But what do we mean by a model or a representation? As I suggested in §4-5, I am using those words to talk about any structure inside one’s brain that one can use to answer some questions about some subject. Of course, those answers will only be useful when your ‘model’ or ‘representation’ behaves enough like the model’s subject does—for the purposes that concern you now.
We sometimes use actual physical objects to represent things, as when we use a picture or map to help us find paths between parts of a city. However, to answer a question about a past event, we must use what we call our “memories.”
But what do we mean by a memory? Each memory must be some kind of record or trace that you made at the time of that prior event—and, of course, you cannot record an event itself. Instead, your brain can only make some records about some of the objects, ideas, and relationships that were involved in that incident. (Indeed, you cannot record an idea itself—and so the best that you can do is to record some aspects of your mental state.)
For example, when you hear a statement like, “Charles gave Joan the book,” you might represent that incident with a script-like sequence of If-Do-Then rules:
However, you also may have wondered about whether that book was a gift or a loan, or did Charles want to ingratiate Joan, or was merely disposed to help a friend. You might have envisioned how the actors were dressed, or some of the words they might have said. Then you might have made several representations for that incident, perhaps including:
A verbal description of that incident.
A visual simulus of the scene.
Some models of the persons involved.
Simulations of how those persons felt.
Analogies with similar incidents.
Predictions about what might happen next.
Why would your brain represent the same event in so many different ways? Perhaps each realm of thought that was engaged left an additional record or trace in some different network inside your brain. This will enable you, later, to use multiple ways to think about that same incident—for example, by using verbal reasoning, or by manipulating mental diagrams, or by envisioning the actors’ gestures and facial expressions.
Today, we still know little about how our brains make those memory traces or how they later retrieve and ‘replay’ them. We do know a lot about how separate brain-cells behave, but we do not yet have good explanations of about how our larger-scale columns and networks of cells manage to represent past events. Nor do our self-reflections help; although as we saw in §8-4, this must involves complex processes, nevertheless it seems to us that we simply ‘remember’ what happens to us.
In any case, one cannot record an event itself, but one can only make some descriptions of how that event affected one’s mental state. Some earlier sections of this book discussed some structures that could used to represent such information. The following section will review some of these, and then speculate about how such structures might be arranged in our brains.
Multiple Ways to Represent Knowledge
This section reviews some structures that researchers have used to represent knowledge inside computers. I will have to leave out most smaller details (many of which are discussed in chapter 8, 19, and 24 of The Society of Mind). Some non-technical readers might do well to skip this section.
Narrative Scripts: Perhaps our most familiar way to represent an incident is to recount it a story or script that depicts a sequence of events in time—that is, in the form of a story or a narrative. The previous section described such a script for the sentence, “Charles gave Joan the book,” and we saw a similar one in §5-3 for Carol’s plan about how to build an arch:
A sequence-script of If-Do-Then Rules
Semantic Networks: However, when we need to describe more details, such as the relations between an object’s parts, it may be better to use the kinds of ‘semantic networks’ we saw in §4-6 to represent a person’s self-model, and in §5-8 to represent the structure of a physical book.
Semantic Networks for ‘Person’ and ‘Book’
Trans-Frames: To represent the effects of an action, it is convenient to use pair of semantic networks to represent what was changed. This is what we did in §5-8 to imagine replacing the top of an arch. This way, one only needs to change the name of a single relationship—instead of altering thousands of points to change a visual picture-like image.
A Trans-Frame for changing the top of an arch
We use the term “Trans-Frame” to name such a pair that represents the conditions before and after some action was done. Then we also can represent the effect of a sequence of actions by linking together a chain of the Trans-Frames to form a story or narrative. Here is a sequence of trans-frames for giving a book:
Such a sequences can describe a script that includes any further details that one might need.
A Script for transferring a Book
Each of these types of representation can answer certain types of questions—but what could enable computers to produce answers so quickly as human brains do? When someone says ‘apple,’ you seem to almost instantly know that a typical apple grows on a tree, is round and red, is about the size of a human hand, and has a certain texture, flavor and taste—yet almost no time seems to elapse between hearing that word and then becoming aware of such things.
To explain how that information could so quickly appear, I conjecture that much of such knowledge is wrapped into structures called Frames. The simplest type of Frame consists of nothing more than a labeled list of some properties of particular object, and you can think of this kind of list as like a printed form that has ‘blanks’ or ‘slots’—each of which can be filled-in with a link to some fragment of knowledge. Then, when you know which slot to inspect, you can quickly retrieve that fragment of knowledge, without need much time to search for it.
A Frame for an Apple’s Properties
Default Assumptions: A valuable feature of a typical frame is that every slot comes already filled in with some ‘default’ or typical value for it. Then you can use such a value to make a good guess whenever you don’t have a more definite answer. For example, you might assume ‘by default’ that an apple is red—but if your particular apple is green, then you will replace ‘red’ by ‘green’ in its color slot. In other words, a typical frame describes a stereotype whose ‘default assumptions’ are usually right—but which we can easily change or replace when we encounter exceptions to them.[173] This would seem to be an important aspect of how we do commonsense reasoning.
173
Where do we get those default assumptions? Answer: we usually make a new frame by making changes in some older one, and values that were not changed at that time will be inherited from those older ones.