CHAPTER 4
The question of totemism
The reception of Ancient Society was generally friendly, an indication that a new consensus had established itself on the question of primitive social structure. The degree of consensus might he obscured by the continuing controversies, which had begun wiih polemical exchanges between Maine, McLennan and Morgan, yet the disputes were increasingly technical and limited in scope. They tended to concern such recondite matters as the meaning of kinship terms. Agreement on fundamentals became more general. Most authors now converged on a single model of primitive sociery. The conflict between the advocates of primitive 'matriarchy' and 'patriarchy' was resolved. 'Matriarchy' was generally held to characterize the most primitive societies; 'patriarchy' typified a higher level of social development. Both forms of organization were based upon group exogamy. Primitive societies had all been organized ■ in these principles of descent and exogamy for many millennia, until at last the revolutionary transition occurred from the original kinship-based polity,  in which property was held in common, to a territorially-based state and a system of private property. The agreement on substantive issues was matched by a convergence on questions of method. Maine's approach - loosely inspired by the example of Indo-European philology - had restricted comparison to societies within a single cultural tradition, which might be identified with a particular 'race'. In the 1860s this versii m of the comparative method lost ground in Britain and America, ai least among those writers who were interested in 'primitive societies' . The more universal, not to say promiscuous, scope of Enlightenment historiography made a comeback. McLennan, Morgan and above all Tylor assumed that successive types of society all over the world had developed out of one primitive ancestral form.
The question of totemism    11
according to a fixed programme.1 In other words, a unilinear evolutionism was an unquestioned component of the new consensus.
But just when a consensus was established on the nature of 'primitive society', the focus of anthropological interest shifted. The first wave of anthropological studies had responded to a variety of political concerns. Its authors belonged to a generation coming to terms with the Reform Bill, the crisis in India, and the Civil War in America. The main theme of their anthropology was the development of political institutions. But, increasingly, political issues began to seem less important than questions of belief and rationality.
The publication in 1871 of Darwin's The Descent of Man served to concentrate the minds of many intellectuals. Its effect was twofold. First of all, anthropologists were bound to ask what the connection was between man's biological evolution and his intellectual development. In The Descent of Man Darwin emphasized the evolutionary significance of language and other cultural techniques, which depended on man's powers of observation, memory, curiosity, imagination and reason.2 He suggested that the growth of language might even have stimulated the development of the brain. The attention of anthropologists, accordingly, was now firmly directed on the evolution of language and cultural techniques, and on man's intellectual development.
Secondly, The Descent of Man reinforced the scientific challenge to the traditional religious view of creation. Many intellectuals now believed that science was actually taking over the explanatory role of religion. Perhaps the intellectual development of mankind was tending to a stage at which faith in magic and religion would be abandoned and man would turn to rationality and science.
The anthropology which emerged to deal with this new agenda was personified above all by an English scholar, E. B. Tylor. Tylor had been born in 1832 and, as his friend Marett remarked, he grew up just when the political preoccupations of the old generation gave way to the religious obsessions of the new.
1  This shift was by no means total. Much German and French writing continued to use the culture-area framework. In France, the rise of Durkheimian sociology brought with it an acceptance of Anglo-American evolutionism, but the German geographical school persisted in their preference for culture-area approaches.
2  This is a central theme of Part 1 of The Descent of Man. Another is the question whether - and to what degree - animals share these capacities.
78    THE CONSTITUTION OF PRIMITIVE SOCIETY
Fortunate in the time of his birth - the hour of the Reform Bill - Tylor had reached his prime just when in England intellectual, following on the heels of political, liberation was calling for recruits in the inevitable struggle with the die-hards of the old order.3
From the first he took little interest in political or social questions: his anthropology dealt with man's intellectual development.
The rise of E.B. Tylor4
Born into a well-to-do Quaker industrial family, Tylor did not attend a university. On a youthful tour abroad a chance meeting on a Havana omnibus with a fellow-Quaker, Henry Christie, led him to accompany Christie to Mexico, where he was engaged in archaeological research. Christie had been an early convert to Darwinism, and his interest in archaeology had been quickened by Boucher de Perthes's discoveries in the Somme valley which established the antiquity of man.
In Mexico Christie and Tylor became fascinated by the complex pre-Conquest Mexican civilization. How had it arisen? Was it the result of diffusion from the Old World, or the fruit of independent development? Did it provide evidence for human progress, or, on the contrary, for degeneration?
The leading theorist in Britain concerned with these questions was Tylor's contemporary, Sir John Lubbock (Lord Avebury)/ Lubbock was even more of a Darwinian than Christie - almost a birth-right Darwinian in fact, as a country neighbour of the Darwins and a lifelong friend. When the storm broke over him with the publication of The Origin of Species, Darwin had written to Lubbock (then only twenty-six years old), 'I settled some time ago that I should think more of Huxley's and your opinion - than ofthat of any other man in England'.6
Lubbock's Prehistoric Times, published in 1865, established a
3  Maren (1936), Tylor, pp. 212-13.
4  An account of Tylor's thought is to be found in J. Leopold (1980), Culture in Comparative and Evolutionary Perspective. Burrow (1966), in Evolution and Society, pp. 234-59, is particularly good on Tylor's 'ulterior motives' which he sees as broadly theological. Marett's Tylor (1936) is still worth reading, as the appreciation of a friend.
5   See Daniel (1950), A Hundred Years of Archaeology. Cf. Burrow (1966), Evolution and Society, pp. 228-9.
6  Quoted by Burrow (1966), Evolution and Society, pp. 228-9.
The question of totemism   79
quasi-evolutionary theory for archaeology. The remains of ancient cultures exhibit a serial progression, like the fossil varieties of contemporary animal species. Strata of social evolution could be uncovered and related in sequences. Technology demonstrated unmistakably that progress had occurred. Later techniques were obviously advances on earlier techniques. This proved that man himself had progressed in his capacities, and put paid to any talk of degeneration from a higher condition. In the late 1860s Tylor formed a close alliance with Lubbock and with the other Darwinians, Huxley and Wallace, and technological progress became one of his major themes.
A second issue which concerned Tylor from the period of his Mexican fieldwork was the development of language. This was an interest of Darwin, but perhaps the key influence on Tylor was Max Müller. Müller had brought from Germany both the narrower philological tradition and also a more speculative interest in the notion of linguistic development. Tylor was especially interested in the evolutionary view of language, and his second major theme was the development of forms of language, including the language of children, gesture language, picture writing, etc. The assumption was that advances in language reflected intellectual progress. The activity which most clearly exhibited these relationships was mythology, to which Tylor, like his German models, devoted considerable attention. In his Researches into the Early History of Mankind and the Development of Civilization (1865), Tylor brought together the themes of linguistic, mythological and technical development to buttress a general argument about the intellectual progress of mankind.
But Tylor skirted the problem of religion in his Researches, rather as Darwin had postponed a detailed consideration of the evolution of the human species itself in Origins. Yet in both cases the next question on the agenda was obvious. In 1866 Tylor published an essay entitled 'The religion of savages', which appeared in The l-'ortnightly Review. The argument was expanded and developed in his Primitive Culture, published in 1871. The first volume ofthat book was essentially a revamp of his earlier Researches, but the second volume was devoted to the development of religion.
It must be remembered that 1871 was also the year that Darwin's Descent of Man was published, applying the theory of evolution to explain the emergence of the human species. Tylor's book added what was potentially an equally devastating challenge to orthodox
80     THE CONSTITUTION OF PRIMITIVE SOCIETY
Christians. He argued that even the earliest men had some form of religious belief. Religions could be ranged in a series according to intellectual sophistication, but later religions all derived from a primitive system of theology, and retained traces of their origins. The clear implication was that classical Christianity might have been outgrown by modern man. Increasingly, religious belief would yield place to scientific theory.
The earliest religion was based upon a series of intellectual (and partly linguistic) confusions between the self and the other. In dreams people saw themelves roaming about in strange places, where they met other people, some of them long dead. This experience led to the belief that every man had a double existence, corporeal and spiritual. And if people had spirits, why not animals, or even inanimate natural objects? The earliest coherent form of religion was based on 'the theory which endows the phenomena i if nature with personal life'. Tylor called it 'animism'.7
This primitive animism was not merely of antiquarian interest. Even the most civilized societies suffered something of a hangover from the animism of their forefathers. Vestiges of outworn primitive cults could be traced in the ceremonies of the most advanced religions. These were 'survivals', the fossils of cultural institution1-. Quite often an ancient rite was still celebrated, even though it might be given a fresh rationale. Tylor's main example of such a rite was sacrifice, and this was a most significant choice. The question of sacrifice was central to the exegesis of Biblical Judaism. In the context of communion, it was also a major point of contention between Catholics and Protestants, and indeed between more ritualistic Protestants and the anti-ritualists, prominent among whom were, of course, the Quakers.
Tylor argued that rituals of sacrifice preserved very primitive religious notions. Sacrifice was so ancient, indeed, that even the explanations given in the Old Testament were anachronistic. To understand the primordial purpose of sacrificej it was necessary to place it in its original context, animism. In animistic religion*, offerings were made to the spirits of the dead after they had appeared in dreams. Later, sacrifices were made to 'other spiritual beings, genii, fairies, gods'. The rationale was quite evident: 'the
7 Tylor (1866), 'The religion of savages', p. 84.
The question of totemism    81
object of sacrificing to the gqds is that they are to consume or enjoy the souls of the things sacrificed'.8
Tylor's ideas had a remarkable success. In 1871, the year of Primitive Culture, not yet forty years old, he was elected a Fellow of the Royal Society. In 1875 Oxford awarded him an honorary degree. In 1881 he published the first general textbook in English on the subject, his Anthropology, which held the field for a generation. In 1884 Oxford created a Readership in Anthropology for him, and in 1896 he was made a Professor by personal title. By now the likes of Max Müller were talking of anthropology as 'Mr Tylor's science'.
Yet however timely, his work was clearly not original. His theory of technological progress was drawn from Lubbock, his notions about the development of language and mythology (despite some differences of opinion) from Müller. His theory of religious development owed a great deal to Comte, and his 'animism' is hardly to be distinguished from Comte's 'fetichism'. Even his ideas about sacrifice owed much to the German biblical scholar Wellhausen.
To a later generation Tylor's most memorable contribution was his idea that technology, language, myth and belief formed a single entity, which, following an already well-established German tradition, he termed 'culture or civilization'. This he defined as 'that complex whole which includes knowledge, belief, art, morals, law, custom and any other capabilities and habits acquired by man as a member of society'.9 Tylor emphasized man's intellectual development, and he increasingly distinguished the history of culture from the history of race (in his practice if not in principle). Yet the introduction of the German idea of culture history did not seem a great novelty to contemporaries. British scholars would certainly have noticed an obvious kinship with Spencer's idea of the 'super-organic'.
Andrew Lang, reviewing his friend's career in his introduction to Tylor's Festschrift, conceded Tylor's lack of originality but suggested that 'his merit lay in his patient, sagacious, well "documented", and, at last, convincing method of exposition'.10 In other words, Tylor was a synthesizer. The reception of his synthesis
8  Op. cit., p. 77.
9  This famous definition is given in the opening sentence of Primitive Culture (Tylor, 1871).
10  In Balfour et dl. (1907), Anthropological Essays Presented to E. B. Tylor.
82     THE CONSTITUTION OP PRIMITIVE SOCIETY
reflected his identification with the Darwinians and also the swing of interest away from political issues and towards religious questions.
Yet despite the new prominence of religious questions in anthropology, the consensus about the form of primitive society was not abandoned. Rather, it now became desirable to relate it to Tyler's ideas about the origin of religion. What was the link between animism and exogamy? The theory which made the strongest bid to resolve this question was put forward by McLennan, who called it 'totemism'.
McLennan and the invention of totemism
Frazer's biographer pointed out that totemism, 'like radar, whiskey and marmalade, was a Scottish discovery or invention, for it was first defined by the Edinburgh lawyer John Ferguson McLennan, and Frazer, Robertson Smith and Andrew Lang were among the first to discuss it'.11 Of these McLennan was the least scholarly but perhaps the most original. At any rate, it fell to him for the second time to launch a theory, one which was to have an even greater long-term impact than his theory of primitive marriage. The new-theory was set out in a two-part essay entitled 'The worship of animals and plants', and was published in The Fortnightly Review in 1869-70.u Tylor's essay, 'The religion of savages', had appeared in the same journal three years earlier.
McLennan took for granted Tylor's thesis that primitive peoples worshipped fetishes, which they believed to be animated by anthropomorphic spirits. Man was conscious of his own spiritual force, and ascribed a similar power to natural objects. These animistic beliefs, however, gave rise to a new religion that McLennan called 'totemism'. 'Fetishism resembles Totemism', he wrote with splendid effrontery, and indeed it turned out that totemism 'is Fetishism plus certain peculiarities. These peculiarities are, (1) ihe appropriation of a special Fetich to the tribe, (2) its hereditary transmission through mothers, and (3) its connection with the jus connubii.n3 In other words, totemism was fetishism but given a sociological anchor in McLennan's primordial society.
11   Downie (1970), Frazer and the Golden Bough, p. 76.
12  But cf. his entry, 'Totemism', in Chambers' Encyclopaedia (1868).
13  McLennan (1869-70), 'The worship of animals and plants', part 1, p. 422.
The question of totemism    83
According to McLennan, some primitive peoples believed that they were of the same species as their totem. They were descended (in the female line, of course) from the original totemic animal. Some had even developed theories about such transformations from animals to men. Totemism might therefore almost be regarded as a first, faint hint of Darwinian evolution. The most primitive religion sowed the seeds not only of later religions, but also of science.
Most of McLennan's paper was given over to illustrations of 'totemism'. The key areas were aboriginal America and Australia. Basing himself on Sir George Grey, and allowing himself remarkable interpretative licence, McLennan deduced the existence in Australia of exogamous matrilineal groups, gentes, whose members share a common totem. Despite the absence of further evidence, he expressed his confident belief that a similar system prevails, or prevailed, throughout Oceania. For America, he relied especially on Gallatin, concluding that here too totems were associated with matriarchal, exogamous groups. A rapid review of world ethnography revealed elements of a totemic system among the tribes of Siberia, Peru, Fiji, and even in classical India. Classical Europe itself exhibited traces of totemism, as did ancient Israel - what else was the serpent story in Genesis? The use of animal terms for constellations of stars was equally a totemic derivation.
Robertson Smith and 'The Religion of the Semites'
In 1866 a group of intellectuals had formed the Edinburgh Evening Club. McLennan was a founder member. Another was the theologian W. Robertson Smith, who became his friend.14 In 1870 Smith was appointed to the chair of Hebrew and Old Testament at the Free Church College at Aberdeen. Here he began to propagate the new critical approach to the Bible which he had learned in Germany from Julius Wellhausen, and which was to lead him to heresy - and also to the adoption of his friend's theory of totemism.
Another Scottish intellectual institution was implicated in Robertson Smith's first crisis. This was the Encyclopaedia Britan-
14 See John Black and George Chrystal (1912), The Life of William Robertson Smith. Another useful source is Beidelman (1974), W. Robertson Smith and the Sociological Study of Religion.
84    THE CONSTITUTION OF PRIMITIVE SOCIETY
nica, which published the first volume of a new edition in 1876, carrying articles by Smith on 'Angeľ and 'Bible'. These expressed his view (derived from Wellhausen) that the Bible was a compilation of sources of various dates, and including mythological as well as historical elements. In 1878 he answered charges of heresy before the Church Assembly, and eventually, in May 1880, he was cautioned. In April, however, he had visited McLennan in Italy, and in June he published an essay entitled 'Animal tribes in the Old Testament', which applied McLennan's theory of totemism to the Bible. A decade after the publication of McLennan's original article, his theory was to receive extraordinary publicity as the central issue in a theological cause célěbre.
In his new essay, Robertson Smith argued that his own researches in the Semitic field confirmed McLennan's theory of totemism. Arabic pre-Islamic sources indicated that tribal groupings were often named after animals, and sometimes after the moon and sun. Since sun and moon were evidently worshipped as gods, animals presumably once had a similar status. Furthermore, tribes worshipping the moon were believed to be descended from their god. The same might well have been true of tribes named after
animals.
This demonstration seems peculiarly thin, and Robertson Smith himself admitted 'that we have very little direct information connecting these facts with animal worship'. But there was a reason for the absence of direct evidence. Greek sources are unreliable on these topics, and Islamic authors censored heathen ideas - 'we must remember the nature of the records'.15
McLennan's theory also required that the original totemic tribes should be matriarchal and exogamous. Robertson Smith claimed that this was indicated by some sub-tribe names, which might 'denote the offspring of one mother'.16 Furthermore, some sources seemed favourably disposed to female infanticide, which again supported McLennan's theory. Strabo also reported traces of polyandry. The marriage rites of some Arab peoples apparently contained survivals of exogamy and marriage by capture. The Queen of Sheba's existence pointed to a pre-patriarchal form of organizal ion. Taken together, 'These facts appear sufficient to prove that Arabia
15  Robertson Smith (1889), Religion of the Semites, p. 84.
16  Op. cit., p. 86.
The question of totemism    85
did pass through a stage in \yhich family relations and the marriage law satisfied the conditionsxof the totem system'.17 There was evidence of the same sort which suggested that totemic elements had survived in ancient Israel, if in an attenuated form. Robertson Smith suggested that the heathen practices against which the prophets inveighed were totemic in origin, and that the second commandment itself was directed against nature worship.
Coming from a man who had just been warned to mind his step by the Church Assembly, this was a provocative argument. The General Assembly did not mince words in its reaction to the paper.
First, concerning marriage and the marriage laws in Israel, the views expressed are so gross and so fitted to pollute the moral sentiments of the community that they cannot be considered except within the closed doors of any court of this Church. Secondly, concerning animal worship in Israel, the views expressed by the Professor are not only contrary to the facts recorded and the statements made in Holy Scripture, but they are gross and sensual - fitted to pollute and debase public sentiment.18
Smith was removed from his professorship in May 1881, but he was not cast into the outer darkness. He became co-editor of the famous ninth edition of the Encyclopaedia Britannica (and was reputed to have read every entry). In 1883 he was appointed Reader in Arabic at Cambridge and in 1889 he became Professor. When he died in 1894, at the age of forty-eight, he was already widely regarded as a great scholar.
In Cambridge Smith acquired his most important apostle, another fellow of Trinity, and another Scot, James George Frazer, whom he at once commissioned to write an entry on 'Totemism' for the Encyclopaedia. He also developed his own ideas on early Semitic religion and social organization, notably in his entry on 'Sacrifice' in the Encyclopaedia, in his book Kinship and Marriage in Early Arabia (1885), and finally in his masterpiece, Lectures on the Religion of the Semites (1889).
Kinship and Marriage in Early Arabia is far more fully argued and documented than the original 1880 essay on 'Animal worship', but the ideas on primitive Semitic society are essentially the same. The main thrust of the book is that, despite obvious indications to the contrary, the strongly 'patriarchal' societies of ancient Arabia
17  Op. cit., p. 88.
18  Cited in Beidelman (1974), W. Robertson Smith, p. 21.
86    THE CONSTITUTION OF PRIMITIVE SOCIETY
were preceded by 'matriarchal' communities. The evidence was (necessarily) indirect and slight. The conclusions which were drawn could be sustained only if McLennan's theory of a universal totemic stage was correct. Smith wrote:
In enquiring whether the Arabs were once divided into totem-stocks, we cannot expect to meet with any evidence more direct than the occurrence of such relics of the system as are found in other races which have passed through but ultimately emerged from the totem stage.19
The stagnation of Robertson Smith's thinking on primitive social structure contrasts strikingly with the vitality of his later work on totemic religion, which was concerned especially with the problem of sacrifice. The traditional theological view derived from the priestly code, according to which sacrifices were essentially acts of atonement. Smith's mentor Wellhausen had rejected this interpretation as anachronistic. Textual criticism revealed that the priestly code was a post-Exilic document. It superimposed a late-priestly theology on earlier ritual practices. Originally sacrifices were not even performed in the Temple. They were associated with what Wellhausen called a natural religion, which was situated within the life of the family. Smith developed this conception and linked it with totemism.
The argument was most fully developed in Lectures on the Religion of the Semites (1889), Smith's last book and by a considerable margin his best. He emphasized particularly two themes. The first was methodological. Rites were the most authentic pointers to earlier religious ideas. The myth or dogma with which a rite is associated may be a later accretion. As Smith put it, the 'ritual was fixed and the myth was variable, the ritual was obligatory and faith in the myth was at the discretion of the worshipper'.20 A later generation read this as a proto-functionalist statement. It was. rather, a methodological principle, derived from Tylor, for the identification of survivals.
Secondly, Roberton Smith developed a sociological argument: 'the fundamental conception of ancient religion is the solidarity of the gods and their worshippers as part of one organic society'. And again: 'gods and men, or rather the god and his proper worshippers, make up a single community, and ... the place of the god in the
19  Robertson Smith (1885), Kinship and Marriage in Early Arabia, pp. 187-8
20  Robertson Smith (1889), Religion of the Semites, p. 18.
The question of totemism    87
community is interpreted on the analogy of human relationships'.21 There was in fact a religion especially appropriate to a clan-based society, and another more appropriate for a state. 'We now see that the clan and the state are both represented in religion: as father the god belongs to the family or clan, as king he belongs to the state.'22
In ancient Israel this relationship of divine fatherhood was conceived of in spiritual terms. In more primitive societies, people believed that they were physically descended from the founding god. This was the original religious conception. It was equally the origin of morality, for 'the indissoluble bond that united men to their god is the same bond of blood-fellowship which in early society is the one binding link between man and man, and the one sacred principle of moral obligation'.23
Totemism represented perhaps the earliest stages of this conception. In a totemic religion, the gods were natural species, generally animals. These gods were associated with natural sanctuaries, which followers had to visit. Even early religions had holy places, shrines. But at certain times a yet more intimate contact with the gods was required. This was achieved through sacrifice, 'the typical form of all complete acts of worship in the antique religions'.24
Sacrifices could take one of two main forms. A vegetable sacrifice was thought of as a tribute or gift (which was Tylor's conception). Alternatively, one could sacrifice an animal, but animal sacrifices were a different matter entirely. They 'are essentially acts of communion between the god and his worshippers'. 'The god and his worshippers are wont to eat and drink together, and by this token their fellowship is declared and sealed.'25 Since pastoralism preceded agriculture, animal sacrifices were anterior to vegetable sacrifices.
But what animals were sacrificed and eaten at these feasts? Originally, one of the sacred animals themselves. A totem animal could not normally be killed or eaten, but this very taboo made the sacrifice sacred. 'The evidence ... is unambiguous. When an unclean animal is sacrificed it is also a sacred animal.'26 Among the
2i Op. cit., pp. 32 and 85.
-2 Op. cit., p. 40.
-3 Op. cit., p. 53.
-4 Op. cit., p. 214.
-5 Op. cit., p. 243 and p. 271.
-6 Op. cit., p. 294.
88     THE CONSTITUTION OF PRIMITIVE SOCIETY
Semites, then, 'the fundamental idea of sacrifices is not that oi' a sacred tribute, but of communion between the god and his worshippers by joint participation in the living flesh and blood of :i sacred victim'.27 Ideas of atonement through sacrifice were a laicr. more sophisticated gloss on a primitive totemic rite.
The argument was clearly leading up to a climax in which something would have to be said about the sacrifices of gods themselves in Semitic religions, perhaps during communal meals. Smith took the step in this passage:
That the God-man dies for His people and that his Death is their life, U an idea which was in some degree foreshadowed by the oldest mystical sacrifices. It was foreshadowed, indeed, in a very crude and materialisme form, and without any of those ethical ideas which the Christian doctrine of the Atonement derives from a profound sense of sin and divine justice. And yet the voluntary death of the divine victim, which we have seen u< be a conception not foreign to ancient ritual, contained the germ of the deepest thought in the Christian doctrine: the thought that the Redeemer gives Himself for his people.
Frazer cited this passage in his obituary essay on Smith and remarked that it was dropped in a later, revised edition of t lie Lectures.2* Yet even if it was left implicit, the theological implications of this view of sacrifice would have been evident to any contemporary scholar. They were certainly clear enough to Robertson Smith. The theological reverberations are enough to account for Smith's obsession with McLennan's theory of totemism; and they account also for the extraordinary fame later enjoyed by Frazer, who developed and popularized Robertson Smith's theories.
Frazer and 'The Golden Bough'
Frazer was a shy young classicist, upon whom the charismatk' Robertson Smith initially exercised a most powerful intellectual influence. Robertson Smith commissioned from him entries on 'Taboo' and 'Totemism' for the Encyclopaedia Britannica, and Frazer fulfilled this commission with the thoroughness which became his hallmark. Characteristically, the resulting entry on
27   Op. cit., p. 345.
28   Frazer's 1894 obituary of Robertson Smith.
The question of totemism    89
totemism was far too lorlg.29 In the event, an abridged version appeared in the Encyclopaedia, but the complete essay was published by Black in book form in 1887. It was to be the authoritative source on the topic for the next decade. Baldwin Spencer himself recorded that when he went into the field, 'my anthropological reading was practically confined to two works, Sir Edward Tylor's "Primitive Culture" and Sir James Frazer's little red book on "Totemism'.'30
McLennan had been working on a refinement of his theory of totemism, but he died before it could be completed. His brother, Donald McLennan, issued some of his unpublished work in a book entitled The Patriarchal Theory, which appeared in 1885, and in his introduction he remarked that his brother had hoped to relate totemism to exogamy. He had come to believe that totemism preceded exogamy. It had existed in all 'rude societies', though its origin remained mysterious.
Frazer's contribution was less imaginative, but more methodical. He distinguished different categories of totem. There were clan totems, sex totems and individual totems. Clan totems were by far the most important, and clan totemism was at once a religious system and a social system. Its social form was a system of exogam-ous clans in which descent was traced in the female line. The religious aspect 'consists of the relations of mutual respect and protection between a man and his totem'.31 Normally there was a prohibition on killing and eating the totem. These religious and social aspects had drifted apart in the course of time, but originally they were inseparable - 'the further we go back, the more we should find that the clansman regards himself and his totem as beings of the same species'.32 The most primitive form of the belief was that the totem was the ancestor of the clan.
29  Robertson Smith wrote to the publishers: I hope that Messrs. Black clearly understand that Totemism is a subject of growing importance, daily mentioned in magazines and papers, but of which there is no good account anywhere - precisely one of those cases where we have an opportunity of being ahead of everyone and getting some reputation. There is no article in the volume for which I am more solicitous. I have taken much personal pains with it, guiding Frazer carefully in his treatment; and he has put about seven months' hard work on it to make it the standard article on the subject. We must make room for it, whatever else goes. (Cited by Beidelman (1974), W. Robertson Smith, p. 24)
30  Spencer (1928), Wanderings in Wild Australia, vol. 1, p. 184.
31   Frazer (1887), Totemism (first edition), p. 3.
32  Ibid.
90     THE CONSTITUTION OF PRIMITIVE SOCIETY
Frazer reviewed the ethnography far more systematically than McLennan or Robertson Smith had done, providing a genuinely encyclopaedic review of the available literature. He suggested thai totemism existed in many culture areas, although without making definite claims for its original universality. But he did not commit himself to a theory of totemism. 'No satisfactory explanation of the origin of totemism has yet been given.'33
In his most famous book, The Golden Bough, first published in 1890, Frazer took up Robertson Smith's central preoccupation, ihc sacrifice of the god.34 He constructed an ethnological detective slory which began with the ritual killing of the priest of Aricia, 'the King of the Wood'. This sacred king was the embodiment of a tree-spirit, and it turned out that he was not simply murdered, but rather sacrificed to ensure the fertility of nature. The clues were drawn from a vast range of ethnographic sources, all tending to show that primitive people identified their well-being with the fate of animistic spirits, whose priests were sacrificed in fertility ritun.K. The precise motive, however, remained a little mysterious. Perhaps it all had something to do with ancient beliefs about the transmigration of human souls into natural objects.35
The Golden Bough was enormously successful. For many educated readers it offered an irresistible combination of classical scholarship, exoticism and daring rationalism. It was far more appealing than his study of totemism, and Frazer soon began to distance himself from totemism and also from Robertson Smith. In ľne preface to the second edition of The Golden Bough, which appeared in 1900, he strengthened his own claims to originality with the statement that 'the worship of trees and cereals ... is neither identical with nor derived from a system of totemism'. And he positively disavowed the theories of Robertson Smith. 'I never assented to my friend's theory, and, so far as I can remember, he never gave me a hint that he assented to mine.'36
33  Op. cit., p. 95.
34  He also drew heavily on the work of the German folklorist, Mannhardt, wlv had tried to construct an 'Aryan' mythology from peasant folktales. Mannhar'.ii'< theory drew upon ideas about animism, and he explained peasant rite in Germany as 'survivals' of ancient fertility rituals. See, e.g., his Der Baumkttiut* der Germanen und ihrer Nachbarstämme (1875).
35  Frazer borrowed this idea from a Dutch scholar, Wilken, who reported a Klic! along these lines in Malaya.
36  Frazer (1890 (1900)), The Golden Bough, 2nd edition, p. 3. Cf. R. Ackermann (1975), 'Frazer on myth and ritual'.
The question of totemism   91
One reason for this shift was presumably simple vanity. However, Frazer preferred to attribute any changes in his thinking to the impression of new ethnographic materials, and he set great store by his accumulation of data.37 He quite compulsively expanded his store of ethnographic analogies, and stimulated ethnographers in the field with his letters and questions. In time he built up what amounted to an international intelligence service, sending out questionnaires on topics which interested him, urging on his proteges, passing on and publishing their letters.
Frazer's network eventually extended into Africa and Asia, but he believed, with other experts, that ethnographic materials from Australia were going to prove of decisive importance. By the turn of the century, totemism had come to be equated with Australian totemism. The speculations of McLennan and Robertson Smith were being put to the test in the central and northern territories of the vast island continent.
'Hypotheses are necessary but often temporary bridges built to connect isolated facts,' he remarked in the Preface to the second edition of The Golden Bough (1900). 'If my light bridges should sooner or later break down ... I hope that my book may still have its utility and its interest as a repertory of facts.'