00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 0Chapter 4 Household Archaeology in Kipapa and Nakaohu, Kahikinui Cindy Van Gilder and Patrick V. Kirch Introduction Archaeology is a unique way of knowing about the human past. Unlike archival research or other historical techniques, archaeology takes the modem researcher physically into the houses and traditional spaces of old; it provides a potential means of understanding the organization and rhythm of ordinary lives in the past all over the world, for which we have no specific written documentation. For example, in a complex and varied society like that of pre-contact Hawai'i, there must have been much local and regional variation in cultural practices. One goal of the Kahikinui Archaeological Project is to document and seek to understand the daily life of the unique communities that once inhabited this moku. To this end, we have selected for study several structures in Kipapa and Nakaohu Ahupua'a, using the approach and methods termed household archaeology. This chapter describes household archaeology, our methods, the preliminary results of our two seasons of work in Ktpapa and Nakaohu, and some of the questions we hope to answer through continued study. Household Archaeology During the earlier part of this century, archaeology in Hawai'i was primarily an ancillary discipline to ethnography and ethnohistory, and all were part of an anthropologi- cal endeavor that sought to document traditional Hawaiian culture as it was presumed to have existed at the time of the first arrival of Europeans. During this period, archaeolo- gists concentrated primarily on the study of known heiau sites and other large monuments such as fishponds, or on petroglyphs (Hiroa 1947). The discovery of radiocarbon dating in the late 1940s revolutionized the practice of archaeology worldwide, by providing a 46 Ni Mea Kahiko o Kahikinul means of giving "absolute" chronological dates to a site. close attention to detail, to the distribution of such mundane Kenneth P. Emory took advantage of this new method and things as fire pits, discarded food remains, and the waste flakes obtained the first radiocarbon date for Hawai'i, from the left by stone tool working on the floors of ancient houses. It rockshelter occupation at Kuli'ou'ou rockshelter, O'ahu. The requires extensive excavation of house floors to expose corrected age of A.D. 1004 ? 180 was older than he or other complete spatial patterns, and not just limited test pits. In areas anthropologists had expected (Emory and Sinoto 1961), and such as Hawai'i where much daily activity may have taken inspired Emory and his colleagues to undertake an extensive place outside of the confines of house walls, it requires program of excavations in Hawaiian sites during the 1950s. excavating seemingly "vacant" spaces between structures. Emory's work focused on chronological issues and on the Household archaeology is still a new approach in definition of stylistic differences among certain artifacts, Polynesia and Hawai'i, although there have been important especially fishhooks (Emory, Bonk, and Sinoto 1959; see applications in New Zealand (Sutton 1990), the Cook Islands also Kirch 1985). (Walter 1993), and on Moloka'i in the Hawaiian Islands In the 1960s a new approach in archaeology, which (Weisler and Kirch 1985). One of the aims of our Kahikinui asked different kinds of questions about the archaeological Archaeological Project has been to apply and develop the record, began to be applied in Polynesia (Green 1967). This methods of household archaeology. As we shall explain, this is "settlement pattern approach" was viewed as a way to move also essential in order to interpret the thousands of stone beyond questions of chronology, to the reconstruction of structures identified by archaeologists in their settlement ancient systems of social and political organization. Settlement pattern surveys, and in its own right as a way to open a window pattern archaeology involves documenting all of the sites-no on the daily lives of those maka'dinana who once claimed matter how small or seemingly unimportant-on a landscape, Kahikinui as its kupa o ka 'aina. in order to determine their interrelationships and gain insight into the organization of the community as a whole. It was this Traditional Households in Hawai'i approach that Peter Chapman, inspired by the work of Roger Green in Mo'orea (Society Island) and elsewhere, began at Before turning to archaeological evidence from Kahikinui (see Kirch, Chapter 1, this volume). Chapman ancient Kahikinui households, it is important to review what ambitiously began a survey of two ahupua'a, Kipapa and has been recorded about the traditional households of ka po'e Nakaohu, and although his work was never published, his data kahiko. Most of our information comes from a limited range of form the foundation of the University of California at sources. The first of these are written descriptions by early Berkeley's Kahikinui Archaeological Project (K.A.P.). European visitors, beginning with Cook's expedition and Settlement pattern archaeology is designed to answer continuing over the next few decades (after the abolition of the questions at a broad regional scale, that of the community and kapu system, in 1819, the organization of Hawaiian households the polity. The distributional patterns of structures and sites and houses began to change in certain significant ways). As across a landscape give testimony to the social and political firsthand eyewitness descriptions, these accounts are very structures of the society that once occupied and organized the significant, yet we must be careful about the potential biases of land. If, however, we are interested in more of a micro-scale, culturally-uninformed Western observers. Second, we have the such as the daily practices of Native Hawaiian maka'dinana in accounts of several Native Hawaiian scholars of the mid- Kahikinui, a different archaeological approach must be applied. nineteenth century (most notably Kamakau [ 1976] and Malo The method and theory of household archaeology is designed [1951]) who wrote from personal experience and cultural to retrieve information about the organization of a society at a perspective about houses, households, and daily life. Yet these smaller scale within a community, the scale at which individual accounts are brief, and also tend to represent an ali'i or chiefly family units lived. In a real sense, household archaeology perspective. Third, there are later, twentieth-century efforts by might also be called "family archaeology." both haole and Native Hawaiian scholars to synthesize a Household archaeology is still a relatively new picture of Hawaiian domestic life. These include Brigham approach within the field of archaeology as a whole, and its (1908), Handy and Pukui (1958), Hiroa (1957), and Forster methods and theory are therefore still developing (see Blanton (1960), whose researches are invaluable, yet to a large extent 1993; Kent, ed., 1990; Wilk and Rathje, eds., 1982). Like dependent on the first two groups of evidence. settlement pattern archaeology, household archaeology is As an example of the first category of evidence, we concerned with spatial patterning of human activities-as may turn to the account of La Perouse, who in 1786 was the evidenced in the material remains they left behind-but at a first European to enter and describe a Maui house, at micro rather than macro scale. Household archaeology requires Keone'6'io in Honua'ula, not too distant from Kahikinui. The Household Archaeology in KMpapa 47 French explorer's journal reads as follows: Such folks only cared for a little shanty, anyway; the On our walk we came upon four small villages fireplace was close to their head, and the poi dish [households, kauhale ?] of 10 to 12 houses; they are conveniently at hand; and so, with but one house, built and roofed with straw [presumably pili grass] they made shift to get along. and resemble [sic] those of our poorest peasants, the 17. People who were well off, however, those of roofs are coupled, the door is usually situated at the respectability, of character, persons of wealth or who gable end, it is only 3 1/2 feet in height and anyhow belonged to the alii class, sought to do everything can open. Their furnishings are merely mats which, decorously and in good style; they had separate' like our carpets, make a very clean floor on which houses for themselves and for their wives. these natives sleep. Their only kitchen utensils are 18. There was a special house for the man to gourds of a large size to which they give the desired sleep in with his wife and children (hale noa), also a shape when they are still green; they varnish them number of houses specially devoted to different kinds and paint various designs on them in black of work, including one for the wife to do her work in (Dunmore, ed., 1994:89). (hale kua). There was the halau, or canoe house; the This is certainly a useful account (despite the bias of the aleo, a kind of garret or upper story in which to stow French nobleman's comparison with "our poorest peasants"), things; also the amana, consisting of three houses yet it leaves out so much that one might wish for. Surely every built about a court. house was not identical? Where were the hearths and cooking 19. This way of living corresponded with what ovens? How many people lived in these little "villages"? These the Hawaiians regarded as decent and respectable and a thousand more questions cannot be answered from (Malo 1951:122). La Perouse's text, or from those of other early European While in contrast to La Perouse's text we now have an observers. "insider" cultural account, even more questions leap to mind. Let us turn then to the accounts of traditional What, for instance, was the "amana," described only elusively households authored by Native Hawaiian authorities of the as "three houses built about a court"? No such definition of nineteenth century. Foremost among these scholars was David amana appears in the definitive Pukui and Elbert Hawaiian Malo, born in 1795 at Keauhou, Kona, and a member of the Dictionary.2 Clearly, not all Hawaiians of Malo's day lived the famous first class of students at Lahainaluna Seminary, same kind of existence, or why else would he make such a established in 1831 (Chun 1993). In his invaluable Mo'olelo point of differentiating the habits of the lapuwale from those of Hawai'i, Malo (1951:118-24) devoted a short chapter to "The po'e ko'iko'i? In Malo's writing we get a clear indication of the House-Its Furniture and Its Construction." (The Hawaiian importance-under the kapu system-of maintaining separate text and a translation are also given in Brigham [1908:76-79, facilities for men and women. It is evident that Hawaiian 122-23].) Most of the account deals with the technical aspects households consisted of clusters of functionally-different of house construction, thatching, the rites of dedication, and structures. Yet we are given no details, no hint of the range of furnishings (as does the similar account by Kamakau [1976:95- variation in daily practice. To Malo, such was probably 108]). With regard to variation in styles of houses and mundane and self-evident, not worthy of his time. Would that households among the Hawaiian population, and to functional we could interview the great scholar today! differences between houses within a household (kauhale), Our third category of established evidence comes Malo's text is vital but brief. Let us quote it in full: from those twentieth century researchers who attempted to 16. People who were of no account (lapuwale) systematize these earlier accounts together with their own did not follow this practice. They went in and observations of Hawaiian culture. None of these later works occupied their houses without any such ceremony. could be more important and influential than that of the At this point, Malo's 1898 editor, Nathaniel B. Emerson (son of missionary parents and a fluent Hawaiian speaker) inserted a lengthy footnote regarding the necessity of "a number of houses, the chief motive being to separate the sexes entirely from each other while eating," an aspect of the kapu system. Emerson goes on to enumerate such houses, including the mua, hale noa, hale 'Gina, hale kua, and hale pe 'a. It must be kept in mind, however, that Emerson was here writing from his own experience and knowledge, gained freom his childhood in at Kawailoa, O'ahu. 2 One of us (PVK) has observed a repeated pattern of three closely adjacent house foundations in parts of southeast Maui, as for example, at Waia'ilio in Kanaio Ahupua'a. Could this possibly be what Malo's reference to the amana is about? We are reminded also of the Russian explorer Golovnin's description that "each Sandwich Islander has to have three houses, cabins, or huts, depending on his status: one is used to sleep in, one for the men to eat in, and one for the women to eat in" (Golovain 1979:178). 48 No WMa Kahiko oKahIkinui collaborative team of E. S. Craighill Handy and Mary Kawena * Hak Noa. "Everybody slept in the Hale Noa (House Pukui, whose study of The Polynesian Family System in Ka'u, freed of kapu), where no restrictions were placed on the Hawai'i remains a classic (Handy and Pukui 1958). Handy, an men and women sharing it together. This house was for anthropologist with experience in the Marquesas, Tahiti, and sleeping and no eating was permitted there" (1958:9-10). other parts of Polynesia, drew upon Pukui's vast knowledge as The hale noa was divided into a sleeping place and a a punahele child raised in rural Ka'u, to jointly author the most sitting or walking place. The sleeping place was raised detailed and nuanced account of traditional Hawaiian house- and covered with finer mats. When unused for sleeping, hold life. Their text is too extensive to quote at length here, but no one was allowed to walk, sit or play on it. It was let us provide a sampling: kapu (1958:10). Within the 'ohana the functional unit is the * Hak Pe'a. "The women had another house that was household. One term used for household was the built near but not too close to the other houses of the word hale, house. In inquiring about the number of kauhale. This was the hale pe'a, a small comfortable "families" or domiciles in a given locality, one would thatched house where the women of the family retired ask "Ehia hale la?" (How many houses?) 'Ohua was when menstruating and remained until the period was a term that signified retainers or dependents in the completely over" (1958:10). household... . The household included members of * Halau. "A fisherman had a halau or long thatched the family proper of all ages plus attached but house where he kept his canoe, fishnets, and other unrelated dependents and helpers. The po'o ("head") paraphernalia. Kapu were enforced there also, for no or functional head of the domicile was not necessar- women were permitted to handle the large nets, or was ily the senior member; it was and is specifically the anyone allowed to step over the lines, hooks or nets" member who assumes responsibility and makes (1958:11). decisions (1958:5) * Hale Papa'a. "An inland dweller would have a Every Hawaiian household had a group of houses house to keep his implements and store his crops until instead of a single house as it is today. A group of such houses needed" (1958:11). was called a kauhale. * Hale Kuku. 'Tapa makers had a thatched shed, It used to be customary, in inviting people to come to called a hale kuku, where they pounded the inner bark of one's dwelling, to say, "E ho'i kakou i kauhale," or "Let mamaki (Pipturus spp.) or wauke (Broussonetia us go to [our] kauhale." A person accustomed to go from papyrifera) into tapa cloth" (1958:12). house to house is said to be: "ma'a i ha hele i kauhale." * Hale Kahumu. 'There was also the hale kahumu This word, kauhale, was used for a dwelling place until (kahu-umu), a thatched shed where cooking was done in recent times when it changed to ka hale, the house, to fit bad weather and cooking materials were stored. The men the modem residence (1958:7). had one, and the women had theirs, until the kapu on Handy and Pukui devote several pages of their text to a eating was abolished [1819]. In good weather, cooking discussion of the various kinds of functionally-differentiated was done in outdoor imu, one for the men and one for the structures that constituted a kauhale. These included the women; but in rainy weather, or if there was not a quantity following: of cooking to do, the hale kahumu was the place to go. * Mua. 'The Mua or men's eating house was a sacred Some of these were small but many were large enough to place from which women were excluded. It was the place include the storage of the utensils and implements" where the men and older boys ate their meals and where (1958:12). the head of the family offered the daily offerings of 'awa * Kamala. "A temporary house. . . was called a to the family 'aumakua. Here men and family gods ate kamala. It was tent shaped. . . the rafters came right to the together, and that was why women, who were periodically ground" (1958:13). unclean, were not allowed to enter here" (1958:9). Handy and Pukui's account of the kauhale, of which * Hale 'Aina. 'The women had their own eating we have only quoted the essentials, is by far the most detailed house, the hale 'dma. Here the women, girls and small available. Yet it is also evident that Handy and Pukui were boys ate together ....... There were prayers in the women's describing a system that neither had observed or experienced eating house for the family 'aumnakua to bless their food firsthand; as they say, the term hauhale had long since given and to come and partake, but the presenting of offerings way to ha hale, the single Western-influenced structure, "to fit belong to the men in the mua (1958:9). the modem residence." Their account of the traditional kauhale Household Archaeology in KTpapa 49 was an amalgam of oral traditions passed down to Pukui by determined with accuracy. The idea of household archaeology, her kupuna, and of earlier written accounts researched in the however, is more than simply "digging in houses;" it is an Bishop Museum Library and archives by Handy. While an investigation of the space of the family, the fundamental unit of "idealized" kauhale might include all of the various function- Hawaiian daily life. It involves far more extensive excavations ally-specific house types listed above, there is no reason to of each structure than would normally be conducted because believe that all kauhale were, indeed, ideal. Just as in our own the goal is to understand the full range of activities that took modem society there is a much variation in individual family place in the household and their distribution within that space. living arrangements, it is likely to have been so in pre-contact The first step in constructing a research plan for Hawai'i. household archaeology in Klpapa-Nakaohu was to examine our Hence we come back to the potential contribution of survey data for potential evidence of household organization. archaeology for writing a fuller history of the Hawaiian Drawing upon the ethnohistoric and ethnographic sources household-the kauhale-so central to the daily life of referred to above, we began with the assumption that the pre- maka'ainana and ali'i alike. There are so many questions contact inhabitants of Kahikinui had lived in kauhale or concerning the kauhale for which-despite the valuable "household clusters," which would be made up of more than a records of early explorers, Native Hawaiian writers, and single structure. When we studied our maps of site distribution twentieth-century scholars-we still have no answers. For it became clear that in some places, particularly along the example, how did the kapu system and especially the 'ai kapu crests of ridges, structures were grouped together in sets of prohibition on cooking and eating together by men and women threes and fours. Our working hypothesis was that such groups actually operate, especially among maka'linana rather than of closely-associated structures had once comprised individual ali'i? To what extent were the households of ali'i and po'e kauhale. ko'iko'i distinguished from those of maka'ainana? There is The next stage was to select several of these clusters also the question of change over time, for Hawaiian culture and for initial testing. Since we had already located more than society were not static. While all Polynesian societies have 1,000 structures in the project area, we choose the first places some form of tapu system, in Hawai'i this was more elaborated to be tested on the basis of variation in structural form, and for in connection with the development of a ranked stratification of variation in geographic location (coastal and upland). Our first the chiefly classes (Kirch 1984). When did this elaborated kapu tests were conducted during the summer of 1995, and consisted system come into effect, and when did it begin to influence the of placing six test pits, none larger than 1 m square, in six physical layout of kauhale clusters? Archaeology may never be different structures (Kirch and Van Gilder 1996:49-50, table 1). able to answer all of these questions, or to answer them to our The structural forms that we tested included two linear shelters, full satisfaction, but it does have the potential to expand our two rectangular enclosures, an L-shaped shelter, and a C- knowledge of the daily household lives of ka po'e kahiko. shaped shelter. In all sites there was some indication of a cultural deposit (usually under some depth of post-occupation Methods windblown sediment), but the cultural content within these deposits varied considerably. Some sites had a great deal of As mentioned earlier, any settlement pattern study charcoal, while others had none. Similar variation was evident needs to be augmented by a program of subsurface excava- in animal bone and shellfish remains, in kukui (Aleurites tions, both to determine the age (chronology) of the various moluccana) nut fragments, in basalt flakes, and in coral. For structures, and to determine their previous functions. After all, example, an upland rectangular enclosure (Site 44) had 60.7 g how can one be sure of patterns of use of any specific structure of charcoal, 8 pieces of animal bone, and no shellfish remains without investigating its interior? In the Kipapa-Nakaohu in a small test pit, whereas a coastal rectangular enclosure (Site survey area, we have now recorded more than 1,300 individual 335) had no charcoal and only 1 fragment of fish bone, but stone structures (see Chapter 2, this volume). The functions of 6.3 g of marine shells. Thus here are two structurally-similar some of these may be evident because of their large size or sites (both rectangular enclosures) with very different subsur- unique form, as in the case of certain heiau. But the large face contents. Their original functions were probably quite majority of sites consist of simple, stacked-stone windbreak different, and only excavation would make this evident. Based walls, terraces, or enclosures of various sorts, all of which on these initial test results, we were able to plan for more could have been used in a variety of ways, ranging from mua, extensive excavations, and a true application of the methods of to hale noa, to hale pe'a, to hale kahumu, and so forth. Only household archaeology, during our 1996 field season. through careful excavation and study of the interior and Since we were applying a household archaeology exterior spaces of these structures can their former functions be approach it was necessary that we choose sites likely to have 50 NM a Kahko oKahikinul been associated together in a meaningful way for a family in Preliminary Results the past. For our 1996 work, we selected three clusters of structures (including those which had been tested in 1995), and In the following pages, we now present some of the planned to excavate significant portions of every structure preliminary results of our studies at three household clusters or making up each cluster. We began by clearing the structures of kauhale in Kipapa and Nakaohu Ahupua'a. Because the lantana and other vegetation, taking extensive photos, and detailed studies of fish bone, shell, charcoal, basalt flakes, and making detailed maps at 1:50 scale. A grid of 1 x 1 m squares other materials from these sites are still in progress, our was then laid across each site for control during excavations. statements and conclusions should be considered tentative. We Excavation was done carefully with hand trowels and brushes, have given names to each cluster according to their relative with the materials in each grid unit kept together and their topographic positions: on the coast (Nakaohu Kai Cluster), in depths below ground surface and datum carefully recorded. the mid-elevation zone (Kipapa Waena Cluster), or in the All of the earth was sieved through nested screens of 1/4- and higher uplands (Kipapa Uka Cluster). The locations of these 1/8th-inch mesh, and all cultural materials such as charcoal clusters are indicated on Figure 4.1. A summary of the pieces, fragments of marine shell, waterwomn pebbles, and so structures within each kauhale cluster is given in Table 4.1. forth were meticulously recorded and bagged by grid unit and depth. All of these materials have subsequently been cataloged Kipapa Waena Cluster and the information organized in a computerized database The Kipapa Waena cluster lies astride a prominent (Paradox for Windows 4.0), which will allow us to analyze this ridge of aa lava a short distance east of the mauka-makai jeep material more readily.3 Few formal artifacts were found road in Kipapa. From this ridge the terrain steeps down steeply despite the extent of our excavations, but to the archaeologist, on three sides, with a large deep swale immediately makai, the scraps of shell and charcoal and their distribution patterns which was probably an excellent location for gardening. A are actually more important than finding a finished adz or small "valley" to the west of the site is occupied today by a fishhook. Our strategy was to open up large portions of the grove of wiliwili trees. The cluster consists of four distinct floor of structure at a time, so that we could see how patterns stone structures (741, 742, 752, and 1011), all much closer to developed on the structure floor. each other than to any other sites in the vicinity. Site 742, the Since our goal was to research intra-site organiza- highest in elevation on the ridge flat, is a linear shelter or long tion-in order to understand what kind of activities took place windbreak wall. A low wall of single course stone slabs within a household cluster-we had to excavate large areas of comprises Site 741. Site 752 is a substantial stone-faced terrace every structure and keep careful records of soil color and located just below 742, on the ridge crest. Finally, Site 1011 is texture, groupings of cultural remains, and of features such as a small C-shaped shelter surrounded by a low wall and located ovens, hearths, or pits, anything that might give us a clue as to on a natural terrace below Site 752. what kinds of things people were doing in that structure. One The wall of Site 742 is of substantial construction, of the most common household features we encountered were including several large, fine pahoehoe slabs, and approximately "combustion features," places where fires had been ignited. We 10.5 meters in length (Figure 4.2). Located on the top of the use the general term combustion feature so as not to bias our ridge, it has the most sweeping view of the surrounding area, interpretation of the functions of such features. Some certainly and indeed from here one can see the Luala'ilua coastline, as functioned as imu or earth-ovens, while others are more formal well as mauka towards Pu'u Pane. After clearing, it was and stone-lined, and were probably kapuahi hearths. But other evident that the linear wall protected a substantial flat area in combustion features do not fall readily into either of these its lee, where we concentrated our excavations. This area Native Hawaiian categories, and may represent cultural contained numerous small combustion features of relatively practices that are not well documented. Whenever possible, the informal construction (Figure 4.3), typically concentrations of materials which were found within a combustion feature were small, fire-altered stones set in a shallow, basin-shaped kept separate from all others; this includes charcoal which will depression (thus not a true imu). They were probably small be identified to botanical species to see what specific woods hearths of some kind, most likely for warmth and light, but were being burned, and for radiocarbon dating. possibly also for cooking of some kind (e.g., roasting). Further 3 By mutual agreement with the Department of Hawaiian Home Lands and Ka 'Ohana 0 Kahikinui, after analysis all of these excavated materials and samples will be curated in the Maui office of the State Historic Preservation Division, pending their ultimate disposition, possibly at a cultural center to be constructed within Kahildnui itself. Household Archaeology In Kipapa 51 * SITE CLUSTER STE Ire~~~~~~~~~~~~~~~AA X /( SITE ~ ~ t't CLSE CLSE .-----NHOAPILI TRAIL ------ PIPELINE ROADS/HIGHWAYS ROCK WALL L - ------STREAM Figure 4.1 Locations of householdclustersdiscuss1 IL.E , O~~~~~~~~~ . 5 1 KILgOMETER 52 No Mea Kahiko o Kahikinul Table 4.1 K-papa-Nakaohu site excavations, 1996. Cluster and Site Type Dimensions Excavated Features Site Number (i) area (m) Klpapa Waena Site 742 Linear shelter 10.5 x 6 28 Multiple small informal hearths; Slab-lined niche Site 752 Stone-faced terrace 14.5 x 6 20 Two slab-lined hearths; internal pit; Cupboard; Exterior /muarea Site 1011 C-shaped shelter 4 x 3.5 8 Well-defined exterior imu Klipapa Uka Site 44 Rectangular enclosure 6 x 6 13 Slab-lined hearth; Lava tube cupboard Site 45 Linear shelter 17.5 x 6 13 Two slab-lined hearths Site 46 U-shaped enclosure 8 x 7.5 8 Partially stone-paved interior Site 48 L-shaped shelter 4.5 x 2 5 No features Nakaohu Kai Site 331 Linear shelter 8.5 x 5 4 One combustion feature Site 334 Rectangular enclosure 12.5 x 6 3 No features Site 335 L-shaped shelter 11.5 x 5.5 4 No features analysis of their contents will help us determine their probable of several meters to the terrace of Site 752. The north wall of use. Site 742 is also notable for the presence of numerous this structure is comprised of the natural outcrop, modified in flakes of fine-grained basalt, some bearing ground edges or places by a stacked stone wall. A small niche is present in this facets, and therefore indicating that they were derived from the rear wall, possibly having been used as a cupboard but nothing process of re-sharpening stone adzes. remained in its interior. The south terrace facing is substantial Of special interest was a small carefully-constructed and well constructed, almost 2 m high in places. The east side niche defined by stone slabs, located in the northeast corner of of this rectangular structure is partially built up into the slope the structure (Figure 4.4). Inside the niche were two small of the ridge, while the south and west sides have little or no waterworn stones. We believe that the position of this feature in visible rock walls. The natural terrace surface has been the northeast corner is quite significant, for the northeast is also substantially added-to and built-up with stone fill to create a the ritually prominent location in most Kahikinui heiau (see larger living space. This site had a significantly deeper deposit Chapter 2, this volume). Samuel Kamakau, in his discussion of of cultural materials than that at 742, including the traces of a ancient religious practice, wrote that "the heiau ipu-o-Lono possible pit in the center of its floor. constantly maintained by the populace was the hale mua, the Perhaps the most remarkable discovery at Site 752 men's eating house, which every household had" (1976:133). was that of two rectangular, slab-lined hearths set close to each Although Kamakau does not describe a stone-lined niche, he other in the eastern half of the main terrace (Figure 4.5). Each does discuss the sacred gourd or ipu kua'aha (ipu 'aumakua) hearth contained charcoal and food related refuse. The cultural containing "food" and "fish" ('ai and i'a, things of the land and deposit in the living floor of this terrace yielded a range of food of the sea), with an 'awa root tied to its cord handle. Kamakau remains and a few simple artifacts (such as coral files and a says that this sacred gourd was variously kept on a "hanging hammerstone). The most likely interpretation of Site 752 is as post, the wall, or the rack," or perhaps as in this case, in a the hale noa or main dwelling house, for this kauhale. The two special niche. The presence of this special niche, plus the hearths, sitting side by side in the house may represent indi- physical location of Site 742 at the highest part of the cluster, vidual women's and men's cooking facilities, as would be and the evidence for stone-tool working (a presumably male consistent with adherence to the 'ai kcapu. If so, however, we activity) are all evidence that this structure may have been the are faced with a seeming contradiction in terms of the mua for this kauhale. ethnohistoric accounts, for most sources suggest that these Heading south down the ridge, there is a steep drop cooking places were in separate structures (e.g., Malo Household Archaeology in Klipapa 53 Figure 4.2 View of Site 742. 1951:27). Here, perhaps, we have a glimpse into local domestic pieces. The excavation profiles showed the outlines of a series practices that varied from what has been described in the of intersecting pits, indicating repeated use of the area for such literature. Although written sources indicate that cooking by ovens. the two genders took place in different buildings, that may not The third structure which we excavated is Site 1011, have been the way that local mnaka'dianana in Kahikinui chose a small C-shaped enclosure. The interior of the C-shape lacked to interpret the kapu, opting instead perhaps for physically any features, but immediately to the west in a flat, terraced area separated kapuahi within a hale noa.4 As we shall see, a was a single, very large earth oven. The shape of the oven was similar set of dual hearths was also uncovered at Kipapa Uka well preserved, and had been outlined with large stones, and Cluster (see below), the interior was filled with layers of ash and charcoal. The site The smaller, western half of the Site 752 terrace was is south (downslope) of 752, on a broad natural terrace over- also excavated, and revealed an area that appeared to have been looking a deep swale. The outline of the terrace appears to have used for construction of large earth ovens or imu. A deep been augmented with a low wall of stone defining a sort of deposit of very dark brown and unusually red earth (from yard-like space around Site 1011. Since we have an oven area cokng hi us a e b 8eeforasting o~rbolng mtos 54 Ni Mea Kahiko o Kahikinul Figure 4.3 A small combustion feature exposed during the excavation of site 742. To sumn up, our current interpretation of the Kipapa Kipapa Uka Cluster Waena Cluster is that the highest structure on the ridge, Site The second cluster that we investigated was situated 742, functioned as the hale mua or men's house, while the 752 in the higher uplands (uka) and consists of four structures terrace supported the common dwelling house or hale noa, (Sites 44, 45, 46, and 48). These are aligned along a low along with a possible men's kahumu or oven house on the pahoehoe ridge just mauka of the first east-west jeep cross western part of the terr ace. Site 1 01 1, at the lowest point in the road. The ridge lies just east of a good-sized swale which cluster, was most likely to have been the women's kahumu. In probably was a cultivation area (see Chapter 2, Figure 2.5). virtually every Polynesian society, the vertical spatial axis is Site 46 is the highest in elevation with 48, 45, and 44 running extremely important as an expression of rank, gender, and down the ridge towards the south. These are a large U-shaped sacredness. Relative height is thus associated with maleness, structure, a small L-shaped structure, a linear shelter, and a seniority, and kapu, while relative lowness is associated with rectangular enclosure, respectively. femaleness, junior status, and noa. Thus in KMpapa Waena Site 46, sitting at the top of the ridge, is a U-shaped Kauhale, it would be culturally consistent for the mua to be at enclosure open to the south. Three low stone walls define a th o ftecmlx n h oe' okigfclt tte lvlae hthdbe at les partial pae ihcrfly bottom W atfas unxece bae on th vial rte eetedfa aatsas uxcvtosddntrvan sources howver wa h rsneo wnsoelndhats hatsorptadvr iteclua aeilwth wihi th prsue haeoprasrpeetn ihro ecpio fanme fvlai ls lks ie h undocumented~~ cutrlpatc fgne-ifrnitdhge lvto ihrgrst h eto h oshl secondary cooking spaces within the main dwelling house. cluster, and the special care taken with the basalt paving, it is~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Household Archaeology in KTpapa 55 possible that this structure was a mua. However, no altar or have some relation to the 'ai kapu. However, we will need to niche was found (as in Site 742), and this interpretation is determine whether the hearths date from the same period, and tentative. what type of food remains each contains, before we can assess Slightly lower in elevation and offset to the east is the likelihood of this interpretation. Site 48, the function of which remains a mystery. It is a small The lowest structure in this cluster, Site 44, is a four- L-shaped shelter, with virtually no cultural materials whatso- walled rectangular enclosure with walls about 1 m high, which ever, or any trace of internal features. Whatever activities took contained a great deal of cultural debris, such as scattered place there, they left no physical remains. The low wall defines charcoal and animal bones. There was such an unusually large a fairly small area so perhaps it was used for storage (e.g., a quantity of charcoal spread throughout the interior of the hale papa a ?); another possibility is that it was a space for structure that we have considered the possibility that the wood activities such as weaving which would not leave and grass superstructure (i.e. walls and roof) may have burnt. archaeologically visible traces. Site 44 also had a single, centrally located slab-lined hearth Moving down the ridge, Site 45 is a linear windbreak containing a carbonized tuber of 'uala (burnt sweet potato). shelter situated on a substantially built-up terrace. As at Site The ridge on which this cluster is located is a pahoehoe flow 752 in Klpapa Waena, Site 45 had two stone slab-lined hearths with a lava tube underneath. In the northeast corner of Site 44, (Figure 4.6) situated close to each other in the eastern part of it appears that the inhabitants deliberately broke through the the house floor Again, we are of the opinion that these features thin rock roof of the underground tube to create a small niche Fiue44Pheo-lblneihstaentenotes onro ie72 56 Ni Mea Kahiko o Kahikinul (or altar?) area within the house. Indeed, it appears that Site 44 in the area that undoubtedly were spaces used by its inhabit- was specifically built in this location so that this opening could ants. At the furthest point mauka overlooking the entire cluster be included in its northeast comer. is a beautifully-preserved ko'a, or fishing shrine. A small Our current interpretation of this complex is that rectangular enclosure with paving and internal space segmenta- Sites 44 and 45 seem to have been the everyday sleeping and tion, its careful construction and many offerings of branch cooking structures while 46 and 48, with virtually no cultural coral speak to the site's importance. materials, may have been more specialized structures. It is We excavated at three of the structures in this possible that the larger of the two, 46, functioned as the hale complex, Sites 331, 334, and 335. Interpreting the pre- mua, or men's house, of this household. European contact settlement patterns along the coast can be difficult, because it appears that these areas continued to be Nakaohu Kai Cluster occupied after contact, and possibly well into the nineteenth The third cluster which we investigated is in the century. This means that some structures were occupied both coastal sector of Nakaohu Ahupua'a. The cluster lies just before and after European contact, others only before, and inland from the imposing cliffs of Puhimake Bay, arranged others only after. In some cases, excavations alone cannot along two aa lava ridges overlooking a small swale or answer this question, it requires the definitive assessment of depression (see Figure 2.6). Due to time limitations, we were radiocarbon dating. Since no radiocarbon dates have been unable to test several small rock shelters and C-shaped shelters processed for these sites yet, the interpretations which follow " | v ~~~~igur .5 Viw o it75afeexvtin Household Archaeology in Kipapa 57 rM Figure 4.6 View of Site 45 after excavation. represent our best assessment based on the architecture and crumbled shell and coral, with some 'ilii'iii gravel evident. We nature of excavated remains at each of the sites. However, we also found a hamnmerstone and several large waterwomn stones suspect that Site 334 is an exclusively post-contact structure on the surface. We excavated a 1 x 3 m trench across the floor, while Sites 331 and 335 are almost certainly pre-contact, but revealing a thin cultural deposit reflecting only a single may or may not have continued to be used into the early occupation phase. This deposit contained pieces of ceramic nineteenth-century. dishes, trade beads, and most notably, a broken slate pencil of Site 334 is a large rectangular enclosure situated in the kind distributed by the Protestant missionaries during the the bottom of the low gully or swale that runs through the 1 820s.5 complex, and is protected from the wind by ridges on three There is no evidence of a pre-contact occupation in sides (Figure 4.7). We suspected that this site had been Site 334, and we think it likely that the household which had occupied in the post-contact period because of its large size occupied the other structures in the Nakaohu Kai Cluster built (12.5 x. 6 in), unusually high walls, and well-defined rectilinear Site 334 to serve as their primary residence in the early doorway at its south end. Such attributes are typically associ- nineteenth century. If we are correct in this interpretation, this ate wihnntet-etr aainhue seBihmi xeletacaooia vdnefrtekn fsitfo [1908] fo maypoorpi xmls.We ecerdte mli-tutr ahl oasnl ahl sdsrbdb structur oflnaa hlo a oee ihamxueo anyadPki(98,adrsligi atfo h ~ hssaepni ayb infcnfranioaercaglr nlsr nahg rs e ude eer ato ie34i lkl ohv ena earl nieenh-etr scolhue(eEhptr2hsvlm) 58 NA Mea Kahiko o Kahikinul Figure 4.7 View of the post-contact stone walled enclosure, Site 334. abolition of the kapu system. nothing. We returned to Site 331 in 1996, however, determined Site 335 is a linear shelter that had been segmented to figure out why there was such a disparity between the into two separate sections or "rooms." It is likely to have been richness above ground and the poverty below. In fact, we soon one of the main living areas for the Nakaohu Kai household in discovered that we had managed to position our 1995 test unit pre-contact times, possibly a hale noa. Like all of the sites at over virtually the only empty section of the structure's interior, the coast there was a fair distribution of crumbled shells and reinforcing our point that household archaeology requires more coral across the floor of the structure. The bend portion of a extensive excavations than mere test pits. bone fishhook was also recovered from this site. No hearth-like Our excavations at Site 331 in 1996 recovered a features were identified in this structure, but due to time significant number of basalt flakes, a small ashy hearth-like constraints our excavations at this site were more limited, and feature, and a greater quantity of pig bones than in any other it is possible that such hearths are present in the unexcavated site so far in our household excavations. Under the 'ai kapu areas. system, pig was traditionally a special food reserved for male Site 331 is a linear shelter located on the ridge crest consumption, and thus our interpretation of Site 331 is that it lokn onoe h eto hshueodcutr hnw fntoe stehl u o h aah a lse.Ti ha tete Sit 33_ihasnl-0x5 mecvto nti nepeato sspotdb h iespyia oiinah 19wefun vitull nohn.oee, wehdpeiul o ftecutr ls otefsigsrn.I ohteKpp note th rsneo aybsl riat uha rknWanadKppaUahueostetrcreht we hav piece of adze an flks ctee vrtesrae n hs tnaivl dniida en h aemawsash were peplexedhenou susrfc inetg tin yild d hihs siei t lseadi wocssteesrcue r Household Archaeology in KWpapa 59 also associated with basalt flaking debris. Thus far we have the close of the eighteenth, or very early nineteenth century. only these three households to compare, but as we continue to The other two charcoal samples yielded valid radiocarbon ages research the nature and patterning of family life in Kahikinui, of 340 ? 90 and 110 ? 50 years before present, when corrected we may find that within the household, sites of a more sacred for what is termed "isotopic fractionation" (the ratio of carbon- or specialized nature were deliberately placed at a higher 12 and carbon-13 isotopes). However, as can be seen in elevation than those used for more mundane tasks. Table 4.2, each radiocarbon age has two or more possible "real" calendar ages associated with it. This is because the Radiocarbon Dating creation of carbon-14 in the upper atmosphere varies over time, The cultural materials present in each of the and in recent centuries has fluctuated widely. Thus a given excavated sites just described provide some indication of radiocarbon age may intersect the curve of calibration at two or whether each is pre- or post-contact. For example, the high- more places. Each intercept has a different probability of being walled rectangular enclosure at Nakaohu Kai (Site 334) yielded correct, as also seen in the table. In the case of the sample from metal, glass, and ceramic objects that clearly date it to the first Site 331, we believe that the most likely true calendar age is half of the nineteenth century. None of the other excavated A.D. 1478-1648, which has a high statistical probability sites yielded such post-contact materials, suggesting that they (p = .98). In the case of Site 742, the most likely true calendar either dated prior to 1778-79 (the dates of Cook's fateful age is A.D. 1692-1729, because the other possible age ranges expedition), or at least before European contact had led to any can be ruled out on independent archaeological evidence. influx of foreign material culture (such as trade goods which Although we have only dated three samples, and began to appear in significant numbers in Hawai'i with the many more will be necessary to understand the chronology of advent of the Northwest Coast fur trade in the late 1780s and households in Kipapa-Nakaohu, these do begin to give us an 1790s; see Kirch and Sahlins 1992). In order to gain a more approximate idea of when these sites were occupied. In terms precise idea of when these sites were occupied, we submitted of the general sequence for Hawai'i outlined by Kirch (1985), three samples of charcoal (one from each cluster) for radiocar- these households would span the very late Expansion Period bon dating.6 The results of radiocarbon analysis are presented and into the Proto-Historic Period. Initial radiocarbon dates in Table 4.2. from upland sites tested by Boyd Dixon (see Chapter 3, this As can be seen in the table, one of the charcoal volume) also date to this same general period. samples (from Site 44) yielded a "modem" age. This does not mean that the site is actually "modem," for it contained Conclusions traditional Hawaiian cultural materials. Rather, this reflects the level of statistical error and uncertainty with radiocarbon ages The landscape at Kahikinui offers a unique opportu- that are very close to the present. Most likely, Site 44 dates to nity to learn about pre-contact Hawaiian history on many Table 4.2 Radiocarbon dates from Kipapa household clusters. Sample Data Measured 12CP3C Conventional Calibrated 14C Age Ratio Radiocarbon Age Age Ranges' Beta-101451 Modern -24.8 Modern Modern Site 45, Unit J9 (101.5 ? 0.7 W/oo) Beta-1 01 452 310 ? 90 -23.1 340 ? 90 A.D. 1454-1458 (p = .02) Site 331, Unit S20 A.D. 1478-1648 (p = .98) Beta-101453 80 ? 50 -23.1 110 ? 50 A.D. 1692-1729 (p = .26) She 752, Unit P4 A.D. 1812-1901 (p = .62) A.D. 1906-1922 (p = .12) 1 Calculated according to Method B of CALIB (Rev. 3.0.3), at one standard deviation. Values given in parentheses are probability distributions. 6 Once charcoal identification and analysis has been completed, we will be submitting a larger number of samples from these clusters, including at least one sample from each excavated site. The three dates reported here are thus only a preliminary indication of age. 60 NA Mea Kahiko o Kahikinul scales. While much research has been done on archipelago- our preliminary findings only. Once more detailed lab analyses wide trends of historical development, little is known about have been completed, we will be able to report in greater detail moku and ahupua'a internal organization in specific areas; still about the rhythms and patterns of daily life in ancient KTpapa. less is known about the activities and patterns of individual The distribution of certain artifact types may allow us to households. The archaeology of the Hawaiian family is reconstruct areas where particular activities regularly took relatively new, and there is much to be learned. place. For example, where did basalt tool making take place? The rich ethnohistoric record for Hawai'i has Inside a house or outside? Did it vary from household to provided us with a starting point from which to understand household? What was the function of the small flakes of household organization. Those records have limitations, though volcanic glass that are found in very specific locations within and tend to be regionally, temporally and class specific. At the household clusters? How did people dispose of refuse time they were written, concern was more for documenting the generated by daily living? Was pig, a food traditionally exemplary lives of the ali'i, than those of the thousands of associated with men, consumed only in the mua? Do mua maka'dinana. Already our excavations at only three household typically have ritual foci in their northeast comers, as at Site clusters have yielded interesting and regionally specific 742? How did men and women segregate their daily round of patterning. In particular, the presence of dual, formal hearths in activities, according to the kapu system? What kinds of both Site 752 (Kipapa Waena Cluster) and Site 45 (Kipapa Uka changes occurred in the layout, architecture, and organizational Cluster), both of which appear to have been hale noa dwell- patterns of kauhale during the several centuries that Kahikinui ings, provides an excellent example of a previously undocu- was inhabited before Western contact, and what changes mented cultural practice. occurred later? These and many other questions will continue The interpretations presented in this chapter represent to orient our research in the years to come.