also expect to find that their cult is modelled there as elsewhere on the cult of the great gods. As totemism had been almost completely metamorphosed by subsequent developments of religion, we need not expect to find much of it in the guardian spirits and family gods of the Old World; and if the idols in which the Chemis of the Antilles dwelt had come to be anthropomorphic in some cases, we need not be surprised if they were invariably anthropomorphic in Greece or Rome, nor if the tutelary deities of families or individuals in those countries were drawn from the ranks of the community's gods, as was the case in Polynesia.

Amongst the Semites, the teraphim, the worship of which was apparently not considered idolatrous amongst the Hebrews, were family gods. They were figures of wood or metal, with heads shaped into the likeness of a human face; they served as house-oracles, and were worshipped by the Chaldæans and by the inhabitants of Syria.1

That at Rome the Genius was the guardian spirit of the individual, and that the Lares and Penates were family gods, no one will question. It is, however, interesting to note that both the Genius and the Lares are associated with animals, the former with the snake and the latter with the dog, and so betray probably their totemistic origin. The life of the individual was in some cases supposed to depend upon the life of the snake in which his genius lived; the man's health depended on his genius," and "when the serpent which was the genius of the father of the Gracchi was killed, Tiberius died." 3 This exactly agrees with the account given of the individual totem amongst the Guatemaltecs: many "are deluded by the Devil to believe that their life depends on the Life of such and such a Beast (which they take to them as their familiar Spirit), and think that when that beast dies they must die; when he is chased, their hearts pant; when he is faint, they are faint; nay, it happens that by the Devil's delusion they appear in the shape of that Beast (which commonly by their choice is a Buck or Doe, a Lion or Tigre, Dog or Eagle), and in that Shape have been shot at and wounded."4

1 Am Urquell, v. 92.

4

Preller, Römische Mythologie,3 ii. 198.

3 Jevons, Plutarch's Romane Questions, xlviii.

Gage, A New Survey of the West Indies, 334.

The resemblance between the Guatemaltec belief and the European belief about the wounding of witches is so close as to suggest that the animal in which the familiar spirit of the European witch appeared may have been a last lingering survival-like the serpent of the Gracchi and the Genius of the Romans-of an individual totem. The dog, with which the Lares are associated, appears in European folk-lore as a form in which ghosts manifest themselves, and the Lar is conceived not only as the house-spirit but as the spirit of a deceased ancestor. Probably we here have ancestor-worship amalgamating with the worship of a guardian spirit, who originally appeared in totem shape. In Polynesia, a deceased ancestor, and not a god, is sometimes chosen as a totem,2 but that is an exception to the general rule, and probably late.

In Greece, the Athenians distinguished between col πάτριοι, ἱερὰ πάτρια and θεοὶ πατρῷοι, ἱερὰ πατρῷα. The former were certainly the national gods. Whether the latter were family gods is less certain. On the one hand, the privilege

of worshipping them seems to have been confined to trueborn Athenians, and to have been a mark of full citizenship,3 which would show that they were the gods of the Athenians as distinguished from other Greeks. On the other hand, their worship was carried on in the private houses of those qualified to worship them, which rather points to their being family gods drawn, as in Polynesia, from the ranks of the community's gods. These θεοὶ πατρῷοι οι ἑρκεῖοι οι μύχιοι were worshipped in the μvxoí of the house, and one of them was apparently Hecate, to whom the dog was sacred; and the dog is, as we saw, associated with the household gods of the Romans also. An apparent trace of guardian spirits in Greece is the Hesiodic doctrine of Saíuoves and what is obviously implied in the word evdaipov, namely, that the man to whom the word is applied has a good Saíuwv. The ȧyalòs Saíuwv, again, like the genius of the Romans, appears as a snake; and there was a variety of harmless snake, the specific name of which was ἀγαθοδαίμονες. We may note that before Hesiod, i.e. in the Homeric poems, there is no

1 Jevons, op. cit. xl.-xlii.

Ar. Ath. Pol.

5 Rohde, op. cit. 233.

2 Waitz, Anthropologie, vi. 317, 321, 324. 4 Eur. Med. 397; Rohde, Psyche, 232.

mention of ancestor-worship, and after him no cult of guardian spirits. Whether we are to connect these two facts, and infer that ancestor-worship, springing up in postHomeric times, amalgamated with the cult of the guardian. spirit (as in Rome with the cult of the Lar), and then overshadowed it altogether, is a point which I will not do more than suggest for consideration. At any rate, it is obviously desirable that we should now go on to consider the question of ancestor-worship in general; and, bearing in mind that it is essentially a private worship and a purely family affair, we may not inappropriately sum up the results of this chapter as affecting cults of this kind. They are as follows. Whenever and wherever cults of this kind are found-and they are found in every quarter of the globe-they are assimilated to the ritual used in the worship of the community's gods; and the tutelary spirits themselves assume the same external form as the public gods. Next, it is more probable that the individual should imitate the community's ritual than the community an individual's; and in some cases it is avowedly the individual that borrows his guardian spirit from the ranks of the community's gods. Finally, the family is an institution which appears relatively late in the history of society. If, therefore, we find points of similarity between the ritual used in ancestor-worship and that used in the worship of the public gods, we shall not fall into the error of treating it as an isolated and unparalleled fact in the history of religion, but shall rather regard it as subject to the same laws and to be explained in the same way as the rest of the class of private cults to which it belongs.

1

CHAPTER XV

ANCESTOR-WORSHIP

A DESCRIPTION has already been given, in Chapter V., of the spontaneous outbursts of sorrow, " the indescribable scenes of lamentation and wailing," as Mr. Turner calls them, which take place amongst savages on the occasion of a death; and of the uncertainty whether death has really supervened, the reluctance to believe that it has, the endeavours to detain the soul of the dying man by offering him his favourite dishes, displaying his most cherished possessions, praising his noble deeds; the attempts to recall the soul, when the man is dead, to induce it to abide with the survivors; in fine, the desire to dwell on the memory and to seek communion with the spirits of those who have been loved and lost. The object of that chapter was to suggest that the avenue of communication thus opened between the savage and the spirits of his dead may have served to suggest to him a way of approaching other beings, who like the dead were spirits, but unlike them possessed supernatural powers; for the dead do not seem, in any of the ceremonies described, to appear as supernatural beings. The being with whom the savage seeks communion in these rites is "the father whom he knew," not a dæmon of any kind. At death, as in sleep, the spirit deserts the body, but does not in either case necessarily thereby gain supernatural powers. After death, indeed, the ghost's relation to the living is rather one of dependence, for food, comfort, and even continuance of existence. In fine, these spontaneous demonstrations of affection, grief, and desire for reunion with the departed do not amount to worship. We have therefore now to trace the process by which they developed into ancestor-worship.

The first condition of any such development is that the demonstrations, at first spontaneous, should become conventional and harden into custom. This is not the same thing as saying that grief ceases to be genuine when the manner of its expression becomes conventional. On the contrary, in the first place, beneath "the outward trappings and the signs of woe" there may be "that which passes show "; and, in the next place, the existence of a conventional mode of expressing the mourner's woe shows that public opinion considers grief in these circumstances right and proper; such demonstrations, in fact, are not the isolated expression of unusual susceptibility, but an indication of the habitual affection even of a savage for those nearest and dearest to him. When, then, it has become the tribal custom for relatives to perform certain acts, on the occasion of a death, which were originally spontaneous and now are the conventional expressions of grief, it becomes possible for fear to operate in support of this as of other tribal customs, though it was not in fear that either it or they originated. Custom is one of the earliest shapes in which duty presents itself to the consciousness of the savage: it is what is expected of him, both by the community and, in his better moments, by himself as a good member of the community. Now, the savage regards all sickness as the work of spirits - not necessarily of evil spirits, as is commonly and carelessly said. When, therefore, he falls ill, he casts about in his mind for the spirit who may be the cause of his sickness; and if, like the African chief mentioned by Lippert, he has been negligent of the rites which it is customary to perform to a deceased parent, he naturally interprets his headache as a reminder from the neglected ghost. In a word, fear of punishment is an indispensable instrument in the education of man, be he savage or be he civilised; but fear of punishment is not the same thing as fear of evil spirits. The latter is irrational, and is sterile both morally and intellectually, while the former implies a standard of duty (or custom), and opens out the possibility of moral and intellectual progress.

That the ceremonies out of which ancestor-worship was 1 Lippert, Kulturgeschichte, iii. 75.