Models Of Cognitive Development

Ken Richardson

• Publisher: Psychology Press
• Year: 2019
• ISBN:
  9781000228014
• Edition: 1
• Topic: Psychologie
• Subtopic: Histoire et théorie en psychologie

In spite of its obvious importance and popularity, the field of cognitive development remains highly fragmented due to the vast diversity of models of what knowledge and reasoning are, and how they develop. This new Classic Edition of Models of Cognitive Development aims to overcome this barrier through its careful introduction, illustrated examples, and approach to helping students think more critically about the subject.

In this significant work, Richardson provides students, researchers, and comparative theoreticians with a cohesive understanding of the area by organizing diverse schools, frameworks, and approaches according to a much smaller set of underlying assumptions or preconceptions, which themselves can be historically interrelated. By understanding these, it's possible to find pathways around the area more confidently as a whole, to see the "wood" as well as the theoretical trees, and be able to react to individual models more critically and constructively.

The Classic Edition of this core text will be essential reading for undergraduate and graduate students of cognitive development.

1
Nativist models

Introduction

Whether in the living world or in that of artefacts, we are generally very quick to assume that almost any organised entity has predetermined form and structure. We see a building slowly taking shape, and automatically infer some basic plan, an agent who drew it, and the resulting structure as the gradual realisation of that plan. We see a complex blossom unfolding from its bud, or an infant from its embryo, and infer that the form must have been there all along in seed or egg, and its development is an automatic materialisation. In the case of physical structures, this view has been around for a very long time. During the exciting early period of microscope development in the 15th–17th centuries, biologists were quick to “see” all manner of structures in their “preformed” states, and the microscopist Hartsoeker duly provided us with a drawing of a little man — a homunculus — with all its little parts, curled up in the head of a sperm (Fig. 1.1). The added implication, of course, is that such homunculi have somehow been copied and passed through the genera tions in the sperm since the beginning of time.

Fig. 1.1. Hartsoeker’s drawing of a homunculus in the head of a sperm.

Much the same views have arisen about the origins and developments of our cognitive abilities, as in knowledge and reasoning. The very complexity, scope, and coherence of these seem strongly to bespeak preformed or predetermined plans, and more or less automatic unfolding. And here, too, such views have a very long history. Although spoken with a modern — even scientific — conviction, they were heard over two thousand years ago in Ancient Greece, when the philosopher Plato argued that knowledge could not possibly be derived from experience, which is unreliable, fragmentary, and incomplete. Instead he said it must be there from conception, put there by the hand of God, and “brought out” by a proper education based on rational discourse (the theory became known as classical rationalism). Over the centuries these Platonic views became expressed and re-expressed by generations of philosopher–psychologists. Indeed, Leibnitz (1646–1716) (quoted in Dobzhansky, 1962, p.25) provided the mental parallel to Hartsoeker’s homunculus when he said that all “the souls which will one day be human souls have been present in the semen of their ancestors down to Adam, and consequently existed since the beginning of time.”

These basic ideas (in various guises) still form the most popular models of human cognitive development. There have been some adjustments, of course. The most significant of these is that the genes have replaced God as the fundamental “designer”. There have also been various other changes in terminology: in particular, modern theorists tend to use the label “nativist” rather than “rationalist” to describe such models. However, the basic principles survive. In this chapter I want to offer an overview of such models, hopefully in a way that will reveal those common basic principles and assumptions. Then I will try to show how the models have been questioned and criticised through questions and criticisms of those assumptions.

Contemporary nativism

Nativist models of cognition and its development have once again become very popular in the last 20 years, and their role in reinstating a genuine, “mentalistic”, psychology is well known (Gardner, 1984). Around the turn of the 20th century, psychology had “lost its mind”. This was due to a “positivist” philosophy, popular in the second half of the 19th century, which insisted that psychology should emulate the physical sciences, and should view the control of behaviour (rather than modelling of mysterious “mentalisms”) as its primary goal. Behaviourism, as it came to be called, insisted that we should ignore unseen processes and focus only on observable behaviours, and the purely statistical rules governing the connections between stimuli and responses (see Chapter 2).

Behaviourism dominated psychology between the wars, but afterwards created increasing dissatisfaction. This, in turn, was due to the gathering evidence for the role of “central processes”, mediating and moderating relations between stimuli and responses (e.g. Hebb, 1949). In the late 1950s and 1960s psychologists once again dared to speak of minds and mental processes, and cognitive psychology was “reborn” (although, in truth, it had never been completely dead). Renewed nativist theorising enlivened this rebirth in several classic forms of expression.

The first expression entailed the re-assertion of predeterminism, and appeared first in Noam Chomsky’s theory of language (Chomsky, 1959). In pointing out that most of what a child hears in everyday language experience is highly diverse, faulty, and piecemeal — that a child actually produces numerous novel sentences that he or she has not heard before — and thus utterly rejecting the idea that the child learns grammar by imitating sentences heard, Chomsky returned to the idea that all the child’s abilities in grammar must be inborn, in the form of a set of innate rules coded in our genes and biologically inherited.

This entailment was based on what has long been known as the argument from the poverty of the stimulus, first put by Plato in Ancient Greece. As with Plato, the essential point is the seeming incredibility of any suggestion that complex, organised mental structures and reliable knowledge can be condensed from unreliable, changeable sense experience. A distinction between perceptual processes and cognitive ones is necessary: there is typically more information in our thoughts and responses than in the perceptual input, which must involve the contributions of the mind itself. According to Chomsky (1980, p.39):

Innate factors permit the organism to transcend experience, reaching a high level of complexity that does not reflect the limited and degenerate environment … we find structures of considerable intricacy, developing quite uniformly, far transcending the limited environmental factors that trigger and partially shape their growth.

Noone argues that our physical organs and their functions come from experience, so why should we imply this for language and cognition, asked Chomsky (1980, p.33).

We may usefully think of the language faculty, the number faculty and others, as “mental organs”, analogous to the heart or the visual system or the system of motor coordination and planning. There appears to be no clear demarcation line between physical organs, perceptual and motor systems and cognitive faculties in the respects in question.

And, thanks to biological natural selection (of which more later) these “organs” are common to humans, presenting universals of structure and form in cognition. For example, Chomsky has detailed aspects of language that are found in all languages all over the world, developing in children with remarkable uniformity, and he extends this idea to cognition and its development. How could such uniformity of structure arise from the disparate experiences of children all around the world, he asked.

The metaphor of unfolding physical organs thus became a compelling model of cognitive development itself. As Chomsky (1980, pp.33–39) states:

My own suspicion is that a central part of what we call “learning” is actually better understood as the growth of cognitive structures along an internally directed course under the triggering and partially shaping effect of the environment … Our biological endowment permits both the scope and limits of physical growth …When we turn to the mind and its products, the situation is not qualitatively different … here too we find structures of considerable intricacy, developing quite uniformly, far transcending the limited environmental factors that trigger and partially shape their growth …

Of course, these “organs” require environmental substances, conditions, and stimulation for them to materialise: but “the environment” in such models, formal or informal, only works as a “trigger” or “fine-tuner” — the medium of activity and exercise, assisting or retarding development, but not substantially altering its course and final form. The evidence for a developmental “plan” or “genetic program”, whose expression is buffered against external perturbations, seems insurmountable. According to Gardner (1984, pp.56–57): “The plan for ultimate growth is certainly there in the genome; and, even given fairly wide latitude (or stressful circumstances) development is likely to proceed along well-canalised avenues.”

These are the general principles of modern nativism. In what follows I offer some slightly more detailed models: in basic information processing using the computational metaphor; in the flow of Chomskian and information-processing ideas into models of more complex aspects of cognition; and how these have more recently been laced with biological (evolutionary) principles to produce another view that has become popular in recent years, that of the “modular” mind.

Natural mechanics and natural computations

Notions of knowledge as parcels of experience, and of reasoning as the logical rules by which these are moved, shuffled, interrelated, converted, or combined to produce more complex parcels, are old ones, and the idea of “natural mechanics” pre-dates the recent resurgence of nativism. Thus, Gruber (1974) notes how Charles Darwin had to theorise about the evolution of complex behaviours in the teeth of Natural Theology, which had declared all complex behaviour as endowed by God, and deterministic machines such as clocks, music boxes, and Babbage’s calculating engine (a famous Victorian computer), as the best “natural” models of cognition.

The advent of the modern computer, which operated even more explicitly for the manipulation of parcels of information, according to pre-programmed rules, thus furnished a potent metaphor for cognitive science in the 1960s and 70s. This signalled a transition from natural mechanics to “natural computations”. In the modern computational models, parcels of information, their more detailed fragments, and their more general composites, are identified and labelled as “symbols”, and then sorted, stored in memory, retrieved for problem solving in a “working memory”, and so on.

By the 1980s, the computer metaphor had inspired a “standard” model of cognition. According to Vera and Simon (1991, p.34): “Sensory information is converted to symbols, which are then processed and evaluated in order to determine the appropriate motor symbols that lead to behaviour”. Gilhooly (1995, p.245) states:

In the standard model, thinking is seen as the manipulation of symbols both within working memory and between long-term and working memory. “Manipulations” of symbols are viewed as consisting of elementary information-processing operations such as comparing symbols, copying symbols, deleting and adding symbols; appropriate combinations of elementary operations may form complex manipulations. The manipulations are presumed to be in accord with rules stored in long-term memory.

A simple information-processing model is shown in Fig. 1.2. Information is input at sensory registers and encoded in some form (usually in symbolic form, such as a letter, word, or image). It is conveyed to short-term or working memory where it is analysed, compared, transformed, or whatever, according to various processes and rules, at least some of which are retrieved from long-term memory. The result of this activity is some motor action (including, possibly, some more active search of input, some strategy for gathering further information, and so on), which creates further sensory input, and so the process continues, until some criterion state has been reached.

Before I go any further it is important to point out that the “cognitive revolution” of the 1960s encompassed almost all of cognitive psychology, including cognitive associationism, which I deal with in the next chapter, and cognitive constructivism, which I deal with in Chapter 3. A number of contemporary models stress learning, development, knowledge-acquisition, and even self-modification (Siegler, 1991), as aspects of developmental change within an information-processing framework, and I will be returning to such aspects in Chapter 2. Here I deal only with nativist aspects of cognition as commonly arising in models of information processing.

Fig. 1.2. The main components of a simple information-processing model.

The point is that the use of nativist assumptions in information-processing models is not always clear. On the one hand, as Bates and Elman (1993, p.628) point out, the computer metaphor of the mind is “by its nature a strongly nativist approach to learning … Everything that really counts is already there at the beginning.” On the other hand, we cannot be too categorical about this, because many information-processing models are implicitly, rather than explicitly, nativist, and many adopt nativist elementary processes themselves as the basis of learning (a “dual” system, which I will be returning to in Chapter 2).

Quite simply, nativism has entered into recent models whenever it has been assumed that, in order to carry out cognitive functions, a number of elementary and more complex processes, as well as some overall controlling or executive function, or even aspects of knowledge itself, are preformed for them. In the strongest expressions, some or all of these processes — encoding, comparing, combining, transforming, mapping, inferencing, and even the “rules” — are “built-in” aspects of our cognitive architecture, present from birth, and functioning like that throughout life. We may describe these aspects in various ways but “Ultimately”, as Klahr (1992, p.151) puts it, “we are forced to make assertions about what I have called the innate kernel” (emphasis in original). Here I will attempt to illustrate how such kernels have been conceived, while describing their implications for cognitive development as we go along.

Such an innate kernel has been hinted at in many models. For example, although the well-known ACT* model of Anderson (1983) is essentially a “learning” model, he nonetheless acknowledges a number of innate principles and processes underlying its operation. These include a set of rules for identifying causal relationships in experience; a set of inductive principles for constraining the causal inference; and a set of other methods (such as using analogies) for applyin...

Table of contents

• Cover
• Half Title
• Series Page
• Title
• Copyright
• Contents
• Preface to the classic edition
• Preface
• 1. Nativist models
• 2. Associationist models
• 3. Constructivist models
• 4. Sociocognitive models
• 5. Models mixed and models new
• References
• Author Index
• Subject Index