16
Into the Netherworld of Island Earth: A Reevaluation of Refuge Caves in Ancient Hawaiian Society Joseph Kennedy and James E. Brady Archaeological Consultants of the Pacific, 59-624 Pupukea Road, Haleiwa, Hawaii 96712 The reexamination of a lava tube that had been modified for defensive/refuge purposes in Hawaii’s early historic period offers insights linking such sites to Polynesian cosmol- ogy, settlement patterns, and status use. Details of site modifications provide method- ological clues for advanced interpretation and analysis. 1997 John Wiley & Sons, Inc. INTRODUCTION This article reports the results of 5 weeks of field investigation conducted in August and September 1994 at Site 5060, located in Keopu 2nd ahupuaa just east of the town of Kailua-Kona on the island of Hawaii (Figure 1). The cave is one segment of a long lava tube network formed within basaltic flows originating from the western flank of Mount Hualalai, the most recent of the major Hawaiian volcanoes to become extinct. The last flows date to as late as A.D. 1801, but Site 5060 was clearly produced by an earlier eruption since radiocarbon dating of the modified portions of the tube’s interior indicates human utilization since roughly A.D. 1400. Lava tubes in general are associated with smooth pahoehoe flows and are formed when the surface of the flow cools faster than the center. The internal core is sometimes released through pressure leaving the outside as a hollow cylinder. Through this process long tubes are often produced which can ac- commodate a range of human activities (Carlquist, 1980:21). Archaeologists have been slow to develop a sophisticated model of the complex and multifac- eted cultural articulation between Hawaiian society and the subterranean darkness of its insular world. The fundamental problem is that the popular European notion of ‘‘cavemen’’ has had such a tenacious hold on archaeologi- cal thinking that a habitational function has become the ‘‘default’’ interpreta- tion whenever anthropogenic materials are discovered in caves. The problem is often exacerbated by those working in Cultural Resource Management (CRM), who routinely check the habitation block on survey forms whenever midden is present in cave assemblages. Since cave function is thought to be understood, there has been little incen- tive to employ even the most basic cave-specific research methods which Geoarchaeology: An International Journal, Vol. 12, No. 6, 641–655 (1997) 1997 by John Wiley & Sons, Inc. CCC 0883-6353/97/060641-15

Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

Embed Size (px)

Citation preview

Page 1: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

Into the Netherworld of Island Earth:A Reevaluation of Refuge Caves inAncient Hawaiian Society

Joseph Kennedy and James E. BradyArchaeological Consultants of the Pacific, 59-624 Pupukea Road,Haleiwa, Hawaii 96712

The reexamination of a lava tube that had been modified for defensive/refuge purposesin Hawai'i’s early historic period offers insights linking such sites to Polynesian cosmol-ogy, settlement patterns, and status use. Details of site modifications provide method-ological clues for advanced interpretation and analysis. 1997 John Wiley & Sons, Inc.

INTRODUCTION

This article reports the results of 5 weeks of field investigation conductedin August and September 1994 at Site 5060, located in Keopu 2nd ahupua'ajust east of the town of Kailua-Kona on the island of Hawai'i (Figure 1). Thecave is one segment of a long lava tube network formed within basaltic flowsoriginating from the western flank of Mount Hualalai, the most recent of themajor Hawaiian volcanoes to become extinct. The last flows date to as late asA.D. 1801, but Site 5060 was clearly produced by an earlier eruption sinceradiocarbon dating of the modified portions of the tube’s interior indicateshuman utilization since roughly A.D. 1400.

Lava tubes in general are associated with smooth pahoehoe flows and areformed when the surface of the flow cools faster than the center. The internalcore is sometimes released through pressure leaving the outside as a hollowcylinder. Through this process long tubes are often produced which can ac-commodate a range of human activities (Carlquist, 1980:21). Archaeologistshave been slow to develop a sophisticated model of the complex and multifac-eted cultural articulation between Hawaiian society and the subterraneandarkness of its insular world. The fundamental problem is that the popularEuropean notion of ‘‘cavemen’’ has had such a tenacious hold on archaeologi-cal thinking that a habitational function has become the ‘‘default’’ interpreta-tion whenever anthropogenic materials are discovered in caves. The problemis often exacerbated by those working in Cultural Resource Management(CRM), who routinely check the habitation block on survey forms whenevermidden is present in cave assemblages.

Since cave function is thought to be understood, there has been little incen-tive to employ even the most basic cave-specific research methods which

Geoarchaeology: An International Journal, Vol. 12, No. 6, 641–655 (1997) 1997 by John Wiley & Sons, Inc. CCC 0883-6353/97/060641-15

Page 2: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

Figure 1. Map of the Island of Hawai'i showing the location of refuge cave sites and importantbattles.

might call those assumptions into question. For instance, archaeologists havenot described areas within caves in terms of their being in the light, twilight,or dark zone. This is no small matter. It is axiomatic in archaeology thatpeople do not live in the dark zone of caves. In prehistoric society cross-cultur-

642 VOL. 12, NO. 6

Page 3: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

ally it is rockshelters rather than caves that are used for habitation. Whenthe cave is utilized in this fashion, habitation is restricted to the mouths ofcaves where there is light and fresh air (Burkitt, 1956:7; Chard, 1975:171).Where access to the cave is difficult or the entrance restricted, the likelihoodthat it was utilized for habitation diminishes.

The dark zone of caves, by contrast, is most often reserved cross-culturallyas a place for religious ritual (Hole and Heizer, 1956:171), and there is acorpus of beliefs and uses associated with aspects of Hawaiian cosmology(Beckwith, 1970). For example, menehunes, the mythical forbearers of mod-ern Hawaiians, are said to have resided in caves (Beckwith, 1970:327), andthe lava tube of Hina-lau-ae on the island of Molokai was referred to as thekumu or root of the island and is still approached with great reverence (Beck-with, 1970:219). The story of Kalama-ula’s cave bears a remarkable resem-blance to the one visited by Ala ad-Din (Aladdin) in The Thousand and OneNight’s Entertainment: In both instances, the cave’s entrance magically opensand closes and great wealth is contained within (Beckwith, 1970:244, 339).

Lava tubes were also the loci of both primary and secondary burial in oldHawai'i. The dark recesses of tubes on all the islands were favored hidingplaces for human bones which contained mana, or supernatural power. Laby-rinthine lava tube systems often foiled arduous and sometimes frighteningjourneys by grave robbers in search of osteological material which could bemanipulated with magic. In addition to an ossuary function, caves may alsohave been places for the deposition of dental material. Our investigation ofcave Site 5060, which will be discussed in more detail below, recovered 11human deciduous teeth, most of which were definitely shed naturally. Itseems likely that the burial of the teeth in a cave is related to the near uni-versal beliefs about the special meaning and handling of deciduous teeth(Frazer, 1922:43–45).

CAVES, REFUGE, AND WARFARE

Like many places in the world, organized conflict was a regular occurrencein almost all Polynesian societies. In Hawai'i, both inter- and intraislandstruggles were reported to be frequent (Kamakau, 1992:230) and were proba-bly a fact of life from, at least, the 12th century A.D. until the consolidationof the island group in the early 1800s (Earle, 1978:192; Kalakaua, 1972:20;Kirch, 1985:7, 306).

Throughout the Polynesian Pacific, refuge was a concept linked with war-fare. Caves as places of refuge are noted in Easter Island (Metraux,1940:368), Mangareva (Laval, 1938:144), Tonga (Collocott, 1919:234–238),Reef Island (O’Farrall, 1904:232), and Samoa (Nelson, 1925:140). Hawaiianswere certainly no strangers to this practice and even had a special term,pe'epao, which meant ‘‘hiding in a cave.’’ Given the relationship betweencaves and warfare not only in precontact Hawai'i but all of Polynesia, a reex-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 643

Page 4: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

amination of refuge caves is in order and will demonstrate the importance ofintegrating aspects of cave utilization into current models of Hawaiian so-cial process.

HISTORY AND METHODOLOGY

Site 5060 was first mentioned in an archaeological inventory report byChing (1980). The most substantial study on the cave was carried outby Schilt (1984) during work conducted between 1980 and 1982 for the Kua-kini Highway Realignment Project. Schilt recorded the presence of two mas-sive barriers blocking entrances and recognized that a portion of the cave‘‘. . . exhibits unique structural modifications which transformed a naturalsubterranean passage into an effective and defendable fort in late prehistorictimes.’’ (1984:116). Unfortunately, the downslope portion of the cave, to whichSchilt ascribed a habitational function, was destroyed during highway con-struction and so was not available for restudy. The upslope refuge area wasalso briefly examined by Dunn and Rosendahl (1991), who concurred withSchilt’s assessment. Finally, data recovery operations encompassing both sur-face sites and the cave were carried out early in 1994 by Archaeological Con-sultants of the Pacific. These served as the basis for our intensive investiga-tion of the cave.

The study was focused on several chambers immediately unslope anddownslope of a large defensive barrier which contained floors covered with acambic A horizon with the consistency of powder. Schilt reported encoun-tering a greater compacted B horizon during previous excavations. Soil condi-tions resembled those encountered in another cave (Brady, 1995) in which ause-floor was discovered beneath a similar layer of loosely compacted powder.Caves tend to be low traffic areas so that soil compaction is often minimalalong use-surfaces, and vestigial remains of these surfaces can easily be de-stroyed using traditional archaeological field techniques. The layer of compac-tion in Site 5060 was completely exposed in two chambers by removing theupper 5–8 cm of powder with soft paint brushes. A use-surface was identifiedand confirmed by the presence of over a dozen hearths at this level and byartifact and cave formations lying on this surface.

In both chambers, floor space was divided by a number of platforms andwalls so that units of the surface sweep were laid out using culturally definedboundaries rather than arbitrary units. The quantity of soil removed fromeach unit was carefully noted so that artifact and midden densities could becalculated. In the final analysis, the surface sweep provided large quantitiesof data and a broad horizontal exposure in which to observe archaeologicalfeatures and contextual relationships. Test excavations were also carried outto observe depositional changes over time.

SITE DESCRIPTION

The focus of the upslope refuge area is a massive stone barrier (Feature 8),some 10 3 5 m which completely filled the circular natural entrance that had

644 VOL. 12, NO. 6

Page 5: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

been formed by the collapse of a roof section. Access to the cave was restrictedto two small openings, designated F and G by Schilt (1984:100), which canonly be negotiated by crawling (Figure 2). Entrance G had been filled withrubble at the time of our investigation, but the configuration of interior con-struction indicates that this opening had once been used. It is assumed thatthese portals were designed to allow only one person at a time to enter thedefended chamber and dictated that one must do so on hands and knees. Atthis measured rate of invasion and in this compromising position, the defend-ers could easily fend off a host of hostile intruders.

A low crawlway along the northern wall of the cave allowed passage be-tween chambers located to the east and west of the barrier. Schilt (1984:102)notes that the barrier is better constructed than the even more massive bar-rier at the now-destroyed downslope entrance to the cave. The stones in thecrawlway and the facing of the barrier in the upslope chamber are smooth-surfaced and closely set which resemble in quality the stonework of two verylarge burial crypts located on the surface, several hundred meters upslope ofthe cave entrance. The quality of the stonework in comparison to constructionin other parts of the cave suggests that ritual specialists in construction, kuh-ikuhipu'uone, may have been employed in the creation of both the barrierand the crypts.

The cave originally contained at least seven entrances so that it was possi-ble to gain access to tunnels both upslope and downslope of the refuge barrier.Although these openings were sealed with rubble, at least three defensivewalls were constructed across each tunnel between the main chambers andthe blocked entrances. The purpose of multiple walls appears to be simpledefensive redundance. In addition, large quantities of rubble were depositedin an area where a small passage connects the two tunnels leading off thewestern end of the main chamber (Figure 3). The rubble obstructs the pas-sage, restricts movement between tunnels, and requires transversing onhands and knees. The lack of breakdown scars on the ceiling clearly indicatesthat the rubble was not the product of ceiling collapse but rather stone thathad been purposely brought to this area for the construction.

With the construction of the barrier, all activity within the cave occurredin the dark zone. On either side of the barrier there are soil-covered chamberswhich measure slightly over 30 m in total length, between 3 and 5 m in widthand almost 1 m deep near the barrier. The larger, unslope chamber is dividedby a number of low stone wall enclosures and stone platforms. In form andlayout the chamber closely resembles refuge caves in Samoa (Freeman, 1943,1944; Green, 1969) suggesting that, morphologically, Site 5060 may be partof a larger Polynesian pattern. The soil-covered area contains high concentra-tions of midden including: fish, mammal and bird bone, echinoderm, marineshell, kukui nut, and volcanic glass. Little or no midden or artifactual mate-rial was encountered beyond the limits of the soil, suggesting that almost allactivity was concentrated in these areas. The nature of the midden is consis-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 645

Page 6: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

Fig

ure

2.P

lan

view

ofth

eu

pslo

pere

fuge

area

ofS

ite

5060

.

646 VOL. 12, NO. 6

Page 7: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

Fig

ure

3.P

lan

view

ofth

eu

pslo

pere

fuge

area

wit

hth

eru

bble

pave

men

tsh

own

atth

ew

este

rnen

dof

the

cham

ber.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 647

Page 8: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

tent with short-term occupation associated with a refuge function. This pointis stressed because archaeologists generally interpret such midden as evi-dence of habitation.

Almost no sediment is encountered either upslope or downslope of thesechambers which raises questions about the origin of the soils. The finegrained texture of the surface soil suggests that these may be water bornsediments and Schilt confirms that a good deal of water moves through thecave after rainstorms. An upslope extension of the same tube system doescontain a deep, uniformly fine grained deposit at the back of the cave whichappears to have been water deposited. However, if large amounts of waterborn sediments were entering Site 5060, one would expect to find sizable sedi-ment accumulations against the upslope side of existing walls. This is not thecase. No more than 10 cm of sediment are found upslope of any of the wallsand those deposits do not cover an extensive area. Therefore, large amountsof water born sediments do not appear to enter the cave. Test excavationswithin the soil-chambers reveal that these are anything but uniform, fine-grained deposits. All test units (Figure 4: TU. 7, 10, 12, 15–16) encountereda core of large rocks mixed with fine soils and smaller, fist-sized stones. Therock was too large to have been transported by water, and most are in areaswhere ceiling collapse can be ruled out by checking for collapse scars on theroof. This appears to rule out any possibility that these accumulations arenatural.

In several cases, 0.5 m rock and earth was found deposited on top of anearlier, hard-packed use-floor which directly overlay bedrock. The accumula-tion, therefore, must have occurred during the 4 centuries of human utiliza-tion of the cave. No internal divisions were found between the upper andlower use-floors, suggesting that the 0.5 m of fill was deposited in a singleconstruction effort. Based on stratigraphic relationships, it appears that thefloor was constructed during the same period that the massive fortificationswere engineered since the refuge barrier itself sits on this rubble core.

There are several possible reasons for the basal construction. Excavationdemonstrated that the tunnel floor slopes sharply so that the upslope side ofa 1 m pit was as much as 15 cm higher than the downslope side. This issignificant because the ground surface of the same pit showed only a 3 cmdifference so the fill may have been designed to level the floor. When waterenters the cave during storms, it quickly percolates through the fill and doesnot form surface puddles. Thus, the construction of the basal platform andraised enclosures afforded the occupants at least moderate comfort even dur-ing wet periods.

DISCUSSION

The use of caves or lava tubes as places of refuge in troubled times wasbriefly discussed earlier in relation to warfare. Since this practice was appar-

648 VOL. 12, NO. 6

Page 9: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

Fig

ure

4.P

rofi

levi

ewof

Sit

e50

60sh

owin

gth

ere

fuge

barr

ier

and

defe

nsi

vew

alls

.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 649

Page 10: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

ently quite common throughout Polynesia, the utilization of Site 5060 is cer-tainly explicable within this larger pattern. What this model fails to explain,however, is the degree of elaboration documented for this site. In the discus-sion of her large, multiyear salvage project, Schilt (1984:116) calls Site 5060‘‘the singularly most significant site examined during recovery phase excava-tions . . . .’’ Recent investigations have shown that the cave modifications arefar more extensive than even she realized so that the cave should be recog-nized as a ‘‘built environment.’’ In terms of labor expenditure, the simplevolume of architectural modifications, including barriers, walls, pavements,and platforms surpass the combined total of all contemporary surface archi-tecture in the immediate area. This point needs to be stressed because suchlavish expenditures should alert the archaeologist to the fact that the focusof such attention is somehow central to the concerns of the time. Site 5060,therefore, must be examined within the framework of its own culturally sig-nificant historic and structural context.

As mentioned above, radiometric dating of Site 5060 strongly suggests thatthe cave was a loci of human activity since A.D. 1400. However, a closerexamination of the dates indicate that there were two distinct periods of utili-zation, one centering around A.D. 1550 and a second during the transitionyears between the 18th and 19th centuries. Test excavations of the barrieritself produced datable material which indicate that these modifications mostlikely occurred during the second phase of utilization, or sometime in the late1700s or early 1800s. Examination of this period in Hawaiian history, whencoupled with information from our reexamination of the refuge cave, bringsthe role of refuge and status into a somewhat better focus.

After the arrival and subsequent murder of Captain James Cook in 1778,Hawai'i entered an era of dramatic change, not only because of contact withthe rest of the world but because of domestic developments as well. Aftercenturies of political intrigue between chiefs, the Hawaiian Islands at thetime of Cook’s visit were divided into four kingdoms: the island of Hawai'iunder the rule of Kalaniopu'u; Maui (except for an eastern section) was thekingdom of Kahekili; O'ahu belonged to Peleioholani; and Kaua'i and Ni'ihauto Kaneoneo (Kuykendall, 1938:30). In the 40 years after Cook’s death, noth-ing in the islands was more significant than the bloody efforts to consolidatethe archipelago under a single paramount chief, and the man who eventuallyaccomplished this goal was Kamehameha I. For the purposes of this article,it is this man’s early rise to power on the island of Hawai'i that may directlyrelate to the construction of Site 5060.

It is well known that after the death of Kalaniopu'u in 1782, Kamehamehaembarked on a course of regional consolidation that resulted in many highprofile battles between men of great rank on Hawai'i. Among others, Kameh-ameha battled Kalaniopu'u’s son Kiwalao as well as his cousin, Keoua, incelebrated conflicts at Keomo and Mokuohai (Figure 1). It is noteworthy thatthese campaigns took place but 12 miles from Site 5060. By this time, Ka-

650 VOL. 12, NO. 6

Page 11: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

mehameha not only had invaded Maui, but had opened several additionalfronts against competing home island chiefs. Because Kona (where Site 5060is located) was land held by Kamehameha and a much sought after part ofthe island, the district could reasonably except immanent attack at any timeby land or sea. In fact, it may be stated that by the year 1790 the entire KonaDistrict was braced for war. It was in this historical setting that the internalmodifications were added to Site 5060. While the radiometric evidence sug-gests the cave had been in use for the previous 300 years, its conversioninto a massive fortress came in response to a heretofore unprecedented levelof conflict.

WARFARE AND SETTLEMENT

The extensive modifications made during the construction of the basal plat-form have disturbed, and in some cases destroyed, evidence of the earlierperiod of cave utilization. What evidence is available suggests a far less inten-sive utilization which was concentrated near the entrance. This cannot ruleout an earlier refuge function since the later construction would have de-stroyed and reutilized stone from earlier walls. Interior cultural deposits mayeven be associated with a nonfortified refuge function since it is quite possibleto hide in the dark recesses of a cave without expending the time and effortto construct impressive barriers or transport soils to create earthen occupa-tion surfaces.

The term pe'epao does not imply any fortifications, and it has already beenshown that cave refuge was traditional throughout Polynesia. If one can en-sure safety in the darkness, then why elaborate defensive fortification to thepoint where the labor expenditure outweighs even that devoted to habita-tional construction? We have already suggested that conflict had reached un-precedented levels and appears to have propelled Hawaiians into a funda-mentally new response.

Excavations on the surface near Site 5060 have found evidence of middenpredating the defensive constructions, indicating the presence of human ac-tivity in the area but without formal architecture or house platforms. Manyrefuge caves in Hawai'i are located away from settlements, which may sug-gest that part of the security lies in the fact that the location is simply un-known to the attackers. During the earlier period, Site 5060’s location wellaway from the coast may have placed it in this category, but as conflict inten-sified, it may not have always been possible to quickly and easily reach theserelatively isolated caves.

There well may have been a settlement pattern response to this situationduring the early years of Kamehameha’s wars of consolidation. A survey ofthe surface in the immediate vicinity of Site 5060 recorded three residentialplatforms located in a tight triangle around entrances F and G of the cave.Radiocarbon dates place the construction of the platforms at the same timeas the creation of the refuge barrier. The massive labor expenditure on the

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 651

Page 12: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

defensive construction may well have been the nucleus around which a newformal settlement sprung up. Ceiling ‘‘drip formations’’ (what would beknown as stalactites in a limestone solution cave) were recovered archaeologi-cally from the basal/foundation levels of surrounding surface features andmost probably came from Site 5060. These ceiling ‘‘drippings’’ do not breaknaturally, so the presence of broken formations within the deposit indicate aclear connection between the cave and surface settlement. The refuge func-tion of the cave, therefore, is clearly linked to issues of demographics andsettlement patterns.

Site 5060 is not the only example of a refuge cave in the area. Just upslopeof Site 5060 another segment (Site 18665) of the same lava tube network wasconverted into a far less elaborate refuge. Downslope portions of the 5060tube network have massive fortifications mentioned by Schilt (1984) and sev-eral simple refuge caves in the area were recorded during her project. Dunnand Rosendahl (1991) documented three other settlements in the immediatearea which appear to have been located with a direct relationship to caves.One of these is identified as possibly a chiefly compound which contains sev-eral caves within an enclosing wall. Thus, the developments in 5060 shouldbe seen as only a special example of a larger phenomenon.

Schilt (1984:278) tends to explain settlement location in the region in termsof ecological/subsistence considerations—harvesting and tending require-ments of certain crops. We know cross-culturally, however, that warfare canbe a strong determinant of settlement location. There are innumerable exam-ples of groups moving to hilltops, cliff dwellings, and other defensible loca-tions that took them significant distances from areas dictated strictly by eco-logical considerations. In the face of escalating violence the presence of asuitable cave may have become a factor of primary importance in site selec-tion. In the case of 5060 it does appear that settlement actually shifted totake advantage of this feature.

Multiple examinations of Site 5060 and several other refuge caves in thearea have failed to produce evidence that any of the defensive barriers hadbeen breached. It is possible, therefore, that Hawaiian offensive militarystrategy may never have developed a successful countermeasure to the refugecave as a defensive form.

REFUGE AND STATUS

The restudy of Site 5060 has documented the fact that much of the cave isa built environment. The degree of modification stands in sharp contrast withthe simple but serviceable defensive walls protecting the nearby cave, Site18665. What accounts for such dramatic differences? The answer is that the‘‘built’’ refuge cave interiors were most likely prepared for, and limited to,chiefs and their entourages. Simply removing oneself to the dark recesses ofa cave does not ensure the type of protection that would be demanded by the

652 VOL. 12, NO. 6

Page 13: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

ranking members of highly stratified Hawaiian society. This was especiallytrue in the late 18th century, where the chiefly system was elaborated to theextent that high born personages were nothing short of powerful demigods,and where these demigods were mobilized for large scale war against eachother.

The evidence to support this hypothesis is compelling and demonstrated bythe reexamination of Site 5060. The area protected behind the stone barrierwas, by design, limited, suggesting that it could accommodate only a fixednumber of individuals and was not meant for the general population. Thiswas not dictated by the physical limitations of the cave. Both upslope anddownslope of the barrier there are suitable tunnels located beyond the seriesof defensive walls, which could accommodate large numbers of people. These,however, do not contain even the simplest defensive walls nor midden onthe floor to suggest that they were ever utilized. This strongly implies thatadmittance was determined by rank or status, and in Hawai'i both of theseelements were personified by the privileged chiefs or ali'i.

We have earlier referred to the quality of the stone work in the refugebarrier and how suggestive it is of the type of craftsmanship associated withthe kuhikuhipu'uone, or professional ritual architects of ancient Hawai'i.These individuals were closely associated with the chiefs, and their effortswere mostly reserved for projects which served the privileged class in a mostdirect way, for example, temples, fishponds, and elaborate house construc-tions. Comparison has already been made between the refuge barrier and thestriking burial monuments on the surface nearby. These twin memorials arealmost unique in Hawai'i and clearly mark the resting place of two very highstatus individuals. Cordy (1974) has argued that (in Hawai'i) the rank ofchiefs is related to the overall labor expenditures in housing construction rel-ative to that of other clusters. If this general measure is applied to the qualityof stonework and general effort that went into the ‘‘built environment’’ of Site5060, the conclusion that the labor expenditure was consistent with otherstructures used by ranked individuals is inescapable.

Finally, the ne plus ultra of portable artifacts relating to rank in Hawai'iwas recovered during excavation of Site 5060. This is the niho palaoa, orstylized fishhook, which was worn exclusively by ali'i as a demonstration oftheir chiefly station.

CONCLUSIONS

The present study has attempted to address a number of issues in Hawai-ian archaeology. First, our analysis has attempted to place caves within thegeneral native Polynesian and Hawaiian social contexts of refuge and war-fare. At the same time, the need for cave-specific research methods and thereading of cave context have been stressed. In particular, it has been demon-strated that the presence of anthropogenic midden is not, in and of itself,

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 653

Page 14: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

KENNEDY AND BRADY

evidence of cave habitation. Blockages at entrances or in passages and/orevidence of dark zone occupation should alert the archaeologist that refugemay be a more appropriate interpretation than habitation.

The implications of this are important since it is expected that the numberof recognized refuge caves would rise dramatically. A dramatic rise in thenumber of refuge caves would force archaeologists to rethink the importanceof both refuge and warfare in precontact Hawaiian society. Furthermore, thedating of refuge cave modifications when coupled with ethnographic, geneo-logical, and mythological events may provide new insights regarding the roleand extent of conflict in emerging societies.

The investigation of Site 5060 has documented a far more thorough-goingmodification of the cave than suggested by previous investigation. When laborexpenditure on cave modifications was compared to that expended on surfacearchitecture, it became clear that the site must have had a far greater impor-tance than archaeologists have previously been willing to ascribe to cave fea-tures. This is consistent with the point just made concerning the need toreevaluate the importance of refuge.

Political and historical details of Kamehameha’s early rise to power havebeen outlined to provide the specific social matrix on the Kona coast ofHawai'i, which led to the construction of the defensive features in Site 5060.The archaeological literature dealing with this area has been reviewed toshow that the transformation of Site 5060 was part of a larger process oc-curring in this area. It was suggested that caves played an important role inthe settlement patterns at that time.

Finally, the archaeological literature for the area immediately surroundingSite 5060 shows that a number of caves were modified for defensive purposes.The degree of modification varies widely from one cave to another. Whilerefuge may have been a concern of all strata of society, evidence for thechiefly status of the occupants of Site 5060 was then produced to explain thespecial elaboration documented for the site.

REFERENCES

Beckwith, M. (1970). Hawaiian Mythology. Honolulu: University of Hawaii Press.Brady, J.E. (1995). A Reassessment of the Chronology and Function of Gordon’s Cave #3, Copan,

Honduras. Ancient Mesoamerica 6, 29–38.Burkitt, M. (1956). The Old Stone Age. New York: New York University Press.Carlquist, S. (1980). Hawaii: A Natural History. Lawai, Kauai, Hawaii: Pacific Tropical Botanical

Garden Press.Chard, C. (1975). Man in Prehistory, 2nd ed. New York: McGraw-Hill.Ching, F.K.W. (1980). Archaeological Reconnaissance, TKM:7-5-03, Parcel 3. Archaeological Re-

search Center Hawaii, Inc., Lawai, Kauai.Collocott, E.E.V. (1919). A Tongan Theogany. Folk-Lore 30, 234–238.Cordy, R.H. (1974). Complex Rank Cultural Systems in the Hawaiian Islands: Suggested Expla-

nations for Their Origin. Archaeology and Physical Anthropology in Oceania 9, 89–109.Dunn, A., and Rosendahl, P. (1991). Archaeological Inventory Survey of TMK:7-5-03:3. Land of

Keopu 2nd, North Kona District, Island of Hawaii. PHRI, Inc., Hilo, Hawaii.

654 VOL. 12, NO. 6

Page 15: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society

REFUGE CAVES IN ANCIENT HAWAII

Earle, T.K. (1978). Economic and Social Organization of a Complex Chiefdom: The Halelea Dis-trict, Kauai, Hawaii. Anthropological Papers of the Museum of Anthropology, University ofMichigan No. 63, Ann Arbor, Michigan.

Frazer, J.G. (1922). The Golden Bough. New York: Macmillan.Freeman, J.D. (1943). The Seuao Cave. Journal of the Polynesian Society 52, 101–109.Freeman, J.D. (1944). Falemauga Cave. Journal of the Polynesian Society 53, 86–106.Green, R.C. (1969). A Lava Tube Refuge at Mulifanua. Reports of the Bernice P. Bishop Museum

18, 267–278.Hole, F., and Heizer, R. (1965). An Introduction to Prehistoric Archaeology. New York: Holt,

Rinehart and Winston.Kalakaua, D. (1972). The Legends and Myths of Hawaii. Rutland, Vermont and Tokyo: Charles

E. Tuttle.Kamakau, S.M. (1992). Ruling Chiefs of Hawaii. Honolulu: The Kamehameha Schools Press.Kirch, P.V. (1985). Feathered Gods and Fishooks. Honolulu: University of Hawaii Press.Kuykendall, R.S. (1938). The Hawaiian Kingdom. Honolulu: University of Hawaii Press.Laval, H. (1938). Mangareva. Braine-le-Comte, Belgium: Maisons des Peres des Sacres-Coeurs.Metraux, A. (1940). Ethnology of Easter Island. Bulletin 160. Honolulu: Bishop Museum.Nelson, O.F. (1925). Legends of Samoa. Journal of the Polynesian Society 34, 24–42.O’Farrall, W.C. (1904). Native Stories from Santa Cruz and Reef Islands. Journal of the Royal

Anthropological Institute of Great Britain and Ireland 34, 223–233.Schilt, R. (1984). Subsistence and Conflict in Kona, Hawai'i: An Archaeological Study of the Kuakini

Highway Realignment Corridor. Department Report Series, Report 84-1. Honolulu: BernicePauahi Bishop Museum, Department of Anthropology.

Received March 18, 1997Accepted for publication May 6, 1997

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 655

Page 16: Into the netherworld of island Earth: A reevaluation of refuge caves in ancient Hawaiian society