IN the foregoing pages we have endeavoured to present some of the more important and characteristic myths from Polynesia. Forced to give undue emphasis to three or four of the many groups because of the paucity of material from all but these, we may, nevertheless, gain a pretty clear impression of the type of tales once current throughout the entire area.
In the presentation of the material and in its discussion resemblances have been pointed out between the various island-groups, both within and without Polynesia; but this has been done only for individual tales or striking incidents, and no attempt has been made to summarize the results. The fact of wide-spread relationship has probably become evident, but the conclusions which may legitimately be drawn are not perhaps apparent. Unless we are to depend entirely upon impressions, some sort of statistical method must obviously be employed. While these are particularly liable to lead to erroneous conclusions because of the fragmentary and unequal quality of the available material, we must use some such method to extract meaning from the mass of individual similarities. All myths, as we have them, may be analyzed into a series of separate incidents. This group of incidents may, and indeed often does, remain intact for long periods, and may be transmitted as a unit from one people or area to another. Very often, however, in the course of time or in transmission one or other of these drops out or is modified, and new ones are added; so that the result may be a tale quite unlike the original, but in which certain of the original
incidents survive. Individual incidents also may be widely transmitted, and by the study of the distribution of these much may be learned as to historic associations, lines of migration, and cultural relationships. Myths, then, as we find them, are of complex origin, the product of long modification, decay, and accretion. If now we consider the mythology of Polynesia from the standpoint of its constituent elements, i. e. its incidents, much light may be thrown on its growth as a whole, on the interrelationship of the mythology of the different island-groups, and on the kinship which the mythology of this area bears to that of adjoining ones. For a really satisfactory study of this sort relatively complete material from the whole region is needed; but unfortunately, as already pointed out, this is not, and probably never will be, available. Incomplete records from certain island-groups inevitably lead to erroneous conclusions in regard to the distribution of incidents, but with all due allowance for these sources of error, and emphasizing the tentative character of the results obtained, it is perhaps worth while to see what conclusions may be drawn from the data which we possess.
Within Polynesia itself such a study of the distribution of myth-incidents leads to results of interest. Perhaps the most striking of these is the apparently close relationship between Hawaii and New Zealand, the two most widely separated groups within the area, since of the Hawaiian episodes occurring elsewhere in Polynesia two-thirds are found in New Zealand, while in the much closer Cook and Society Groups only about one-third appears. New Zealand's similarities are closest with the Cook Group (as indeed they should be, seeing that the bulk of the historic immigration came from there), but the number of agreements with Hawaii is very nearly as great, and a strong relationship to Samoa is also apparent. Considering other groups, Samoa is most closely affiliated with the Cook Group and New Zealand, and only secondarily with Tonga. Central Polynesia, i. e. the Cook Group, Society,
and Marquesas, seems to form more or less of a unit with affiliations running in all directions.
If the character of the incidents themselves be considered, and not merely the number of agreements, it appears that in the case of Hawaii and New Zealand the episodes which are common to these two groups are, for the most part, other than those which either shares with the geographically intermediate Cook or Society Groups. Similarly, although New Zealand's affiliation with Samoa is nearly as strong as with Hawaii, the incidents which it possesses in common with the former are, generally speaking, quite distinct from those which it shares with the latter. The logical explanation of such a condition, would seem to be that Polynesian mythology is, as a whole, a complex of incidents derived from different sources, one portion of the area having received its material mainly from one source, another from another. Thus the myths of any individual group, such as New Zealand, would be the result of a blending of two or more streams of incidents, or, to vary the figure, would be composed of different strata superimposed in a definite historical order.
In the presentation of the myths, as given in the preceding pages, frequent reference has been made to the occurrence of similar incidents in Melanesia and Indonesia; whence a consideration of the number and proportions of these similarities in different parts of Polynesia may be expected to throw light on this question of sources. The Melanesian area lies immediately adjacent to Polynesia on the west, and we may first consider how far incidents found in Polynesia also occur in Melanesia. Theoretically, any community of episodes discovered between these two areas might be due to transmission in either direction, i. e. from Polynesia to Melanesia, or from Melanesia to Polynesia. Inasmuch, however, as a great mass of evidence derived from other sources points to the drift of peoples from west to east in the Pacific area, we may reasonably regard the bulk of the similarities as due to transmission
from west to east; i. e. that the incidents common to Melanesia and Polynesia are at least in part of Melanesian origin. Assuming for the moment that this is true, it is obvious that we may have two sorts of agreement: incidents of Melanesian origin (or at least of wide Melanesian distribution) which occur only in a single group or in a restricted area in Polynesia; and Melanesian incidents which are current over a considerable portion or the whole of the Polynesian area. Beginning with a consideration of the first of these types, it appears that about one-sixth of the myth-incidents peculiar to the Hawaiian group, and not found elsewhere in Polynesia, occur also in Melanesia. As regards Samoa, however, almost half of the episodes which are purely local and confined to Samoa, so far as Polynesia is concerned, are recorded in Melanesia. In New Zealand the comparable figure rises to nearly three-fourths; but, on the other hand, there are practically no episodes of this type in the Society and Cook Groups. It is clear, then, that from this point of view there is a very strong Melanesian. element in New Zealand and Samoa, while it is weak in Hawaii and apparently absent from the Society and Cook Groups. The individual incidents of Melanesian similarity are, moreover, different in each case, one series being found in New Zealand, another in Samoa, and a third in Hawaii. Moreover, we must note that the Melanesian incidents showing similarity with the Hawaiian are current, so far as our present information goes, only in the Admiralty Islands and New Britain; whereas those occurring in Samoa and New Zealand are more widely distributed and are especially characteristic of eastern Melanesia. The influence, therefore-, exerted on Hawaiian mythology by that of Melanesia would seem to have been not only slight, but localized, as if the wave of Polynesian immigrants which settled in Hawaii had merely touched the northern edge of Melanesian territory. On the other hand, the ancestors of those who reached Samoa and New Zealand must have passed through
much of eastern Melanesia and been subjected to a contact of greater length and intensity.
If we now examine the second type of agreements the results are somewhat different. We are here dealing with myth-incidents which are not confined to single portions of Polynesia, but are common to two or more island-groups. Of this class of episodes Hawaii shows a fifth which are of Melanesian origin, the Society Group slightly less, the Cook Group and Samoa slightly more, and New Zealand nearly one-half. The latter area, again, reveals by far the strongest Melanesian affinities, while Hawaii, Samoa, and the Cook Group have a much smaller proportion, with the Society Group showing the minimum. It is fairly well established that Hawaii received a considerable influx of population from central Polynesia between the twelfth and fourteenth centuries, and the obvious inference is that the Melanesian incidents which Hawaii shares with this group are in large part to be traced to this migration. The great proportion of Melanesian incidents in New Zealand would argue a strong infusion of this darker blood among the Maori.
Nearly all the recognized theories as to the origin of the Polynesian peoples bring them in one way or other from the Indonesian area, and ascribe to them only a temporary stay in Melanesia en route to their homes in historic times. In pointing out similarities of incident during the presentation of Polynesian mythology, the Indonesian affinities have frequently been mentioned, and we must now examine these in the same way in which the Melanesian resemblances have just been considered. Following our former order of treatment, we may first investigate those myth-incidents which, although localized in some one island-group in Polynesia, also have an Indonesian or extra-Polynesian distribution. Of the important incidents of this type in Hawaii, fully half occur also ill Indonesia; but in Samoa and New Zealand, on the other hand, the proportion sinks to about one-eighth. Here again, as
with the Melanesian incidents, the series of episodes in common are different for each group; but the conditions are exactly reversed, for whereas in regard to Melanesian affinities Hawaii shows but few, though New Zealand and Samoa possess a large number, in respect to Indonesian similarities Hawaii is strong, while New Zealand and Samoa are weak. The inference would seem to follow, therefore, that Hawaii has preserved a larger proportion of its original Indonesian inheritance than the other Polynesian groups, while in New Zealand and Samoa this original element has been largely lost or overlaid with borrowed Melanesian incidents. If instead of taking the localized incidents we consider those of general Polynesian distribution which are also found outside its bounds) much the same general results are obtained, although the disproportion between the different island-groups is not so marked. Of this type of incidents Hawaii has nearly a fifth that are also found in Indonesia; the Society and Cook Groups, taken together, about one-tenth; and Samoa and New Zealand even less. The relatively high proportion of Indonesian incidents in central Polynesia is worthy of note in this connexion, as indicating that here the ancestral material was not so largely overlaid by elements of Melanesian origin as was the case in Samoa, which is geographically nearer to Melanesia and which for many generations had had close trade relationships with its eastern margin.
One other line of investigation throws some light upon the course of development of Polynesian mythology. The Indonesian incidents, whose general distribution in Polynesia has just been discussed, have been such as occur in Indonesia and Polynesia, but not in the intervening areas of either Melanesia or Micronesia. If the Polynesian ancestors passed through either of these regions in the course of their movement from west to east, we might expect to find the evidences of such Migration in the presence of Indonesian incidents in Melanesian and Micronesian mythologies. This is precisely what
does occur, and thus one class of incidents is found in Indonesia, Melanesia, and Polynesia, and another in these areas and in Micronesia as well. Of the first, Hawaii shows the smallest proportion, followed by central Polynesia, Samoa, and New Zealand, in the order given. Of the second, Hawaii shows the largest proportion, followed by the other island-groups in the same order as above, only with much greater differences between the extremes, i. e. Hawaii shows five times as many incidents of this most widely distributed type as does New Zealand. It thus becomes once more apparent that Hawaii has had less Melanesian influence brought to bear upon it than the rest of Polynesia, and also that it shows close relationship to Micronesia.
To sum up, then, it may be said that from a study of the distribution of myth-incidents Polynesian mythology, as known to us today, bears evidences of a composite origin. The facts may be reasonably explained by assuming that the ancestors of the Polynesian people were immigrants from the west and that they came into the area in at least two waves: an earlier, one branch of which, barely touching the edge of northern Melanesia, passed northward into Hawaii, while the main body delayed longer in Melanesian territory and extended over the remainder of the area; and a later, which, after traversing Melanesia, spread mainly through western and central Polynesia to New Zealand, and afterward sent an offshoot from the central region northward to Hawaii. The latter group and New Zealand, owing to their comparative isolation, preserved more of their early inheritance, whereas in the remainder of the area this original material was much changed and largely overlaid by the tales introduced by the later immigrant wave. There are, to be sure, various legends which do not exactly fit with this theory, but it at least serves as a working hypothesis and harmonizes remarkably with the data obtained from the study of other aspects of Polynesian culture.
In the foregoing discussion of Polynesian mythology no attempt has been made to explain or to interpret the various myths. Although some of them undoubtedly show features characteristic of sun-myths, moon-myths, and so forth, and although certain scholars have recognized a solar and lunar cycle of tales of supposedly separate origin, it seems wise to go very slowly in any such investigations. It has been so clearly demonstrated that, in the transmission and migration of myths, the original form of the tale may become so greatly modified by the elimination of some incidents and the absorption of others as quite to change its meaning and application, and it has been demonstrated that myths originally told to account for or to explain one phenomenon ultimately come to be applied to a very different one. Consequently we need a much more detailed knowledge of the whole Oceanic area before trustworthy conclusions can be reached.