Chapter 10
Morphic Resonance in Human Learning
People, like other animals, usually learn what other members of the species have learned before. Most human behaviour is habitual. By morphic resonance, learning something that has been learned before should be easier than if it had not been learned before. The more people that have learned it, the easier it should become.
The acquisition of physical skills
Generally speaking, we learn things from people who can already do them. In the case of physical skills, such as swimming, bicycle riding or violin playing, the skills are transmitted to us by people whom we imitate. In this transmission words play a subordinate role: it is notoriously difficult to learn such skills from books. By beginning to do these things we tune in to the morphic fields of the skills, and our learning is facilitated by morphic resonance not only with our teachers, but also with many other people who have previously practised these skills.
In the 1980s, neuroscientists discovered that when animals watched other animals doing something, for example a monkey peeling a banana, changes in the motor part of their brains mirrored those in the brains of the animals they were watching. These responses were described in terms of ‘mirror neurons’. But this term is misleading if it suggests that special kinds of nerves are involved. Instead, it is better thought of as a kind of resonance. In fact, Vittorio Gallese, one of the discoverers of mirror neurons refers to the imitation of movements or actions by another individual as ‘resonance behaviour’.1
In traditional societies, a great variety of skills have remained more or less the same for many generations – for example those involved in hunting, cooking, agriculture, weaving, and potting. The same is true of the traditional crafts such as those of blacksmiths, goldsmiths, and carpenters. Even in our own society, most trades are still learned through a system of apprenticeship. All such skills, according to the present hypothesis, involve nested hierarchies of morphic fields which are strongly stabilized by morphic resonance from countless people in the past.
In the case of long-established skills it would be difficult to investigate empirically the role of morphic resonance in the facilitation of learning. Even in skills of relatively recent origin, such as bicycle riding, there are no objective data that would enable us to compare average rates of learning of children today with, say, those at the beginning of the twentieth century. According to the hypothesis of formative causation, other things being equal, the rate of learning should be faster now than it was then because many millions of people have learned to ride bicycles in the intervening period. And indeed anecdotal evidence suggests that children do tend to learn to cycle more readily than they used to. But of course other things have not remained equal: bicycle designs have altered, special children’s bicycles are now common, trainer wheels are standard, motivation and opportunities have changed, and so on.
In a few instances, detailed data on the performance of physical skills at various times in the past are available, most notably in the form of records for various athletic events. These generally show an increasing level of performance. An obvious example is the running of the mile. Since Roger Bannister first broke through the four-minute ‘barrier’ in 1954, many others have run the mile as fast and the top performances have continued to improve. In 1967 the record was held by Jim Ryun at 3 minutes 51.1 seconds, in 1985 by Steve Cram at 3 minutes 46.3 seconds, and in 1999 by Hicham El Guerrouj in 3 minutes 43.1 seconds. A similar pattern of improvement is shown in practically all other athletic events. Not only have records repeatedly been broken, but the average levels of performance in international competitions have also improved.
Here too many factors may have played a part: improved nutrition, better training methods, psychological factors, greater motivation, selection of athletes from a larger population of potential competitors, and so on. Any contribution of morphic resonance cannot be teased apart from these other influences.
Experimental tests for the effects of morphic resonance need to be designed in such a way that other factors are kept as constant as possible. Several experimental designs and actual experimental results are considered below.
Morphic resonance in language learning
Human babies have an inherited disposition to learn languages; the young of other mammalian species do not. In conventional terms this is thought of as a kind of programming in the DNA. From the present point of view it is due to morphic resonance from innumerable people in the past. This resonance underlies the general tendency to acquire language, and also facilitates the acquisition of particular languages, such as Swedish and Swahili, by resonance from previous speakers of these languages.
Languages have hierarchically ordered structures which, as René Thom has pointed out, can be thought of as a hierarchy of chreodes, or ‘canalized pathways of change’ (Fig. 6.2). The sentence chreodes organize the phrases, the phrase chreodes the nouns, verbs, adverbs, conjunctions, and other parts of speech in the phrases, the word chreodes the syllables, and the syllable chreodes in turn organize the lowest-level chreodes, the phonemes.2
Such hierarchical patterns of organization are found in all languages; the grammatical ways in which the words are arranged and interrelated constitute the syntax of a language. But the syntax alone does not confer meaning. It is perfectly possible to construct grammatically correct sentences that are meaningless; and of course meaningful sentences are not necessarily true. Beneath the grammatical structures of sentences, which Noam Chomsky called the ‘surface structure’, are further levels of organization that give a sentence its ‘deep structure’; and the deep structures of sentences are connected with further levels of organization and interrelationship that are the basis of meaning.
There are a variety of theories about the structure of language, and a consideration of the patterns of organization that give rise to meaning inevitably leads beyond the realm of linguistics into the territories of psychology and philosophy. These invisible and inaudible structures that lie below or beyond the surface structure of the language have been extremely difficult to characterize. After all, what kinds of things can these structures be? They are certainly patterns of organization, but what is their nature? From a mechanistic point of view they are somehow connected with patterns of nervous activity within the brain. The hypothesis of formative causation complements this assumption by enabling one to think of these structures as nested hierarchies of morphic fields that act on and through the patterns of electrical activity in the nervous system.
An understanding of these structures is not merely a matter of academic or philosophical interest, but is of great practical importance, for example in attempts to program computers to translate from one language to another automatically and in the development of ‘artificial intelligence.’ These attempts have so far met with limited success. One reason for the slow progress in these fields, despite a vast investment of money and effort, is that they take for granted the mechanistic theory of the organization of language and intelligence, which may well be wrong. The organization of language and learning may depend on morphic fields, which these computer models do not take into account.
Chomsky argued that the rapid learning of language by children, including their grasping of grammatical rules, cannot possibly be explained in a behaviourist manner in terms of stimuli and conditioned responses. One of the most striking features of the use of language is its creativity: by the age of five or six children are able to produce and understand an indefinitely large number of utterances that they have not previously encountered.3 Chomsky thinks of this as an organic rather than mechanical process: ‘Language seems, to me, to grow in the mind, rather in the way that the physical systems of the body grow.’4 In order to account for the remarkable ability of children to pick up languages, he proposed that the basic organizing structures of languages are innate: children inherit them.5 But since babies of any race seem to have the capacity to learn any human language, Chomsky was driven to the conclusion that these inherited structures must be common to all languages: they represent what he calls a universal grammar, and he regards one of the tasks of linguistics as the determination of the universal and essential properties of all human languages. He regards these as ‘genetically programmed’.
This notion of a universal grammar is the most controversial aspect of Chomsky’s system. It is by no means clear that all languages share common generative organizing principles and deep structures of the kinds that Chomsky proposes on theoretical grounds – especially since many of these have never been specified.
A general morphic resonance from all of past humanity would indeed reinforce any organizing fields and chreodes that are common to most if not all languages, and this would be in harmony with Chomsky’s proposal. However, it is not necessary from this point of view to suppose that the grammatical structure of all languages depends on a single universal grammar. The general morphic resonance gives young children a general tendency to learn language, but as they begin to speak a particular language, such as Swedish, they enter into morphic resonance with the people they hear speaking it; their learning of its particular grammar and vocabulary is facilitated by this resonance. Speaking this language tunes them in to speakers of the same language, including many millions of speakers in the past.
Chomsky pointed out that his theory makes a prediction that could be tested in principle, if not in practice. If an artificial language that violated the universal grammar were constructed, ‘it would not be learnable under normal conditions of access and exposure to data.’6 The same result would be expected on the basis of the hypothesis of formative causation (assuming for the purposes of argument that the universal grammar can indeed be specified). However, this hypothesis makes a prediction that differs from Chomsky’s: an artificial language constructed in accordance with the universal grammar but differing in important respects from all natural languages, past and present, would be much more difficult to learn than any natural language. This would be because the chreodes are not stabilized by morphic resonance from past speakers of the language, for there would have been no such people.
The hypothesis of formative causation also predicts that languages spoken by very many people in the past should on average be easier to learn than those spoken by very few people, other things being equal. Why, then, cannot we all learn languages such as Mandarin and Spanish very easily, since so many millions have spoken them? In the case of adults, the situation is obviously complicated by the fact that the deep-seated habits associated with our own language strongly interfere with the acquisition of others. The more strongly these habits shape our speaking, listening, and understanding, the less easily will we be able to acquire the new patterns of the language we are learning. Perhaps this is why people with a ‘good ear’ and an unusual ability to imitate the speaking of others are often particularly gifted in the learning of foreign languages. Through skilful imitation, they tune in more effectively to speakers of the language than most of us can, and their further learning of it is facilitated to a greater extent by morphic resonance.
In the case of babies, where there are no established habits to interfere with the acquisition of language, there could indeed be differences in the ease with which they acquire common and rare languages: other things being equal, English, for example, might be easier to learn than a rare tribal language from the Amazon, just because so many more people have spoken it. In practice, of course, this would be very difficult to investigate because other things are unlikely to be equal; any effect of morphic resonance would be difficult to separate from differences in genetic constitution, cultural environment, methods of child-rearing, and so on.
In his book The Language Instinct, Steven Pinker gives several examples of the rapid evolution of new languages. When speakers of different languages need to communicate but do not learn each other’s language, as when colonizers or traders come into contact with unrelated people, they develop a makeshift called a pidgin, choppy strings of words borrowed from the language of the colonizers, with little in the way of grammar. But in many cases, pidgins have been transformed into a full complex language, a creole, at one fell swoop. All it takes is for a group of young children to be exposed to the pidgin at the age when they acquire their mother tongue. Historically, this probably happened in groups of children of slaves who were tended collectively by a worker who spoke to them in pidgin. ‘Not content to reproduce the fragmentary word strings, the children injected grammatical complexity where none existed before, resulting in a brand new, richly expressive language.’7
Even more remarkable is the evolution of new sign languages. For example in Nicaragua deaf people were isolated from each other and there were no sign languages at all until the Sandinistas came to power in 1979, when the first schools for the deaf were created.
The schools focussed on drilling the children in lip reading and speech, and as in every case where that is tried, the results were dismal. But it did not matter. On the playgrounds and school buses the children were inventing their own sign system, pooling the makeshift gestures that they used with their families at home. Before long the system congealed into what is now called the Lenguaje de Signos Nicaragüense (LSN).8
This pidgin sign language was used by deaf children who joined the school when they were aged ten or older. But younger people, who joined the school about the age of four, when LSN was already around, are quite different. They speak a far more complex and expressive language, referred to by a different name, Idioma de Signos Nicaragüense (ISN). This creole language with its consistent grammar was created in one leap. As Pinker remarked, ‘A language has been born before our eyes’.9
The inherited plans that facilitate the learning of existing languages and the evolution of new ones are not general principles that for logical reasons have to be present in all languages. They are more like arbitrary conventions that could have been different. As Pinker wrote, ‘It is as if isolated inventors miraculously came up with identical standards for typewriter keyboards or Morse code or traffic signals.’10
Chomsky and Pinker both assumed that the ability to learn languages must depend on a coding in the DNA for universal structures common to all languages. They took it for granted that all hereditary information was inscribed in the genes, and were forced to assume the existence of a universal grammar because young children of any ethnic group seem to be able to learn any language; for example, a Vietnamese baby adopted by a Finnish family can easily learn Finnish.
Morphic resonance provides a simpler explanation. The young child resonates with the speakers around it, and with millions of speakers of the language in the past. Morphic resonance facilitates its learning of the language, just as it facilitates other kinds of learning. Likewise, morphic resonance facilitates the acquisition of sign languages by deaf people, who tune in to past users of these languages. There is no need to suppose that genes for ordinary languages or for sign languages lie latent in everyone’s DNA.
The interpretation of language acquisition in terms of formative causation is speculative. But so is the theory of genes for a hypothetical universal grammar. As Pinker himself remarked: ‘No one has yet located a grammar gene.’11
Experimental tests with foreign languages
There are two types of experimental test for morphic resonance in human learning. First, old-field tests attempt to detect the influence of morphic resonance from many people in the past who have learned these skills. Second, new-field tests involve the acquisition of a new skill during the experiment itself. Such tests may involve the solving of novel puzzles, for example, or the playing of new kinds of video games. The average rate at which groups of inexperienced subjects can learn them is monitored at regular intervals in one location. Meanwhile, these puzzles are solved or the video games are played by thousands of people elsewhere. The average speed of learning for naïve subjects in one place should increase as more and more people learn them somewhere else. Tests of both kinds have already been carried out. I first discuss old-field tests.
In 1982, the British magazine New Scientist held a competition for experimental designs that could be used in the testing of the hypothesis of formative causation.12 The winner, Richard Gentle, proposed an ingenious test involving a Turkish nursery rhyme.13 He suggested that English-speaking people could be asked to memorize two short rhymes in Turkish under standard conditions, one a traditional nursery rhyme, known to millions of Turks over the years, the other a new rhyme made by rearranging the words in the genuine nursery rhyme. The subjects would not be told which was which. After equal periods spent in memorizing each of the rhymes, they would be tested to find out which one they remembered better. If the learning of the genuine rhyme was facilitated by morphic resonance from millions of Turks, then it should be easier to memorize than the newly constructed rhyme.
I took up Gentle’s suggestion, but used Japanese rather than Turkish rhymes. A leading Japanese poet, Shuntaro Tanikawa, kindly supplied me with three rhymes for this purpose: a genuine nursery rhyme known to generations of Japanese children, and two others specially composed to resemble it in structure. One of these was meaningful and the other meaningless in Japanese.
In a series of experiments with groups in Britain and America who learned the rhymes by chanting each of them a fixed number of times (without knowing which was which), 62 per cent of those tested found the genuine rhyme easiest to recall half an hour later. This result was far above chance expectation: if the rhymes were of equal difficulty, by chance about 33 per cent of those tested would have been expected to recall the genuine rhyme better than they recalled the new ones. In another experiment in which people were supplied with the rhymes in a written form 52 per cent found the genuine rhyme easiest to learn, again a highly significant result. Learning the written material was not as effective as chanting, which is of course much closer to the way that Japanese children learn the genuine rhyme. There were no consistent differences in the ease with which the two newly composed rhymes were learned.
These results, although encouraging, are open to the criticism that the new rhymes may be intrinsically harder to learn than the traditional nursery rhyme, in spite of the poet’s efforts to make them of comparable difficulty. This argument gains additional force when we consider the history of nursery rhymes: presumably they are subjected to a process resembling natural selection, and maybe easy-to-remember rhymes are more likely to survive. If, on the other hand, one of the new rhymes had been memorized better than the nursery rhyme, it could have been argued that for some reason it was intrinsically easier. Consequently, this type of experiment cannot give conclusive results one way or the other.
To overcome this problem, a different experimental design would be necessary, using a new-field procedure. This could be carried out using several new rhymes, in Japanese for example, of similar metre and sound structure, and of similar difficulty. The rate at which they could be memorized would then be determined by testing people in, say, the United States. Then one of these rhymes, chosen at random, would be learned by many people in Japan. Subsequently, new sets of subjects in the United States would be asked to memorize the rhymes, and the rates at which they did so would again be measured. The one that had been learned by the Japanese should now be easier to memorize than before, but the other rhymes, which serve as controls, should not. Such an experiment might be practicable if one of the rhymes were used in a popular song in Japan, or even in an advertising jingle.
In a second competition, initiated by the Tarrytown Group of New York, three prizes were offered for the best actual tests of the hypothesis of formative causation: with a $10,000 first prize.14 The winners were selected by an international panel of judges: Professors David Bohm of London University, David Deamer of the University of California at Davis, Marco de Vries of the Erasmus University, Rotterdam, and Michael Ovenden of the University of British Columbia. The prizes were awarded in New York in June 1986. Two entrants tied for first place. Both independently carried out similar experiments that involved words written in foreign scripts, Hebrew in one case and Persian in the other.
These tests were based on the idea that words that have been read by millions of people should be associated with morphic fields that facilitate the perception of the patterns of the words. Hence people who are entirely unfamiliar with a foreign language and its script may find it easier to recognize or learn real words in this language than false words made up of letters in meaningless sequences. These non-words will not have been written or read by millions of people in the past, and hence the perception of their patterns will not be stabilized by morphic resonance. Note that such tests depend entirely on the visual patterns of the words: they do not involve hearing the words, nor do they involve any attempt to pronounce them; they are carried out in ignorance of the phonetic values of the letters.
Gary Schwartz, who was then a professor of psychology at Yale University, selected 48 three-letter words from the Hebrew Old Testament, 24 of them common and 24 rare. He then scrambled each word to produce a meaningless anagram containing the same three letters. This gave 96 words in all, half real and the other half false.
Over 90 students who were ignorant of Hebrew were shown these 96 words one by one, projected on a screen in a random order. They were asked to guess the meaning of each word and to write down the first English word that came to mind. They were then asked to estimate on a 0-to-4 scale the confidence they felt in their guess. The subjects were not told the purpose of the experiment, nor that some of the words were scrambled.
A few of the subjects did in fact correctly guess the meanings of some of the Hebrew words. Schwartz excluded these subjects from his analysis on the ground that they could possibly have had some knowledge of Hebrew in spite of the fact that they said they had none. He then examined the replies of the subjects who had always guessed the wrong meanings. Remarkably, on average they reported feeling more confident about their guesses when they were viewing the real words than the scrambled words, even though they did not know that some of the words were real and others were fake. This effect was roughly twice as strong with the common words as with the rare words. These results were highly significant statistically.15
After Schwartz had tested his subjects in this way, he informed them that half the words were real and the other half scrambled. He then showed them all the words again, one by one, asking them to guess which were which. The results were no better than chance: the subjects were unable to do consciously what they had already done unconsciously. Schwartz interpreted the greater confidence subjects felt about their wrong guesses of the meanings of the real words in terms of an ‘unconscious pattern recognition effect’.
Alan Pickering, a psychologist at the former Hatfield Polytechnic in England (now the University of Hertfordshire), used two pairs of real and scrambled Persian words, written in Persian script (which resembles Arabic). He tested 80 students, showing each of them only one of the words. They were asked to look at the word for ten seconds and then to draw it after the viewing period was over. Several independent judges subsequently compared these reproductions of the actual and false words. The judges were not told the purpose of the experiment, nor did they (nor Pickering himself) know which words were real and which scrambled.
The real words were reproduced more accurately than the false words. For example, in one method of judging, which involved comparing pairs of answers (paired at random) with the corresponding real and false words, on average in 75 per cent of the pairs the real words were judged to be reproduced better than the false. This effect was highly significant statistically. Pickering, like Schwartz, concluded that his results were in good agreement with the hypothesis of morphic resonance.
A possible alternative explanation is that real words tend to have certain aesthetic or other qualities that false words do not, for reasons that have nothing to do with morphic resonance. This argument is, however, vague, as becomes apparent when we try to apply it in detail. Schwartz found a greater ‘unconscious pattern recognition effect’ with common Hebrew words than rare ones. If this effect is due not to morphic resonance but to aesthetic or other properties of the real words, then why should common words be more aesthetic when written down than rare words? Do words become more frequently used within a language because their written forms are pleasing to the eye? Or do common written words become easier to recognize, owing to morphic resonance, because they are frequently used? Or are common words more recognizable both because of morphic resonance and for aesthetic reasons unconnected with morphic resonance?
Suitbert Ertel, a professor of psychology at Göttingen University, Germany, investigated the possible effects of morphic resonance on the recognition of Japanese hiragana script, a phonetic component of the Japanese writing system. Participants were shown nine different hiragana characters in a random order, projected on a screen for eight seconds. They then turned to an answer sheet with twenty hiragana characters among which the nine characters they had just seen were randomly mixed. They were asked to mark the characters they thought they had just seen. The same test was repeated with the characters in different random orders. Each participant did six trials, and the recognition of the hiragana characters generally improved trial by trial.16
Ertel predicted that if morphic resonance were playing a part, hiragana characters should be recognized better when they were the right way up than when they were upside down, because millions of Japanese people were used to recognizing these characters in their normal position. Sure enough, this is what he found.
In a further experiment, he used artificial hiragana characters invented by a graphic designer. Before running the learning tests, he and his students showed participants genuine and artificial hiragana characters and asked them to pick the genuine ones. They could not tell the difference. The Göttingen team then carried out their standard memory tests, and found that the real characters were remembered better than the false ones, in accordance with the predictions of the hypothesis of morphic resonance.
Ertel and his team then carried out a further test, which they regarded as crucial. They compared the effect of putting the real characters upside down with putting fake characters upside down. Ertel argued that with fake Hiragana characters, rotation should have no effect because morphic resonance plays no part in the recognition of these characters either way up.
The results were confusing and Ertel’s interpretation was hard to follow. In the first two trials, there was indeed almost no difference in the recognition of the upside down and right way up fake hiragana characters (Fig. 10.1). But in the subsequent trials, the false hiraganas were remembered better right way up. Ertel argued that the faster rate of learning in the later trials with the fake hiraganas right way up was because of ‘intrinsic factors’ that had nothing to do with morphic resonance. Surprisingly, he provided no statistical analysis to show whether this effect was significant or not.
Figure 10.1 The results of Suitbert Ertel’s experiment on the recognition of hiragana characters. The vertical axis shows the number of words recognized and the horizontal axis the six successive trials. (Reproduced by courtesy of Suitbert Ertel)
However, Ertel’s fake hiragana characters were designed to look like real hiragana characters when they were the right way up. Insofar as they resembled real hiragana characters, it may be that they did so precisely because they had a ‘right way up’ feel to them, because of a generic resemblance to real characters. This ‘intrinsic factor’ may not be an alternative to morphic resonance, but depend on their generic resemblance to right-way-up hiragana characters, which was built in from the start. But the complexity of Ertel’s interpretations illustrates how difficult it is to obtain clear-cut results in old-field experiments.17 A further complication was that Ertel unwittingly used an obsolete hiragana script that went out of use in the 1930s, that even Japanese people find hard to recognize today.
Several subsequent projects using unfamiliar scripts have given results that support the initial finding of Schwartz and Pickering, and much of Ertel’s work. At the University of Leipzig, Germany, Dirk Baumeier carried out a study in which subjects memorized real and false Japanese words, written phonetically in Roman characters, without knowing that some were real and the others were meaningless anagrams of the real words. Subjects remembered the real words significantly better than the false ones.
In a second study, Baumeier presented his subjects with a series of real Japanese words (written in Kanji script) as well as lists of real words in Indonesian, Swahili and Sanskrit, written phonetically in Roman script. Beside each real foreign word were two German words, one of which had the same meaning as the foreign word, while the other was completely unrelated. They had to try and guess which German word had the same meaning. By chance they would have been right 50 per cent of the time, but in a total of 11,200 guesses, 54 per cent were correct, which were very significantly above chance. He concluded that his results supported the hypothesis of morphic resonance.18
At the University of Innsbruck, in Austria, Robert Schorn, Gottfried Tappeiner and Janette Walde compared real Russian words written in Cyrillic script with meaningless anagrams of these words. The real and false stimuli were presented in pairs, and the participants asked to judge which had more ‘spirit’. The real words were selected significantly more than the anagrams. Some of these tests took place through the internet, illustrating the potential for widespread public participation in automated morphic resonance tests.
The same researchers also carried out another old-field test using political, religious and economic symbols such as flags, emblems and trademarks that were once well known but have now fallen into oblivion, or ones that are familiar to many people in foreign countries, such as the Chinese Coca-Cola sign, Indian trademarks, or Far Eastern religious symbols. For each of the symbols, a designer created a corresponding control symbol with a similar general pattern and similar complexity.
In order to find out whether the new symbols were indeed comparable, the experimenters conducted seven pre-tests with more than 200 participants who were asked to indicate which of the symbols in each pair they found less credible or real. In their main experiment, they employed only false symbols that seemed as real as the originals, if not more real. Participants were shown pairs of symbols, one real and one false, in a random order, and asked to judge which of each pair had more ‘spirit’. They selected the real symbols significantly more often than the fake ones.19
At the University of Northampton, England, Kimberly Robbins and Chris Roe, carried out an old-field experiment using genuine and false Chinese characters, with an experimental design similar to Ertel’s. Participants were first shown a powerpoint presentation consisting of five real and five false Chinese characters in a random sequence, seeing each character for three seconds. They were not told that some characters were real and others false. They were then given a sheet with twenty characters on it, and asked to circle the ten they had just seen. The other ten characters were ‘decoys’, and again five were real and five false. The participants recognized the real characters significantly better than the false ones. In addition, with the decoys, participants had significantly more false memories of real than false characters, consistent with a morphic resonance effect.20
Nevertheless, all old-field tests face the difficulty of controlling for ‘intrinsic factors’ that might make old symbols, words or rhymes more memorable or more attractive than newly invented ones. But again, it is hard to know if intrinsic factors and morphic resonance are genuine alternatives. Intrinsic factors may themselves depend on morphic resonance.
A test with Morse Code
Samuel Morse invented his eponymous code in the mid-nineteenth century for use in telegraphy. It has been learned and used by many people over the years, and is still in use today, especially by amateur radio operators. Does morphic resonance from all these people make it easier to learn?
An experiment to test this possibility was carried out by Arden Mahlberg, an American psychologist who received the third prize in the Tarrytown competition. He constructed a new version of this code by reassigning the dots and dashes to different letters of the alphabet. His subjects were people who did not know Morse code, and he compared their ability to learn the new code with their learning of the real Morse code. The test used letters and their associated dots and dashes in a written form. (The letters S and O were excluded because many people who do not know Morse code are nevertheless familiar with the code for S.O.S.) The subjects were exposed to the new code and the genuine Morse code one after the other, in random order, for equally brief periods.
In his first trials, Mahlberg found that, on average, subjects learned the real Morse code significantly more accurately than the new code.21 In subsequent tests with new subjects, he found that the average accuracy of learning of the new code progressively increased until it was learned almost as well as the real Morse code. He suggested that the original difference could have been due to morphic resonance from past practitioners of Morse code, resulting in a significant facilitation of the learning of this code compared with the newly constructed code. But as the tests were repeated with fresh subjects, those in the subsequent tests were influenced by morphic resonance from those who were tested before them. This effect, owing to the high specificity of resonance with previous subjects tested in the same way under the same conditions, swamped the more subtle effects of morphic resonance from users of the real code, and led to a progressive equalization of scores of the two codes. But Mahlberg recognized that this was only a tentative explanation, and emphasized the need to explore this possibility in further experiments.
One way in which the experimental design could be improved would be to use sounds for the dots and dashes, rather than presenting them in written form. This method would correspond far more closely to the habitual experience of telegraph and amateur wireless operators. For this purpose, a computer could relatively easily be programmed to show a standard sequence of Roman letters on the screen and at the same time beep the dots and dashes.
A possible test with Hindi keyboards
The first commercially successful typewriters were made by Remington in the 1870s. The keyboard was constructed not so much for ease of use or ease of learning, but for mechanical reasons connected with the way the pivoted type-bars swung. The layout of the keys was designed to prevent the most frequently used letters from jamming together at the printing point. This original arrangement, called the QWERTY layout after the first letters in the top row, was retained almost unaltered in subsequent machines. It still survives in computer keyboards, in spite of the fact that the original mechanical reasons for it have long since disappeared. Over the years, many improved keyboard layouts have been advocated, designed for greater ease of use, but none have so far succeeded in displacing the traditional QWERTY format.
Hundreds of millions of people have used QWERTY layouts since the 1870s. Morphic resonance might therefore be expected to facilitate markedly the learning of this skill and strongly stabilize the associated morphic fields. Typing has in fact intrigued experimental psychologists for decades, because the ‘rate at which typists (even average ones) perform exceeds by far the rate that laboratory tests quite common in psychology would lead a psychologist to predict.’22
No doubt one reason why new, improved keyboard designs have failed to catch on is the difficulty of retraining people and scrapping existing machines; but despite its inefficient design, the QWERTY layout may be easier to learn and use precisely because so many people have already become familiar with it. There is in fact empirical evidence that non-typists find the standard QWERTY layout easier to learn than a random layout;23 and alphabetical ABCDE keyboards, designed for ease of learning by novices, have in some experiments proved harder24 and in others at least no easier than the standard layout.25 ‘Operators with little or no typing skill, for whom alphabetical arrays are often intended, were as fast or faster with the standard typewriter arrangement’.26
In experiments specifically designed to test for the effects of morphic resonance, the rate at which novices learn to type on the QWERTY keyboard could be compared with other layouts that from the point of view of contemporary psychological theory should be of equivalent difficulty. But a conclusive experiment would be difficult to perform in the Western world, because almost everyone has already been exposed to QWERTY keyboards.
However, this experiment could be carried out with students of Hindi (or Tamil or Russian) who are familiar with the relevant script but who have not previously been exposed to typewriters using this script. In such an experiment, the rate of learning to type with, say, a standard Hindi keyboard would be compared with the rate with a different layout, designed to be of similar difficulty according to conventional theories. The hypothesis of formative causation would predict that the standard layout should be easier to learn, just because so many people in India have already learned it.
This experiment could be done quite easily using suitably programmed computers whose keys are labelled with the appropriate patterns of Hindi letters. A standardized learning procedure could also be programmed into the computer and the rate of learning recorded automatically.
These examples, only a few of the many conceivable ways in which the hypothesis of formative causation could be tested in the realm of human psychology, illustrate that such experiments can be carried out with facilities and equipment that are readily available in universities and schools.
New-field tests
The simplest new-field tests start with two new patterns. The first step is to find out how easily they can be learned or recognized. The second step is to build up morphic resonance from one and not the other. If morphic resonance is playing a part, the one that has been ‘boosted’ should subsequently be easier to learn or recognize; there should be no such change with the control. The fourth step is to test how easily the two patterns can now be learned or recognized. Has the boosted pattern indeed become easier relative to the control?
The first new-field test was carried out with hidden images, following a suggestion by Nicholas Humphrey. Such pictures seem to make no sense at first, or contain only vague hints of patterns (Fig. 10.2). Seeing the underlying image (Fig 10.3) involves a sudden Gestalt shift; the picture takes on a definite meaning. After this has happened it is difficult not to recognize the hidden image, and hard to believe that others cannot see it. If morphic resonance is at work, a hidden image should become easier to recognize if many people have already seen it.
Figure 10.2 A hidden image used in a television test for morphic resonance. The image is revealed in Fig 10.3.
In 1983, a British television company, Thames Television, made it possible for me to conduct an experiment of this kind. Two puzzle pictures were specially produced by an artist and designed to be difficult; only a small minority of people could spot the hidden images. Before the television broadcast in Britain, I sent both these pictures to collaborators in Europe, Africa, and North America. Each experimenter showed both pictures for one minute each to a group of participants before the transmission, and afterwards to another group of comparable participants. The number of people who recognized the hidden image was recorded.
The experimenters did not know which of the pictures was going to be shown on television, and nor did I. On the TV show itself, one was picked at random and shown to about two million viewers. After several seconds the answer was revealed, and this then ‘melted’ back into the puzzle picture so that the previously hidden image was now readily apparent. The same picture was shown once more at the end of the programme.
Figure 10.3 The image hidden in Fig 10.2.
The percentage of participants recognizing the control picture before and after the TV broadcast did not change, while the percentage recognizing the image shown on TV in Britain increased to a statistically significant extent.27
The experiment was repeated, using new puzzle pictures, on BBC television in 1984 on a popular science programme called Tomorrow’s World. Experimenters in other countries tested groups of participants to find out what proportion could recognize the hidden images within 30 seconds. Tests were carried out before the TV transmission in Britain, and again with comparable participants afterwards. On the TV show, one of the two images was selected at random and shown to 8 million viewers, to whom the answer was then revealed.
This picture did in fact become significantly easier to recognize elsewhere while there was no change with the control. But this positive effect was confined to participants in continental Europe; there was no effect in North America. The disparity was surprising. Morphic resonance should not be distance-dependent. One possible explanation was that in Europe, where the time difference from Britain was only one hour, people were more ‘in phase’ with the British TV audience then people in America, with a five to eight hour time difference.
A similar hidden image experiment was carried out in February 1985, with a TV transmission in Germany by Norddeutscher Rundfunk. Again there were two pictures of which only one was shown on television and thousands of participants were tested in other countries before and after the German TV show.28 As in the previous experiments, there was no significant change in the proportion of people recognizing the control picture, but the proportion recognizing the test picture declined in Britain and elsewhere after being seen by about half a million people in North Germany! From the point of view of morphic resonance, there should have been an increase. From a sceptical point of view, there should have been no change. Nobody predicted a decrease.
This result showed that other factors were coming into play, but what were they? No one knew. This puzzling finding discouraged anyone from doing more tests of this kind on television.
In 1987, the Institute of Noetic Sciences (IONS), near San Francisco, California, offered an award for the best student research on morphic resonance. An independent panel of judges assessed the entries, and the results were announced in 1991.29
The winner of the undergraduate award was Monica England, a psychology student at the University of Nottingham, England. Her test was stimulated by anecdotal evidence that some people find it easier to do newspaper crosswords the day after they have been published than when they first appear, an effect that could be due to morphic resonance from thousands of people who have already done the puzzle.
The experiment involved two puzzles from a London newspaper, the Evening Standard, which was not distributed in Nottingham. The newspaper kindly co-operated by supplying two unpublished puzzles a week before they appeared: the ‘easy crossword’ and the ‘quick crossword’. The easy crossword had simple cryptic clues, and the quick crossword single-word clues that required synonyms as answers.
Monica England tested about 50 students the day before the crossword puzzles were published in London, and a further 50 the day after. Both groups of participants were also given two control puzzles, which had been published in the Evening Standard two weeks earlier. The participants were given ten minutes with each crossword to solve as many clues as possible.
On average, participants solved significantly more clues with the easy puzzle after it had been published than before. There was no change with the control crossword. By contrast, with the quick crossword there was no significant difference in the test crossword relative to the control.
I repeated this experiment in 1990, again using easy and quick crosswords from the Evening Standard, and testing people with the help of experimenters who lived far from London, where the participants would not have seen this London newspaper. Again, the scores with both crosswords were compared with controls. There was a slight improvement in scores with the easy crossword after it was published, but this change was not statistically significant. By contrast, there was a statistically significant increase with the quick crossword. Thus the results were inconsistent, giving a positive effect with one crossword but not the other, as in Monica England’s experiment. In her test the easy crossword showed a positive effect, and in mine the quick crossword.
While reflecting on these results, I realized that I had taken it for granted that all the crossword puzzles were new, and I had assumed that they would be unaffected by morphic resonance from crosswords in the past. I then enquired how the crosswords were put together, and found that the compilers frequently recycled clues from previous crosswords. Hence these simple crosswords did not provide a good test for morphic resonance, since some of the clues were in fact not new.
Zoltan Dienes, then in the Psychology Department at the University of Oxford, won the IONS award for graduate students. His participants were required to decide quickly whether a string of letters they saw on a computer screen was a meaningful English word or a non-word. This experiment involved a phenomenon known to psychologists as ‘repetition priming’, which occurs when a word (or other stimulus) is recognized more quickly after repeated exposure to it. Dienes reasoned that later participants might find it easier to recognize stimuli if others had done so earlier.
The participants saw strings of letters flashed on a computer screen and had to indicate whether the string was a real word or a non-word by pressing computer keys as fast as possible. Dienes used two sets of words and non-words. All 90 participants were presented with a ‘shared’ word set, while only every tenth participant was shown a ‘unique’ set. The experiment thus involved 80 ‘boosters’ who viewed only the shared stimuli and ten ‘resonators’ who saw the shared stimuli together with the unique ones. If morphic resonance were at work in this test, the speed at which the shared stimuli were correctly judged should increase relative to the speed at which unique stimuli were correctly judged. In order to maximize the resonance between participants, all experimental trials were conducted in an environment with a distinctive smell, sound and visual background.
The outcome was positive and statistically significant. The more often a non-word had been seen before, the faster subsequent resonators responded to it. However, when Dienes repeated this experiment at the University of Sussex, there was no significant effect.30
Suitbert Ertel carried out two new-field experiments in addition to the old-field tests discussed above. In one there was a small effect in the direction predicted by the hypothesis of morphic resonance, but it was not statistically significant.31 In the second test there was no detectable effect.32
At Leipzig University, Dirk Baumeier analysed the results of his old-field test with Japanese words, discussed above, to find out if subjects became better at learning the real and false Japanese words after others had already learned them. In other words his old-field test also served as a new-field test. He found that the more people that had learned these transliterated Japanese words, the easier they became for others to learn them, an effect that was highly significant statistically. There were no significant differences in intelligence between the subjects tested earlier and later.33
In summary, some small-scale new-field tests have given positive results and others have shown no significant effects, perhaps because the resonance was too weak to be detectable with relatively few boosters.
Morphic resonance can be investigated on a much larger scale by studying changes in human performance over time. Does the performance of new skills show a tendency to improve as time goes on? Do video games get easier to play? Do new sports such as skateboarding and windsurfing become easier to learn? Anecdotal evidence suggests that they do, but such changes are not documented quantitatively, and the situation is complicated by other factors like improvements in equipment, fashion, better teaching methods, and so on.
One of the few areas in which detailed data are available over many years is for the scores of IQ (Intelligence Quotient) tests.
Rising IQ test scores
IQ tests should be getting easier to do as a result of morphic resonance from the millions of people who have done them before. Hence IQ scores should be rising, even though people are not really becoming more intelligent.
In 1982, it was discovered that average IQ test scores in Japan had been increasing by three per cent a decade since the Second World War.34 Psychologists invoked a combination of radical environmental changes to try to explain this unexpected finding: massive urbanization, a cultural revolution from feudal to Western attitudes, the decline of inbreeding, and huge advances in nutrition, life expectancy, and education.35 But a similar increase occurred in the United States over the same period, without such dramatic environmental changes as in Japan; changes of a comparable magnitude occurred nearer the beginning of the twentieth century in the United States.
The psychologist James Flynn then studied the results of intelligence tests by the US military authorities over a period of 70 years. He found that recruits who were merely average when compared with their contemporaries were above average when compared with recruits in a previous generation who had taken exactly the same test. No one had noticed this trend because testers routinely compared an individual’s score with others of the same age tested at the same time; at any given time the average IQ score was set to 100 by definition.36
Comparable increases occurred in at least twenty other countries, including Australia, Britain, France, Germany and Holland,37 and this phenomenon is now known as the Flynn effect (Fig 10.4). Many attempts have been made to explain it.38 One hypothesis postulates an increased sophistication in performance on standardized tests, and a ‘rising level of educational achievement.’39 Studies on the effects of repeated testing with parallel forms of IQ test have in fact shown that subjects can gain up to five or six IQ points through practice, but not much more than this.40 A second hypothesis was ‘enhanced educational achievement’.
Figure 10.4 The Flynn Effect. Changes in average IQ test scores in the US from 1918 to 1989, relative to 1989 values. (Data from Horgan, 1995)
Both these explanations were thrown into question by the fact that from 1963 onwards there was a decline in the average performance of American high-school students on the standard Scholastic Aptitude Test (SAT), taken by over a million students a year. Among several subtests, the greatest decline occurred in the verbal SAT. An official advisory panel was appointed to examine this decline and found that about half could be explained by the broadening of the sample of candidates; but the other half reflected a downward trend in test results of the general population and showed up in all socioeconomic groups of students.
Various personal traits contribute to SAT scores: intellectual ability, study habits, motivation, self-discipline, and acquired verbal and writing skills. The panel suggested that these traits could have been influenced by causes such as less demanding school standards, student absenteeism rates of over 15 per cent, the erosion of the nuclear family, and the influence of television.
But why did average IQ scores go up while average SAT scores had gone down? Some people suggested increasing exposure to television may have played a part. But IQ scores were rising decades before the advent of television in the 1950s, and as Flynn commented wryly, television was usually considered ‘a dumbing down influence until this effect came along.’41 He added: ‘The combination of IQ gains and the decline in Scholastic Aptitude Test scores seems almost inexplicable … IQ gains of this magnitude pose a serious problem of causal explanation.’42
In the 1980s, Flynn described the rise in IQ scores as ‘baffling’43 and he continued to wrestle with this problem for years. In 2007 he published a new analysis based on a detailed comparison of the different components of IQ tests. He found very little increase in the scores in the subtests for information, arithmetic and vocabulary, and only modest gains in comprehension. There had been large increases in ‘performance’ subtests, which involved solving puzzles. One of these, for example, involved completing pictures by indicating missing parts from incomplete pictures. The biggest increases of all were in a Similarities Subtest in which subjects had to indicate similarities between different items, and also in a subtest called Raven’s Progressive Matrices, which demand ‘that you think out problems on the spot without a previously learned method for doing so’.44
Flynn pointed out that the smallest gains occurred in subtests that depended on previously acquired knowledge, which he described as ‘crystallized’ skills. Tests of crystallized skills depended on knowledge acquired before the subject did the test: ‘you either know that Rome is the capital of Italy or you know only of Rome, Georgia; you know what ‘delectable’ means or you do not.’ In other words, the scores hardly changed when people were being tested on general knowledge that was independent of the test procedure, but they increased when people were solving problems on the spot under the same conditions that other people had previously solved them. These are exactly the results that would be expected on the basis of morphic resonance.
Flynn did not consider the possibility of morphic resonance. Instead, with his colleague William Dickens, he put forward an explanation in terms of ‘social multipliers’ suggesting that the population as a whole, all over the world, has become more imbued with abstract, scientific thinking, more used to using their minds in new ways, and more adept at on-the-spot problem solving. Flynn and Dickens saw the rise in IQ as a vindication of more and better education in an increasingly technological society. The massive gains in IQ ‘represent nothing less than a liberation of the human mind. The scientific ethos, with its vocabulary, taxonomies, and detachment of logic and the hypothetical from concrete referents, has begun to permeate the minds of post-industrial people’.45 However, the Dickens-Flynn hypothesis was greeted with deep suspicion by other psychologists and is not generally accepted.46
Flynn concedes that testing the Dickens-Flynn model will take a long time, involving large numbers of people over many years. However, testing the morphic resonance hypothesis could happen much faster. Here is a possible experimental design. First, construct a new, shortened, IQ-type test containing several questions depending on arithmetic, information and vocabulary, and several questions that involve on-the-spot puzzle solving. Then test large numbers of people with this test in the space of a few months. If morphic resonance is at work, there should be a larger increase in the scores for on-the-spot puzzles than for the other items. This should happen rapidly, and will not depend on social changes extending over many years, as in the Dickens-Flynn hypothesis.
An improvement of this design would be to prepare two versions of the test, with different but similar questions. Version A will be given to 90 per cent of the people taking part, and version B to the other 10 per cent. By morphic resonance, there should be larger increases in test scores in version A than version B because there will be a bigger boost by morphic resonance for those who take the test later. By contrast, the Dickens-Flynn hypothesis would predict no difference between the results with versions A and B.
In this chapter we have explored the possibility that the learning of language and of physical and mental skills is facilitated by morphic resonance from people who have already learned them. If further experiments provide a convincing weight of evidence in favour of this hypothesis, it could have far-ranging implications for education and training. For example, new methods of teaching could be designed to maximize the facilitation of learning by morphic resonance.
I now consider the role of morphic resonance in personal memories.