P As T Probabilities

383 _NA_TIJ_RE_VO_L_•• 3_12_22_N_O_VE_M_B_E_R_l984______BOOK REVIEWS------

feels this shows that events call forth the necessary person. Perhaps, but hold on. P as t probabilities We have at least five birth-orders (only child, eldest, youngest of two, middle of Patrick Rabbitt three or more and youngest of three or more) and if we take only the four political Genius, Creativity & Leadership: Historiometric Inquiries. scenarios distinguished by Stewart (civil By Dean Keith Simonton. conflict; international crisis and warfare; Harvard University Press: 1984. Pp.231. $22, £17.60. peace; and revolution), to show that one of these variables (birth-order) predicts even DEAN Simonton is a prescient man. While tions from which it was drawn. To do this lOOJo of the variance in assignment to the the rest of us glumly accept that our speci we must take into account enormous popu other (political scenario) would require alisms are of little interest to colleagues lation growth since 1450, the consequent, data from a far larger sample than Stewart trapped in their own obsessions, he has likely, changes in the percentages of the provides. There are, indeed, well-estab developed a career as a "historio population who could achieve historical lished associations between birth order and metrician". In any Common Room in the recognition of pre-eminence in successive intelligence test scores (5% of variance or world, after the initial difficulty of centuries, and the changes in the percen- less) but samples of thousands have been explaining what this means, he can emerge necessary to detect them. as an expert on the Great Academic Post - The book is shot through with intriguing Prandial Topics of All Time. Do higher examples. For instance, can we find out degrees actually hinder creativity? Can we whether originality in musical composition quantify charisma? Is history, especially is a function of historical period? the history of science, shaped by Simonton considered the first six notes of individuals or by Zeitgeists? Does the 15,618 themes produced by479 "classical" physical presence of a commander really composers between 1500 (Josquin de Pres) affect the outcome of a battle? Are and 1950 (Schoenberg and Shostakovitch). important scientific discoveries typically He examined the first-order transitional made independently by several people, or probabilities between notes in each of these by isolated individuals? And much, much note strings and related them to a (here un more. defined) baseline computation for all tran Simonton offers more than dinner-table sitional probabilities between notes. Note expertise. He is widely read, and has strings were classified as less original if training in statistical methods the their transition probabilities coincided principles of which he tries to transmit in with, and more original if they deviated five appendices. Perhaps these will not help from, the frequency of two-note transi people without, or impress people with, a tional probabilities in the (undefined) similar background, but he undertakes to source transition matrix. On this basis bring recent analytical ideas to his material. Simonton plots his composer-originality This is interesting, since the post hoc detec by year from 1500 to 1950 and finds, tion of statistical associations in sub-sets overall, a steady increase in thematic ori from vast bodies of potential data can ginality over time, with sinusoidal varia easily persuade us of very strange things. tions peaking at about 1650 (Palestrina) Does Simonton's use of some up-to-date and 1920 (Schoenberg), and with a severe statistical techniques reduce uncertainties trough between 1750 and 1820 (Haydn, about assumptions of causal relationships Mozart, Beethoven and Schubert). in very large, uncontrollable, post hoc data I leave others to argue how far two-note samples? Well, not as such. When the logic transitional-probabilities can be a valid behind selection of numbers for analysis is tages of the population examined who ever index of thematic originality because I am dubious, elaborate statistical models experienced higher (let alone "higher entranced by the possibility of a splendid cannot help. For example, on the whole higher") education from 1450 to 1900. computational tautology: if the reference Simonton favours the view that higher edu Even if such a tedious computation were transition probability matrix is, as one cation stunts intellectual flexibility. He re attempted, the Cox sample of 301 spread must suspect, computed from the entire analyses Cox's (1926) collection of bio over four-and-a-half rapidly changing cen corpus of 15,618 themes collected over 450 graphical details on "301 geniuses" extant turies would certainly be too small to allow years, this would mean that the earlier a between 1450 and 1900. Very few of them any definite conclusion. It is not that his composer flourished the more chance his took higher degrees - which, therefore, toriographers do their sums wrong, it is work had to influence his successors and so are at least uncorrelated with originality?· that apparently they would rather analyse the greater was his likely indirect influence The conclusion may well be correct, but inadequate data than hold their peace. on the baseline against which he was Simonton's argument requires some odd Historical evidence will always be judged. The steady increase in "origin assumptions. patchy. This makes it important for histori ality" over 450 years is at once explained. One is that the content, and the social ographers to recognize when they do not Further, the more eminent a composer the and career implications of "higher have enough data to answer a question. more he would influence his contem degrees" have not changed since 1450. Simonton is intrigued (pp. 165-166) by poraries and successors, and so the more Simonton apparently reassures himself by Stewart's analysis that the familial birth substantial would be his indirect contribu computing that the incidence of higher de order of "British Prime Ministers" (how tion to the "baseline" matrix. Hence the grees within Cox's sample did not change many examined?) and Presidents of the historical plunge from "originality" between 1450 and 1900. This is interesting, United States (38 studied) appears to cor shown by Mozart et al., who are very like but hardly germane. We can only discover relate with the types of political situation the matrix which their influence deter whether incidence of higher degrees in the faced by their cabinets. Elsewhere he con mined. If we leave out of consideration Cox sample is atypical by comparing it vinces himself that birth order is a potent such perturbing factors as the increased against incidence in the reference popula- force in determining personal style, and diffusion of musical ideas between geo- © 1984 Nature Publishing Group 384 ------BOOK REVIEWS------NA_ltJ_RE_VO_L_. 3_12_22_NO_VE_M_BE_R_I_CJ84_ graphically remote cultures as communi probable complexity of most mammalian cations improved over the centuries, it may Cue for cooperation chemical cues. The next six chapters deal be that we can extract one conclusion from with the potential sources of the cues and Simonton's elaborate computation: those F .H. Bronson the types of compounds that may arise whose works have survived longest, and from these sources. Thus the skin, scent those whose works are most often per Mammalian Semiochemistry: The glands, urine, breath and the reproductive formed, shape the "norms" in musical Investigation of Chemical Signals tract are considered in detail. composition! Between Mammals. Unfortunately much of this discussion Sophisticated mathematical models may By Eric S. Albone. relies upon data collected in human beings, be misleading in another way - they may Wiley: 1984. Pp. 360. £29.50, $57. and in only a few cases are sources and provide precise descriptions of numerical compounds related directly to functions data which are quite implausible in terms MosT mammals are nocturnal and by (this, it should be said, is not the fault of the of real-world processes. Simonton ele human standards they are quite small. author, but an accurate reflection of the gantly reviews data on changes in human Thus they rely heavily upon their chemical state of our knowledge). In a further creativity in politics, arts and sciences senses when interacting with other chapter that should be read by everyone between the ages of 20 and 70. Alas, though members of their own species, avoiding interested in the topic, Dr Alb one considers cheerful exceptions abound, the unmis predators or searching for food. Chemical the important role played by micro takeable trend is for peaks to be reached cues play many crucial roles even in organisms in shaping odours. Finally, a between 35 and 45 with an accelerated mammals that are larger, more diurnal, guest contributor, Dr Stephen Shirley, has decline thereafter. He neatly fits an alge and hence more visually-orientated. The written a nice primer on the neuro braic model to the data (pp.110 and 111), term "semiochemical" encompasses all physiology of the chemical senses. and on p.109 describes the functional im such cues, whether they originate from While this is a historically important plications of this model: plants or animals, and incorporates the book, dealing as it does with a young and concept of communication via pheromones. rapidly developing area, it will probably Each creator uses up [sic) this supply of creative By its very nature, the broad study of fail in its stated objective of bringing potential [determined individually - perhaps close together biologists and chemists. Much of at birth?] during his or her productive career chemical cueing requires by transforming potential into contribu collaboration between the chemist and the what the biologist finds exciting about the tions. biologist; as witnessed by its impact on subject is treated only superficially - the entomology, the result can be immensely integration of chemical cues with inform I wish that I could believe that, by prudent rewarding. Research on mammals, ation provided by the mammal's other inertia, I may yet reserve some "creative however, has thus far been dominated by senses; the shaping of responses by potential", undissipated by book reviews behaviourists and physiologists interested experience and context; the adaptive and unclouded by time, to astound my strictly in communication between functions of signalling systems; and so clapped-out contemporaries when I reach members of the same species. Dr Albone is forth. On the other hand, one can hope my 80s. one of the few chemists whose primary that the book will attract more chemists Incidentally it is encouraging for com interest is in mammalian chemical cueing, into semiochemical research, and all mon sense to note that Simonton's and this is the first account of the subject biologists interested in chemical cueing algebraic statement of his model does not written by a chemist. The strengths and certainly should read it because it will help actually have this, or any other, clear func weaknesses of the book are those that them understand how a chemist tional implication. It seems to be a neutral might be expected when a scientist of one approaches many of the problems on equation for fitting empirical data using discipline looks at a complex subject that which they work. Ultimately, real progress one invariant constant (e), two other con has its roots in another, markedly different will occur only through cooperation, and stants empirically derived for particular discipline. scientists in both disciplines need to disciplines (a and b) and a fourth constant The first two chapters of the book understand much better how their opposite empirically derived for individuals (c). It is introduce the subject and note the numbers view this challenging field. D not a parsimonious equation. potential contributions and limitations of Simonton is clearly a very hard-working chemical investigations; of particular F. H. Bronson is Professor of Zoology at the academic, and seems sincerely anxious that interest here is the author's emphasis on the University of Texas, Austin. his subject should help us to understand ourselves better, to learn to anticipate and to avoid political violence, and to give each attractive features in their own right, with other just credit for real achievements. His SOM introduction the result that commercial scanning book will be marvellous bed-time reading electron microscopes have been with us for for any academic with a curiosity about his P.W. Hawkes nearly 20 years and commercial scanning torical or statistical matters. Only part of transmission electron microscopes for its entertainment value lies in the detection Theory and Practice of Scanning Optical about a decade. Now, as this book of charming errors and non sequiturs. Microscopy. convincingly demonstrates, the scanning Simonton has an acute sense of really By Tony Wilson and Colin Sheppard. optical microscope (SOM) is beginning to interesting, though possibly insoluble, Academic: 1984. Pp.213. £25, $39.50. supplement its conventional ancestor. problems, and whenever he seems to go No two people are better qualified to wrong he sets us the important task of won THERE are not many types of radiation for write a book on SOMs than Tony Wilson dering whether, and precisely how, some which high-quality, two-dimensional, and Colin Sheppard, for almost all the thing useful may yet be said. In spite intensity-sensitive recording media exist. work on the theory and performance of (surely?) of the most extreme temptations Light and electrons are exceptional in this these instruments has come from their he brings to his material no detectable respect, so that (apart from a few early Department of Engineering Science in humour. But, for the reader with his own scanning electron instruments) scanning Oxford. In extremely simple language, and sense of fun, this will be a delightful microscopes were developed long after using mathematics that anyone familiar book. D their fixed-beam counterparts had become with Sections 9.5-10.5 of Born and Wolf's commonplace. It is, however, gradually Principles of Optics will find easy to of Patrick Rabbitt is Professor ofGerontology and being recognized that the optics follow, they take us through the theory of Cognitive Psychology at the University of scanning systems and the sequential nature image formation in the SOM. The two Manchester. of the image-forming process possess operating modes - essentially large