- 65

In the beginning was the word. And the word became flesh, but the God-like power to make the word flesh is not common. Not all representations are created equal, nor are they equally creative. Power is the ability to make the world conform to one's representations of it, to make one's word become flesh. To be powerful is to be able to impoverish the Poor and enrich the Wealthy, to feminise the Female and masculinise the Male, and to make those who are denominated “savages” deserve the title.

No sooner did Captain Cook and his crew arrive in Hawai'i than Hawaiians and Haole 1 began to denominate each another. Marshall Sahlins (1981, 1985) emphasises the ways Hawaiians denominated Haole. Sahlins argues that there is no such thing as immaculate perception: every perception is mediated by a pre-existing conception. Thus, according to Sahlins, Hawaiians perceived Cook according to their concept of the god Lono, the Hawaiian category that best fit Cook given how and when he arrived. The fit, however, was far from perfect. As Sahlins (1985:138) observes, “the world is under no obligation to conform to the logic by which some people conceive it”. Hawaiians had some success at coaxing Cook to conform to their concept of Lono—for example, getting the captain to perform some of the rituals that Lono is supposed to perform—but their success was less than total. On the one hand, perceiving Cook as Lono was a way in which Hawaiians reproduced their society. The arrival of Cook and his crew was thereby seen not as something new, but as a reproduction of something old. On the other hand, perceiving Cook as Lono put the category of Lono “at risk”. Every actor brings something new to the role, so every performance not only reproduces, but also transforms the play, and Cook's unwitting performance of Lono was no exception. Indeed, Cook's unfamiliarity with - 66 the part made for a particularly free interpretation, and thus for a particularly dramatic transformation. Once Cook played the role of Lono, the role and the play itself would never be the same again.

Sahlins (1985:138) explains that “burdened with the world the cultural meanings are thus altered”. The world's burdens, however, are not evenly distributed. Beginning with Cook's arrival in Hawai'i, Haole and Hawaiians attempted to perceive each other according to their own pre-existing conceptions, but over time Haole were able to make Hawaiians conform to pre-existing Haole conceptions, while Hawaiians did not have as much success at making Haole conform to pre-existing Hawaiian conceptions. Accordingly, Hawaiian conceptions of Haole were more apt to be corrected than were Haole conceptions of Hawaiians. If, as Gananath Obeyesekere (1992a) argues, Hawaiians were more practically rational than were Haole, it is because the world is more burdensome to the colonised than to the colonisers. One effect of the unequal power relations that define the colonial encounter is that Hawaiian conceptions were burdened with the world in ways that Haole conceptions were not. Thus, Obeyesekere argues that when Cook and his crew first sighted Hawai'i, they already expected to find savages there who would, for example, mistake the Captain for a god (Obeyesekere 1992a). At first Hawaiians did not mistake Cook for a god, but very soon, according to Obeyesekere, terrible displays of English fire power taught Hawaiians to express properly apotheosising reverence toward the man who commanded those weapons. This was the first of many Haole scripts that Hawaiians were obliged to perform. Haole preconceptions about Hawaiians were thereby reinforced, while at the same time Hawaiian conceptions of Haole were altered. Rather than adjust their conceptions to fit the world, Haole adjusted the world to fit their conceptions of it. Through their colonial encounter, Hawaiians came to approximate the savages that Haole always already perceived them to be.

Sahlins and Obeyesekere have devoted book-length treatises to debating the Hawaiian apotheosis of Cook. Sahlins argues that Hawaiians spontaneously apotheosised Cook, whereas Obeyesekere argues that the Hawaiians apotheosised Cook as a practical response to Haole military might. Since the first Haole barely set foot on Kaua'i before shooting a Hawaiian dead, it is probably impossible to distinguish spontaneous apotheosis from practical apotheosis. Sahlins cites Haole accounts that Hawaiians immediately apotheosised Cook as evidence that Hawaiians immediately apotheosised Cook, whereas Obeyesekere cites many of the same accounts as evidence that Haole expected Hawaiians to apotheosise Cook, and that they therefore reported the apotheosis even before it could have occurred. The apotheosis of Cook is thus a methodologically problematic issue. I find - 67 that the issue of infanticide is more instructive regarding the early post-Haole history of Hawai'i than is the issue of apotheosis. Infanticide, like the apotheosis of Cook, was a myth that Cook and his crew carried to Hawai'i, but while Hawaiians came to accept the apotheosis of Cook within a few months, several decades passed before Hawaiians accepted the myth of their own practice of infanticide. Therefore, the processes by which a Haole myth about Hawaiians was made true are easier to see in the case of infanticide than in the case of Cook's apotheosis. Whereas in the case of the apotheosis of Cook the same texts can be used to demonstrate the Haole myth and the Hawaiian acceptance of that myth, in the case of infanticide there is ample, independent evidence of the pre-existing Haole myth, its initial falseness and its subsequent “truth”.

Sahlins observes that one “problem with understanding the sources of infanticide is that, in Hawaiian categories—and in those of some missionaries—it was not distinguished from what we now call abortion” (Sahlins 1992:202, n. 12). Thus, some of the 19th-century accounts of “infanticide” in Hawai'i involve foetuses, not infants. Of course, this distinction between foetuses and infants is still contested, but I assume, along with Sahlins, that we, 20th century anthropologists, agree to distinguish between a foetus and an infant, regardless of our various positions on abortion. 2 We agree that infanticide, as opposed to abortion, involves the death of an infant after birth, though there may be differences of opinion regarding viable and unviable foetuses, and regarding the age at which infancy ends. It is important to note that definitions of infanticide are historically and culturally variable, so that 19th century accounts of Hawaiian “infanticide” may or may not pertain to what we, late 20th century anthropologists, would call “infanticide”.

Nancy Scheper-Hughes calls attention to another challenge in the anthropology of infanticide. She problematises the definition of “infanticide” by focusing attention on some of the socio-economic factors that accompany high infant mortality in Northeast Brazil (Scheper-Hughes 1992). Scheper-Hughes writes about women who, living in poverty, make choices about which infants to feed and which to allow to die. Is this infanticide? As Nietzsche (1967:29) argued, “‘sinfulness’ is not a fact, but merely the interpretation of a fact”. So it is with infanticide. Infanticide is not a transparent or natural category, but is a culturally—and historically—bound moral judgement. The fact of a single infant's death might be interpreted to be the result of infanticide, euthanasia or poverty. Or the social fact of a society's high infant mortality might be interpreted to be the result of the - 68 widespread practice of infanticide, an epidemic disease or famine. In either case, to select the “infanticide” label is to privilege a moralising discourse of individual responsibility and blame over a critical discourse of social and political analysis. I suggest that infanticide should be considered not as the practice of murdering infants, but as a discourse by which infant mortality is interpreted to be murder. I am not denying that infanticide exists. To be sure, there have been individual mothers and fathers who, in a rage, murdered their infant children even when no threat of disease or starvation was present, and there have been societies in which infants of the “wrong” race or sex have been put to death. My point is that both individual and social practices of infanticide are relatively rare, and that the denomination “infanticide” usually serves to assign responsibility for an infant's death to a defective parent or society when other factors—namely, poverty, slavery, famine or disease—are present. In such cases, the issue should not be whether the immediate cause of death was famine or suffocation, disease or neglect. The issue should be the context that made mercy killing or benign neglect thinkable options.

Sahlins makes a related observation. He points out that many contemporary Western writers are inappropriately influenced by Western moralities regarding infanticide:

A certain level of detachment may allow anthropologists to speak of infanticide without intending a moral judgement, but the judgement is still there. “Infanticide” is simply not a morally-neutral label. The word implies that some individual is responsible for an infant's death. Anthropologists can abstain from joining the missionaries in condemning infanticide, but we already join them if we focus attention on the practice of infanticide instead of on the social, economic and political context in which high infant mortality and the discourse of infanticide co-exist. To argue that Hawaiians practiced infanticide is to reproduce a missionary discourse and to do so—not coincidentally—at the expense of attending to the context that makes the discourse of infanticide relevant.

- 69

At the time that Haole arrived in Hawai'i, infanticide was regularly attributed to lower-class mothers back in England. David Stannard argues that missionaries were motivated to find infanticide in Hawai‘i because “white children were being slaughtered by the tens of thousands in the very centers of the missionaries’ own proud Christian civilization” (Stannard 1991:410-11). Thus, missionary rumours of Hawaiian infanticide should be read sceptically. Infants in England were dying, and those deaths compelled Haole to seek and find evidence of infanticide in Hawai'i, but scepticism regarding Hawaiian infanticide should be extended to include English infanticide. The finding that poor women in England “slaughtered” their infant children was just as motivated as the finding that savage women in Hawai'i “slaughtered” theirs. Thus, the “political myth” of Hawaiian infanticide did not compensate for the actual slaughter of infants in England so much as it reproduced the equally “political myth” of English infanticide. Unlike England, Hawai'i did not have foundling hospitals to which infants were consigned to die, nor burial clubs that could be exploited to pay parents many times over for the death of an infant. Nevertheless, the fact in England was high infant mortality among the poor, and the interpretation was the sinfulness of poor mothers. Abject poverty, disease and malnourishment were excluded from this English discourse. In their place, medical and legal professionals diagnosed a lack of the maternal instinct, laziness or greed. The death of a lower-class infant was interpreted to be the mother's moral failure, and not to be the failure of the society to which the mother belonged—albeit as a marginal member. Poverty was not a sin, but the poor were inevitably constituted as sinners. The poor were a different species, or at least a different race, and what the poor were at home, the savage and the slave were abroad.

Particularly in 18th century English discourse, far-off savages were represented suspiciously like the squalid masses in England. For example, Jonathan Swift (1955) referred to the Irish poor as “our Savages,” and Thomas Malthus (1798) placed accounts of “savages” and those “in the rudest state of mankind” alongside those of the English “poor” and “the lower orders of society”. The same English myths of infanticide that were told about the Poor in England and Ireland, were told about Natives in the Pacific and North America, and about African slaves in North America and the Caribbean. John Locke (1894), for example, used accounts of infanticide—gleaned from travel journals—as evidence that morality is not innate. In his Notes on the State of Virginia, Thomas Jefferson wrote of Native American women ingesting a vegetable to abort foetuses because “childbearing becomes extremely inconvenient to them” (Jefferson 1955:60). Janet Schaw, writing in the West Indies, complained that “[Negro] wenches - 70 become licentious and insolent past all bearing, and as even a mulattoe child interrupts their pleasures and is troublesome, they have certain herbs and medicines, that free them from such an incumbrance” (Schaw 1939:112-13). Matthew Lewis was more baffled by the declining number of slaves on his plantation in Jamaica: “[H]ow they manage it so ill I know not, but somehow or other certainly the children do not come.… Say what one will to the negroes, and treat them as well as one can, obstinate devils, they will die!” (Lewis 1969:381-82). 3 Victims of capitalism—whether Native, Negro, or Poor—were constituted as lazy, licentious, obstinant or lacking in maternal instinct, and were thereby held responsible for their low fertility and high infant mortality. The economic institutions and material conditions that might have better explained their demise were thereby mystified.

Being heirs to the English poor, savages—including Hawaiians—were presumed to practice infanticide even before they were “discovered” by Haole. Thus, Sahlins' claim that “Cook was a tradition for the Hawaiians before he was a fact” (Sahlins 1985:148) could be rephrased, “Hawaiian infanticide was a tradition for the Haole before it was a fact.” When Cook and his crew arrived in Hawai 'i, they looked for evidence of infanticide and were surprised not to find it. Over the ensuing decades, however, Haole visitors to Hawai'i continued to look for confirmation of their myth of Hawaiian infanticide, and eventually they did find it. The Haole tradition of Hawaiian infanticide became a fact.

Some of Cook's crew believed they saw signs of infanticide in Tahiti. William Monkhouse recorded an enigmatic tale about the corpse of an infant that the father “readily bartered” (Monkhouse [1778] in Beaglehole 1967, I:584-85). Beaglehole believes this was the basis for the crew's belief that the Tahitians practiced infanticide, but it is not clear that Monkhouse thought the infant was an infanticide victim, nor is there an explanation for why Monkhouse bartered for the body. Still, Cook's crew formed the opinion that infanticide was practised in Tahiti, and they looked for signs of it in Hawai'i as well. Thus, David Samwell, a surgeon on the Discovery, wrote of the Hawaiians on February 4,1779–10 days before his captain's death—that “their general behavior is confined much more within the bounds of Decency than that of the Otaheiteans, whose horrid Custom of destroying their newborn infants these women are also totally unacquainted with” (Samwell [1779] in Beaglehole 1967, III:1182). Given Samwell's inability to speak Hawaiian, it is difficult to imagine how he could have determined that Hawaiians did or did not practice infanticide. As Obeyesekere (1992b) observes, Haole reports of conversations with Pacific Islanders cannot be - 71 taken at face value. Sometimes there was a translator present, but in the case of Cook's visits to Hawai'i, translation was always tenuous. Some of Cook's crew had learnt a little Tahitian, which they used to try to communicate with Hawaiians. There were also Tahitians on board who may have had more success communicating with Hawaiians, but who had their own difficulties communicating with the Englishmen. More often than not, the earliest conversations between Haole and Hawaiians depended on gestures and mimicry of dubious effectiveness. It is difficult to imagine what gestures Haole might have performed to ask Hawaiians whether or not they practised infanticide. Still, Samwell expected to find evidence of infanticide but did not. Given the intrinsic ambiguity of gestural communication, we might assume that it is biased towards confirming rather than contradicting preconceptions, which makes Samwell's non-finding of infanticide all the more compelling.

John Turnbull, a sailor who visited Hawai'i in 1802, also noted “the absence from Owhyhee of the horrid practice of infant murder” (1810:229). Turnbull's statement indicates that he also looked for but did not find evidence of infanticide. Turnbull's investigations may have taken the same dubious form as Samwell's. Alternatively, Turnbull may very well have formed his opinion on the absence of infanticide in Hawai'i from conversations he had with John Young or Isaac Davis. They were Englishmen who had lived in Hawai'i since 1790 and who were therefore fluent in Hawaiian and well-versed in Hawaiian customs by the time of Turnbull's visit. If this is the case, Turnbull's statement is a strong indication that infanticide was not a traditional Hawaiian practice. In any case, Samwell's and Turnbull's accounts confirm that Englishmen—if not Hawaiians—were interested in the topic of infanticide and that they failed to find evidence of it in Hawai'i despite their expectations.

The earliest reports of Hawaiian infanticide arrived in Hawai'i in 1820, along with the first wave of missionaries. Writing in their journals, Daniel Chamberlain and Elisha Loomis were the first to attribute infanticide to Hawaiians (Daws 1968). In 1779, Samwell had been impressed that there were “such a great number of fine lively children” (Samwell [1779] in Beaglehole 1967, III:1182). In 1820, Chamberlain encountered a very different situation:

- 72

Chamberlain had only recently arrived in Hawai'i. His reference to a Haole who was a long-time resident of the Islands lends credence to his account, though this must be balanced with the indirect nature of the communication. Chamberlain reports hearing secondhand that mothers strangled their children, but he reports seeing directly only cases in which children “die for want of care” from mothers “not able to support them”. It is unclear that such reports pertain to a practice of infanticide. Note, too, that Chamberlain's account stands in stark contrast to those of Samwell and Turnbull in that the small number of children observed by Chamberlain is a drastic change from the great number of children who, according to Samwell, filled the islands 42 years before. Elisha Loomis also mentioned infanticide in his journal. In the course of discussing an abandoned Hawaiian child, Loomis wrote, “A large proportion dies in infancy—Nay the mothers sometimes kill their infants to ease the trouble of taking care of them” (Loomis 1821: 7 May). Loomis did not report witnessing infanticide, nor did he mention his source, but in his account of an abandoned child he referred to the mother's death and to the father's absence—in the mountains collecting sandalwood for the ali'i (chiefs). Given the economic strains implicit in the father's absence, supposed Hawaiian indolence might rather be indigence, while the mother's death was all too typical: hundreds of thousands of Hawaiians died from diseases introduced by Haole. From Chamberlain and Loomis we get no compelling evidence that infanticide was practised, but we do get a picture of a society very different from that encountered by Cook and Samwell: a society marked by disease and economic hardship, in which children died from “want of care” and “to ease the trouble of taking care of them”. Moreover, Chamberlain's and Loomis's references to the trouble of caring for infants is suspiciously consistent with contemporaneous Haole explanations for infanticide among American Indians and African slaves, such as the encumbrance and inconvenience of motherhood identified by Jefferson and Schaw.

The missionary William Ellis is usually considered a valuable source of information on the Hawaiian society of the 1820s because of his apparent linguistic skills. He came to Hawai'i after spending some years in Tahiti, so he had a much easier time than his U.S. counterparts in learning the related language. Writing around 1824, Ellis reported a specific case of infanticide “about twelve years ago”. Ellis reported that John Young brought an infanticidal father before King Kamehameha. The murder occurred in the heat of a moment, as the result of a dispute between the parents. Kamehameha dismissed Young's charge, saying,

“To whom did the child he has murdered belong?” Mr. Young answered, - 73 that it was his [the murderer's] own son. “Then,” said the king, “neither you nor I have any right to interfere; I cannot say any thing to him” (Ellis 1963:230-31).

This case seems to establish that infanticide was a domestic matter, outside of the king's jurisdiction and that it went unpunished, but it does not reveal that infanticide was a common practice. Rather, the fact that Young would think to bring the case to Kamehameha around 1812—that is, more than 20 years after Young's arrival in Hawai'i—suggests that the murder of a child was a very unusual occurrence. Nevertheless, Ellis uses the abovementioned example to introduce his discussion on the “prevalence of infanticide in Hawaii”. The following passage in which Ellis recounts a conversation he had during his 1823 tour of the Island of Hawai'i has been especially influential in scholarship on Hawaiian infanticide.

Given that the published accounts reported that Hawaiians did not practise infanticide, we have to wonder on what authority Ellis long knew that they did. The identities of Ellis's informants are noteworthy. Kalanimoku, Kapiolani and Governor Kuakini were not only recent converts to Christianity, they were also all ali'i nui (high chiefs). According to these ali'i nui, infanticide was practiced by everyone in Hawai'i except for “the higher class of chiefs”, that is, themselves. Thus, this Native Hawaiian discourse of infanticide corresponds with the contemporaneous English discourse of infanticide. In Hawai'i, as in England, infanticide was a practice the upper class attributed to the lower. Ellis provides corroboration for this thesis with his explanation that the “principle motive” for Hawaiian infanticide was idleness: “like other savage nations, they are averse to any more labour than is absolutely necessary” (Ellis 1963:233). Again, this is a notion imported from the English discourse of poverty directly to that of savagery. The English poor, just like the Hawaiian lower-class savages, were presumed to practise infanticide due to idleness. 4

- 74

Ellis's account of Hawaiian infanticide is further compromised by his subsequent testimony before the House of Commons Committee on Aborigines. On 6 June 1836, Ellis testified that “The practice of infanticide, which destroyed so many in the Southern islands [i.e., Tahiti], did not prevail to any extent in the Sandwich Islands [i.e., Hawai'i]” (Ellis in Coates et al. 1837:50). This statement is consistent with the observation in Ellis's earlier, 1825, work that the Tahitians “probably practised infanticide more than any other natives in the Pacific” (Ellis 1963:232), but in that earlier work Ellis also wrote of the Hawaiians that “from the prevalence of infanticide two-thirds of the children perished” (Ellis 1963:232). In his 1842 Polynesian Researches, Ellis (1969) repeated almost verbatim the passages pertaining to infanticide from his 1825 Narrative of a Tour of Hawaii (Ellis 1963). Thus, it is difficult to reconcile the 1836 statement that infanticide “did not prevail to any extent” in Hawai'i with the 1825 and 1842 statements that from its “prevalence” two-thirds of the Hawaiian children died.

Looking elsewhere for accounts that might corroborate Ellis's claims that infanticide was prevalent in Hawai'i usually leads back to Ellis. For example, Charles Stewart, a missionary who lived in Hawai'i from 1823 to 1825, purported to “have the clearest proof” that “two-thirds of the infants born perish by the hands of their parents” (Stewart 1970:251, emphasis in the original), but he did not state the nature of this proof, and throughout his work he referred to Ellis's Narrative. Similarly, Louis Freycinet, who visited Hawai'i in 1819, also appears to corroborate much of Ellis's higher estimate of Hawaiian infanticide, but he prefaced his discussion of Hawaiian infanticide with the statement “Mr. Ellis, who gives us these details” (Freycinet 1978:67). In sum, there is no corroborative evidence for Ellis's claims that two-thirds of the Hawaiian children were victims of infanticide; Ellis's principal informants were ali'i nui who attributed the practice of infanticide to maka 'ainana (commoners); and Ellis's published journals of 1825 and 1842 are contradicted by his parliamentary testimony of 1836.

The first and only account of Hawaiians' confessing to repeated infanticide is found in the journal of Laura Fish Judd, the wife of a missionary doctor. Though the journal was not published until 1880, it includes the description of a church meeting that falls between two other entries that are dated August 1828 and November 1829:

The scene that followed I can never forget. Why are you childless? we inquired. Very few had lost children by a natural death. One woman replied in tears, holding out her hands: “These must answer the question. I have been mother of eight children, but with these hands I buried them alive, one after another, that I might follow my pleasures, and avoid growing old. Oh, - 75 if I had but one of them back again to comfort me now! If tears and penitence could restore the dead!” She was followed by others, making the same sad confessions of burying alive, of strangling, of something, until sobs and tears filled the house (Judd 1880:34).

What is most striking about this passage is the reported speech of a Hawaiian woman. Earlier references to infanticide were more indirect. Judd's account stands out for its putative inclusion of a Hawaiian woman's voice, directly attesting to what she herself did, not to what other Hawaiians did. As with all cases of reported speech, Judd's account begs for a reading of the relative roles of speaker and scribe. Was the conversation in English or Hawaiian, or was there an interpreter translating between the two? Also, a woman might conceivably be driven to kill her child, but it is not believable that she would choose to do so by burying him or her alive. Was the Hawaiian woman telling Judd what she wanted to hear, or was Judd telling her readers what they wanted to hear? Furthermore, the Hawaiian woman's reported speech conforms to Calvinist confessional conventions. Is this an indication that Judd took the liberty to adapt the speech to her readership's taste, or that the Hawaiian woman was already schooled in confessional language? Buck (1986), following Foucault, emphasises the process by which Hawaiians were taught to speak “within the true” (Foucault 1971:16). She traces the fate of Hawaiians relative to the colonial discourse of Hawai'i. At first, Hawaiians were excluded from the Haole discourse of Hawai'i; as “savages” they were not allowed to represent themselves and were represented only through the mediation of Haole experts. Buck cites accounts like Judd's to show that Hawaiians were admitted directly to the dominant discourse only when they had learned to speak within it by conforming to its conventions.

David Malo, a Hawaiian historian, published an essay in 1839 “On the decrease of population”, which provides a particularly poignant example of a Hawaiian being taught to speak within the English discourse of infanticide. The text is especially revealing because it was published in English translation with corrective notes by its missionary translator, Lorin Andrews. For example, in one passage Malo wrote that the maka 'ainana “are more oppressed at the present time than ever they were in ancient times” (Malo 1839:126). Here, Andrews was compelled to amend Malo's statement:

- 76

Perhaps as a result of Andrews's instruction, Malo's text is remarkably similar to contemporaneous texts produced in response to similar social ills in England. Malo's essay especially resembles Thomas Carlyle's “Chartism,” an essay that was also published in 1839. For example, Carlyle lamented an “abdication on the part of the governors” (Carlyle 1899:156, emphasis in the original), while Malo wrote that “[i]n former times, before Kamehameha, the chiefs took great care of their people” (Malo 1839:125), and both writers lamented the rise of a cash economy. These are the “sentimental tears” shed for the passing of feudalism that Marx noted and refused to join in (Marx 1975:318). Accordingly, whereas Malo declared that the “appropriate business” of ali'i was “to seek the comfort and welfare of the people” (Malo 1839:125), Carlyle asserted that what the poor labourer wants is “a superior that should lovingly and wisely govern” (Carlyle 1974:134). By 1839, this same nostalgia for paternalism lost served to lament the waning of oligarchies in England and in Hawai'i. Hawaiians had not only acquired social ills such as poverty and high infant mortality, they had also acquired a ready-made discourse to accompany those ills.

When Haole first arrived in Hawai'i, they were struck by the great numbers of children and the apparent absence of infanticide. Fifty years later, a group of Hawaiian women reportedly confessed to practising infanticide at a church meeting. Did Samwell, Turnbull and the other early explorers fail to find evidence of infanticide due to their inability to speak Hawaiian? Did Hawaiian women tell Judd what she wanted and expected to hear about the horrors of pre-Christian Hawai'i? I suggest that Haole visitors to Hawai'i failed to find evidence of infanticide before 1820—despite their expectations—because in those years Hawai'i had a relatively low rate of infant mortality and because Hawaiians did not practise infanticide. Moreover, Haole who were in Hawai'i after 1820 did find evidence of infanticide because by then the material conditions of poverty, disease and high infant mortality, upon which the English discourse are predicated, were reproduced in Hawai'i. At the same time, Hawaiians were taught to speak and write about these conditions in very much the same ways that similar conditions were spoken and written about in England. Many Hawaiian infants - 77 died, and some of their parents apparently accepted blame for their deaths.

Hawaiians, like indigenous people throughout the Pacific and the Americas, suffered catastrophic depopulation in the years immediately following their first contact with Europeans. Some of the diseases introduced by Europeans had direct, mortal effects on adults and children. Other diseases caused reduced fertility. In addition, the relatively rapid incorporation of Hawai'i into the global capitalist economy undermined the subsistence agriculture of the Islands. Thus, Hawai'i underwent changes that made the formerly irrelevant English discourse of infanticide all too relevant. As David Stannard observes in his book Before the Horror,

It is possible that epidemic disease and economic hardship caused some parents to murder their infant children, but in such cases it would be difficult to distinguish between the practice of infanticide and the discourse of infanticide. In the discourse of infanticide, parents are (mis)interpreted as responsible for their infants' deaths, regardless of the socio-economic context in which those deaths occurred. In the environment of epidemic disease that Stannard documents so well, to denominate high infant mortality as “infanticide” is to reproduce the missionary discourse by which the disastrous effects of the Haole presence in Hawai'i are obscured.

There is no evidence in the historical record of a pre-Haole Hawaiian discourse of infanticide and, not coincidentally, the infant mortality rate in pre-haole times was relatively low. There is evidence of a discourse of infanticide in post-Haole, post-1819 Hawai'i, and, also not coincidentally, diseases introduced by Haole caused the rates of miscarriages, stillbirths, and infant mortality to soar. What is being tracked here is not necessarily—or even very likely—an increase in baby-killing, but an increase in infant mortality and the concomitant increase in the relevance of the discourse of infanticide. Haole introduced to Hawai'i not only diseases and a market economy, but also a discourse from which medical and economic explanations were excluded in favour of moral explanations focused on idleness and sin.

Alfred Crosby has employed a demographic analysis to argue that Hawaiians practised infanticide from soon after the arrival of Haole until - 78 about 1880. Crosby's argument is based on uneven sex ratios. He cites a relatively high ratio of males to females in Hawai'i throughout much of the 19th century as evidence of widespread infanticide (Crosby 1994:129). Crosby observes that the naturally occurring sex ratio at birth “is something like 103 or 105 boys to 100 girls” (Crosby 1994:128). Thus, Crosby cites as evidence of infanticide an 1831 census in which there were 12 percent more males than females on O'ahu. Stannard observes, however, that a sex ratio of even 12 percent more males or females “is certainly well within the expected differential for smaller population communities” (Stannard 1991:405). Moreover, Crosby fails to note that according to the same 1831 census there were negligibly more females than males on Maui—the only other island for which there were separate counts of males and females—and that in the 1835 census of the Island of Hawai'i females outnumbered males by 7 percent (Stannard 1991:405). As Stannard argues, O'ahu “was a magnet for internal migration of male sailors and laborers within the archipelago” (Stannard 1991:405). Clearly, observed internal migrations are a more likely explanation than an unobserved practice of infanticide for the uneven sex ratio on a single island. In sum, Crosby does not provide a sound argument that Hawaiians practised infanticide before 1831.

Crosby's argument is stronger when he turns to the all-island censuses, in which internal migration is not a factor. In 1850, there were 8 percent more males than females who had been born in 1832 or earlier, and there were 25 percent more males than females who had been born between 1833 and 1850 (Crosby 1994:129). As in the 1831 census, the sex difference in the cohort born in 1832 or earlier is not significant, but the male-to-female ratio in the 1833-1850 cohort does exceed what could be expected even for smaller populations. Crosby also cites the all-island census of 1884, according to which the male-to-female ratio decreased with age (Crosby 1994:129). Among Hawaiians born in 1833 or earlier, there were 43 percent more males than females who survived to be counted in 1884. Among those born between 1834 and 1853, there were 27 percent more males than females as of 1884. Among Hawaiians born after 1853, the differences were not significant—ranging from 8 percent more males than females in the 1854-1868 cohort to 1 percent more females in the 1879-1884 cohort. Crosby uses these figures to argue that infanticide occurred in Hawai'i with decreasing frequency from early in the 19th century up until about 1880. Since there were only 8 percent more males than females in the pre-1832 cohort as of 1850, the 43 percent difference in the pre-1833 cohort as of 1884 cannot be taken as evidence of infanticide. 5 Crosby would be on surer ground if he argued that infanticide was practised in Hawai'i from about 1830 to about 1850. For this period, the 1850 and 1884 censuses are in - 79 agreement. As of 1850 and as of 1884, there were roughly 25 percent more males than females who were born between about 1833 and 1850. Thus, this is the only period for which there are significantly fewer females than males, which is what Crosby takes as evidence of infanticide.

The hole, though, in Crosby's argument—and in all arguments for Hawaiian infanticide that are based on uneven sex ratios—is the assumption that a relatively high male-to-female ratio is evidence of infanticide. Crosby argues (1) that “in many societies males are valued more than females” and (2) that “in periods of severe stress … limiting the birth rate may seem desirable, even necessary; and it is plain that eliminating girls is a more efficient means of doing so than eliminating boys” (Crosby 1994:129). Regarding the first thesis, males are valued more than females in many societies, but Crosby offers no evidence—beyond uneven sex ratios—that such was ever the case in Hawai'i. Instead, Jocelyn Linnekin demonstrates that Hawaiian females gained status precisely in the decades that Crosby argues Hawaiians were “eliminating” female infants (Linnekin 1990). Regarding the second thesis, there is no evidence that during the mid-19th century Hawaiians wanted to limit their birth rate. Rather, as Stannard observes, there is evidence that they “were desperate to have children, children who survived, and they did everything in their power to counter the devastating effects of introduced disease on their low birth rates and on their high levels of infant mortality” (Stannard 1991:404). In sum, Crosby's argument lacks a motive. Or, rather, its motive—that females were less valued than males—has nothing to do with Hawaiians and much to do with Haole. Moreover, the relatively high male-to-female ratio in the 1850 census does not pertain specifically to infants, but to Hawaiians up to 18 years of age (those born in 1833 and later). Given that there is no evidence that Hawaiians undervalued female infants, there is no basis for arguing that the relatively high female mortality rate for Hawaiians up to 18 years of age was due to infanticide.

Robert Schmitt has also argued that infanticide was practised in Hawai'i in the 1820s and 30s, and he argues that this practice had its roots in pre-Haole Hawaiian society. In an otherwise abundantly documented work, Schmitt writes, “[a]bortion and infanticide, known to have existed in precontact times, reached new highs in 1819-1825 and 1832-1836” (Schmitt 1968:37). Schmitt's claim is detailed and therefore it sounds definitive, but it is presented without evidence. The closest footnote cites an 1838 article by Artemas Bishop. Bishop wrote of high infant mortality, the destitution of the commoners, the general lack of children among the population, and the dramatic decline in the population. He specifically mentioned the years 1832-1836 as years of depopulation. “In 1832 the population amounted to rising - 80 130,000; and in 1836, only to about 110,000” (Bishop 1838:53). Thus, Schmitt appears to have taken the years 1832-1836 from Bishop, but Bishop did not connect the population decline in these years to an increase in abortion and infanticide. Bishop did mention infanticide later in his article, not, however, in connection with the years 1819-1825 or 1832-1836, but as a practice that had “existence from remote antiquity” (Bishop 1838:58). Apparently Bishop is Schmitt's sole source for stating what is “known to have existed in pre-contact times”, but we cannot accept Bishop's incidental, unsupported statements as authoritative. Considering their otherwise solid scholarship, it is noteworthy that Schmitt and Crosby are so careless in dealing with the question of infanticide.

One might expect the archaeological record to be more helpful than the historical record regarding pre-Haole infanticide. For example, there are very few foetuses or neonates in the archaeological record of Hawai'i (Collins in Han et al. 1986:240). It might be argued that this absence in the archaeological record suggests the absence of abortion and infanticide in traditional Hawaiian society, but arguments based on absence, especially in archaeology, are tenuous. If infanticide was practised, there is no good reason to assume its victims would appear in burial sites alongside other children and adults. In contemporary Haole societies, for example, aborted foetuses are not normally given burials, nor are they likely to appear in the future archaeological record. Their absence from contemporary Haole burial sites does not correspond with an absence of abortion. Aborted foetuses are under-represented in our burial sites for the same reasons that they are aborted: foetuses are not considered fully human. A similar logic could explain the absence of infant corpses in the burial sites of any society that practised infanticide. Indeed, the premise of mortuary analysis is that the treatment that an individual received post-mortem corresponds in some way with the treatment that individual received pre-mortem. Thus, if foetal or neonatal bodies are not protected by societal prohibitions regarding murder, they should not be expected to be protected by their society's burial rites either.

Despite the absence of evidence of high infant mortality in pre-Haole Hawai'i, archaeologist Patrick Kirch has argued that pre-Haole Hawaiians widely practised infanticide. Kirch claims that

- 81

In the work cited, Kirch fails to specify what evidence there is for Hawaiians' practising infanticide at the time of Cook's arrival. In another work Kirch (1984:234) cites an 1867 text by the Hawaiian historian Samuel Kamakau, but this text hardly provides even minimal evidence of infanticide nearly a century earlier. As Stannard observes, the published English translation of this text incorrectly states that infanticide was “another evil practiced in pagan days and still made use of today” (Kamakau 1961:234). The original Hawaiian text refers to “olden times”, not to “pagan days” (Stannard 1991:396). Moreover, Kamakau observed that infanticide “became more common in late days” (Kamakau 1961:235), which is consistent with my finding that a discourse of infanticide increased in the 1820s and thereafter.

Kirch's argument that infanticide was a pre-Haole Hawaiian practice is derived from his theory that the Hawaiian population began to decline before the arrival of Haole—and their diseases—because it had reached the “carrying capacity” of the islands. Kirch reasons that Hawaiians had to practise infanticide because they were not able to feed all of the children who were born. This theory of infanticide as a method of population control can be traced back to Malthus' Essay on the Principle of Population (1798:46). For Malthus, as for Kirch, infanticide was a practical response to inevitable population pressures. Charles Darwin was the first to apply this Malthusian logic to the case of (purported) infanticide in Hawai'i. In The Descent of Man, Darwin cited Ellis's (1969) accounts of Hawaiian infanticide, and he concluded that “these practices appear to have originated in savages recognizing the difficulty, or rather the impossibility of supporting all the infants that are born” (Darwin 1887:46). Darwin also received confirmation that Hawaiians practised infanticide from Titus Munson Coan, the son of the Reverend Titus Coan, a Hilo missionary. In 1899, the younger Coan published a “Hawaiian Ethnography” in The Bulletin of the American Geographic Society. In this work Coan employed Darwin's infanticide theory to praise Hawaiians and other Polynesians:

- 82

Kirch's theory of Hawaiian infanticide is not so romantic as Coan's. Rather than see infanticide as a method to preserve an idyllic lifestyle, Kirch sees infanticide as a last resort in the face of famine. Still, both Kirch's and Coan's theories of Hawaiian infanticide can be located in the Malthusian discourse that was imported to Hawai'i by Darwin.

Kirch's theory about the pre-Haole depopulation of Hawai'i is derived from counting and dating habitation sites, “based on the reasonable assumption that the number of permanent habitation sites at any given time is allometrically related to the size of the population inhabiting them” (Kirch 1982:4). A decrease in the number of sites dated in the late “pre-contact” period leads Kirch to conclude that the population must have decreased just before “contact”. Kirch attributes this decrease to a lowered “carrying capacity,” which was due to “centuries of intensive exploitation, modification, manipulation, and, frequently, degradation [of the environment]” (Kirch 1982:10). Kirch's theory of pre-Haole population decline in Hawai'i has been challenged on numerous fronts. Jeffrey Clark has identified major flaws in Kirch's method (Clark 1988). Some of the flaws are inherent in the data Kirch took from others' work, namely Hommon (1976) and Cordy (1981). These flaws include the failure to distinguish between permanent and temporary residences, and the dubious assumptions that the average number of people per residence remained constant over several centuries and that the latest reported date coincides with the abandonment of a residence. Most importantly, sites that were inhabited in post-Haole times were eliminated from the sample. For example, an accurately dated site that was inhabited from 1650 to 1820 would not be included in Kirch's data. According to this method, as old dwellings are abandoned and replaced with newer dwellings, the population appears to decline. If this method were applied to 20th century Hawai'i—discarding habitation sites that were occupied after 1959, for example—a similar decrease in population would be found to have occurred just before 1959 despite the actual and dramatic population increase during the period. Clark considers these methodological problems severe enough to conclude that “there is no supporting evidence to indicate a prehistoric population decline in the Hawaiian Islands beyond the atypical islands of Nihoa and Necker” (Clark 1988:28).

Michael Graves and Thegn Ladefoged have added to the debunking of Kirch's depopulation theory. Graves and Ladefoged note that volcanic glass dating varies with the environment and the source of the volcanic glass, and that volcanic glass dates have yet to be calibrated specifically for Hawai'i. Kirch makes relatively uncritical use of volcanic glass dating in his analysis of Hawaiian demography even though he is well aware of its potential - 83 problems (Kirch 1984:105). Comparing radiocarbon dates to volcanic glass dates from the same sites on the Island of Lana'i, Graves and Ladefoged find a pattern. Volcanic glass dates tend to be one to two centuries earlier than radiocarbon dates. “If the midpoint volcanic glass dates are at least 150 years too early, as suggested by the Lana'i data, the actual peak in the population curve would be shifted to a later time period, at about A.D. 1800” (Graves and Ladefoged 1991:24-25). Thus, a preliminary recalibration of Kirch's dates indicates that the population of Hawai'i began to decline not a century before the arrival of Haole, but immediately afterward. Considering the documented (and the undocumented) epidemics that plagued Hawai'i in the first 50 years of Haole-contact, there is no need to look to infanticide or environmental degradation to explain this population decline.

Stannard (1989) contributes ethnohistorical evidence to the case against Kirch's theory. Stannard observes that Hawaiians displayed none of the characteristics associated with people living on the ecological fringe. The earliest written accounts of Hawai'i contradict the claim that Hawaiians suffered from population pressures. Stannard notes that there is no evidence of malnutrition. Quite the opposite, the earliest Western explorers wrote that the Hawaiians were remarkably healthy and that they had an abundance of food (Stannard 1989:67). Stannard also points out that Hawaiian culture did not conform to anthropological models for peaked populations. For example, Hawaiians did not place severe restrictions on mating or on leisure activities (Stannard 1989:68; cf. Kirch 1984:118).

In addition to lacking evidence, Kirch's argument about the relation between infanticide and depopulation contains a logical paradox. Kirch—like Ellis and Darwin—considers infanticide to be both an effect of overpopulation and a cause of depopulation. Let us suppose that Kirch is right, that Hawai'i did reach its “carrying capacity” in pre-Haole times and that infanticide began to be practised as a result. Let us also suppose with Kirch that the “carrying capacity” of the islands was in a continuous decline due to environmental degradation. Thus, infanticide would have to continue even as the population began to decrease. This decrease in the population would translate into a reduction in the labour force necessary to maintain agricultural and aquacultural intensification, thereby leading to a further decrease in the “carrying capacity”. The question i: Is there no end to this downward cycle? The population of Hawai'i was reduced to about 130,000 people by 1828 (Schmitt 1968)—the time at which Judd recorded hearing Hawaiian women confess to practising infanticide. By all accounts this is a small fraction of what it was at its peak—probably close to a million, or more (Stannard 1989), but at the very least 300,000 (Schmitt 1968). Given this tremendous depopulation, why would Hawaiians in 1828 continue to - 84 practise infanticide? It is inconsistent to argue that infanticide was both a coldly practical response to overpopulation and a widely impractical cause of catastrophic depopulation.

Sahlins offers an argument that serves to correct the biological reductionism implicit in Kirch's analysis. Sahlins observes that the practice of infanticide cannot be positivistically associated with either overpopulation or depopulation. Here, Sahlins takes up a line of thought he launched in Culture and Practical Reason (1976) to assert that needs are culturally constructed. Concepts such as “carrying capacity” and “population”—and their derivative, “overpopulation”—must be thought of as cultural constructions and not as independent variables.

Biological explanations about the “carrying capacity” of a particular valley are inappropriate because human organisation is not optimal. In traditional Hawai'i, individuals did not have direct access to land. Land tenure was mediated by an individual's ties to kinship groups and by those groups' ties to ali'i. Thus, Sahlins argues that a family might simultaneously practise infanticide and adoption. The important variable is not how many children there are per acre, but how each particular child fits—or does not fit—into existing kinship and political orders. This analysis has the virtue of disassociating infanticide from population per se, allowing Sahlins to avoid the paradox that confounds Kirch.

Still, Sahlins, like Kirch, argues that infanticide was a pre-, as well as a post-Haole Hawaiian practice. On one level, since Sahlins assumes that people inevitably act according to structured principles, he can—and does—let the principles stand for the behaviour. Thus, rather than provide evidence that pre-Haole Hawaiians actually practised infanticide, Sahlins offers an explanation for why they would have found infanticide structurally necessary. On another level, Sahlins' thesis that “every transformation is a reproduction” allows him to discover evidence about pre-Haole Hawai'i in even the most - 85 peculiarly post-Haole texts. Thus, Sahlins is able to take accounts of infanticide from the 1820s and 30s (and later) as evidence of pre-Haole Hawaiian infanticide. Change itself is structured and thereby contained, so every present, according to Sahlins, reveals much of its past.

Sahlins's powerful model works better for European societies than for Hawai'i or other Pacific Island societies. Over several centuries, Europeans had contact with one another, with Near Easterners and North Africans, with Sub-Saharan Africans and Far Easterners, and eventually with Native Americans and Pacific Islanders. Europeans perceived each of these Others according to established categories of Otherness, including that of the “savage”. To some extent, such categories were altered by their application to new groups, though I would emphasise the ways that the groups were altered to fit the categories. Still, Sahlins' model seems to fit the contact experiences of Europeans better than it fits the contact experiences of Pacific Islanders in general and of Hawaiians in particular. To be sure, Hawai'i was “colonised” by Polynesians, whose descendants were later contacted by other Polynesians. The tradition of Polynesian voyaging brought about periodic contact experiences, for example between Hawaiians and Tahitians, but Hawaiian contact with Haole was radically different from Hawaiian contact with other Polynesians. Contact experiences with Marquesans or Tahitians may have caused significant technological changes in Hawai'i—including new fishhook designs, for example—and these experiences are fairly covered by a model of structural reproduction and transformation, but the Hawaiian contact with Haole was something new and different.

To focus on structural reproduction and transformation in the early post-Haole period in Hawai'i is to overlook the peculiar effects of European and U.S. colonialism. Sahlins presents Hawaiian history as structurally continuous despite the discontinuities brought about by contact with Haole,and their wholly unprecedented metallurgy, weaponry and diseases. For Hawaiians, however, contact with Haole was not a case of continuous mythical realities, but of discontinuous realised myths. There is, for example, no pre-Haole Hawaiian discourse of infanticide—no story like that of Abraham and Isaac or that of poor women lacking in maternal instinct. The post-1819 Hawaiian discourse of infanticide is not best understood as the transformation of a traditional myth because infanticide was not a mythical Hawaiian reality. Rather, infanticide was a realised Haole myth about Hawai'i. Haole introduced the mythical discourse of infanticide to Hawai'i, and it was the presence of Haole in Hawai'i that made that discourse true, but the discourse did not become true in the sense that Hawaiians truly practised infanticide. There may have been Hawaiian parents who killed their infants, either out of desperation or mercy, and there were certainly - 86 cases of parents who helplessly and hopelessly watched their children die, but the discourse of Hawaiian infanticide became true in the sense that it became relevant: high infant mortality occurred, and its occurrence required an explanation.

Sahlins's method has proven to be tremendously productive, but its limits must be marked. Because so many non-Western societies were so changed by contact with the West, it is necessary to employ extreme caution in using “post-contact” sources to speculate about “pre-contact” societies. To some extent, in the case of Hawai'i, this sort of salvage ethnography or salvage ethnohistory is unavoidable. All written records were produced in post-Haole times: text produced by Haole sailors beginning in 1778, by missionaries beginning in 1820, and by Hawaiian scholars beginning in the 1830s. There is a large and growing archaeological record pertaining to pre-Haole Hawai'i, but there are limits to the sort of questions on which the archaeological record by itself can shed light. Most of the questions that interest cultural anthropologists require recourse to written records. Thus, it is necessary to make inferences based on post-Haole texts to answer questions about pre-Haole Hawaiian culture. Given Hawai'i's extreme geohistorical isolation, these sources must be used with caution. Hawai'i may have been more drastically transformed in the decades following contact with Europeans than any other place. For these reasons, the very earliest sailors' journals are of special importance in reconstructing the pre-Haole culture of Hawai'i. We should be especially skeptical of observations made in the 1820s or later that are not corroborated—or that are contradicted—in these earliest texts.

A more complex approach to post-Haole Hawaiian history would supplement Sahlins' structural method, by which continuities are identified, with Foucault's archaeological method, by which discontinuous layers are left unconnected (Foucault 1970). Much as European society underwent an epistemic shift at the end of the 18th century, Hawaiian society underwent an epistemic shift at just the same time, but while the causes of the shift are unremarked in Foucault's study of Europe, there is no mystery regarding the causes of the shift in Hawai'i. The arrival of Haole and their diseases caused obviously abrupt changes in Hawaiian society. The catastrophic depopulation of Hawai'i that followed from the presence of Haole and the effects of that depopulation on Hawaiian society were not primarily structural—and structured—transformations. Through contact with Haole, the structures of pre-Haole Hawaiian society were not reproduced so much as they were replaced.

The danger in using post-Haole texts to make inferences about pre-Haole - 87 Hawai'i is that peculiarly post-Haole phenomena—such as the discourse of infanticide—are wrongly projected into the pre-Haole past. Or, more insidiously, a post-Haole present acquires a corresponding pre-Haole past that has much to do with Europe and little to do with Hawai'i. This process by which Hawaiian history is Europeanised is what Hawaiian scholars Lynette Cruz and J. Kalani English refer to as “historical colonialism” (Cruz and English 1990:280). For example, I have argued that Malo's essay “On the decrease of population…” (Malo 1839) and the speech of Hawaiian women reported by Judd (1880) reflect the introduction of Haole discourses to Hawai'i. Malo's essay and other texts by 19th century Hawaiian historians reproduce a European nostalgia for feudal times gone by. In these cases, the capitalist present carried with it the construction of a peculiarly pre-capitalist past. Consider the case of William Richards, the missionary-turned-political-advisor most responsible for teaching Hawaiians to be capitalists (see Kame'eleihiwa 1992:176-80). Richards believed that traditional Hawaiian government had been “mainly of a feudal character” (Richards 1973:21). Richards obviously found it convenient to conceive of the society to which he was introducing capitalism as feudal. The construction of a pre-capitalist, feudal past helped Richards naturalise the Hawaiian adoption of capitalism. What is surprising is that 20th century anthropologists have repeated Richards' characterisation of pre-Haole Hawaiian society. Valerio Valeri, in particular, has argued at length that pre-Haole Hawaiian society was feudal (Valeri 1985), while Matthew Spriggs has found class formation, as well as feudalism, in Hawai'i before Cook's arrival (Spriggs 1988). Valeri and Spriggs have projected into pre-Haole Hawai'i a past that corresponds with the “pre-capitalist” stage in Europe, but capitalism in Hawai'i did not develop in the same way that capitalism developed in Europe. Capitalism in Europe may be seen as having developed out of feudalism, but capitalism in Hawai'i did not develop out of anything Hawaiian, though it did require Hawaiian collaboration and submission. Capitalism in Hawai'i was wholly imposed from the outside, by Haole, as part of a colonising project.

Hawaiian historian Lilikal Kame'eleihiwa has made a similar argument regarding the anthropologically famous 'Aikapu (eating restrictions) abolition of 1819 (Kame'eleihiwa 1992). Anthropologists from Kroeber (1948) to Valeri (1982) have attempted to locate the abolition within a continuous Hawaiian history. Thus, Valeri argues that the abolition was a structural transformation, which followed a logic set in motion by the apotheosis and death of Cook. Kame'eleihiwa, by contrast, uses the data presented by Stannard (1989) regarding the depopulation of Hawai'i between 1778 and 1819 to argue that the 'Aikapu abolition was a radical break as - 88 opposed to a logical continuation. Where Valeri sees Hawaiians as slaves to their culture and its reproduction, Kame'eleihiwa sees them as self-conscious agents, who chose deliberately to abandon the 'Aikapu because they determined that it was no longer working.

In Kame‘eleihiwa’s account of the 'Aikapu abolition—as in Obeyesekere's account of the supposed apotheosis of Cook—Hawaiians are represented as being practically rational. Kame'eleihiwa and Obeyesekere recognise that pre- and post-Haole Hawaiians were capable of recognising something as new—for example, the arrival of Haole or the massive depopulation that followed. Thus, Kame'eleihiwa argues that Hawaiian society underwent a radical, unprecedented shift as a result of the Haole presence in Hawai'i. She explains both the 1819 'Aikapu abolition and the 1848 Māhele as highly deliberate attempts by Hawaiians to respond effectively to the disease, depopulation and destitution that followed from the arrival of Haole in Hawai'i. In her account, Hawaiians were not bound to reproduce their cultural categories. Rather, they were capable of abandoning those categories that were not working, and of adopting Haole categories that their observations—and their Haole advisors—gave them reason to trust.

I thank Beth Fowkes Tobin for directing me through the English discourse of poverty in Great Britain, North America, and the Caribbean. I thank Michael Graves, Joseph Tobin, Marta Savigliano and James Faubion for commenting on various drafts of this paper.

- 89

• Beaglehole, J.C. (ed.), 1967. The Journals of Captain James Cook on His Voyages of Discovery. Published for the Hakluyt Society. Cambridge: Cambridge University Press.
• Bishop, Artemas, 1838. An inquiry into the causes of decrease in the population of the Sandwich Islands. Hawaiian Spectator, 1:52-66.
• Buck, Elizabeth, 1986. English in the linguistic transformation of Hawaii. World Englishes, 5:141-52.

• Carlyle, Thomas, 1899 [1839]. Chartism. In Critical and Miscellaneous Essays, Vol. IV. London: Chapman and Hall, pp.118-204.
• Chamberlain, D., 1820. Journal. Typescript by Roland B. Dodge. Hawaiian Mission Children's Society.
• Clark, Jeffrey T, 1988. Palaeodemography in leeward Hawaii. Archaeology in Oceania, 23:22-30.
• Coan, Titus Munson, 1899. Hawaiian ethnography. Bulletin of the American Geographic Society, 31:24-30.
• Coates, D., John Beecham and William Ellis, 1837. Christianity: The Means of Civilization. London: R.B. Seals and W. Burnside.
• Cordy, R.H., 1981. A Study of Prehistoric Social Change. New York: Academic Press.
• Crosby, Alfred W., 1994. Hawaiian depopulation as a model for the Amerindian experience. In A.W. Crosby (ed.), Germs, Seeds, and Animals: Studies in Ecological History. Armonk, NY: M. E. Sharpe, pp.120-47.
• Cruz, Lynette and J. Kalani English, 1990. Review of Before the Horror by David Stannard. Pacific Studies 13:279-84.

- 90

• Darwin, Charles, 1887 [1871]. The Descent of Man and Selection in Relation to Sex. New York: D. Appleton and Co.
• Daws, Gavan, 1968. Shoal of Time: A History of the Hawaiian Islands. Honolulu: University of Hawaii Press.

• Ellis, William, 1963 [1825]. Narrative of a Tour of Hawaii, or Owhyhee; with Remarks on the History, Traditions, Manners, Customs and Language of the Inhabitants of the Sandwich Islands. Honolulu: Advertiser Publishing.
• —— 1969 [1842]. Polynesian Researches: Hawaii. Rutland, Vt: Charles E. Tuttle.

• Foucault, Michel, 1970. The Order of Things. New York: Random House.
• —— 1971. Orders of discourse. Social Science Information, 10:7-30.
• Freycinet, Louis Claude de Saulces de, 1978 [1825-1839]. Hawaii in 1819: A Narrative Account. Edited by Marion Kelly. Pacific Anthropological Records No. 26. Honolulu: Bernice P. Bishop Museum.

• Graves, Michael W. and Thegn N. Ladefoged, 1991. The disparity between radiocarbon and volcanic glass dates: New evidence from the island of Lāna'i, Hawai'i. Archaeology in Oceania, 26:70-77.

• Han, Toni L., Stephen D. Clark, Anne Garland and Sara L. Collins, 1986. Moe Kau Ho'oilo: Hawaiian Mortuary Practices at Keōpū, Kona, Hawai'i. Honolulu: Bernice P. Bishop Museum, Department of Anthropology Report 86-1.
• Hommon, R. J., 1976. The Formation of Primitive States in Pre-contact Hawaii. Ph.D. dissertation, University of Arizona.

• Jefferson, Thomas, 1955. Notes on the State of Virginia. Edited by William Peden. Chapel Hill: University of North Carolina Press.
• Judd, Laura Fish, 1880. Honolulu: Sketches of Life. New York: Anson D.F Randolph.

• Kamakau, Samuel M., 1961 [1866-1871]. Ruling Chiefs of Hawaii. Honolulu: Kamehameha Schools Press.
• Kame 'eleihiwa, Lilikala, 1992. Native Land and Foreign Desires: Ko Hawai 'i 'Āina me Nā Koi Pu'umake a ka Po'e Haole. Honolulu: Bernice P. Bishop Museum Press.
• Kirch, Patrick V., 1982. The impact of the prehistoric Polynesians on the Hawaiian ecosystem. Pacific Science, 36:1-14.
• —— 1984. Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press.
• —— 1985. Feathered Gods and Fishhooks: An Introduction to Hawaiian Archaeology and Prehistory. Honolulu: University of Hawaii Press.
• Kroeber, A.L., 1948. Anthropology. New York: Harcourt, Brace and Co.

• Lewis, Matthew Gregory, 1969 [1817]. Journal of a West India Proprietor, Kept During a Residence in the Island of Jamaica. New York: Negro Universities Press.
• Linnekin, Jocelyn, 1990. Sacred Queens and Women of Consequence: Rank, Gender, and Colonialism in the Hawaiian Islands. Ann Arbor: University of Michigan Press.
• Locke, John, 1894 [1690]. An Essay Concerning Human Understanding. Edited by Alexander Campbell Fraser. Oxford: Clarendon Press.
• Loomis, E., 1821. Journal. Microfilm, Hamilton Library, University of Hawai'i-

- 91

• Mānoa.
• Malo, David, 1839. On the decrease of population on the Hawaiian Islands. Translated by L. Andrews. Hawaiian Spectator, 2:121-30.
• Malthus, Thomas Robert, 1798. An Essay on the Principle of Population as it Affects the Future Improvement of Society. London: J. Johnson.
• Marx, Karl, 1975 [1844]. Economic and philosophical manuscripts. In Early Writings. New York: Vintage Books, pp.279-400.
• Melville, Herman, 1965 [1844]. Typee: A Peek at Polynesian Life. New York: Airmont Publishing Co.

• Nietzsche, Frederich, 1967 [1887]. On the Genealogy of Morals, and Ecce Homo. Edited by Walter Kaufmann. New York: Vintage Books.

• Obeyesekere, Gananath, 1992a. The Apotheosis of Captain Cook: European Mythmaking in the Pacific. Princeton and Honolulu: Princeton University Press and Bernice P. Bishop Museum Press.
• —— 1992b. “British cannibals”: Contemplation of an event in the death and resurrection of James Cook, explorer. Critical Inquiry, 18:630-54.

• Richards, William, 1973 [1841]. William Richards on Hawaiian culture and political conditions of the Islands in 1841. Edited by Marshall Sahlins and Dorothy Barrère. Hawaiian Journal of History, 7:18-40.

• Sahlins, Marshall, 1976. Culture and Practical Reason. Chicago: University of Chicago Press.
• —— 1981. Historical Metaphors and Mythical Realities: Structure in the Early History of the Sandwich Islands Kingdom. ASAO Special Publications, No. 1. Ann Arbour: University of Michigan Press.
• —— 1985. Islands of History. Chicago: University of Chicago Press.
• —— 1992. Historical Ethnography. Volume One of Anahulu: The Anthropology of History in the Kingdom of Hawaii, by Patrick V Kirch and Marshall Sahlins. Chicago: University of Chicago Press.
• Schaw, Janet, 1939 [1774]. Journal of a Lady of Quality; Being the Narrative of a Journey from Scotland to the West Indies, North Carolina, and Portugal in the years 1774 to 1776. Edited by Evangeline Walker Andrews in collaboration with Charles McLean Andrews. New Haven: Yale University Press.
• Scheper-Hughes, Nancy, 1992. Death Without Weeping: The Violence of Everyday Life in Brazil. Berkeley: University of California Press.
• Schmitt, Robert, 1968. Demographic Statistics of Hawaii, 1778-1965. Honolulu: University of Hawaii Press.
• Spriggs, Matthew, 1988. The Hawaiian transformation of ancestral Polynesian society: Conceptualizing chiefly states. In J. Gledhill et al. (eds), State and Hierarchy. London: Unwin Hyman, pp.57-73.
• Stannard, David E., 1989. Before the Horror: The Population of Hawai'i on the Eve of Western Contact. Honolulu: Social Science Research Institute, University of Hawaii.
• —— 1991. Recounting the fables of savagery: Native infanticide and the functions of political myth. Journal of American Studies, 25:381-418.
• Stewart, C.S., 1970 [1830]. Journal of a Residence in the Sandwich Islands During

- 92

• the Years 1823, 1824, and 1825. Honolulu: University of Hawaii Press.
• Swift, Jonathan, 1955 [1729]. A modest proposal. In The Prose Works of Jonathan Swift, Volume the Twelfth: Irish Tracts, 1728-1733. Edited by Herbert Davis. Oxford: Basil Blackwell.

• Tobin, Beth Fowkes, 1993. Superintending the Poor: Charitable Ladies and Paternal Landlords in British Fiction, 1770-1860. New Haven: Yale University Press.
• Turnbull, John, 1810. Voyages in the Pacific Ocean. London: A. Maxwell.

• Valeri, Valerio, 1982. The transformation of a transformation: A structural essay on an aspect of Hawaiian history (1809-1819). Social Analysis, 10:3-41.
• —— 1985. The conqueror becomes king: A political analysis of the Hawaiian legend of 'Umi. In A. Hooper and J. Huntsman (eds), Transformations of Polynesian Culture. Memoir No. 45. Auckland: Polynesian Society, pp.79-103.