The Flaw in Progressivism


Kieran Egan

Chapter 2: Part 1


Learning according to nature's plan



Like everyone else remotely involved with children and their education, Spencer observed that in "the household, the streets and the fields" (1911, p. 24) they are able to learn all kinds of things effortlessly, with eager pleasure, yet these same children often have great difficulty learning quite elementary things in formal educational settings. How to explain this puzzle? Why should children who learn to talk fluently, later find it so difficult to learn to read and write fluently or learn a second language as easily as the first? Why should children who rapidly become so easily initiated into the norms and values of one culture find it so difficult to accommodate to those of another culture later in life? Why should children who find it easy to learn the sometimes complex rules of games find it difficult to grasp simple mathematics?

Spencer believed that his studies of evolution, biology, and psychology had given him the answer. He confidently claimed "that the evolution of intelligence in a child . . . conforms to laws; and it follows inevitably that education cannot be rightly guided without knowledge of these laws" (1911, p. 23). When children fail to learn in schools, the fault lies in the methods of instruction or in the contents of the curriculum failing to conforming with the laws whereby children's intelligence and learning works.

The answer Spencer proposed was to devise methods of instruction, learning environments, and a curriculum that did conform with the underlying laws of children's learning and development. Once methods and curricula more hospitable to children's natural modes of learning were in place, their desire for knowledge would be released, and an educational revolution would take place.

The progressive movement in particular, but many others too, have been convinced of this idea, and the twentieth century saw immense amounts of time, energy, ingenuity, and money expended on trying to make learning in schools match children's spontaneous learning in household, street, and field. The Holy Grail of progressivism&emdash;to let the metaphors run free&emdash;has been to discover methods of school instruction derived from and modeled on children's effortless learning, and so bring about the revolution promised by Spencer and by progressivists throughout the twentieth century. Despite all the ingenuity, effort, and money, the revolution hasn't shown much sign of occurring.

I'm not in the business of arguing that significant improvements in general education are impossible, especially as the final chapter of this book indicates how we might sensibly move towards such a goal. I am in the business of showing why Spencer's progressivist prescription for education's problem hasn't worked and can't work. Ironically, one of the main road-blocks to real educational progress during the twentieth-century was the very wide acceptance of Spencer's general idea.

So what is the error? Well, it is more like two related errors&emdash;a bad one and a worse one. The bad one first. Consider Jerry Fodor's model of the human mind as composed of a set of fast and "stupid" input systems, each of which has its own dedicated processes and is focused on specific stimuli in the environment. The input system concerned with language learning, for example, attends to linguistic data and ignores other data, and it works so "stupidly" that we can't not learn a language, unless there is severe brain damage. These fast input systems then reformulate their data to make them accessible to the general&emdash;purpose central processor. In Fodor's model, then, we have input systems and a slower, more deliberative central processor. I will use Fodor's terms here as a short-hand way of indicating the bad error, even though there may be reasons to question Fodor's model, as Karmiloff-Smith (1992), among others, has argued. But his model is useful just to indicate why one might begin to worry about the "common-sense" objective of making children's learning in schools conform with their effortless learning in households, streets, and fields. What Spencer is requiring, in Fodor's terms, is to make the central processor work like an input system. It won't and can't. The century and more of attempts to make school-learning more like children's early effortless learning has been misdirected. Well, that's to oversimplify the argument, and the following pages will involve us in some complexifying.

The worse error I want to expose, in Spencer's writings and today, is connected with the common belief that children's minds have some preferred natural kind of learning and if we can isolate and understand it we can then make the educational process more efficient and effective. I will show why we should abandon this belief and look elsewhere for the key to enhancing the efficiency of children's learning.


Nature's paradigm of human learning

The year after Spencer's death in 1902, William James published in the New York Evening Post an essay&emdash;"Herbert Spencer dead"&emdash;that tried to give as favorable an interpretation of the Englishman's work as James could bring himself to write (cf. Lampert, 1997). Among his judgements of the massive Spencer oeuvre, James concluded that The Principles of Biology, The Principles of Sociology, and The Principles of Psychology "must soon become obsolete books" (James, 1978, p. 100). (Mind you, just a few years earlier James had been using the last as a textbook in his Harvard introductory psychology class (cf. Kennedy, 1978)). The Principles of Biology is rarely referred to now, except as a horrible example of what can go wrong if one builds complex theoretical edifices on factually flawed foundations. The Principles of Sociology does rather better, in texts about the origins of the discipline; Spencer's doctrines are considered obsolete, but he gets points for originality and for trying to bring social processes under some law-like survey. The Principles of Psychology seems, like a big stone rippling a pond, to have over time faded completely away.

Reading the two volumes of The Principles of Psychology today can be a tad dispiriting. Apart from the ubiquitous racism, the organization of all facts, and "facts," around the dogma that all the process of the changing world follow a path from a state of relatively indefinite, incoherent homogeneity to a state of relatively definite, coherent heterogeneity, is like watching a fairly clumsy conjuror constantly repeating the same trick with far too much worn-out equipment. But Spencer claimed that he was exposing, amongst other things, the fundamental nature of human learning.

My aim here is not to hammer Spencer for failing to have incorporated into his theorizing all the findings of late twentieth-century neurophysiology and cognitive science. Rather, I want to show that the principles so many educators still find attractive are as problematic as their source in his psychology.

Nature's plan, as Spencer outlines it, is to deliver a baby with just a few basic reflexes but with a power of learning, "which lies latent in the brain of the infant" (1897, vol. 1, p. 471), created and furnished by the long process of evolution. The simplest actions of the infant impinge on the environment and, from the reaction of the environment, the child begins to construct an understanding of the world. Let me give a couple of brief examples of Spencer's way of putting it:

The organism is placed amid immeasurable relations of all orders. It begins by imperfectly adjusting its actions to a few of the simplest of these. To adjust its actions more exactly to those few simplest, is one form of advance. To adjust its actions to a greater variety of these simplest, is another form of advance. To adjust its actions to successive grades of the more complicated, is yet another form of advance (Spencer, 1897, Vol. 1, p. 418).

Gradually we adjust our inchoate and incoherent senses to the environment, and as our perceptions become sharper and clearer and take in more of the environment, they lead eventually to ideas and adult understanding. Our early experiences lead to concepts which progress by differentiation and integration (1897, Vol. I, pp. 624-628.) Our internal sense of the world gradually mirrors ever more precisely the world outside, following "the law which Intelligence fulfills more and more the further it advances" (1897, Vol. I, p. 418):

Only by supposing such a law to exist can we explain the facts, that relations which are absolute in the environment are absolute in us, that relations which are probable in the environment are probable in us, that relations which are fortuitous in the environment are fortuitous in us (1897, Vol. I, p. 417).

Spencer does not try to chart in any detail how the operation of the basic law of learning he identifies in these simplest interactions between infant and environment reaches full adult understanding: "Everyone must . . . admit that the steps by which these simplest inferences of the infant pass into these inferences of high complexity drawn in adult life, are so gradual that it is impossible to mark the successive steps" (1894, Vol. I, p. 461). What is clear to Spencer, though, is that the same fundamental law applies equally from the earliest and simplest learning to the most mature and complex.

Some of Spencer's ideas are, of course, at least as old as Aristotle, and can be traced through Locke, Hume, and others. One of his distinctive themes, though, was the manner and degree of environmental responsibility for the conception of the world built up in the child's mind. This was one of Spencer's excesses noted by William James. The result of Spencer's study of psychology is an odd sense of the mind passively accepting the impress of the environment's regularities, while requiring an active organism to attract as many impressions as possible.

Armed with his law, and the acknowledgement that the successive steps of its gradual realization in each child was too complex to describe in any detail, what we can do, Spencer suggest, is observe the child's most characteristic forms of learning in the natural environment when no deliberate teaching is taking place. The trick then would simply be to "systematize the natural process" (1966a, p. 84) and adopt the methods of teaching that Nature has been perpetually thrusting upon us, "If we but had the wit to see it, and the humility to adopt it" (1966a, p. 84).

From his conclusions about the mind and learning, Spencer made those proposals, with which we are now familiar, for making what happens in educational institutions more like what happens in natural learning in the child's everyday environment: the child must be active not passive; learning occurs best through play in the early years; new knowledge must be connected with what children already know and will initially be concerned with the local, the concrete, and the simple; learning should be pleasurable and not forced; and so on.

(Eyes closed I reached out and took a book from my Early Childhood shelf, and spent a few minutes flipping through the pages; a pity it is an older one, but I didn't feel I should go back having used the eyes-closed rule. One finds such observations as "children have a natural inclination to explore and manipulate" (Hildebrand, 1981, p. 37); "The children become personally and actively involved in their own education. They are helped to find concrete ways of relating new knowledge to what they already know" (p. 80); "play is very valuable" (p. 38), with 16 particular values of play then listed; "Children are more likely to understand and remember information that they discover themselves" (p. 80); "learning takes place following sensory stimulation from people, things, and activity in the environment" (p. 172); and so on, including pervasive promotion of activity.)

Spencer's Principles of Psychology is spotted with examples of his conclusions applied to pedagogy, in the form of appealing brief scenarios of a kind still familiar in Education texts. Reflections on logic lead him to invite his readers to suppose "I am giving a child a lesson in Mathematics, carried on after that concrete method which teachers, were they wise, would habitually adopt as an initiation" (1888, Vol. II, p. 91). He sits with the child playing marbles. The child has fifty marbles and Spencer demonstrated that one can make squares of four marbles on each side and see that these squares have sixteen marbles in them. With fifty marbles one can make three such squares and have two marbles left over. Or one can make five sided squares, two of which will use up all fifty marbles. One can classify the marbles and form sets based on those that are chipped or have striped designs, and so on. Thus Spencer begins the child's understanding of mathematics in play with familiar and simple concrete objects, showing how "sensible experience" carries the child's mind "to the numerical truths" (1887, Vol. II, p. 91).

Of course this use of play as a paradigmatic form of natural learning is hardly new. Plato observed that one shouldn't "use force in training the children in the studies, but rather play. In that way you can also better discern what each is naturally directed towards" (Plato, The Republic, 537a). Plato also described this as a procedure practiced long ago by the Egyptians.

The ideal example of an educational practice that violated all Spencer's principles, but was still commonly used, was "the vicious system of rote learning" (1966a, p. 30). Rote-learning has remained for progressive educators the epitome of the unnatural, forced, and meaningless: "To repeat the words correctly was everything; to understand their meaning nothing" (1966a, p. 61), and it was a method that could operate only by "threats and bribes" (1966a, p. 61).

The related epitome of bad educational practice was to teach rules first and examples later, as in "that intensely stupid custom, the teaching of grammar to children" (Spencer, 1911, p. 50). "The particulars first and then the generalization, is the new method" Spencer declared, in the 1850s (1966a, p. 61). The student should also experience the particulars not simply be told them.

The educational principles he had outlined would, Spencer believed, overcome the horrors of the daily Press and the useless, ornamentally-educated minds that were prevented from seeing so much of what Nature was thrusting on their attention. Like Karl Marx in 1848 announcing the specter of Communism stalking Europe, before most people had even heard of it, Spencer announced the imminent universal success of the new principles of learning:

The forcing-system has been, by many given up . . . The once universal practice of the learning by rote, is daily falling more into discredit. All modern authorities condemn the old mechanical way of teaching the alphabet. The multiplication table is now frequently taught experimentally. In the acquirement of languages, the grammar-school plan is being superceded by plans based on the spontaneous process followed by the child in gaining its mother tongue (1966a, pp. 60/61).

Spencer might be saying the same today, as many do. Why, he and his modern followers wonder, cannot educators attend to the principles of learning exemplified so straightforwardly when children are active and at play in their everyday environment? Why, instead, do teachers become caught up forcing an artificial curriculum into children's minds using out-dated methods, making the learning painful and inefficient for the child and the teaching frustrating and inefficient for the adult?

Part of Spencer's success in stimulating a revolution in educational theory if not in practice was due to his rhetorical strategies. Apart from his own conviction, and his presenting learning as some kind of binary moral choice between the traditional, passive, forced, and vicious, and the progressive, active, effortless, and pleasurable, he constantly intersperses his arguments with attractive and plausible images of the new methods at work. In contrast with an image of frustrated and unhappy children toiling indoors over dull textbooks. Consider this typical example of Spencer's descriptions of his proposed methods at work:

Every botanist who has had children with him in the woods and lanes must have noticed how eagerly they joined in his pursuits, how keenly they searched out plants for him, how intently they watched while he examined them, how they overwhelmed him with questions (1966a, p. 86).

Who could be against methods that promised such scenes of eager learning? But consider, as something of an astringent drawn from reality, Beatrice Webb's account of just such an outing with Spencer. When Beatrice and her siblings were released from their governess' traditional classes and given over to Spencer's educative care, she exuberantly describes how "the philosopher found himself presently in a neighbouring beech-wood pinned down in a leaf-filled hollow by little demons, all legs, arms, grins and dancing eyes, whilst the older and more discrete tormentors pelted him with decaying beech leaves" (p. 49). This might indeed have been of more value in some ways to the children than lessons with the governess, but Webb's account of Spencer's pedagogy leaves one to wonder:

"We agreed [as children] with his denunciation of the 'current curriculum', history, foreign languages, music and drawing, and his preference for 'science'&emdash;a term which meant, in practice, scouring the countryside in his company for fossils, flowers and water-beasties which, alive, mutilated or dead found their way into hastily improvised aquariums, cabinets and scrap-books&emdash;all alike discarded when his visit was over" (p. 50).

Obviously it is unfair to use this example as typical of what happens when Spencer's principles are put into practice. But the gap between the predictions of what will result from engaging children's "natural" learning and the result from trying to do so in practice has been consistently very wide. To committed progressivists, it is also something of a puzzle, and each generation who recommends much the same principles has to give a new explanation why their predecessors failed to produce the promised results in the past. Spencer himself took many of his principles from Pestalozzi. Spencer acknowledged that the Pestalozzian system held great promise but "we hear of children not at all interested in its lessons&emdash;disgusted with them rather" (1911, p. 55). How to explain this failure? Even with "the choicest tools, an unskilled artisan will botch his work" (1911, p. 56). It was the teacher's fault&emdash;a reason we have heard again and again for the gap between what ought to happen and what does. The other common excuse is that the previous generation got "nature" wrong&endash;&endash;they misunderstood the nature of learning or the nature of development, but an ever-advancing science will, or has, exposed more precisely or clearly or correctly the real nature of learning or process of development&endash;&endash;which is then used to justify the same progressivost scheme.


What is wrong?

Derek Bickerton bemusedly concludes that "for a species that spends so much of its time thinking, we don't have much of a handle on the subject" (1995, p. 57). We might say the same about "learning," while adding that we don't have much of a handle on how to distinguish the things we call "thinking," "learning," and "development." When studying infant cognition, "learning" and "development" overlap and intermingle. I mention this because I will introduce my single point about learning by stealing a few paragraphs that might equally well appear in the following chapter on development.

Spencer's homogeneous to heterogeneous law requires that human beings start life outside the womb with virtually nothing in the way of mind and then gradually learn from the environment's tutelage the array of knowledge of the typical adult. In his scheme, we start from "low and vague beginnings" (1897, Vol. I, p. 627), when our thinking is "simple, vague, and incoherent" (p. 617); this general condition is evident, he assures us, to anyone who has observed the "vacant stare of the infant" (p. 617). Perhaps the most famous exposition of this notion is William James's description, in his Principles of Psychology (1890) of infants perceiving only a booming, buzzing confusion. (If vagueness and confusion are the prime criteria of intellectual development, I fear that I'm heading in the wrong direction; "booming, buzzing confusion" seems an increasingly good description of the state of my mind.)

No-one familiar with research on infant cognition during the past quarter century or so can take Spencer's views seriously. It is precisely the infant's intelligent staring at particular features of their environments that has provided a tool to investigate their early learning. In the next chapter, I will detail some of the cognitive competencies of infants, which decisively undermine all those schemes based on babies' supposedly blank minds. What is particularly interesting, for my present purpose, is that the accumulating evidence of infants' surprising range of knowledge and intellectual competence should not be seen as evidence of general cognitive abilities. That is, what infants can do, while surprising in light of millennia of theories that asserted the "blank slate" or empty condition of the mind at birth, is nevertheless limited and quite precise. As Howard Gardner puts it: "Try to get an infant to recognize faces upside down, or a toddler to speak a language which does not make phonemic distinctions or which requires that the child attends to every other word. You will soon discover the powerful, specific constraints on cognition in Homo sapiens sapiens" (1997, p. 26). That is, effortless ease in one kind of intellectual act goes together with incapacity in acts that can look very similar.

How are we to explain this? My purpose is to show that the process Spencer described, and on which he then based his central claims about methods of learning and teaching, is wrong. And I want to focus on its wrongness in such a way as to raise problems for the methods of learning and teaching that remain largely taken for granted today.

So I want to emphasize the irregularities, the peculiarity, and the differentiations in forms of human learning. The mind we have inherited from our evolutionary history is not one whose accumulation of knowledge and understanding proceeds in some gradual, regular, spontaneous, and undifferentiated way. This should be easier to recognize now as it is generally accepted that our evolutionary history is also quite unlike the gradual process Spencer imagined. Rather, in the process of our evolution we have experienced "radical reengineering of the whole brain" (Deacon, 1998, p. 45). Our brain is like the product of an ingenious bricoleur, a scavenging odd-job tinkerer, reshaping and reusing old parts for new purposes, adding bits, attaching new to old bits, gerrymandering a functional, occasionally inspired, occasionally clumsy, organ. As Merlin Donald puts it: "modern human cognitive architecture is highly differentiated and specialized . . . the modern mind is like a mosaic structure of cognitive vestiges from earlier stages of human emergence" (1991, p. 2/3), and it has gone through "a series of radical evolutionary changes . . . rather than a continuous or unitary process" (Donald, 1993, p. 737); see also Bickerton, 1990; Lieberman, 1984; Plotkin, 1988).

The most straightforward challenge to the belief that we have a single, undifferentiated learning process comes from theories such as Fodor examines, which I mentioned earlier. That the mind deploys some different modes for learning different things, particularly early in life, is an idea whose current prominence owes much to Noam Chomsky's work. Influential in the crumbling of behaviorist beliefs about learning was Chomsky's argument that children's experience of language was too fragmented and sparse for them to build up the complexity of regular syntactic structures evident in the average three or four-year-old. Also one can observe patterns in children's language-learning that make no sense if one assumes their learning is shaped entirely by their environment. They make some mistakes&endash;&endash;like saying "I seed your feets"&endash;&endash;that are a result of misapplying grammatical rules, but are not learned from their environment as no adult says such things. And in learning certain forms of language, they virtually never make mistakes. For example, if one child has a single doll and another child has three dolls, and you say "Give me a doll" or "Give me the doll," even three year olds will not be confused whom you are addressing.

Chomsky argues that as these features of language-learning cannot be accounted for in classical theories, they must be explained by something else. The best candidate for "something else," he and others have argued, is a part or parts of the brain "pre-programmed" for language-learning. That is, we don't simply run some general learning program that we use to master language and then turn to learning algebra or phone numbers or faces or names or what hairstyle is fashionable. Language and faces seem to require very little effort. It is as though experience acts less as the whole source of such learning and more as a trigger with minimal information that kicks the program into action. Similarly, learning faces just isn't, notoriously, like learning names, or phone numbers, even though the last two contain much less information than the first.

The power of our predisposition to learn language is evident in the cases of deaf children who have no signing parent. Such children invent visuomanual systems that display several of the features of natural languages (Goldin-Meadow & Feldman, 1979; Feldman et al. 1978). The relative specificness of language centers in the brain is indicated by results of localized brain damage in deaf adult signers. They can imitate the physical movements involved in a sign but have become incapable of using such signs for linguistic purposes (Poizner et al., 1987). Karmiloff-Smith suggests that such findings support the theory that we possess a "domain-specific, innately guided process that can get language acquisition off the ground even in the absence of a model . . . And when a linguistic model is available, young children are clearly attentive, not to some domain-general input, but to domain-specific information relevant to language" (199 , p. 38).

For many cognitive psychologists, the specialization of brain functions is no longer open to dispute. As Howard Gardner sums it up: "recent research on early infancy provides the strongest clues to the inherent modularity of human cognition" (1997, p. 26). He goes on to make the point that: "Empirical evidence shows that the mind&endash;&endash;human or prehuman&endash;&endash;is distinguished precisely by the fact that it does not . . . harbor all-purpose rules or operations" (p. 26). Gardner's own work on Multiple Intelligences is, of course, a vivid indicator of diversity in human learning (1983; 199x).

Whether modularity is the best account of the evident differentiation in human forms of learning remains to be seen. It would have been convenient to wait fifty years before writing this chapter so that greater definition of the neurophysiological evidence might be given (but my dependants are howling for pearls and caviar). I am not entirely convinced by the modular thesis, at least in some of its versions, and find Karmiloff-Smith's account in Beyond Modularity (199 ) more attractive in some regards. (I've also cut out chunks of refined qualifications about just how extensive are the differences between, say, learning a first language and later learning literacy, but I'll put them on my www Home Page). The overall accumulation of evidence at present, however, at least establishes that human beings do not have a single mode of learning whose character has been captured in Spencer's principles.

So Spencer's Nature, which has been urging sound methods of teaching on us, is obviously itself a capricious pedagogue. It puts immense investment into ensuring we learn a language, track moving objects efficiently with our eyes, classify flora, and so on, but seems to care hardly at all about our mastering irregular Latin verbs (unless we are Roman children long ago) or basic algebra. For many educational thinkers, it is the artificiality of schooling tasks that is the problem. Learning ought to be "natural"; it is one of the things we have evolved to do superbly well. Spencer and most modern educational researchers concerned with learning believe that improvements in school learning will follow understanding the nature of human learning. Once we get that clear or clearer, we can devise methods of teaching, a curriculum, learning environments, and so on, that will better conform with how children naturally learn.

So far I have been able to make my argument arm in arm with a host of researchers, but I fear the following section is going to be a fairly lonely hike, unless you join me. I want to show that we have inherited from Spencer an even worse error than his belief that he was able to identify some general form of natural learning from children's effortless learning in fields, streets, and homes, and that this natural learning could be used as a basis for all forms of teaching&endash;&endash;hence his seven principles of the previous chapter. He also passed on to us, partly derived from Rousseau and partly from his positivistic science, a belief that the mind can be studied much as the body is studied. The scientific study of learning that he, among others, promoted, remains in the forefront of methods of research today, and I will show why we might sensibly conclude that it, too, is mistaken.




Return to Home Page

Go on to Chapter 2: Part 2