What is that “real work”? If the neural networks that make up our brains are not in the business of applying rules, vast libraries-full of them, just what business are they engaged in? How should we think of what they are doing, if not as administering rules? What is the positive alternative to this traditional construal, expressed in non-technical language?
What those networks are doing is (trying to) interpret any new experience or situation as being an instance of some prior category that the brain already understands. They are trying to assimilate each new social/moral situation to an already grasped prototype situation, a template or prototype that has been incrementally created by the brain’s prior experience with its surrounding social/moral reality. They are trying to grasp each of the endless novelties that they encounter as being just a modified case of some kind-of-thing that they have already encountered many times, and with which they have already become familiar. They are trying to interpret the fleeting here-and-now (which is always specific) in terms of their comparatively enduring background concepts (which are always general). They are trying to identify which of their various categories, categories that past experience has constructed for them, is the one into which their current experience fits most closely and most accurately. In sum, they are trying to apply their acquired conceptual and practical wisdom to their current situation.
Why should they, or rather, you, be trying to do that? For the very good reason that your acquired concepts or prototypes are precisely what contains your accumulated information about the world, information beyond what your current and highly specific experience happens to make evident. Those abstract prototypes contain presumptive information about the wide range of features that any instance of an applied concept can typically be expected to display, about the wide range of relations it will typically bear to other things, about the ways in which it will typically unfold or behave over time, and about the ways in which it can typically be controlled or steered. That is the point, after all, of having a conceptual framework in the first place. It embodies your accumulated understanding of the world’s enduring background structure, your grasp of the unchanging background framework within which the ephemeral and the changeable are always constrained to unfold.
Consider, for an example of moral perception in particular, the arrival of lunchtime in a typical elementary-school classroom. Every student retrieves a paper-bag lunch from the cloakroom and settles down to consume its contents. You are one of those students and, while eating, you perceive Johnny surreptitiously attempt to remove a banana from the lunch-bag next to Michael. On the face of it, you are witnessing a case of theft. And that interpretation implies many things: that the banana belongs to Michael, that Michael will be seriously aggrieved when he discovers Johnny’s affront, that Johnny has inadequate self-control, that a noisy conflict will ensue if events are left to themselves, and so on and so on. This situation, as described, warrants some immediate intervention.
Most obviously, you might just openly berate Johnny in front of the other students. Or, more boldly, you might seize the banana from Johnny and quietly return it to Michael. Or you might call the teacher and rat Johnny out. These hardly exhaust your possible responses, but they are all typical sorts of responses to a typical sort of problem, and which response you choose will depend on contextual factors such as how big and mean Johnny is, how susceptible he is to collective disapproval, and how reliable the teacher is at dispensing justice. Perhaps the first path is the best response, with the second and third left as backups if the first path fails to return the situation to a just equilibrium.
And so that is what you do: berate him on the spot. All within a second of witnessing the presumed theft. Because your eight-year-old brain is already keenly tuned to that sort of possibility and to thousands of other social possibilities as well. Given your well-trained neural networks, it takes only the external perceptual situation itself to provoke the interpretation of theft. And it takes only that conceptual interpretation itself, in the context of one’s ever-present character and background information, to activate your overt social response.
Your interpretation, of course, might be incorrect. Perhaps Johnny was just trying to retrieve his own banana, earlier stolen by the avaricious Michael. Perhaps your openly berating Johnny was inappropriate, since everyone in the class except you witnessed Michael’s earlier theft but was too frightened of Michael to do anything about it. If so, Johnny has now been victimized twice over, once by Michael and once by you.
To be sure, there are many other convoluted possibilities, in addition to or beyond this one. But they are increasingly unlikely, compared to your first take on the situation. This is why your brain fell so swiftly and easily into that straightforward interpretation: theft is the simplest, most obvious, most probable explanation of what you have actually seen, and that’s why it’s the explanation that the brain tries first. Furthermore, once that explanatory assumption is in place, an immediate attempt at restitution is the most natural expression of your antecedent character and your acquired social skills.
What is impressive here is not just the swiftness with which your cognitive resources get tapped. It is the enormous range of alternative possibilities to which your brain is/was no less prepared to respond, and with equal swiftness, insight, and know-how. If, instead of a banana theft, you had witnessed Mary accidentally press her hand against the point of a newly sharpened pencil, your recognition of her pain and your comforting response would have been just as quick. If you saw the class’s pet rabbit escape from its (poorly locked) cage, you would know to retrieve it and return it to its proper home. If you had turned to see a small fire blazing in the classroom’s bookcase-corner library, with Johnny (him again!) slipping a plastic lighter into his pocket, you would grasp the significance of the event instantly and let out a loud warning to everyone in the room. If (here we deliberately choose something unlikely) Superman, with cape swirling, then bursts through the open classroom window and asks, “Which way did the fire-bug go?!,” you would know to point to Johnny’s fleeing backside as he hightails it out the classroom door. If . . . if . . . if . . . for a thousand thousand “ifs” and more, even your eight-year-old self would be competent to recognize the situation and to respond to it swiftly and appropriately.
This extraordinary breadth of capacity is a consequence, in part, of the combinatorics of the already large number of neurons the brain uses to represent any sort of social situation. It is the same trick, once again, used by your television screen, in order to display an almost endless variety of possible pictures, despite a (large but) finite set of pixels with which to portray them. The retina of the eye uses the same trick, recall, but boasts many more “pixels” than a TV screen. Your perceptual capacities, accordingly, far exceed the modest range of that familiar technology.
Of course, simply representing something at the perceptual level does not mean that you understand it, and that is strictly what concerns us here. To understand a perceptual input is, as we saw above, to assimilate it to one of the brain’s learned prototype situations, to one of the standard, recurring, well-patterned kind of circumstances that one’s past experience has impressed upon your memory and your habits of behavior. That memory and those habits, you will recall, are a matter of the acquired configuration of the brain’s synaptic connections and their synaptic strengths or “weights.” For it is those collective synaptic “filters” or “transformers”—the ones that intervene between each of the brain’s many reality-portraying neuron populations— that steer the initial perceptual representations into the higher-level prototype patterns that fit those percepts most closely.
Look again at the toy network of figure 1 and note that its 327,680 synaptic connections were there adequate to steer a wide variety of possible input face images into one or other of exactly five emotional prototypes. If we suppose that this ratio (i.e., 327,680 synapses for every five categorial prototypes) is roughly typical, then a brain like yours, with 100 trillion synaptic connections, should be able to learn, and to deploy immediately (when appropriate), something in the neighborhood of (5 / 327,680) × 100 trillion = 1.5 billion distinct categorical prototypes! Now, presumably we don’t have quite that many distinct categories awaiting activation. That number is best viewed as a theoretical upper limit on what we might achieve. But in light of how our cognitive systems evidently do their jobs, it is small wonder that even your grade-school self is hair-trigger ready for such an astonishing range of situations, social and otherwise—and ready, note well, with an astonishing range of understanding and relevant skills.