Memo 3

Marvin Minsky

Role Models, Mentors, and Imprimers and Thinking

You can’t think about thinking without thinking about thinking about something. —Seymour Papert

Many children dislike the subject called Math because they don’t find significant links between those classroom ideas and their everyday lives—so they’re left with scattered fragments of knowledge. We all know what frequently happens then: because we failed to provide them with adequate “cognitive maps,” many students end up envisioning “Math” as an endless progression of unpleasant tasks.

Now let’s take a larger-scale view of what we want our schools to do. Of course we want them to teach the contents of subjects—but we also want schools to “socialize” children—by fostering such civil skills as cooperation and courtesy. And we also ask schools to infuse our children with morals and ethical values, along with suitable goals and ambitions. All this amounts to a very large order—but it seems to me that something important still is missing: teaching children good ways to think about thinking. After all, thinking is the principal tool we use to accomplish everything we do.

Accordingly, our schools should try to help their pupils get ideas about how their thinking works and I’m sure that good teachers already do this. However, I’m not proposing to add a Psychology course to the elementary curriculum, because there’s no consensus among psychologists about how to augment our mental resourcefulness. So I’ll propose a different approach: to develop ways to get children to think of themselves as though they were programmed computers!

You might see this as a dreadful idea: how it could possibly help to think of oneself as being like an unfeeling machine. But consider the alternative views that our popular cultures tell us to use:

“Each individual is born endowed with certain talents and aptitudes, and those are ‘gifts that’ we can’t give back and exchange—so education can only help us to make the most of what we happened to get."

"Our best ideas come from processes called Intuition and Inspiration, which are magical things that can’t be explained—and examining them can damage them."

Such views might seem harmless, but I think they’re pernicious: they imply that our minds can’t do much by themselves, but can only make selections among the ideas that happen to “come to us.”

You might argue that the idea of being a kind of machine should cause even worse feelings of helplessness. However, I’m not suggesting that you should envisage yourself as like some kind of unchangeable object—such as a toaster or sewing machine. Instead, I’m suggesting that if you think of yourself as a thing you can program, all those constraints will suddenly weaken! For example, when you recognize that you have a “bug,” you can imagine that your trouble is caused by some steps in the programs you call your mind. And now, instead of feeling powerless, you can imagine things to find and repair—inside some particular parts of your mind. Of course, this doesn’t mean that a person can actually do such things—but this viewpoint could serve as an antidote to the older belief that we each have a fixed set of “aptitudes,” or an unchangeable dose of “intelligence.”

Thinking about Thinking about Ways to Think

Suppose you get stuck at achieving some goal. This could lead you to conclude that you’re simply not suited for that kind of job—perhaps because you lack certain ‘talents’ or ‘aptitudes’. However, if you can learn to recognize the particular way in which you got stuck, or the particular kind of trouble you’re in, that diagnosis can suggest more appropriate ways to think. Here are a few examples of these from chapter 7 of The Emotion Machine.

If a problem seems familiar, try reasoning by Analogy. If you solved a similar one in the past, and can adapt to the differences, you may be able to re-use that solution.

If the problem still seems too hard, divide it into several parts. Every difference you recognize may suggest a separate subproblem to solve.

If it seems unfamiliar, change how you’re describing it. Find a different description that highlights more relevant information.

If you get too many ideas, then focus on a more specific example—but if you don’t get enough ideas, make the description more general.

If a problem is too complex, make a simpler version of it. Solving a simpler instance may suggest how to solve the original problem.

Reflection. Asking what makes a problem seem hard may suggest another approach—or a better way to spend your time.

Impersonation. When your ideas seem inadequate, remember someone more expert at this, and imagine what that person would do.

Resignation. Whenever you find yourself totally stuck, stop whatever you’re doing now and let the rest of your mind find alternatives.

Knowing How: The best way to solve a problem is to already know how to solve it—if you can manage to retrieve that knowledge.

If none of these methods work, you can ask another person else for help.

Do we really need to teach such things, in view of the fact that all normal people (including children) already use all or most of these methods? Yes, because, unless we have names for them (or other ways to refer to them) it will be too hard to think about them (so that we can think about improving them). So our conjecture is that many children could greatly augment their resourcefulness if we could provide them with more effective ways to think about their mental processes. (I don’t know of any good evidence for this—but that could be because no one has done such experiments.)

Now let’s look at possible applications. Every person has some weaknesses that could be seen as ‘disabilities.’ For example, some persons are relatively inept at dealing with 3-D spatial relationships, but once we recognize such a deficiency, we can suggest alternative ways to envision things (provided that we can make adequate theories of what might be wrong with that the data-structures that person is using). A child once complained to me that she couldn’t see how to draw a perspective cube—but when I showed her how to draw the appropriate lines, but she objected “That looks like a cube but it must be wrong, because cubes don’t have any slanted edges.” However, this discomfort was relieved when we pointed out that a picture need not correctly represent all the features of what it portrays and, often, the essence of making good drawings is finding the right kinds of simplifications. In other words, this child had simply become too ambitious!

Similarly, some children have trouble with memorization—and psychologists know many different ‘mnemonic’ techniques. However, we don’t know nearly enough about which techniques could help which children. But we probably could make good suggestions if we guess which kinds of ‘store’ and ‘load’ procedures each particular child tries to apply to the data-structures she’s trying to use.

Likewise, many youngsters appear to be deficient in physical dexterity and coordination—and in some cases when physical training doesn’t help, this might happen because those children have incorrect mental models of how their bodies work. Then again, if we could diagnose those kinds of ‘bugs’, we might be able to teach those children to improve their ‘body-images’ by using mental, rather than physical training.

It also might even be possible to improve the ways that a person makes higher-level decisions. For example, one can imagine computer programs that could help a person to learn to avoid the common mistakes described in Dan Ariely’s Predictably Irrational book. Similarly, many people are prone to logical fallacies—and we might be able to develop game-like programs that help young children to acquire better kinds of reasoning.

How do children acquire self-images?

Some readers may object that intervening to teach about Thinking could be too intrusive (and possibly harmful). Indeed, one might argue that, instead, we should protect young children from such concerns and let them enjoy their carefree childhoods. However, whether we try to guide them or not, our children will still develop ideas about themselves—and the less we influence those self-images, the more they will get them from somewhere else. So now let’s explore some questions about how children acquire their ideas about themselves.

We each construct representations of our abilities, goals, aversions, and tastes— as well as of our dispositions, talents, and traits—and of our physical appearances and of our present and future social roles. We often use the term “self-images” for these descriptions of our own characteristics. Of course, we all would like to understand how our brains embody these structures, but no one has yet discovered much about how our brains construct those representations. Chapter 9.1 of The Emotion Machine makes some suggestions about what neuroscientists could try to find in those networks.

From where do our children’s self-images come? Of course, they copy a lot from their parents, siblings, teachers, and friends but (as noted in Memo 2) they also tend to emulate familiar public “celebrities,” so that many children come to know a lot about athletes, pop-stars and actors, but few can recognize the name of a single philosopher, scientist, or mathematician, because such achievers are rarely mentioned either in classrooms or media. The images of those celebrities must have substantial effects on our children’s goals—yet those descriptions are mainly fictitious, crafted by publicists to grip our children’s attentions for countless thousands of valuable hours. And even when those biographies are accurate, they don’t often demonstrate qualities that we should want our children to admire.

For example, consider the extent to which athletic prowess depends on speed and strength—which mainly come from inherited genes. Of course, sports also entail clever strategies—but even in the mental realm, the sport-stories in the media tend to emphasize injunctions like Try Harder, Don’t Quit, and Use Guts and Grit—which emphasize raw doggedness instead of skill and resourcefulness. Similarly, entertainment-scripts tend to glorify leading-role characters whose successes stem from using deceptions or exploiting attractive appearances. More generally, our popular cultures tend to venerate warriors over intellectuals, and rarely depict productive careers. The trouble is that some values acquired in childhood can remain in our minds for the rest of their lives.

We often assess a school’s qualities in terms of how many of its students go on to get college degrees. But what of the “drop-outs” who depart from that track? We tend to attribute those “academic failures” to deficiencies in persistence, talent, or self-esteem—or to lack of “scholastic aptitude,” or to not having enough “intelligence.” However, I suspect that it often might be more productive to attribute the trouble to the self-models those children have built. For consider how hard it would be for young children, by themselves, to invent good ideas about themselves; instead, it is far more convenient for them to absorb such ideas from other people they happen to meet. Therefore, we need to be concerned about which of those acquaintances will most influence our children’s ambitions, goals, and future roles—because those are the persons who, by default, will be the main sources of those children’s self-images. How do children acquire their goals and ambitions?

You’re almost always pursuing goals. Whenever you’re hungry, you try to find food. When you sense danger, you strive to escape. When you’ve been wronged, you may wish for revenge. Sometimes you’re aiming to finish some work—or perhaps seeking ways to escape from it. We have a host of different verbs for desiring, such as try, strive, wish, want, aim, and seek—but we rarely ask ourselves questions about why some goals seem strong and others weak, what decides which goals should be active when, what processes govern how long they’ll persist, what happens when several ambitions conflict, and what makes some goals get "too strong to resist”?

(As for what goals are and how they work, I like the explanations offered by Allen Newell and Herbert A. Simon in the early1960s. See the summaries of those ideas in sections 2-2 and 6-3 of The Emotion Machine.)

How do people acquire their goals? Most 20th century theories of learning assumed that each animal begins with certain ‘basic instincts’ that determine what that animal “wants.” Later, that animal may learn to connect new sub-goals to those instinctive ones, using processes through which those connections get “reinforced” by rewarding success. Some educational schemes have been based on those theories, but I don’t see this as a sound approach, because those old ideas about learning were mainly based on the behaviors of pigeons, dogs, and rats—whereas it seems clear that humans evolved additional levels that go far beyond the “reactive” thinking of animals.

For examples of those higher levels, consider that you often ”deliberate” before you react, by first envisioning some alternatives, and going on to evaluate them. Furthermore, you frequently think “reflectively” about your previous mental activities. Finally, you may even go on to “self-conscious reflection” about whether what you have done (or are about to do) is in harmony with what you call your morals, values, or ideals; this level considers not just what one wants, but also what a person ought to want. So clearly, those simple, reward-based theories of how most animals learn do not equip us for the task of educating human beings, whose minds have all those additional levels.

In any case, we need to learn more about how each person develops their values and goals. To what extents does this happen by chance or result from making deliberate choices? How much of our children’s desires and beliefs are shaped by their inherited genes, and how much by the cultural memes that pervade that child’s community? Chapter 2.3 of The Emotion Machine conjectures that we mainly acquire our values from the persons to whom we become “attached.” Long ago, an outstanding psychologist recognized this:

"Now since shame is a mental picture of disgrace, in which we shrink from the disgrace itself and not from its consequences, and we only care what opinion is held of us because of the people who form that opinion, it follows that the people before whom we feel shame are those whose opinion of us matters to us. Such persons are: those who admire us, those whom we admire, those by whom we wish to be admired, those with whom we are competing, and whose opinion of us we respect.”— Aristotle, in Nicomachean Ethics, Book VIII

This suggests that our high-level values are formed in ways that depend on the feelings of pride and shame that come when we receive praise or blame from certain persons. But strangely, we have no conventional name for those particular acquaintances—so we need to introduce a new word!

Imprimer: An Imprimer is one of those persons to whom a child has become attached. For animals that raise their young, the function of infant attachment seems clear: remaining close to parents helps to nourish, teach, and protect their offspring. However in humans, attachment has another effect—on the children’s ultimate values and goals; when your Imprimer praises you, you feel a special thrill of pride that elevates the goal of your present activity to have some kind of higher priority. Similarly, a sense of shame depresses your present goal’s desirability.

When you meet someone whose work you admire, you might want to acquire their skills—but in the case of an imprimer, you may also want to acquire their values, and more generally, to want to become more like that person! It seems clear that our children’s values and goals are greatly influenced by those of their imprimers—who are likely to include some of their parents and teachers, as well as some of their classmates and friends. Also, of course, their values and goals may be affected by their encounters with other people—and even by the fictitious heroes and villains they read about. And because all these may influence which kinds of ideas each child will like, our thinking about education must be especially concerned with the attachments our children will make.

What kinds of reasoning do children use to answer the questions they ask themselves? How do they classify the situations they face, and which ways to think will they use in each? What kinds of evidence will they accept or reject, and what will they see as succeeding or failing? We often hear suggestions that children should play a large role in deciding what they should learn—and of course, it will be hard to teach any subject that they have no interest in. However, few children will have enough knowledge to make good decisions about such things, so we look to teachers to provide such help, but the economics of typical schools leaves almost all teachers with too much to do. As for the child’s parents and friends, few of them will have the required skills—so most children will need other mentors.

Dictionary entries for “mentor” mention these kinds of qualities: “A trusted counselor who provides advice and support to, and watches over and fosters the progress of, a younger, less experienced person. Someone you trust, respect and admire—and has a direct interest in the your development.” Again, a very large order—and we can add to it more: a good mentor must also be a Tutor, to teach the factual content of a field—and should also be a Coach, to train up the needed set of skills—and also should be a Role-Model who can inspire a good set of values and goals.

However, perhaps most important of all is for a mentor to teach the postponement of pleasure—that is, paradoxically, to enable you to enjoy the discomforts that come during any significant exploration. Also, good mentors should not just reward you, but should help find ways to learn from your failures, because on one’s expertise comes learning “what not to do” —that is, from learning to know (and then to avoid) the most common hundreds or thousands of bugs! See the link to Negative Expertise" at http://web.media.mit.edu.

Perhaps the most valuable assets of a culture are its potential specialists—so we should put high priority on helping children pursue their interests. When I was a child, adults often said “you shouldn’t be so serious” or “take it easy and relax” instead of encouraging commitment and intensity. Such injunctions sometimes also come from other children—and in the microcultures of many schools, well-focused children get called ‘geeks’ or ‘nerds’—and become the targets of bullies. (Some of those nerds become scornful of ‘regular’ or ‘normal’ kids—and see them as victims of a disorder that might be called “insufficient obsessiveness.”) All these are problems our schools should face—but also are ones for which networks can offer escapes.

Finding Mentors in Network Communities

Of course we want to improve our schools, but each teacher can only do so much. However, our new interactive networks offer new ways to freely connect young people to other kinds of mentors and friends, who can offer new opportunities to enter new kinds of communities. Furthermore such networks could help many children to escape into more mature cultures and environments—options with special values to children who would otherwise be confined to grow up in noxious local neighborhoods.

This is also important because no small school can teach all possible subjects, or serve the needs of individuals who have atypical abilities. If a child in a small community develops some specialized interest, that child is unlikely to find any local help. The same applies to children with unusual disabilities; no small community can afford the range of resources available to the worldwide network. But as we develop more global connections, it will become more possible to find others with similar interests, and all sorts of specialized services.

Today, our networks are rapidly growing while, at the same time, the world-population is rapidly aging—and this could be a huge new source for mentoring. Soon we’ll have hundreds of millions of retired (and child-free) persons, and this will include great numbers of wise and experienced ones who’ve been left with more than enough “spare time” to mentor great numbers of children. Memo 2 discussed some problems that come from having mainly having same-age friends—whereas huge networks of people of varying age will enable young students to interact (and even become apprenticed to) older persons with wider experience.

How can we make productive connections to those millions of faraway mentors? This already happens spontaneously in many thousands of special-interest groups that exist on the World Wide Web. One trouble is that a good many such groups have had cycles in which they flourish, and then deteriorate—so we need to find ways to stabilize them. Perhaps we can get some ides about this from the websites that succeed at helping consumers with shopping and dating: could we modify that software to make systems that help children to shop for appropriate mentors?

Of course, such suggestions will raise concerns about protecting kids from predators. Clearly, some such problems exist—and many sites already try to prevent adults from joining juvenile groups. However, too much emphasis on safety would reject too many potential mentors, so we should be concerned about the cost-ineffectiveness of hyper-excessive protectiveness.

I’ll welcome all suggestions from readers. Please email them to minsky at mit.edu.