"Emergent Learning" Is an Oxymoron

In the introduction to Steven Johnson’s oft-referenced but seldom understood book Emergence: The Connected Lives of Ants, Brains, Cities, and Software, he describes emergent systems as follows:

In the simplest terms, they solve problems by drawing on masses of relatively stupid elements, rather than a single, intelligent “executive branch.” They are bottom-up systems, not top-down. They get their smarts from below. In a more technical language, they are complex adaptive systems that display emergent behavior. In these systems, agents residing on one scale start producing behavior that lies one scale above them….[emphasis added]

The key idea here is that the individual agents do not generally understand and consciously orchestrate the observed emergent behavior. Individual ants don’t coordinate the foraging activities for the colony through weblog-like networked communication; they respond individually and mechanically to certain cues from other ants in their immediate surrounding who, in turn, are responding mechanically to cues (including but not limited to communications from other ants) in their own local environment. The individual ants don’t know about coordinated foraging; the hive as a whole knows. Emergence is when some totally new phenomenon…well…emerges out of the collective behavior of much simpler parts where the individual simpler parts are responding through simple rules to their local environment. For example, some philosophers of mind suggest that consciousness is an emergent property of brains. Each individual neuron is simply a mechanical switch responding to triggers in its immediate environment. But when you string a bunch of these switches together in the right way, you suddenly have an aware being. The neurons aren’t individually conscious; it’s the brain as a collective entity that posesses the emergent property of consciousness.

When people talk about “emergent learning” these days, this is not generally what they mean. What they generally mean is some form of rapid consensus-building in which a group of people can share observations and make coordinated decisions without any one person filling the role of executive command and control. This is, no doubt, an important phenomenon to understand and try to cultivate. However, it is not emergence. A democratic decision-making process is not sufficient for an action to be called “emergent.” Almost by definition, if you have the kind of self- and group-awareness that is usually entailed when we use the word “learning”, you can’t have emergence. You can say that a colony of ants “learns” what the best foraging strategy is, but it is the colony as a whole that “learns,” not the individuals. If the individual ants were able to learn the best foraging strategy, communicate it throughout the hive, and consciously arrive at a consensus, then their adaptive foraging would not be an emergent behavior. So “emergent learning” as the term is currently being used is actually an oxymoron.

Re-reading Johnson’s book, it seems to me that he’s fairly clear on this point throughout. And yet, people seem to consistently mis-understand him. Apparently, Johnson’s readers are not the only ones with this difficulty. He describes molecular biologist’s Evelyn Fox Keller’s decade-long struggle to convince her colleagues that, in fact, there is no “pacemaker” cell directing the behavior of slime mold colonies. “It amazes me how difficult it is for people to think in terms of collective pheonomenon [sic],” he quotes her as saying. Seeing this quote again triggered a memory for me and compelled me to go back and revisit Howard Gardner’s book The Unschooled Mind. Without quoting him at length, Gardner’s basic hypothesis is that, as children, we all develop certain intuitive theories about how the world works based on our own experiences and observations. These basic theories (which, fascinatingly enough, often fit very well with early scientific theories such as Aristotelean physics and Lamarckian evolutionary biology) have to be undone in order for students learn theories that contradict them. For example, I can tell you from personal experience that students tend to cling to the notion of Lamarckian evolution (e.g., the giraffe has a long neck because it stretched it to reach leaves high in the tree and then passed this stretched neck on to its children) even after they seem on the surface to have understood the Darwinian alternative. If you give them a test asking them what Darwin’s theory was, they will give the correct answer. However, ask them to give you an example hypothesis of how a particular animal may have evolved and their Lamarckian ideas peek through.

I suspect that emergence rubs directly against the grain of one of the intuitive theories that we develop as children. The idea of a system exhibiting judgment when its components are dumb just seems…weird. Intelligence feels like it should be irreducible and tied in a one-to-one relationship with living beings. The idea that the people are dumb (and remain dumb) but the crowd as a whole exhibits intelligence that none of the individual members have is just something that we have trouble processing. Nevertheless, if we want to figure out how the concept of emergence can (and can’t) be applied to the domains of group decision-making, then we need to get it through our individually thick heads that peer-to-peer education in a networked environment is not the same thing as emergence.