Let the remains ask the questions : In search for prehistoric relations on a Samoan settlement pattern through a correspondence analysis

(1)

LET THE REMAINS ASK THE

QUESTIONS

In search for prehistoric relations on a Samoan settlement

pattern through a correspondence analysis

Essay

Advanced Course in Archaeology At Gotland University 2006 Author: Joakim Wehlin Supervisors: Helene Martinsson-Wallin

(2)

ABSTRACT

My interest in ethno-archaeology and a Minor Field Study grant brought me to Samoa. There I had a chance to participate in a multinational project “The Pulemelei project, the

origins and Development of Monumental architecture”. On the former copra plantation

Letolo situated on the SE part of the Island of Savai’i, is an extensive pre-historic settlement with over 3000 remains situated. The remains are now heavily overgrown but when the vegetation was cleared an extensive survey was carried out (1977-78).

By using these records my aim is to understand more about how chiefdom society works in pre-historic times as well as in the present. The archaeological study is carried out with the aid of a correspondence analysis using the survey data as well as through a practical fieldwork (excavations) study of the settlement pattern at the Letolo plantation. The settlement shows large variations between the inland and the coastal region and those actions opened for whole new questions and ideas about the Samoan prehistory. During my journey I also got a better understanding for ethno-archaeology, and the problems that can meet us using these analogies and carrying out archaeological studies in a global setting and traditional society. To date, quite limited archaeology has been carried out on Samoa but the archaeological “revival” boosted by the project which started in this area in 2002 has created an opportunity to train the first generation of

Samoan archaeologists and give them a good platform to stand on! Keywords: Polynesia, Samoa, Letolo, Pulemelei, Settlement pattern,

(3)

For the love to my Dad and my only true Hero

Lars Wehlin

1956-11-06

2006-07-23

(4)

Preface

This study is financed by SIDA (Swedish International Development Cooperation Agency) as a Minor Field Study and is a part of the multinational project, The Pulemelei project, the

origins and Development of Monumental architecture. This project was initiated in 2002 by

Drs. Helene Martinsson-Wallin and Paul Wallin (The Kon-Tiki Museum) as a part of the larger multidisciplinary project; Identity, Matter, Movement and Place, concerning past and present migration in the Pacific. The Pulemelei project has eventually lead to a local interest for cultural heritage management and archaeology. As a result a teacher/student exchange (Linnaeus-Palme exchange sponsored by SIDA) was possible to be initiated between Gotland University and The National University of Samoa in 2005.

While I was starting to write my undergraduate thesis (2004) about a questioned Scandinavian Bronze Age remain I was directed towards a discussion about social organisation during this period of the Scandinavian prehistory. A general idea of the Bronze Age society is that it was organised as a chiefdom. The model for what they looked like is taken from the anthropological studies of Polynesian chiefdoms. This led me towards the study of ethno-archaeology and the problems that can meet us using these analogies and carrying out archaeological studies in a global setting became visible.

Trough globalisation we have become more aware of the world that exists outside our own resulting in a growing interest in other peoples and their culture. In archaeology this interest started with the use of ethnographic records in search of “our” past. A current growing interest for the cultural heritage focus on protecting and preserving “our” (humanity’s) cultural heritage for coming generations. Archaeology thereby play an active role in contemporary society. One problem is that the global protection and management (as for example concerning World heritages) aspires to be universally related but ultimately reflect western values. I suggest that it is together and in embracement with the local people that solutions to cultural heritage management should be sought (Skeates, 2000:84).

Thanks to my supervisor Dr. Helene Martinsson-Wallin, Gotland University and a scholarship from SIDA I got the opportunity to work, meet and interact with the local people of a village in Samoa. This gave me an opportunity to develop as a person and as a messenger of archaeology during a project to protect one of the most interesting prehistoric places in Polynesia.

(5)

1. INTRODUCTION

”Who controls the past controls the future: who controls the present controls the past…”…Is it your opinion, Winston, that the past has real existence?”

Orwell, 1965:253f

How much are the Polynesians really in need of archaeology? Samoa, is a “new” state of the Pacific, and as such it has an apparent need for “new” histories to build up their identity. Obvious sources for such “new” histories are the writings of anthropologists and archaeologists. I agree with Matthew Spriggs (1999:116) stating that; “What has already been

written and what is said in the future will be read by audiences we have only recently acknowledged.”

The thoughts about prehistory are probably not the same for an archaeologist as for the indigenous people of Samoa. It is therefore important that the interest in the Samoan society from archaeologist coming from the outside not become ethnocentric. The thoughts about the past from an archaeological point of view is often linear, but it is important to see how this past can become a part of contemporary society both on an economic and political scale. According to Johnson, (1999:107); “…Interpreting the past is always a political act.”.

My hope is that this study makes me more aware of the role of archaeological research in a global setting. I also hope that it benefits the local people in aiding an increased awareness of the importance to protect and preserve their past cultural remains as well as building up a sustainable cultural tourism. That will provide work opportunities for the local community. Focusing on the value of the Samoan cultural heritage, and a just started (2005) exchange project financed by the funding Linnaeus-Palme (SIDA) including Gotland University and the National University of Samoa, the hope is to contribute to the development of an awareness in cultural heritage and archaeology among the local people and the development of a third world country.

It is important to respect and obtain a cultural understanding of indigenous societies. These studies should not be done just for the western science, but a way to move forward and to encourage exchange and understanding. Essentially; It is theirs history but we can all learn something from it!

(7)

1.1 Aim and questions

My aim is to study Samoan settlement patterns at the Letolo plantation through a correspondence analysis, as well as, through archaeological practical fieldwork. By those actions I will be open for new questions and ideas about the Samoan prehistory and changes in the settlement pattern over time. An intention is also to get a notion about the condition of the settlement and the local relation to it. My hope is to get a better understanding for ethno-archaeology, and the problems that can meet us using these analogies and carrying out archaeological studies in a global setting.

1.1.1 Questions

• What did the prehistoric Samoan settlement pattern look like?

• Which archaeological observations/considerations on the settlement pattern can be traced?

• What is the present view about the Samoan Household units and high status clusters, (wards) in general, and the Letolo settlement Household units and wards in particular?

• Have the Letolo settlement been disturbed since the survey of Jackmond 1977-78?

• What relations do the current local people have to the plantation and its pre-historic remains?

I will also investigate if the correspondence analysis with its large quantity of observable records can generate some answers or new questions about changes in the Samoan settlement pattern over time?

How about the archaeological future in Samoa; is the people of Samoa interested in, or in need of, archaeology? If so, how should this been done? should it be; the Samoan way? 1.2 Method and Material

My study is based on a correspondence analysis. The Correspondence analysis considers the relationships (modalities) between individuals and the variables of the individuals in a specific field, region or area. That gives the possibility to handle the individual (here, the ancient find) in the same manner as the variables connected or unconnected to it. That gives a high-quality and understandable product in the visual graphic manufacture. A field creates approximately a system of relations between the positions taken by the individuals and the variables in the analyses. The result becomes new hypothesis about different relations in the material, showed through the analysis. The correspondence analysis deals with types of variables that are common in archaeological work and the analysis is a good theoretical

(8)

method to handle this large quantity of data. The correspondence analysis produce hypothesis on the basis of this large quantity of archaeological data and I agree with Wallin and Martinsson-Wallin, nd., that the use of correspondence analysis can provide new essential information to the archaeological debate. Important in every archaeological theory and method is the foundation in the material remains studied.

The information, in this study from the Letolo Survey made by G. U. Jackmond 1977-78, is tabulated and analysed with a correspondence analysis carried out in the computer programme WINBASP (available on the web through the University of Cologne: http://www.uni-koeln.de/~al001/, 18-05-2006). The results are also worked out and showed graphically by the computer program WINBASP.

To get a practical background to my correspondence analysis and a reality and first hand understanding about the settlement pattern and the Samoan society was a journey to the Samoan archipelago necessary. During this journey an excavation was carried out by us and local assistants at the Letolo plantation 13-29 of Mars 2006. Together with me and of great help and support, was the graduated archaeologist/osteologist Ilse Vuijsters. She was the field assistant to my supervisor Helene Martinsson-Wallin. The main aim with 2006 years excavation was to find, uncover and excavate a Household unit containing a large number of different features. This could give me a good practical knowledge about the settlement pattern studied. This practical admission in the social organisation pattern used in my correspondence analysis gives a better platform for the theoretical discussion.

1.3 Facts about Samoa

Western Polynesia constitutes of the Samoan Archipelago together with Tonga, Niue, Úvea and Futuna. The Samoan Archipelago consists of American Samoa (US territory) and the Independent state of Samoa (1962) previous called Western Samoa (NZ). The small country of Samoa is mainly composed of the two volcanic islands of Savaií and Úpolu comprising around 93% of the whole land area in the archipelago. Savaií and Úpolu are much larger than any of the other islands and together with a few other small islands they reach an area of almost 3000 sq. km and a population of approximately 180´000 (http://sv.wikipedia.org /wiki/Samoa 15-05-06).

The islands of Samoa are composed of lavas and pyroclastics from a series of basaltic shield volcanoes. Volcanism is most recent in the east where Ta´u (American Samoa) dates 100´000

(9)

BP and older to the west with Savai´i over 2.5 million years old (http://www.volcano.si.edu /world/volcano 15-05-06). The volcanic origin makes the landscape of Samoa very spectacular with over 200 volcanic craters, crater lakes, deep valleys, gorges and waterfalls. This complex geology has over time affected the settlement pattern. Half of the coastline is surrounded with coral reefs with shallow lagoons inside. The archipelagos close location to the equatorial line makes the climate tropical with abundant rainfall and humidity averages 80 percent yearly. The average monthly temperature lies between 20-30 degrees Celsius with small seasonal variations (http://www.govt.ws/gi_listing.cfm 15-05-06).

The islands are covered with thick tropical vegetation, except a quite recent lava flow (1905-11) on the north side of Savai´i. The interior mountains are covered with dense rainforest, tall trees, large ferns and vines. Down on the shores and the lower slopes it is instead the second growth woodland, coconut and pandanus and other coastal trees that dominate. On the tidal mudflats there are areas of mangroves. Early settlers brought plants such as bananas and breadfruit. They also brought root vegetables as taro, taamu, yams, tapioca and sweet potatoes (http://www.govt.ws/gi_listing.cfm 15-05-06). Village areas and crop growing cover the major parts of the coastal regions. Closest to the villages grow crops for daily food and higher up the mountains grow more rhizome crops and crops for cash. The areas uphill are also used for keeping of cattle (Lindgren, Lodin and Schönfeldt, 1997:10). The large animals such as

(10)

cattle and horse have been taken to the islands in later days, but early immigrants brought the pigs, dogs and chickens today running around everywhere. Otherwise there are few large animals and the only indigenous mammal is the flying fox, today classed as endangered specie. Only animals like birds, insects and fish have found their own way to Samoa (Tamua, 2000:1ff).

In the year 1722 the first European, Roggeveen, observed the Samoan island of Manu´s, but he did not land on it. The French explorer Bougainville sailed past the islands in 1768. The first real European contact with Samoa or the Navigators Islands (dubbed by Bougainville) came 19 years later when the French under La Pérouse landed on Tutuila (American Samoa). This contact ended with 13 dead sailors and the Western contact after this experience was limited over the succeeding decades. More significant European contact began in 1830 with the missionary John Williams. The first permanent station was built in 1836. After this event the Western influence increased and a large change took place in the Samoan socio-cultural system. In 1899 the Samoan archipelago was divided in two by Germany (Western Samoa) and United States (American Samoa). New Zealand took over the administrative control from Germany in 1914 and in 1962 Western Samoa became an independent nation (Clark, J. T, 1996:447).

The Independent state of Samoa is today a democracy with a British based parliament system. The parliamentary, the fale fono, has its own essence revised to accommodate local customs and Christian principles. The Constitution provides for a Head of State, a Prime Minister and a Cabinet of Ministers who comprise the administrative council and a legislative assembly. There are twelve Ministers of cabinet currently led by the Prime Minister Tuilaepa Lupesoliai Sailele Malielegaoi. Today’s Head of state (since the independence) for the rest of his life is His Highness Malietoa Tanumafili II. In the future the Head of State will be elected by the Legislative Assambly and hold office for five year terms. Parliamentary elections are held every five years and the number of seats in the fale fono is today 49. These seats (47) is only available for matai (chiefs) and the other two seats are held by elected representatives if ethic minorities who are citizens of Samoa but not members of any family (aiga). Today (since 1991) everyone with the age of 21 years are eligible to vote (http://www.govt.ws/gi_listing.cfm 15-05-06). During recent election (March 2006) the HRP party (The Human Rights Party) kept most of their seats.

(11)

Samoa is one of the least developed countries in the world, according to the UN criteria. The country has a limited economic resource but hardly anyone starves and very few are homeless because of their customs. Cash crop agriculture employs two thirds of the labour force and contributes to 90 percent of the export (coconut cream, coconut oil and copra). The tourism sector has risen to 25 percent of GDP in the latest years. Fishing is a contributing sector but there are problems with the decline fish stocks in the area. Important for the economy is also aid and remittance from Samoans living overseas (World Fact Book, 15-05-06).

(12)

2. RELATIONS TO A LOCAL PLANTATION AND TO NATIONAL

PREHISTORIC REMAINS

2.1 No history no law?

There is no established archaeological profession in Samoa, and foreign archaeologists have only made a few real attempts to archaeological work. There is neither a legislation about cultural remains, but in 2005 a policy about heritage was formulated by the cabinet. The policy provides a framework for the conservation, preservation, use, allocation and sustainable management of the resources the heritage constitutes. The policy recognises that:

“cultural and national heritage is among the priceless and irreplaceable possessions not only of each nation but also of mankind as a whole.” (Ministry of National Resources, 2005:2).

This policy is in my opinion a pointing in the right direction showing that the Samoan Government now recognised the importance of protecting the cultural heritage, but the policy is not a law and is therefore not being obliged to be re-enforced. The theory is not a practise and the way towards such a goal is through education. During my time at the Letolo plantation and the village of Satupaitea I got the idea to make a smaller questionnaire. My intention was to look at the villagers relations to the plantation today. My first weeks experience was that the knowledge in the English language was not that good. Therefor the daughter in the priest family I lived with helped me working out a formula in the Samoan language. One morning before we started the days work I gave out my small formula with a few questions about the plantation to the workers. It took them 1,5 hour to answer them, the two chiefs was a little bit faster, but I understood that the education in their own language was not that good either. I heard later that many of the children in the village just made one or two years in school. The chiefs and the later questioned children of the priest family have better education, which also was visible at the answers. This was especially shown in the answers on the last question; What is history for you? where the chiefs and the priest daughters answered with traditional oral histories, but one of the workers at Letolo plantation answered quite directly on this issue; I have no history! Or maybe he does not know about it. The education about the country’s history in Samoan schools today, if you have the economy to go there, is almost only about the colonisation. With such a background it is not difficult to understand the answer: I have no history!

The cultural heritage of Samoa is their “day by day living”, to live the Samoan way (Fa´a Samoa) and many Samoans proclaim that the cultural heritage of Samoa is the living history

(13)

and the story of the ancestors. Moa Nord discuss in her essay: Linking Local and Universal

Values –Managing and protecting cultural heritage in Samoa (2006) which is based in

interviews with Samoans, the importance to understand and incorporate the Samoans and their cultural heritage. With “no history” and “no law” it is difficult to protect the prehistoric remains in Samoa. My hope is that the archaeology work embracing the local people together with a development of archaeological education at NUS could create a base for an interest about the Samoan material Prehistory and knowledge about the importance of its protection in the building of a strong Samoan nation. The cultural heritage, and the conservation of it, are resources, which strengthen the knowledge and contribute to the understanding of the Samoan identity as a people. As Tuiolo Schuster in an interview (Nord, 2006) suggests; “…But I guess

at this point, the first step that we are taking is actually making people aware.”

2.2 Samoan Prehistory –a review

Archaeology on today’s Independent State of Samoa (earlier Western Samoa) has been made in different epochs. Roger Green and Janet Davidson made the first surveys and excavations on the islands in the 60´s. The two publications from their work (1969 and 1974) are today the foundation of Samoan archaeology. A more detailed investigation of the settlement pattern were made by Jesse Jennings and Richard Holmer in the 70´s and published in two volumes 1976 and 1980. Jennings and Holmer implemented among other things a detailed survey of the Letolo plantation.

The first evidence of human settlement in the Samoan archipelago dates 2850 years before present. The period today called: ”ancestral Polynesian culture” dates from 2500-1800 BP. About 1000 years ago the mound-building tradition started and interactions in the archipelago can be detected.

2.2.1 The period of decorated Lapita Ceramics

The Lapita people with there characteristic decorated pottery are described by Patrik Vinton Kirch (2000:90). With a wide selection of tools, and described as a social organisation with fundamental social units in economic competition with each other is the Lapita people, according to Spriggs (1999:111f), a culture complex with South-East Asian origin.

The Lapita pottery found in Samoa dates 2900-2700 BP. So far referring to Green 2002 only one site has been located near the coast, namely at Mulifanua on the island of ´Upolu. Lapita sites on Samoa are today difficult to find because of the reconstructed shoreline represent a mean sea level for the Lapita time that is 2,6-3 metres lower than today and is therefore

(14)

difficult to detect. However; at today’s knowledge it seems as the colonisation of Savai´i and ´Upolu were quite limited in the time for the Lapita colonisation.

2.2.2 The period of Polynesian Plainware Ceramics

This period dates very roughly from 2700-2600 BP to around 1500 BP. There are several sites from this period in the Samoan archipelago. Referring to Green 2002 appears Plainware pottery often under later mound constructions or buried by alluvium and has never been found during surface surveys. Plainware sites may, like the Lapita sites, lie under today’s sea level, but finds of pottery at inland sites on Uplou shows that some human occupation had moved to the inland at that time (Green, 2002:136f). The excavation at the Pulemelei mound in 2003 showed that the makers of the plainware pottery also settled inland on Savai´i (Martinsson-Wallin, 2003). Green agrees with Davidson (1974b:161) that almost the whole coast were colonised at about AD 200. Green also suggests that the occupation of the inland started during the plainware era and that the inland settlements are an extended coastal community rather than a settlement entirely located to the inland (Green, 2002:138). According to Green 2002 and Kirch 2000 were the Plainware people organized in ancestral line social groups and led by a senior(ranked) male with both the secular and the ritual functions, and into a lower order social group designed the Proto-Polynesian kainga (senior male).

2.2.3 The dark ages

As in European history, Samoa also has a time called the “dark ages”, with limited evidence visible in the settlement patterns. Green and Davidson (1974) document how meagre the archaeological evidences are in the period from about AD 400-500 and AD1000. Remains found during the dark ages mainly consist of earth ovens and some agricultural traces. Clark (1996:451) suggest, opposite to the model provided by Green and Davidson (1974), that pottery was widely spread and used in Samoa throughout the first millennium AD. Clark discusses that the use of pottery became uncommon after AD 1000 and around 1300-1600 it became really rare. After AD 1600 the pottery was abandoned. This theory can however be questioned, since the pottery may be secondary on the sites (Clark, 1996:451, Green, 2002:139f).

If this gap is a thing created of the early site recording and excavation sampling strategies or from constructed sequences on these islands is difficult to say. Fact is that this is a period when much of the Samoan landscape became occupied but it is still a period that we today know very little about (Green, 2002:140).

(15)

2.2.4 The settlement pattern before European contact (1000-200 years ago)

During 800 years, beginning at c. AD 1000, the Samoan settlement pattern took the shape (as observed by the archaeologists) that is visible on the landscape today. The majority of the archaeological records derive from this period. The basic settlement structure continues to be the Household unit and visible on the landscape are large settlements containing platforms, walkways, walls, ovens and piles (Green, 2002:140). A more detailed description of this pattern will be described later.

2.3 The Letolo Plantation

Located between the two villages of Satupaitea and Vailoa and situated on the southeast part of the island Savai´i a coconut (copra) plantation known as Letolo is situated (see map p.8). The plantation is owned by the family company Nelson Inc. The plantation covers an area that extends from the coast and four kilometres inland. The area is bordered to the west by the Seugagogo river and to the east by the Faleata river. The land slopes to the south from an elevation of 250 metres in the northern part of the plantation. Situated on this plantation are over 3000 features of human manufacture which makes the Letolo plantation of great interest for studies concerning the Samoan prehistoric settlement pattern. These features consist of stone fences, walkways, foundation platforms, earth ovens and refuse piles with a natural “organic” like development. In the middle of this pattern and as the heart of an organic creature stands Pulemelei one of the largest prehistoric monuments in the Pacific. This great stone mound has a height of c.12 metres and a base measurement of 60x65 metres. Located at the plantation and falling properly from the Faleata River is Olemoe, one of the most visited waterfalls in Samoa.

In the 1960´s the area around Pulemelei was described by Scott (1969:77ff). This was also the time for the first detailed drawings and mapping of the area. No more archaeological investigations were done at that time.

During more than one year, from May 1977 until November 1978, G. U. Jackmond did a survey of the (in that time open) cultural landscape at the Letolo plantation. Interested in archaeology, he worked as a volunteer for the Peace Corps organisation (Jennings, Holmer and Jackmond, 1982:85, Jackmond 1977-.78). This survey resulted in a very good map of the settlement pattern at the plantation with a short description to every mapped and numbered feature (report). Since the first archaeological excavation at Pulemelei 2002 has archaeological investigations (2003 and 2004) in the area resulted in a better knowledge of the

(16)

Samoan prehistory, social organisation and the settlement pattern (Martinsson-Wallin, Wallin and Clark, 2003, Martinsson-Wallin, 2004, Brødholt and Vuijsters, 2004, Wallin and Martinsson-Wallin, nd)

A comparative study of the settlement pattern at the Letolo plantation and the organisation of a modern village on Savaií were made by Jennings, Holmer and Jackmond (1982). The aim of the mapping of the prehistoric settlement was to determine if earlier observations of archaeological patterns was valid also on other settlements. To see if the social and behavioural interpretation of that pattern was valid Jennings et al. also did a study of the modern village of Faáala on Savaií. The main question they wanted to find the answer to was concerning changes that appeared in the settlement patterns during the abandonment of inland villages around 1750-1800 (Davidson, 1969). There has been a shift in location of the villages between 1500-1800 AD. The interesting thing in this, thought Jennings et al., was to see whether there were any comparable changes within the villages in the organisation of social space, and distance on household relationships (Jennings, Holmer and Jackmond, 1982:81).

Using the Survey of G. U. Jackmond, Paul Wallin and Helene Martinsson-Wallin worked out a correspondence analysis of the different mounds at the plantation. Compared to surrounding features and using of the five Wards (described later) in the settlement, discussed by Jennings et al. (1982:87ff), Wallin and Martinsson-Wallin made circles with the central places of the Wards in the middle and a diameter of 400 metres. The aim was to see if it was possible to see differences in location and relation between for example different types of mounds; oval and squared, small and large, inland and coast. The correspondence analysis showed that Ward A and B (located more to the coast and the lowlands) associates more with round and oval platforms, but also small low platforms. Also the umuti (big round oven probably for the preparation of the ceremonial ti-plant; Cordyline fructicosa L) of the small type is associated with Ward A and B. The Wards situated more inland (C and E) showed higher association with squared and rectangular platforms, large and high platforms and the large type of umuti. It was also possible to see a higher association to roads in the Wards C and E. This opens for the possibility that the inland was the place for ritual and ceremonial activities (not really discussed earlier in the west part of Polynesia), and with this also the association with larger squared and rectangular platforms (which is the shape of the ceremonial places; Marae, in eastern Polynesia) (Wallin and Martinsson-Wallin nd).

(17)

2.3.1 In the jungle

During the German colonial time c.1870, the Letolo plantation was owned by the Germans. The plantation was later sold to the Swedish merchant August Nelson married to the daughter of leading chief from Safune, Savai´i (Davidson, 1983:60). With hundreds of workers the Letolo Plantation was in historic days a big economic source for the near villages. Today the larger part of the huge coconut plantation is overgrown by secondary jungle and the employees are few. Now and then some tourists find the way to the plantation and for the small amount of five tala (c. 15 SEK) they take a swim in the waterfall pool and maybe, if the interest is there, take the hike up to Pulemelei.

The company still has a few cattle on the plantation, but besides that no work is carried out on the plantation. Today, Letolo is only considered a place to pick some wood, river stones and coconuts. Recent archaeological excavations, which was initiated 2002, have attracted new attention to this place. Many villagers hope that the excavations can increase to a big tourist attraction and for that reason give more income to the villages in the closest area. One interesting thing is that only few of the villagers have in fact visited the Pulemelei mound and one of the workers on the plantation even called it “cursed”.

In Samoan economy the productivity of the forest are related to the regeneration power of the native forest. The Christian missionaries met a spiritual Samoan Cosmology; Aitu ´spirits, spirit-gods or deities, which appears in large variations and “incarnated” in animals, but also in material like stones or trees. According to historical sources a large amount of forest, trees, plants and animals could be seen as “sacred”. The fact that sacred is to be untouchable gives us an idea of a spiritual controlled productivity of recourses in the Samoan social organisation (Olson, 1997:10, 15, 18f).

When the Christians came with the new religion they could see the sacredness of the forest. Wesleyan missionary Dyson mention that the Samoans did not touch the forest, not even years after that all the inhabitants have been nominally Christians. This large change of religion in the Samoan nation got consequences like “temples were destroyed, the sacred

groves left to be overrun by the bush, the shells and stones and divining cups were thrown away…” (Turner 1861, p.b243)

There are not many people that know about the over 3000 other stone constructions and features situated under the vegetation of the Letolo plantation. These remains of an abandoned

(18)

prehistoric settlement with one of the largest prehistoric stone structures in Polynesia are the foundation of this thesis. Maybe the possibilities to find the relevant questions are to be found under the guava trees and green ferns of the Letolo plantation. So why not ask the questions about changes in the Samoan settlement pattern?

2.3.2 Archaeological evidence for changes in the settlement pattern at the Letolo plantation

Davidson and Green suggest that mound constructions may not have begun until around 900 BP, but other features as house platforms and earth fortifications can have a longer chronology (Davidson, 1974:227). Jennings et al. agree with Green that the dating of the very large mounds is to the 16th and 17th centuries. On the fact that the platforms on the Letolo plantation are larger then the ones at Mt. Olo (which is dated to the early 17th century) Jennings et al.dates the Letolo settlement to the 17th century. This may be the time for the abandonment of the Letolo settlement, but the 2002, 2003 and 2004 years investigations at the Letolo plantation show us a long time sequence of the settlement. The investigations at the Pulemelei mound shows a place inhabited 2000 years ago and the time for the building of Pulemelei and constructions like it seems to be around AD1100-1300. There is furthermore evidence that Pulemelei was rebuilt or changed in the years around AD1400 to 1600 (Martinsson-Wallin, Wallin and Clark, 2003, Martinsson-Wallin, 2003, 2004).

An interesting thing with the study by Jennings et al., is the refinement that none of the three archaeologically examined villages: Mt. Olo, Sapapali´i and Letolo, are isolated inland villages. All of them have connections with the coast even if the settlement extends more than 3 km inland. This originating of the large villages on the coast does not change the fact of inland abandonment. Interesting is the different perspective on the nature of the abandonment that the villages can give us. It seems as it has been a gradual removal rather than a total abandonment of the villages, but Letolo was entirely abandoned (Jennings, Holmer and Jackmond,1982:100).

These indications support the idea that the coastal areas have been used since the primary settlers of the island. When the population was growing the villages moved further inland. With the remarkable reduction of population, often discussed as a result of European diseases, the remaining villages moved to coastal areas again. This was also the area where the trade with the Europeans took place (Jennings, Holmer and Jackmond, 1982:100).

(19)

2.4 House Hold Units and the Ward idea

In the early Samoan archaeology all settlement patterns were called villages. This is according to Jennings et al. (1982), just an adoption of the Samoan practice where all, small or large, settlements are called villages. Jennings et al., recognised a new terminology in the Samoan settlement pattern, when they were inspecting maps from old surveys and from investigations made by themselves they saw units of use that they called Household units (HHU) (see picture below) and larger units called Wards. The earliest representation of the Household units has been found in the Polynesian Plainware site at Sasoa´a. (Davidson, 1974c:232, Green, 2002:138f).

Jennings et al., definition of a HHU is that it; “consisting of an area more than 75% enclosed

by walls and paths containing one large or two small stone platforms with one stone-free area within the enclosure”. (1982:82). This HHU discussed as single family residences and the

place for domestic activities have the Samoan name fuai i ala (measurement along a path). Many of these HHUs are also, certainly, bounded by a path. In the pattern of HHUs it was possible to see, and generate by the statistics, clusters of HHUs. Jennings, et al., called these clusters residential Wards. Ethno-historical descriptions suggests, that the high chief or sub chief having larger house platforms and a larger domestic use area (Jennings, Holmer and Jackmond 1982:82, 84).

The Wards were, in the study of Jennings et al. (1982) detected visually and confirmed statistically. These Residential Wards are, according to Holmer 1974:100, just a part of the larger village; Nu´u. The Nu´u is not, as mentioned before, a village in today’s meaning. It is Typical Samoan Household unit,

(20)

instead a large tract of land extending from the coast to inland and may contain a number of different isolated settlements all belonging to the same “village”. Recently, Clark mentions, that nuú taken for a nucleated village, consists of a number of smaller kin groups; áiga (family) (1996:451). Clark suggests that it is possible that the nuú in the past referred to a geopolitical unit which territory reached from the coast to the inland and “…whose population could be either nucleated or widely dispersed.” (p. 451).

Tautala Asaua proclaim for the use of a Samoan terminology in her thesis: Samoan

Archaeoratory: Bridging the Gap between the Spoken and the Scientific, 2005. She discuss

the Samoan terms from Ethnohistorical sources where the whole Samoan archipelago first are divided in eleven districts or itū, which means “side”. The island of Savaií have six itū and Satupaitea is one of them. These itū´s can then be further subdivided into the earlier mentioned Nuú. Later on the Nuú can be divided into Pitonuú and Fuiala. From ethnographic records it is indicated that the chief or matai was ruler over a village Pitonuú and Nuú (Jennings, et al. 1982). Pitonuú could be further divided at the fuiala level; which has been defined (Pratt, 1893) as one division of a village. Each pitonuú is separated by

fuiala. By introducing the terms Household unit and wards (Jennings, et al. 1982) they

appeared comparable with the Samoan Nu´u, Pitonu´u and Fuiala (Green 2002:138).

Because of the reason that the terminology worked out by Jennings et al., Wards and House hold units, are the one used by archaeologist until present day, and they are the terms that I am going to use in my thesis.

(21)

3. THE ANALYSE

The family or aiga is the central part of traditional Samoan culture, so instead of studying individual features in a larger settlement pattern the correspondence analysis is based on variables such as: the enclosures or family Household unit, high status units, their enclosure status, size and the including or connecting to features (Wards, platforms, walkways, walls, pits and ovens). From this “break down” of the settlement pattern study the relations between the variables and out of that make way for new ideas and discussion about the Letolo settlement.

3.1 Relational Archaeology and the Correspondence analysis

Everything has relations and the classical Butterfly effect, (Edward Lorenz, 1979) is all about relations,- the idea in meteorology that the flapping of a butterfly's wing will create a disorder that in the chaotic motion of the atmosphere will become enlarged. The flapping of the butterfly could eventually change the large-scale atmospheric motion, so that the long-term behaviour becomes impossible to calculate (www.dn.se 17-05-2006). The world will always feel the flapping of the prehistoric butterfly.

3.1.1 Relations on the archaeology path of theory

After the step into the “New archaeology” and the “wake up” (Colin Renfrew) of archaeology a living debate started, about theory and how it could be used in archaeological work. This debate made it possible and “legitimate” to ask new questions with more variation to the material. With the New Archaeology followed the ethno-archaeology and the study of modern material cultures (Olsen, 2003:49).

As a reaction against the Processual Archaeology, and the using of methods as statistics to prove hypotheses, the Post-Processual Archaeology developed in the 80´s. The generalizations of the statistical methods were questioned, as well as the “objectivity” of interpretations about the past and the archaeological material remains (Olsen, 2003:50ff).

The consequences of the Post-Processual Archaeology ideas might, as Wallin and Martinsson-Wallin (nd.) discuss, have reduced the using of statistical method and also supported relativistic thoughts. The past is reduced to the present, but with the Processual era started a larger acknowledge about context and relations. The “open” Post-Processual archaeology may have neglected the opportunity for contextual analyses on

(22)

complex archaeological interactions that can possibly be indicated through relational statistics.

In all times the search for relations, similarity, and differences in material remains have been central in the science of Archaeology. Relations can be observed both in the small local context, as well as in the larger regional cohesion. I agree with Wallin and Martinsson-Wallin (nd.) that the statistics can be used in a relational kind of archaeology. It is however difficult to find pertinent patterns in individual or few observations, and it is therefore of immense importance that the material is adequate enough for a statistical research. In a large material it is possible to see patterns and the individuals can be sorted out and analysed as “different” and get relevance from the “normal”. Through statistical and relational analyses on the archaeological material it is possible to see the general and the peculiar in a settlement pattern and place it in a relation with the general and the peculiar from another settlement pattern.

3.1.2 French Relations with Bourdieu and Benzécri

Underlying the theory of a relational archaeology is the French relational sociology and in front line Pierre Bourdieu. Bourdieu suggested that the essential characteristics of an element depend on the elements position among the other elements. More precise, writes Broady (1991:462); the elements position in the larger system of relations between elements.

Ethnographic records, often used by the Post-Processual archaeologists as a complementary method is very useable as a multidisciplinary approach, Bourdieu remarks (1991:477). Bourdieu maintain that the ethnographic records are not enough for a study and understanding of more complex and changeable societies, like the chiefdom mentioned as the social organisation of the Polynesian islands in prehistory. Because of the large area that the Pacific Ocean with the islands constitutes it is impossible to talk about one general chiefdom society since there are great differences between a chiefdom society in the east of Polynesia and a similar one in the west part. Marshall Sahlins classified the Polynesian societies according to their grade of hierarchy or stratification in his classic study Social Stratification in Polynesia

(1958). Sahlins classification was built on productivity, but in contrast to Sahlins Irving

Goldman (1970) suggested that the chiefly competition or rivalry in status is the agent of the differentiation seen on chiefdom societies. Goldman classified the Polynesian Chiefdoms in three types: Traditional, Open and Stratified. These ideas was applied for example in the Scandinavia archaeology (Burenhult, 1999, Harding, 2000, Renfrew & Bahn, 2001) and used for studies about the organisation of the Bronze Age society, something I discussed in my

(23)

undergraduate thesis Holy Garbage in Scandinavian Bronze Age –Fire cracked Stones from a

Polynesian Perspective. In this work I was studying ethno-historical Polynesian chiefdoms

and theories about them. Sahlins’ and Goldman’s works are primarily based on ethnographic records from the European contact period and their ideas might be considered somewhat out of mode today, but their work still have a great impact. Today´s knowledge mainly discussed by P. V. Kirch indicates a larger difference in the Polynesian chiefdoms springing from the basis of an Ancestral Polynesian culture (2000:249f).

I agree with Broady (1991:477) proclaiming that the only line to attack multivariate complex contexts is through statistical methods. Bourdieu have shown that humans are not simply passive actors as some would argue, and he therefore showed with different ethnographic situations that these “actors” had their own ideas about rules and situations. Bourdieu argued for a theory of how these individual social “actors” practised the living, reproducing and transformation of the culture that surrounds them. We need, Bourdieu argues, to study rules that not just are followed, but these that are creatively manipulated by its social “actors” (Johnson, 1999:105).

The development of the correspondence analysis as a method started with the Frenchman Juan-Paul Benzécri during the 60´s and 70´s (Broady, 1991:473). The difference between the French and Anglo-American Philosophy at that time is probably the reason why the correspondence analyse was not extensively used. The Anglo-American Philosophies dominated the archaeology in the 60´s and 70´s and there is a vast difference between the two philosophies concerning the use of statistics. The French school sees the statistics as a hypothesis producer. The result of the analyses became, according to the French Philosophy, a basis for new discussions (Wallin and Martinsson-Wallin, nd.). Benzécri with his strict ideas about induction principles suggested that the systems of relation should appear as a fruit grown from the analysis. Bourdieu has suggested that a structure of a field contains its prehistory. The structure of a studied field reflects the events until the present day, the tradition is showed in the expression (Broady, 1991:465).

3.2 The Letolo Wards, Household units and enclosures in Theory

Jennings et al., interprets the basic village structural units at the Letolo settlement; the Wards, as centres for local authority and the family Household units (HHU). The settlement consists of almost four kilometres of walkways. It seems as two parallel primary walkway system is going through the upper part of the surveyed area (inside the two oblong rectangular figures

(24)

above). Jennings, et al., discuss these walkways as represents of the principal access roads in and out from the different Wards. According to Jennings et al., there are approximately around 300 HHUs in the surveyed area, but only 50 are clearly defined. Like the HHUs at Sapapali´i (Jennings et al.1982:86) some of them were more difficult to identify because since some enclosures did not really enclose the stone platforms. There are 1059 platforms constructed of basalt stone and boulders recorded in the area, which vary in size and quality of construction. The largest is Pulemelei and none of the other larger platforms approaches Pulemelei in size. The platforms are rectangular to oval in shape except of two small star mounds near the Seugagogo river. Also 41 raised rimmed ovens and plenty of stone piles were recorded (Jennings et al.,1982:87-89, Jackmond, 1977-78).

Letolo settlement (Green, 2002:145)

The suggested Wards; darker areas (Jennings, et al., 1982:90)

(25)

Based on platform size and on the ethno-historical literature and observations at Mt. Olo (Upolu, Samoa), Jennings et al., suggests that the larger platforms indicate central points in clusters of local authority and high status. At the Letolo plantation five of these clusters are visible (right picture, p.23). Certainly, a few other large platforms occur in the area, but these appear to have a random distribution. Like the settlement at Mt. Olo all of the suggested clusters of larger platforms at the Letolo plantation fall along the primary walkways. A Ward is, based on archaeological observations, consisting of 50-75 fence-enclosed HHU´s. Associated with the five clusters on Letolo are also, as at the Mt. Olo settlement, earth ovens (umuti). However, a great difference does occur between the both settlements. At the Letolo plantation many of the huge ovens or umuti also occur away from the “high status” clusters (Jennings, Holmer and Jackmond, 1982:89ff).

3.3 Variables

I am now going to do a general description of the variables used in this study. For a more detailed catalogue about the used variables see appendix 8.2 and 8.3.

3.3.1 Wards

The study is based on three of the five suggested Wards or high status clusters at the Letolo plantation (Jennings, et al., 1982) situated from the coast to inland named A, B and D. I have used the suggested area limitations put forward by Wallin and Martinsson-Wallins (nd.) with the central places of the Wards in the middle of an area with a diameter of 400 metres. I have subsequently used the maps of Jackmond (1977-78) to identify enclosures appearing inside these circles (see pictures following).

Ward A, reworked after Jackmond 1977-78

(26)

During the survey of the Mt. Olo settlement (see appendix 8.1) Holmer mentions that the more densely populated Wards have more clearly defined HHUs. This is probably a result of a larger need to differentiate land. This observation is of great use in the discussion about the Letolo settlement pattern. At the Mt. Olo settlement the more densely populated Wards have also a more consistent distribution of platforms (Holmer, 1974:99).

3.3.2 Household units and Enclosures

This study include 133 enclosures or Household units (HHU) and as discussed earlier, only 50 of the 300 Household units at the Letolo plantation are clearly defined, according to Jennings et al., (1982:82). As the HHUs at Sapapali´i (see appendix 8.1), Holmer, (1974:150) and Jennings, et al. (1982:86), some of them are more difficult to identify because some enclosures does not enclose stone platforms.

Interesting is also that in the modern village of Fa´aala (see appendix 8.1) the HHUs are not always delineated by a physical border. A few stones, banana plants or fence posts, can instead mark the boundaries. It seems as, the families know where their land ends and where the neighbours land begin. The HHUs located outside the village has a proper surrounding of walls made of stone. Local villagers informed Jennings, et al. that the fenced HHUs are outside of the “pig wall” and therefor needs walls to keep out the free-roaming pigs from their gardens. This could also be a product from the contact with the missionaries who felt it improperly to live with the pigs (Jennings, Holmer and Jackmond, 1982:95).

Ward D, reworked after Jackmond 1977-78

Key

(27)

According to these records I have made an attempt to define every possible enclosure as a HHU, irrespective of size, or if the enclosure include features. HHUs no:1-31 are situated in Ward A, HHUs no:32-77 in Ward B, and HHUs no:78-133 in Ward D. Every enclosure or/and HHU that is in contact with the fictive Ward circle are included in my study (see pictures p.24 and 25) The HHUs are divided in three enclosure variables; closed, party closed (75 percent or more) and open (75 percent or less). The HHUs are also divided in three groups according to area size; small (300-1200 square metres), medium (1201-4000 square metres) and large (4001-8400 square metres).

3.3.3 Walls and Fences

There are 64,6 kilometres of recorded stone fences or/and walls at the Letolo plantation. These contained in 1982 of stone ruins shaped as linear mounds that separate the area in roughly rectangular household units. The fences/wall ruins are today even more collapsed, and have a great variation in width and height. At the Sapapali´i settlement they have, according to Jackmond and Holmer (1974:147), a large variation in construction. Everything from quickly built, unstable (often just one stone thick); one-meter high walls to really solid walls were recorded. Based on these facts the walls or/and fences at the Letolo plantation are not used in this study as a variable to definite the HHU. They are only used as markers for the enclosures.

Important to have in mind are the possibilities that stone fences at the Letolo, just as suggested for the Sapapali´i settlement, could have been used for subdividing land for use as pig pens and agricultural plots (Jackmond and Holmer, 1974:148). To have in mind is also the use of walls and fences during modern disturbance.

3.3.4 Walkways and accesses

3,76 kilometres of walkways are recorded from the Letolo settlement, and have a kind of “skeletal” structure. Two “primary” walkways can be seen which traverse the length of the settlement pattern (see picture p.23). These “primary” walkways, discussed by Jennings et al., as representing principal accesses routes to the Wards, are used together with other walled or/and elevated walkways in my study as variables for access. Two different access variables are used depending on the HHUs contact with a walled or/and elevated walkway. The pathway or auala are significant in the Polynesian culture and can, according to Tamasese 2004:2, be associated with the funeral ritual with same name; the pathway to lagi (heaven) or

(28)

3.3.5 Platforms, shape and size

Stone platforms are remains after foundations from houses or meeting places often scattered with pebbles or river stones (ili ili). According to the traditions the dwelling house or “meeting” platforms have rounded corners. This oval shape can therefore, suggests Wallin and Martinsson-Wallin, nd., be interpreted as the social arena. The round house, fono, were (and are today) the place for the chiefs important socio-political meetings. From this point of view it is interesting with the meaning of rectangular and squared platforms in the settlement pattern. Wallin and Martinsson-Wallin, nd., discuss the possibility that these rectangular and squared platforms are remains after the ceremonial or/and religious activities. To compare with the shape of the ceremonial marae places in Polynesia which also are squared (Wallin 2001). The shape of the platform does however not necessarily indicate the shape of the house placed on top of it. The platforms inside the enclosures/HHU in this study are divided into variables according to shape, height and size. Two main platform shapes are used; round/oblong and rectangular/squared. According to these shapes the platforms are divided in fictive numbers with the largest one, the shape disregarded, “named” (1). I have used this fictive variable like: the first (largest) platform in the HHU is round/oblong and has the height (1) or (2). One variable is also the numbers of platforms inside the HHU; (1), (2), (3) or (4 and more). According to that the platforms are also counted in the two main shapes described earlier; (1), (2) or (3 and more) round/oblong or rectangular/squared platforms inside the HHU.

According to the variation in platform sizes at the plantation these are divided in three groups: small (up to 700 square metres), medium (701-1300 square metres) and large (1301 square metres and larger). The two variables in height are: Height 1 (0,1-0,5 metres) and height 2 (0,6 metres or more). The height is, according to the problem with the sloping ground, calculated after the medium measurement.

To take into account during the discussion is the study of the modern village of Fa´aala made by Jennings, et al. (1982:94). showing that the total of 314 structures had an additional of 92 abandoned platforms. These platforms are today decayed or removed and their function are today unknown.

The dwellings of the chiefs in the modern village is associated with other structures, such as guest and meeting houses, today ranked as “high status” features (Jennings, Holmer and Jackmond, 1982:94). Of interest is the observation, made by Holmer (1974:96) and by myself

(29)

during the journey 2006, that many of the kitchen houses constructed today are without a stone platform.

3.3.6 Large earth ovens or umuti

The ti-plant, Lau ti, (Cordyline fructicosa) native to tropical Asia, but today widely spread throughout the Pacific, is probably an ancient introduction to Samoa. Ti has, since ancient times, been a central plant in the Samoan culture. The leaves are good for making skirts, dance costumes, roofing thatch, medicine and food wrappers. Of great importance, and interest for the study of the prehistoric settlements, are the large, sugar laden, tuberous roots that were baked in underground ovens (Whistler, 1996:73f). The ovens, often in presence with the high status clusters, are sometimes suggested as used in ritual or luxury purposes (Holmer, 1974:102). Another possibility, also suggested by Holmer (1974:102), is that these ovens could have several uses, for example, the preparation of large amounts of food for a big feast.

The earth ovens or umuti are in this study distinguished by two variables according to size when they appear inside a HHU. Size 1 (small to medium sized under 12 metres in diameter) and Size 2 (large sized over 12 metres in diameter).

Jennings et al. (1982:92) discuss the fact that big ovens or umuti not always occur in association with the “high status” clusters on the Letolo plantation. Therefore it is not possible to use the ovens alone as indicators for “high status” areas (Jennings, Holmer and Jackmond, 1982:92).

3.3.7 Stone piles

Stone piles are to be found everywhere at the Letolo settlement area and are, because of there different designs, in this study only used as one variable; if there is a stone pile/or piles in the HHU.

3.4 Result

Appendix 8.2 shows the explanations and symbols for the variables used in the correspondence analysis. All used data are showed in tables appendix 8.3.

(30)

Graph 1; All used variables without the platforms (with names)

The correspondence analysis shown in Graph 1 (above) all variables used, except the platforms, to compare with Graph 3 (p.30) where all the Household units are showed. The internal relation between the HHUs and the variables used gives us their internal dependence which is expressed in the graphs.

(31)

In Graph 2 (above) the platform variables are showed together with all the variables used. It is possible to see two groups; all rectangular/squared platform variables located inside the dark circle to the down left in the graph. In the white circle and located more to the middle right are the round/oblong platform variables.

In Graph 3 the Household units are showed. Except from HHU 001 are all the HHUs belonging to Ward A (circles) located to the upper half of the graph. The HHUs belonging to Ward B and D appear most around the middle line of the graph, but is not visible separated in any way. The HHUs that considerably diverge from the norm in the graphic are sorted out as outliers by the program. The outliers are located inside the dark circles showed in Graph 3. Outliers in Ward A are 001, 026 and 030, in Ward B 056, 063 and 076, and in Ward D 092, 093, 096, 103, 113, 114, 133.

Graph 3; All Household units (in circles the by WINBASP elected outlier HHUs)

3.4.1 Outliers

To look for the explanation to the outliers found by the correspondence analysis, I will describe them in detail below. These HHUs can be seen as in some way exceptional or different:

(32)

HHU 001: Partly closed and large sized HHU in Ward A with more than four platforms. The HHU includes both round/oblong (2) and rectangular/squared (3 or more) platforms. The sizes of the platforms are, according to the variables, 1 both in size and height, except from the first rectangular/squared platform, that has the size 2.

HHU 026, 030: Open medium sized HHUs in Ward A without further variables.

HHU 056: Partly closed medium sized HHU in Ward B with two platforms. The HHU includes a large umuti and two rectangular/squared platforms with size and height value 1.

HHU 063: A closed large sized HHU in Ward B with four or more platforms. The HHU also includes one or more pile/piles, access to a walled walkway, 3 or more round/oblong platforms with variable 1 in both size and height, 2 rectangular/squared platforms with variable 1 in both size and height, except from the first rectangular/squared platform, which has the size 2.

HHU 076: Partly closed large sized HHU in Ward B with more than four platforms. The HHU includes access to a walled walkway, three or more round/oblong platforms with variable 1 both in size and height, except from the second round/oblong platform, which has the height 2, and three or more rectangular/squared platforms with variable 1 in both size and height, except from the first one, which has the height 2.

HHU 092: A closed medium sized HHU in Ward D with three platforms. The HHU includes access to a walled walkway, one round/oblong platform with variable 1 in both size and height, and two rectangular/squared platforms with variable 1 in both size and height, except from the first one, which has the size 2.

HHU 093: Partly closed large sized HHU in Ward D with four or more platforms. The HHU includes one or more pile/piles, one round/oblong platform with variable 1 in both size and height, and three or more rectangular/oblong platforms with variable 1 in both size and height.

HHU 096: Partly closed medium sized HHU in Ward D with two squared platforms with variable 1 in both size and height.

(33)

HHU 103: A closed medium sized HHU in Ward D with two squared platforms with variable 1 in both size and height, except from the first platform, which have the size 2. The HHU also includes access to a walled walkway.

HHU 113: Partly closed large sized HHU in Ward D with four or more platforms. The HHU includes a small umuti, one or more pile/piles, one round/oblong platform with variable 1 in both size and height, and three or more rectangular/squared platforms with variable 1 in both size and height

HHU 114: Partly closed medium sized HHU in Ward D with four or more platforms. The HHU includes one or more pile/piles, one round/oblong platform sized 1 and the height variable 2, and three or more rectangular/squared platforms with variable 1 in both size and height, except the second rectangular/squared platform, which has the height variable 2.

HHU 133: A closed large sized HHU in Ward D with four or more platforms. The HHU includes access to a walled walkway, one round/oblong platform sized 1 and with height variable 2, and three or more rectangular/squared platforms with variable 1 in both size and height, except the second rectangular/squared platform, which has the height variable 2.

3.4.2 Inland vs Coast

The result of the correspondence analysis shows differences in variables appearing in the coastal versus inland located Household units. Graphic 4 shows the variables positions versus their general location in the landscape, according to the location of the Wards. The variables down left in graphic 4 are more common inland and the variables upper right are more common near the coast. The size of the HHU appears to increase towards the coast, which is shown in light bordered rectangular in Graph 4. From left to right in the graphic the size of the Household units increase (light bordered rectangular). The numbers of platforms inside the HHUs also increase from the left to the right, except from the variable HHU No4 (four or more platforms), which falls down in the graphic (light bordered ovals).

(34)

Graph 4; All variables with names

Ward B and D appear to have more “access” through walkways than Ward A (dark bordered diamond in Graph 4). Piles seem to be common in the whole area and the variable falls near the norm. The type of enclosure goes from more closed to the left and more opened to the upper right. Large umuti appears to be more common inland.

(35)

Graph 5; All variables and HHUs (The HHUs involved in 2006 years excavation marked)

The HHUs located central in the suggested Wards are in A:018, B:055 and D:103 (see picture p.24 and 25). Just one of them (103) falls in the number of outliers. HHU 055 is situated near the norm and HHU 018 locates almost opposite HHU 103 (see Graph 5, above).

Let the remains ask the questions : In search for prehistoric relations on a Samoan settlement pattern through a correspondence analysis