Commentary: The birth of the twin study--a commentary on Francis Galton's 'The History of Twins'

Abstract

In 1875, the English scientist Francis Galton published an article entitled ‘The History of Twins’. Now little remembered, it nonetheless has considerable historical significance. This 137-year-old paper, reproduced in this volume, represents the first detailed attempt to use the phenomenon of twinning to estimate the relative powers of nature and nurture. Fittingly for a study of twins, Galton wrote a pair of very similar papers on the phenomenon in the same year. This one was printed in a general and literary journal called Fraser’s Magazine.1 There was nothing unusual about cutting-edge science appearing in a popular periodical: science was only gradually emerging as a profession in its own right and Galton already enjoyed a wide reputation as an African explorer. In the mid-1870s, however, he was also a rising man of English science. Galton had an excellent scientific pedigree as the grandson of the poet and evolutionist Erasmus Darwin and the half-cousin of Charles Darwin. Even so, he had taken some time to find his vocation as a scientist. After a severe mental breakdown while a student at Cambridge University, in 1849 Galton turned to a phrenologist to discover where his talents might lie. His cranial topography suggested a flair for adventure.2,3 Partly for this reason, Galton had spent a year exploring the regions of southwestern Africa where the imperial powers of Europe were yet to extend their reach. On his return he wrote an account of his travels, which won him a membership to the highest circles of geographers and anthropologists.4 But even as he made the charmlessly racist observations about indigenous people typical of the mid-Victorian Briton, Galton was already showing the bent for scientific measurement and compulsive counting which were to be the hallmarks of his scientific career. In Africa, he delighted in making detailed measurements of distance, calculating latitude and longitude, estimating altitude by testing the temperature of boiling water, even working out the vital statistics of a beautiful indigenous woman whom he measured from a distance using a combination of sextant, trigonometry and logarithms. This fascination for rendering complex natural phenomena down to simple numerical expressions would lead him to a series of significant advances. The nascent fields of statistics and criminal identification would all profit from his endeavours. And he would die, in 1911, a grand old man of science, able to reflect on the part he had played in helping to secure victories for secular science and evolution over theology and revelation.5,6 But most of his career was devoted to trying to better understand the nature of heredity and the degree to which inheritance determines the way we think, feel and act. By 1875, Galton had been puzzling for over a decade about how to evaluate the roles of nature and nurture. This was not an entirely new endeavour. Since classical antiquity scholars had been making claims about the importance of either birth or education. Only in the preceding century, however, had men of science begun to suspect that it might be possible to disentangle the two. They had not enjoyed much success. With dogged determination, Galton now entered the fray. Like so many of those who preceded him, he already knew exactly what he wanted to find.7,8 So convinced from the outset was he of the primacy of heredity that Galton argued repeatedly for the need for the state or its wiser citizens to seek to encourage the best to outbreed the rest.9 In 1883, he even coined the term ‘eugenics’.10 The historian Raymond Fancher has speculated that Galton’s conviction that human intelligence is largely a matter of nature arose from his experience at school and university of desperately trying hard and still falling short of his personal expectations: a belief in the power of heredity perhaps rationalized his recurrent disappointments.11 Galton may also have been influenced by the mid-Victorian biological racism, which asserted that human races can be arrayed on a linear scale of inborn intelligence from the White European down to the Australian Aboriginal. Perhaps most importantly, Galton was inspired by the publication of The Origin of Species in 1859. He seized on the chance to support his half-cousin’s theory of evolution by arguing that human intelligence is a matter of brain rather than soul and that mental qualities are highly heritable. For obvious reasons, Darwin applauded and welcomed his kinsman’s efforts. In 1869, Galton published a book entitled Hereditary Genius, which claimed, on the basis of hundreds of potted genealogies, that ‘illustrious’ men are usually born to highly able fathers.12 The stuff of good brains, he concluded, is passed faithfully down through lineages. Galton was keenly aware of the obvious objection to his argument that the children of ‘great’ men also tend to receive superior education, social advantages and opportunities for emulation. In Hereditary Genius, he tried to argue his way around this thorny problem. He alleged, for instance, that when a pope adopts his nephew as his own son, the young man typically turns out to be mediocre in spite of receiving all the advantages of position and place. The papal nephew, Galton explained, simply does not have enough of his uncle’s hereditary material to really excel. But such arguments convinced neither Galton nor the reviewers of Hereditary Genius. Only Charles Darwin and Alfred Russel Wallace, who had their own agendas, greeted the book with enthusiasm. And so, in his next book, the 1874 statistical account of scientists entitled English Men of Science, Galton felt he had to concede that the ‘effects of education and circumstance are so interwoven with those of natural character’ that it is ‘impossible to treat them wholly apart’.13 This admission left undiminished Galton’s determination to come up with a credible strategy for separating nature from nurture. In English Men of Science, he had briefly mentioned the phenomenon of twins.14 On further reflection, he now wondered if it might not offer just the kind of natural experiment that he needed. He threw himself earnestly into the collection of data. Having obtained names and addresses from a leading English actuary, in late 1874, Galton sent out 600 questionnaires to families that included twins. He asked a series of questions about a wide range of physical and mental traits. He also sent queries to asylum superintendents asking about the occurrence of insanity in twins under their care.15,16 In a preamble to his questionnaire, Galton stated his purpose. ‘My object’, he wrote, ‘is to collect data for estimating the respective shares that “Nature” and “Nurture” ordinarily contribute to the body and mind of adults, meaning by “nature” everything that is inborn, and by “nurture”, every influence subsequent to birth’. He received back questionnaires relating to an impressive 94 sets of twins. Galton did not think about twinning in the way we do today.17 Abreast of recent research, he did know that some twins are the result of two eggs being separately fertilized and that others are the product of a single egg splitting in two. But Galton remained a long way from a clear comprehension of the differences between monozygotic and dizygotic twins. According to the theory of heredity through which he interpreted his evidence, Galton assumed that all children in a family receive exactly the same suite of hereditary particles from their parents. He called it the ‘stirp’ from the Latin for ‘root’. He went on to argue that some of these elements, the ‘patent elements’, are used to form the body of the child whereas others remain latent in testes and ovaries and are then passed on to future generations. Children in a single family differ, said Galton, because each is made out of a different sample of the elements contained in the original stirp bequeathed by their parents. Although wrong, Galton had come up with an intuitively plausible way of getting past the confusion that had always surrounded discussions of heredity. This model shaped how Galton understood the nature of twinning. He reasoned that children who arise from the same egg are not necessarily identical. This view derived from the distinction he made between patent and latent elements. He imagined that when eggs are released the patent elements, destined to be the building blocks of the future body, have already been randomly selected and separated from the latent elements. Now, if the egg splits at an early stage, the patent elements in the cell have barely begun to differentiate and so its division will produce two highly similar sister eggs. But, Galton elaborated, if the egg cell divides later it will have developed more, and so the two halves will contain less similar selections of patent elements. In other words, children who arise from the same egg can be very different by nature in terms of physique and character. Galton’s ideas of heredity were obviously a long way from those of Gregor Mendel, and he was as ignorant as virtually every other late 19th-century biologist of the labours of the mathematical monk of Brno. Galton had no sense of the neat regularities that actually underpin genetics and twinning. He could not, therefore, conceive of the classical twin study, which involves comparing the level of variation among monozygotic and dizygotic twins. Instead, Galton took his questionnaire15 data and asked a series of simple and pertinent questions. It did not matter for his purposes what happened during the development of eggs and sperm. There were for him two key empirical issues to be addressed. First, do twins who are similar at birth remain alike or do they diverge from one another over time? To Galton, the more the resemblances of twins persisted, especially if they experienced contrasting environments, the stronger the evidence for heredity. Secondly, do twins who differ at birth become more or less similar over time? High levels of convergence between initially dissimilar twins would provide evidence for the primacy of nurture over nature. Galton’s chief concern, then, was what happened after birth. What did he find? A lot of the data Galton received were too vague to be analysed. One return was perfectly useful but too indelicate to quote: a father told of how different his two twin sons were, explaining that if ‘there is any dirty trick—any nastiness I mean—done among the children it is he alone who does it (e.g. dipping a stick into a privy hole and daubing his brothers with it)’.15 Galton did, however, elicit a lot of quite serviceable information. He received dozens of accounts of remarkable levels of resemblance between twins: siblings who died, went mad or fell sick in the same way and at about the same time of life. There were stories of the wrong sibling being whipped for his brother’s misdemeanours, of fiancées who could not tell their betrothed from her twin, of mothers who mistook one son for another. In 35 out of 94 twin sets, the similarities were very striking in some way. Usually, the clearest correspondences were in physical appearance. But in 16 of these 35 pairs, Galton found that the twins were also alike in taste and disposition. Seldom were they identical in character; nevertheless, Galton concluded that whereas many twins do differ in personality, it is ‘in the keynote, not in the melody’. What about evidence of stability, divergence or convergence over time? Here lay the crux of Galton’s argument, and he posed this question with ‘no little interest’. He pointed out that the highly similar sets of twins had been raised together until early childhood and had then gone on to live quite different lives. Galton regarded this as a neat test of the power of heredity: had the twins stayed the same in later life or become unlike one another? Galton claimed to have been surprised by the strength of the evidence in favour of a hereditarian explanation. For, he asserted, except for those whose life-course was altered by a serious ailment, the twins ‘had continued unaltered up to old age, notwithstanding very different conditions of life’. It is as if, said Galton, they go on ‘keeping time like two watches’. Everything in life was mere sound and fury, which scarcely rippled the surface of the individual’s inborn nature. Next, Galton looked at 20 cases of twins who differed greatly at a young age. Had they become more similar because of being raised in the same home? Galton quoted a parent’s testimony to make his case: ‘They have had exactly the same nurture from their birth up to the present time; they are both perfectly healthy and strong, yet they are otherwise as dissimilar as two boys could be, physically, mentally, and in their emotional nature’. Galton alleged that not one set of twins that was dissimilar in childhood had become more similar ‘through identify of nurture’. Ergo, he went on, ‘there is no escape from the conclusion that nature prevails enormously over nurture when the differences of nurture do not exceed what is commonly to be found among persons of the same rank of society and in the same country’. Galton felt that, at last, he had what had been eluding him: a solid foundation on which to say that nature routinely triumphs over nurture. What are we to make of Galton’s foray in the science of twinning? Not surprisingly, his 1875 paper does not look very impressive when held up against a modern twin study. Galton lacked the statistical tools and the biological insights that would, in the 20th century, make the twin study the primary means of exploring the roles of nature and nurture. The evidence Galton invoked was also highly anecdotal and usually impossible to verify. Moreover, he made no attempt to find out if the similarity in behaviour of young children was the result of their shared upbringing rather than their shared heredity: similarities in childhood are always chalked up by Galton as the result of nature rather than nurture. One could, however, give his data the opposite interpretation: children receive their characters as young children because of social and familial influences, and they do not change much when older because their personalities and preferences have already been largely fixed. Like most studies, Galton’s was open to multiple interpretations. One could go on detailing the many alternative explanations for Galton’s data, which he blithely ignored. For instance, Galton did not acknowledge that the parents of twins often ignore the actual differences among their children because they expect to see conformity. In fact, he failed to consider any of the multiple ways in which observer bias could have come into play in the completing of his questionnaires. In short, a paper with the rigour of ‘The History of Twins’ would not, at least one hopes, pass muster in the modern peer review process. The paper also contains some more basic empirical flaws and non-sequiturs. In particular, Galton had several cases of twins who did become more different over time. This did not tally with his hereditarian convictions. So what did he do? He made the argument that as their parents could not point to anything that happened in their children’s lives that could explain the magnitude of the divergence, their growing distinctiveness must have been due to inherited differences. These innate differences, Galton explained, simply took a while to express themselves. This is a classic ‘ad hoc manoeuvre’: a slight adjustment made to a preferred theory so as to neutralize a piece of awkward data. Another striking example from the history of science is the Renaissance cardinal who was determined to uphold, against the evidence of Galileo’s telescope, Aristotle’s view that all bodies above the earth are perfectly round. The cardinal argued that the moon's surface might seem uneven but that it is actually coated in an invisible but perfectly smooth fluid. In the same way, if not as egregiously, Galton knew what he wanted to find and was an unimaginative critic of his own preconceptions. Unquestionably clever, Galton could also be rather naive in his interpretation of data. Yet, it would be unjust to reject ‘The History of Twins’ as the substandard science of a bygone age. For a couple of reasons, this is an article worth recognizing as a classic. First, some of the evidence it contains is genuinely suggestive even if it is far from conclusive. It may well be significant that, according to Galton’s questionnaire data, unlike twins did not become more similar in spite of receiving the same or similar upbringing. This does, as Galton pointed out, make it more reasonable to suppose that the twins who were similar as children were alike because of heredity rather than parental influences. Secondly, ‘The History of Twins’ testifies to Galton’s recognition that discussions of nature and nurture ought to be carried out on the basis of rational criteria for separating the effects of heredity and environment. Although he set out with a firm conviction that nature mattered most, Galton was not content to simply make bald assertions. He tried hard to be systematic: collecting a large quantity of data, subjecting it to at least basic numerical analysis and trying to find real-life situations—in this case twinning—which suggested the possibility of isolating the two key variables. Unlike most of those who waded into these treacherous waters, Galton faced up to the He knew that it would considerable to the effects of nature and nurture. Galton’s on twins, and the statistics of all arose out of this determination to go the vague and which most of his The modern of twin look Galton’s 1875 paper with a of and For all its ‘The History of Twins’ the questions and some of the basic which later would and In a and a Galton’s paper, we have an from twin while also aware of the of the and by which bodies and are To his in not from the of using twins to disentangle the effects of heredity and Galton to give a science the strength to go on to become a of modern of

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.