It's a Jungle in There: How Competition and Cooperation in the Brain Shape the Mind (33 page)

What mechanism underlies this differentiation? Whatever mechanism it is, it must have the property that it doesn’t allow concepts to leapfrog. The mechanism can’t allow for the formation of a more advanced concept before its predecessors take hold. In addition, the mechanism must take time to work—as long, in fact, on average, as observed rates of concept formation. The rates at which concepts are formed occupy a range of values that must never be so high that 8-month-olds can do tensor calculus or design integrated circuits. In this sense, cognitive development, like species development, is evolutionary rather than revolutionary.

A hint about the neural mechanism that underlies concept development comes from the other end of development, when the neural system falls apart rather than comes together. In a syndrome known as
semantic dementia
, which is seen mainly but not exclusively in the elderly, there is a breakdown
of concept knowledge. A key feature of semantic dementia is that the later a concept is acquired in early life, the earlier it’s lost. Patients with semantic dementia forget what distinguishes dachshunds from beagles before they forget what distinguishes dogs from cats, and they forget what distinguishes dogs from cats before they forget what distinguishes mammals from birds.
¹⁵
This feature of semantic dementia suggests that the most basic concepts are welded into place more firmly than are the concepts that are less basic. This feature of concept learning is decidedly Darwinian in the sense that, over the course of evolution, adaptations that prove useful in a very wide range of circumstances—for example, wherever gravity exists—tend to be preserved.

The evolutionary nature of concept learning has been noticed by others. An influential developmental psychologist, Robert Siegler, of Carnegie Mellon University, advocated a Darwinian view of cognitive development. In the Preface (p. v) of his book on this subject,
Emerging Minds
, Siegler wrote:

In both evolutionary and cognitive developmental contexts, adaptive change seems to require mechanisms for producing new variants (species or ideas), mechanisms that select among the varying forms available at any given time, and mechanisms that lead to the more successful forms becoming increasingly prevalent over time. This analogy leads to such assumptions as that children will generally think about any given phenomena in a variety of ways, rather than only having a single understanding; that they will choose adaptively among these alternative understandings; and that their thinking will change continuously and become increasingly adaptive over time. These assumptions prove to have large implications for the questions we ask about children’s thinking, the methods we use to study it, and the mechanisms we propose to account for changes in it.
¹⁶

Siegler was mainly concerned with higher-level cognition, but his proposal can also apply to lower-level cognition, to perception, to action, and, potentially, to all the cognitive functions afforded by the brain.

After Siegler advanced his theory, another book came out that propounded the selectionist view that Darwin’s theory exemplified and that Siegler endorsed. This other book built on the neural network modeling approach that the book’s second author (along with others) had pioneered nearly two decades earlier.
¹⁷
The specifics of the approach needn’t be spelled out now. All that needs to be said at this juncture is that the model relies on neural or neural-style elements interconnected via excitatory and inhibitory links
whose strengths change with experience. In the model, the earliest-learned and most often-used connections are strongest, whereas the latest-learned and least-often-used connections are weakest. This feature of the model makes it a good candidate for explaining the feature of semantic dementia referred to above—that the most basic concepts turn out to be the most secure.
¹⁸

What makes this model especially exciting for understanding concept learning is that it can be used to simulate the formation of concept hierarchies. In addition, it can be sensitive to the ways the typicality of examples and the frequency of explicit propositions bollix the predictions of a strictly rule-based, tree-structure model. The same general neural-network architecture can also be used to model important features of visual perception, such as the word superiority effect (see
Chapter 6
), patterns of data for reaction times (see
Chapter 5
), and patterns of data for motor control (see
Chapter 7
).

Because the neural network approach is so powerful, you might wonder whether it already fills the bill as a general theory of cognitive psychology. In many ways, I think it does, though the model’s emphasis is on neural mechanisms or, in particular, on neural mechanisms that exemplify parallel distributed processing, or PDP, as it is known among cognitive psychologists. The core concept of the PDP model lines up with the idea advanced by Darwin and other selectionists that there is no hidden overseer in the brain who determines what should be done, when, and by whom. Instead, as Darwin would have liked, all the elements of the system are always active to varying degrees, with the strengths of the connections between them, both excitatory and inhibitory, changing in ways that tend to promote greater adaptation to challenges that arise.
¹⁹

The way I introduced concept formation was to say it was a topic I had neglected, and then I devoted a section to concept formation. I did so for a couple of reasons. First, I wanted to redress the imbalance between my relatively heavy coverage of lower-level aspects of cognitive psychology compared to my relatively light coverage of higher-level aspects of cognition. This imbalance reflects my own interests, which are especially focused on human performance and the cognitive control of physical action.
²⁰
Second, I couldn’t resist the low-hanging fruit provided by research on concept formation. Because concept formation has been modeled in a way that allows concepts to be differentiated based on experience, I wanted to share that result because of its similarity to the branching structures of Darwin’s “tree of life” (
Figure 15
).

FIGURE 15.
The “tree of life” diagram from Darwin (1859). Two branches of the tree are shown. The sole hypothesized origin is implied. It lies somewhere beneath the base of the picture.

The fact that the diversification patterns are similar for species and for concepts could, of course, be coincidental, but I think that’s unlikely. What’s more probable is that the same dynamic is at work in both domains. The notion that similar dynamics can be played out in different domains—in different scales of space and time and with different material substrates—is one of the core tenets of science. Scientists generally believe they’re on the right track toward grasping core principles when they see the same principles playing out in different arenas.
²¹

Darwin’s dynamic applies to many other facets of cognition that I didn’t yet cover in this book. Here’s a brief overview of some of them. The overview is incomplete, though I can assure you no topic I omitted was left out because I thought the jungle principle couldn’t accommodate it. I daresay every topic in cognitive psychology, when looked at from the standpoint of the jungle principle, can be understood with the principle.

Emotion wasn’t covered very much here, though of course it’s an immensely important topic. Likewise for affect and motivation, which include topics like hunger, thirst, aggression, addiction, and sex. Are those behaviors governed by processes that reflect inner Darwinism? I think so. Darwin himself analyzed the expression of emotion in terms of his own theory.
²²

Emotions play a pivotal role in experience, often accounting for why some memories survive and others don’t. Emotionally charged events tend to be
more resistant to forgetting than are emotionally neutral events. Frightening or life-threatening events activate the amygdala, a brain region that is strongly activated when emotions run high, and the amygdala plays an important role in memory consolidation.
²³
This is just one indication of how emotion might be approached and understood from the perspective offered here.

Regarding sex, as I said in
Chapter 2
, I have deliberately shied away from sexualized Darwinism here; that is, I have avoided a Darwinism that emphasizes mating and creation of progeny within the brain. I did so not out of prudishness, but instead because of my uncertainty about what mating and reproduction could be like in the nervous system. Reproduction may be clear enough in that there could be replicas of memories, each a bit different from others. The notion that there are multiple memories of the same basic event is often raised in theories of cognitive psychology—especially in connection with so-called instance theories of memory, where the main idea is that each and every instance (or at least most or many) are stored.
²⁴
Mating, on the other hand, may not apply, though one approach to concept formation relies on the mating analogy via genetic algorithms, as mentioned earlier. I have refrained from emphasizing this approach because I think it can distract from the broader selectionist perspective I have advocated. Neural mating in one form or another may occur, but I think the broader view put forth in this book is better advanced by remaining agnostic on this point.
²⁵

Altered states of consciousness also got short shrift in this volume, as they do in most books on cognitive psychology. Sleeping, dreaming, and otherworldly states shouldn’t be a sideshow to the study of mental function, especially considering that, as regards sleep, we spend a third of our lives in that state. Are altered states, or the predisposition to fall into them, understandable from the perspective of the jungle principle? I’d be surprised if they weren’t.
²⁶

There is one kind of altered state for which the jungle dynamic seems very clear: sensory deprivation. You might think that people floating in baths of warm water with their eyes closed and their ears plugged would enter a state of blissful tranquility. What actually happens is the opposite. These people hallucinate—hearing, seeing, and feeling things not there. Ghostly, self-invented stimuli come and go in an eerie glide. People deprived of sensory stimulation find the experience unsettling, to say the least. Unmoored from external reality, they feel themselves losing control and sometimes terrified. As a result, sensory deprivation is sometimes used for torture.
²⁷

What can be said about the effects of sensory deprivation from a theory that’s all about the ruckus of everyday life? Sensory deprivation may unleash
the metaphorical beasts within. Deprived of normal sensory input, they feud with one another, so to speak, in ways normally kept under wraps. When sensory input is severely limited, neural creatures freed from their normal duties strike out looking for new connections. Like kids with too much time on their hands, the result can be psychologically destabilizing.

Social cognition was also omitted from this book despite its obvious importance in everyday life.
²⁸
In recent years, there has been a surge of interest in this topic, with the most growth of interest occurring in the field of social neuroscience. An insight from social neuroscience has been that neuroscientific methods can be usefully applied to the study of social understanding.

One result from this field that I find particularly interesting is that an area of the brain becomes activated when people see other people. This area of the brain—the medial prefrontal cortex—is sensitive to other people as
social
beings. This brain site becomes activated when observers see people of reasonably high status but not when they see people of low status. A picture of a man wearing a suit and tie causes this brain area to light up, but a picture of the same man dressed as a vagrant does not.
²⁹

Does this sort of finding have anything useful to say about the inner jungle claim, and does the claim that it’s a jungle in there help in any way to edify social neuroscience? If social appraisal is important, one would expect neural circuits to evolve accordingly. Socially useful neural niches would develop because humans rely on social interactions for their survival.

Two other topics that deserve more discussion are consciousness and qualia. Consciousness is a topic that many people have written about and some have even said they’ve explained.
³⁰
Defining consciousness is a challenge, but there has been no shortage of definitions. For example, the following definition appeared in a letter to
Scientific American
: “At its core, consciousness is a term we use to refer to our common human perception that we exist, are aware of ourselves, and are aware of our being part of the environment with which we are interacting.”
³¹