The phrase comes from David Hume, a philosopher who was rarely wrong. Among “perceptions” he included everything buzzing inside our heads: thoughts, feelings, imagination, memory, as well as true sense perceptions. Hume opposed the identity of the “self” — myself, yourself — which philosophers of his day believed to be a “simple and continued” entity. Instead of an integrated, pilot-like controller of our actions, he found flow and fragmentation. “I never can catch myself at any time without a perception,” he wrote, “and never can observe any thing but the perception.”
Much of what we have learned since Hume’s day appears to confirm his claim. If we equate — as I think we must — the self with the conscious mind, we bump into the current understanding of consciousness as a highly fragmented set of capabilities. These capabilities include short- and long-term memory, attention, a theorizing “interpreter,” various regulatory emotions and feelings, all interacting with a massive inflow of sense data, yet exchanging little information.
How this adds up to a sense of “myself” is a total mystery to science. And the mystery has been given a name: the binding problem. If consciousness isn’t simple and continued, what makes it feel that way?
Until recently, I considered the binding problem of the self to be just another gap in human knowledge. By and large, normal people don’t behave in a fragmented manner, so I supposed a binding force, though not yet discovered, must exist.
These were comfortable assumptions. I have just finished a fascinating but disturbing book by Andy Clark, Being There, that undermines them by providing a method whereby a fragmented being might behave like an integrated character.
Clark’s thesis is simple and persuasive. Suppose the human mind “offloads” much of its cognitive needs onto the environment — and by environment, Clark means nature but also language and the social world. (That we do offload such needs is apparent: think of the calculator, or a list of groceries, or this blog, or the white house my wife always uses as a landmark.)
In essence, we have then built a “scaffolding” around us that tremendously magnifies and integrates our thinking capacity. Much as the natural world constrains our actions — we stop short of the precipice — so the social and linguistic environments drive us to behave in specified, predictable ways. For example, I buy the items on the grocery list, and my wife invariably turns right at the white house.
The disturbing aspect of this explanation is that it doesn’t require a self to explain human behavior. (Clark himself takes no position on the matter.) Each of our fragmented capabilities can play off the natural and cultural scaffolding, needing only a few simple rules of priority to interact with the other capabilities.
Long-term memory may bring up the white house as a landmark, commanding attention and initiating a right turn in the road. Short-term memory may scan the grocery list, initiating the search and retrieval of an item. No self or central command is needed for these outwardly integrated activities, any more than for stopping short of the precipice.
Even the idea of the self as a pilot or central command, which receives signals from the sensory organs and orchestrates the appropriate actions, has been shown to be logically untenable. Thinkers observe that an internal “homunculus” or mini-me isn’t much of an explanation for human behavior: it just pushes the problem one layer back. Who decides for the homunculus? Another, inside him? And another? The central command theory appears to dissolve into an infinite regress — commands inside commands.
If the self — the traditional source of moral agency and responsibility — fragments and disintegrates, what is left on which to base moral judgments, or morality itself?
We have suddenly entered a Twilight Zone episode in which nothing, least of all ourselves, is what it appears to be. But the story isn’t done. Here, submitted for your consideration, is the remainder of the plot.
We are not mechanical devices. We are self-reflecting organisms. Second-level cognition — what Clark calls “thinking about thinking” — is, Clark himself observes, the most powerful cognitive capability available to the human race, and the one that separates us from other animals. It involves self-criticism: testing one’s ideas from a variety of perspectives. Clark offers an example of how we use written language to this end:
By writing down our ideas, we generate a trace in a format that opens up a range of new possibilities. We can then inspect and reinspect the same ideas, coming at them from many different angles and in many different frames of mind. We can hold the original ideas steady so that we may judge them, and safely experiment with subtle alterations.
The consequence is self-motivated change in thinking and behavior. I think about my plan, consider the facts, change the plan, and act. The reflecting agent, however fragmented, is the self. Clark, like most cognitive scientists, is enamored with the mechanisms of thought and mind, described in terms of “computations” and “algorithms.”
But Being There was published in 1997 — since then, brilliant researchers like Antonio Damasio and Jonathan Haidt have shown the emotions to play a leading role in practical problem-solving and moral judgment alike. An organism capable of self-reflection and self-criticism, endowed with feelings about right and wrong behavior, possesses all the equipment necessary for morality.
We each inhabit a subjective consciousness: the consistent, persistent, integrated feeling that I am me. Some schools of thought, like the behaviorists, downplay this feeling. But that won’t do. Thinking without consciousness is impossible to a normal human being. In his Chinese room metaphor, John Searle illustrated the fallacy of identifying mechanical computation with human thought. We are not calculators. We are not robots. The difference lies precisely in our ability to externalize ourselves, and reflect on what we see — who we are, how things might be different, how we might become better.
What can I assert with confidence about the feeling that I am me? This me isn’t an illusion, but neither is it a pilot-like figure, commanding my actions. I learn about my actions much as I learn about the actions of others: by watching me do things. On occasion, I have no clue why I acted a certain way. It might have been the action of a total stranger. I am, no doubt, fragmented. But I can reflect on my behavior and its consequences, and I have some power over my actions — as I surely don’t over the actions of others — to change and align them with some model in my mind.
The me, I submit, is not myself but a model of myself: an aspiration, not a reality. And this slow march toward perfection is only possible because of the scaffolding provided by our moral traditions.
Clark’s message is that we aren’t the brilliant and far-seeing individuals we fondly imagine ourselves to be. Private thought stands on the stilts of the ancient wisdom of the community. By far the most powerful example is language. Individuals contribute variably to the evolution of language, but only the crazed or the perverse would try to invent a personal idiom. By the same token, the only languages that survive are those that are easy for children to learn, and that easily expand the thinking power of the user relative to his environment. The rest die out early, or are never born.
The key factor is selection. That is true for all forms of cultural scaffolding, including morality. What we call tradition is the triumph across centuries of chosen behavior over chance and necessity. We achieve integrity — we become ourselves — by standing on the shoulders of those who came before us.