- 11

THIS PAPER 1 describes some differences noted in a comparison of cognate folktales from the islands of Kapingamarangi and Nukuoro, discusses possible relations of the folktale differences to differences in social structure so far as known, and presents for further investigation some hypotheses suggested by these data.

Nukuoro and Kapingamarangi are both small atolls located in Ponape District of the U.S. administered Trust Territory of the Pacific Islands. In land area Nukuoro is a little larger (.64 square miles as against .52 square miles) although in 1949 it was reported to have had about half the population of Kapingamarangi (273 to 540). 2 Both islands are south of the Carolinian typhoon belt. The principal environmental difference between them is that Nukuoro has more rainfall, Kapingamarangi being subject to occasional severe droughts.

Culturally both islands may be regarded as Polynesian outliers in Micronesia. Nukuoro is the nearer of the two to the main line of the Caroline Islands and seems to have had more Micronesian cultural influence and racial admixture. The closest cultural similarities of each island appear to be with each other. The dialects are mutually intelligible. On discovery neither people possessed ocean-going canoes, although they were aware of the existence of other islands from traditions of the travels of their ancestors and from occasional castaways.

The Kapingamarangi tales discussed below were collected in 1947 by Drs. Samuel H. Elbert and Kenneth P. Emory during their study of the island under the Co-ordinated Investigation of Micronesian Anthropology. Summaries of these tales have previously been published, along with a description of the circumstances of collection. 3 The native texts and full translations of the tales collected by Elbert appear in his mimeographed Linguistic Study of Kapingamarangi. 4

The Nukuoro tales compared with these were written in English for me in 1951-52 by a Nukuoroan intellectual named Aiter or, as he prefers to spell his name, Ideal. The Kapingamarangi collection used here is superior to this Nukuoro collection on several grounds: the Kapingamarangi tales are from a number of informants; a majority of the tales were recorded verbatim in the native language, which means that the texts tend to be fuller and lack ambiguities due to use of a foreign language; also Drs. Elbert and Emory had the opportunity to - 12 ask for explanations at the time of the recording, while the Nukuoro texts were simply written in English by the informant on Nukuoro and later sent to me. Nevertheless, Ideal's texts generally give the important points of the tales clearly and I feel their quality is adequate for the sort of cultural comparison between Nukuoro and Kapingamarangi which is the purpose of this article.

The comparison made here is of tales which appear to be told primarily for amusement and not taken too seriously by the people. This is explicitly stated for the Kapingamarangi tales and is probably safe to assume for the Nukuoro tales. Origin myths and realistic traditional history have been excluded from consideration as possibly having a different psychological meaning. By coincidence the number of tales available to me from both islands was twenty.

In this sample I find eight pairs of tales with sizeable cognate sections. Contrary to what one would think, this does not mean that there are eight Nukuoro tales with Kapingamarangi cognates and vice versa. The number of mutually cognate tales is actually seven for Nukuoro and four for Kapingamarangi. This is because, as shown in the diagram below, more than one tale in a culture may be cognate with different parts or even the same part of a single tale in the other culture. Thus the beginning of the Kapingamarangi story of Hina is cognate with two Nukuoro stories (each with a similar beginning but distinctive conclusion), while the ending of Kapingamarangi Hina is cognate with the Nukuoro story of King Kaneki, although the beginnings of Hina and King Kaneki are distinctive.

Illustration

NOTE: A line connecting two of the tales indicates that they possess a major segment or episode in common. The letters A and B following a tale's name indicate the initial and final segments of a tale, not two separate versions.

- 13

It may be objected that where two tales from Culture A are cognate with a single tale from Culture B there is something questionable about using the B tale twice in comparison, for if the B tale should happen to be non-representative in some way the differences between the cultures may become unduly magnified. Against this view I would make two points:

• (1) assuming the null hypothesis that all changes in the tale are random, then it is quite possible that the two A tales have varied in opposite directions from both the original shared tale and the current B tale, and that consequently when each is compared with the single B tale the two sets of differences will “bracket” the B tale and cancel out;
• (2) if what was originally a single tale in both cultures has in Culture A been “fractured” into two tales and remained intact in Culture B this may be because in Culture A the original tale has been more subject to pressure for change than in Culture B; if the B version has in fact remained more stable this would be a reason for assuming that it is also reasonably representative of Culture B.

Because of the recombination of episodes we may distinguish at least two integrational levels of differences between the cognate tales. On a microscopic level one may distinguish changes made within cognate episodes: e.g., in the Nukuoro tale of Girl and Clothes the turtle carrying the fleeing girl offers her a coconut to eat and warns her to crack it only on the end of his shell, which she does; in the Kapingamarangi tale this turtle tells the girl to drink the liquid herself but give him the nut to split open; the girl disobediently cracks it on his head, so the turtle appropriately douses her in the sea. Changes such as these may be regarded as intra-episodic.

But there are also macroscopic differences involving the context of a cognate segment or episode in a pair of complete tales from the two cultures. That is, for any given cognate segment one may ask the question: what non-cognate segment(s) have the people in the two cultures, over time, come to attach to the cognate segment (either following it or preceding it).

Perhaps this large scale shuffling of sizeable segments between tales is more culturally diagnostic than the small scale intra-episode changes, since the large scale changes presumably require greater psychological pressures to produce. Folk audiences are often ready to allow each narrator to make minor embellishments and deletions in a familiar tale but are likely to reject or criticize major tampering with the plot unless the result is definitely of superior appeal to the modal audience personality. 5

However, because of the small size of the sample I have not handled large and small scale differences separately and have treated them alike for most of the statistical comparisons.

- 14

In addition to the categorization of differences as large and small scale—episode-contextual versus intra-episodic—the small scale differences may also be classified as they involve differential additions or losses of elements on the one hand and alterations of shared elements on the other. (The episode-contextual differences by definition involve additions or losses.) In a comparison of tales from two cultures it is of course usually impossible to tell whether the presence of item X in the Culture A version of a tale and its absence in the B version is due to the addition of X to the A version after the two cultures diverged or to the original presence of X in the ancestral version with its retention in A and loss in B. Accordingly any difference in which a given item is present in one culture and absent in another is regarded as an addition/loss difference. An alteration difference on the other hand involves the presence of comparable but differing items in both A and B versions. For example, if the heroine in the A version gave birth to two children while the B heroine did not, this would be an addition/loss difference. But if the A heroine bore a son while the B heroine bore a daughter this would be an alteration difference.

In counting up differences between the Nukuoro and Kapingamarangi versions of cognate tales I have generally treated addition/loss differences involving a continuous unshared text segment as a single difference, regardless of the length or complexity of the text segment. For a long segment a short summary is made of the general outcome and this is used for comparison. A slight exception to this rule is made for some medial unshared segments which both follow and precede cognate or functionally comparable segments. 6 In such cases the unshared segment may be divided in two by the convention that the first half is part of the “context” of the preceding segment and the second half part of the “context” of the following segment.

These rules may seem to reduce the amount of data unnecessarily but if we omitted them and counted each distinguishable element or motif in an unshared continuous text segment as a separate addition/loss difference we would really be partly abandoning our goal of comparison of cognate tales; the criticism could be justly raised in some cases that we were really comparing unrelated tales which happened to have acquired minor cognate episodes. Moreover, some addition/loss differences between cultures may be simply due to incomplete texts. By counting each unshared continuous segment as one or at the most two differences regardless of length, somewhat less weight is given to personal omissions of informants in the cultures for which the segment is missing.

I would remark that the method of tale comparison used here is in practice much the same as the one used in a previous article. 7 In this - 15 paper I had not, however, explicitly recognized the recombination of episodes noted above as a distinct type of difference, although I had done so implicitly. A further change in the present paper is that I have also classified the differences in some additional ways.

The specific differences and major similarities of plot and characterization found in the comparison of cognate tales are shown in Appendix A. Differences of names of characters and places and other linguistic and stylistic differences are ignored. I would note in passing, however, that the Kapingamarangi tales seem to include more widespread Polynesian names than those from Nukuoro. This suggests that of the two groups the Nukuoro tales may be more modified from the original. Some general trends in the differences between the tales are summarized below with respect to the relative status of the sexes, generations, and spouses versus siblings.

One respect in which the tales of the two groups show a contrast is that of the relative competence of the sexes, a comparison which I have also made for Truk and Ponape. 8 In making a count of this I have interpreted a difference as signifying greater competence for the character of a given sex in one of the cultures if that character does something more efficiently or quickly, succeeds where his cognate character in the other culture fails, tries something difficult which the cognate character does not, does something by himself where the cognate character has a helper, where a successful character is of opposite sex in the two cultures, etc. Results are as follows:

Chi2=9.31. .01 > P > .001.

NOTE: For judgments of specific differences compare Appendices A and B.

This tabulation shows that where changes have occurred in the cognate tales they have had the overall trend of making the women in Nukuoro tales and the men in Kapingamarangi tales appear more competent, the intercultural difference being especially strong for women.

My previous study using tales from Truk and Ponape, where more cultural and psychological data were available to me, suggests that these differences are probably an index of the relative status and modal psychological security of the sexes. While data published to date on the social structures and psychological characteristics of Nukuoro and Kapingamarangi cannot be regarded as entirely conclusive, it is worth - 16 noting that on Kapingamarangi political succession to the secular chieftainship was patrilineal, whereas on Nukuoro it depended on membership in an unusual type of descent group wherein the children of a given couple were assigned alternately, regardless of sex, one to the father's group and one to the mother's. Moreover, on Nukuoro women were able to hold the office of secular chief themselves whereas on Kapingamarangi they were not. 9 I would suspect that historically the Nukuoro type of descent group (which lacks obligatory exogamy, incidentally) originated from normally patrilineal descent groups in which there was an option of exercising a right of affiliation with the mother's patrilineage. Presumably as the position of women grew stronger, more and more children took up this option until the conflict between the parents over the affiliation of their children was settled by the reported compromise. Once the community was well established and the way of life adapted to the local circumstances, the position of secular chief probably required little effort and the time which women owed to household and child care duties was not a serious hindrance to their assumption of the chieftainship on Nukuoro.

Regarding the relative psychological security of the sexes on the two islands, it is an interesting and perhaps significant fact that during four years of residence in the Eastern Carolines I heard of no cases of male sexual impotence except on the island of Nukuoro. On Nukuoro there was, however, a woman practitioner of a remedy for male impotence. (The chief effective elements of the remedy appeared to be suggestion and massage of the male organ.) I did not investigate specifically whether the same medicine was known on Kapingamarangi but would suspect it is not.

The question of relative competence of tale characters in the two cultures may also be investigated with respect to age differences. The questions asked here include: where there are age differences between parallel characters who are equally successful (or unsuccessful) in the two versions, do the successful characters tend to be older in one culture and younger in the other? Again, where there are parallel conflicts between older and younger characters in the two cultures, both pairs of roughly comparable age, do the elder tend to be more successful (or less unsuccessful) in one culture than in the other? Even where there is no parallel action in tales (that is, in cases of addition/loss differences resulting in non-cognate segments) do the oldest characters in the tales in one culture and the youngest in the other tend to be successful?

Of course, there can be no census of folktale characters giving the exact chronological age of each. There are several criteria in the tales, however, which may be used to judge age differences. The most obvious is that of position in the life cycle and marital status. In this respect I assumed the following order of statuses in direction of increasing age: unmarried and living with parents, married in the course of the tale, married throughout the tale, married throughout tale but children born in tale, married with children throughout tale, widowed in tale, widowed throughout tale (children mentioned but not spouse).

- 17

I also assumed that characters of high social status were older than ordinary people, on the grounds that these positions were usually acquired at least at middle age and were held until death. Supernatural human-like spirits were assumed to be older than ordinary humans, on the grounds that culturally they are assimilable to ghosts of the aged and in psychoanalytic terms probably represent a negative split parental image. Personalized animals were also regarded as older than ordinary humans on analogous grounds. This assumption may seem questionable as a general principle but the instances in these tales may also be supported by a reasoning from context: the animals appear immediately after a conflict of a young girl with her parents and interact with the girl but not directly with the parents proper, for whom they are presumably a disguised representation.

In future studies it would be desirable to ask informants their opinion on the probable age or position in life cycle of specific animal and spirit characters in tales. A similar inquiry might be made about the sex of such characters where not apparent in the text. It is conceivable that even if there is no specific tradition on these points a decided consensus of “guesses” might appear among a number of informants. The definite absence of such a consensus would also be a fact not without interest.

Results of the generation comparison are as follows:

Chi2=5.01. .05 > P > .02.

NOTE: For judgments of specific differences compare Appendices A and C.

These figures indicate that the differences between cognate tales in the two cultures are significantly in the direction of Nukuoro attributing more competence to the younger generation and Kapingamarangi attributing more competence to the older generation. These differences suggest that Nukuoro is characterized by more emphasis on individualistic precocity and achievement, and that parental and kin authority generally is weaker. This conclusion is reached because, to put it perhaps over-simply, we would expect that a society emphasizing individualistic achievement would begin by inculcating in the children a belief in their own competence while a society emphasizing group co-operation and authority of leaders or parents would inculcate in the children a belief in their own dependence on more competent seniors.

Again, while conclusive data are lacking, I can offer anecdotal support for this view. Ideal, the man who prepared the Nukuoro texts used here, is an example of a man with a strong personal achievement - 18 motive for whom I know of no equivalent among Kapingamarangi men. His preparation of these tale texts for me is typical of his interest in improving himself through association with foreigners. Another instance of Ideal's achievement motive was his sending, entirely on his own initiative, a model Nukuoro house to President Truman. Another comparable Nukuoro man was Ideal's aged father, who had been a sailor in his youth and was fond of recounting his visit, many years before, to “the Market Street Theatre, the best theatre in San Francisco.” Similar anecdotes could be told about other Nukuoro men, while the Kapingamarangi men with whom I am acquainted uniformly give at least a first impression of being comparatively modest and retiring.

Another difference into which I inquired was the moral quality of characters by sex in tales in the two cultures. A character was considered to have committed an immoral act if he attempted or indulged in theft, intentional cannibalism or unprovoked or insufficiently provoked murder and the like. Commensurate revenge, however, was considered as morally neutral. A character was considered to have committed a positive moral act if he forgave an injury, rescued another from danger, or gave things to another or otherwise helped another in need. Two separate comparisons regarding moral quality of the sexes are possible, for good acts and for bad ones. These two comparisons were consistent with each other (e.g., men in Kapingamarangi had more good acts and fewer bad acts than Nukuoro men) but the size of the sample of paired tale differences when treated in this fashion was too small for statistical significance. To increase the size of the sample of moral differences I combined the two kinds of differences by a sort of conversion of moral sign: treating both cases of immoral acts by a category of characters in one culture and moral acts by the same category in the other culture as a positive moral difference in favour of the second culture. Results were as follows:

Chi2=7.77. .01 > P >.001.

NOTE: For judgments of specific differences compare Appendices A and D.

As can be seen, there is a significant difference in the direction of Nukuoro men and Kapingamarangi women as being depicted as morally inferior to their counterparts in the other culture. This moral deficiency parallels the deficiency in general competence, which is another way of saying that the tales are on the whole comic rather than tragic: the good win and the bad lose. Our interest here of course is not in the fact that in both cultures the tales are comic but in the differences between - 19 the cultures in attribution of moral character to the sexes. The moral differences reinforce the competence differences and their implication of a relatively higher status for women in Nukuoro and men in Kapingamarangi.

However, a similar comparison of moral quality by age for the two islands does not reinforce the distribution of competence by age. Rather, as shown below, there is a significant trend for the less competent Nukuoro elders to act more morally than the children, and the reverse for Kapingamarangi elders and children.

Chi2=4.53. .05 > P > .02.

NOTE: For judgments of specific differences compare Appendices A and E.

These results were contrary to my original expectations but since noting them they seem to be consonant with the hypothesis that Nukuoro is a more individualistic culture than Kapingamarangi. One would expect that in a kin-group oriented culture the parents would put increasing pressure on older children and adolescents to work for the kin-group. One would, moreover, expect this pressure to be in fair measure successful. Accordingly younger characters in tales would tend to be pictured as virtuous, while the elders would tend to be pictured as exploitative. On the other hand, in the individualistic society, the parents are presumably trying with some effect to encourage the child to “look out for himself,” even if at the expense of others, and, occasionally, themselves. Accordingly younger characters in tales would tend to be pictured as rather aggressive while the elders would be pictured as indulgent and assisting the advance of their children.

A final comparison was made of differences in the strength of spouse versus sibling relationships in the overall plot of the tales. A spouse relationship was said to be stressed more in one culture if in the version from it a couple became married, helped each other in a crisis, or survival a marital dispute, whereas an analogous event was missing or less clearly developed in the cognate tale from the other culture. Analogous considerations held for scoring stress on sibling relationships, which included those between both same sex and cross-sex siblings. Individual pairs of cognate tales differing in these respects might thus be scored once or twice but not more: for a greater stress on one of the relationships in one of the cultures, or this plus a greater stress on the other relationship in one of the cultures. Results were fairly significant in view of the small size of the sample:

- 20

.05 > P > .025, using tables for the Fisher exact test.

NOTE: For specific judgments on tale differences for this table compare Appendices A and F.

These results are logically consistent with the interpretation of the differences in age competence made above. One would expect that in the society emphasizing co-operative kin groups (here Kapingama-rangi) the individual competence of young people would be played down and the mutual dependence of siblings encouraged as an alternative means of acquiring psychological security and organizing work. Where marital ties threatened to conflict with sibling ties one would expect the marital ties, in fantasy at least, to give way. In the more individualistic society (Nukuoro) one would expect to find more competition between siblings and a greater dependence on the marital relationship. Parents might well more or less consciously encourage a certain amount of sibling rivalry in the hope of spurring their children on to greater achievement. Spouse relationships in such a society might also be subject to much internal competition and strain and would not necessarily be more stable in an absolute sense than in the group-oriented society—that is, in the individualistic society perhaps social ties are generally weaker. However, within the individualistic society a person should tend to find the marital relationship more reliable than the sibling relationship.

Differences in cognate tales in the competence and moral quality of the generations and in stress on sibling versus spouse relationships are all consistent with the notion that Nukuoro is a more individualistic society than Kapingamarangi. I would suggest that in view of probable historical developments in the social structures of the two islands the previously noted differences for the sexes in competence and moral quality are also consistent with this interpretation. If we assume that both societies were originally organized in limited bilateral descent groups with a patrilineal bias then we may deduce that it was the brother-brother relationship which was most strongly stressed in the ancestral culture. An increasing individualism, as is posited for Nukuoro, would have the effect of improving the status of women in conflicts with men by breaking up the male opposition and making it harder for a man to get support from his brother (or father) in arguments with his wife, sister, or mother. Thus, starting from the particular historical situation posited at the time of settlement, which seems in line with other cultures in Western Polynesia, one would expect a growing individualism to result in a relative rise in women's position. (Of course, starting with a matrilineal society with strong sister ties, increasing individualism should by the same reasoning produce a lowering of women's position.)

- 21

What might explain the inferred increase in individualism on Nukuoro? A diffusionist explanation might attribute this to the disruptive influence of matrilineal castaways from the Caroline and Marshall Islands on an originally semi-patrilineal system. It is almost surely true that Nukuoro has received more Micronesian admixture as one would expect from its geographical position but I find this explanation insufficient. The culture and language of Nukuoro are still recognizably Polynesian, and castaways must have come in small enough numbers so that they had to adjust to the existing culture rather than being able themselves to make major changes in the social customs.

One might argue also that any local social changes in Nukuoro have not had enough time to become reflected in folktales, and that the folktale differences are the result of direct diffusion of tales from matrilineal parts of Micronesia along with the castaways. The castaways certainly told folktales, some cognate with previous Nukuoro tales and some not, but I personally doubt that the castaway versions of tales would have displaced the Nukuoro versions without adaptations to Nukuoro culture. To be sure, one might cite instances of primitive peoples who have taken over European folktales with little modification. However, I suspect that where modification of imported tales is minor there has usually also been much acculturation or at least there has been established a firm conviction of the superiority of the Europeans. The Micronesian castaways of Nukuoro might have occasionally been regarded as approximate equals but probably never as vastly superior.

I feel that an environmental and demographic explanation of Nukuoroan individualism as reflected in folktales is more powerful than a diffusionist one. Because of the more stable and favourable environment of Nukuoro as compared to Kapingamarangi, I suggest that this island at some time after settlement had to face a prolonged over-population. Such a situation would, I believe, be eventually self-correcting in that it would give rise to increased individualism and a lowered birthrate accompanying this. Prolonged overpopulation should produce individualism since the kin groups would have inadequate resources with which to reward their members for group loyalty. A temporary solution without an increase in individualism might be competition between kin groups for the inadequate resources with banishment of the losers. Traditions on many Pacific islands indicate that this has probably occurred frequently enough, although it should be remembered that Nukuoro evidently lacked sea-going canoes when discovered. However, this solution would be only temporary since it would leave the remainder of the population more unified than before and a new cycle of overpopulation would soon result. Within a culturally homogeneous, face-to-face community such as Nukuoro periodic expulsion or extermination of part of the population is not, I submit, as likely an adjustment to overpopulation as a simultaneous increased individualism and lowered birthrate.

The postulated association between individualism and a low birth-rate deserves some comment. I assume that parents in an individualistic - 22 competitive society find children more of a burden than parents in a kin-oriented society, where relatives share child care more. Also in a kin-oriented society, as Lorimer has pointed out, 10 a larger kin group size may help the group in its competition with other groups, so that having many children is regarded not as a burden but as an asset. The exact mechanisms by which parents limit birthrates in individualistic societies may vary considerably but presumably if the desire is there it can be and is put into effect, at least to some degree.

It seems unlikely that Kapingamarangi is a very extreme example of kin-group oriented society. Probably motivation to increase the kin-group size through many births is weaker there than in many societies with strong lineages on larger land masses. But my argument here does not require that Kapingamarangi parents should actively desire many children—it is sufficient for my point to suggest why we might expect Kapingarangi parents to have less motivation to limit births than Nukuoro parents. It appears to me that this situation has obtained because the Kapingamarangi environment has controlled the population through occasional severe droughts and consequent famine without any human decision being required. It would appear that for some time after each famine there was an abundance of resources per capita; that the underpopulated inter-famine periods were typical of most of the island's history and that a fairly high birthrate would actually be adaptive, by building up the population enough between famines so that a single severe famine would not wipe it out. This adaptive value of a high birthrate for the society would not in itself presumably serve as much of a motive for individual parents to desire many children. However, under these circumstances the ordinary economic advantages of fairly strong kin co-operation and large family size could give people enough reward most of the time to make individualism and a small number of children seem less attractive than on Nukuoro during the postulated drawn-out period of overpopulation there. The severe famines on Kapingamarangi were presumably short enough so that they did not have much lasting direct effect of this sort on the social structure.

In conclusion I would caution the reader that the historical reconstruction and some of the inferences about the social structure of the two islands must be regarded as imperfectly substantiated at this point. It is hoped that further information from traditions, contemporary ethnographic and psychological study and perhaps also archaeology will some day be available to confirm or reject the inferred explanations with more certainty than is now possible.

In any case the tale comparison has established that many of the differences between pairs of cognate tales from the two cultures are not random but are instances of certain generalized divergent trends which are statistically significant. If the explanations given for these differences are in some respects incomplete or incorrect it at least seems likely to me that the eventual “final” explanation will be related to - 23 some kind of differences in culture and social structure, further dependent at least in part on differences in local environment.

The application of statistical techniques to the comparison of socio-cultural variables in folktales and the systematic comparison of cognate folktales and other forms of oral literature for these variables are methods of investigation which so far lack personnel more than data. The variables selected for comparison here do not, I am sure, exhaust the patterned differences between Nukuoroan and Kapingamarangian tales, and others with different interests can no doubt try other comparisons.

Anthropologists have long recognized that the study of the diffusion of folktales and tale motifs can contribute to general studies of cultural origin and contact. The findings of this paper suggest that through studies of oral literature we can not only establish the bare facts of contact and genetic relationship but can also gain additional evidence as to trends in the internal development of cultures.

• EILERS, Anneliese, 1934. Inseln um Ponape. In G. Thilenius, ed., Ergebnissa der Sudsee—Expedition 1908-10, Series IIB, v. 8. Hamburg, Friedrichsen, De Gruyter.
• ELBERT, Samuel H., 1948. Linguistic Study of Kapingamarangi. Washington, Pacific Science Board. Mimeographed.
• —— 1949. “Utamatua and Other Legends from Kapingamarangi.” Journal of American Folklore, 62:240-246.
• EMORY, Kenneth P., 1949. “Myths and Tales from Kapingamarangi.” Journal of American Folklore, 62:230-239.

• FISCHER, J. L., 1956. “The Position of Men and Women in Truk and Ponape.” Journal of American Folklore, 69:55-62.

• LORIMER, Frank, et. al., 1954. Culture and Human Fertility. Paris, Unesco.

• U.S. NAVY DEPARTMENT, 1949. Trust Territory of the Pacific Islands, Annual Report to U.N. Washington, D.C.

- 24

- 25

- 26

- 27

- 28

- 29

- 30

- 31

- 32

Arabic numerals refer to specific differences within the tales.

Notes on interpretation of certain differences:

• I 1—On the assumption that the girl's brother in Kapingamarangi is there for helping her, as the story indicates, and is not required in Nukuoro.
• II 6—Nukuoro wife does not try to kill husband, submits to reconciliation. Kapingamarangi wife succeeds in killing husband.
• II 10—Nukuoro girl gets help of animal on own initiative; conflict of males over Kapingamarangi girl, ultimately beneficial to her, but she is passive.
• III 3—On the assumption that the girl in Nukuoro is there to help her brother, as the story indicates, such is not required in Kapingamarangi.
• VI 2—Nukuoro husband is superior to Kapingamarangi husband; Nukuoro girl superior to Kapingamarangi girl for acquiring husband.
• VI 3—On the possibly questionable assumption that girl desires children as potential helpers.
• VIII 6-8—Male Kapingamarangi demon is more competent than analogous female Nukuoro demon.

- 33

Explanations of certain less obvious interpretations:

• I 1—Assuming that the Kapingamarangi brother is needed to help the girl in her troubles with parents and parental surrogates.
• I 2,5,6; II 2,5,6,7,8,10—Assuming these supernaturals and humanized animals represent disguised elders in these and most contexts.
• III 2—Assuming that demon=dead person=older person disguised as young wife.
• V 3—Assuming forced marriage to chief is better than imprisonment and that Nukuoro girl achieves marriage by superior sexual attractiveness.
• V 4—Assuming husband, being married, is older than brother.
• VI 3—On grounds that Nukuoro girl is superfluous to plot and one Nukuoro boy=two Kapingamarangi boys in later achievements in tale.
• VI 4—Assuming demon has superior magical powers to ordinary humans.
• VI 5—Assuming demon unwittingly betrays own interest by telling husband of boys.
• VII 3—Assuming servant, being of low status, is younger than wife.

- 34

• V 1—Assuming the Nukuoro girl's disobedience was more serious than the Kapingamarangi girl's since the Nukuoro girl's is a sin of commission, the Kapingamarangi girl's of omission.
• V 7—Assuming mother's help to son is commendable, even though son's aim is not; perhaps questionable.
• VI 1—Assuming Kapingamarangi girl's separation primarily due to malicious female (demon), while Nukuoro girl's flight due equally to malicious male as well as female (both parents).
• VI 2—Assuming Kapingamarangi husband must accept a little responsibility for failing to discriminate wife from demon; perhaps questionable.

- 35

Notes on certain less obvious interpretations:

• I 2,4,6—Assuming humanized animals and demon in this context represent elders.
• II 2,4,6,8,10—Previous assumption.
• IV 2—Assuming that men represent elders, perhaps unjustifiably; also that Kapingamarangi husband is partly at fault for wife's fate.
• V 3—Assuming that Kapingamarangi girl wrongly violated taboo, but Kapingamarangi chief punished with undue severity; also that forced marriage preferable to imprisonment.
• V 4—Assuming Nukuoro husband older than Kapingamarangi (unmarried) brother.
• V 7—Assuming mother's help to son is moral even though son immoral; perhaps unjustified.
• VII 1,3—Assuming Nukuoro and Kapingamarangi man are same age.
• VII 5—Bad Nukuoro father and good mother cancel out.
• VII 6—Assuming Kapingamarangi man morally neutral, Nukuoro girls' mother's and mother's brother's help to girls outweighs Nukuoro father's murderousness.

- 36

Explanation of certain less obvious interpretations:

• IV—A sibling pair (brothers) exists only in the Kapingamarangi version.
• VI—Nukuoro girl is jealous of her sister; her son kills his younger brother.
• VII—Two co-operative sibling pairs in Nukuoro version (two daughters and their mother and mother's brother) versus one questionable one in Kapingamarangi version (two women not specified as sisters).
• VIII—Sibling relation of 10 Nukuoro men unspecified; in any case Kapingamarangi brother rescuing sister indicates stronger ties than difference in Kapingamarangi and Nukuoro loan of adze.