Imagine that you plan to take a trip. You start to think of how you'll pack your traveling case and start some spatial problem-solving agency — call it Packer — to see how to fit the larger items in. Then you interrupt yourself to think about your smaller things, perhaps of how to pack your jewelry in a smaller box. Now Packer has to reapply itself to a new and different box-packing problem. The problem of keeping track of what is happening is hard enough when one agency calls on another one for help. Until the other's job is done, the first agency has to keep some temporary record of what it was doing. In Packer's case the problem is even worse because it interrupts itself to pack the smaller box. And here is the important point: when that second packing job is done and we return to the first, we mustn't go all the way back to the very beginning, or else we would be caught in a circular loop. Instead, we must return to the point where we were when we interrupted ourselves — which means the system needs some memory to keep track of what it was doing before. This is exactly the same problem we mentioned long ago when Find and See had several different jobs to do at the same time.
Why do we so often get confused when we're interrupted? Because then we have to keep our place in several processes at once. To keep things straight, our memory-control machinery needs intricate skills. Yet psychologically, we're unaware that ordinary thinking is so complicated. If someone asked, What was your mind just doing? you might say something like this:
I was thinking about packing that suitcase, and started to wonder if the umbrella might fit. I remembered that on an earlier trip, I managed to fit my camera tripod in the same case, and I tried to compare the umbrella and the tripod in my mind, to see which one was longer.
Now, this might be a true account of some of the things you were thinking about. But it says little of how the mental work was actually done. To understand how thinking works, we'd really need descriptions of the processes themselves:
A few moments ago, I activated two micromemory-units inside Packer, one of my space-arranging agencies, while also activating one of Packer's memory-control scripts. This script proceeded to use the information inside those two micromemory-units as cues to fetch certain partial states from the long-term memory-system attached to Packer. Next, the script controlling Packer's memory-system requested a certain higher-level planning-agency to record most of Packer's present state. Then it interchanged the contents of the two active micromemory units, and then used other cues to fetch another, second script from long-term memory — and thus erased the present copy of itself. The last step of that second script caused yet another micromemory-unit to restore Packer to its previous state, so that the original script could continue on its interrupted course. Then . . .
But no one ever says such things. The processes are too many levels away from those we use to work the short-term memories involved with language and consciousness. We couldn't think so if we wanted to — without knowing more about the anatomy of our memory-
machinery. Even if we had ways to represent those processes at higher levels, our memory-controls would probably be overloaded by attempting, both at the same time, to solve a difficult problem and to remember everything that was done while solving it.