an ant colony. We do, however, see moods and trends spread through networks - if you know people who are getting fatter then you are far more likely to become fat yourself, for instance, and depression is contagious. It is reasonable to surmise that values spread in much the same way: indeed, if we look at extremes such as the spread of Naziism or the growth of fundamental religions, there is a very strong sense in which networks act as conduits for value. But I think that's it: they are conduits, not containers of value. Whatever has values may consist of networks that facilitate the spread or even the formation of those values, but it is the thing, not the network, that is what we care about here.

All of this leads me to suspect that the social forms that Terry and I identified as different in their pedagogical uses and affordances have some fundamental characteristics that go quite a way beyond that and relate to and intersect with one another in quite distinctively different ways. When we picture them as a Venn diagram it homogenizes these differences and makes it seem as though there are simply overlaps between vaguely similar entities, but there is more to it. Networks provide conduits for the spread of value between and within sets and groups. They are not the only conduits by any means: for example, if the human race were attacked by an alien civilization then I think it unlikely that a network would be needed to spread a range of values that would surface fairly ubiquitously (as a set characteristic), though it might help spread attitudes to how we should respond to such a threat. The same is true of many things in the more mundane realms of broadcast media, city planning and publication, not to mention the effects of natural features of the environment. Part of the reason for the distinctive culture and values in Canada, for example, is surely related to its dangerously cold climate that makes assistance to and from others a very strong necessity, plus a million other things like the opportunities afforded by its abundant natural resources and its proximity to other places. Prairie people are not quite the same as mountain people for reasons that go beyond historical happenstance and path dependencies. This is all about sets: shared characteristics and features. Sets can help to generate values: the fact that shorter-than-average people have to interact differently with the environment than taller-than-average people in many different ways leads to (at least) greater tendencies to share some values. The fact that people are collocated in a region, quite apart from network and group facets that emerge, means they are likely to share some attitudes and tendencies. It's simple evolutionary theory. It's why the finches in the Galapagos Islands have evolved differently: they have to interact with their different environments, and those environments have varied constraints and affordances. Other factors like path dependencies play an enormous role. Networks have a crucial part here too as co-evolution occurs not only in response to the environment but in response to the interconnections between agents in the system. In human systems, groups are both containers of networks and are themselves nodes in networks, so there are layers of scale that make this quite a complex thing.

The complexity becomes much more manageable if, instead of focusing on the social forms of aggregation, we think of values as being attached not to the aggregations themselves but to the collectives that emerge from them. Collectives are, by definition, behaviours that emerge from multiple interactions and are different from those interactions. A human can be viewed as a net, a set or even a group (there are hierarchies of organisation in which the brain might be seen as a controller) but it is the collective, the emergent entity that arises out of sets, nets and groups that is recognisably an individual, that has values. In the development of nationalist or religious values, it is the operation of algorithms that makes the set, net or group of which it is comprised into something distinct and potentially able to embody values, typically resulting from a mix of interactions combined with intentional categorisation by individuals - a collective.

I don't see any of this as suggesting even a glimmer of consciousness but it does seem at least possible that collectives can, at least sometimes, be described as having tropisms and to talk, perhaps loosely, in terms of intentionality. Whether this is enough to ascribe values to them is another matter, but it is not entirely absurd. We sometimes talk of plants as 'liking the sun' or 'liking the shade' in ways that probably have more to do with metaphor than beliefs about plant feelings, but there is a sense that it is true. It is even more obviously true in animals: even single-celled organisms are slightly more than just billiard balls bouncing round in reaction to their surroundings. They have purposes, aversions, likes and dislikes. Some exhibit fascinatingly complex behaviours - slime moulds, for example. It is not a great stretch from there to talking about human collectives in similar terms. Financial markets, for instance, are archetypal examples of human collectives that in principle need little or no machine mediation, yet move in complex ways that are not simply the sum of their parts. And, interestingly, we talk blithely of bull and bear markets as though they were in some way alive and, in some sense, imbued with feelings and even emotions. And maybe, in some sense, they are.